Advanced in Geosciences, V2: Solar Terrestrial(St) (2006)(en)(368s)

A d v a n c e s i n Geosciences Volume 2: Solar Terrestrial (ST) This page intentionally left blank A d v a n c e...

Author: Wing-huen Ip

26 downloads 634 Views 10MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

A d v a n c e s

i n

Geosciences Volume 2: Solar Terrestrial (ST)

This page intentionally left blank

A d v a n c e s

i n

Geosciences Volume 2: Solar Terrestrial (ST)

Editor-in-Chief

Wing-Huen Ip

National Central University, Taiwan

Volume Editor-in-Chief

Marc Duldig

Department of the Environment and Heritage, Australian Antarctic Division, Australia

World Scientiﬁc NEW JERSEY

•

LONDON

•

SINGAPORE

•

BEIJING

•

SHANGHAI

•

HONG KONG

•

TA I P E I

•

CHENNAI

Published by World Scientific Publishing Co. Pte. Ltd. 5 Toh Tuck Link, Singapore 596224 USA office: 27 Warren Street, Suite 401-402, Hackensack, NJ 07601 UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE

British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library.

ADVANCES IN GEOSCIENCES A 5-Volume Set Volume 2: Solar Terrestrial (ST) Copyright © 2006 by World Scientific Publishing Co. Pte. Ltd. All rights reserved. This book, or parts thereof, may not be reproduced in any form or by any means, electronic or mechanical, including photocopying, recording or any information storage and retrieval system now known or to be invented, without written permission from the Publisher.

For photocopying of material in this volume, please pay a copying fee through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to photocopy is not required from the publisher.

ISBN 981-256-456-X (Set) ISBN 981-256-984-7 (Vol. 2)

Typeset by Stallion Press Email: [email protected]

Printed in Singapore.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 fm

EDITORS

Editor-in-Chief:

Wing-Huen Ip

Volume 1: Solid Earth (SE) Editor-in-Chief: Yuntai Chen Editor: Zhong-Liang Wu Volume 2: Solar Terrestrial (ST) Editor-in-Chief: Marc Duldig Editors: P. K. Manoharan Andrew W. Yau Q.-G. Zong Volume 3: Planetary Science (PS) Editor-in-Chief: Anil Bhardwaj Editors: Francois Leblanc Yasumasa Kasaba Paul Hartogh Ingrid Mann Volume 4: Hydrological Science (HS) Editor-in-Chief: Namsik Park Editors: Eiichi Nakakita Chulsang Yoo R. B. Singh Volume 5: Oceans and Atmospheres (OA) Editor-in-Chief: Hyo Choi Editor: Milton S. Speer

v

This page intentionally left blank

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 fm

REVIEWERS

The Editors of Volume 2 would like to acknowledge the following referees who have helped review the papers published in this volume: Sobhana Alex Ashok Ambastha Nanan Balan Thanasis Boudouridis Tom Chang Paul Charbonneau Leroy Cogger Ian Craig T. E. Cravens Marc Duldig Clive Dyer Peter Dyson Brian Fraser Wally Friedberg Ted Fritz Claus Frohlich Suiyan Fu Tim Fuller-Rowell Americo Gonzalez-Esparza Nat Gopalswamy Bogdan Hnat David Holdsworth John Humble Emi Ito H. Gordon James Andrew Klekociuk P. Venkata Krishnan Hing-Lan Lam

Manfred P. Leubner Brent Lewis Chris Lewis Xinlin Li Tony Lui P. K. Manoharan Yoshi Miyoshi Isao Morishita David Moss Keran O’Brien Andrew Parﬁtt Zuyin Pu Ian Richardson S. K. Satheesh Brigitte Schmieder Karel Schrijver Gordon Shepherd Jiankui Shi Zdenka Smith Willie Soon R. Sridharan Sunny W. T. Tam Masayuki Ugai Ilya Usoskin Ioannis Vogiatzis Yongli Wang Andrew W. Yau Q.-G. Zong

vii

This page intentionally left blank

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 fm

CONTENTS

Editors

v

Reviewers

vii

The Solar Cycle at the Maunder Minimum Epoch Hiroko Miyahara, Dmitry Sokoloﬀ and Ilya G. Usoskin

1

Solar Coronal Plumes: Theoretical Concepts and Results Manfred Cuntz

21

Preﬂare Features in Microwaves and in Hard X-Rays Ayumi Asai, Hiroshi Nakajima, Masumi Shimojo and Stephen M. White

33

Nonextensive Entropy Approach to Space Plasma Fluctuations and Turbulence M. P. Leubner, Z. V¨ or¨ os and W. Baumjohann Simulation of Interplanetary Shock Wave Caused by CME on August 25, 2001 Tomoya Ogawa, Mitsue Den, Takashi Tanaka and Kazuyuki Yamashita Observation of the Inﬂuence of the January 15–17 Solar Storms to the Magnetic Field and Ionosphere of Indonesia Clara Y. Yatini, Jiyo and Mamat Ruhimat Observational Study of Solar Magnetic Active Phenomena by Huairou Vector Magnetograph Hongqi Zhang

ix

43

65

73

83

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 fm

x

Contents

Development of Storm-Time Proton Total Energy Based on Multiobservation of NOAA Satellites Keiko T. Asai, Tsutomu Nagatsuma, Tomoaki Hori and Yoshizumi Miyoshi On the Cross-Field Diﬀusion of Galactic Cosmic Rays into an ICME K. Munakata, S. Yasue, C. Kato, J. Kota, M. Tokumaru, M. Kojima, A. A. Darwish, T. Kuwabara and J. W. Bieber On the Upper Limiting Energy of the Solar Diurnal Anisotropy of Galactic Cosmic Ray Intensity K. Munakata, S. Yasue, C. Kato, S. Akahane, M. Koyama, S. Mori, A. A. Darwish, H. Tsuchiya, H. Onuma, K. Mizutani, T. Yuda, M. Takita and J. Kota Sector Boundary Crossings and Geomagnetic Activities Shinichi Watari and Takashi Watanabe

105

115

125

135

Energetic Particle Composition Signatures in the Earth Magnetotail S. Y. Fu and Q.-G. Zong

143

The High Latitude Boundaries Under Extreme Solar Wind Conditions: A Cluster Perspective H. Zhang, T. A. Fritz, Q.-G. Zong and P. W. Daly

163

The Magnetospheric Cusp: Structure and Dynamics Q.-G. Zong, T. A. Fritz, H. Zhang, S. Y. Fu, X. Z. Zhou, M. L. Goldstein, P. W. Daly, H. Reme, A. Balogh and A. N. Fazakerley Initial Results from the Simulation of the Halloween 2003 Storms Ramon E. Lopez, Salvador Hernandez, Michael Wiltberger, John Lyon and Charles Goodrich

173

191

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 fm

Contents

Round-the-Clock Ground-Based Imaging Spectroscopy of Space Weather Eﬀects on the Thermosphere and Ionosphere Supriya Chakrabarti and D. Pallamraju Auroral Equatorward Boundary Observed by the NOAA-17 Satellite L. Xie, T. A. Fritz, Q.-G. Zong, Z. Y. Pu, X. Z. Zhou and X. L. Li Aurora-Associated Phenomena and the ePOP Mission Ludmila M. Kagan On the Importance of the Cross-Body Approach in Planetary Aeronomy Marina Galand, Anil Bhardwaj and Supriya Chakrabarti Solar Terrestrial and Planetary Science Missions in Asia–Oceania: Opportunities for Collaborative Research Andrew W. Yau, Anil Bhardwaj, Iver H. Cairns, C. Z. Cheng, Wing H. Ip, Yasumasa Kasaba, Kyoung W. Min, Masato Nakamura and Yoshifumi Saito

xi

201

219

229

239

249

Incoherent Scatter Radar in Ionospheric Research: Past Contribution and Future Promise T. Hagfors

265

Vertical Geomagnetic Cutoﬀ Rigidities for Epoch 2000 — Deviations from Expected Latitude Curves D. F. Smart and M. A. Shea

277

Ultra Long Range Aircraft Operations and Space Weather I. L. Getley and M. L. Duldig Measurements and PCAIRE for Monitoring the Cosmic Radiation Exposure of Aircrew L. G. I. Bennett, B. J. Lewis, A. R. Green, A. Butler, M. Takada and I. L. Getley

287

303

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 fm

xii

Contents

Measurement and Modeling of High Latitude Flights in the Southern Hemisphere I. L. Getley, A. R. Green, L. G. I. Bennett, B. J. Lewis and M. L. Duldig

311

Link Between Cosmic Rays and Clouds on Diﬀerent Time Scales Ilya G. Usoskin and Gennady A. Kovaltsov

321

The Eﬀect of Solar Activity on Annual Precipitation in Delingha Region, Tibetan Plateau for the Last 1000 Years Lei Huang, Xuemei Shao, Hongbin Liu, Eryuan Liang and Lily Wang

333

Climate and Extreme Events in Central-Southern Region of Eastern China During 1620–1720 Jingyun Zheng, Quansheng Ge, Xiuqi Fang and Zhimin Man

341

Eﬀects of Typhoon on the Ionosphere Yi-Mou Liu, Jing-Song Wang and Yu-Cheng Suo Formation and Observations of Shadow Bands During the Total Solar Eclipse of November 23, 2003 Near Maitri, Antarctic Hari Om Vats, S. P. Bagare and S. M. Bhandari

351

361

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch01

THE SOLAR CYCLE AT THE MAUNDER MINIMUM EPOCH HIROKO MIYAHARA∗ , DMITRY SOKOLOFF† and ILYA G. USOSKIN‡ ∗Solar-Terrestrial Environmental Laboratory, Nagoya University Nagoya, Aichi 464-8601, Japan †Department of Physics, Moscow State University, Moscow 117588, Russia ‡Sodankyl¨ a Geophysical Observatory (Oulu unit), University of Oulu P. O. Box 3000, FIN-90014, Finland

Here, we present a brief review of the current status of the Maunder minimum study. The Maunder minimum is considered as an example of occasionally occurring Grand minima, when the solar dynamo was in a special mode. We review available sets of direct and indirect data covering the period during and around the Maunder minimum. The start of the minimum was very abrupt and was followed by a gradual recovery of the activity. The data suggest that while the sunspot activity was greatly suppressed during the deep phase of the minimum, the cyclic dynamo kept working around the sunspot formation threshold level, leading to seemingly sporadic occurrence of sunspots. The majority of proxy data depict the dominant 22-year periodicity during the Maunder minimum with the sub-dominant 11-year cycle. The length of the cycles was probably slightly enhanced. We also discuss theoretical models and speculations concerning the solar dynamo as well as the heliosphere during the Maunder minimum. Comparison with other minima (Sp¨ orer and Dalton) suggests that these features are common.

1. Introduction Solar magnetic activity usually exhibits cyclic behavior with about 11-year period, which is also subject to great secular variations. While the contemporary level of the activity is high, the normal cyclic activity is sometimes interrupted by periods of unusually low activity known as Grand minima. The most recent Grand minimum is the so-called Maunder minimum (MM) that took place between 1645 and 1715 (the deep phase in 1645–1700). The idea of “a prolonged sunspot minimum” was suggested already in the 19th century.1 –3 The initial concept of a Grand minima was introduced based on rather nonsystematic and indirect evidence and observations (see a brilliant review of the history of the Maunder minimum discovery and underlying physical and astronomical ideas4 ). However, it took almost a century before this idea became widely accepted5 –7 after careful consideration of 1

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch01

2

H. Miyahara, D. Sokoloﬀ and I. G. Usoskin

other related solar proxy data, including cosmogenic 14 C. Although the very existence of the MM is beyond doubt now, it is much less clear what exactly had happened during the MM and whether the MM is similar to other Grand minima. The existence of such long (several decades) periods of suppressed solar activity is very important for understanding dynamo processes in the Sun and other stars. A particular important question is whether the physical mechanism responsible for the solar activity cycle was still operating during the MM, and if yes — in what mode? Other Grand minima of solar activity are known (e.g., those called after Sp¨ orer, Wolf, etc.) using cosmogenic isotope data, but the MM is very special since it is the most recent one and amazingly well covered by direct and indirect data. Since the 1980s, from archival data on instrumental solar observations performed by the French astronomical school in (17–18)th centuries the solar activity have been restored.8,9 Together with precisely measured cosmogenic isotope data this allows for a systematic quantitative study of solar activity during the MM. Here we aim to review the recent observational facts and theoretical speculations related to the MM. In the present paper, we ﬁrst consider various observational data and evidence (Sec. 2) and present theoretical models to deal with the solar dynamo and heliosphere during a Grand minimum (Sec. 3). Then, we compare the available data for other known Grand minima with the pattern of the MM (Sec. 4). Conclusions are summarized in Sec. 5.

2. Observational Data Here, we consider the bulk of direct and indirect data available for the MM. These data form proxies for diﬀerent solar and heliospheric parameters.10 In particular, sunspot number is a proxy for the toroidal magnetic ﬁeld in the convection zone, aurorae provide information about transient interplanetary phenomena and local heliospheric state, and data on cosmogenic isotopes 14 C and 10 Be are governed by the global heliospheric modulation of cosmic rays. 2.1. Solar observations By initial deﬁnition, the MM is a peculiar epoch in solar activity that took place at the time of Louis XIV, during which sunspots were almost absent. After a systematic investigation of archive data by Ribes and Nesme-Ribes9

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch01

The Solar Cycle at the Maunder Minimum Epoch

3

we know that, during the decade 1660–1670, the French astronomer Picard supported by the interest of Louis XIV observed the Sun as often as it was possible at that time and ﬁxed the results in a way reasonably comparable with modern standards of scientiﬁc presentation. The result is that only one sunspot was observed during that period. The corresponding conclusion is twofold. On one hand, the level of solar activity was extremely weak in comparison with periods after and just before the MM. On the other hand, the high reliability of the data implies that the only observed sunspot manifests some level of solar activity during the most deep phase of the MM. The point is that the sunspot formation is associated with a substantial toroidal magnetic ﬁeld located somewhere deep inside the Sun. A maintenance of such ﬁeld needs some excitation mechanism which is usually identiﬁed with solar dynamo which in turn is usually cyclic. At present, we know that sunspots were observed during 368 days within the deep MM (1645–1700),11 which is less than 2% of all days during the MM (see Fig. 1). Note that about 95% of days during the deep MM were covered by reported solar observations. A careful analysis and conservative consideration of the available data12 show that, despite uncertainties in the data, the level of sunspot activity was indeed extremely low during the MM: yearly sunspot numbers are below 4 for the deep MM, and below 8 for the cycle 1700–1712. While the sunspots were very scarce in the ﬁrst half of the MM, the recovery of solar activity after 1670 became more and more noticeable in sunspot data and a clearly distinguishable regular solar cycle took place in 1698–1712. Because of the scarcity of sunspot occurrence, standard time series analysis methods cannot be applied13 to study the question whether the sunspot activity was still cyclic during the MM. By means of a special method for analysis of the occurrence of sparse events, Usoskin et al.14 ,15 studied time clustering of sunspot occurrence11 (see Fig. 1). This ﬁgure shows clustering of sunspot occurrence as a function of the scale (see Ref. 14 for details). One can see that there are two major clusters, around 1660 and 1680, which are persistent through all scales. Together with the neighboring sunspot cycle maxima in 1640 and 1705, it implies that the dominant 22-year periodicity in sunspot activity was present during the MM. Some sub-clustering is visible at small scales. It is important that sunspot formation is a threshold eﬀect and the absence of spots during the major part of the MM does not necessarily imply switch-oﬀ of the dynamo process. The solar cycle probably existed but at a sub-threshold level and was only barely visible in sunspot data.16

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch01

4

H. Miyahara, D. Sokoloﬀ and I. G. Usoskin

100

1650

1660

1670

1680

1690

1700

1650

1660

1670

1680

1690

1700

80

No

60 40 20

Fig. 1. Sunspot occurrence during the deep MM. Upper panel depicts individual days (vertical bars) when sunspots were recorded.11 Lower panel shows the concentration of sunspot occurrence (colors) as a function of the scale (vertical axis), after Ref. 15.

Another interesting fact is that it is possible to build butterﬂy diagrams for some periods of the MM, thanks to the archives of the French astronomy school. The butterﬂy diagram was strongly asymmetric during the MM with the majority of spots observed in the Southern solar hemisphere.9 This fact is very important for understanding the Grand minimum scenario as discussed in Sec. 3.1. Basing on archive observations of Mutton, Richerd, Picard and La Hire, Ribes et al.8 presented a time series for the apparent solar diameter for the MM period. A wavelet analysis of these data17,18 shows that an approximately 11-year cyclicity exists in the apparent solar diameter data. Note however, that the apparent solar diameter observations are much more uncertain than sunspot data, and the level of cyclic variations is very low. Accordingly, the question of the cyclic variations of the solar diameter during the MM is still controversial in general.19 The data available suggest that solar magnetic cyclic activity might have been in a special mode during the MM, which is diﬀerent from the modern activity. 2.2. Cosmogenic isotopes Cosmogenic isotopes provide the most extendable indirect data on the cosmic ray ﬂux, the state of the heliosphere, and hence on the solar magnetic

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch01

The Solar Cycle at the Maunder Minimum Epoch

5

activity during the past. The most commonly used cosmogenic isotopes are radiocarbon (i.e., 14 C) and 10 Be, which are measured in tree-rings and in ice cores, respectively. Both tree-rings and ice cores form stratiﬁed structures and retain the time variations of the abundance of isotopes in each layer. Cosmogenic isotopes are produced in the Earth’s atmosphere mainly by galactic cosmic rays, which originate from outside of the heliosphere and are modulated by the solar wind and interplanetary magnetic ﬁeld. The basics of the modulation process are well understood (see, e.g., Refs. 20 and 21), and the attenuation level of cosmic rays in the heliosphere depends on the strength and level of turbulence of solar magnetic ﬁeld and on the global structure of the heliosphere. Basically, the ﬂux of cosmic rays impinging on the Earth is inversely correlated with the solar activity, but shows also variations depending on, for example, the polarity of solar magnetic ﬁeld. These variations may become important during the Grand minimum. Generally, cosmogenic isotopes can serve as an index of the poloidal magnetic ﬁeld which is carried out by the solar wind and ﬁlls the heliosphere. This is fully applicable for the 11-year cycle averaged data,22 while inside the solar cycle, an important role is played also by the tilted heliospheric current sheet and drift processes. 14 C and 10 Be are produced in the atmosphere as a result of nuclear reactions of cosmic rays with the atmospheric nuclei. Then 14 C is oxidized to form carbon dioxide and circulates within the carbon cycle between diﬀerent reservoirs, some of which are very inertial, and it gets eventually absorbed by trees by means of photosynthesis. On the other hand, 10 Be becomes attached to aerosols, precipitates with snowfall and is accumulated in the ice in polar regions. These diﬀerences in the transportation system decide the advantages and disadvantages of each isotope, and sometimes cause different behaviors in the time series of the two isotopes. In this way, using both 14 C and 10 Be, we can trace the history of cosmic rays, the heliosphere and the Sun. Since cosmogenic isotopes are measured contemporarily, their data series are homogeneous, i.e., with nearly constant quality and resolution throughout the recorded period, contrary to direct observables (sunspots, aurorae) which were quite unevenly observed/recorded in early times. Due to distinguishable layers in tree-rings and ice cores, high-temporal resolution can be achieved (one year in tree rings and 2–3 year in ice cores). Here we review two 14 C records and one 10 Be record from around the time of the MM. It was ﬁrst shown by Stuiver23 that variations of the 14 C content are inversely correlated with the level of solar activity. The ﬁrst systematically measured annual 14 C data for the epoch since 1510 were

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch01

6

H. Miyahara, D. Sokoloﬀ and I. G. Usoskin

reported by his group in 199324 and greatly improved in 1998.25 A record of 10 Be in polar ice as a tool for investigating the solar activity was made by Beer et al.26,27 The 10 Be record obtained from the Dye-3 site27 provides annual data since 1424 AD and shows signiﬁcant 11-year cyclic variation, persisting through the MM.28 On the other hand, 14 C data exhibit quite a different cyclic behavior during the MM. Peristykh and Damon29 reported the disappearance of the 11-year cyclic variation during the MM. They divided the 14 C record24 into three periods with 90-year interval: before (1540–1630), during (1630–1720), and after (1715–1805) the MM, and analyzed them using the Maximum Entropy method. A signiﬁcant approximately 11-year signal was found before and after the MM, but only signals of 24.2, 15.6, and 6.22 years were detected during the MM with the lack of the 11-year cycle. Recently, a new independent record of precisely measured biennial 14 C data from a Japanese cedar tree was obtained for the MM by Miyahara et al.30 –32 Using these two independent 14 C records (Fig. 2), a cross spectrum of the frequency analysis was obtained (Fig. 3) aiming to remove regional climate eﬀect and measurement systematic errors of 14 C data.30 In this cross spectrum, two main periods of 13–15 and 24–29 years were detected, and it was suggested that the 11-year cycle and consequentially the 22-year cycle were lengthened by a few years. Figure 4 shows wavelet spectra of the two 14 C records from 1617 to 1745 AD. The spectra were obtained using the S-transform.33 The spectra consistently show the period of about 14 years from 1660 to 1715 AD. The quasi-periodic signal of about 26 years, which probably corresponds to the lengthened 22-year cycle, is seen through the period in both spectra. However, the 11-year signal is quite weak from 1640 to 1660 AD, suggesting strong suppression or discontinuity of solar cyclic activity. A lengthening of the 11-year cycle is also visible in 14 C data around the Dalton minimum (Fig. 5), while other data sets (sunspot, aurora,22 and 10 Be34 ) yield controversial results for that period.

2.3. Magnetospheric phenomena Magnetospheric disturbances can be observed by means of aurorae (polar lights), which are caused by transient heliospheric phenomena such as coronal mass ejections or high-speed solar wind streams. Therefore, easily observable aurorae, records of which can be found in archives, form a proxy

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch01

The Solar Cycle at the Maunder Minimum Epoch

7

Fig. 2. Variation of 14 C content in tree-rings around the time of the MM. The solid line shows the biennial record from Refs. 30 and 32 and the dotted line shows the annual record from Ref. 25.

Fig. 3. Cross spectrum of the S-transform of the 14 C records shown in Fig. 4. Horizontal lines correspond to the periods 11, 13, and 15 years, respectively. The solid line denotes the boundary eﬀects.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2

8

ch01

H. Miyahara, D. Sokoloﬀ and I. G. Usoskin

Fig. 4. Spectra of the S-transform of the 14 C records for the MM. Left and right panels correspond to the 14 C series by Miyahara et al.30–32 and by Stuiver et al.25 Horizontal lines correspond to the periods 11, 13, and 15 years, respectively. The solid lines denote boundary eﬀects.

Fig. 5. Spectrum of the S-transform of the ∆14 C record25 for the period from 1650 to 1850 AD. Horizontal lines correspond to the periods 11, 13, and 15 years, respectively. The solid lines denotes the boundary eﬀects.

for major local interplanetary disturbances and, hence, for the solar magnetic activity over both the solar cycle and longer time intervals (see, e.g., Refs. 15 and 35). While the 11-year cycle is dominant in the auroral series during normal solar activity times, it was suppressed during the MM.35

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch01

The Solar Cycle at the Maunder Minimum Epoch

9

Fig. 6. Auroral activity at mid-latitudes according to Refs. 36 (solid “KP88” line) and 37 (dotted “S92” line). All data are ﬁve-year smoothed.

Instead, signiﬁcant peaks with longer periods of approximately 19–20, 25, and 15 years appear in the power spectrum during and around the MM.35 Figure 6 presents two ﬁve-year smoothed series of auroral activity observed in central Europe: combined series36 from sites with latitude below 55◦ N (mostly Czech and German sites) and solely German observations.37 Aurorae were very rare (0–4 aurorae/year) during the MM, in accordance with the suppressed solar activity.4,5 However, it is important to note that some aurorae were still observed at mid-latitudes, implying that the solar activity did not vanish completely even during the deep phase of the MM. Periods of increased auroral activity agree fairly well with sunspot occurrence during the MM (see Sec. 2.1). This implies a dominant 22-year cycle in auroral activity during the MM in phase with the similar pattern in sunspot activity. A small increase in the auroral series occured in 1695, corresponding to the isolated sunspot group and denoting a subdominant 11-year cyclicity at the end of the MM. Accordingly, the time behavior of auroral occurrence in central Europe, and, hence, of the major transient irregularities in the inner heliosphere, is in good agreement with sunspot activity during the MM. 2.4. A grand minima scenario Let us summarize the main features of the MM according to the available data. The MM is considered as an example of Grand minima when the intensity of solar activity cycle diminishes drastically and a speciﬁc state of solar activity occurs. A Grand minimum epoch is substantially longer than a solar activity cycle, and therefore cannot be related to as an

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch01

10

H. Miyahara, D. Sokoloﬀ and I. G. Usoskin

unusual cycle. Transition from normal activity to the MM was very abrupt compared to the typical time scale of the solar cycle. On the other hand, the recovery of activity at the end of the minimum was gradual and took several decades.38 Because of the gradual recovery, the total duration of a minimum is not well-deﬁned (Eddy5 deﬁned the MM duration as 1645– 1715). Roughly we can deﬁne the deep phase as 1645–1700, when sunspots occurred seemingly sporadically without an apparent cyclic behavior, while the whole minimum was extended until ca. 1712, including the very tiny but regular solar cycle 1700–1712. Cyclic activity did not completely disappear even during the deep minimum but was reduced to a level that is sub-threshold for sunspot formation. The solar cycle is clearly visible at the end of the MM, see cycle in 1700–1712. Because of the scarcity of the sunspot occurrence, traditional methods of spectral analysis (e.g., wavelet analysis) applied to a particular data set (e.g., sunspot data) can be unable to reveal the periodicities in the deep phase of the MM. On the other hand, special methods of data processing demonstrate that a cyclic behavior was present during the whole MM. Both nominal 11- and 22-year cycles can be found inside Grand minima with a possible lengthening of the periods. A substantial North–South asymmetry appears typical for the Grand minima events. During the MM, sunspots appeared predominantly in the Southern hemisphere with the apparent lack of spots in the Northern hemisphere. The 22-year cycle was notable in the sunspot data, implying that the magnetic Hale cycle kept working during the MM, with the phase locked. It is important that, despite the seemingly sporadic occurrence of sunspots, the regular dynamo kept working although in a special mode. Heliospheric parameters (solar wind, interplanetary magnetic ﬁeld, and the size of the heliosphere) were reduced during the MM. Suppressed auroral activity implies that heliospheric transient phenomena were quite rare. This conclusion is consistent with results available for the Sp¨ orer minimum. It looks plausible to suggest that, apart from typical Grand minima, weaker and shorter suppressions of the solar activity sometimes take place. These might include, for example, the Dalton minimum and a minor event that occurred at the end of 19th century. 3. Theoretical Speculations Here we would like to note that the interpretation of the above data can hardly lead to a unique result. We try to avoid pushing forward a speciﬁc interpretation favoring our own scientiﬁc preferences but rather to present a survey of the relevant theoretical ideas and speculations.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch01

The Solar Cycle at the Maunder Minimum Epoch

11

3.1. Solar dynamo Let us brieﬂy remind the basic scheme of solar dynamo. The solar activity cycle is presented as a manifestation of a dynamo wave propagating inside the Sun. In ﬁrst approximation, the dynamo wave corresponds to the concentration of the toroidal magnetic ﬁeld propagating from middle latitudes towards the solar equator. As was noted by Larmor as early as in 1919, the only realistic way to excite such magnetic ﬁeld in the framework of Maxwell equations is associated with the Faraday induction eﬀect. However, according to the Lenz rule, toroidal magnetic ﬁeld BT cannot be excited without poloidal magnetic ﬁeld BP . A particular scheme of solar dynamo suggests a physical mechanism connecting toroidal and poloidal magnetic ﬁelds. An obvious way to obtain toroidal magnetic ﬁeld from poloidal one is the solar diﬀerential rotation. It is, however, much more diﬃcult to obtain BP from BT . Parker39 suggested that this can be done by means of cyclonic motions in the solar convective zone. A joint action of Coriolis force and density gradients results in an excess of right-hand vortices in one hemisphere and left-hand vortices in the other hemisphere. In turn, a component of the mean magnetic ﬁeld B parallel to the mean electric current J appears due to this excess. A consistent theory of this eﬀect was developed in 1960s by Krause and R¨ adler40 who used the notation α for the proportionality coeﬃcient between B and J. This eﬀect is known now as the α-eﬀect. This scheme results in self-excitation of a dynamo wave similar to that one known from observations. The toroidal magnetic ﬁelds in Northern and Southern solar hemispheres usually have opposite polarities. This toroidal magnetic ﬁeld conﬁguration is referred to by theoreticians as dipolar. The Maxwell equations admit however another conﬁguration with the toroidal magnetic ﬁeld of the same polarity in both hemispheres, which is called quadrupolar conﬁguration. In practice, phases of the dynamo waves propagating through Northern and Southern hemisphere can be shifted in respect to each other. This displacement can be presented as an admixture of the quadrupole conﬁguration with the dipole.41 The toroidal magnetic ﬁeld is hidden inside the solar convective zone and is inaccessible for direct observation. Fortunately, the toroidal magnetic ﬁeld can be traced by sunspots. On one hand, the sunspots are not an inevitable component of solar dynamo. One can imagine a star with a dynamo wave propagating somewhere inside the convective zone where due to some reason the sunspot production is strongly suppressed. It would

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch01

12

H. Miyahara, D. Sokoloﬀ and I. G. Usoskin

be very diﬃcult to recognize the existence of toroidal magnetic ﬁeld on such a star. In contrast, poloidal magnetic ﬁeld is present on the solar surface directly. On the other hand, the toroidal magnetic ﬁeld inside the Sun known via sunspot data is much more intense than the relatively weak poloidal magnetic ﬁeld. The most spectacular data concerning solar and stellar activity cycles are indirect and represent the toroidal magnetic ﬁeld behavior. Direct data related to the poloidal magnetic ﬁeld behavior are more obscure. As a matter of fact, comparisons between dynamo models and observations are based mainly on sunspot data. Cosmogenic isotope data are particularly important because they reﬂect properties of the poloidal magnetic ﬁeld, i.e., they are complementary to the sunspot data. The importance of Grand minima events for the solar dynamo theory is twofold. On one hand, the very existence of Grand minima is a challenge for the theory. For instance, Grand minima may be simulated in some numerical models (see, e.g., Refs. 42–45) of the solar dynamo but the theoreticians yet cannot straightforwardly explain why the Grand minima occur. It is very important to select those features of the Grand minima phenomenology which are relevant for a confrontation with solar dynamo models. On the other hand, studying the Grand minima allows understanding of how the solar dynamo machine works in unusual regimes. In principle, one could suppose that the occurrence of a Grand minimum can be related to a suppression of sunspot formation without changing the dynamo mechanism itself. This possibility is unfavorable for dynamo interpretation and can be declined because of the fact that the magnetic activity recovery was strongly asymmetric at the end of the MM46 (see Sec. 2.1). This argument is however not completely decisive because of the threshold nature of sunspot formation, which could amplify a small random North– South asymmetry of the toroidal magnetic ﬁeld to a seemingly asymmetric butterﬂy diagrams. The pattern followed from cosmogenic isotope data and auroral records during the MM (Sec. 2.2) rejects this interpretation on a more solid way. This indicates that not only sunspot formation but also the global solar/interplanetary magnetic ﬁeld was reduced during the MM. As a result, we conclude that Grand minima are associated with some disturbances in the solar dynamo machine, although this machine keeps working during a Grand minimum. The cyclic component of the activity during the MM is most pronounced in the cosmogenic isotope data while sunspot and aurora data provide only indicative support. Let us summarize, according to the spectral analysis of isotope data, the main features which are crucial for the dynamo interpretation. The nominal 11-year signal is

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch01

The Solar Cycle at the Maunder Minimum Epoch

13

intermittent around the Maunder minimum in the 14 C records, sometimes it is not visible or is discontinuous. A signal for the nominal 22-year cycle is weak but regular in 14 C data. This may imply that during the MM something happened with the structure of poloidal ﬁeld oscillations. Simultaneously, the conﬁguration of toroidal magnetic ﬁeld was strongly asymmetric with respect to the solar equator. The scenario of the MM (generally valid also for the Dalton minimum38 ) suggests that the dynamo is most suppressed in the beginning of the minimum, followed by a gradual recovery. Note that such a behavior is consistent with a stochastically forced return map model of the dynamo.47 We note that some events in the solar activity history may be interpreted as not fully successful attempts of the solar dynamo machine to switch into the Grand minimum state. For example, data on 14 C favor such an interpretation for the Dalton minimum. Unfortunately, sunspot data are quite incomplete for that period and not fully instructive: the quantity of the data is much worse than for the MM. Although a suppression of the cyclic activity is clearly visible, the interpretation of the event is not straightforward. In particular, a butterﬂy diagram cannot be constructed for the Dalton minimum, making it impossible to study the North–South asymmetry. The Dalton minimum was shorter than the MM but similar to it in its overall structure (sharp decrease followed by a gradual recovery22,48 ). A tiny suppression of the activity (called sometimes the Modern minimum) around 1890 was even less pronounced and shorter than the Dalton minimum. The whole bulk of cosmogenic isotope data provides, however, additional support to the idea that the solar dynamo machine tries from time to time to go into the Grand minimum state, and that only a fraction of such attempts is successful.49 In addition to the above, an event occurred circa 1633 AD, for which observations by Gassendi imply asymmetric butterﬂy diagrams, provides another example of this kind.19

3.2. The heliosphere and cosmic rays Because of the very low level of solar magnetic activity, the heliosphere (the region totally controlled by the solar wind and magnetic ﬁeld) is expected to be quite diﬀerent during the MM from its present state. It has been shown that while the modulation of cosmic rays was reduced during the MM, cosmic rays were still modulated,15,50,51 which together with the fact that some magnetic activity was still observed during the MM, implies that the heliosphere did exist during that time, but probably its size was reduced.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch01

14

H. Miyahara, D. Sokoloﬀ and I. G. Usoskin

Some estimates of the heliospheric parameters have been performed based on the available data sets discussed above. It is supposed52 –54 that the solar wind was signiﬁcantly slower during the MM, 200–350 km/s, compared to the presently measured 400–800 km/s. The interplanetary magnetic ﬁeld (actually its Bz component)54 and the axial dipole strength55 were also estimated to be essentially lower (by a factor 4–7) than presently. Applying a heliospheric model of cosmic ray transport to the measured 10 Be in polar ice, Scherer et al.56 ,57 have shown that the diﬀusion coeﬃcient of cosmic rays in the heliosphere should be increased during the MM, which implies decreased level of the interplanetary magnetic ﬁeld and/or interplanetary turbulence. However, these numbers were obtained using regression or other models based on sunspot numbers and ﬁtted to modern conditions and, therefore, can be considered only as rough estimates. It is not straightforward to interpret the observational facts discussed in the previous section in terms of the heliosphere. For instance, the dominating 22-year periodicity visible in 14 C data during the MM can be understood in two ways. First, as suggested in Ref. 15, it may be due to a natural 22-year periodicity in the global solar magnetic parameters (solar wind, poloidal ﬁeld, etc.). On the other hand, the modulation of cosmic rays can be drift-dominated contrary to the diﬀusion dominated modulation during normal solar activity times. Being only a relatively minor factor nowadays, polarity-dependent drifts58,59 may become dominant during the MM when diﬀusion/convection modulation is suppressed, so that the regular 22-year change of the global magnetic ﬁeld polarity (assuming it was maintained throughout the MM) may alone lead to a basic 22-year cycle in low energy cosmic rays. A direct modeling of this eﬀect forms a challenging task since the usual heliospheric models, which work well for the recent activity, cannot be applied to a Grand minimum. For example, a heliospheric model corresponding to the recent minima of solar cycle can not adequately reproduce the MM conditions because of some major distinctions. Within a solar cycle, there is a strong correlation between the level of sunspot activity and tilt of the heliospheric neutral sheet. However, this correlation does not correspond to a real physical link, and suppressed sunspot formation does not imply a ﬂat neutral sheet during a Grand minimum. If the global magnetic ﬁeld reversal keeps working during the MM, the tilt is expected to vary within its full extent.20 Also, a residual modulation beyond the termination shock may become important during a Grand minimum.57,60

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch01

The Solar Cycle at the Maunder Minimum Epoch

15

Resolving this problem is a challenging task, and a proper modeling of the heliospheric conditions during a Grand minimum is required to address this issue. Such work is in progress.

3.3. Other minima Occurrence of a Grand minima is a rare but not unique phenomenon. The most recent Grand minima, other than the MM, are Sp¨ orer (around 1500 AD), Wolf (around 1300 AD) and the tiny Oort (around 1050 AD), but they appear more or less regular through millennia.61 However, only few of them can be studied in some detail, in addition to the MM. The Sp¨ orer minimum (1415–1535 AD) was not covered by direct sunspot observations, but annual cosmogenic isotope records allow some conclusions about cyclic activity to be drawn. The data on 10 Be in polar ice (Dye-3, Ref. 27) depict the dominant cyclicity with 20–25-year periodicity (see Fig. 7) with some power in the 5–10-year interval during the minimum.51 Radiocarbon 14 C data31 imply periodicities at about 22 years (continuous), and 7–11 years (intermittent) during the Sp¨ orer minimum. It is important that the 11-year signal is not observed during 1460–1500, corresponding probably to the deep phase of the minimum. Therefore, both isotope series show the dominance of the 22-year cycle on the background of a greatly suppressed 11-year cycle during the Sp¨orer minimum. This pattern is very similar to the MM, supporting the idea that Grand minima correspond to

Fig. 7.

The power spectrum (FFT) of the annual

10 Be

data from Dye-3 series.27

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch01

16

H. Miyahara, D. Sokoloﬀ and I. G. Usoskin

a speciﬁc state of the dynamo machine. No deﬁnite conclusion can be made on the exact cycle length though. Another suppression of the solar activity is the Dalton minimum taken place during 1790s–1820s. Although we do not regard it as a Grand minimum (see discussion in Sec. 3.1), some of its features are quite similar to those of the MM. In particular, its start was also quite abrupt,38 especially if the lost cycle in 1790s is taken into account.48 The recovery to the normal level was gradual, through fairly regular 11-year cycles. Unfortunately, the quantity of the sunspot data available was very small in the beginning of the Dalton minimum, and no deﬁnite information on cycle lengths can be obtained.22,38 A general similarity between Maunder and Dalton minima may imply that the latter, while not a Grand minimum, corresponds to an attempt of the dynamo machine to switch into the minimum mode.

4. Conclusions The main conclusion of our paper can be summarized as follows. (1) The MM of solar activity is considered as a period of extraordinarily low sunspot occurrence. The activity did not, however, completely vanish. A vestige of the nominal 11-year cycle is visible in sunspots at the end of the MM, and traces of the 11- and 22-year cycles can be found in proxy data as approximately 14- and 28-year cycles during the whole MM. As far as we can distinguish activity tracers for the Northern and Southern hemispheres, the activity demonstrates a high level of North– South asymmetry during the MM. The Southern hemisphere appeared to be much more active than the Northern. While the transition into the deep minimum phase was abrupt — nearly instantaneous comparing to the typical cycle length — the recovery from the minimum was gradual and took several decades. (2) The above phenomenological features of the MM give some hints about behavior of the solar dynamo machine during the MM times. On one hand, the start of the MM looks like a strong deviation/excursion of the solar dynamo action. On the other hand, it looks plausible that the cyclic dynamo kept working through the whole MM but producing rather unusual magnetic conﬁgurations with a strong North–South asymmetry. (3) The MM and other long-term Grand minima should be distinguished from shorter events like, e.g., the Dalton minimum. We believe that

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch01

The Solar Cycle at the Maunder Minimum Epoch

17

the solar dynamo machine tries to pass into a Grand minima state from time to time. Some of such attempts are successful and Grand minima occur while the unsuccessful attempts result in shorter events like Dalton minimum or even in minor events like a phase deviation that occurred circa 1900. In this content, in order to distinguish from such events, we would like to deﬁne a Grand minimum as a period with long-term (compared to the typical solar cycle length) and strong suppression of the solar magnetic activity. Let us ﬁnish our presentation with some remarks concerning possible perspectives in the Grand minima study. An extensive comparative study of the Sp¨ orer and Maunder minima as well as the Dalton minimum would reveal features which are common and typical for a Grand minimum. According to the nature of the available data sources, the investigation is most promising in the ﬁeld of analysis of cosmogenic isotope data. However, the most direct information comes from sunspot observations. We would like also note that even the most recent group sunspot number series,11 which replaces the famous but outdated Wolf sunspot series, is not complete, and some additional sunspot data are still to be restored from astronomical archive. In particular, we refer to the talk by Hoyt62 who stated that some sunspot observations by Scheiner, Alischer, Musano, Soemmering, Chevallier, and Williamson are still not included in the available databases. In particular, the Soemmering data known from63 look adequate to construct the butterﬂy diagram for the period 1826–1829, adjacent to the Dalton minimum. Note that a butterﬂy diagram, even if it is based on isolated data or obtained for a limited temporal interval, can still be extremely useful for the historical reconstruction of solar activity. For instance, the butterﬂy diagram64 built using Manfredi and Salvago’s observations at Bologna during 1703–1709 strongly supports the conclusion of a very strong North–South asymmetry of solar activity known from the La Hire data from Paris. Some archival data can be found also outside major observatories. For example, earlier unknown sunspot data from the Mexican astronomer J. A. Alzate65 and Portuguese astronomer Sanches Dorta66 have recently been discovered for the period 1784–1785, i.e., before the Dalton minimum when observations were quite scarce. Concluding, the MM forms a challenging problem for both experimental and theoretical studies. The bulk of direct and indirect data about solar, heliospheric and terrestrial systems, together with theoretical developments lead to a better understanding of the phenomenon. During the last decades,

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch01

18

H. Miyahara, D. Sokoloﬀ and I. G. Usoskin

a substantial progress has been achieved in this direction but the puzzle is yet far from being solved. Acknowledgments DS acknowledges the support from the Academy of Finland and Russian Foundation for Basic Researches (Grant No. 040216068). HM’s work is supported by the Grant-in-Aid for JSPS Fellows. We thank the Organizing Committee of the AOGS-2005 conference for giving us the possibility to prepare and present this review. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.

F. W. G. Sp¨ orer, Vierteljahrsschr. Astron. Ges. (Leipzig) 22 (1887) 323. F. W. G. Sp¨ orer, Bull. Astron. 6 (1889) 60. E. W. Maunder, Mon. Not. R. Astron. Soc. 50 (1894) 251. W. W.-H. Soon and S. H. Yaskell, Maunder Minimum and the Variable SunEarth Connections (World Scientiﬁc Printers, Singapore, 2003), p. 278. J. A. Eddy, Science 192 (1976) 1189. J. A. Eddy, Sci. Am. 236 (1977) 80. J. A. Eddy, Solar Phys. 89 (1983) 195. E. Ribes, J.-C. Ribes and R. Barthalot, Nature 326 (1987) 52. J.-C. Ribes and E. Nesme-Ribes, Astron. Astrophys. 276 (1993) 52. I. G. Usoskin and G. A. Kovaltsov, Solar Phys. 224 (2004) 37. D. V. Hoyt and K. Schatten, Solar Phys. 179 (1998) 189. G. A. Kovaltsov, I. G. Usoskin and K. Mursula, Solar Phys. 224 (2004) 95. P. Frick, D. Galyagin, D. V. Hoyt, E. Nesme-Ribes, K. H. Schatten, D. Sokoloﬀ and V. Zakharov, Astron. Astrophys. 328 (1996) 670. I. G. Usoskin, K. Mursula and G. A. Kovaltsov, Astron. Astrophys. 354 (2000) L33. I. G. Usoskin, K. Mursula and G. A. Kovaltsov, J. Geophys. Res. 106 (2001) 16039. I. G. Usoskin, K. Mursula and G. A. Kovaltsov, Solar Phys. 199 (2001) 187. P. Frick, E. Nesme-Ribes and D. Sokoloﬀ, Acta Astron. Geophys. Comen. 39 (1997) 113. E. Nesme-Ribes, P. Frick, D. Sokoloﬀ, V. Zakharov, J.-C. Ribes, A. Vigouroux and F. Laclare, Comptes Rendus Acad. Sci. Paris 321 (1995) 525. D. Sokoloﬀ, Solar Phys. 224 (2004) 145. J. K´ ota and J. R. Jokipii, Astrophys. J. 265 (1983) 573. M. Potgieter, R. A. Burger and S. E. S. Ferreira, Space Sci. Rev. 97 (2001) 295. I. G. Usoskin, K. Mursula and G. A. Kovaltsov, Geophys. Res. Lett. 29, 24 (2002) CiteID 2183. M. Stuiver, J. Geophys. Res. 66 (1961) 273.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch01

The Solar Cycle at the Maunder Minimum Epoch

24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49.

50. 51. 52. 53. 54. 55.

19

M. Stuiver and T. F. Braziunas, Holocene 3 (1993) 289. M. Stuiver, P. J. Reimer and T. F. Braziunas, Radiocarbon 40, 3 (1998) 1127. J. Beer et al., Nature 331 (1988) 675. J. Beer et al., Nature 347 (1990) 164. J. Beer, S. Tobias and N. Weiss, Solar Phys. 181 (1998) 237. A. N. Peristykh and P. E. Damon, Solar Phys. 177 (1998) 343. H. Miyahara, H. Masuda, Y. Muraki, H. Furuzawa, H. Menjo and T. Nakamura, Sol. Phys. 224 (2004) 317. H. Miyahara, K. Masuda, H. Furuzawa, H. Menjo, Y. Muraki, H. Kitagawa and T. Nakamura, Radiocarbon 46 (2004) 965. H. Miyahara, K. Masuda, H. Menjo, K. Kuwana, Y. Muraki and T. Nakamura, Proc. Int. Cosmic Ray Conf., Pune, India (2005). R. G. Stockwell, L. Mansinha and R. P. Lowe, IEEE Trans. Sign. Proc. 44 (1996) 998. M. Fligge, S. K. Solanki and J. Beer, Astron. Astrophys. 346 (1999) 313. S. M. Silverman, Rev. Geophys. 30 (1992) 333. L. K˘rivsk´ y and K. Pejml, Astron. Inst. Czech. Acad. Sci. 75 (1988) 32. W. Schr¨ oder, J. Geomagn. Geoelectr. 44 (1992) 119. I. G. Usoskin and K. Mursula, Solar Phys. 218 (2003) 319. E. Parker, Astrophys. J. 122 (1955) 293. F. Krause and K.-H. R¨ adler, Mean-Field Magnetohydrodynamics and Dynamo Theory (Pergamon Press, Oxford, 1980). D. Sokoloﬀ and E. Nesme-Ribes, Astron. Astrophys. 288 (1994) 293. A. Brandenburg, F. Krause, R. Meinel, I. Tuominen and D. Moss, Astron. Astrophys. 213 (1989) 411. R. L. Jennings and N. O. Weiss, MNRAS 252 (1991) 249. A. Tworkowski, R. Tavakol, A. Brandenburg, J. M. Brooke, D. Moss and I. Tuominen, MNRAS 296 (1998) 287. J. Brooke, D. Moss and A. Phillips, Astron. Astrophys. 395 (2002) 1013. J. Brooke, J. Pelt, R. Tavakol and A. Tworkowski, Astron. Astrophys. 332 (1998) 339. P. Charbonneau, Solar Phys. 199 (2001) 385. I. G. Usoskin, K. Mursula and G. A. Kovaltsov, Astron. Astrophys. 370 (2001) L31. E. Nesme-Ribes, D. Sokoloﬀ, J.-C. Ribes and M. Kremliovsky, The Solar Engine and Its Inﬂuence on Terrestrial Atmosphere and Climate, Proc. NATO Advanced Research Workshop, ed. E. Nesme-Ribes (Springer, Berlin, 1994), p. 71. W. R. Webber and P. R. Higbie, J. Geophys. Res. 108 (2003) CiteID 1355. K. G. McCracken, F. B. McDonald, J. Beer, G. Raisbeck and F. Yiou, J. Geophys. Res. 109 (2004) CiteID A12103. S. T. Suess, Planet. Space Sci. 27 (1979) 1001. B. Mendoza, Ann. Geophys. 15 (1997) 397. E. W. Cliver, V. Boriakoﬀ and K. H. Bounar, Geophys. Res. Lett. 25 (1998) 897. Y.-M. Wang and N. R. Sheeley, Jr., Astrophys. J. 591 (2003) 1248.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch01

20

56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66.

H. Miyahara, D. Sokoloﬀ and I. G. Usoskin

K. Scherer and H. Fichtner, Astron. Astrophys. 413 (2004) L11. K. Scherer, H.-J. Fahr, H. Fichtner and B. Heber, Solar Phys. 224 (2004) 305. J. R. Jokipii and B. Thomas, Astrophys. J. 243 (1981) 1115. S. E. S. Ferreira, M. Potgieter, B. Heber and H. Fichtner, Annal. Geophys. 21 (2003) 1359. V. Florinski, G. P. Zank and N. V. Pogorelov, J. Geophys. Res. 108, A6 (2003) CiteID 1228. S. K. Solanki, I. G. Usoskin, B. Kromer, M. Sch¨ ussler and J. Beer, Nature 431 (2004) 1084. D. V. Hoyt, Applied Historical Astronomy, abstract book of 24th IAU meeting, Joint discussion 6 (2000). R. C. Carrington, MNRAS 20 (1860) 71. E. Baiada and R. Merighi, Solar Phys. 77 (1982) 357. J. M. Vaquero, Solar Phys. 219 (2004) 379. J. M. Vaquero, R. M. Trigo and M. C. Gallego, Astron. Nachr. 326 (2005) 112.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch02

SOLAR CORONAL PLUMES: THEORETICAL CONCEPTS AND RESULTS MANFRED CUNTZ Department of Physics, Science Hall University of Texas at Arlington, Arlington, TX 76019-0059, USA [email protected]

In the following, I describe the status of theoretical research on solar coronal plumes. Emphasis is placed on the relevance of slow magnetosonic waves, considering results from 1D, 2D, and analytical magnetohydrodynamic models, in addition to results from empirical models and observations. Theoretical models, taking into account the combined eﬀects of plume spreading, heat conduction, and radiative damping, have shown that the waves nonlinearly steepen as they propagate, resulting in the formation of shocks at relatively low coronal heights. Consequently, slow magnetosonic waves are relevant for the energy budget at most heights, even though they do not constitute a solely operating energy supply mechanism.

1. Introduction Solar plumes have been identiﬁed within coronal hole regions and, in particular, are a well-known feature of coronal holes during solar cycle minima. Solar plumes are present in that part of the corona where the fast solar wind originates (e.g., Refs. 1 and 2). Observationally, solar plumes appear as quasi-radial rays in coronal holes (e.g., Refs. 3 and 4), visible between one and several solar radii as found by the Ultraviolet Coronograph Spectrometer (UVCS), the Large Aperture Solar Coronograph (LASCO), and the Extreme Ultraviolet Imaging Telescope (EIT) on the Solar and Heliospheric Observatory (SOHO) (e.g., Ref. 3) as well as by other space instruments and ground-based observations. These experiments have added greatly to our knowledge of the empirical properties of plumes, which are now interpreted in conjunction with theoretical studies (e.g., Refs. 5–7). Nonetheless, the plasma parameters and the dynamics of plumes are still not fully understood. DeForest et al.3 have ﬁnally unambiguously shown that all plumes lie over photospheric magnetic ﬂux concentrations, although not all ﬂux concentrations have plumes. Modeling, SOHO/UVCS, SPARTAN-201, and Interplanetary Scintillations 21

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch02

22

M. Cuntz

(IPS) have shown that plume ﬂow speeds are typically ∼ 200–300 km/s at 5.5 R ,5,8 while the interplume ﬂow may already be as high as 750 km/s,9 albeit controversies about the contributions from plume and interplume material to the fast wind have been conveyed (e.g., Refs. 10 and 11). Plumes are of low-β plasma and are considered to be in pressure balance with the ambient medium. They are visible to EIT, UVCS, and LASCO because they are more dense than the interplume plasma. There is continuing interest in plumes precisely because they (1) are a tracer of structures in the corona, (2) contribute to the mass and energy balance of the solar wind, and (3) exhibit a range of interesting dynamic phenomena. Observations of polar plumes led to the detection of quasi-periodic density variations,12–14 which were identiﬁed through quasi-periodic perturbations in the brightness of Fe IX and Fe X line emission near 171 ˚ A. The brightness perturbations amount to 10–20% of the plumes’ overall intensity and were found to propagate outward at 75–150 km/s with periods of 10–15 min. These waves are now interpreted as slow magnetosonic waves,14,15 a synonym for longitudinal tube waves. More recent observations by Banerjee et al.16 also provide unequivocal evidence for waves with longer periods, i.e., 20–30 min or more. These waves, including those with shorter periods, are expected to contribute to the heating and wind acceleration within the plumes and are therefore relevant for the overall plume dynamics. There are also studies focusing on the source of slow magnetosonic waves in coronal plumes. Ofman and Davila17 showed that this type of wave can be generated highly eﬀectively via mode coupling from Alfv´en waves for conditions pertinent to solar coronal holes. However, Alfv´en waves have not been observed directly, owing to the absence of density ﬂuctuations in the linear MHD limit, even though they can have substantial energy density.2 Mode coupling between diﬀerent types of MHD wave modes has previously been explored for solar and stellar convective zones (e.g., Refs. 18 and 19), noting however that important diﬀerences between both settings exist due to the role played by (magneto-)convection (e.g., Ref. 20; see also reviews by Narain and Ulmschneider21 and Musielak22 ). The MHD waves generated in convective zones are able to provide signiﬁcant chromospheric heating. Models based on slow magnetosonic waves (i.e., longitudinal ﬂux tube waves) for solar chromospheric structure have been given by Fawzy et al.,23 Rammacher and Ulmschneider,24 Bogdan et al.,25 Rammacher and Cuntz,26 and others. In the following, I will discuss the status of theoretical models of slow magnetosonic waves in solar plumes. In that respect, I will consider results from 1D, 2D, and analytical magnetohydrodynamic models, noting that

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch02

Solar Coronal Plumes: Theoretical Concepts and Results

23

some important features considered in 1D and analytical models have not yet been taken into account in the 2D models. I will also brieﬂy describe studies to identify solar plume geometry, expressed through global and local spreading functions.

2. Solar Plume Geometry The calculation of solar plume models requires the determination of plume geometry, typically expressed by the area function A(r). The area function describes the dilution (or concentration) of both the wave and wind energy ﬂuxes as function of height. The plume geometry also aﬀects the height of shock formation, and therefore the propagation and steepening of the waves;27 for earlier discussions on the eﬀects of geometry on shock formation at solar chromospheric heights (see, e.g., Ref. 23). In spherical coordinates, it is found that A(r) = A0 (r/r0 )2 f (r) with f (r) as spreading factor and A0 f (r0 ) as the area of the plume at its base, i.e., at radius r0 . The magnetic ﬁeld strength inside of plumes follows from the conservation of the magnetic ﬂux density φ given as φ = A(r)B(r) = A0 f (r0 )B0 with B0 as magnetic ﬁeld strength at the base. Theoretical predictions for f (r) in coronal plumes and coronal holes have been obtained by Kopp and Holzer,28 Habbal et al.,5 Hu et al.,29 Suess et al.,30 and others. Suess et al. assumed that the plume spreading can be considered as consisting of two parts, i.e., the local spreading fl (r) and the global spreading fg (r) with f (r) mathematically given by f (r) = fl (r)fg (r). The local spreading fl (r) is important below 35 000 km (1.05 R ), varies rapidly at these heights, and is constant above that height. On the other hand, fg (r) varies much more slowly at those heights. Suess et al.30 found that the geometrical spreading of plumes can be computed with acceptable accuracy independent of the ﬂow because β 1 in plumes and throughout the surrounding coronal holes from the base of the corona to at least 10 R . Conversely, the global spreading fg (r) can be computed from a global model of the corona.31 The spreading function of coronal plumes by Suess et al.30 has been used in the one-dimensional (1D) time-dependent as well as the analytic plume models (see Secs. 3.2 and 3.3) (Fig. 1).

3. Theoretical Solar Plume Models In the following, I will describe the status of 1D, 2D, and analytical MHD models for solar plumes, which are especially aimed at the description of slow magnetosonic waves, sometimes also called longitudinal ﬂux tube

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch02

24

M. Cuntz

Fig. 1. Plume geometry between 1.01 and 5 R (solid lines) following Suess et al.30 The spreading according to spherical geometry is given for comparison (dotted lines).

waves. These waves seem to correspond to previously observed compressive waves, even though there are apparently only responsible for a (relatively) small fraction of the energy ﬂux identiﬁed in solar plumes. It is noteworthy that important features considered in 1D and analytical models have not yet been considered in the 2D models, and vice versa. 3.1. 2D time-dependent solar plume models Arguably, the most sophisticated models for solar plumes have been presented by Ofman et al.15,32 They started by deriving a linear dispersion relation for the waves. Next, they solved an evolutionary equation of the Burgers type, incorporating the eﬀects of radial stratiﬁcation, quadratic nonlinearity, and viscosity. They also employ the WKB approximation, which is mostly adequate for waves of weak nonlinearity and dissipation, while neglecting the wave reﬂection from the radial inhomogeneity. Finally, they modeled the propagation and dissipation of the slow magnetosonic waves in polar plumes using their 1D and 2D codes in spherical symmetry. They found that the waves nonlinearly steepened as they propagate away from the Sun into the solar wind. The nonlinear steepening of the waves resulted in enhanced dissipation. This eﬃcient dissipation of the waves led to damping of the waves within the ﬁrst solar radii above the surface. In particular, Ofman et al.32 investigated the parametric dependence of the wave properties. Their models considered diﬀerent values for the

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch02

Solar Coronal Plumes: Theoretical Concepts and Results

25

wave periods and amplitudes in accordance with EUV intensity ﬂuctuations observed by the EIT and UVCS. The models of Ofman et al. especially focussed on the radial dependence of the wave amplitude, viscosity, density perturbation, and the viscous heating rate in the limit of small-amplitude waves. The authors conclude that “(t)he slow magnetosonic waves may contribute to the acceleration of the solar wind close to the Sun by two separate mechanisms. First, by dissipating and heating the coronal plasma, the waves will increase the thermal pressure, which will lead to the acceleration of the solar wind. Second, the waves may transfer momentum to the ambient plasma by accelerating the particles directly by Landau damping farther from the Sun, when collisions are less frequent”. Nevertheless, the energy ﬂux associated with these waves appears to be small, indicating that almost certainly they do not constitute a solely operating energy supply mechanism. Ofman et al.15 estimated an energy ﬂux between 2 × 103 and 3 × 104 ergs/cm2/s for slow magnetosonic waves in coronal plumes, assuming typical plasma conditions, compared to ∼ 105 ergs/cm2 /s, required to accelerate the fast solar wind. This estimate for the wave energy ﬂux also corrects a previous estimate by DeForest and Gurman14 that has been notably higher, but was aﬀected by a computational error. It should be noted that the work by Ofman et al.15,32 is especially important because of their choice of treating the wave propagation in a time-dependent, 2D setting. On the other hand, the dampening of the waves revealed in their study may be heavily inﬂuenced by the adopted treatment of energy dissipation (“weak quadratic nonlinearities”), possibly producing unreliable results. In any event, the adopted dissipation formalism is not based on a self-consistent implementation of the 2D Rankine-Hugoniot relations in accordance with the quasi-discontinuous structure of the shocks. Moreover, the authors refrain from considering detailed plume geometry (see Secs. 2, 3.2, and 3.3).

3.2. 1D time-dependent solar plume models So far, 1D time-dependent solar plume models had been pursued to describe the height of shock formation in coronal plumes for slow magnetosonic waves.27 This is relevant because shock formation is decisive for the energy and momentum input by these waves. Furthermore, the onset of shock formation is directly related to nonlinear eﬀects, which are associated with various thermodynamic, MHD, and radiative processes (Fig. 2).

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch02

26

M. Cuntz

Fig. 2. Snapshot of a slow magnetosonic wave with a period of 600 s and an initial amplitude of 0.10 Mach in a solar coronal plume after 970 s. Shown are: the ﬂow speed v (solid line), temperature T (dashed line), magnetic ﬁeld strength B (dashed-dotted line), and the thermal heating rate E˙ (dotted line), which is almost entirely determined by heat conduction (see Ref. 27 for details).

The solar plume models by Cuntz and Suess27 take into account plume geometric spreading, heat conduction and radiative damping. The wave parameters as well as the spreading functions of the plumes, adopted from Ref. 30 and the base magnetic ﬁeld strength follow empirical constraints mostly from SOHO/UVCS data. The shock formation is calculated using the well-established wave breaking condition given by the intersection of C+ characteristics in the space–time plane (e.g., Ref. 33). The models by Cuntz and Suess27 consider wave periods between 10 and 15 min,14 with initial wave amplitudes between 0.05 and 0.2 Mach (7.1 and 28.3 km/s, respectively). Shock formation is found to occur at relatively low coronal heights, i.e., within 1.3 R , depending on the model parameters (see Table 1). The shock formation height is lowest for large amplitude waves and waves with relatively short periods. The height of shock formation is signiﬁcantly aﬀected by the assumed geometry of the plumes. It is found that the large plume spreading close to its base (see Fig. 1) increases the height of shock formation compared to models with spherical geometry. This result is a direct consequence of the dilution of the wave

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch02

Solar Coronal Plumes: Theoretical Concepts and Results Table 1.

Shock formation heights.

P = 10 min

0.05 0.10 0.15 0.20

27

P = 15 min

PL

SPH

PL

SPH

1.255 1.158 1.115 1.090

1.196 1.119 1.084 1.065

1.325 1.207 1.154 1.122

1.248 1.152 1.109 1.083

Note: The table depicts shock formation heights for diﬀerent wave amplitudes (in units of the sound speed) for diﬀerent types of models. The models assume wave periods of P = 10 min or 15 min and spherical (SPH) or plume (PL) geometry (see Ref. 27).

energy ﬂux due to a larger spreading of the plume near the plume base. A similar phenomenon has been found in models of chromospheric ﬂux tubes as discussed by Fawzy et al.23

3.3. Analytical solar plume models A further type of model, which in principle is much more basic than the 1D and 2D time-dependent magnetohydrodynamic models previously described, is the analytic solar plume model by Cuntz and Suess.34 This model assumes a complete linearization of the basic equations (see Refs. 35 and 36 for details), except that the shocks are treated as discontinuities, assuming a saw-tooth proﬁle for the ﬂow speed and the relevant thermodynamic quantities. The focus of the models is to study the behavior of the shock strength as function of height. The models consider (1) the geometrical structure of plumes expressed through the geometrical spreading functions of Suess et al.30 and (2) eﬀects of the ﬂow patterns in solar plumes, given by the fast wind originating in solar plumes. The assumed density distribution is given by ρ ∼ r−8.785 , closely following the empirically deduced height distribution of the electron density in White-Light polar plumes (Ref. 4; see also results by Giordano et al.37 ), allows constraints on the ﬂow speed of the wind. The ﬂow patterns are found to alter both the (height-dependent) shock strengths and the energy damping length (“scale height of dissipation”) of the wave energy ﬂux. The outline of this study is consistent with the ﬁnding that slow magnetosonic waves form shocks at relatively low coronal heights (Ref. 27; see

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch02

28

M. Cuntz

Fig. 3. Wave energy damping length for a variety of models following Cuntz and Suess.27 They show two sets of curves consisting of three lines each, belonging to the following models: plume geometry without wind, P = 10 min (dashed lines) and plume geometry with wind, P = 10 min (solid lines) and P = 15 min (dashed-dotted lines). For the upper set of curves (1), the wave amplitudes are 7.5 km/s, whereas for the lower set of curves (2) they are 15 km/s instead. For comparison, Cuntz and Suess also show the case of spherical symmetry, no wind, for waves of P = 10 min and 7.5 km/s (dotted line), see curve (3).

Sec. 3.2). Assuming wave amplitudes and wave periods in accord with observations, Cuntz and Suess34 found that the damping length of the wave energy ﬂux increases up to 85% through the “carry-along eﬀect” by the wind (see Fig. 3), initiated by another heating mechanism as, e.g., Alfv´en waves.17,38,39 This eﬀect was found to be largest for high-amplitude waves. The carry-along eﬀect of the wind eﬃciently oﬀset the dilution of the wave energy by the plume geometry, owing to the increase of the plume spreading with height. The results of this study are a strong motivation for future studies of wave propagation in solar-type winds, ideally based on self-consistent time-dependent simulations. The envisioned simulations should also explore the relationship, if any, between the initiation and propagation of the slow magnetosonic waves, on one hand, and the physics of the underlying generation mechanism of the fast solar wind, which appears to emerge from solar plumes, on the other hand (e.g., Refs. 1 and 2). This eﬀort should be undertaken in conjunction with 2D or 3D magnetohydrodynamic models.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch02

Solar Coronal Plumes: Theoretical Concepts and Results

29

4. Conclusions and Outlook Based on previous observational and theoretical results, I would like to summarize the following eﬀorts: • Plumes are found to exist in solar coronal holes, and are denser and cooler than the surrounding medium. They have been seen from 1 to at least 30 R , although they fade with respect to the background above 10 R (e.g., Ref. 40). • Plumes are impacted by compressive waves, identiﬁed as slow magnetosonic waves, with periods of 10–30 min, and possibly longer (e.g., Refs. 12, 14, and 16). The likely source of these waves is due to mode coupling from torsional Alfv´en waves in coronal holes.17 • The energy input by the waves appears to be relevant at most heights, especially if long-period waves are included in the estimations. However, slow magnetosonic waves almost certainly do not constitute a solely operating energy supply mechanism, noting that based on estimates by Ofman et al.,15 the energy ﬂux of these waves is about a factor of 3.5–50 lower than the total energy required to accelerate the fast solar wind, depending on the adopted plasma and wave parameters. • 2D models by Ofman et al.15,32 show that 10–15 min slow magnetosonic waves steepen and propagate away from the Sun into the solar wind, and show signiﬁcant dissipation at the wave fronts. The models incorporate the eﬀects of radial stratiﬁcation and an approximate treatment of nonlinearities. However, they refrain from considering detailed plume geometry. Because of the adopted treatment of the wave energy dissipation, the dampening of the waves may be inaccurate. • 1D models by Cuntz and Suess,27 considering a highly accurate method for the treatment of shock formation as well as detailed plume geometry, show that 10–15 min slow magnetosonic waves form shocks within 1.3 R . This implies that the wave energy and momentum input is highly nonlinear, and is possibly in excess of previous estimates, in particular when waves with periods beyond 20 min16 are considered. • Analytic models by Cuntz and Suess34 indicate that a “secondary” solar wind mechanism can result in a spreading out of the energy dissipation of magnetosonic waves, thus increasing the relevance of these waves at larger coronal heights. • Future models should address the combined eﬀects of 2D wave propagation, (nonlinear) shock formation, realistic plume geometry, radiative heating, heat conduction, and the eﬀect(s) of additional coronal heating/wind acceleration mechanism(s).

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch02

30

M. Cuntz

Acknowledgments This work has been supported by the UAH/USRA/NASA Cooperative Agreement on Research in Space Plasma Physics and by the National Science Foundation under grant ATM-0087184. M. C. also acknowledges the receipt of an International Travel Grant provided by the American Astronomical Society (AAS) in support of his attendence of the 2nd Annual Meeting of the AOGS in Singapore.

References 1. K. Wilhelm, I. E. Dammasch, E. Marsch and D. M. Hassler, Astron. Astrophys. 353 (2000) 749. 2. S. R. Cranmer, Space Sci. Rev. 101 (2002) 229. 3. C. E. DeForest, J. T. Hoeksema, J. B. Gurman, B. J. Thompson, S. P. Plunkett, R. Howard, R. C. Harrison and D. M. Hassler, Solar Phys. 175 (1997) 393. 4. S. Koutchmy and K. Bocchialini, Solar Jets and Coronal Plumes, ESA SP421 (1998), p. 51. 5. S. R. Habbal, R. Esser, M. Guhathakurta and R. R. Fisher, Geophys. Res. Lett. 22 (1995) 1465. 6. L. Del Zanna, A. W. Hood and A. W. Longbottom, Astron. Astrophys. 318 (1997) 963. 7. S. Casalbuoni, L. Del Zanna, S. R. Habbal and M. Velli, J. Geophys. Res. 104 (1999) 9947. 8. Y.-M. Wang, Astrophys. J. 435 (1994) L153. 9. R. R. Grall, W. A. Coles, M. T. Klinglesmith, A. R. Breen, P. J. S. Williams, J. Markkanen and R. Esser, Nature 379 (1996) 429. 10. A. H. Gabriel, F. Bely-Dubau and P. Lemaire, Astrophys. J. 589 (2003) 623. 11. A. H. Gabriel, L. Abbo, F. Bely-Dubau, A. Llebaria and E. Antonucci, Astrophys. J. 635 (2005) L185. 12. L. Ofman, M. Romoli, G. Poletto, G. Noci and J. L. Kohl, Astrophys. J. 491 (1997) L111; Erratum Astrophys. J. 507 (1998) L189. 13. L. Ofman, M. Romoli, G. Poletto, G. Noci and J. L. Kohl, Astrophys. J. 529 (2000) 592. 14. C. E. DeForest and J. B. Gurman, Astrophys. J. 501 (1998) L217. 15. L. Ofman, V. M. Nakariakov and C. E. DeForest, Astrophys. J. 514 (1999) 441. 16. D. Banerjee, E. O’Shea, J. G. Doyle and M. Goossens, Astron. Astrophys. 380 (2001) L39. 17. L. Ofman and J. M. Davila, Astrophys. J. 476 (1997) 357. 18. B. Roberts, Solar Phys. 69 (1981) 27. 19. T. J. Bogdan, C. S. Rosenthal, M. Carlsson, V. Hansteen, A. McMurry, E. J. Zita, M. Johnson, S. Petty-Powell, S. W. McIntosh, ˚ A. Nordlund, R. F. Stein and S. B. F. Dorch, Astron. Nachrichten 323 (2002) 196.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch02

Solar Coronal Plumes: Theoretical Concepts and Results

31

20. P. Ulmschneider and Z. E. Musielak, Astron. Astrophys. 338 (1998) 311. 21. U. Narain and P. Ulmschneider, Space Sci. Rev. 75 (1996) 453. 22. Z. E. Musielak, Stars as Suns: Activity, Evolution and Planets, IAU Symposium 219, eds. A. K. Dupree and A. O. Benz (Astr. Soc. Pac., San Francisco, 2004), p. 437. 23. D. E. Fawzy, P. Ulmschneider and M. Cuntz, Astron. Astrophys. 336 (1998) 1029. 24. W. Rammacher and P. Ulmschneider, Astrophys. J. 589 (2003) 988. 25. T. J. Bogdan, M. Carlsson, V. H. Hansteen, A. McMurry, C. S. Rosenthal, M. Johnson, S. Petty-Powell, E. J. Zita, R. F. Stein, S. W. McIntosh and ˚ A. Nordlund, Astrophys. J. 599 (2003) 626. 26. W. Rammacher and M. Cuntz, Astron. Astrophys. 438 (2005) 721. 27. M. Cuntz and S. T. Suess, Astrophys. J. 549 (2001) L143. 28. R. A. Kopp and T. E. Holzer, Solar Phys. 49 (1976) 43. 29. Y.-Q. Hu, R. Esser and S. R. Habbal, J. Geophys. Res. 102 (1997) 661. 30. S. T. Suess, G. Poletto, A.-H. Wang, S. T. Wu and I. Cuseri, Solar Phys. 180 (1998) 231. 31. A. H. Wang, S. T. Wu, S. T. Suess and G. Poletto, J. Geophys. Res. 103 (1998) 1913. 32. L. Ofman, V. M. Nakariakov and N. Sehgal, Astrophys. J. 533 (2000) 1071. 33. P. Ulmschneider, W. Kalkofen, T. Nowak and U. Bohn, Astron. Astrophys. 54 (1977) 61. 34. M. Cuntz and S. T. Suess, Astron. Astrophys. 424 (2004) 1003. 35. M. Cuntz, Astron. Astrophys. 350 (1999) 1100. 36. M. Cuntz, Astron. Astrophys. 420 (2004) 699. 37. S. Giordano, E. Antonnucci, G. Noci, M. Romoli and J. L. Kohl, Astrophys. J. 531 (2000) L79. 38. R. Esser, E. Leer, S. R. Habbal and G. L. Withbroe, J. Geophys. Res. 91 (1986) 2950. 39. K. K. Ong, Z. E. Musielak, R. Rosner, S. T. Suess and M. E. Sulkanen, Astrophys. J. 474 (1997) L143. 40. C. E. DeForest, S. P. Plunkett and M. D. Andrews, Astrophys. J. 546 (2001) 569.

This page intentionally left blank

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch03

PREFLARE FEATURES IN MICROWAVES AND IN HARD X-RAYS AYUMI ASAI∗ , HIROSHI NAKAJIMA and MASUMI SHIMOJO Nobeyama Solar Radio Observatory National Astronomical Observatory of Japan Minamisaku, Nagano 384-1305, Japan ∗[email protected] STEPHEN M. WHITE Department of Astronomy, University of Maryland College Park, MD 20742, USA

We present a detailed examination on the nonthermal emissions during the preﬂare phase of the X4.8 ﬂare which occurred on July 23, 2002. The microwave (17 and 34 GHz) data obtained with Nobeyama Radioheliograph (NoRH), at Nobeyama Solar Radio Observatory, National Astronomical Observatory of Japan, and the hard X-ray (HXR) data taken with Reuven Ramaty High Energy Solar Spectroscopic Imager (RHESSI ) distinctly showed nonthermal features. We examined the temporal, spatial, and spectroscopic characteristics of the emission sources, and found loop-top sources during the preﬂare phase both in HXRs and in microwaves. Moreover, we found that the electron spectral index derived from microwave emission closely corresponds to that obtained from the HXR emission. We also discuss the energy release mechanism in the preﬂare phase.

1. Introduction Nonthermal emissions from accelerated particles are often observed in hard X-rays (HXRs), γ-rays, and microwaves at the beginning of a solar ﬂare. These nonthermal emissions are associated with intense energy release processes. The particle acceleration mechanism has been one of the most important and diﬃcult problems in solar physics (see reviews by, e.g., Ref. 1). Nonthermal emissions are associated with even a small energy release process such as in a microﬂare.2 However, it has been thought that particle acceleration works eﬃciently only in the impulsive phase.3 On the other hand, it is also interesting to study preﬂare activity, since this may hold the key for understanding how the catastrophic energy release of the ﬂare is triggered. In the preﬂare stage we sometimes ﬁnd ﬂarepredictive phenomena, such as a gradual enhancement of soft X-ray (SXR) 33

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch03

34

A. Asai et al.

emission, rise of SXR plasmoids and/or Hα ﬁlaments, etc. Even in the preﬂare stage of a solar ﬂare, some energy release process is probably occurring at a low level, although the energy release is much milder. It is not widely accepted that nonthermal particles are present in signiﬁcant numbers prior to the impulsive phase of a ﬂare, rather it is common to speak of preﬂare heating implying thermal behavior. Therefore, the reports on the nonthermal emissions during the preﬂare phases have been mostly negative. Recently, Holman et al.4 examined the HXRs features of the July 23, 2002 ﬂare, and reported that the nonthermal energy was large even before the impulsive phase. Motivated by their work, we analyzed this ﬂare, and found suﬃcient emissions both in HXRs and in microwaves that can be candidates for nonthermal emissions during the preﬂare phase. In order to derive information on the energy release in the preﬂare phase, we examined in detail the features of the emission sources spatially, temporally, and spectroscopically. We report the results of the investigations of the emissions in HXRs and in microwaves during the preﬂare phase, and discuss the relation between the nonthermal emissions and other observed phenomena.

2. Observations and Results The intense solar ﬂare, X4.8 on the GOES scale, occurred in NOAA Active Region 10039 (S12◦ , E72◦ ) at 00:18 UT, July 23, 2002. This ﬂare showed many spectacular features5 in HXR and γ-ray wavelengths obtained with the Reuven Ramaty High Energy Solar Spectroscopic Imager (RHESSI 6 ). This ﬂare was also observed in microwaves with the Nobeyama Radioheliograph (NoRH7 ).9 NoRH observes the Sun in two frequencies, 17 and 34 GHz, which allows us to derive a spectral index α (Fν ∝ ν α ; Fν is the ﬂux at frequency ν) with a temporal resolution of 1 s. The spatial resolutions (FWHMs) of NoRH data are 14 for 17 GHz and 7 for 34 GHz. We synthesized the HXR images obtained with RHESSI by using grids 3–8 which gives the spatial resolution (FWHM) of about 7 . We integrated over 20 s to synthesize each image used in this paper. We also determined the temperature and the emission measures of the thermal plasma in the corona by using the ratios of the two of GOES channels. EUV images of the ﬂare were obtained with the Transition Region and Coronal Explorer A images, in which the Fe XII line formed at (TRACE10,11 ). We used 195 ˚ ∼ 1 MK is normally dominant. The pixel size of the CCD is 1· 0, and the temporal resolution is about 9 s.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch03

Preﬂare Features in Microwaves and in Hard X-Rays

35

We focus on the nonthermal emissions in HXRs and in microwaves of the preﬂare phase, from about 23:00 UT, July 22, 2002 to about 00:30 UT, July 23, 2002. We divide the preﬂare phase into four sub-phases, and examine each phase in more detail. The bottom panel of Fig. 1 shows the expanded light curves of the preﬂare phase of the ﬂare (from 00:10 UT to 00:30 UT, July 23, 2002), which corresponds to the time between the

GOES Flux [W m-2]

10-3 10-4

GOES

10-5 10-6

23:30 GOES Flux [W m-2]

10-3 10-4

GOES

00:00

(1)

00:30

(2)

01:00

01:30

(3) (4)

10-5 10-6

17 GHz 34 GHz

12 - 25 keV 25 - 40 keV 60 - 100 keV

00:12

00:16 00:20 00:24 Time (2002-July-23) [UT]

00:28

˚ channel of the July 23, Fig. 1. Top panel: Soft X-ray ﬂux in the GOES 1.0–8.0 A 2002 ﬂare. Two dashed vertical lines show the time range of the preﬂare phase. Bottom panel: Light curves of the preﬂare phase. From top to bottom: Soft X-ray ﬂux in the GOES 1.0–8.0 ˚ A channel; radio correlation plot observed at 17 and 34 GHz with NoRH (scaled arbitrary); hard X-ray count rate measured with RHESSI in 12–25 keV and in 60–100 keV (scaled arbitrarily). The dotted vertical lines divide the preﬂare phase into four sub-phases as numbered in the top frame.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch03

36

A. Asai et al.

-180

23-Jul-2002 00:02:33.000

23-Jul-2002 00:22:29.000

23-Jul-2002 00:24:39.000

23-Jul-2002 00:26:52.000

-200 -220 -240 -260 -280

(a)

-180

(e)

(

b)

(c)

(

d)

(f)

(g)

(

h)

-200 -220 -240 -260

α index

-280 -4

-960 -940 -920 -900 -880 -860

-960 -940 -920 -900 -880 -860

-960 -940 -920 -900 -880 -860

-2

0

2

-960 -940 -920 -900 -880 -860

Fig. 2. Images of the ﬂare for each phase. The top panels ((a)–(d)) show the EUV images taken with TRACE 195 ˚ A. The contours show the NoRH 34 GHz brightness temperature (red), the RHESSI 25–40 keV intensity (yellow), and the 40–60 keV intensity (blue, only in the panel d), respectively. The bottom panels ((e)–(h)) show the maps of NoRH α index overlaid with the RHESSI 12–25 keV intensity (white). The 34 GHz contours are 15, 30, 50, 70, and 90% of the peak intensity. Contours for the RHESSI HXR images are at 20, 40, 60, 80, and 95% of the peak intensity.

dashed lines in the top panel. The time ranges of the four sub-phases are numbered at the top of the bottom panel of Fig. 1. Figure 2 shows the evolution of the ﬂare. The top panels show the TRACE 195 ˚ A images overlaid with the contour images of the NoRH 34 GHz (red), the RHESSI 25–40 keV (yellow), and the 40–60 keV (blue, only in the right panel), respectively. The bottom panels show the maps of NoRH α index overlaid with the RHESSI 12–25 keV intensity (white).

2.1. Before the ﬂare (sub-phase 1) First, we focus on the time period from about 23:30 UT, July 22, 2002 to 00:16 UT, July 23, 2002. In Fig. 2(a) we can see a large loop-like bright region in the NoRH image.9 We derived the spectral index α of the emission source, and found it to be within the range of −0.4–0.6, which indicates that the optically thin (free-free) thermal emission is dominant for the source. Moreover, the polarization of the sources is no more than 10%, which eliminates the possibility of the emission from the gyroresonance near the sunspot umbrae.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch03

Preﬂare Features in Microwaves and in Hard X-Rays

37

The GOES temperature shows the existence of hot plasma even in this phase of about 5 MK. This could be related to a small ﬂare which occurred at 22:00 UT in the same active region, although we could not conﬁrm it due to the lack of image data for the small event. Although TRACE was observing the region, we cannot see any active phenomena at the time. This is presumably because the large structure is too hot to be observed in the EUV range, and we suppose that the large loop-like structure is a kind of sigmoid which is often observed at the preﬂare phase in SXRs (see, e.g., Ref. 8).

2.2. Preﬂare phase (sub-phase 2) Second, we examine the time for sub-phase 2 marked in Fig. 1 (bottom panel) which includes the ﬁrst ﬂare emissions (from 00:16 UT to 00:23 UT). We can see some features of thermal emission and also clear signatures of nonthermal emission. The GOES temperature rapidly increases from about 4.5 MK at 00:15 UT to above 10 MK at 00:22 UT. At the same time, the RHESSI count rate in 12–25 keV increases. Such HXR brightenings in lower energy bands, associated with a hotter GOES source, are often observed in a preﬂare phase, and the emissions are thought to be thermal. Moreover, Holman et al.4 performed a spectroscopic analysis of the ﬂare with RHESSI data, and reported that the thermal component of the region has high temperatures up to 20– 30 MK. After a short delay, from 00:18 UT the NoRH 17 GHz emission starts to rise, and its temporal evolution resembles the RHESSI 25–40 keV light curve. The NoRH 34 GHz emission and the RHESSI 40–60 keV count rate start to rise at 00:22 UT almost simultaneously. Figures 2(b) and 2(f) show the images of this phase taken at 00:22:30 UT. In the TRACE images, we can see that a large two-ribbon structure9 brightens from 00:20 UT. The brightening of the ribbons implies that a larger structure rather than the core of the ﬂare is destabilized in this phase. The TRACE EUV images in Figs. 2(b) and 2(c) were taken a few minutes after this phase. We can see a diﬀuse loop-like structure that is identiﬁed as Fe XXIV emission from 20 MK plasma, as is often observed in TRACE 195 ˚ A images during ﬂares. A new microwave source appears above the ﬂare ribbons at (−878, −243) arcsec heliocentric. This site corresponds to the post ﬂare loops, which became visible in the later phase in the TRACE images connecting the ﬂare ribbons. The α index is about −3.0, which implies that this source is emitting nonthermal-gyrosynchrotron radiation. The index is quite small and shows a steep (soft) power-law spectrum. An

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch03

38

A. Asai et al.

HXR source also appeared at this site, although at slightly higher location (−890, −240) than the microwave emission source. The HXR source is visible in both 12–25 and 25–40 keV bands. These sources could resemble to the “loop-top” HXR source.12 On the other hand, we can also see footpoint sources which are located on the TRACE ﬂare ribbons mentioned above both in the microwave (34 GHz) and in the HXRs (12–25, 25–40 keV). The energy release probably occurred in the corona, a part of which is deposited at the footpoints to produce the EUV brightenings. The HXR emissions from the ﬂare ribbon are thought to be generated by thick-target emission by nonthermal electrons. This is an evidence for the existence of the nonthermal particles in this phase. 2.3. Ejection (sub-phase 3) Next, we focus on the small microwave burst and the faint EUV ejection which occurred at about 00:23:30 UT. The apparent speed of the EUV ejection on the time slice image is roughly about 250 km/s. A halo coronal mass ejection (CME) associated with the ﬂare was also observed with Large Angle Spectrometric Coronagraph (LASCO) on board the Solar and Heliospheric Observatory (SOHO: see the SOHO LASCO CME online cataloga ). The combination of ejections and HXR bursts has been observed in impulsive ﬂares,14–16 and is consistent with the so-called “plasmoid-induced reconnection model”.17 The plasmoid ejections correspond in detail with nonthermal emissions, and CME acceleration may also show this pattern.18 Associated with the EUV ejection, a tiny burst was observed in microwave (see the NoRH 17 GHz light curve in Fig. 1). We synthesized the images in microwaves and in HXRs around the time of this peak, and examined the spectroscopic features of the coronal emission sources. Figure 3 shows the results. In the right panel we present a NoRH 34 GHz image. We overlaid the counter image with the gray line to show the coronal emission source. The ﬂux spectral index α of this emission source is about −3.1. Then, we can estimate the electron spectral index δ (N (E) ∝ E −δ ) to be about 4.8, if we assume the gyro-synchrotron emission for the microwave emission source. We also overlaid the HXR contour images on the TRACE (left) and NoRH (right) images with black lines. We performed the spectral ﬁtting with a power law distribution for the core of the HXR emission (black box), and found that the spectral index γ (IHXR () ∝ −γ ) is about 5.3. If we assume the thin-target model to explain the emission source, the a See

http://cdaw.gsfc.nasa.gov/CME list/.13

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch03

Preﬂare Features in Microwaves and in Hard X-Rays

39

23-Jul-2002 00:23:56.000 23-Jul-2002 00:22:46.028 -200

-220

-240

-260 -900

-880

-860

-900

-880

-860

Fig. 3. Left: an EUV image of the ﬂare observed with TRACE. Right: a microwave (34 GHz) image obtained with NoRH, overlaid with the counter image with the gray line to show the position of the coronal emission source. In both images the HXR (25–40 keV) counter images are overlaid with the black lines. The black box in the right panel shows the region used for the imaging spectroscopy.

electron spectral index δ is estimated to be 4.8, which completely corresponds to that derived from the microwave emission. These results suggest that the nonthermal electrons which generate the microwave emission sources are accelerated with the same mechanism as that for the HXRemitting electrons. 2.4. Impulsive phase (sub-phase 4) Finally, we examine the time after the TRACE ejection and before the start of the impulsive phase (from 00:24 UT to 00:27 UT). Roughly speaking, the physical features are the same as in the impulsive phase, as Krucker, Hurford and Lin19 reported. The NoRH 34 GHz source moved slightly northward (−875, −230), and showed a loop like structure (Figs. 2(c) and 2(d)). Then the NoRH 34 GHz source gradually localize on the upper section of the loop. This loop structure corresponds to the most intense post-ﬂare loop which appeared later in the TRACE 195 ˚ A images. The HXR RHESSI 12– 25 keV and 25–40 keV emissions still appear at the top of the NoRH loop. The HXR coronal sources ascend slightly as the ﬂare progresses. A notable result is that an HXR loop-top source is observed even in 40–60 keV (Fig. 2(d)). In this phase, the α index increased (became harder) to about −1.5 (Fig. 2(h)).

3. Discussion and Summary We examined in detail the nonthermal emissions in the preﬂare phase spatially, temporally, and spectroscopically. We also examined the relation

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch03

40

A. Asai et al.

between the nonthermal emissions and other observed phenomena. We identiﬁed a faint EUV ejection in the TRACE data which was associated with a nonthermal microwave burst, just before the fast energy release process occurred in the impulsive phase. In the phase before the ejection, we found observational evidence of both thermal and nonthermal emissions in the corona above the ﬂare ribbon structure. Furthermore, the imaging spectroscopic analyses for the emission sources suggest that they are generated by the accelerated electrons with the same spectral indices. Under the standard reconnection model, the current sheet reduces its thickness in the preﬂare phase,20 which leads to the fast magnetic reconnection and the violent energy release in the impulsive phase. This process is associated with the slow reconnection and/or the low-level energy release, and leads to the heating of the coronal plasma as often observed. Our results, on the other hand, indicate that the process also releases enough energy with the right conditions to accelerate particles to nonthermal energies. This suggests that energy release mechanism in the preﬂare phase of a typical ﬂare may be accompanied by particle acceleration, although it is much milder than that in the impulsive phase and therefore diﬃcult to detect in ﬂares smaller than this event.

Acknowledgments We ﬁrst acknowledge anonymous referees for their useful comments and suggestions. We wish to thank Drs. H. S. Hudson, P. R. Lin, G. D. Holman, L. Sui for fruitful discussions and their helpful comments. We made extensive use of TRACE and RHESSI Data Center.

References 1. M. J. Aschwanden, Space Sci. Rev. 101 (2002) 1. 2. A. O. Benz and P. C. Grigis, Adv. Space Res. 32 (2003) 1035. 3. S. T. Wu et al., Energetic Phenomena on the Sun, eds. M. Kundu and B. Woodgate (NASA CP-2439, Washington DC, 1986). 4. G. E. Holman, L. Sui, R. A. Schwartz and A. G. Emslie, ApJL 595 (2003) L97. 5. R. P. Lin et al., ApJL 595 (2003) L69. 6. R. P. Lin et al., Sol. Phys. 210 (2002) 3. 7. H. Nakajima et al., Proc. IEEE 82 (1994) 705. 8. P. K. Manoharan, L. van Driel-Gesztelyi, M. Pick and P. Demoulin, ApJL 468 (1996) L73.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch03

Preﬂare Features in Microwaves and in Hard X-Rays

41

9. S. M. White, S. Krucker, K. Shibasaki, T. Yokoyama, M. Shimojo and M. R. Kundu, ApJL 595 (2003) L111. 10. B. N. Handy et al., Sol. Phys. 187 (1999) 229. 11. C. J. Schrijver et al., Sol. Phys. 187 (1999) 261. 12. S. Masuda, T. Kosugi, H. Hara, S. Tsuneta and Y. Ogawara, Nature 371 (1994) 495. 13. S. Yashiro, N. Gopalswamy, G. Michalek, O. C. St. Cyr, S. P. Plunkett, N. B. Rich and R. A. Howard, JGR 109 (2004) A07105. 14. R. Kano, X-Ray Solar Physics from Yohkoh, eds. Y. Uchida, T. Watanabe, K. Shibata and H. S. Hudson (Universal Academy Press, Tokyo, 1994), p. 273. 15. H. S. Hudson, L. W. Acton and S. L. Freeland, ApJ 470 (1996) 629. 16. M. Ohyama and K. Shibata, PASJ 49 (1997) 249. 17. K. Shibata, Ap&SS 264 (1999) 129. 18. J. Zhang, K. P. Dere, R. A. Howard, M. R. Kundu and S. M. White, ApJ 559 (2001) 452. 19. S. Krucker, G. J. Hurford and R. P. Lin, ApJL 595 (2003) L103. 20. T. Magara and K. Shibata, ApJ 514 (1999) 456.

This page intentionally left blank

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

NONEXTENSIVE ENTROPY APPROACH TO SPACE PLASMA FLUCTUATIONS AND TURBULENCE M. P. LEUBNER Institute of Astrophysics, University of Innsbruck, Austria [email protected] ¨ OS ¨ and W. BAUMJOHANN Z. VOR Space Research Institute, Austrian Academy of Sciences, Graz, Austria

Spatial intermittency in fully developed turbulence is an established feature of astrophysical plasma ﬂuctuations and in particular apparent in the interplanetary medium by in situ observations. In this situation, the classical Boltzmann– Gibbs extensive thermo-statistics, applicable when microscopic interactions and memory are short ranged and the environment is a continuous and diﬀerentiable manifold, fails. Upon generalization of the entropy function to nonextensivity, accounting for long-range interactions and thus for correlations in the system, it is demonstrated that the corresponding probability distribution functions (PDFs) are members of a family of speciﬁc power-law distributions. In particular, the resulting theoretical bi-κ functional reproduces accurately the observed global leptokurtic, non-Gaussian shape of the increment PDFs of characteristic solar wind variables on all scales, where nonlocality in turbulence is controlled via a multiscale coupling parameter. Gradual decoupling is obtained by enhancing the spatial separation scale corresponding to increasing κ-values in case of slow solar wind conditions where a Gaussian is approached in the limit of large scales. Contrary, the scaling properties in the high speed solar wind are predominantly governed by the mean energy or variance of the distribution, appearing as second parameter in the theory. The PDFs of solar wind scalar ﬁeld diﬀerences are computed from WIND and ACE data for diﬀerent time-lags and bulk speeds and analyzed within the nonextensive theory, where also a particular nonlinear dependence of the coupling parameter and variance with scale arises for best ﬁtting theoretical PDFs. Consequently, nonlocality in ﬂuctuations, related to both, turbulence and its large scale driving, should be related to long-range interactions in the context of nonextensive entropy generalization, providing fundamentally the physical background of the observed scale dependence of ﬂuctuations in intermittent space plasmas.

1. Introduction Leptokurtic, long-tailed probability distribution functions (PDFs) subject to a non-Gaussian core and pronounced halo are a persistent feature in a variety of diﬀerent astrophysical environments. Those include 43

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

44

M. P. Leubner, Z. V¨ or¨ os and W. Baumjohann

the thermo-statistical properties of the interplanetary medium where the electron, proton and even heavy ion velocity space distributions show ubiquitously suprathermal halo patterns (see Ref. 1 for a general review or Refs. 2 and 3 and references therein), well described by the empirical family of κ-distributions, a power law in particle speed, recognized ﬁrst by Vasyliunas.4 In continuation, signiﬁcant progress was provided by Treumann5,6 who developed a kinetic theory, demonstrating that powerlaw velocity space distributions are a particular thermodynamic equilibrium state. Similarly, scale invariant power-law distributions are manifest in any systems relying on self-organized criticality (SOC)7–11 or gravitationally bound astrophysical stellar systems.12,13 Moreover, recently Leubner14 developed a theory representing accurately the hot plasma and dark matter density proﬁles in galaxies and clusters in the context of scale invariant power-law distributions. In all cases the standard Boltzmann–Gibbs–Shannon (BGS) statistics does not apply. Remarkably, we have to add to this diversity also the PDFs of the turbulent ﬂuctuations of the magnetic ﬁeld strength, density and velocity ﬁelds in space and astrophysical plasmas.15,16 In particular, the analysis of the PDFs of the solar wind plasma is of considerable interest to study intermittency and multiscale statistical properties in fully developed turbulence, since high-resolution in situ observations are available. The characteristics of the spectral properties of ﬂuctuations in the incompressible interplanetary medium were provided in classical statistical theory via the phase space distribution obtained from ideal MHD invariants by Matthaeus and Goldstein17 and followed by a conﬁrmation of the existence of solar wind multifractal structures.18,19 Furthermore, solar wind observations were also able to study the diﬀerences between ﬂuid and MHD turbulence.20 The non-Gaussianity of the PDFs of the magnetic ﬁeld and plasma ﬂuctuations was analyzed and linked to intermittency and the fractal scaling of the solar wind MHD ﬂuctuations,21,22 followed by detailed investigations of the non-Gaussian fractal and/or multifractal characteristics of solar wind and related magnetospheric parameters.23 Their multiscale coupling properties and signiﬁcance in view of magnetospheric response was analyzed,24,25 including studies of multiscale intermittency and anisotropy eﬀects in the near-Earth magnetotail dynamics.26,27 WIND, ACE, and Voyager observations of solar wind multiscale statistical properties verify that the leptokurtic, long-tailed shapes of the PDFs at small scales represent the characteristics of intermittent turbulence and approach a Gaussian, reﬂecting a decoupled state, on large scales.28–30 In other words, the probability of rare events is raised on small scales, where

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

Nonextensive Entropy and Space Plasma Turbulence

45

the spatial separation scale is characterized commonly by the diﬀerences δX(t) = X(t + τ ) − X(t), X(t) denoting any characteristic solar wind variable at time t and τ is the time lag. Recently, intermittency was considered to appear as result of an interplay between stochastic Alfv`enic ﬂuctuations and coherent two-dimensional (2D) structures.31 The empirical Castaing model32–34 introduces intermittency through ﬂuctuations of log–normal distributed variances based on the idea that for constant energy transfer between spatial scales all variables obey a Gaussian distribution of ﬂuctuations δX and hence assuming that ﬂuctuations on diﬀerent scales are independent. This convolution of Gaussians of diﬀerent variances was introduced to model the non-Gaussian energy cascade character of intermittency in turbulent ﬂows28,35–37 where the log–normal distribution of variances through the inertial scales was found to provide excellent ﬁts to the observed leptokurtic PDFs in solar wind ﬂows. Due to the ﬁtting accuracy the Castaing model achieved high popularity, but appears to be subject to two signiﬁcant shortcomings on physical grounds: the model provides (1) no link to nonlocality and long-range interactions present in turbulence and (2) no justiﬁcation for direct energy coupling between separated scales. Other known shortcomings of Castaing or log– normal models are related to the predicted features of high-order moments, which violate the fundamental assumptions of turbulence theory.38,39 The global leptokurtic non-Gaussian shape of the increment PDFs requires theoretically a corresponding unique global distribution function. This condition can be formulated on a general level by considering the basic feature of turbulent ﬂows, i.e., multiscale coupling or nonlocality in physical or in Fourier space, where nonlocality appears due to the presence of long-range forces implying direct nonlocal interactions between large scales and small-scales. Due to long-range interactions small and large scales are strongly coupled indicating that small-scale ﬂuctuations in each time/space point depend on the large scale motions in the whole time/space domain and vice versa.40 Accounting for long-range interactions is a particular feature of nonextensive systems and available from pseudo-additive entropy generalization. The classical Boltzmann–Gibbs extensive thermo-statistics constitutes a powerful tool when microscopic interactions and memory are short ranged and the environment is an Euclidean space–time, a continuous and diﬀerentiable manifold. However, in the present situation we are dealing with astrophysical systems, generally subject to spatial or temporal long-range interactions evolving in a non-Euclidean, for instance, multifractal space– time that makes their behavior nonextensive. A suitable generalization of

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

46

M. P. Leubner, Z. V¨ or¨ os and W. Baumjohann

the BGS entropy for statistical equilibrium was ﬁrst proposed by Renyi41 and subsequently by Tsallis,42 preserving the usual properties of positivity, equiprobability, and irreversibility, but suitably extending the standard extensivity or additivity to nonextensivity. The main theorems of the classical Maxwell–Boltzmann statistics admit profound generalizations within nonextensive statistics (sometimes referred to as q-statistics where q characterizes the degree of nonextensivity of the system), wherefore a variety of subsequent analyses were devoted to clarify the mathematical and physical consequences of pseudo-additivity, for an early review (see, e.g., Ref. 43). Those include a reformulation of the classical N -body problem within the extended statistical mechanics44 and the development of nonextensive distributions45,46 where a deterministic connection between the generalized entropy and the resulting power-law functionals,47 as well as the duality of nonextensive statistics were recognized.48 Relating the parameters q and κ by the transformation κ = 1/ (1 − q) Leubner49 provided the missing link between nonextensive distributions and κ-functions favored in space plasma physics, leading to the required theoretical justiﬁcation for the use of κ-distributions from fundamental physics. Since the parameter κ, a measure of the degree of nonextensivity of the system, is not restricted to positive values in the nonextensive context, the commonly observed core–halo twin character of the interplanetary electron and ion velocity space distributions was veriﬁed theoretically upon generalization to a bi-κ distribution, subject to a less pronounced core along with extended tails, as compared to a Maxwellian.50,51 Recently, the PDF of the Tsallis ensemble was linked to the analysis of fully developed turbulence providing a relation between the nonextensive parameter q and the intermittency exponent m, that is, a manifestation of multifractality of the distribution of eddies52,53 as well as of scaling of the velocity structure functions.54 Moreover, the context of generalized thermo-statistics provides analytical formulas for PDFs of distancedependent velocity diﬀerences, linking the entropic index to the cascade like structure of the turbulent dynamics.55 We relate in the following nonlocality in turbulent ﬂows to the presence of long-range forces in nonextensive systems and demonstrate in the context of entropy generalization the consistency of the theoretically derived bi-κ distribution15,50 with observed, scale-dependent PDFs of characteristic variables in the intermittent, turbulent solar wind, where both, slow and high-speed conditions are analyzed separately.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

Nonextensive Entropy and Space Plasma Turbulence

47

2. Theory The standard BGS statistics is based on the extensive entropy measure pi ln pi , (1) SB = −kB where pi is the probability of the ith microstate, kB is Boltzmann’s constant and SB is extremized for equiprobability. As physical background one assumes that particles move independently from each other, i.e., there are no correlations present in the system considered. This implies isotropy of the velocity directions and thus the entropy appears as additive quantity yielding the standard Maxwellian distribution function. In other words, microscopic interactions are short ranged and we are dealing with an Euclidean space–time. The assumptions behind standard BGS statistics are not applicable if one needs to account for nonlocality and long-range interactions in a fractal/multifractal physical environment. It is required to introduce correlation within the system, which is done conveniently in the context of nonextensive entropy generalization leading to scale-free power-law PDFs. Considering two subsystems A and B one can illuminate nonextensivity by the property of pseudo-additivity of the entropy such that 1 Sκ (A)Sκ (B) , (2) κ where the entropic index κ is a parameter quantifying the degree of nonextensivity in the system. For κ = ∞ the last term on the right-hand side cancels leaving the additive entropy of standard BGS statistics. Hence, nonlocality or long-range interactions are introduced by the multiplicative term accounting for correlations between the subsystems. As a measure for entropy mixing the entropic index κ quantiﬁes the degree of nonextensivity in the system and thus accounts for nonlocality and long-range interactions or coupling and correlations, respectively. In general, the pseudo-additive, κ-weighted term may assume positive or negative deﬁnite values indicating a nonextensive entropy bifurcation. Obviousely, nonextensive systems are subject to a dual nature since positive κ-values imply the tendency to less organized states where the entropy increases whereas negativ κ-values provide a higher organized state of decreased entropy (see Ref. 14). The general nonextensive entropy consistent with Eq. (2), replacing the classical BGS-statistics for systems subject to long-range interactions, takes the form (Refs. 42 and 50) 1−1/κ −1 . (3) pi Sκ = κkB Sκ (A + B) = Sκ (A) + Sκ (B) +

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

48

M. P. Leubner, Z. V¨ or¨ os and W. Baumjohann

In order to link the κ-notation deﬁned within −∞ ≤ κ ≤ +∞, commonly applied in space plasma modeling in terms of the family of κ-distributions, to the Tsallis q-statistics one may perform the transformation 1/(1 − q) = κ to Eq. (3).49 κ = ∞ corresponds to q = 1 and represents the extensive limit of statistical independency. Consequently, the interaction term in Eq. (2) cancels recovering with respect to Eq. (3) the classical (BGS) entropy (Eq. (1)). Equation (3) applies to systems subject to longrange interactions or memory and systems evolving in a non-Euclidean and multifractal space–time. A further generalization of Eq. (2) for complex systems, composed of an arbitrary number of mutually correlated systems, is provided by Milovanov and Zelenyi,56 where appropriate higher-order terms in the entropy appear. Once the entropy is known the corresponding probability distributions are available. In Maxwell’s derivation the velocity components of the distribution f (v) are uncorrelated where lnf can be expressed as a sum of the logarithms of the one-dimensional (1D) distribution functions. In nonextensive systems one needs to introduce correlations between the components accounting for the long-range interactions, which is conveniently done by extremizing the entropy under conservation of mass and energy yielding the corresponding 1D power-law distributions as −κ 1 v2 , (4) f ± = A± 1 + κ vt2 where vt corresponds to the mean energy or thermal speed of the distribution. Hence, the exponential probability function of the Maxwellian gas of an uncorrelated ensemble of particles is replaced by the characteristics of a power law where the sign of κ, indicated by superscripts, governs the corresponding entropy bifurcation. Note that the distribution (4) can be derived entirely using general methods of statistical mechanics without introducing a speciﬁc form for long-range interactions. Incorporating the sign of κ into Eq. (4) generates a dual solution with regard to positive and negative κ-values, respectively, resulting also in two diﬀerent normalizations A± . The entropy bifurcation appears to be manifested also in higher order moments yielding for the second moments κ-dependent generalized temperatures (see Ref. 50). Furthermore, the positive solution is restricted to κ > 3/2 whereas the negative solutions are subject to a cut-oﬀ in the distri√ bution at vmax = vt κ. Both functions, f + and f − in Eq. (4) approach the same Maxwellian as κ → ∞. Figure 1, left panel, demonstrates schematically the nonthermal behavior of both, the suprathermal halo component

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

Nonextensive Entropy and Space Plasma Turbulence

49

Fig. 1. Left panel: a schematic plot of the characteristics of the nonextensive bi-κ distribution family: with κ = 3 the outermost and innermost curve correspond to the halo fh and core fc distribution fraction. For increasing κ-values both sets of curves merge at the same Maxwellian limit, indicated as bold line, fh from outside and fc from inside. Right panel: A nonextensive bi-κ ﬁt (solid line) with κ = 1.8 of an observed PDF (dashed line) obtained from ACE magnetic ﬁeld amplitude data. A Gaussian (dasheddotted line) and a conventional κ-distribution are provided for comparison.

and the reduced core distribution, subject to ﬁnite support in velocity space, where the case κ = ∞ recovers the Maxwellian equilibrium distribution. Any unique and physically relevant nonextensive PDF must obey the following three conditions: (a) the distribution approaches one and the same Maxwellian as κ → ∞, (b) a unique, global distribution must be deﬁnable by one single density and a unique temperature, and (c) upon variation of the coupling parameter κ particle conservation and adiabatic evolution are required, such that a redistribution in a box (a source free environment) can be performed. Subject to these constraints the appropriate mathematical functional, representing observed core–halo structures in nonextensive astrophysical environments, is available from the elementary combination fch = Bch (fh + fc ), Bch being a proper normalization constant. In this context the full velocity space bi-κ distribution, compatible with nonextensive entropy generalization and obeying the above constraints, reads −κ κ N 1 v2 1 v2 1+ + 1− . (5) Fch (v; κ) = 1/2 G(κ) κ vt2 κ vt2 π vt The last term on the right-hand side denotes an expression subject to a √ thermal cutoﬀ at the maximum allowed velocity vmax = κvt , which limits also required integrations. Figure 1 (right panel) indicates also that the

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

50

M. P. Leubner, Z. V¨ or¨ os and W. Baumjohann

core of the distribution, corresponding to the last term on the right-hand side of Eq. (5), contributes only for small δB. For details regarding the corresponding second moments or generalized temperatures (see Refs. 50 and 51). The function G(κ) is deﬁned by G(κ) =

κ1/2 Γ(κ − 1/2) κ1/2 Γ(κ + 1) + Γ(κ) Γ(κ + 3/2)

−1 (6)

from the normalization and is subject to a particular weak κ-dependence where G(κ) ∼ 1/2 (see Ref. 50 for a graphical illustration and discussion). Hence, the normalization is independent of the parameter κ and the factor 1/2 reﬂects consistently the superposition of the two counter-organizing contributions in Eq. (5). For κ = ∞, G(κ) = 1/2 and the power laws in the brackets on the right-hand side of Eq. (5) turn each into the same Maxwellian exponential. With regard to this speciﬁc mathematical feature the energy levels Ei of the turbulent spectrum can be related to the corresponding kinetic energy of velocity diﬀerences δv(t) = v(t + τ ) − v(t) between two points of separation τ allowing to transform the 1D Maxwellian particle distributin of mean energy vt into the mathematical form of a Gaussian of variance σ. Upon normalizing the 1D bi-κ particle distribution (5) to unity and assigning the distribution variance σ to the thermal spread vt the “Maxwellian form” of the bi-κ distribution transforms to a “Gaussian form” of a global bi-κ PDF as −κ κ 1 δX 2 δX 2 + 1− . (7) 1+ Pch (δX; κ, σ) = √ κσ 2 κσ 2 2 πσ This two-parameter PDF is applicable to the diﬀerences of the ﬂuctuations δX(t) = X(t + τ ) − X(t) of any physical variable X in the astrophysical system considered. Here, κ assumes a physical interpretation deﬁning the degree of nonextensivity or nonlocality in the system, thus being a measure of the degree of organization or intermittency, respectively,15 and σ denotes the distribution variance. Again, depending on the particular choice of the parameters, the last term on the right-hand side is subject to a cut-oﬀ, treated by limiting the integration to the proper interval. Figure 1, left panel, illuminates also that large values of δX, corresponding to the tails of the distribution, are represented by the ﬁrst term on the right-hand side of Eq. (7). As κ → ∞ the bi-κ distribution Pch (δX; κ, σ) approaches a single Gaussian.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

Nonextensive Entropy and Space Plasma Turbulence

51

Note that Eq. (5) represents the full velocity space bi-κ distribution, while Eq. (7) is related to the time shifted diﬀerences of any physical variale X. Physically, Eq. (5) describes particles, which motion is controlled by long-range forces and interactions. The basic assumption for deriving the velocity space bi-κ distribution was the pseudo-additivity of the entropy of particle subsystems expressed through Eqs. (2) and (3). It is important to recognize that the same type of expression for bi-κ distribution is obtained, if instead of interacting particles we assume interacting coherent structures with the same pseudo-additivity property of the entropy as in Eq. (2). In the context of MHD, nonpropagating multiscale coherent structures or ﬂux tubes can interact, deform and produce new sites of nonpropagating ﬂuctuations. Coherent structures of the same polarity merge into a structure with lower local energetic state, while structures of opposite polarities may repel each other.57,58 These coherent structures can be considered as discrete interacting “particles” in MHD ﬂows, responsible for the particular entropy within the system and validating the analogy to the kinetic level of PDFs.16 A usual way of statistical analysis of intermittency due to the occurrence of coherent structures in turbulence is through two-point diﬀerences of ﬂuctuations. Therefore, in the expression of the bi-κ distribution for turbulent interactions (Eq. (7)) diﬀerences of physical variables can be used instead of the velocity notation. Moreover, passive scalars as the magnetic ﬁeld follow the dynamics of V or δV (see Ref. 59). One of the basic features of turbulent ﬂows is multiscale redistribution of energy during which interacting coherent structures appear, reducing the entropy of the system and leading also to negative κ-values. At the same time, turbulence enhances dissipation and mixing of the plasma, which increases entropy and can be described in terms of positive κ-values. The presence of processes which increase entropy and those which decrease entropy in turbulence advocates therefore, the introduction of bi-κ distributions. We proof in the following the relevance of the nonextensive, global bi-κ PDF (7) on the observed scale dependence of the PDFs of the diﬀerences of magnetic ﬁeld, velocity, and density variables in the intermittent, turbulent interplanetary medium.

3. Application: Scale Dependent Interplanetary PDFs Based on the nonextensive two parameter bi-κ distribution (7) we compare the PDFs obtained from slow and fast solar wind data with particular

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

52

M. P. Leubner, Z. V¨ or¨ os and W. Baumjohann

attention to the scale dependent changes of the two physically interpretable parameters (κ, σ) involved. We do not consider any occurrence of discontinuities or shocks in the system. The problem of interaction of turbulence with large-scale structures as shocks is investigated elsewhere.59 For each data set the magnetic ﬁeld and plasma parameter increments were calculated at a given time lag τ by δX(t) = X(t + τ ) − X(t), where X represents any physical variable considered. For each realization the empirical probability distribution function (histogram) was then computed. δX(t) is binned into n equal spaced boxes and the number of elements in each box was computed where the robustness of the histograms against n is tested. δX(t) represents characteristic ﬂuctuations at the time scale τ or, equivalently, across eddies of size l ∼ vτ . Hence, by changing τ it is possible to analyze the statistical features of ﬂuctuations in diﬀerent time scales, which roughly correspond to those statistical characteristics across turbulent eddies of size l ∼ vτ . In the following we demonstrate by means of Eq. (7) from ﬁrst principle statistics that the strong non-Gaussianity of the PDF of small-scale ﬂuctuations should be associated physically with long-range interactions provided in nonextensive systems by pseudo-additive entropy generalization. The scale dependence of the PDF in the solar wind can be represented accurately via the tuning parameters κ and σ of the bi-κ functional. In Fig. 1, right panel, we focus on the bi-κ model demonstrating that the nonextensive context, generates a precise representation for the observed solar wind PDF characterizing the intermittency of the small-scale ﬂuctuations. For comparison also a Gaussian and the conventional κ-function,49 subject to the same κ-value but not able to reproduce the structure of the PDF of small-scale ﬂuctuations, are provided. It is also possible to assume that the smallest ﬂuctuations are random and uncorrelated, and this might be the reason why the central part of the distribution near the maximum is well-ﬁtted by a Gaussian, as in the right panel in Fig. 1. Note, however, that the characteristic scale of the diﬀerences of ﬂuctuations is introduced through the time-lag τ . Therefore, δB → 0 near the maximum of the distribution simply means that, at the scale τ the diﬀerences between the corresponding values of B(t + τ ) and B(t) are very small. Combined with the ﬁnite precision of PDFs estimation, small two-point diﬀerences can really produce uncorrelated ﬂuctuations near PDFs maxima. This does not mean, however, that the Gaussian distribution is the correct answer for the proper description of the central part of the distribution at the small scales τ . If it was true, we should suppose that, at smaller and smaller scales τ , the

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

Nonextensive Entropy and Space Plasma Turbulence

53

ﬂuctuations become more and more random and we are getting closer and closer to a Gaussian distribution. Actually the opposite is true, the peakedness of the distribution increases as the scale τ decreases (see later). 3.1. Slow speed solar wind As an example, in Fig. 2 undisturbed solar wind ACE magnetic ﬁeld amplitude data of 16 s time resolution are analyzed where the dimensionless τ is multiplied by the resolution to generate an eﬀective time-lag. In particular, the scale dependent PDF evolution of magnetic ﬁeld ﬂuctuations is subject to a two-point separation scale of τ = 100, 2000, and 10 000. The corresponding best ﬁts of the bi-κ distribution are obtained for κ = 1.8, 3, and ∞, measuring the degree of nonextensivity, or coupling, respectively, through long-range interactions and the dotted lines refer to the standard deviation. The accuracy of the bi-κ distribution ﬁt demonstrates that nonlocality in turbulence, when introduced theoretically by long-range interactions through the nonextensive context, generates a precise representation for the observed PDFs characterizing the intermittency of the ﬂuctuations at all scales. Based on WIND velocity ﬁeld magnitude data Fig. 3 presents an analysis of the scale dependence of interplanetary PDFs of the velocity ﬁeld magnitude showing in three plots from left to right the decreasing kurtosis with increasing time-lags, where the observational uncertainty is again indicated by the standard deviation (dotted lines). Constraint by a time resolution of 92 s the two-point time separation τ assumes the values 10, 70, and 900

Fig. 2. The PDF of the increments of observed ACE magnetic ﬁeld ﬂuctuations for τ = 100 and a resolution of 16 s as compared to the bi-κ function with κ = 1.8. Based on the same data, the central panel provides the characteristics for increased τ = 2000, where κ assumes a value of 3.0 for the best representation. The PDF of large-scale magnetic ﬁeld ﬂuctuations, τ = 10 000, are well modeled by a Gaussian with κ = ∞, right panel. The dotted lines correspond to the standard deviations.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

54

M. P. Leubner, Z. V¨ or¨ os and W. Baumjohann

Fig. 3. The PDF of the increments of observed WIND velocity ﬁeld magnitude ﬂuctuations for τ = 10 and a resolution of 92 s as compared to the bi-κ function with κ = 2. Based on the same data, the central panel provides the characteristics for increased τ = 70, where κ assumes a value of 3.5 for the best representation. The PDF of largescale velocity ﬂuctuations are well modeled by a Gaussian with κ = ∞, right panel. The dotted lines correspond to the standard deviations.

from left to right. The solid lines represent best ﬁts to the observed PDFs with nonextensive distributions, where the corresponding κ is determined as κ = 2, 3.5, and ∞. The strong non-Gaussian character of the leptokurtic PDFs (left panel), exhibiting pronounced tails associated with solar wind turbulence and intermittency in small-scale ﬂuctuations, ﬁnds again an accurate analytical ﬁt and hence a physical background in the nonextensive representation. The non-Gaussian structure is somewhat softened for enhanced τ = 70 (central panel) but again precisely modeled within the pseudo-additive entropy context, turning into the Gaussian shape of large scale ﬂuctuations, which is independent of the increment ﬁeld. Figure 4 provides the corresponding nonextensive analysis of the scale dependence of the density ﬂuctuations obtained from WIND data (92 s resolution). Evidently, the scale-dependent characteristics of the observed PDFs of the increment ﬁelds δX(τ ) = X(t + τ ) − X(t) for all solar wind variables evolve simultaneously on small-scales approaching independency of the increment ﬁeld in the large-scale Gaussian. With separation scales of τ = 10, 70, and 900 the corresponding evolution of the PDFs of observed density ﬂuctuations are best represented by the same values of κ = 2, 3.5, and ∞, as for the velocity ﬁeld magnitude, accounting for nonlocal interactions in the nonextensive theoretical approach. 3.2. High-speed solar wind The four panels in Fig. 5 show PDFs of high-speed associated magnetic ﬁeld magnitude ﬂuctuations. The two-point statistics is demonstrated in

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

Nonextensive Entropy and Space Plasma Turbulence

55

Fig. 4. Left panel: the PDF of the increments of observed WIND density ﬂuctuations for τ = 10 and a resolution of 92 s as compared to the bi-κ function with κ = 2. Based on the same data, the central panel provides the characteristics for increase τ = 70, where κ assumes a value of 3.5 for the best representation. The PDF of large-scale density ﬂuctuations with τ = 900 are well modeled by a Gaussian with κ = ∞, right panel.

Fig. 5. The PDF of the increments of observed ACE high-speed associated magnetic ﬁeld magnitude ﬂuctuations (16 s time resolution). Top-left: ﬂuctuations at the scale τ = 10 as compared to the bi-κ function with κ = 1.4 and σ = 0.05; top-right: τ = 40, κ = 1.4, and σ = 0.12; left-bottom: τ = 400, κ = 1.2, and σ = 0.6; left-right: τ = 10 000, κ = ∞, and σ = 15 (Gaussian ﬁt).

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

56

M. P. Leubner, Z. V¨ or¨ os and W. Baumjohann

the subplots from top-left to bottom-right for the scales τ = 10, 40, 400, and 10 000. The eﬀective time-lag is obtained after multiplying τ with time resolution of 16 s. The corresponding best ﬁts reveal diﬀering statistical features of high-speed associated magnetic ﬂuctuations. In comparison with low-speed data the degree of nonextensivity does not change during highspeed intervals, κ = 1.4, 1.4, and 1.2, over the range of scales τ = 10, 40, 400, and only for τ = 10 000, κ reaches ∞. In contrast to the slow wind data, however, good quality high-speed ﬁts can be achieved only when the standard deviation σ is changed. This indicates that the abundance of large-scale energy content of highspeed ﬂows may facilitate to maintain the degree of nonextensivity and self-organization unchanged over the considered scales. On the other hand, we have no clear explanation yet for the observed changes of the standard deviation. We can speculate that changes in sigma appear because of the ample changes in the amplitudes of two-point ﬂuctuations in the solar wind, having solar origin or being generated by local processes absent in the slow wind. Obviously, further comparative case studies on multiscale ﬂuctuations are needed to answer the question of the relative contribution of local processes versus processes originating in the solar corona to the observed behavior of statistical moments in the fast and slow solar wind. The four panels in Figs. 6 and 7 provide the same qualitative behavior for high-speed associated ACE density (64 s resolution) and WIND magnetic ﬁeld (3 s resolution).

4. Discussion and Conclusions The WIND and ACE solar wind data analysis unambiguously manifests that the PDFs of large scale density, velocity, and magnetic ﬁeld ﬂuctuations are well represented by a Gaussian, turning into leptokurtic peaked distributions of strong non-Gaussianity in the center along with a pronounced tail structure at smaller scales. In particular, the PDFs of large-scale magnetic ﬁeld ﬂuctuations, not related to the increment ﬁeld are known to be subject to relatively small deviations from the Gaussian statistics and are well ﬁtted by the Castaing distribution, a convolution of Gaussians with variances distributed according to a log–normal distribution.32,60 Assuming a constant energy transfer rate between spatial scales all quantities exhibit a Gaussian distribution of ﬂuctuations in this context. Independent of the physical situation considered, the Castaing distribution provides a multiparameter description of observed PDFs, plausible in this case, since

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

Nonextensive Entropy and Space Plasma Turbulence

57

Fig. 6. The PDF of the increments of observed ACE high-speed associated density ﬂuctuations (64 s time resolution). Top-left: ﬂuctuations at the scale τ = 10 as compared to the bi-κ function with κ = 1.5 and σ = 0.3; top-right: τ = 40, κ = 1.5, and σ = 0.5; left-bottom: τ = 400, κ = 1.05, and σ = 0.9; left-right: τ = 3000, κ = ∞, and σ = 25 (Gaussian ﬁt).

the large-scale ﬂuctuations of the interplanetary magnetic ﬁeld are generated by a variety of discrete coronal sources. If individual coronal sources evoke Gaussian distributed magnetic ﬁelds, the net magnetic ﬂuctuations can be modeled by their superposition with a spread of the corresponding variances. Contrary, small-scale ﬂuctuations are associated with local intermittent ﬂows where ﬂuctuations are concentrated in limited space volumes. Consequently, the PDFs are scale dependent and intermittency generates long-tailed distributions. It is customary to use nth-order absolute powers of the plasma variables and magnetic ﬁeld increments (nth-order

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

58

M. P. Leubner, Z. V¨ or¨ os and W. Baumjohann

Fig. 7. The PDF of the increments of observed ACE high-speed associated magnetic ﬁeld magnitude ﬂuctuations (3 s time resolution). Top-left: ﬂuctuations at the scale τ = 10 as compared to the bi-κ function with κ = 1.4 and σ = 0.003; top-right: τ = 40, κ = 1.4, and σ = 0.006; left-bottom: τ = 400, κ = 1.8, and σ = 0.06; left-right: τ = 4000, κ = ∞, and σ = 1 (Gaussian ﬁt).

structure functions)22,61 allowing to investigate the multiscale scaling features of ﬂuctuations. Direct studies of observed PDFs of the increment ﬁelds δX(t) = X(t + τ ) − X(t) for any characteristic solar wind variable at time t and time lag τ revealed departures from a Gaussian distribution over multiple scales28 and an increase of kurtosis (intermittency) towards small scales.21 The PDFs are also found to be leptokurtic, which indicates the turbulent character of the underlying ﬂuctuations. Sorriso-Valvo et al.28 have shown that the non-Gaussian behavior of small-scale velocity and magnetic ﬁeld ﬂuctuations in the solar wind can also be described well by a Castaing

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

Nonextensive Entropy and Space Plasma Turbulence

59

distribution where the individual sources of Gaussian ﬂuctuations appear at small-scales in turbulent cascades. From the corresponding nonextensive WIND data analysis of the density and magnetic ﬁeld ﬂuctuations,15 it is evident that the scale dependent characteristics of the observed PDFs of the increment ﬁelds δX(t) = X(t + τ ) − X(t) for solar wind variables evolve simultaneously on small scales, approaching independency of the increment ﬁeld in the large-scale Gaussian. Highly accurately, the overall scale dependence appears as a general characteristic of quiet astrophysical plasma environments indicating a universal scaling dependence between density, velocity, and magnetic ﬁeld intermittency within the experimental uncertainties. This strong correlation implies that the scale dependencies of all physical variables are coupled, where the solar wind Alfv´enic ﬂuctuations provide a physical basis of the velocity and magnetic ﬁeld correlations. On the other hand, according to recent analyses, the magnetic ﬁeld intensity exhibits a higher degree of intermittency than the solar wind bulk velocity, both in fast and slow winds.62 However, Veltri and Mangeney63 found that the most intermittent structures in the slow wind are shock waves, displaying similar intermittency in the magnetic ﬁeld intensity and bulk velocity. Furthermore, the proportionality between density ﬂuctuations and the magnetic ﬁeld and velocity ﬂuctuations is already maintained in the solar wind by the presence of weak spatial gradients.64 Figure 8 (left panel) provides an estimation of the functional dependence between the time-lag τ and the nonextensive parameter κ for best ﬁtting bi-κ functions to the observed PDFs for slow speed solar wind conditions

Fig. 8. The functional dependence of the spatial scale τ (κ) (left panel) for slow speed and τ (σ) (right panel) for high-speed conditions.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

60

M. P. Leubner, Z. V¨ or¨ os and W. Baumjohann

(v ≤ 400 km/s). As signiﬁcant global behavior the scale dependence of the PDFs for quiet conditions appears to be independent of the variance or mean energy of the distribution. The best ﬁtting bi-κ functions are found for a constant corresponding parameter σ, indicating that the parameter κ, measuring the degree of coupling within the system, governes primarily the scale dependence of the PDF in the slow solar wind. As τ increases from small scales to the intermediate regime a pronounced plateau formation is established, i.e., the relative increase in κ-values with enhanced scales appears reduced. In other words, the PDF shape appears at intermediate scales to be independent of the spatial separation scale. Such a behavior may indicate the presence of a transitional dynamical element characterizing a balance between long- and short-range interactions. Contrary, in high-speed streams (v ≥ 400 km/s) best ﬁts of bi-κ functions to the observed scale dependent PDFs are found when keeping κ constant and varying only the parameter σ, right panel in Fig. 8. Hence the scaling features in the fast wind appear independent of the degree of coupling controlled by κ, but rely predominantly on changes of the variance or mean energy. In summary, the theoretical nonextensive context indicates a signiﬁcant and physically contrary scaling behavior in slow and fast wind. For slow solar wind conditions the correlations/intermittency governed by κ decrease with increasing scale, whereas the characteristic energy governed by σ remains constant. Contrary the scaling properties in high-speed streams are characterized by constant correlations (κ) but enhanced variance with increasing scale. κ-distributions reproduce the Maxwell–Boltzmann distribution for κ → ∞, a situation identifying κ as an ordering parameter that acounts for correlations within the system. Highly correlated turbulent conditions characterized by κ-distributions represent stationary states far from equilibrium where a generalization of the Boltzmann–Shannon entropy, as measure of the level of organization or intermittency, applies.65,66 Physically this can be understood considering a system at a certain nonlinear stage where turbulence may reach a state of high-energy level that is balanced by turbulent dissipation. In this environment equilibrium statistics can be extended to dissipative systems, approaching a stationary state beyond thermal equilibrium.67 Since turbulence is driven in the solar wind by velocity shears we have choosen for the data analysis intervals of low-speed solar wind with limits in velocity space, where the driving and dissipation conditions do not change signiﬁcantly, maintaining therefore the dynamical equilibrium condition. High-speed intervals may depend more on the occurence of dynamical processes on the Sun.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

Nonextensive Entropy and Space Plasma Turbulence

61

A multiscale cascade mechanism is not the only way for a realization of long-range interactions. Let us provide a physical situation where large and small scales are directly coupled and the nonlocal energy transfer is not induced by cascading processes, following, e.g., a log–normal model.68 An example is found in the context of MHD where Chang57 and Consolini and Chang69 proposed an intermittent turbulence model for the solar wind and for the Earth’s magnetotail, which comprises neither cascades nor requires local interactions in Fourier space. In this scenario nonpropagating or convected ﬂuctuations generate multiscale coherent structures (e.g., ﬂux tubes), which can interact, deform, and produce new sites of nonpropagating ﬂuctuations. Coherent structures of the same polarity merge into a structure with lower local energetic state, while structures of opposite polarities may repel each other. This coherent structures can be considered as discrete interacting “particles” in MHD ﬂows, responsible for the particular entropy within the system and validating the analogy to the kinetic level of PDFs. Chang et al.70 have computed PDFs of the intermittent ﬂuctuations from direct numerical simulations of interacting coherent structures. The resulting PDFs have typical leptokurtic shapes, which can be well ﬁtted again by a variety of models, including those with predominantly local interactions in Fourier space. Since all models provide similar ﬁtting accuaracy it is required to focus on the underlying physical situation in turbulent ﬂows. Hence, with regard to nonlocal interactions not based on cascade processes the nonextensive entropy approach provides physically a justiﬁcation for nonlocal interactions and should therefore, be favored over cascade models in such processes. The entropy quantiﬁes the degree of structuring in intermittent turbulence expressed through singular multifractal measures, where also the parameter κ (or q) is related to the extremes of multifractal distributions.52,53,55,71 Since the nonextensive entropy approach is independent of the mechanism leading to the structures — in both situations, cascading processes and multiscale interacting coherent structures — or even in coexisting situations, the entropy concept can be applied for the analysis and quantiﬁcation of resulting characteristics in turbulence. In summary, the majority of hitherto existing models of intermittency in the solar wind essentially correspond to the cascade picture of turbulence. Smallscale intermittency, however, can be associated also by emerging topological complexity of coherent sturctures in turbulence, which might be understood better through entropy concepts, disregarding the goodness criteria of different ﬁts.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

62

M. P. Leubner, Z. V¨ or¨ os and W. Baumjohann

We provided speciﬁc examples based on multiscale interacting coherent structures where the traditional cascade — and thus, e.g., the log–normal approach beside others — should not be applied for physical reasons (not in terms of PDF ﬁtting accuracy). Therefore, the proposed context based on entropy ganeralization has a potential to describe the underlying physics suitably and thus justiﬁes the nonextensive approach. Certainly this must be viewed as an incomplete concept in view of the complexity of intermittence in turbulence,72 in particular regarding the interactions with large scale structures. Summarizing, a bi-κ distribution family turns out theoretically as consequence of the entropy generalization in nonextensive thermo-statistics. The two-parameter global bi-κ function provides theoretically access to the scale dependence of the PDFs observed in astrophysical plasma turbulence. The redistribution of a Gaussian on large-scales into highly non-Gaussian leptokurtic and long-tailed structures, manifest on small scales, is theoratically well described by the family of nonextensive distributions. Pseudo-additive entropy generalization provides the required physical interpretation of the parameter κ in terms of the degree of nonextensivity of the system as a measure of nonlocality or coupling due to long-range interactions whereas the variance σ measures the mean energy in the system. The scale dependence in the slow speed solar wind is sensitive to variations of κ and in high-speed streams to variations of σ. We argue that multiscale coupling and intermittency of the turbulent solar wind ﬂuctuations must be related to the nonextensive character of the interplanetary medium accounting for long-range interaction via the entropy generalization. References 1. D. A. Mendis and M. Rosenberg, Ann. Rev. Astron. Astrophys. 32 (1994) 419. 2. M. P. Leubner, Planet. Space Sci. 48 (2000) 133. 3. M. P. Leubner and N. Schupfer, J. Geophys. Res. 106 (2001) 12993. 4. V. M. Vasyliunas, J. Geophys. Res. 73 (1968) 2839. 5. R. A. Treumann, Physica Scripta 59 (1999) 19. 6. R. A. Treumann, Physica Scripta 59 (1999) 204. 7. P. Bak, How Nature Works the Science of Self-organized Criticality (Copernicus, New York, 1996). 8. P. Bak, C. Tang and K. Wiesenfeld, Phys. Rev. A 38 (1988) 364. 9. N. W. Watkins, M. P. Freeman, S. C. Chapman and R. O. Dendy, J. Atmosph. Solar-Terrestr. Phys. 63 (2001) 1435. 10. S. C. Chapman, N. W. Watkins, R. O. Dendy, P. Helander and G. Rowlands, Geophys. Res. Lett. 25 (1998) 2397.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

Nonextensive Entropy and Space Plasma Turbulence

63

11. S. C. Chapman and N. W. Watkins, Space Sci. Rev. 95 (2001) 293. 12. A. Nakamichi, I. Joichi, O. Iguchi and M. Morikawa, Chaos, Solitons and Fractals 13 (2002) 595. 13. P. H. Chavanis and F. Bouchet, Astron. Astrophys. 430 (2005) 771. 14. M. P. Leubner, Astrophys. J. 632 (2005) L1. 15. M. P. Leubner and Z. V¨ or¨ os, Astrophys. J. 618 (2005) 547. 16. M. P. Leubner and Z. V¨ or¨ os, Nonlin. Proc. Geophys. 12 (2005) 171. 17. W. H. Matthaeus and M. L. Goldstein, J. Geophys. Res. 87 (1982) 6011. 18. L. F. Burlaga, Geophys. Res. Lett. 18 (1991) 1651. 19. L. F. Burlaga, J. Geophys. Res. 97 (1992) 4281. 20. V. Carbone, P. Veltri and R. Bruno, Phys. Rev. Lett. 75 (1995) 3110. 21. E. Marsch and C.-Y. Tu, Ann. Geophys. 12 (1994) 1127. 22. E. Marsch and C.-Y. Tu, Nonlin. Proc. Geophys. 4 (1997) 101. 23. B. Hnat, S. C. Chapman, G. Rowlands, N. W. Watkins and M. P. Freeman, Geophys. Res. Lett. 29 (2003) 2078. 24. Z. V¨ or¨ os, D. Jankoviˇcov´ a and P. Kov´ acs, Nonl. Proc. Geophys. 9 (2002) 149. 25. Z. V¨ or¨ os and D. Jankoviˇcov´ a, Nonl. Proc. Geophys. 9 (2002) 425. 26. Z. V¨ or¨ os et al., Ann. Geophys. 21 (2003) 1955. 27. Z. V¨ or¨ os et al., Phys. Plasmas 11 (2004) 1333. 28. L. Sorriso-Valvo, V. Carbone and P. Veltri, Geophys. Res. Lett. 26 (1999) 1801. 29. L. F. Burlaga, J. D. Richardson and C. Wang, J. Geophys. Res. 107, A10 (2002) SSH20. 30. L. F. Burlaga, J. D. Richardson and C. Wang, J. Geophys. Res. 107, A11 (2002) SSH18. 31. R. Bruno, V. Carbone, P. Veltri, E. Pietropaolo and B. Bavassano, Planetary Space Sci. 49 (2001) 1201. 32. B. Y. Castaing, Y. Gagne and E. J. Hopﬁnger, Physica D 46 (1990) 177. 33. B. Y. Castaing and B. Dubrulle, J. Phys. II France 5 (1995) 895. 34. B. Y. Castaing, J. Phys. II France 6 (1996) 105. 35. G. Consolini and P. De Michelis, Geophys. Res. Lett. 25 (1998) 4087. 36. G. He, B. Dubrulle and F. Graner, Phys. Lett. A 245 (1998) 419. 37. F. Schmitt and D. Marsan, Eur. Phys. J. B 20 (2001) 3. 38. U. Frisch, Turbulence (Cambridge University Press, Cambridge, 1995). 39. R. Arratia, L. Goldstein and L. Gordon, Stat. Sci. 5 (1990) 403. 40. A. Tsinober, An Informal Introduction to Turbulence (Kluwer Academic Publishers, Dodrecht, 2001). 41. A. Renyi, Acta Math. Hungaria 6 (1955) 285. 42. C. Tsallis, J. Stat. Phys. 52 (1988) 479. 43. C. Tsallis, Physica A 221 (1995) 277. 44. A. R. Plastino, A. Plastino and C. Tsallis, J. Phys. A: Math. Gen. 27 (1994) 5707. 45. R. Silva, A. R. Plastino and J. A. S. Lima, Phys. Lett. A 249 (1998) 401. 46. M. P. Almeida, Physica A 300 (2001) 424. 47. J. S. Andrade, M. P. Almeida, A. A. Moreira and G. A. Farias, Phys. Rev. E 65 (2002) 036121.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch04

64

48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63.

64. 65. 66. 67. 68. 69. 70. 71. 72.

M. P. Leubner, Z. V¨ or¨ os and W. Baumjohann

I. V. Karlin, M. Grmela and A. N. Gorban, Phys. Rev. E 65 (2002) 036128. M. P. Leubner, Astrophys. Space Sci. 282 (2002) 573. M. P. Leubner, Astrophys. J. 604 (2004) 469. M. P. Leubner, Phys. Plasmas 11 (2004) 1308. T. Arimitsu and N. Arimitsu, Phys. Rev. E 61 (2000) 3237. T. Arimitsu and N. Arimitsu, J. Phys. A: Math. Gen. 33 (2000) L235. T. Arimitsu and N. Arimitsu, Prog. Theor. Phys. 105 (2001) 335. C. Beck, Physica A 277 (2000) 115. A. V. Milovanov and L. M. Zelenyi, Nonlin. Proc. Geophys. 7 (2000) 211. T. Chang, Phys. Plasmas 6 (1999) 4137. R. Bruno and V. Carbone, Living Rev. Solar Phys. 2 (2005) 1. Z. V¨ or¨ os, M. P. Leubner and W. Baumjohann, J. Geophys. Res. 111 (2006) A02102, doi:10.1029/2005JA011479. N. S. Padhye, C. W. Smith and W. H. Mathaeus, J. Geophys. Res. A106 (2001) 18635. C. Pagel and A. Balogh, Nonlin. Proc. Geophys. 8 (2001) 313. L. Sorriso-Valvo, V. Carbone, P. Giuliani, P. Veltri, R. Bruno, V. Antoni and E. Martines, Planet. Space Sci. 49 (2001) 1193. P. Veltri and A. Mangeney, Solar Wind 9, eds. S. R. Habbai, R. Hesser, J. V. Hollweg and P. A. Isenberg (American Institute of Physics, 1999), p. 543. S. R. Spangler, Nonlin. Proc. Geophys. 21 (2003) 113. S. Goldstein and J. L. Lebowitz, Physica D 193 (2004) 53. R. A. Treumann, Phys. Plasmas 11 (2004) 1317. T. Gotoh and R. H. Kraichnan, Physica D 231 (2002) 244. U. Frisch, Turbulence (Cambridge University Press, Cambridge, 1996). G. Consolini and T. S. Chang, J. Atm. Sol. Terr. Phys. 64 (2002) 541. T. Chang, S. W. Y. Tam and Ch. Wu, Phys. Plasmas 6 (2004) 4137. M. L. Lyra and C. Tsallis, Phys. Rev. Lett. 80 (1998) 53. E. G. D. Cohen, Physica D 193 (2004) 35.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch05

SIMULATION OF INTERPLANETARY SHOCK WAVE CAUSED BY CME ON AUGUST 25, 2001 TOMOYA OGAWA∗ , MITSUE DEN† , TAKASHI TANAKA‡ and KAZUYUKI YAMASHITA§ ∗Space

Simulation Group, National Institute of Information and Communications Technology, Tokyo 184-8795, Japan †Computer and Information Network Center National Institute for Fusion Science, Japan

‡Department §The

of Earth and Planetary Sciences, Kyushu University, Japan Center for Educational Research, University of Yamanashi, Japan

We simulated an interplanetary shock wave caused by a coronal mass ejection (CME) on August 25, 2001. A three-dimensional adaptive mesh reﬁnement code was used and calculated propagation of the shock wave. A CME model was input into inner boundary of simulation, and ﬂuctuation of solar wind at the position of ACE spacecraft was reproduced. The passage times of observed and simulated shock waves was almost the same. The predicted shock amplitude in density and velocity was consistence with observation. The temporal proﬁle of ﬂuctuation of density was well reproduced. In post-shock region, the velocity ﬂuctuation had some enhancement. We show that the CME has the angular diameter of about 100◦ .

1. Introduction Prediction of passage time of interplanetary shock waves is important for space weather. Numerical simulation is a powerful tool for the prediction. There are some factors which makes it diﬃcult to execute such simulations. A large range of scales between the sun and the orbit of the earth is one of diﬃculties. The distance between the sun and the earth is 1.5 × 1011 m, that is 215 times as long as the solar radius of 7.0 × 108 m. The shock waves have small scales in a short time after they arise, then they expand into interplanetary space and eventually get large extent. Ones who research interplanetary shock waves by using numerical simulations make eﬀorts to overcome the hard conditions. A spherical mesh is one of solutions. For example, Odstrcil and Pizzo1 and Odstrcil et al.2 used a fan-shaped spherical mesh to investigate propagations of interplanetary shock waves. The spherical mesh has a good point. Its mesh intervals in horizontal direct are 65

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch05

66

T. Ogawa et al.

short near the center and becoming smoothly longer in more distant region from the center. On the other hand, the spherical mesh has weak points. The ﬁrst good point above contains an opposite weak point in resolving tilted structures in a distant region. And its longitudinal interval is becoming shorter near the polar line, which is unfavorable feature when one is not very interested in high latitudes. The use of an adaptive mesh reﬁnement (AMR) technique is another solution. The AMR technique adapts dynamically the meshes to suit the physical state and is able to monitor running structures by ﬁne meshes. Only coarse meshes is used on quiet regions. This feature is an advantage in investigating interplanetary shock waves. A weak point of AMR is that artiﬁcial waves may arise when disturbances places at a boundary between large and small cells. But one can avoid such a situation by keeping the boundary distant from any large disturbances. Manchester et al.3 performed AMR simulations of interplanetary coronal mass ejection (CME) propagation. They, however, localized their ﬁne meshes for a shock wave to a region along the sun-to-earth line. We choose the AMR technique in this paper. We do not restrict ﬁne meshes onto the sun-earth line but spread these over a shock wave surface. We simulate a CME that occurred on August 25, 2001. Resulting time ﬂuctuations of density and velocity are compared with data observed by the ACE spacecraft ﬂying at the ﬁrst Lagrange point, that is, at 1.5 × 109 m from the earth toward the sun. In Sec. 2, we outline our three-dimensional (3D) AMR simulation code and models for solar wind and a CME. Results are shown in Sec. 3 and summarized in Sec. 4.

2. Simulation and Models We use 3D AMR simulation code. In this code, a mesh structure is managed by a fully threaded tree.4 Eight cells are bundled up into a unit block, and the blocks are assembled to a tree structure. Decision whether cells should be split/joined or not is according to reﬁnement indicators. We use two indicators as follows: 1 if dx(j)/r(j) > tan(9◦ ) , ξr (j) = 0 if dx(j)/r(j) ≤ tan(9◦ ) , 1 if max(vr (i1 ) − vr (i2 )) > 40 km/s , ξv (j) = 0 if max(vr (i1 ) − vr (i2 )) ≤ 40 km/s , max(ξr (j), ξv (j)) , ξ(j) = 27 blocks

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch05

Simulation of Interplanetary Shock Wave Caused by CME on August 25, 2001

67

where j is index of a block, r is distance from the solar center, dx is side length of a cell, vr is radial velocity, i1 and i2 represent all combinations of adjacent two cells in the block, and represents the sum of the block j and 26 surrounding blocks (see Refs. 5–7) for more information of our AMR code). In this study, we aim dynamics of the CME. There is relatively weak magnetic ﬁeld and kinetic energy of a CME-caused shock wave is dominant in the region distant from the sun. We therefore adopt hydrodynamic approximation. The equations for the system are ∂ρ + ∇ · (ρu) = 0 , ∂t ∂(ρu) + ∇ · (ρuu) = ρg , ∂t ∂e + ∇ · (e + P )u = ρg · u + (γ − 1)Q , ∂t

(1)

where ρ, u, g, e, P , γ, and Q are the mass density, velocity, gravitational acceleration, total energy density, pressure, speciﬁc heat ratio, and heating function, respectively. The speciﬁc heat ratio of γ = 5/3 and point source gravity g = −(GM /r3 )r are set, where G, M , r, and r ≡ |r| are the gravitational constant, the solar mass, the position vector from the solar center, and the distance from the solar center, respectively. We set the heating function as (r − R )2 , Q = ρq0 (T0 − T ) exp − σ02

(2)

where T , r, and R are the temperature, distance from the sun, and the solar radius, respectively. We set values σ0 = 4.5 R and q0 = 106 erg/g/s/K, which are the value used in Ref. 3. We choose T0 = 1.8 × 106 K and ρ = 3 × 106 protons/cm3 to reproduce the velocity and the density of the pre-shock solar wind observed by the ACE spacecraft at point L1. The equations are solved by the 3rd order accurate Roe-MUSCL algorithm.8–11 We reserve 1.6 × 107 cells for the simulation. A needed number of cells are distributed over a simulation box. The simulation box is a cube whose side length is (500 R )3 = (2.3 AU).3 The sun was positioned at the center of the box. The inner boundary places at a height of 0.15 R above the solar surface. The cell size was (0.12 R)3 near the sun and (7.8 R )3 in

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch05

68

T. Ogawa et al.

the coarsest-mesh regions. The cell size on the shock surface is (0.24 R )3 at a beginning, and then it is enlarged to twice before the reserved cells are used up. The ﬁnal cell size on the shock surface is (0.98 R )3 . We ﬁrst obtain a steady-state solar wind and then input the model CME into the inner boundary. In general, solar wind is not homogeneous nor stationary. But we introduce homogeneous and stationary solar wind, because variation of the solar wind on which the shock wave has underwent is not be observed or it is diﬃcult to be inferred. In ACE’s data, the solar wind at this time seem to be relatively quiet. It may indicate that our assumption is not so bad. The CME model is described as

πξ , A(ξ) = cos 2ξ0 t  , (0 < t < τr ) ,     τr (3) B(t) = 1, (τr < t < τr + τd ) ,    τr + τd + τa − t   (τr + τd < t < τr + τd + τa ) , τa V (t, ξ) = Vmax A(ξ)B(t) , where Vmax is the maximum velocity of a CME, ξ is angle to the axis of a CME cone, ξ0 is angular radius of a CME, and τr , τd , and τa are standup time, duration time, and attenuation time of a CME, respectively. The model resembles Odstrcil and Pizzo1 in the velocity function, but diﬀers in others. In our model, the input density is constant and the eruption time is decided by observed emission of the associated X-ray ﬂare described below. A CME occurred at 16:50 UT on August 25, 2001. The SOHO LASCO CME Catalog estimates that velocity of the CME is 1327 km/s at 20 R . We set Vmax of 1772 km/s in the formula (3), that reproduces the observed velocity in our model. An associated X5.3 X-ray ﬂare is detected at S17E34. We assume that the CME occurs just above the X-ray ﬂare. We adopt ξ0 = 50◦ , τr = τd = 0.1 h, and τa = 1 h. These values well reproduced behaviour of the CME event in traveling time, time variation of density, and peak of velocity observed by the ACE spacecraft. 3. Results We observed ﬂuctuation of solar wind by an “imaginary” spacecraft at the ﬁrst Lagrange point in the simulation. Time plots of resulting density and velocity are displayed with ACE’s data in Figs. 1 and 2, respectively.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch05

Simulation of Interplanetary Shock Wave Caused by CME on August 25, 2001

69

22 density

20

simulation observation

18 16 14 12 10 8 6 4 2

239.5

240

240.5

241

Fig. 1. Density ﬂuctuation at the ﬁrst Lagrange point for two days from 07:19 on August 27. Units of the vertical and the horizontal axes are protons/cm 3 and day of year, respectively. Solid line is simulation result and dashed line is ACE’s observation data.

600 velocity

580

simulation observation

560 540 520 500 480 460 440 420 400

239.5

240

240.5

241

Fig. 2. Velocity ﬂuctuation at the ﬁrst Lagrange point for two days from 07:19 on August 27. Units of the vertical and the horizontal axes are km/s and day of year, respectively. Solid line is simulation result and dashed line is ACE’s observation data.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch05

70

T. Ogawa et al.

Observed passage time of the shock wave is at 19:19 UT on August 27, and predicted one is 19:21 UT on that day. Traveling times from the sun to the spacecraft are 50.5 h. Degrees of jump at the shock surface show good agreements with observation in both density and velocity. Observed number density of proton is 6/cm3 in pre-shock region, the peak is 20/cm3 , and then damps to 3/cm3 . The resulting density jumps from 6/cm3 to 19/cm3 , it agrees with observation, and then damps down to about 2/cm3 . The damping is some slower than observation. Observed velocity is 450 km/s just before the shock wave arrives, reaches about 570 km/s and maintains the high speed for 10 h, and then drops to 470 km/s. The resulting velocity jumps from 450 to 560 km/s, maintains the high speed for about 17 h, and then drops to 490 km/s. The jumping altitude shows good agreement with observation, but high speed is prolonged in the post-shock region.

4. Summary We simulated propagation of interplanetary shock wave raised by the CME occurred at 16:50 UT on August 25, 2001. The AMR technique make it possible to resolve solar corona and a shock wave with ﬁne meshes in a simulation box concluding the orbit of the earth. The resolution of the shock surface is 1/2048 at the beginning and 1/512 at the ﬁnal. The CPU time is 3.7 h when the shock wave reaches the L1 point and 4.4 h at the right end of Figs. 1 and 2, including I/O time. This shows that our code is able to be used for forecast of passage time of a interplanetary shock wave in respect of calculation time. Predicting passage time is 50.5 h, that agrees with observation. Our preliminary simulations indicate that CMEs’ angular radius of about 45◦ gives statistically good results of passage time. A preliminary simulation of the same CME event, in which we adopt the angular radius of 45◦ , resulted in the passage time of 52.8 h. In the present work, we chose the angular radius of 50◦ . That is near to previous indication. The result may indicate the aimed CME have the angular radius of about 50◦ . However, there are uncertainties. While the CME seems to experience relatively quiet solar wind on its 1 AU journey, ACE’s data in the previous solar cycle indicates that there is a fast wind region, which has the maximum velocity of 600 km/s. In our simulation, solar wind has constant 450 km/s, that is observed value of the pre-shock solar wind. If solar wind has varied largely, then it should have inﬂuenced propagation of the shock wave and an optimum parameters of a CME model should be changed. It is diﬃcult to infer a property of solar wind on which a shock wave has underwent.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch05

Simulation of Interplanetary Shock Wave Caused by CME on August 25, 2001

71

A realistic solar wind should be introduced in our future work. Magnetic ﬁeld was not introduced in this paper. Behavior of magnetic ﬁeld is crucial for geomagnetic storm. Introduction of magnetic ﬁeld is our next subject.

Acknowledgments We use the SOHO LASCO CME Catalog. The CME catalog is generated and maintained by NASA and The Catholic University of America in cooperation with the Naval Research Laboratory. SOHO is a project of international cooperation between ESA and NASA. We are very thankful to compilers Drs. S. Yashiro and G. Michalek.

References 1. D. Odstrcil and V. J. Pizzo, J. Geophys. Res. 104, A1 (1999) 483. 2. D. Odstrcil, P. Riley and X. P. Zhao, J. Geophys. Res. 109 (2004) A02116. 3. W. B. Manchester, T. I. Gombosi, I. Roussev, A. Ridley, D. L. De Zeeuw, I. V. Sokolov and K. G. Powell, J. Geophys. Res. 109 (2004) A02107. 4. A. M. Khokhlov, J. Comp. Phys. 143 (1998) 519. 5. A. Noro, T. Ohta, T. Ogawa, K. Yamashita and S. Miyaji, Information Processing Society of Japan Symposium Series 2002, 4 (2002) 9. 6. A. Noro, T. Ogawa, T. Ohta, K. Yamashita, S. Miyaji and M. Den, The Lecture Notes in Computer Science 2327 (2002) 207. 7. T. Ogawa, Thesis, Chiba University (2003). 8. K. Fujii, Numerical Methods for Computational Fluid Dynamics (University of Tokyo Press, 1994) (in Japanese). 9. P. L. Roe, J. Comp. Phys. 43 (1981) 357. 10. B. van Leer, J. Comp. Phys. 23 (1977) 276. 11. B. van Leer, J. Comp. Phys. 32 (1979) 101.

This page intentionally left blank

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch06

OBSERVATION OF THE INFLUENCE OF THE JANUARY 15–17 SOLAR STORMS TO THE MAGNETIC FIELD AND IONOSPHERE OF INDONESIA CLARA Y. YATINI∗ , JIYO and MAMAT RUHIMAT National Institute of Aeronautics and Space of Indonesia Jl. Dr. Junjunan 133 Bandung 40173, Indonesia ∗c [email protected] ∗[email protected]

The Coronal Mass Ejection (CME) associated big ﬂares were ejected from solar surface on January 15–17, 2005. This extreme condition is caused by a single sunspot group of active region NOAA 0720. Together with the ﬂares and associated CMEs several proton events occurred. As the consequence of the solar storms, the earth’s magnetic ﬁeld and ionosphere were disturbed. These CMEs led to the geomagnetic storms. In addition we also noted some depression in the height of critical frequency of ionosphere’s F2 layer, the increase of its height and a disturbance in radio communication in Indonesia. Geomagnetic ﬁeld was observed at the low latitude location of Indonesia; from Biak (−1.08◦ South and 136.5◦ East or geomagnetic −12.8◦ South and 207.30◦ East) and Pontianak (−0.05◦ South and 109.25◦ East or geomagnetic −11.37◦ South and 180.85◦ East) observatories. The variation of geomagnetic ﬁeld during the magnetic storms was recorded, and the delay between the onset of the ﬂares and the geomagnetic storms could be obtained. The eﬀect of the storms to the ionosphere is analyzed using ionospheric data over Kototabang and Biak observatories after the storms.

1. Introduction The research of the inﬂuence of solar activity to the earth’s environment, such as ionosphere condition and geomagnetic ﬁeld, is very complicated. It is so because of the indirect nature of the inﬂuence and the properties of the ionosphere and magnetic ﬁeld itself. Earth’s ionosphere responds markedly to varying solar and magnetospheric energy input. It is also well accepted that solar ﬂare and coronal mass ejection (CME) event could produce large geomagnetic storm. When a ﬂare or CME occurs on the sun, a large increase in electromagnetic radiation follows. The direct response of upper atmosphere is a temporary increase in ionization of minutes to hours. Recently, a primary

73

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch06

74

C. Y. Yatini, Jiyo and M. Ruhimat

cause of geomagnetic activity are believed to be CME. The CME result in a cloud of charged particles which will interact with the earth’s magnetic ﬁeld. In this study, we use one kind of investigation: the temporal-spatial study of one deﬁnite event having high energy that connected with the disturbance of ionosphere and earth’s magnetic ﬁeld, especially at low latitude.

2. Data and Instrumentation Observations of geomagnetic ﬁeld were made in Biak and Pontianak observatories, while the ionospheric observations were performed on Biak and Kototabang observatories. The location of each station is shown in Table 1. As the solar data we are very grateful to the open data on solar websites. 2.1. Solar data The data of solar active activity on January 15–17, 2005 comes from Transition Region and Coronal Explorer (TRACE)1 for active region white light data and 1600 ˚ A, Solar Environment Center2 for GOES ﬂux data and ﬂare list, Yohkoh Solar Observatory3 for X-ray ﬂare list. 2.2. Geomagnetic observation Observation of geomagnetic ﬁeld has been performed on two observatories. The ﬁrst located on Biak, the other is performed on Pontianak. The observation has been made by ﬂuxgate magnetometer with time interval 1 min and noise level 0.02 nT. 2.3. Ionosphere observation The monitoring of ionosphere of the solar storms has been performed on Biak and Kototabang Observatories. On Biak, observation is made by use of Canadian Advance Digital Ionosonde (CADI) to observe height versus frequency, phase and amplitude of echo, angle of arrival, and polarization of the echo. Table 1. Station Biak Pontianak Kototabang

Location of each observing station.

Geographic location

Geomagnetic location

1.08◦ S 136.5◦ E 0.05◦ S 109.3◦ E 0.3◦ S 100.3◦ E

12.8◦ S 207.3◦ E 11.37◦ S 180.85◦ E 10.63◦ S 171.93◦ E

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch06

Observation of the Inﬂuence of the January 15–17 Solar Storms

75

On Kototabang the Frequency Modulation Continuous Wave (FMCW) Ionosonde is used to observe the height. This ionosonde provides ionograms for the transmitter frequencies from 2 to 22 MHz and also data of virtual height at a ﬁxed frequency.

3. Observations and Results Here, we will describe the result of the observations. First, we look at the sun as the source of the disturbance. We use the solar data from Solar Environment Center NOAA, GOES data for X-ray, and Transition Region and Coronal Explorer (TRACE). 3.1. Solar activity Strong ﬂares erupted from solar surface on January 15–17, 2005 (see Table 2). Three of them were long duration events (LDE) ﬂare, which accompanied by CMEs and proton events. These ﬂares originated from active region NOAA 0720 (Fig. 1). The ﬁrst LDE ﬂare erupted on January 15; M8.6 X-ray class ﬂare which started on 05:54 UT, ejected a large amount of protons (Fig. 2). Protons ﬂux with energies > 10, > 50, and > 100 MeV increased, and the ﬂux seems to become normal level when then the second LDE ﬂare erupted. It also increased the proton ﬂux as well as the third LDE ﬂare. The ﬂux peaked on January 17, 2005. On that day the peak proton ﬂux was greater than 4000 pfu. Table 2.

X-ray ﬂares greater than M1 class erupted on January 15–17, 2005.

No.

Date

Start (UT)

Max. (UT)

End (UT)

X-ray class

Active region

1 2 3 4 5 6 7 8 9 10 11

15 15 15 15 15 15 15 15 16 17 17

00:22 04:09 04:26 05:54 11:41 14:08 22:01 22:25 21:55 03:10 06:59

00:43 04:16 04:31 06:38 11:48 14:23 22:08 23:02 22:03 03:21 09:52

01:02 04:22 04:36 07:17 11:50 14:39 22:16 23:31 22:22 03:32 10:07

X1.2 M1.3 M8.4 M8.6 M1.2 M3.2 M1.0 X2.6 M2.4 M2.6 X3.8

0720 0720 0720 0720 0720 0718 0720 0720 0720 0720 0720

Remarks

LDE, CME

LDE, CME

LDE, CME

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch06

76

C. Y. Yatini, Jiyo and M. Ruhimat

Fig. 1. The LDE ﬂares on January 15–17, 2005. Upper panel is ﬂare no. 4, middle is ﬂare no. 8, and the lowest is ﬂare no. 11 (see Table 2) observed by TRACE. Left ﬁgures are as observed in 1600 ˚ A, right ﬁgures are in white light (courtesy of TRACE).

3.2. Geomagnetic ﬁeld observation Biak and Pontianak observatories have geomagnetic data according to solar storms on July 15–17, 2005. The observations are conducted to the H, D, and Z components of geomagnetic variations. The change in geomagnetic activity appeared only on H component. Figure 3 shows a ﬁveday record of the horizontal (H) component in 6 min average. The solar

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch06

Observation of the Inﬂuence of the January 15–17 Solar Storms

77

Fig. 2. Plot of proton ﬂux with energy ≥ 10 MeV (upper), ≥ 50 MeV (middle), and ≥ 100 MeV (lower) on January 15–20, 2005 (courtesy of Solar Environment Center2 ).

Fig. 3. Geomagnetic variation of H component from Biak (solid line) and Pontianak (dotted line) observatories on 6 min average. From top to bottom: H component for January 15–19, 2005.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch06

78

C. Y. Yatini, Jiyo and M. Ruhimat

storm signatures for the Biak station are appeared in the ﬁgure which is shown by the decrease in the H component to lower level (down about 100 nT), and then recovery. The geomagnetic substorms are appeared on January 17, 2005 (around 15 UT) and on 18 January at around 0 and 07 UT.

3.3. Ionospheric response 3.3.1. Ionospheric storms Ionospheric storm is a global electrodynamic disturbance in the ionosphere. The magnetic storm can lead to the ionospheric storm, which is indicated by the depression of critical frequency of F2 layer to its median. This phenomenon is observed over Biak on January 19, 2005, as shown in Fig. 4. Such depression lasted all day. In the other day, on 17th January which is the period of the solar storm the critical frequency shows a sharp peak. It is suggested that the energetic electromagnetic radiation accompanying ﬂares arrived at earth just a few minutes after leaving the ﬂare site, and led to the increase in ionization. After the depression, the critical frequency increased again to the higher level on 22nd. The other response of the solar storm in the ionosphere is the increase of the height of ionospheric layer. This increase of the layer which caused by January solar storms was observed over Biak observatory. It moved up to more than 800 km in height for a couple hours, and went down after 15.00 (local time) (see Fig. 5.) In the other hand, this kind of ionospheric response was not seen on the data over Kototabang. Figure 6 shows the height of F2 layer over Kototabang.

3.3.2. Spread F Irregularities in the equatorial regions on the bottom and the top sides of the ionosphere can cause the disturbance condition of the ionosphere known as spread F .4 According to January 15–17, 2005 solar storms, we observed the Equatorial Spread F (ESF) on Biak on January 19, 2005 from 23.00 to January 20, 05.30 in the local time or on January 19 from 14.00 to 20.30 UT as seen in Fig. 7. Such spread F was not observed over Kototabang.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch06

Observation of the Inﬂuence of the January 15–17 Solar Storms

79

Fig. 4. Top: plots of critical frequency of ionospheric layer over Biak from January 17–22, 2005 (dotted line). Solid line shows the median. Bottom: the depression of the frequency on January 19, 2005 (ﬁlled circle) as the consequences of the solar storms. The open circle denotes the median.

3.3.3. Radio communication The condition of radio wave propagation in Jakarta to some places in Indonesia in the period of January 2005 was not good. However, on 16th and 17th the condition was worst. It is appeared in the plot of propagation

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch06

80

Fig. 5.

C. Y. Yatini, Jiyo and M. Ruhimat

The increase of the height of F2 layer over Biak on January 19, 2005 (triangle).

Fig. 6. The height of F2 layer over Kototabang from January 15 to 21, 2005 (in local time). There was no signiﬁcant change in the height of layer during and after solar storms.

index as seen in Fig. 8. It is suggested that this condition is caused by the ﬂares which erupted in January 15–17.

4. Summary and Discussion Equatorial ionosphere–magnetosphere irregularities are the scientiﬁc interest and to understand for more eﬃcient use of the ionosphere as a communication medium. It has been known that the geomagnetically induced storms and ionospheric storm are forms of space weather resulting from solar activity. The plasma cloud ejected from ﬂares caused the interplanetary shock, and then generated the disturbance of the H ﬁeld.5

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch06

Observation of the Inﬂuence of the January 15–17 Solar Storms

Fig. 7.

Spread F as observed over Biak observatory on January 19, 2005.

Fig. 8.

High-frequency wave propagation index in January 2005.

81

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch06

82

C. Y. Yatini, Jiyo and M. Ruhimat

This report only summarizes the phenomenology of remarkable geomagnetic storm caused by solar ﬂare eruption on low latitude. The January 15–17, 2005 solar storms led to geomagnetic disturbance on the H (NorthSouth) component on January 17–18. The time delay was 58–59 h (two days). These geomagnetic storms also inﬂuence the critical frequency and the height of ionospheric layer and led to F -spread type ionospheric disturbance. The delay between geomagnetic and ionospheric disturbance is about 19 h, as found by Rastogi6 and references therein.

Acknowledgments We want to express our thanks to the observers of Pontianak, Biak, Kototabang Observatory of National Institute of Aeronautics and Space of Indonesia (LAPAN).

References 1. 2. 3. 4. 5. 6.

Transition Region and Coronal Explorer, http://vestige.lmsal.com/TRACE/. Solar Environment Center, http://www.sec.noaa.gov/. Yohkoh Solar Observatory, http://www.lmsal.com/. K. Cole, Australian & New Zealand Physicist 28, 2 (1991) 263. B. G. Fejer, M. F. Larsen and D. T. Farley, Geophys. Res. Lett. 10 (1983) 5337. R. G. Rastogi, Ann. Geophys. 17 (1999) 1426.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

OBSERVATIONAL STUDY OF SOLAR MAGNETIC ACTIVE PHENOMENA BY HUAIROU VECTOR MAGNETOGRAPH HONGQI ZHANG National Astronomical Observatories Chinese Academy of Sciences, Beijing 100012, China [email protected]

We mainly present some works based on the study of photospheric vector magnetograms of solar active regions and also chromospheric longitudinal magnetograms obtained at Huairou Solar Observing Station near Beijing. The conclusions of the analysis of the formation process of complex and delta magnetic conﬁguration in some super active regions are the following: (1) The magnetic shear and gradient provide the nonpotentiality of the magnetic ﬁeld of active regions reﬂecting the existence of electric current. (2) The possibility of twisted magnetic ropes generated in the subatmosphere can be conﬁrmed from the evolution of photospheric vector magnetic ﬁelds. Some of large-scale delta active regions could be due to the emergence of highly sheared nonpotential magnetic ﬂux bundles from the subatmosphere. (3) We also present some results of a study of the magnetic (current) helicity. Measurements of the chromospheric magnetic ﬁelds are discussed in the frame of a three-dimensional magnetic conﬁguration extending from the photosphere to the high levels in the ﬁbrial-like features. The analysis of chromospheric magnetic ﬁeld is useful to diagnose the possible extending form of the magnetic nonpotentiality also.

1. Observations Photospheric vector and chromospheric longitudinal magnetograms have been observed at Huairou Solar Observing Station of National Astronomical Observatories of China since 1987.1–3 Two working lines are used in the Huairou Vector Magnetograph with a birefringent ﬁlter of 1/8 ˚ A bandpass. One line is FeIλ5324.19 ˚ A for the measurements of the photospheric vector A magnetograms and Dopplergrams,1 and the another line is Hβλ4861.34 ˚ for the measurements of the chromospheric longitudinal magnetograms and Dopplergrams.3 The Lande g factor of Hβ line is about 1 and its equivalent width is 4.2 ˚ A.

83

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

84

H. Zhang

2. Accuracy of Measured Photospheric Magnetic Field The comparison of vector magnetograms obtained by diﬀerent magnetographs and Stokes polarimeters at diﬀerent observatories is also important. A series of such works has been made at Huairou Solar Observing Station.4–7 Figure 1 shows the photospheric vector magnetograms obtained at Huairou Solar Observing Station and Marshall Space Flight Center, which was a powerful ﬂare producing delta region.8–11 The basic correlation between the diﬀerent vector magnetograms is analyzed. The diﬀerences of the mean azimuthal angles for both transverse magnetograms ◦ i ∆ϕi Bzi ∆ϕi = ∆ϕi = 6. 4 , ∆ϕwi = = −2.◦ 3 and B zi i i i ∆ϕi |Bzi | = −2.◦ 3 , ∆ϕwai = i |Bzi |

Fig. 1. The photospheric vector magnetogram of active region NOAA 6659. The white (black) areas indicate the positive (negative) polarity and black arrows mark the direction of the transverse component of the ﬁeld in the Huairou vector magnetogram. The solid (dashed) contours indicate the positive (negative) polarity and white arrows mark the direction of the transverse component of the ﬁeld in the MSFC vector magnetogram.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

200

4

100

2

Bz (1000G)

MSFC (degree)

Observational Study of Solar Magnetic Active Phenomena

0

-100

-200 -200

85

0

-2

-100

0 Huairou (degree)

100

200

-4 -200

-100 0 100 Az(MSFC)-Az(HR) (degree)

200

Fig. 2. The statistical correlation between the azimuthal angles of transverse magnetic ﬁeld obtained at Huairou and MSFC in active region NOAA 6659 on June 9, 1991.

where Bzi and ∆ϕi are the longitudinal magnetic ﬁeld and azimuthal difference of transverse ﬁeld at ith pixel in the vector magnetograms (Fig. 2). Except the basic agreement for diﬀerent vector magnetograms, some diﬀerence can be found even between data obtained by diﬀerent magnetographs of the same solar observatory.7 The diﬀerences of the magnetograms are probably caused by the following possibilities: (a) The resolution of vector magnetograms, which also relates to the diﬀerent seeing conditions at diﬀerent observing sites. (b) The inﬂuence of magneto-optical eﬀects, especially on the transverse magnetograms. Even if one observes the magnetic signals at the wing of the spectral lines, this eﬀect still exists. (c) The crosstalk of diﬀerent Stokes parameters on the measurement of magnetic ﬁeld in the magnetographs and/or Stokes polarimeters due to losing accuracy of their magneto-modulators, etc. (d) The residual polarizations in the optical systems of magnetographs or Stokes polarimeters. (e) The magnetic sensitivity of magnetographs in comparison with observational noise. All diﬀerences also infect the calculation of the current and current helicity using photospheric vector magnetograms,12 even if we excluded of the time evolution of magnetic ﬁelds of the active regions. Moreover, the inversion code of Stokes polarization proﬁles obtained by the Huairou Vector Magnetograph to recover the vector magnetic parameters has been presented by Su and Zhang.13,14

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

86

H. Zhang

3. Magnetic Shear, Twist, and Relationship with Electric Current The electric current reﬂects the nonpotentiality of magnetic ﬁeld in the solar atmosphere and can be inferred from the magnetic ﬁeld in the form J=

1 ( × B) , µ0

(1)

where µ0 = 4π × 10−3 Gm/A. The magnetic shear or twist normally reﬂects the existence of electric current and helicity in the active region. The electric current can be written in the form15 1 B J= (B) × b + ×b, (2) µ0 µ0 where B = Bband b is the unit vector along the direction of magnetic ﬁeld and B = Bx2 + By2 + Bz2 . This indicates that the electric current in solar active regions is related to the properties of gradient and chirality of magnetic ﬁeld. The ﬁrst term in Eq. (2) depends on the heterogeneity and orientation of magnetic ﬁeld, i.e., the shear of vector magnetic ﬁeld. The second term in Eq. (2) depends on the twist of unit magnetic lines of force and the intensity of ﬁeld. We can artiﬁcially deﬁne that the ﬁrst term is the shear current and the second term is the twist current in Eq. (2). According to Eq. (2), the vertical current can be inferred from

∂bx ∂B ∂B 1 B ∂by − bx − by + . (3) Jz = µ0 ∂x ∂y µ0 ∂x ∂y

4. The Relationship Between the Magnetic Shear, Gradient, and Current The magnetic shear is an important parameter to measure the nonpotentiality of magnetic ﬁeld in solar active regions,16 while the nonpotential ﬁeld can also be measured from the strong magnetic gradient of active regions, which is strongly correlated with active region CME productivity.17 Figure 3 shows the distribution of magnetic shear in the active region 6659 on June 9, 1991. The shear is deﬁned from the shear angle weighted by the transverse magnetic ﬁeld θT = BT · cos−1

BT · BpT , BT BpT

(4)

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

Observational Study of Solar Magnetic Active Phenomena

87

Fig. 3. The distribution of magnetic shear in active region NOAA 6659 on June 9, 1991. The white arrows mark the observed transverse ﬁeld and the black arrows show the transverse components inferred from the calculation of magnetic charges.

where BT and BpT are the observed transverse ﬁeld and that calculated from the magnetic charges in the approximation of potential ﬁeld. The amplitude of the shear angle reﬂects the nonpotentiality of the active region. It is noticed that the deﬁnition of the shear angle of the transverse ﬁeld in Eq. (4) is slightly diﬀerent from the ﬁrst term of Eq. (2), but the physical meaning is almost the same. The gradient of the photospheric longitudinal magnetic ﬁeld in active regions can be inferred from

2

2 ∂Bz ∂Bz + . (5) | (Bz )| = ∂x ∂y The distribution of the corresponding magnetic gradient in active region NOAA 6659 inferred from the photospheric vector magnetogram is shown in Fig. 4. The main contribution of magnetic shear in the active region comes from the deviation of the transverse ﬁeld from the potential ﬁeld inferred by magnetic charges in the photosphere, while the magnetic gradient comes from the nonuniformity of the longitudinal ﬁeld. By comparing Eqs. (4) and (5) with Eq. (2), it is found that the basic information on the magnetic shear and gradient are contained in the shear term of the current. The deﬁnition of the angle of observed magnetic shear

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

88

H. Zhang

Fig. 4. The magnetic gradient in active region NOAA 6659 in 1991 inferred from the photospheric vector magnetogram.

in Eq. (4) relative to the direction of transverse ﬁeld inferred from the potential ﬁeld is slightly diﬀerent from the orientation of transverse ﬁeld relative to B, but it reﬂects the diﬀerence of the reference frames in the analyzing the nonpotentiality of magnetic ﬁeld only. This means that the shear component of the current actually provides both magnetic shear and gradient of ﬁeld in active regions. The directions of magnetic shear and gradient are lost in Eqs. (4) and (5). Moreover, it is also noticed that the contribution of twist component of the current is not contained in the analysis of the magnetic shear of transverse ﬁeld and the gradient of longitudinal ﬁeld in the active regions. This means that the electric current is relatively more complete quantity than the magnetic shear and gradient in the study of the nonpotential ﬁeld. The magnetic shear and gradient are not enough to describe the magnetic nonpotentiality completely.

5. Nonpotentiality of Magnetic Field in Active Regions Active region NOAA 9026 was a fast developing region in June 2000, and it was noticeable in the northern hemisphere in the 23rd solar cycle, which

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

Observational Study of Solar Magnetic Active Phenomena

89

Fig. 5. The soft X-ray image and the extrapolated magnetic ﬁeld (top) inferred from the full disk MDI magnetogram (bottom) on June 7, 2000. The white (black) in the magnetogram indicates positive (negative) polarity. North is at the top, west is at the right.

produced a series of powerful ﬂares and the corresponding geophysical eﬀects.18,19 Figure 5 shows the soft X-ray conﬁguration of this active region overlapped by the magnetic lines of force above the photosphere. The lines of force are extrapolated using a constant force free ﬁeld based on the MDI full disk magnetogram obtained by SOHO satellite. The α factor is −3.83 × 10−8 /m, which is relevant to the value of αbest presented in the following. We ﬁnd that the magnetic ﬁeld of active regions actually includes the strong magnetic ﬁeld in the vicinity of spots and also the enhanced network one. The latter spreads into a large area near the active region. The powerful soft X-ray ﬂare occurred above the active region and the past ﬂare loops connected both polarities of main photospheric magnetic ﬁeld of the active region.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

90

H. Zhang

The mean tilt angle of the photospheric magnetic ﬁeld of active region is about 14◦ in the MDI synoptic magnetogram of June 7, inferred from the weighting centers of positive and negative magnetic ﬁeld of the active region. It is roughly consistent with the estimation (about 20◦ ) from the connection line of the footpoints of the extrapolated magnetic lines of force of opposite polarities, where the lines of force mainly connected the largescale photospheric enhanced network magnetic ﬁeld of opposite polarities and are almost consistent with the morphology of soft X-ray features near the middle of the active region in Fig. 5. It is found that on July 5 the magnetic ﬁeld of active region decayed to the enhanced network ﬁeld and the tendency of the tilt angle of magnetic ﬁeld is almost the same. Figure 6 shows the development of photospheric vector magnetic ﬁeld in the active region on June 4–8. It is found that the local polarity distribution of sunspot ﬁeld is complex and shows a delta conﬁguration. The shear of transverse ﬁeld relates to the evolution of sunspots in the active region, i.e., the transverse ﬁeld is almost parallel to the inversion line of opposite polarities in the middle of active region. From the orientation of the transverse ﬁeld, one probably can infer that the magnetic poles p1 and f 1 consist of a pair of opposite polarities. The magnetic pole p1 moved westward relative to f 1 in the development of the active region. It probably means that the magnetic pair p1 and f 1 is independently relative to magnetic pole P in the formation process from the subatmosphere, in other words, the magnetic pair p1 and f 1 probably was a magnetic rope in the subatmosphere and emerged in the solar surface to form a pair of magnetic poles of opposite polarities. The shear of transverse magnetic ﬁeld between magnetic poles P and f 1 reﬂects the interaction of diﬀerent magnetic systems. The vertical current and current helicity density hhc in the active region NOAA 9026 can be inferred from the vector magnetograms of Fig. 6. We may ﬁnd that the electric current normally ﬂows from negative polarity to the positive one. The corresponding helicity picture shows that in most areas of the active region the photospheric current helicity density (hcz = Bz ( × B)z ) is with a negative sign. The values of αbest of force free ﬁeld are −3.5, −3.5, −3.8 and −2.8 inferred from vector magnetograms on June 4, 6–8, respectively (the unit is ×10−8/m). It is consistent with the left handedness of the large-scale soft X-ray loops in Fig. 5, the magnetic lines of force show the reversal sigmoid conﬁguration above the photosphere in the active region basically. Although the conﬁguration of the large-scale magnetic ﬁeld, extrapolated from the photospheric magnetic ﬁeld in the approximation of constant force free ﬁeld, is relative simple, the existence

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

Observational Study of Solar Magnetic Active Phenomena

91

Fig. 6. Photospheric images (left) and vector magnetograms (right) in active region NOAA 9026 in 2000. The white (black) in magnetograms indicates positive (negative) polarity. The arrows indicate the transverse components of ﬁeld. North is at the top, west is at the right. The size of images is 2. 62 × 1. 82.

of local photospheric reverse magnetic structures near the center of the active region actually means the complexity of magnetic ropes from the subatmosphere. The tangential component of the velocity ﬁeld near the footpoints at the solar surface can be normally detected by local correlation tracking techniques. The change of magnetic helicity in the solar atmosphere relates

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

92

H. Zhang 80

dH/dt (1040 Mx2 h-1)

60 40 20 0 -20 -40 -60 -80 0

30

0

30

60

90

120

150

180

150

180

Helicity accumulation (1040 Mx2)

1000 750 500 250 0 -250 -500 -750 -1000

Fig. 7.

60 90 120 Time (UT) 00:00UT 00 Jun 04 -- 20:48 UT 00 Jun 10

Injection of magnetic helicity in active region NOAA 9026 in June 2000.

to the motion of footpoints of magnetic ﬁeld in the solar surface.20–29 dHm = −2 [(Vt · Ap )Bn − (Ap · Bt )Vn ]ds , (6) dt S where the magnetic ﬁeld B and velocity ﬁeld V are observable in the solar atmosphere. The ﬁrst term in Eq. (1) provides the contribution from the motion of footpoints of magnetic ﬁeld in the solar surface, while the second term does that from the emergence of magnetic ﬂux from the subatmosphere. The change rate of magnetic helicity in active region NOAA 9026 is shown in Fig. 7. It is found that an order of 1043 M x2 helicity is injected in this active region when it passed in the solar disk.

6. Measurements of Magnetic (Current) Helicity Analyzes of the magnetic (current) helicity in active regions show: (a) The evolution of active regions and the relationship with magnetic helicity, as well as the photospheric current helicity density with emerging

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

Observational Study of Solar Magnetic Active Phenomena

(b)

(c)

(d)

(e)

(f)

(g)

93

magnetic ﬂux, are interesting topics.15,30 Because only photospheric vector magnetograms can be observed, one still cannot observationally infer how the twist of magnetic ﬂux ropes forms in the subatmosphere immediately. The important evidence is the emergence of highly sheared magnetic ﬂux near the magnetic neutral line and the motion of its photospheric footpoints in the nonpotential active regions. It is found that the mean current helicity density and twist of active regions normally shows a negative sign in the northern hemisphere and positive sign in the southern one.31–35 Even if the sign rule has been statistically found, the deviations between the diﬀerent data sets are probably caused by observational errors, etc. Pevtsov et al.36 found that in the most of active region samples obtained by the Video Vector Magnetograph at Huairou and Stokes Polarimeter at Hawaii the αbest shows the same sign. The relationship between the current helicity and the solar activity cycle has been analyzed by Bao and Zhang34 and Hagino and Sakurai.37,38 The sign rule probably does not change with solar activity cycles,12,36 while the reversal sign in the beginning of solar cycle was proposed by Choudhuri et al.39 based on a solar dynamo model. The longitudinal distribution of the mean current helicity density of active regions with opposite signs relative to most ones have also been studied (e.g., Ref. 40). The possible origin of the twisted magnetic ﬁeld in solar active regions is notable (e.g., Refs. 41 and 42). Kuzanyan et al.43 proposed the possible relationship between the formation depth of the helical magnetic ﬁeld inside the Sun and the α-eﬀect of the solar dynamo. The α-eﬀect probably changes the value and sign of magnetic helicity near the bottom of the convection zone. The observational study on the hemispheric distribution of current helicity of solar active regions provides an important widow to analyze the basic mechanism of solar dynamo actually.44,45 A consistency between the soft X-ray helical conﬁguration and the photospheric vector magnetic ﬁeld can be found (e.g., Ref. 46). The change of photospheric current helicity density with solar ﬂares47 provides a similar means of the relaxation of the magnetic energy in the active regions, which is actually equivalent to the change of the photospheric vector magnetic ﬁeld during the ﬂares.48 The ﬂare occurrence in the solar active regions with the reversed helicity sign was presented by Bao et al.47,49 The theoretical interpretation of the ﬂare activity and variability of electric current helicity was

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

94

H. Zhang

presented by Kim et al.50 The relationship between helicity evolution and δ-conﬁgurations was studied by Deng et al.51 Zhang,30 and Liu and Zhang.52 The helicity patterns of CME-associated delta active regions were analyzed by Wang et al.53 (h) In recent years, the injection of magnetic helicity from subatmosphere with ﬂares, ﬁlament eruptions, CMEs and MCs has been computed by the local correlative tracking technique and other methods.20–29

7. Measurements of Chromospheric Magnetic Field The chromosphere is an important layer for the study of solar activity. Due to the frozen-in eﬀect of the plasma in the solar atmosphere, it is normally believed that the chromospheric features reﬂect the distribution of chromospheric magnetic ﬁeld.54 The observations of the chromospheric magnetic ﬁeld can be made by the Zemann and Hanle eﬀects of the spectral lines. They are normally used to measure the magnetic ﬁeld on the solar disk or that of the prominences near solar limb respectively. We mainly discuss the measurements of chromospheric magnetic ﬁeld on the solar disk. It is noticed that a few of solar spectral lines can be used for the measurements of chromospheric magnetic ﬁeld, and these lines normally have lower sensitivity for the magnetic ﬁeld. The measurements of chromospheric magnetic ﬁeld had been paid attention on the areas where the magnetic ﬁeld is strong, such as sunspot regions. Moreover, due to amount of technical problems for the measurements of coronal magnetic ﬁeld, the measurement and study of the chromospheric magnetic ﬁeld are very important for the reconstruction of the 3D magnetic ﬁeld in the extrapolation of the photospheric ﬁeld. The study of chromospheric magnetic ﬁeld demonstrates that the magnetic ﬁeld of active regions extends up in the canopy form.55 Due to the variation of the solar atmospheric parameters, such as density, temperature, etc., one tries to infer the common model of the magnetic ﬁeld56 and interpret its possible spatial distribution ﬁeld in the solar quiet region.57 Some new noticeable results in the observational study of the ﬁne morphological conﬁguration in the transition region and corona have been done by TRACE.58 These achievements improved the knowledge and imagine for the spatial conﬁguration of the magnetic ﬁeld. At Huairou Solar Observing Station of National Astronomical Observatories in Chinese Academy of Sciences, a great of observational data of chromospheric magnetic ﬁeld with high spatial resolution has been obtained,

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

Observational Study of Solar Magnetic Active Phenomena

95

and a series of works in the areas has been made. We will discuss the observational results and questions on the study of chromospheric magnetic ﬁeld. 7.1. Nonuniformity of the chromospheric magnetic ﬁeld The observations at Huairou Solar Observing Station demonstrate that the sunspot magnetic ﬁeld extends up in the form of the ﬁbril-like features in the chromosphere and runs along the direction of the features in the Hα and Hβ monochromatic images.59 It provides the possible spatial extension form of the ﬁne features. The relationship between the chromospheric and photospheric magnetic ﬁeld is shown in Fig. 8. In the study of the chromospheric magnetic ﬁeld in the quiet Sun, Zhang and Zhang60–62 ﬁrst obtained the high spatial resolution chromospheric magnetograms with the method of deep integration. The spatial resolution of the magnetic elements is about 3–4 . It is found that the distribution of chromospheric magnetic ﬁeld is similar to the photospheric one. This similarity not only includes the magnetic ﬁeld near the boundary of the network, but also the intranetwork ﬁeld. By analyzing the chromospheric magnetic ﬁeld near the solar limb, Zhang and Zhang60–62 do not ﬁnd the

Fig. 8. Photospheric and Hβ chromospheric images (left) and corresponding longitudinal magnetograms (right) in a solar active region.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

96

H. Zhang

stronger horizontal component of the chromospheric magnetic ﬁeld in the quiet Sun. The former models of the magnetic ﬁeld in the quiet Sun were established on the magnetostatic equilibrium of the solar atmosphere, which was suggested that the magnetic ﬁeld in the quiet Sun extends from the boundary of the magnetic networks to the intranetwork and forms the magnetic canopy model. By analyzing the spectral lines in the solar magnetic atmosphere and considering the inﬂuence of the ﬁlling factor, some authors suggest that the size of the real magnetic elements probably are less than 0.2 .63 The nonuniformity of the chromospheric magnetic ﬁeld is also questionable in the magneto-hydrodynamic conditions of solar atmosphere. As we analyze the Huairou photospheric and chromospheric magnetograms in the quiet Sun, we know that these magnetograms have not reach the best spatial resolution in the earth bases, i.e., less than 0.5 . This means that the observational results by Zhang et al.64 cannot be used to conﬁrm how does the extension of the ﬁeld of individual magnetic elements, but the ﬁeld is conﬁned in the observational size pixel. On the other hand, one can say that the photospheric magnetic ﬁeld extends into the chromosphere in the quiet Sun and does not return to the photosphere in the lower solar atmosphere or does not contain the strong horizontal components. As we know that the ﬁne structure of coronal magnetic ﬁeld cannot be detected well until now, the results obtained by Zhang et al.64 are consistent with the 171 monochromatic images obtained by TRACE satellite. From the 171 monochromatic images, it is found that the features in the 171 monochromatic images extend from the photospheric elements in the ﬁbril-like form and do not diﬀuse signiﬁcantly. As we believe the frozen-in condition in the solar magnetic atmosphere, we can infer that these ﬁbril-like features in the TRACE images reﬂect the information of the magnetic lines of force basically. Of course, the distribution of coronal ﬁne conﬁguration is not equal to that of the magnetic ﬁeld, a part of coronal magnetic ﬁeld probably cannot be detected due to the nonexcitation of the neighboring plasma (Fig. 9). Moreover, the modeling chromospheric and coronal magnetic ﬁeld was also achieved based on the analysis of observations of TRACE by some authors (e.g., Ref. 65). Bao and Zhang66 obtained the Hβ magnetograms nearby a quiescent prominence and the corresponding photospheric magnetograms, and try to analyze the distribution of the magnetic lines of force in the vicinity of prominences. It is probably useful for our understanding the forming environment of prominences in the high solar atmosphere supported by the surrounding magnetic ﬁeld.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

Observational Study of Solar Magnetic Active Phenomena

97

TRACE 171 Fibrils

Chromosphere Photosphere

Chromospheric magnetic element

Photospheric magnetic element

Fig. 9. The possible extension of the magnetic ﬁeld nonuniformly. The gray features are 171 ﬁbril-like features observed by TRACE satellite. The solid arrows mark the magnetic lines of force along the 171 ﬁbrils. The dashed arrows mark a part of possible magnetic lines of force.

7.2. The possibility of reversal features in Hβ chromospheric magnetograms By analyzing Hβ chromospheric magnetograms, we can ﬁnd that in some areas the diﬀerence between the chromospheric and photospheric magnetograms can be detected. It probably results by the extension of the magnetic ﬁeld from the photosphere.59 Chen et al.48 pointed out that some reversal structures, relative to the photospheric one, probably exist nearby the regions of the strong magnetic ﬁeld in the chromospheric magnetograms obtained at the wavelength of Hβ-0.24 ˚ A. It is normally called CAZJ reversion of the chromospheric magnetic ﬁeld.67 A similar case has been presented by Wang and Shi.68 The observational results have been systematically analyzed, such as some of data analysis has been made by Li et al.69 By analyzing the formation of Hβ line, several possibilities for the reversal structures in the chromospheric magnetograms can be inferred: (a) the real reversal magnetic structures in the solar atmosphere; (b) the reversal sign of the Stokes parameter V caused by the reversion of the Hα line proﬁle in the local areas of the chromosphere67; (c) the disturbance of the blended lines in the wing of the Hβ.70 We need to point that the formation height of the Hβ core is less than 2000 km from the photosphere. As the size of our analyzed reversal magnetic conﬁguration is larger than the height

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

98

H. Zhang

diﬀerence between the photosphere and chromosphere, the conservation of the magnetic ﬂux in the both layers should be considered71 : B · ds = 0 , (7) S

where B is the magnetic ﬁeld strength and S is the area passed by the magnetic ﬂux in the photosphere and chromosphere. If this deformed magnetic conﬁguration is relative stable, it should also consistent with the static equilibrium condition of the magnetohydrodynamics.72 On the other hand, it is diﬃcult to guess that the chromospheric reversal magnetic structures form above the sunspot umbrae, because the umbral magnetic ﬁeld normally is vertical to the solar surface. Thus, if the chromospheric reversal magnetic structures exist really, the size of these kinds of twisted magnetic lines of force would be limited and can be reﬂected in the photospheric vector magnetograms. As pointed by Alfven and Falthammer,73 the condition of the energy decrement for the twisted magnetic rope is72 : a a Bφ2 r dr > 2 Bz2 r dr , (8) 0

0

where Bz is the axial ﬁeld, Bφ toroidal one and a is the radius of the magnetic rope. This means that as the density of magnetic energy of the toroidal ﬁeld is larger than twice of the axial one, the magnetic rope becomes instable. On the other hand, the analysis of the Stokes monochromatic images obtained at diﬀerent wavelengths in the wing of Hβ line is useful for the diagnostic of the blended lines. The blended lines in the wing of Hβ line probably cause the reversed sign in the magnetograms obtained at the single wavelength of the Hβ line. A similar evidence on the disturbance of blended lines in the wing of Hα is demonstrated by Hanaoka.74

8. Formation of Nonpotentiality Inferred from Chromospheric and Photospheric Magnetograms Figure 10 shows a example of Hβ ﬂares in the active region NOAA 6619 on May 10, 1991. The Hβ ﬂare ribbons occurred in the both sides of the magnetic neutral line. A ribbon consisted of two cores marked by f 1 in the umbra of negative polarity and another one marked by f 2, f 3, and f 4, where f 2 was located in the umbra and f 3 and f 4 extended into the enhanced network of positive polarity. It is found that the extrapolated magnetic lines of force connect the both ribbons of opposite polarities under

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

Observational Study of Solar Magnetic Active Phenomena

99

Fig. 10. The Hβ ﬂare in active region NOAA 6619 on May 10, 1991 overlaid by the extrapolated magnetic lines of force (top). The solid (dashed) contours mark the positive (negative) polarities of the photospheric magnetic ﬁeld. The chromospheric longitudinal magnetic ﬁeld and photospheric transverse magnetic ﬁeld (bottom). The white (black) marks the positive (negative) polarity.

the approximation of linear force free ﬁeld, where the α factor is −4.0 (the unit is 10−8 /m). In the observational photospheric vector magnetogram, the highly sheared transverse component is formed along the magnetic neutral line between the ﬂare ribbons with opposite magnetic polarities marked by N and S in Fig. 10.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

100

H. Zhang

As comparing the relationship between the Hβ magnetogram and photospheric transverse ﬁeld, it is found that the chromospheric magnetic features around the magnetic main poles N and S show a tendency that these features are almost parallel to the direction of the photospheric transverse ﬁeld. It reﬂects a evidence on the basic consistence between the distribution of chromospheric magnetic ﬁbrils and direction of observational transverse ﬁeld resolved 180◦-ambiguity near the center of solar disk. The reversal structure in the umbra N in the chromospheric magnetograms probably is caused by the disturbance of photospheric blended lines in the wing of Hβ line.70 A limb Hβ ﬂare in the active region occurred on May 16, 1991 which ﬂare ribbons show the similar pattern to that of May 10, even if the projective eﬀect of the longitudinal magnetic ﬁeld inﬂuences the adjustment of the relationship between the position of ﬂare ribbons and magnetic polarities. As the comparison with development of transverse magnetic ﬁeld near magnetic neutral line in the active region, we ﬁnd the shear of magnetic ﬁeld increased after ﬂares on May 10, 16, and the X-2.8 ﬂare on May 18. It means that the increase of highly sheared magnetic ﬁeld near the magnetic neutral line and the interaction with the existed ﬁeld are basic processes in the development of active region NOAA 6619 before a series of powerful ﬂares. The similar observational evidence was also demonstrated by Wang et al.75 and Zhang et al.11 The shear increases after powerful ﬂares immediately contradicts with kink model of twisted magnetic ropes. It is hard to infer that high intense kinked magnetic ropes formed after the reconnection of magnetic ﬁeld related to the power ﬂares.

9. Summary We present some works based on the study of photospheric vector and chromospheric magnetograms of solar active regions obtained at Huairou Solar Observing Station near Beijing. The main conclusions of the analysis of the formation process of complex and delta magnetic conﬁguration inferred from photospheric vector magnetograms in some super active regions and the relationship between magnetic ﬁelds in the photospheric and chromospheric layers are the following: (1) The measurements of magnetic ﬁeld are still a challenge. It concerns diagnostic of the magneto-optical eﬀects and instrumental polarization for the measurement of photospheric vector magnetic ﬁeld and the identiﬁcation of blended lines in the wings of Hα and Hβ lines or the

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

Observational Study of Solar Magnetic Active Phenomena

(2)

(3)

(4)

(5)

101

reversal of the lines in the period of solar ﬂares for the measurement of the chromospheric magnetic ﬁeld, etc. It is found that the magnetic shear and gradient are two basic quantities to bring some information on nonpotentiality of the magnetic ﬁeld of solar active regions reﬂecting the existence of electric current. The interaction between diﬀerent magnetic ropes probably causes the high shear and gradient near the magnetic neutral line between them in the solar surface. The emergence of twisted magnetic ropes is one of candidates relative to the transfer of free magnetic energy from the subatmosphere only, which can be conﬁrmed from the study of the evolution of large-scale twisted photospheric transverse magnetic ﬁelds and the relationship with the magnetic shear in some delta active regions. The formation of delta active regions in the solar surface is due to the emergence of highly sheared nonpotential magnetic ﬂux bundles, relative to surrounding ones, generated in the subatmosphere. The magnetic (current) helicity is important for the study of magnetic topology in the solar atmosphere. As the observable quantity, it relates to the distribution of current helicity density in the photosphere and the injection rate of magnetic helicity from the photosphere into the interplanetary space. Measurements of the chromospheric magnetic ﬁelds and the comparison with photospheric ﬁeld proved the basic information on the frame of a 3D magnetic conﬁguration extending from the photosphere to the high levels in the ﬁbrial-like features and the possible nonpotentiality in solar active regions. It does not ﬁnd the the intense horizontal component of the magnetic ﬁeld in the chromospheric layer in the quiet sun, such as near the boundary of the networks and also in the internetwork.

Acknowledgments The author would like to thank Dr. M. Adams for providing the observational data at Marshall Space Flight Center and the kindly comments for the manuscript by Dr. B. Schmieder. This study is supported by grants 10233050, 10228307, 10311120115 and 10473016 of National Natural Science Foundation of China and TG 2000078401 of National Basic Research Program of China.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

102

H. Zhang

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34.

G. Ai, W. Li and H. Zhang, Acta Astronomica Sinica 23 (1982) 39. G. Ai and Y. Hu, Scien. Sinica A 30 (1987) 868. H. Zhang and G. Ai, Acta Astronomica Sinica 27 (1986) 226. H. Wang, J. Varsik, H. Zirin, R. Canﬁeld, K. Leka and J. Wang, Solar Phys. 142 (1992) 11. S. Bao, A. Pevtsov, T. Wang and H. Zhang, Solar Phys. 195 (2000) 75. H. Zhang, Solar Phys. 197 (2000) 235. H. Zhang, B. LaBonte, J. Li and T. Sakurai, Solar Phys. 213 (2003) 87. M. Hagyard, E. West, F. M. Trussart and E. G. Kenney, Como. Work (1992) 490H. T. Sakurai et al., PASJ 44 (1992) L7. B. Schmieder, P. Demoulin, M. Hagyard, M. E. Machado, G. Ai, H. Zhang, Q. Fu, Z. Li, T. Luan, B. Kalman and L. Gyori, AdSpR 13 (1993) 123. H. Zhang et al., Astrophys. J. 423 (1994) 828. S. Bao, G. Ai and H. Zhang, J. Astrophys. Astron. 21, 3–4 (2000) 303. J. Su and H. Zhang, Solar Phys. 222 (2004) 17. J. Su and H. Zhang, Chin. J. Astron. Astrophys. 4, 4 (2004) 365. H. Zhang, ApJ 557 (2001) L71. H. Wang, M. Ewell, H. Zirin and G. Ai, Astrophys. J. 424 (1994) 436. D. A. Falconer, JGR 106 (2001) 25185. H. Kurokawa, T. Wang and T. Ishii, Astrophys. J. 572 (2002) 598. H. Zhang, Chin. J. Astron. Astrophys. 4 (2004) 563. J. Chae, Astrophys. J. 560 (2001) L95. Y. J. Moon, J. Chae, G. S. Choe, H. M. Wang, Y. D. Park, H. S. Yun, V. Yurchyshyn and P. R. Goode, Astrophys. J. 574 (2002) 1066. Y. J. Moon, J. Chae, H. M. Wang, G. S. Choe and Y. D. Park, Astrophys. J. 580 (2002) 528. A. Nindos and H. Q. Zhang, Astrophys. J. 573 (2002) L133. K. Kusano, T. Maeshiro, T. Yokoyama and T. Sakurai, Astrophys. J. 577 (2002) 501. A. Nindos, J. Zhang and H. Q. Zhang, Astrophys. J. 594 (2003) 1033. P. Demoulin and M. A. Berger, Solar Phys. 215 (2003) 203. A. Nindos and M. D. Andrews, Astrophys. J. 616 (2004) L175. G. Yang, Y. Xu, W. D. Cao, H. M. Wang, C. Denker and T. R. Rimmele, Astrophys. J. 617 (2004) L151. E. Pariat, P. Demoulin and M. A. Berger, Astron. Astrophys. 442 (2005) 1105. H. Zhang, Monthly Notices for the RAS 326 (2001) 57. N. Seehafer, Solar Phys. 125 (1990) 219. A. A. Pevtsov, R. C. Canﬁeld and T. R. Metcalf, Astrophys. J. 440 (1995) L109. V. I. Abramenko, T. J. Wang and V. B. Yurchishin, Solar Phys. 168 (1996) 75. S. Bao and H. Zhang, Astrophys. J. 496 (1998) L43.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

Observational Study of Solar Magnetic Active Phenomena

35. 36. 37. 38. 39. 40. 41.

42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66.

103

H. Zhang and S. Bao, Astron. Astrophys. 339 (1998) 880. A. A. Pevtsov, J. Dun and H. Zhang, Solar Phys. 234 (2005) 203. M. Hagino and T. Sakurai, PASJ 56 (2004) 831. M. Hagino and T. Sakurai, PASJ 57 (2005) 481. A. Choudhuri, P. Chatterjee and D. Nandy, Astrophys. J. 615 (2004) L57. H. Zhang and S. Bao, Astrophys. J. 519 (1999) 876. L. Tian, H. Zhang and S. Bao, Understanding Solar Active Phenomena, eds. C. Fang, J. C. Henoux and M. D. Ding (World Publishing Corporation, International Academic Publishers, Beijing, P. R. China, 1999), p. 131. L. Tian, H. Zhang, S. Bao and H. Wang, Astron. Astrophys. 374 (2001) 294. K. Kuzanyan, S. Bao and H. Zhang, Solar Phys. 191 (2000) 231. N. Kleeorin, K. Kuzanyan, D. Moss, I. Rogachevskii, D. Sokoloﬀ and H. Zhang, Astron. Astrophys. 409 (2003) 1097. H. Zhang, D. Sokoloﬀ, I. Rogachevskii, D. Moss, K. Kuzanyan and N. Kleeorin, Monthly Notices for the RAS 365 (2006) 276. H. Zhang, T. Sakurai, K. Shibata, M. Shimojo and H. Kurokawa, Astron. Astrophys. 357 (2000) 725. S. Bao, H. Zhang, G. Ai and M. Zhang, Astron. Astrophys. Suppl. Ser. 139 (1999) 311. J. Chen, G. Ai, H. Zhang and S. Jiang, Publ., Yunnan Astron. Obs., Suppl. 33, Special Issue I (1989) 108. S. D. Bao, G. X. Ai and H. Q. Zhang, IAUS 203 (2001) 247. J. S. Kim, H. Q. Zhang, J. S. Kim, K. S. Kim and X. M. Bao, Chin. J. Astron. Astrophys. 2 (2002) 81. Y. Deng, J. Wang, Y. Yan and J. Zhang, Solar Phys. 204 (2001) 11. Y. Liu and H. Zhang, Astron. Astrophys. 386 (2002) 656. J. X. Wang, G. P. Zhou and J. Zhang, Astrophys. J. 615 (2004) 1021. H. Zirin, Solar Phys. 22 (1972) 34. R. G. Giovanelli, Solar Phys. 68 (1980) 49. R. G. Giovanelli and H. P. Jones, Solar Phys. 79 (1982) 267. A. H. Gabriel, Chromospheric ﬁne structure, IAU Symp. 56, ed. R. G. Athay (D. Redel Publ. Company, Dordrecht, 1974), p. 295. A. Title, RSPTA 358 (2000) 657. H. Zhang, G. X. Ai, T. Sakurai and H. Kurokawa, Solar Phys. 136 (1991) 269. M. Zhang and H. Q. Zhang, Astrophys. Reports (Publ. Beijing Astron. Obs.), Special Issue 4 (1998) 85. H. Zhang and M. Zhang, Solar Phys. 196 (2000) 269. M. Zhang and H. Q. Zhang, Solar Phys. 194 (2000) 29. J. O. Stenﬂo, Solar Magnetic Fields (Kluwer Academic Publishers, Dordrecht, 1994). M. Zhang, H. Q. Zhang, G. X. Ai and H. N. Wang, Solar Phys. 190 (1999) 79. T. E. Berger, B. De Pontieu, L. Fletcher, C. J. Schrijver, T. D. Tarbell and A. Title, Solar Phys. 190 (1999) 409. X. M. Bao and H. Q. Zhang, Chin. J. Astron. Astrophys. 6 (2003) 491.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch07

104

H. Zhang

J. S. Almeida, Astron. Astrophys. 324 (1997) 763. J. X. Wang and Z. X. Shi, Solar Phys. 140 (1992) 67. W. Li, G. X. Ai and H. Q. Zhang, Solar Phys. 151 (1994) 1. H. Zhang, Astrophys. Astrophys. Sup. 119 (1996) 205. H. Zhang, Astrophys. J. 471 (1996) 1049. H. Zhang, Publ. Beijing Astron. Obs. 26 (1995) 13. H. Alfven and R. C. Falthammer, Cosmical Electrodynamics (Oxford University Press, Oxford, 1963), p. 100. 74. Y. Hanaoka, PASJ 57 (2005) 235. 75. T. J. Wang, A. A. Xu and H. Zhang, Solar Phys. 155 (1994) 99. 67. 68. 69. 70. 71. 72. 73.

May 30, 2006 14:40 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch08

DEVELOPMENT OF STORM-TIME PROTON TOTAL ENERGY BASED ON MULTIOBSERVATION OF NOAA SATELLITES KEIKO T. ASAI∗ , TSUTOMU NAGATSUMA and TOMOAKI HORI National Institute of Information and Communications Technology 4-2-1 Nukuikita, Koganei, Tokyo 184-8795, Japan ∗[email protected] YOSHIZUMI MIYOSHI Solar-Terrestrial Environment Laboratory Nagoya University, Japan

Since July 2002, three polar orbiting NOAA/POES satellites (N15, N16, and N17) have observed particles in a wide range of local time at altitudes of about 850 km. The proton ﬂuxes in the energy range of ∼ eV to MeV, which can cover the energy spectra of the ring current particles are obtained from these satellites. We estimate the total energy content of protons in the inner magnetosphere and examine its time variation during storms. As the results, the time proﬁles of the total proton energy roughly correspond with those of the Dst index. However, the total energy estimated from the satellite observations is one-order less than that estimated from the Dst index. Based on the analysis of 24 storm events (Dst < −100 nT), it is found that the peak of the total energy tends to precede the minimum peak of Dst by a few hours. The time diﬀerence between two peaks is 3.6 h on average and shows a correlation with the magnitude of minimum Dst in a storm event.

1. Introduction We analyzed the multiobserved data sets from the NOAA/POES satellites, and examined the total proton energy considering the local time distribution based on the particle observation. The total energy of ring current particles is proportional to the decrease of geomagnetic ﬁeld intensity caused by the ring current according to the Dessler–Parker–Sckopke (DPS) relation.1 The DPS relation describes a linear relationship between perturbation of the geomagnetic ﬁeld on the ground and the total energy of ring current particles. The NOAA satellites monitor charged particles with polar orbits at altitudes above 800 km.2 Charged particles are bouncing in geomagnetic ﬂux tubes and detected near the mirror points by the satellites. There 105

May 30, 2006 14:40 WSPC/SPI-B368 Advances in Geosciences Vol. 2

106

ch08

K. T. Asai et al.

is an observational limit that particles detected by the satellites are the part of ring current particles. However, it is expected that time variation of proton ﬂux observed at low-altitude reﬂects that of ring current during geomagnetic storms.3 In the present paper, we report the comparison between the total proton energy and the Dst index regarded as the averaged H-components of geomagnetic ﬁelds. During a storm-time period, the H-components are reduced strongly by the developed ring current for a few days.

2. Data Three satellites, NOAA-15, NOAA-16, and NOAA-17, are orbiting in the local time regions from dawn to dusk, from post-noon to post-midnight, and from pre-noon to pre-midnight, respectively. Their orbital period is about 100 min. The combined data set obtained by all of these satellites covers all magnetic local times during a storm event. It is available since July 2002 after the launch of NOAA-17. These satellites have the detector set of the same version, called SEM-2, monitoring plasma and energetic particles in a very wide energy range. The instrumental details have been given by Evans and Greer.2 SEM-2 includes two sets of directional proton and electron sensors. TED-0 and -30 sensors cover the low-energy range (< 20 keV) and MEPED-0 and -90 sensors cover the high-energy range (> 30 keV). TED-0 sensors are directed outward along the zenith angle and TED-30 sensors are in an angle of 30◦ from TED-0. MEPED-0 sensors are directed outward almost between the view directions of TED-0 and -30. MEPED-90 sensors are directed perpendicular to MEPED-0. Field-of-view of these directional sensors is 30◦ . Therefore, MEPED sensors monitor both precipitating and trapped particles while TED sensors monitor mainly precipitating particles. The MEPED proton sensor has six energy channels: 30–80 keV (P1), 80–240 keV (P2), 240–800 keV (P3), 0.8–2.4 MeV (P4), 2.4–6.9 MeV (P5), and > 6.9 MeV (P6). The TED proton sensor has 16 energy channels but gives count rates for only four channels whose energy rages are centered at 189 eV (4), 844 eV (8), 2595 eV (11), and 7980 eV (14). The maximum count rate and its channel number are also given. The data from three satellites are sorted into time bins of 1.5 h in UT, which is the time interval near the orbital period of the satellites. The count data for each time bin are then averaged into spatial bins each of which is 0.2 in L-value and 4 h in magnetic local time (MLT). The coordinates

May 30, 2006 14:40 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch08

Development of Proton Total Energy During Storms

107

used for analysis, such as latitude, MLT, and L-value, are based on the corrected geomagnetic coordinates.4 The combined data from the three satellites shows good coverage for all MLT sector. Figure 1 shows an example of L-value versus time (L–T ) diagrams using the combined data obtained by the three satellites for September 3–9, 2002. The six color panels are for the six MLT sectors; 0–4, 4–8, 8–12, 12–16, 16– 20, and 20–24 MLT, respectively. Each panel has two L–T diagrams for protons in diﬀerent direction; trapped (upper) and precipitating (lower). The color code shows proton energy density (J/m3 ) in log-scale, and white bins correspond to the data missing periods. The energy density is estimated from the all proton channels of MEPED (30 keV to 6.9 MeV). The details of the proton energy density estimation are described in Sec. 3. For reference, the hourly values of the interplanetary magnetic ﬁeld (IMF) and solar wind parameters from the OMNI2 database of the NSSDC are plotted at the bottom of Fig. 1. In this period, two storms occurred: the ﬁrst minimum peak of Dst was –104 nT at 6 UT on September 4 (doy 247), and the second minimum peak was –170 nT at 0 UT on September 8 (doy 251). It is seen from Fig. 1 that the energy density increases during the two storms showing clear local time asymmetry. The intensity of the energy density is strongest in the premidnight sector (20–24 MLT) and weakest in the pre-noon sector (8–12 MLT). It is also found that the region of intense energy density dynamically changes during the period, particularly in the nightside sectors. The L-values of the peak energy density become lowest during the storm main phases.

3. Time Development of Proton Total Energy In the present study, we compare time variations of the total proton energy from the NOAA observations with those of the Dst index. The total energy is estimated from energy densities shown in Fig. 1. The details of the calculation are described in Sec. 3.1, and the statistical results on storm-time development of the total proton energy are shown in Sec. 3.2. 3.1. Calculation of proton energy density and total energy First, we assume that an observed count rate in unit of /cm2 /str/s keeps in the ﬂux tube. For every energy channels i, the number density Ni (/m3 ) is calculated from the observed directional count rate Ci by considering its energy width of the channel. Then, the energy density (J/m3 ) is locally

May 30, 2006 14:40 WSPC/SPI-B368 Advances in Geosciences Vol. 2

108

ch08

K. T. Asai et al.

Fig. 1. L–T diagrams for six local time sectors from September 3 to 9, 2002. The color code shows energy density (10−12 to 10−9 J/m3 in log-scale) of trapped (upper diagram of each panel) and precipitating (lower) protons in the energy range of 30 keV to 6.9 MeV. The bottom three panels show the geomagnetic indices and the hourly solar wind parameters: Dst (black) and Kp (green histogram) in the upper panel, IMF Bt (black), By (green), and Bz (red) in the middle panel, and solar wind velocity (black) and density (blue) in the bottom panel.

May 30, 2006 14:40 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Development of Proton Total Energy During Storms

ch08

109

Fig. 2. Energy–time diagrams show the count rates of precipitating protons averaged for all MLT sectors and L-values from 2.5 to 7. The upper spectrogram shows data from six channels of MEPED: P1 to P6 channels from the bottom. The lower spectrogram shows data from four channels of TED: numbers 4, 8, 11, and 14 from the bottom, and the additional maximum count rates. The part enclosed by red dots is the data of P1 (30–80 keV) channel.

Fig. 3. Proton total energies integrated from L = 1 to 10 in the top panel and are shown with the Dst index in the bottom panel; MEPED trapped protons (black), MEPED precipitating protons (green), and MEPED and TED precipitating protons (blue). Red lines are marked at the peaks for the each of the two storms.

given as the integral of Ni multiplied by the center energy Ei (J) for all energy channels. For example, the top panels of Fig. 2 show the energy versus time (E–T ) diagrams of precipitating protons for each energy channel of TED and MEPED. The vertical axis roughly shows the log-scaled energy (Sec. 2). The color-coded values are the count rates averaged for all MLT sectors and L-values from 2.5 to 7 (in log-scaled/cm2/str/s). It is found that the protons in the P1 channel (30–80 keV) show the strongest counts during the storm periods.

May 30, 2006 14:40 WSPC/SPI-B368 Advances in Geosciences Vol. 2

110

ch08

K. T. Asai et al.

Next, the total energy of protons is derived from integrating the energy densities of each bin. An energy density of each bin of 0.2 L × 4 h MLT is multiplied by the diﬀerential ﬂux tube volume of the dipole ﬁeld, and then integrated for L-values from 1 to 10 and for all MLT sectors. In this process, we assume that the density is constant within the dipole ﬂux tube based on the Liouville theorem. We examine the data of both precipitating and trapped protons. The precipitating protons are with pitch angles < 45◦ (> 135◦ ) in the northern (southern) hemisphere while the trapped protons are with pitch angles between 45◦ and 135◦ . The count rates of the energy spectrograms in Fig. 2 are of precipitating protons detected by both MEPED and TED. The data of trapped protons detected by TED are not used in the present study because the count rates are negligible compared with the other sensors. In the upper panel of Fig. 3, three lines are time variations of the total energies calculated from trapped protons detected by MEPED (black), precipitating protons detected by MEPED (green), and precipitating protons detected by both MEPED and TED (blue), in the linear scale. It is found that a quantitative agreement between the black and green lines indicates that the total energies calculated with the ﬁrst assumption have almost the same values for the trapped and precipitating protons of MEPED. The blue dot lines indicate the sum of total energies of precipitating protons by MEPED and TED. These are almost twice as the green dashed lines. As a note, data gaps as shown with colored white in Fig. 1 are not interpolated to calculate these total energies. The storm-time depression of the geomagnetic ﬁelds on the ground represented from the Dst variation suggests the energy over 1015 J of the ring current.3,5 However, the result shown in Fig. 3 indicates that the total proton energy of ∼ 1014 J is one order smaller than the expectation. It should be noticed that the major component of the ring current is in the magnetospheric equator while the NOAA observation is at low altitudes. The particles at the NOAA altitudes all become parallel to the ﬁelds at the equator. If the equatorial pitch angle distribution is isotropic, our calculation shows quantitative agreement. However, in situ observations and simulations in the past studies indicated that the ring current particles have possibly a pancake-type pitch angle distribution which has a larger perpendicular component.1,6,7 The averaged count rates shown in Fig. 2 are one- or two-order smaller than the POLAR observations.8,9 Therefore, our calculation based on the low-altitude observation shows an underestimation of the energy density.

May 30, 2006 14:40 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Development of Proton Total Energy During Storms

ch08

111

3.2. Time development of proton total energy In Fig. 3, we can identify two energy peaks near the Dst peaks as marked by vertical red lines in the upper panel. This indicates that the time variation of the total energy based on the NOAA observation generally corresponds to the time proﬁle of the equatorial ring current energy. Comparing the time variations of the total proton energy with the Dst variation, it is found that the ﬁrst peak of total energy almost coincides with the ﬁrst minimum of Dst, while the second peak of total energy seems to precede the second minimum of Dst for about 5 h. Using 24 cases of storm events (Dst < −100 nT), we investigate the time diﬀerences between the peak of energy density and the minimum peak of the Dst index for each storm event. The hourly Dst index includes variations on shorter time scales, for example, ionospheric currents, currents owing to auroral activity at high latitudes.5,10,11 In our analysis, however, these eﬀects are neglected because of the analytical time resolution of 1.5 h. We identiﬁed 24 minimum peaks of Dst and these neighboring maximum peaks of total proton energy. Each of the minimum peaks of Dst is the minimum for the period of two days. It is found that the energy peaks precede Dst peaks in the most cases. The time diﬀerences between the peaks of Dst and of total proton energy versus the values of minimum Dst are plotted in Fig. 4. The negative value of the time diﬀerence indicates 0

Minimum Dst [nT]

-50 -100 -150 -200 -250 -300 -350 -400 -450 -500 -15

-10

-5

0

5

Interval [hour] between proton energy peak and min-Dst

Fig. 4. Time diﬀerences between peaks of total proton energy and minimum peaks of Dst are plotted versus the magnitudes of the minimum Dst for 24 cases. The negative values of intervals indicate the preceding of proton energy to the Dst peaks. A correlation between intervals and magnitude of storms is seen as roughly enclosed by a grey ellipse.

May 30, 2006 14:40 WSPC/SPI-B368 Advances in Geosciences Vol. 2

112

ch08

K. T. Asai et al.

that the peak of total proton energy precedes the Dst peak. The time diﬀerence is 3.6 h on average. In addition, we ﬁnd a correlation between the time diﬀerences and the storm intensities. In particular, the points enclosed by the grey ellipse show a good correlation for storms with the Dst minimums > −200 nT. Some points for storms with Dst < −250 nT outside the ellipse do not show any obvious correlation, however, the largest time diﬀerence was obtained for the well-known intense storm with Dst < −350 nT so-called the Halloween event in 2003. This result suggests that the time intervals can be longer for stronger storms. We comment on cases with Dst > −100 nT, which were excluded from the present study. Identifying the minimum peaks of Dst for such cases is diﬃcult because their Dst variations are complicated including eﬀects of HILDCAAs.10,12 We will investigate carefully them in the future.

4. Discussion and Conclusions It is well known that the storm-time ring current has MLT asymmetry, based on the ground geomagnetic observations and the particle simulations.12–16 The Dst index is the average of such asymmetrical geomagnetic variation. In the present study, the total proton energy of ring current was calculated from the NOAA data covering all MLT sectors. The distribution of protons has an obvious MLT asymmetry as shown in Fig. 1. We compared the time variations between the total proton energy and the Dst index because both of them were regarded as the MLT averages. As a result, the peak of total proton energy tends to precede the minimum peak of Dst . The time diﬀerence of the precedence is 3.6 h on average and shows a correlation with the storm intensity. Considering that the storm-time ring current is centered at L-values of 3–4 in the magnetospheric equator,17 the interpretation of these results are as follows: (1) The preceding peak of total energy is contributed largely from protons at high-L values. For example as shown in the L–T diagrams of Fig. 1, the distribution of protons move from higher-L (maybe auroral region) to lower-L during the initial and main phase of storms. (2) The main part of the ring current particles making geomagnetic depression can not be observed by NOAA at low altitudes. (3) There is diﬀerent developing mechanism for intense storms than for small storms. In the future, a study considering the other ion species will be necessary because it was reported that those ion ﬂuxes (He+ , O+ , etc.) increase during storms, especially super storms.8,18–20 It is interesting to examine

May 30, 2006 14:40 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Development of Proton Total Energy During Storms

ch08

113

the relation to the solar wind and IMF Bz conditions, which control the magnitude of storm.21,22

Acknowledgments We thank NSSDC for providing the OMNI2 data, WDC Kyoto for providing the geomagnetic indices, and specially NGDC for providing the NOAA/POES data.

References 1. M. W. Liemohn, J. Geophys. Res. 108 (2003) 1251. 2. D. S. Evans and M. S. Greer, NOAA Technical Memorandum OAR SEC-93 (Boulder, Colorado, 2000) . 3. F. Søraas, K. Aarsnes, K. Oksavik and D. S. Evans, J. Geophys. Res. 107 (2002) 7. 4. G. Gustafsson, N. E. Papitashvili and V. O. Papitashvili, J. Atmos. Terr. Phys. 54 (1992) 1609. 5. N. E. Turner, D. N. Baker, T. I. Pulkkinen, J. L. Roeder, J. F. Fennell and V. K. Jordanova, J. Geophys. Res. 106 (2001) 19149. 6. A. T. Y. Lui, R. W. McEntire, D. G. Sibeck and S. M. Krimigis, J. Geophys. Res. 95 (1990) 20839. 7. J. D. Perez, X.-X. Zhang, P. C. Brandt, D. G. Mitchell, J.-M. Jahn, C. J. Pollock and S. B. Mende, J. Geophys. Res. 109 (2004) A09292. 8. T. I. Pulkkinen, N. Y. Ganushkina, D. N. Baker, N. E. Turner, J. F. Fennell, J. Roeder, T. A. Fritz, M. Grande, B. Kellett and G. Kettmann, J. Geophys. Res. 106 (2001) 19131. 9. Y. Ebihara, M. Ejiri, H. Nilsson, I. Sandahl, A. Milillo, M. Grande, J. F. Fennell and J. L. Roeder, Geophys. Res. Lett. 29 (2002) 1969. 10. W. D. Gonzalez, J. A. Joselyn, Y. Kamide, H. W. Kroehl, G. Rostoker, B. T. Tsurutani and V. M. Vasyli¯ unas, J. Geophys. Res. 99 (1994) 5771. 11. T. P. O’Brien and R. L. McPherron, J. Geophys. Res. 105 (2000) 7707. 12. F. Soraas, K. Aarsnes, K. Oksavik, M. I. Sandanger, D. S. Evans and M. S. Greer, J. Atmos. Solar-Terr. Phys. 66 (2004) 177. 13. N. E. Turner, D. N. Baker, T. I. Pulkkinen and R. L. McPherron, J. Geophys. Res. 105 (2000) 5431. 14. K. K. Hashimoto, T. Kikuchi and Y. Ebihara, J. Geophys. Res. 107 (2002) 1337. 15. V. K. Jordanova, A. Boonsiriseth, R. M. Thorne and Y. Dotan, J. Geophys. Res. 108 (2003) 1443. 16. F. Søraas, K. Oksavik, K. Aarsnes, D. S. Evans and M. S. Greer, Geophys. Res. Lett. 30 (2003) 1052. 17. A. M. Jorgensen, H. E. Spence, W. J. Hughes and H. J. Singer, J. Geophys. Res. 109 (2004) A12204.

May 30, 2006 14:40 WSPC/SPI-B368 Advances in Geosciences Vol. 2

114

ch08

K. T. Asai et al.

18. I. A. Daglis, J. U. Kozyra, Y. Kamide, D. Vassiliadis, A. S. Sharma, M. W. Liemohn, W. D. Gonzalez, B. T. Tsurutani and G. Lu, J. Geophys. Res. 108 (2003) 1208. 19. C. G. Mouikis, L. M. Kistler, W. Baumjohann, E. J. Lund, A. Korth, B. Klecker, E. M¨ obius, M. A. Popecki, J. A. Sauvaud, H. R`eme, A. M. Di Lellis, M. McCarthy and C. W. Carlson, Geophys. Res. Lett. 29 (2002) 1432. 20. W. L. Liu, S. Y. Fu, Q.-G. Zong, Z. Y. Pu, J. Yang and P. Ruan, Geophys. Res. Lett. 32 (2005) L15102. 21. T. P. O’Brien and R. L. McPherron, J. Geophys. Res. 107 (2002) 1341. 22. R. P. Kane, J. Geophys. Res. 110 (2005) A02213.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch09

ON THE CROSS-FIELD DIFFUSION OF GALACTIC COSMIC RAYS INTO AN ICME K. MUNAKATA∗, , S. YASUE∗ , C. KATO∗ , J. KOTA† , M. TOKUMARU‡ , M. KOJIMA‡ , A. A. DARWISH§ , T. KUWABARA¶ and J. W. BIEBER¶ ∗Department of Physics, Shinshu University Asahi 3-1-1, Matsumoto, Nagano 390-8621, Japan †Lunar ‡Solar

Terrestrial Environmental Laboratory, Nagoya University Nagoya, Japan

§Physics ¶Bartol

and Planetary Laboratory, University of Arizona Tucson, AZ, USA

Department, Faculty of Science, Alexandria University Alexandria, Egypt

Research Institute and Department of Physics and Astronomy University of Delaware, Newark, Delaware, USA [email protected]

We develop a numerical model of the cross-ﬁeld diﬀusion of galactic cosmic rays into an interplanetary coronal mass ejection (ICME), on the assumption that the local part of the ICME is an expanding straight cylinder. It is found that the spatial distribution of cosmic ray density in the cylinder rapidly reaches a stationary state due to the balance between inward diﬀusion and adiabatic cooling in the expanding cylinder. By ﬁtting the model to the Halloween ICME event observed with the network of muon detectors in October 2003, we evaluate the magnitude of the cross-ﬁeld diﬀusion coeﬃcient to be 1.7 × 1021 cm2 /s at ∼ 50 GeV.

1. Introduction When the Interplanetary counterpart of a Coronal Mass Ejection (ICME) accompanied by a strong shock travels through interplanetary space, it often forms a depleted region of Galactic cosmic rays behind the shock and within the ICME, changing dramatically the pre-existing spatial distribution of cosmic rays. When Earth enters the depleted region, ground-based cosmic ray detectors record a Forbush Decrease.1 This change in the spatial distribution can be observed by cosmic ray (CR) detectors at Earth as a dynamic variation of the directional anisotropy of CR intensity (or the CR streaming), since the CR anisotropy corrected for convection by the solar wind is solely due to spatial diﬀusion and drift ﬂuxes that are 115

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch09

116

K. Munakata et al.

proportional to the spatial gradient of the cosmic ray density (the isotropic component of the intensity) in local space. Bieber and Evenson2 reported such strong enhancements of the anisotropy observed by a network of seven neutron monitors and concluded that “B × ∇n” drift is a primary source of ICME-related anisotropies. They demonstrated for the ﬁrst time that the evolution of CR density and density gradients is closely linked to magnetic properties of the ejecta, and provides information on the structure and orientation of the ICME as it approaches and passes Earth.3 Hofer and Fl¨ uckiger4 also analyzed the anisotropy observed by neutron monitors during a large Forbush Decrease in March 1991 and demonstrated the potential capability of CR observations for providing information on complex transient structures in the near-Earth interplanetary medium. Kuwabara et al.5,6 modeled the CR depleted region in the Halloween ICME observed by the muon detector network on October 29, 2003, by a straight cylinder and deduced the three-dimensional (3D) geometry of the cylinder. They also compared the geometry derived from cosmic rays with that derived from in situ interplanetary magnetic ﬁeld (IMF) observations using an expanding Magnetic Flux Rope (MFR) model, and demonstrated that these two geometries based on independent observations are in a reasonable agreement. Cane et al.7 presented for the ﬁrst time a quantitative study of the cross-ﬁeld diﬀusion of CRs into an ICME and derived the density distribution in the ICME. Their model, however, assumed a stationary ICME with a constant radius ignoring the adiabatic cooling of CRs that occurs in an expanding ICME. Such a model is not applicable to the Halloween ICME, in which the in situ observations of the IMF clearly indicate expansion.5,6 In the present paper, we develop a numerical model for the cross-ﬁeld diﬀusion of CRs into an expanding ICME, taking account of the adiabatic cooling eﬀect due to expansion. In Sec. 2, we ﬁrst derive the CR density distribution in the ICME based on the transport equation of CRs. In Sec. 3, we apply the model to the Halloween ICME event observed by the muon detector network and deduce the magnitude of the cross-ﬁeld diﬀusion coefﬁcient appropriate to the observation. Due to the closed ﬁeld geometry, CRs can penetrate in the MFR only through cross-ﬁeld diﬀusion. This provides us with a unique opportunity to precisely evaluate the cross-ﬁeld diﬀusion coeﬃcient, which is one of the most diﬃcult physical parameters to estimate from observations. We note that MFRs are a subset of ICME with a special magnetic structure (see Ref. 5 and references therein for details). Our Forbush decrease

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch09

On the Cross-Field Diﬀusion of Galactic Cosmic Rays into an ICME

117

model derived in Sec. 2, does not explicitly assume MFR structure, and thus should apply generally to any ICME that can be modeled as an expanding cylinder. However, our method used to analyze the Halloween event in Sec. 3, employs parameters derived from a MFR analysis of the ICME, and thus can only be applied to the MFR subset of ICME.

2. Model and Numerical Solutions 2.1. Transport equation The axisymmetric distribution of the CR density in a cylinder is governed by the following transport equation for the cross-ﬁeld diﬀusion of CRs into the ICME, which is assumed to be a cylinder in this paper.

κ⊥ ∂ 1 ∂ ∂f ∂f ∂f ∂f = + (rV ) , (1) r −V ∂t r ∂r ∂r ∂r 3r ∂r ∂ ln p where f (r, p, t) is the omnidirectional phase space density of CRs with momentum p at a radial distance r from the ICME (cylinder) axis and time t and V is the radial expansion velocity of the ICME. The ﬁrst and second terms on the right hand side denote respectively the cross-ﬁeld diffusion and the convection in the expanding plasma, while the third term denotes the adiabatic cooling due to expansion. We rewrite (1) for f (x, p, s) = f (r, p, t) by replacing r and t, respectively with dimensionless quantities x and s, deﬁned as,

t r and s = log , (2) x= R(t) tc with R(t) denoting the radius of the ICME envelope at time t and tc denoting an arbitrary reference time. We assume self-similar expansion of the ICME3 with radius R(t) and expansion velocity V deﬁned as r , t Rc t R(t) = , tc

V (r, t) =

(3) (4)

with Rc denoting R(t) at t = tc . In the following analyses, we deﬁne t = tc as the time of Earth’s ﬁrst contact with the ICME envelope. We also assume κ⊥ independent of x, but proportional to the radius of the ICME envelope. This seems to be a reasonable assumption, since the self-similar expansion

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch09

118

K. Munakata et al.

of the ICME will increase the diﬀusion mean-free-path as well. We assume κ⊥ can be expressed as κ⊥ = κ0 Vc R(t) ,

(5)

where κ0 is a dimensionless parameter denoting the degree of the cross-ﬁeld diﬀusion, and Vc the expansion velocity of the ICME envelope at tc . We ﬁnally assume a single power momentum spectrum for f , as f (x, p, s) = p−(2+γ) F (x, s) ,

(6)

with the spectral index γ set equal to 2.7, as appropriate for high-energy Galactic cosmic rays. We thus obtain the equation to be solved numerically, as

2 ∂ F 1 ∂F ∂F 2(2 + γ) = κ0 F. (7) + − ∂s ∂x2 x ∂x 3 Note that the convection term in (1) does not appear in this equation. 2.2. Numerical solutions We solve (7) numerically with an initial condition that the CR density is zero inside and uniform outside the ICME (i.e., starting from an “empty cylinder”). More practically, we set F = 0 for x < 1.0 and F = 1 for x ≥ 1.0 as the initial condition at s = −4.605 (t = 0.01 × tc ). Figure 1 shows numerical solutions of (7). As seen in this ﬁgure, the spatial distribution F (x) for κ0 > 1 rapidly reaches an equilibrium due to the balance between inward diﬀusion (causing an increase of F ) and adiabatic cooling (causing a decrease of F ) within s < −2.303 (t < 0.1 × tc ). Since we deﬁne s = 0.0 (t = tc ) as the time of Earth’s ﬁrst contact with the ICME surface, Fig. 1 implies that F is already stationary when the ICME arrives at Earth. The magnitude of the maximum density depression (1 − F (0, ∞)) in this stationary distribution is shown as a function of κ0 in Fig. 2(a). As the maximum density depression in Forbush Decreases observed by muon detectors is typically 1 ∼ 10% (0.01 ∼ 0.1), this ﬁgure implies that the magnitude of κ0 appropriate to the observation should be 10–50. The stationary distribution F stat (x) is given by (7) with the left hand side set equal to zero, as

2 stat ∂ F 1 ∂F stat 2(2 + γ) stat F + . (8) κ0 = 2 ∂x x ∂x 3

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch09

On the Cross-Field Diﬀusion of Galactic Cosmic Rays into an ICME

1

1

0.001 ~ 1.0

119

1

0.0005

0.8

3 5

F(x,t)

1-F(0,t)

0.0002

0.6 0.4

10 30

κ0 = 10

0.2 0

0.1

50

0.01

100

0.00002

-1

-0.5

0

0.5

0.01

1

0.012

x

t/t c

(a)

(b)

0.014

Fig. 1. Numerical solutions of (7). (a) The density distributions are plotted as a function of normalized radial distance x from the ICME axis at diﬀerent times. Each number attached to the curve indicates the normalized time after tc (t/tc − 1.0). The numerical distribution for x ≥ 0 is repeated for x < 0 to clarify the physical distribution. (b) The temporal evolution of the magnitude of maximum density depression at x = 0 on the ICME axis. Each number attached to the curve indicates the value of κ0 .

1

0.98

numerical polynomial

0.96

κ0 =10

0.1

F stat (x)

1-F(0, ∞ )

1

0.01

0.94

0.001

10

κ0 (a)

100

0.92 -1

-0.5

0

0.5

1

x (b)

Fig. 2. Stationary solutions of (7). (a) The magnitude of maximum density depression in numerical solution is plotted as a function of κ0 . (b) The stationary distribution for κ0 = 10 is plotted in the same manner as Fig. 1(a). The full circles connected by a line display the numerical solution, while the open circles represent the polynomial solution (see text).

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch09

120

K. Munakata et al.

The solution is a Bessel function which we approximate with a polynomial: ∞ F stat (x) = an xn , (9) n=0

then we obtain,

an =

Γ n2

an−2

for n = 2, 4, 6, . . .

(10)

and an = 0

for n = 1, 3, 5, . . .

(11)

where Γ=

2(2 + γ) . 3κ0

(12)

Figure 2(b) shows this polynomial solution for κ0 = 10 and n ≤ 6, together with F (x, ∞) obtained by solving (7) numerically. It is seen that the numerical solution is well reproduced by the polynomial expression with n ≤ 6. In Sec. 3, we will use this expression for best-ﬁtting to the observed data. 3. Best-Fitting to the Halloween ICME Event In this section, we derive the magnitude of κ0 appropriate to the Halloween ICME event observed on October 29, 2003, by best-ﬁtting the model in the previous section to the observed data. Although a substantial CR depression also commenced at the time of the passage of a strong shock ahead due to the modulation by the post-shock region, we restrict ourselves in modeling only the modulation in the ejecta behind the shock in this paper. A more complete model will be given elsewhere in the future. Using the polynomial solution F stat (x) in (9), we get the expected fractional density depression I exp (ti ) and the fractional density gradient vector exp (ti ) at time ti , as g⊥ Γ2 Γ exp 2 4 x(ti ) + · · · , (13) I (ti ) = a0 1 + x(ti ) + 4 64 a0 Γ2 Γ exp x(ti ) + x(ti )3 + · · · e⊥ (ti ) , g⊥ (ti ) = − exp (14) I (ti ) 2 16 where e⊥ (ti ) is the unit vector pointing to the Closest Axial Pont (CAP) on the cylinder axis from Earth at ti . For detailed deﬁnitions of the fractional density depression and gradient, readers can refer to Kuwabara et al.5 Note exp that g⊥ (ti ) is independent of the parameter a0 denoting the magnitude exp of I (ti ). We can calculate x(ti ) and e⊥ (ti ) in (13) and (14), as follows.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch09

On the Cross-Field Diﬀusion of Galactic Cosmic Rays into an ICME

121

The position vector of the CAP as viewed from Earth, PE (ti ), is given by ¯ SW − (eaxis · V ¯ SW )eaxis }(ti − tc ) + Pc , PE (ti ) = {V

(15)

¯ SW is the average radial solar wind velocity, eaxis is the unit vector where V parallel to the axis derived from the MFR analysis5 and Pc is the CAP position at t = tc . Pc is calculated from the impact time (td ) and location (d) of the MFR, as ¯ SW (td − tc ) . Pc = d + V

(16)

With PE (ti ) in (15) and R(t) in (4), we get x(ti ) and e⊥ (ti ), as |PE (ti )| , R(t) PE (ti ) e⊥ (ti ) = . |PE (ti )| x(ti ) =

(17) (18)

We repeat the calculations for the expected density and gradient vector by introducing (17) and (18) into (13) and (14) for various values of κ0 , and ﬁnd the best-ﬁt κ0 minimizing the residual S deﬁned, as N 1 obs (t ) − gexp (t )|2 } . {|I obs (ti ) − I exp (ti )|2 + |g⊥ (19) S= i i ⊥ 4N i=1 Note that the best-ﬁt value of a0 for each κ0 is uniquely given from the least-squares requirement, ∂S 2 /∂a0 = 0, as exp N obs (ti ) ∂I ∂a0(ti ) i=1 I (20) a0 = exp 2 . N ∂I (ti ) i=1

∂a0

We perform the best-ﬁtting calculation described above using the MFR parameters listed in Table 1, which were derived from our analysis of the Table 1. MFR parameters used for the best-ﬁt calculation for Halloween ICME event. These parameters were derived from in situ IMF observations using an expanding MFR model.5,6 MFR period Time of the ﬁrst encounter with MFR (tc ) Radius of MFR at tc (Rc ) Time of the impact with MFR (td ) Location of MFR at the impact (dx , dy , dz ) (AU)∗ Latitude and longitude of the MFR axis direction (θ, φ)* Average solar wind velocity (V¯SW ) Expansion velocity of MFR (Vc ) ∗ Values

in the GSE coordinate system.

302.47–303.09 doy 302.47 doy 0.174 AU 302.679 doy (0.000, −0.077, 0.060) AU (46◦ , 54◦ ) 1323 km/s 0.209 AU/day

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch09

122

K. Munakata et al.

density (%)

93 92 91 90 89

observed best-fit

2 0 -1

0.8

S (%)

2

⊥y

0

Rg

L

1 -1

2

0.7

-2

⊥z L

1 0 -1

Rg

0.9

-2

(%)

(%)

Rg

⊥x

1

L

(%)

88

0.6

-2 302.4

302.6

302.8

0

10

20

κ0

30

40

50

303

doy Fig. 3. Best-ﬁtting model to the Halloween ICME event. Left panels show the density and three GSE components of the gradient vector, each as a function of time in the day of year (doy) in 2003. Open circles display the observed data, while the full circles connected by a line show the model best-ﬁt to the data (see text). Right panel shows the mean residual of the best-ﬁtting as a function of κ0 . The minimum residual (S = 0.65%) is found at κ0 = 18.

in situ IMF observations of the Halloween MFR.6 The best-ﬁt I exp (ti ) and exp (ti ) are compared with the observed data in Fig. 3. In this ﬁgure, we g⊥ exp (ti ) by the eﬀective particle’s gyroradius (RL = 0.044 AU) multiplied g⊥ derived from the CR cylinder analysis of the Halloween event for the direct comparison with the gradient vector deduced from the observed CR anisotropy.6 It is seen in this ﬁgure that a reasonable agreement between the observed and modeled values is achieved even with such a simple model. Figure 3 also displays S as a function of the parameter κ0 and indicates that the best-ﬁt is achieved with κ0 = 18 (a0 = −12%) and S = 0.65%. The actual value of κ⊥ at t = tc can be deduced from (4) and (5) with Vc

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch09

On the Cross-Field Diﬀusion of Galactic Cosmic Rays into an ICME

123

and Rc in Table 1, as κ⊥ = κ0 Vc Rc = 1.7 × 1021 cm2 /s .

(21)

4. Summary and Conclusion We developed a model of the cross-ﬁeld diﬀusion of galactic CRs into an ICME based on the assumption that the local part of the ICME is an expanding straight cylinder. It is found that the spatial distribution (as a function of the normalized radial distance from the cylinder axis) of CR density in the rope rapidly reaches an equilibrium due to the balance between inward diﬀusion and adiabatic cooling in the expanding cylinder. This implies that the distribution is already stationary when the ICME arrives at Earth. By best-ﬁtting the model distribution to the data observed by the muon detector network during the Halloween ICME/MFR event, the magnitude of the cross-ﬁeld diﬀusion coeﬃcient is evaluated to be κ⊥ = 1.7 × 1021 cm2 /s. According to analyses of the diurnal anisotropy observed by muon detectors, the long-term average of the parallel mean free path of CRs is ∼ 2.0 AU.8,9 This implies κ// =

1 ∼ 3 × 1023 cm2 /s and 3λ// c

κ⊥ ∼ 0.0057 κ//

(22)

for CRs with the median energy of ∼ 50 GeV, to which surface muon detectors have major responses. This value of κ⊥ /κ// is consistent with theoretical expectations for the pitch angle scattering of CRs in the turbulent magnetic ﬁeld in interplanetary space.10 Note that the mean free path in MFR is likely to be longer than average, due to the exceptionally strong and smooth magnetic ﬁelds. Thus, the value in (22) might be regarded as an upper limit.

Acknowledgments This work is supported in part by US NSF Grant ATM-0207196, and in part by Scientiﬁc Research (JSPS) in Japan and by the joint research program of the Solar-Terrestrial Environment Laboratory, Nagoya University.

References 1. H. V. Cane, Space Sci. Rev. 93 (2000) 55. 2. J. W. Bieber and P. Evenson, Geophys. Res. Lett. 25 (1998) 2955.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch09

124

K. Munakata et al.

C. J. Farrugia et al., J. Geophys. Res. 98 (1993) 7621. M. Y. Hofer and E. O. Fl¨ uckiger, J. Geophys. Res. 105 (2000) 23085. T. Kuwabara et al., Geophys. Res. Lett. 31 (2004) L19803-1. T. Kuwabara, Ph.D. thesis, Shinshu University (2005) (in Japanese). H. V. Cane, I. G. Richardson and G. Wibberenz, Proc. 24th International Cosmic Ray Conference 4 (1995) 377. 8. K. Munakata et al., Proc. 25th International Cosmic Ray Conference 2 (1997) 77. 9. K. Munakata et al., Adv. Space Res. 29 (2002) 1527. 10. J. W. Bieber, W. H. Matthaeus and A. Shalchi, Geophys. Res. Lett. 31 (2004) L10805-1. 3. 4. 5. 6. 7.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch10

ON THE UPPER LIMITING ENERGY OF THE SOLAR DIURNAL ANISOTROPY OF GALACTIC COSMIC RAY INTENSITY K. MUNAKATA∗,†† , S. YASUE∗ , C. KATO∗ , S. AKAHANE∗ , M. KOYAMA∗ , S. MORI∗ , A. A. DARWISH† , H. TSUCHIYA‡ , H. ONUMA§ , K. MIZUTANI§ , T. YUDA¶ , M. TAKITA and J. KOTA∗∗ ∗Department of Physics, Shinshu University Matsumoto 390-8621, Japan †Physics Department, Faculty of Science, Alexandria University, Egypt ‡RIKEN, Wako 351-0198, Japan §Department of Physics, Saitama University, Saitama, Japan ¶Faculty of Engineering, Kanagawa University, Yokohama, Japan Institute for Cosmic Ray Research, University of Tokyo, Kashiwa, Japan ∗∗Lunar and Planetary Laboratory, University of Arizona Tucson, AZ, USA ††[email protected]

The long-term variation of the solar diurnal anisotropy of the galactic cosmic ray intensity in the sub-TeV region is studied by using data recorded over 20 years with the deep underground muon detector at Matsushiro in Japan since 1984. The harmonic dial of the diurnal variation shows a clear 11-year change. At around 1987 and 1998, the variation has a maximum at 06:00 local solar time (LST) being fairly consistent with the Compton–Getting anisotropy due to the Earth’s orbital motion around the Sun, while in 1992 and 2003, it contains an additional variation with the maximum around 15:00 LST. The observed amplitude of this additional variation varies with 11-year cycle and the amplitude reaches a maximum of about 0.04% a few years after sunspot maximum. We discuss these results in the relation to the upper limiting energy of solar modulation in the heliosphere.

1. Introduction Galactic cosmic rays (CR) are high-energy nuclei (mostly protons) accelerated in our galaxy and continuously arriving at Earth after traveling through the heliosphere. From the observations of CR at Earth, deep inside the heliosphere, we can deduce information on the physical condition of space through which CR traveled to Earth. After entering the heliosphere, CR interact with the solar magnetic ﬁeld being carried outward by the solar wind. The interaction with the large-scale ordered ﬁeld causes the gradient- and curvature-drift motions of CR in the inner heliosphere, while 125

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch10

126

K. Munakata et al.

the interaction with the irregular (or disordered) ﬁeld component results in the pitch angle scattering of CR. The pitch angle scattering by the magnetic irregularities embedded in expanding radial solar wind causes the outward convection and the deceleration (called the adiabatic cooling) of CR and produces nonuniform distribution, in which the density decreases with decreasing distance from the Sun.1 This radial density gradient causes the inward diﬀusion of CR mainly along the ordered ﬁeld. The inward diffusion together with the outward convection produce the CR streaming called the convection–diﬀusion–drift (CDD) anisotropy, which is observed as the solar diurnal variation (SDV) of the CR intensity recorded in a ﬁxed directional channel of the detector on the spinning Earth. The typical CDD anisotropy in space has an amplitude of ∼ 0.5% with the maximum intensity at 12:00–18:00 local solar time (LST); the average interplanetary magnetic ﬁeld (IMF) lies along 09:00–21:00 LST at Earth’s orbit. The CDD anisotropy is most notable at Earth in the intensity of the relatively low energy CR below several 10 GeV. Those CR are completely scattered, losing their original directions of motion outside the heliosphere. In the higher-energy region above ∼ 100 GeV, CR become less sensitive to the modulation eﬀects because the pitch angle scattering by the small scale irregularities in the IMF are no longer signiﬁcant. This is seen in the energy dependence of the SDV; the amplitude of which decreases with increasing energy above ∼ 100 GeV. This implies that there must an upper limiting energy (ULE), above which the CDD mechanism cannot produce the solar diurnal anisotropy.2 In addition to the CDD anisotropy mentioned above, there is another type of diurnal anisotropy called the Compton–Getting (CG) anisotropy due to the orbital motion of Earth around the Sun.3 When a detector on Earth moves with respect to the rest frame of CR, the fractional enhancement due to the CG anisotropy is expressed as υ ∆I = (γ + 2) cos θ , I c

(1)

where I is the cosmic ray intensity, γ the power-law index of the cosmic ray energy spectrum, υ/c the ratio of the detector’s velocity to the particle speed which is practically the speed of light, and θ is the angle between the arrival direction of CR and the direction of detector’s motion.4 This anisotropy is expected to have an amplitude of ∼ 0.05% independent of CR energy with the maximum intensity at 06:00 LST in space. Note that this anisotropy is extremely stable as it is expressed in terms of γ and υ in (1),

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch10

On the Upper Limiting Energy of the Solar Diurnal Anisotropy

127

both of which show only negligible temporal variations for CR with energy above ∼ 10 GeV. Since the amplitude of the CG anisotropy is 10 times smaller than the CDD anisotropy in the low energy region below ∼ 10 GeV, it becomes notable only in the high energy region, near or above the ULE where the CDD anisotropy diminishes. An air-shower experiment at Yangbajing in Tibet has recently succeeded, with the highest precision ever reported, in observing the daily variation caused by the CG anisotropy of ∼ 10 TeV CR intensity.5 The Tibet experiment also reported that the variation observed in the lower portion of the primary energy deviates signiﬁcantly from the expectation, suggesting the solar modulation eﬀect possibly extending up to multi-TeV energies. In the present paper, we examine the ULE of the solar modulation by analyzing the solar diurnal anisotropy observed over 20 years at sub-TeV energies with a Japanese deep underground muon detector.

2. Data and Analysis Method An underground muon detector has been in operation at Matsushiro, Japan since 1984, recording the muon intensity produced by the hadronic interaction of primary CR with atmospheric nuclei in the high-altitude atmosphere. The vertical depth of the overburden at the observation site is 220 m of water equivalent. The multi-directional muon detector (hereafter referred to as Zohzan) consists of two horizontal layers of plastic scintillators separated by 1.50 m. Each layer comprises a 5 × 5 array of 1 m2 unit detectors (1 m × 1 m × 0.1 m plastic scintillator viewed by two photomultipliers each of 12.7 cm diameter) giving 25 m2 total detection area. A muon passing through a unit detector thus produces a single count. A coincidence between certain detectors from opposite layers give the directional information of the incident muon. Zohzan has been continuously monitoring muon rates in 17 directional channels. The hourly count rate in the vertical channel is 19 500 cph (counts per hour). The primary CR median energy producing muons recorded in the vertical channel is calculated to be 659 GeV using the response function of the atmospheric muons to the primary CR.6,7 For more detail of the instrument, readers can refer to a paper by Mori et al.8 To properly eliminate the spurious variations due to the atmospheric eﬀect and extract the solar diurnal anisotropy from the data near the ULE, where the anisotropy is expected to be very small, we adopt the following “East–West” method. Let I(t) be the percent deviation of the intensity in

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch10

128

K. Munakata et al.

a vertical directional channel at time t in hours, deﬁned as, I(t) =

¯ N (t) − N × 100 , ¯ N

(2)

¯ is the 24-hour cenwhere N (t) is the pressure corrected muon count and N tral moving average of N (t) at t. Also let R(t) be the percent deviation due to the solar diurnal anisotropy expected in the vertical directional channel. By taking the diﬀerence between I(t)s in the East- and West-viewing channels, I E (t) and I W (t), we can deduce the “diﬀerential” variation of R(t), as I E (t) − I W (t) dR(t) = , (3) dt ∆t where ∆t is the hour angle separation between the mean East- and Westincident directions averaged over the East- and West-incident events. Note that the “diﬀerential” variation D(t) in (3) is free from the spurious variation due to the atmospheric eﬀect which is expected to be common for I E (t) and I W (t). Let us consider the SDV for R(t), as D(t) =

R(t) = aR cos{ω(t − tR )} ,

(4)

where aR and tR are, respectively, the amplitude in percent and the maximum phase in hours and ω = π/12. We deduce from the data the amplitude aE−W and phase tE−W of the observed diﬀerence variation by ﬁtting a trigonometric function to I E (t) − I W (t), as I E (t) − I W (t) = aE−W cos{ω(t − tE−W )} .

(5)

By introducing (4) and (5) into (3), we get aR and tR in terms of aE−W and tE−W , respectively, as aR =

aE−W , ω∆t

tR = tE−W + 6 .

(6)

In the present paper, we use I E (t) and I W (t) respectively derived from (2) with N all E (t) and N all W (t), each composed by adding hourly muon counts in ﬁve directional channels available from Zohzan, as N all E (t) = N E (t) + N NE (t) + N SE (t) + N 2E (t) + N 3E (t) , N all W (t) = N W (t) + N NW (t) + N SW (t) + N 2W (t) + N 3W (t) .

(7)

For the directional channels in Zohzan, readers are referred to Mori et al.8 The average count rates of N all E (t) and N all W (t) are 25 700 cph and 35 600 cph, respectively, while the median energies of primary CR recorded in these two composite channels are calculated to be 660 GeV and 549 GeV

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch10

On the Upper Limiting Energy of the Solar Diurnal Anisotropy

129

on the basis of the response function of the atmospheric muons to the primary CR. The hour angle ∆t is also estimated to be 5.3 h. We ﬁrst derive the monthly harmonic vector (with the amplitude aE−W and phase tE−W in (5)) from the harmonic analysis of I E (t) − I W (t) averaged over every month. We then calculate the yearly mean harmonic vector by averaging 12 monthly vectors. Averaging over a year cancels out the sidereal diurnal variation arising from the stationary galactic anisotropy, which becomes signiﬁcant in higher energies.

3. Results The yearly mean harmonic vectors of I E (t) − I W (t) in (5) and R(t) in (6) observed for 20 years between 1985 and 2004 are shown in Figs. 1(a) and 1(b), respectively. The amplitude of each vector is represented by its length, while the phase is represented by the angle measured clockwise from vertical (the local times of +y, +x, −y, and −x directions are, respectively, 00:00, 06:00, 12:00, and 18:00 LST). To demonstrate the long-term variation, we plot in this ﬁgure the “vector summation dials”, in which yearly mean vectors are summed up one by one starting from the ﬁrst year of 1985. It is clearly seen that the harmonic vector shows a systematic change with 0

0.1% '85

'00

'04

0.1%

00:00 '95

'95

12:00

'90

IE- IW

0.1% '85 0

06:00 '90

'00

'04

R

06:00

0.1%

(a)

(b)

Fig. 1. The summation dial of the yearly mean harmonic vectors of the solar diurnal variation observed by Zohzan over 20 years between 1985 and 2004. (a) The harmonic vector of I E (t) − I W (t). (b) The harmonic vector of R(t). The representative years are indicated beside the corresponding yearly mean vectors displayed by full circles (see text). The statistical errors deduced from the counting rates are also displayed by horizontal and vertical error bars.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch10

130

K. Munakata et al.

an 11-year period. The harmonic vector of R(t) in Fig. 1(b) shows a phase around 06:00 LST in 1986–1988 and 1997–1999 in agreement with the CG anisotropy, with a later phase in 1989–1996 and 2000–2004, including the two maxima of the solar activity in 1989 and 2000, respectively. Such systematic variation has been reported earlier and is conﬁrmed by the present analysis with the data covering a much longer period, which allows more quantitative analyses.9,10 The long-term change in Fig. 1(b) implies that the observed R(t) is partly due to an additional anisotropy changing every 11-years. By using the CG anisotropy deﬁned in (1) together with the response function of atmospheric muons, we calculate that the diurnal variation R(t) due to the CG anisotropy has an amplitude of 0.035% and a phase of 05:56 LST. Figure 2 shows the vector summation dial in Fig. 1(b) after subtracting the CG anisotropy. The summation dial after the correction is almost a straight line, indicating that the additional diurnal variation has signiﬁcant amplitude with an almost constant phase at around 15:00 LST in 1989– 1996 and 2000–2004, while it has almost zero amplitude in 1986–1988 and 1997–1998, following (with a lag) the respective solar minima. To quantitatively examine the additional anisotropy change, we plot in Fig. 3 the amplitude and phase of each yearly mean vector from Fig. 2 as a function of year. We excluded from this plot the phases in years when the mean amplitude is smaller than its error. It is clearly seen that all the phases (open circles) are between ∼ 1200 and ∼ 1800 h LST, except 1997 when the diurnal variation has phase at 03:15 ± 02:51 LST and amplitude

18:00

0.1%

0

'90 '95

0.1%

'00

'04

12:00

R-CG Fig. 2. The summation dial corrected for the diurnal variation due to the CG anisotropy. Same as Fig. 1(b), but the contribution from the CG anisotropy is subtracted (see text).

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch10

On the Upper Limiting Energy of the Solar Diurnal Anisotropy

24

R-CG

0.06

18

0.04

12

0.02

6

0

1985

1990

1995 year

2000

phase (hours)

amplitude (%)

0.08

131

0 2005

Fig. 3. The amplitude and phase of the additional diurnal variation plotted as functions of time (year). The amplitudes of the vectors in Fig. 2 are displayed by full circles (left vertical axis), while the phases are displayed by open circles (right vertical axis). The trigonometric best-ﬁt function to the amplitudes is also shown by the solid curve. The amplitude vanishes following solar minima (see text).

of 0.013 ± 0.009%. The average of phases in this ﬁgure is 14:53 ± 03:24 LST. The amplitude of the additional diurnal variation in Fig. 3, on the other hand, shows a clear year-to-year change, reaching maxima and minima separated by 11 years. By ﬁtting a trigonometric function A{1+cos Ω(t−T )} with time, t, in years and Ω = 2π/11 to the observed amplitudes, we get A = 0.023 ± 0.002% and T = 1992.4 ± 0.2. The best-ﬁt function, displayed by the solid curve, indicates that the additional diurnal variation reaches a maximum in 1992–1993 and 2003–2004, whilst vanishes in 1986–1987 and 1997–1998. 4. Conclusions and Discussions We analyzed the SDV observed by a Japanese deep underground muon detector Zohzan over 20 years from 1985 to 2004. The East–West technique was adopted to eliminate possible spurious variation due to the atmospheric eﬀect. The obtained SDV shows a clear 11-year change. The variation in 1986–1988 and 1997–1998 can be fully attributed to the CG anisotropy due to Earth’s orbital motion around the Sun. This implies that the ULE of the solar modulation in these years lies somewhere below ∼ 600 GV, that is, below the major energy response of Zohzan. The variation, on the other hand, contains a signiﬁcant additional component in 1989–1996 and 2000–2004. All the signiﬁcant additional variations have the phase between

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch10

132

K. Munakata et al.

∼ 1200 and ∼ 1800 h LST, with the average at 14:53 ± 03:24 LST. The amplitude of the additional diurnal variation shows a clear year-to-year change with maxima of 0.046% in 1992 and 2003 and minima of 0% in 1987 and 1998, each separated by 11 years. From the long-term observations with neutron monitors, which have the maximum energy response to ∼ 10 GeV galactic CR, it has been reported that the CDD anisotropy has the average amplitude of ∼ 0.5% and the phase between ∼ 1500 and ∼ 1800 h LST.11 The observations using the surface muon detectors monitoring the cosmic ray intensity in higher energy region (∼ 50 GeV) have also reported similar anisotropy, although the amplitude is smaller (∼ 0.3%) and the phase slightly earlier (between ∼ 13 and ∼ 18 h LST).12,13 Most of these observations have consistently reported 11-year variations of the anisotropy, such as the smaller (larger) amplitude around the minimum (maximum) periods of the solar activity. It is also known that the phase of the anisotropy shows a 22-year variation, shifting toward early hours during the minimum period in the “positive” polarity state of the solar dipole magnetic ﬁeld, in which the polar magnetic ﬁeld of the Sun directs away from (toward) the Sun in the northern (southern) hemisphere. All these 11- and 22-year variations of the anisotropy have been interpreted by the CDD theory and are regarded as reﬂecting the long-term variations in the modulation parameters, such as the radial and latitudinal components of the local gradient of the cosmic ray density and the particle drifts and diﬀusion tensor in the regular and irregular magnetic ﬁelds.11–13 Since the phase (14:53 ± 03:24 LST) of the additional anisotropy deduced in Fig. 3 is consistent with the typical phase expected from the CDD anisotropy due to the solar modulation eﬀect, one might conclude that the ULE in 1989–1996 and 2000–2004 extends somewhere up to or over ∼ 600 GV. However, the standard CDD theory of cosmic ray transport is unable to properly describe solar modulation in the high-energy region where the gyroradius and mean free path of particles become large and the diﬀusion picture breaks down. Kota14 has presented an attempt to extend the theoretical description of solar modulation up to ∼ 100 GeV and succeeded in reproducing the gradual decreases of the anisotropy as the particle energy increases. The small amplitude of the observed additional anisotropy in Fig. 3, which is less than a tenth of that observed at ∼ 10 GeV with neutron monitors, might indicate such a soft energy spectrum below the ULE. It is also very interesting to look at the SDV observed above ∼ 600 GV during one of the periods when Zohzan observes the additional diurnal

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch10

On the Upper Limiting Energy of the Solar Diurnal Anisotropy

133

variation. An air-shower experiment at Yangbajing in Tibet, China has recently reported the solar diurnal variation of the multi-TeV CR intensity observed over four years between November 1999 and November 2003.5 The SDV recorded in the higher energy event samples (6.2 and 12 TeV) were fairly consistent with the CG anisotropy, while the SDV in the lowest energy sample (4.0 TeV) showed a signiﬁcant deviation, probably suggesting the extension of the ULE up to multi-TeV region. This seems to be consistent with the observation by Zohzan presented in this paper. The additional SDV observed by the Tibet experiment has an amplitude of 0.043 ± 0.005% and a phase at 08:20 ± 00:25. This amplitude is comparable with the average amplitude (0.033 ± 0.015%) of the additional SDVs observed by Zohzan in four years between 2000 and 2003. The phase observed by Tibet experiment, on the other hand, is ∼ 6.5 h earlier than the average phase by Zohzan (14:54 ± 01:12). This might suggest that the phase of the SDV around the ULE shifts to earlier hours from ∼ 15:00 LST below the ULE. Zohzan also observed such a shift in 1997, though the SDV was less signiﬁcant in that year. Further development of the high-energy modulation theory is awaited for a better understanding of the physical mechanisms responsible for the observed anisotropy around the ULE. Further coordinated observations by Zohzan and Tibet experiment are also required to clarify the behavior of the additional SDV. It is of particular interest to see the SDV from the Tibet experiment in 2008 and 2009, when Zohzan is expected to observe the CG anisotropy alone, as predicted by extrapolating the solid curve in Fig. 3. Acknowledgments This work was supported in part by Grants-in-Aid for Scientiﬁc Research on Priority Areas (MEXT) and by Scientiﬁc Research (JSPS) in Japan. The observation by Zohzan has been supported by Shinshu University. References 1. H. Moraal, Nucl. Phys. B (Proc. Suppl.) 33A,B (1993) 161. 2. K. Munakata et al., Proc. 25th International Cosmic Ray Conference 2 (1997) 77. 3. A. H. Compton and I. A. Getting, Phys. Rev. 47 (1935) 817. 4. L. J. Gleeson and W. I. Axford, Astrophys. and Space Sci. 2 (1968) 431. 5. M. Amenomori et al., Phys. Rev. Lett. 93 (2004) 061101–1. 6. K. Murakami et al., IL Nuovo Cim. 2C (1979) 635. 7. K. Fujimoto et al., Rep. of Cosmic-Ray Res. Lab., Nagoya University 9 (1984) 7.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch10

134

8. 9. 10. 11. 12. 13. 14.

K. Munakata et al.

S. Mori et al., J. Fac. Sci., Shinshu University 27 (1992) 47. S. Yasue et al., Lett. J. Geomag. Geoelectr. 43 (1991) 771. S. Mori et al., Proc. 23rd International Cosmic Ray Conference 3 (1993) 633. J. Chen and J. W. Bieber, Astrophys. J. 405 (1993) 375. D. L. Hall, M. L. Duldig and J. E. Humble, Astrophys. J. 482 (1997) 1038. K. Munakata et al., Adv. Space Res. 29 (2002) 1527. J. Kota, J. Geophys. Res. 104 (1999) 2499.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch11

SECTOR BOUNDARY CROSSINGS AND GEOMAGNETIC ACTIVITIES SHINICHI WATARI National Institute of Information and Communications Technology 4-2-1 Nukuikita, Koganei, Tokyo 184-8795, Japan [email protected] TAKASHI WATANABE Ibaraki University, 2-1-1 Bunkyo, Mito Ibaraki 310-8512, Japan [email protected]

There are periods of azimuthal interplanetary magnetic ﬁelds (IMFs) outwards from or inwards towards the Sun. The patterns of IMFs are called sector structures. Two or four sectors are often observed during a solar rotation period. Boundaries of the sectors of opposite IMF polarities are called sector boundaries. We statistically studied a relation between sector boundary crossings and geomagnetic activities. Quiet periods of geomagnetic activity often occur before the sector boundary crossings which can be predicted from observations of solar magnetic ﬁeld.

1. Introduction The earliest satellite observations of interplanetary magnetic ﬁelds (IMFs) reveal the sector structures of inward and outward azimuthal magnetic ﬁeld.1−7 The interface of sectors, where the magnetic ﬁelds alternate between inwards and outwards, is called a sector boundary. Previous studies suggested that there is a relationship between sector boundary crossings and geomagnetic activity.1−11 Forecasting geomagnetic activity is a subject of space weather. It has been shown in Ref. 12 that sector boundary crossings may be forecastable from solar magnetic ﬁeld observations. We re-visited this problem. We statistically studied a relation between sector boundary crossings and geomagnetic activity using continuous, in situ solar wind observations at 1 AU.

135

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch11

136

S. Watari and T. Watanabe

2. Analysis We used daily average values of solar wind data from the OMNI2 database between 1995 and 2004 because data coverage was high during this period. We determined the sector boundaries using azimuthal angle of IMFs observed at 1 AU. We selected boundaries where the polarity was the same at least four days before the reversal and remained reversed for at least four days following the reversal. As a result large scale sector structures are selected in this study. During this period, 341 sector boundaries are identiﬁed. We used daily average Kp-index from the OMNI2 database for the indicator of geomagnetic activity in this study. Figure 1 shows plots of yearly sunspot number and the annual occurrence number of sector boundary crossings. According to Fig. 1(b), number of the sector boundary crossings increased during the declining phase (2003 and 2004) of solar cycle. This suggests that stable large sector structures were formed during the declining phase. Figure 2 shows a Bartels diagram of daily average azimuthal angles of IMF from the OMNI2 database. White denotes angles in the range 90–180◦, away polarity. Black denotes angles in the range 270–360◦, toward polarity. Gray denotes other azimuthal angles. According to this ﬁgure, stable two-sector structure appeared between 2002

sunspot number

160 120 80 40 0

1995

number

20

1997

1999 year (a)

min.

15

2001

2003

max.

10 5 0

1995

1997

1999

2001

2003

year (b) Fig. 1. Plots of (a) yearly sunspot number and (b) annual occurrence number of sector boundary crossings.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch11

Sector Boundary Crossings and Geomagnetic Activities

Fig. 2.

137

Bartels diagram of IMF polarity (white, away, black, toward).

and 2004. During the maximum (2000 and 2001) of solar cycle, the number of the crossings was lower. The increase of interplanetary coronal mass ejections (ICMEs) during the solar maximum largely modiﬁed the sector structures. As a result, the number of observed stable large sector structures decreased. Figure 1(b) shows increase of sector boundary crossings in 1999. Figure 2 suggests that this increase was caused by a stable two sector structure appeared in the ﬁrst half of 1999.

3. Results and Discussion Figure 3 shows superposed epoch plots of average values of daily average Kp-index and solar wind parameters (magnetic ﬁeld, speed, density, and temperature) for the Earth’s crossing of sector boundaries between 1995 and 2004. The error bars indicate values of one standard deviation. According to Fig. 4, the daily average Kp-index increase after the sector boundary crossings. However, the increases of geomagnetic activities after the crossings are not statistically signiﬁcant. The change in the solar wind parameters (increases in magnetic ﬁeld, increase of speed and temperature, and decrease in density) associated with the boundary crossings suggest interaction regions which are inﬂuenced by highspeed streams from coronal holes. The quiet conditions of geomagnetic

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch11

138

S. Watari and T. Watanabe

Fig. 3. Superposed epoch plots of daily average values of Kp-index and solar wind parameters (magnetic ﬁeld, speed, density, and temperature) for the Earth’s crossing of sector boundaries.

activity often occurred before the sector boundary crossings. This suggests that high-speed solar wind from coronal holes is dammed up by sector boundaries. Figure 4 shows superposed epoch plots of daily average values of Kp-index and solar wind parameters (magnetic ﬁeld, speed, density, and temperature) for the Earth’s crossing of sector boundaries during the minimum phase (1995–1996), the rising phase (1997–1998), the maximum phase (1999–2002), and the declining phase (2003–2004) of solar activity. According to these ﬁgures, enhancements of geomagnetic activity are smaller during the rising phase of solar cycle than during other phases of solar cycle. A signiﬁcantly geomagnetically quiet day occurred on December 4, 2004. The sum of eight Kp-indices was recorded at 0. Figure 5 shows Dst-index and solar wind parameters around this signiﬁcant quiet day which is indicated by the two vertical dotted lines. This ﬁgure suggests that the quiet day occurred before the boundary crossing, where magnitude of the IMF

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch11

Sector Boundary Crossings and Geomagnetic Activities

139

Fig. 4. Superposed epoch plots of daily average values of Kp-index and solar wind parameters (magnetic ﬁeld, speed, density, and temperature) for the Earth’s crossing of sector boundaries during the minimum phase (1995–1996), the raising phase (1997–1998), the maximum phase (1999–2002), and declining phase (2003–2004) of solar cycle.

was weak and the solar wind speed was slow. After the quiet day, the strong IMF was caused by the ICME near the boundary. It is possible to predict timing of sector boundary crossings based on observations of solar magnetic ﬁeld.12 Our result suggests that we can infer geomagnetically quiet periods based on sector boundary crossings. It is a

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch11

140

S. Watari and T. Watanabe

Fig. 5. Dst-index and solar wind parameters around signiﬁcantly geomagnetically quiet day December 4, 2004.

further subject to study a relationship between a signiﬁcantly geomagnetically quiet period and a sector boundary crossing. Acknowledgment We thank NASA/Goddard Space Flight Center for the OMNI2 database. References 1. N. F. Ness and L. F. Burlaga, J. Geophys. Res. 106 (2001) 15803. 2. E. J. Smith, J. Geophys. Res. 106 (2001) 15819.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch11

Sector Boundary Crossings and Geomagnetic Activities

141

J. M. Wilcox and N. F. Ness, J. Geophys. Res. 70 (1965) 5793. J. M. Wilcox, K. H. Schatten and N. F. Ness, J. Geophys. Res. 71 (1967) 19. J. M. Wilcox, Space Sci. Rev. 8 (1968) 258. J. M. Wilcox and D. S. Colburn, J. Geophys. Res. 74 (1969) 2388. J. M. Wilcox and D. S. Colburn, J. Geophys. Res. 77 (1972) 751. B. Mendoza and R. Perez-Enriquez, J. Geophys. Res. 98 (1993) 9356. B. Mendoza and R. P. Enriquez, J. Geophys. Res. 100 (1995) 7877. N. U. Crooker, J. T. Gosling and S. W. Kahler, J. Geophys. Res. 103 (1998) 301. 11. E. Echer and W. D. Gonzalez, Geophys. Res. Lett. 31 (2004) L09808. 12. G. K. H. Schatten, J. M. Wilcox and N. F. Ness, Sol. Phys. 6 (1969) 442.

3. 4. 5. 6. 7. 8. 9. 10.

This page intentionally left blank

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch12

ENERGETIC PARTICLE COMPOSITION SIGNATURES IN THE EARTH MAGNETOTAIL S. Y. FU∗,† and Q.-G. ZONG‡ Peking University, Beijing 100871, China †Key Laboratory for Space Weather, CAS, Beijing 100080, China ‡Center for Space Physics, Boston University, Boston, MA 02215, USA ∗ISPAT,

Energetic particle species with emphasis on singly charged oxygen ions in the tail and their related geomagnetic activities are investigated by using GEOTAIL, CRRES, as well as Cluster data. The O+ ions are a tracer of the ionospheric population in the magnetotail during dynamic process. More ionospheric origin ions participate in magnetospheric storms/substorms than solar wind origin ions during strong geomagnetic activities. The solar wind source is saturated in providing ions into magnetosphere during very intense geomagnetic activities (|Dst| > 180 nT, Kp > 7). When the near-Earth reconnection occurs and expands from the central plasma sheet to the plasma sheet boundary layer or even to the lobe region, more ionosphere origin O + ions could participate and be energized in the magnetospheric dynamic process. This leads to oxygen-rich injections in the near earth tail region and plasmoids in the distant tail. The ionospheric outﬂow oxygen ions (initial accelerated at the topside auroral ionosphere) are subsequently energized in the plasmoid. The low O+ abundance observed in some plasmoids is consistent with the idea that only pre-existing ions in the near-Earth plasma sheet are accelerated thus a rather low abundance of O+ ions can be observed. Multiple ﬂux ropes/plasmoids observed in the tail plasma sheet can be interpreted as strong evidence for multiple X-lines.

1. Introduction During the ﬁrst decade of investigation of the space environment in 1970s, energetic ions in the magnetosphere are predominantly regarded as protons coming from the solar wind.1,2 The limited capability of plasma instruments at that time did not allow discrimination of particle species except for very high energies. A possible ionospheric source for the inner magnetosphere was suggested by Axford.3 Indeed, Shelley et al.4 found that there is a signiﬁcant component of low energy O+ ions in the magnetosphere for which the ionosphere is the most likely source. CHEM (Charge–Energy–Mass) instrument onboard AMPTE/CCE for the ﬁrst time allowed the characterizing ring current ions in the energy range of 1–300 keV/e in detail. Now it is widely accepted that the particle populations in the magnetosphere 143

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch12

144

S. Y. Fu and Q.-G. Zong

have two main sources: the Earth’s ionosphere and the solar wind. The ionosphere provides plasma through the polar wind, the cleft fountain, the auroral regions and the plasmasphere. The solar wind supplies particles through the high latitude cusp region, the mantel and the low latitude boundary layer (LLBL) region. The plasma sheet and the magnetotail are two regions that contain a mixture of particles from these two originals. The relative contributions vary with the solar wind conditions and the magnetospheric state, since the direct source of particle injected to the ring current is the plasma sheet that is populated by the solar wind and ionosphere. The variation of the ion composition in the inner magnetosphere is an interesting topic, since the history of the more energetic particles reﬂects a great deal of the dynamical processes which eﬀect the coupling of the solar wind to the Earth magnetosphere. The relative contributions of solar wind particles and ionospheric particles to the ring current have been studied using the ion composition information acquired by the CHEM during solar minimum. An overview of the AMPTE composition measurements has been given by Gloeckler et al.5 A storm case study by Hamilton et al.6 and statistical studies by Daglis et al.7 established the fact that ionospheric ions play an important role during ring current build-up and decay phases. For solar maximum cases, Fu et al.8,9 have examined the spatial and temporal variations of oxygen ions in the ring current for storm and substorm cases, as well as statistic studies. It has now been well established that oxygen ions of terrestrial origin are found almost everywhere in the magnetosphere. In the magnetotail region, energetic particles together with the plasmoids have been studied.10–12 Using CRRES data, Fu et al.8,9 have shown that the oxygen abundance in the magnetosphere is greatly increased in periods with enhanced activity. Recently, with the newly launched Cluster mission, energetic particles, especially oxygen ions in the near earth plasma sheet have been studied.13,14 The two major aspects of energetic composition study are their origin and related acceleration processes. Reconnection processes have been thought to be one of plausible candidates for the energization of energetic particles. Reconnection of nearly anti-parallel tail magnetic ﬁeld lines leads to the formation of plasma bubbles that eventually disconnect from the Earth magnetic ﬁeld and move tailward. Therefore, the relation of energetic particles and plasmoids are one of important topics in the magnetotail study. In this paper, we will review some of the observations made by GEOTAIL, CLUSTER, and CRRES and try to present a global picture of the energetic particles in the tail region.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch12

Energetic Particle Composition Signatures in the Magnetotail

145

2. Observations 2.1. Geotail plasmoid observations From 1992 to 1994, GEOTAIL was operated in the distant tail orbit. The energy ranges of the particle detector HEP-LD on board Geotail for hydrogen, helium, and oxygen are approximately 40–3100, 70–4000, and 140– 4000 keV, respectively.15 Figure 1 gives an overview of the HEP-LD and MGF measurements16 in a typical plasmoid for the time interval 02:00– 04:00 UT on January 15, 1994, when Geotail was located approximately at (GSE: −96, 7, −5 Re).10 The bipolar waveform in the By component is

Fig. 1. Geotail HEP-LD energetic particle and MGF magnetic ﬁeld observations in a plasmoid event from 02:00 to 04:00 UT on January 15, 1994. From the top: integral counting rates for oxygen and all ions; hydrogen and helium ions; color coded azimuthal intensity distributions of the counting rate for “all ions” (look direction) and GSE components of the magnetic ﬁeld and its magnitude (in nT). The circles in the third panel mark the magnetic ﬁeld direction projected on the equatorial plane.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch12

146

S. Y. Fu and Q.-G. Zong

evidence for a three-dimensional (3D) nature of this plasmoid. Inspection of the time proﬁles for particles shows clearly that hydrogen and helium ions appear at the very beginning of the event when the magnetic ﬁeld measurement indicates the transition from the lobe ﬁeld into the plasmoid structure. With a delay of approximately 30 min relative to hydrogen, strong oxygen abundance developed (top panel). In fact, the azimuthal distributions in the third panel (look directions) indicate a rather narrow angular band around the magnetic ﬁeld. This means that the oxygen ions form a narrow beam, strongly conﬁned to the interior of the plasmoid, moving away from Earth. During the oxygen burst (02:47–03:10 UT), however, the oxygen distribution increased substantially at smaller velocities as documented in Fig. 2. The high-velocity tail of the distribution function, beyond 2500 km/s, remained nearly unchanged. On the other hand, the slopes for hydrogen and helium became somewhat harder by the same factor. The oxygen peak occurred at a velocity of about 1700 km/s that corresponds to 250 keV in energy space. A database of 167 oxygen events identiﬁed between October 2, 1992 and June 4, 1994 was created to study the statistical properties of these events as a function of position in GSE coordinates and geomagnetic activity. Event selection was based on the abrupt increase of counting rates (at least

Fig. 2. Distribution functions f (v) for supra-thermal hydrogen (top), helium (middle), and oxygen (bottom) ions versus energy before (left) and during (right) the plasmoid event on January 15, 1994.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch12

Energetic Particle Composition Signatures in the Magnetotail

147

Fig. 3. Histograms of normalized occurrence probabilities for the dawn–dusk (Y ) (top panel) and south–north (Z) (middle panel) distributions of oxygen bursts. The third panel shows the distribution of energetic oxygen events versus geomagnetic activity (Kp-index).

20 counts/min) of energetic oxygen ions and with duration time longer than 20 min. The main results are: (a) Distribution along the Y - and Z-axis. The probability for oxygen events in a given bin is shown in Fig. 3. The results indicate a strong dawn–dusk asymmetry in the occurrence frequency of oxygen bursts, with the highest probability occurring on the dusk side, whereas the distribution along the GSE-Z axis shows roughly asymmetry. (b) Distribution versus the planetary Kp-index. The 167 events were sorted according to the mean Kp value prevailing throughout the event interval. Each bin was properly normalized by the relative frequency of occurrence of a given value for Kp in the time interval of interest. It is obvious that the probability for observing an oxygen event increases strongly (nearly in a linear fashion) with Kp. Figure 4 shows the dependence of occurrence probability on the Y coordinate for four diﬀerent intervals in the X-coordinate. The frequency of occurrence of plasmoid/oxygen events in the tail is not at all evenly distributed as one may assume. About 80% of all events were found in

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch12

148

S. Y. Fu and Q.-G. Zong

Fig. 4. Histograms of normalized probabilities of occurrence for oxygen events as a function of GSE-Y (dawn–dusk variation) obtained for selected values of GSE-X.

the dusk side, and the peak probability shifts toward dusk with increasing distance from the Earth. The statistical survey of these oxygen events in the tail established that substorms are a necessary but apparently not a suﬃcient condition for these beams to occur. The absence of oxygen beams can imply diﬀerences in the morphology of substorms, i.e., a fraction of substorms is associated with oxygen beams and others are not. Those studies are consistent with previous case studies.10–12,17 Indeed, Lui et al.17 is the ﬁrst one who found the oxygen abundance in the plasmoid depends on the geomagnetic Kp-index based on 12 case studies. When Kp-index is small (Kp < 3), a plasmoid/ﬂux

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch12

Energetic Particle Composition Signatures in the Magnetotail

149

rope may have low oxygen ion content, however, if the geomagnetic activity becomes stronger, one ﬁnds signiﬁcant amounts of oxygen ion embedded in plasmoids. In the oxygen bursts events, the partial number density ratio is much higher than that found in the normal plasma sheet. In addition, it is established that the oxygen bursts are closely related to substorms and can be understood as a direct product of the substorm process. By invoking the neutral line model one can explain the observations. First, earthward ﬂow with positive magnetic ﬁeld (Bz) is caused by a “constant” quiet time neutral line in the distant tail.18 Second, magnetic reconnection at the nearEarth neutral line eventually results in the formation and rapid tailward motion of a plasmoid. Magnetic tension causes the plasmoid to move downtail at high speed (−400 to −600 km/s).19 Oxygen ions, extracted by the substorm process from the polar ionosphere, drift slowly (about 150 km/s) tailward in the lobe ﬁelds20 toward the active reconnection region which is located earthward of Geotail. This leads to the observed delay of 20–30 min relative to substorm onset assuming the reconnection site is between 30 and 40 Re. In the post-plasmoid plasma sheet, plasma and oxygen ions are accelerated to velocity greater than 1000 km/s, which is far greater than the speed of plasmoids.21 This process leads eventually to the observed bursts of energetic ions in the post plasmoid plasma sheet (PPPS) characterized by a consistently negative Bz component.

2.2. CRRES observation The CREES data presented here are obtained from the spectrometer MICS on board CRRES spacecraft. The energy range of MICS is from 1.2 to 426 keV/e.22 Statistic studies on several selected large storms and a total of 398 substorm injection events from February 4 to October 10, 1991 have been carried out.8,9 Figure 5 gives an example of the changing of energy spectra of diﬀerent ion species before activity and during storms. It is clear that the ﬂuxes of all ions (H+ , He+ , He++ , and O+ ) often increase simultaneously. However, the abundance of these ions varies from case to case. Fluxes of oxygen ions increased extraordinarily strong, whereas that of protons only shows some enhancement below 200 keV/e. The averaged He++/H+ , O+/H+ energy density ratio obtained by CRRES for diﬀerent Dst and Kp-indices are shown in Fig. 6. In all injections, ﬂux enhancements of O+ ions are usually coincident with the ﬂux intensiﬁcations of high energy (E/q > 100 keV/e) H+ and He++ ions.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch12

150

S. Y. Fu and Q.-G. Zong

Fig. 5. Diﬀerential intensity spectra at L = 4 versus E/Q for O+ (point), He+ (cross), He++ (diamond) and H+ (star). (a) At quiet time ring current (orbit = 836), and (b) during a storm maximum (orbit = 847).

These results (bottom panels in Fig. 6) demonstrate that the abundance of O+ ions in an injection strongly depends on the geomagnetic activity. The abundance of O+ ions in terms of energy density in injections increases roughly linearly with Kp and Dst-indices. As we can see from Fig. 6, the ratio increases substantially with enhanced geomagnetic activity from about 20% to more than 65%. The relation between the oxygen ion energy density ratio and the Dst and Kp-indices are found to be ED-ratio (oxygen) = 22.9 + 0.15 × Dst ,

R = 0.95 ,

ED-ratio (oxygen) = 0.68 + 6.6 × Kp ,

R = 0.97 ,

where R is the correlation coeﬃcient. However, the averaged He++ /H+ energy density ratio shown in the top panel varies much less compared with oxygen ions, which is only from about 7–11%. The implication of the increase in He++ as high as (11–7)/7 = 60% is that more solar wind origin ions have been put into the inner magnetosphere during geomagnetic activity period. However, for the most intense geomagnetic activities (Dst < −150 nT, or Kp-index > 7), the He++ energy density indeed decreases. This implies that much more ionosphere origin ions participate in magnetospheric storms/substorms than solar wind origin ions during very intense geomagnetic activities. The solar wind source is saturated in providing ions for very intense magnetospheric storms/substorms.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch12

Energetic Particle Composition Signatures in the Magnetotail

151

Fig. 6. The averaged fractional energy density of O+ and He++ ions as functions of Dst and Kp-indices for all 398 substorm injection events. The bottom panel is the averaged oxygen ion energy density ratio and the top panel is that of He++ ion energy density.

The relation between the He++ ion energy density ratio and the Dst and Kp-indices are found to be ED-ratio (Helium) = 7.1 + 0.029 × Dst (Dst > −150) ,

R = 0.93 ,

ED-ratio (Helium) = 20.6 − 0.04 × Dst (Dst < −150) ,

R = −0.95 ,

ED-ratio (Helium) = 3.8 + 1.0 × Kp (Kp < 7) ,

R = 0.97 ,

ED-ratio (Helium) = 13.8 − 0.54 × Kp (Kp > 7) ,

R = −0.92 ,

where R is correlation coeﬃcient. 2.3. Plasmoid observed by cluster Plasmoids are often observed by Cluster in the tail plasmasheet. Figure 7 gives an example on October 28, 2002. The plasmoid event lasting from 19:42 to 19:52 UT can be identiﬁed by the bipolar signature in Bz. Prior to the arrival of the plasmoid at 19:47 UT, a compression of the lobe magnetic ﬁeld and lobe plasma density in front of the Earthward moving plasmoid

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch12

152

S. Y. Fu and Q.-G. Zong

cause the total magnetic ﬁeld intensity increased from 20 to 23 nT. In the plasmoid, the Bz component shows (−/+) bipolar signatures at 19:47 UT with a Bz variation of 8.5 nT. There is no indication for the presence of a core ﬁeld at the time of the inﬂection point (where Bz changes its sign). During this event, the peak ﬂuxes of both energetic ions and electrons occurred at the reﬂection point where the total magnetic ﬁeld was around a local minimum. In particular, the magnetic ﬁeld of this event is consistent with a “closed loop” plasmoid. The reﬂection point of the bipolar signature does not coincide with the local maximum as shown in Fig. 7. Both ion and electron spikes started about 2 min after the magnetic ﬁeld signature was observed. It should be mentioned that the (−/+) bipolar signature

Fig. 7. An overview of RAPID, CIS, and FGM data from 19:30 to 20:00 UT, October 28, 2002 for S/C1 (left) and S/C4 (right). From the top the panels show: integral electron and proton ﬂux; plasma density; V x and magnetic ﬁeld components (in GSE), and magnitude. The vertical dashed lines at 19:47 UT mark the reﬂection point of the ﬁrst plasmoid (P1). The shade area is the core of the earthward ﬂowing plasmoid. The dashed lines at 19:54 UT mark the center of the second plasmoid (P2).

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch12

Energetic Particle Composition Signatures in the Magnetotail

153

passed over the spacecraft essentially during the interval for which the By component was negative. Between 19:52 and 19:58 UT the Bz component observed by Cluster S/C1 was essentially negative. This is indicative of the fact that the other plasmoid was traveling toward to the Earth. In fact, this can be proved by measurement of S/C4 that another smaller plasmoid was observed at around 19:54 UT. Another important point to be noticed is that TCR signatures were observed by Cluster S/C1 which is furthest from the plasma sheet center between 19:55 and 20:00 UT and that small plasmoids were observed by S/C4 which is close to the plasma sheet center. This is additional evidence for (1) a continuing existence of a neutral line tailward of Cluster spacecraft; (2) the connection of the post-plasmoid conﬁguration directly to this neutral line; (3) curved lobe ﬁeld lines caused by the huge size of the plasmoid in the X–Z plane. RAPID ion composition observations from spacecraft 1 and 4 are shown in Fig. 8. The panels display the ratios of the ﬂux of helium ions (70–1500 keV) to that of protons (30–1500 keV) and the ﬂux of CNO ions (140–1500 keV) to that of protons, respectively. It is interesting to note that just within the plasmoid, these two ratios showed a signiﬁcant decrease, with a value around 0.2 and 0.03, respectively. The ratios on either side of

Fig. 8. The two panels display the ratio of the ﬂux of helium ions to that of protons (J(He)/J(p)) and the ﬂux of oxygen ions to that of proton (J(O)/J(p)) by the four Cluster spacecraft during the time period of 19:30–20:00 UT.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch12

154

S. Y. Fu and Q.-G. Zong

the plasmoid encountered were found to be between 0.6 and 0.1. This indicates that the ion composition in the plasmoid is signiﬁcantly diﬀerent from that in the ambient plasma. The much less heavy ions have been “trapped” in the plasmoid that formed in the distant tail region. The plasmoid should preserve the ion composition information where it was formed. As suggested by Fu et al.,9 diﬀerent ion population escape from the ionosphere, a mass ﬁlter mechanism (the ionospheric origin ions will be dispersed with their different mass) may exist in the magnetotail during strong magnetic activities. The observed earthward moving plasmoids are probably related to the substorm injections observed in the geosynchronous orbit. Proﬁles of differential ﬂux of energetic electrons versus time from three Los Alamos geosynchronous satellites are shown in Fig. 9. The particle data indicate the development of an isolated substorm with a moderate level of activity starting around 20:08 UT. The event was embedded in a relative quiet period as we can see from Fig. 9. Two geostationary spacecraft located at 00:48 LT (LANL-02A) and 03:02 LT (LANL-97A) detected at least two

Fig. 9. Diﬀerential electron data (channels centered at 60, 90, 125, 190, and 270 keV) from six Los Alamos National Laboratory (LANL) spacecraft. The local midnight are indicated with dark green vertical lines. The arrows mark two injections at around 20:08 and 20:24 UT.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch12

Energetic Particle Composition Signatures in the Magnetotail

155

injections at around 20:08 and 20:24 UT, which are about 20 min after the earthward moving plasmoid detected by Cluster spacecraft. The spacecraft LANL-01A located at 20:07 LT did not catch the injections. A spike-like dispersionless energetic electron ﬂux increase occurred about 20 min after the plasmoid detected by Cluster in all the energy channels. About 16 min later at 20:24 UT, a sharp impulsive ﬂux variation occurred simultaneously in all electron energy channels. There is a same evidence for multiple dispersionless injections between 20:00 and 21:00 UT followed by a period of ﬂux oscillations. 3. Discussion and Conclusion 3.1. Ion composition in the tail and in the plasmoid structure Both in the deep tail and in the near-Earth ring current region, oxygen ions have been observed frequently during diﬀerent geomagnetic activities. A striking character in the inner region is that the O+ rich injection events occurred in large Kp, and more than 80% of them were observed in storm times. The oxygen-void events were detected predominately for small Kp values. It is interesting that the He++ energy density ratio increases for moderate activity, but decreases for intense activity. In fact, this can be interpreted as evidence that the oxygen ions (O+ ) and helium ions (He++ ) have diﬀerent origins. The He++ ions come from the solar wind and more solar wind origin ions could participate in the inner magnetosphere during geomagnetic activity period. However, much more ionospheric oxygen ions are escaped from the Earth into the injection region during intense geomagnetic activities. It is worth to note here that O+ ion injections are usually associated with enhancements of high energy (above 100 keV) H+ and He++ . Injections without oxygen ions often include only lower energy H+ and He++ ions (< 100 keV/e). This can be explained as these particles (higher energy H+ and He++ ions) probably have been accelerated together with the ions of ionospheric origin in the magnetotail. This scenario is consistent with Geotail observations in the distant magnetotail.23 In the magnetotail region oxygen ions with substantial intensities occurred only in conjunction with substorms and in plasmoid-like ﬁeld topologies. The appearance of these O+ ions is thought to be a result of substorm related extraction of oxygen from the polar ionosphere and subsequent acceleration in reconnection regions in the magnetotail. At the beginning of the growth phase, the tail magnetic

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch12

156

S. Y. Fu and Q.-G. Zong

ﬁeld is compressed and forms a more tail-like topology. Ions, mainly H+ and He++ from the solar wind are convected to the plasma sheet forming the strong cross tail current. In the late growth phase, slow merging of the magnetic ﬁeld begins in the central plasma sheet on closed ﬁeld lines and ions (mainly H+ and He++ ) are accelerated to a relative lower energy (< 100 keV/e), almost no oxygen ions directly from the ionosphere is involved. At the same time, with the beginning of geomagnetic activity, O+ ions in the ionosphere are being dragged out by parallel electric ﬁeld or conics and transported tailward along open magnetic ﬁeld lines into the lobe and/or plasma sheet boundary layer. Although some acceleration mechanisms in the polar ionosphere have energize these O+ ions during their outﬂow process, the energy of escaping oxygen ions can only reach about hundreds of eV or several keV, which is still far below the energy observed in the tail region. Therefore, additional mechanism is needed. If the activity continues to develop and the magnetic ﬁeld lines in the plasma sheet boundary or even in the lobe (where the Alfven speed is much higher) begin to merge, then all ions, H+ , He++ , together with the newly coming O+ ions are energized to relative high energies by the reconﬁguration of magnetic ﬁeld lines. In this case, high abundance of O+ ions, together with H+ and He++ ions can be observed. This is consistent with both CRRES and Geotail observations during geomagnetic activities. CRRES often observed events with two sequential injections in which the time interval between the two is relative short. The ﬁrst injection is not well developed and thus very weak, known as pseudo-breakup. When the activity continuously develops, a following strong injection takes place and real substorm onset occurs. In fact, as an additionally evidence, such energetic oxygen ions will also be injected into the tailward moving plasmoid. A plasmoid/ﬂux rope may have low oxygen ion content with Kp < 3, however, if the geomagnetic activity becomes stronger, one ﬁnds signiﬁcant amounts of oxygen ion embedded in plasmoids. As pointed out by Baumjohann,24 the lobe magnetic pressure is decreased only during the expansion phase of storm-time substorms. This implies that the lobe magnetic ﬁeld lines have indeed participated in the reconnection process during the storm-time substorms. In contrast, only closed magnetic ﬁeld lines are involved in the process of weak substorms. The occurrence frequency of oxygen burst events in the tail exhibits a strong dawn–dusk asymmetry, whereas the distribution about the GSE-Z axis is roughly symmetric. The highest probability for such events is observed on the dusk side. The asymmetry in the Y -direction seems to

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch12

Energetic Particle Composition Signatures in the Magnetotail

157

vary with the distance along the –X-axis. Most of these oxygen bursts are tailward ﬂowing in the post plasmoid plasma sheet. Statistical analysis also shows signiﬁcant dawn–dusk asymmetries of lower energy beams (5 keV) in the near-Earth plasma sheet with a strong preference in the dusk side of the plasma sheet.25 A strong dawn–dusk asymmetry in the occurrence frequency of auroral ion acceleration events has been proposed by Ghielmetti et al.26 Likewise the inverted V electron precipitation events and auroral zone VLF hiss also show a strong dawn–dusk asymmetry favoring the dusk side. Thus, the observed asymmetry in the distant tail could represent a simple expansion of an asymmetry already existing in the auroral oxygen extraction processes. On the other hand, the plasmoids are found more frequent in the duskside tail in the Geotail distant tail data in the almost same interval.27,28 This may also explain that the oxygen distribution asymmetry in the Y -direction.

3.2. Multiple plasmoids and their implications The phenomenon of substorms as a means of energy release in the magnetosphere may occur in a manner that is in some way analogous to a dripping faucet. Such a substorm model has been proposed by Hones.29 It is also proposed to modify this single-plasmoid model in order to explain the multiple-plasmoid formation in the course of substorm event.23,30 The occurrence of multiple reconnection sites after substorm expansion onset is also consistent with the current disruption model.31 In the October 28, 2002 event, the sequence of plasmoid at about 19:54 UT and TCRs is an indication for the repeated development of NENLs. Inspection of Figs. 7 and 9 shows that no convincing one-to-one correlation can be established for the multiple-plasmoids observed by Cluster and the injections observed by the geostationary spacecraft. However, it is nevertheless conceivable that the chain of plasmoids and TCRs is related to the possible multiple injections between 2000 and 2100 UT in Fig. 9. The lack of clear correlation may be due to unknown parameters in the plasmoid motion, e.g., deceleration of the earthward moving plasmoid and the exact corresponding time. Previous simulation studies, assuming both constant and time-varying driving forces delivered by the solar wind, indicated the formation and subsequent convection of magnetic islands (plasmoids) in the magnetotail. The plasmoids occur intermittently and repeatedly every 2–4 h,32 Lyon et al.33 also showed that a near-Earth X-line could occur

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch12

158

S. Y. Fu and Q.-G. Zong

repeatedly, the time between the formations of two successive near-Earth X-lines was found to be 40 min. The signiﬁcance of multiple-plasmoids/ﬂux ropes in the plasma sheet is that their formation can be understand in terms of simultaneous reconnection at multiple X-lines (MRX) in the magnetotail, as illustrated in the Fig. 10.30 The observations made in this paper also raise some interesting questions regarding the “fate” of earthward moving plasmoid in the plasma sheet. The earthward moving plasmoids pushed up against the geomagnetic ﬁeld will probably dissipate quickly since the orientation of their magnetic ﬁelds is favorable for reconnection with the closed ﬁeld lines in the near Earth. In particular, multiple earthward moving plasmoid/ﬂux ropes could be possible triggers for the substrom injections observed in the geostationary orbit. The relationship between these earthward moving plasmoids in near-tail and the larger plasmoid (if any) observed much farther down the tail is still not clear. As illustrated in Fig. 10, in addition to the small plasmoids formed closer to the Earth, the closed ﬁeld line region between the furthest of the MRX reconnection sites and the distant neutral line (DNL) should give rise to a much larger “plasmoid” containing loosely wound helical ﬁeld lines. MRX reconnection in the near-tail would appear to lead naturally to tailward ejection of multiple plasmoid-type ﬂux ropes for each substorm.

Fig. 10. Schematic depiction of the formation of earthward and tailward moving ﬂux ropes as a result of multiple, simultaneous reconnection neutral lines in the X–Z plane (adopted from Ref. 30).

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch12

Energetic Particle Composition Signatures in the Magnetotail

159

Acknowledgments The present study has been supported by CNSF under projects 40528005, 40425004, 40374060, and 40390150. This research was supported at Boston University by NASA grant NAG 5-10108. We are grateful to the Cluster CIS, FGM, and RAPID team.

References 1. T. W. Hill, Origin of the plasma sheet, Rev. Geophys. Space Phys. 12 (1974) 379–388. 2. T. A. Fritz and D. J. Williams, Initial observations of geomagnetically trapped alpha particles at the equator, J. Geophys. Res. 78 (1973) 4719– 4723. 3. W. I. Axford, On the origin of radiation belt and auroral primary ions, Particles and Field in the Magnetosphere, ed. B. M. McCormac (D. Reidel, Norwell, MA, 1970), pp. 46–59. 4. E. G. Shelley, R. G. Johnson and R. D. Sharp, Satellite observations of energetic heavy ions during a geomagnetic storm, J. Geophys. Res. 77 (1972) 6104–6110. 5. G. Gloeckler and D. C. Hamilton, AMPTE ion composition results, Phys. Scr. T18 (1987) 73–84. 6. D. C. Hamilton, G. Gloeckler, F.-M. Ipavich, W. Studemann, B. Wilken and G. Kremser, Ring current development during the great geomagnetic storm of February 1986, J. Geophys. Res. 93 (1988) 14343–14355. 7. I. A. Daglis, E. T. Sarris and B. Wilken, AMPTE/CCE CHEM observations of the ion population at geosynchronous altitudes, Ann. Geophys. 11 (1993) 685–696. 8. S. Y. Fu, B. Wilken, Q. G. Zong and Z. Y. Pu, Ion composition variations in the inner magnetosphere: Individual and collective storm eﬀects in 1991, J. Geophys. Res. 106, A12 (2001) 29683–29704. 9. S. Y. Fu, Q. G. Zong, T. A. Fritz, Z. Y. Pu and B. Wilken, Composition signatures in ion injections and its dependence on geomagnetic conditions, J. Geophys. Res. 107, A10 (2002) 1299–1313. 10. Q.-G. Zong et al., Geotail observation of energetic ion species and magnetic ﬁeld in plasmoid-like structures in the course of an isolated substorm event, J. Geophys. Res. 102 (1997) 11409–11428. 11. Q.-G. Zong et al., Energetic oxygen ion bursts in the distant magnetotail as a product of intense substorms: Three case studies, J. Geophys. Res. 103, A9 (1998) 20339–20364, 10.1029/97JA01146. 12. B. Wilken, Q.-G. Zong, T. Doke, T. Mukai, T. Yamamoto, G. D. Reeves, K. Maezawa, S. Kokubun and S. Ullaland, Substorm activity on January 11, 1994: Geotail observations in the distant tail during the leading phase of a corotating interaction region, J. Geophys. Res. 103, A8 (1998) 17671–17690, 10.1029/98JA00673.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch12

160

S. Y. Fu and Q.-G. Zong

13. Q.-G. Zong et al., Cluster observations of earthward ﬂowing plasmoid in the tail, GRL 31 (2004) L18803, doi:10.1029/2004GL020692. 14. R. Ruan, S. Y. Fu, Q.-G. Zong, Z. Y. Pu, X. Cao, W. L. Liu, X. Z. Zhou and P. W. Daly, Ion composition variations in the plasma sheet observed by Cluster/RAPID, Geophys. Res. Lett. 32 (2005) L01105, doi:10.1029/2004GL021266. 15. T. Doke et al., The energetic particle spectrometer HEP onboard the Geotail spacecraft, J. Geomagn. Geoelectr. 46 (1994) 713–733. 16. S. Kokubun, T. Yamamoto, M. Acuna, K. Hayashi, K. Shiokawa and H. Kawano, The Geotail magnetic ﬁeld experiment, J. Geomagn. Geoelectr. 46 (1994) 7–21. 17. A. T. Y. Lui, D. J. Williams, R. W. McEntire, S. P. Christon, T. E. Eastman, T. Yamamoto and S. Kokubun, Ion composition and charge state of energetic particles in ﬂux ropes/plasmoids, J. Geophys. Res. 103, A3 (1998) 4467–4476, 10.1029/97JA02256. 18. R. L. McPherron, Physical processes producing magnetospheric substorms and magnetic storms, Geomagnetism 4 (1991) 593–739. 19. I. G. Richardson and S. W. H. Cowley, Plasmoid-associated energetic ion bursts in the deep geomagnetic tail: Properties of the boundary layer, J. Geophys. Res. 90 (1985) 12133–12158. 20. D. N. Baker, T. I. Pulkkinen, P. Toivanen, M. Hesse and R. L. McPherron, A possible interpretation of cold ion beams in the earth’s tail lobe, J. Geomagn. Geoelectr. 48 (1996) 699–710. 21. M. B. Moldwin and W. Hughes, On the formation and evolution of plasmoids: A survey of ISEE 3 geotail data, J. Geophys. Res. 97 (1992) 19259–19282. 22. B. Wilken, W. Weiss, D. Hall, M. Grande, F. Sraas and J. F. Fennell, Magnetospheric ion composition spectrometer onboard the CRRES spacecraft, J. Spacecraft Rockets 29 (1992) 585–591. 23. Q.-G. Zong and B. Wilken, Energetic oxygen ion bursts in the distant magnetotail, MRAT Proceedings of the COSPAR Series, ed. A. T. Y. Lui (Elsevier Science, UK, 1997), pp. 23–32. 24. W. Baumjohann, Storm–substorm relationship, Proceeding of the Third Conference on Substorms (ESA Publications, 1996), pp. 627–632. 25. R. D. Sharp, D. L. Carr, W. K. Peterson and E. G. Shelley, Ion streams in the magnetotail, J. Geophys. Res. 86 (1981) 4639–4648. 26. A. G. Ghielmetti, R. G. Johnson, R. D. Sharp and E. G. Shelley, The latitudinal, diurnal, and altitudinal distributions of upward ﬂowing energetic ions of ionospheric origin, Geophys. Res. Lett. 5 (1978) 59–62. 27. T. Nagai, K. Takahashi, H. Kawano, T. Yamamoto, S. Kokubun and A. Nishida, Initial GEOTAIL survey of magnetic substorm signatures in the magnetotail, Geophys. Res. Lett. 21 (1994) 2991–2994. 28. A. Ieda, S. Machida, T. Mukai, Y. Saito, T. Yamamoto, A. Nishida, T. Terasawa and S. Kokubun, Statistical analysis of the plasmoid evolution with Geotail observations, J. Geophys. Res. 103, A3 (1998) 4453–4466, 10.1029/97JA03240.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch12

Energetic Particle Composition Signatures in the Magnetotail

161

29. E. W. Hones, Jr., Transient phenomena in the magnetotail and their relation to substorms, Space Sci. Rev. 23 (1979) 393–410. 30. J. A. Slavin, R. P. Lepping, J. Gjerloev, D. H. Fairﬁeld, M. Hesse, C. J. Owen, M. B. Moldwin, T. Nagai, A. Ieda and T. Mukai, Geotail observations of magnetic ﬂux ropes in the plasma sheet, J. Geophys. Res. 108, A1 (2003) 1015, doi:10.1029/2002JA009557. 31. A. T. Y. Lui, Current disruption in the Earth’s magnetosphere: Observations and models, J. Geophys. Res. 101 (1996) 13067–13088. 32. L. C. Lee, Z. F. Fu and S.-I. Akasofu, A simulation study of forced reconnection processes and magnetospheric storms and substorms, J. Geophys. Res. 90 (1985) 10896–10910. 33. J. G. Lyon, S. H. Brecht, J. D. Huba, J. A. Fedder and P. J. Palmadesso, Computer simulation of a geomagnetic substorm, Phys. Rev. Lett. 46 (1981) 1038–1041.

This page intentionally left blank

May 30, 2006 14:31 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch13

THE HIGH LATITUDE BOUNDARIES UNDER EXTREME SOLAR WIND CONDITIONS: A CLUSTER PERSPECTIVE H. ZHANG∗,‡ , T. A. FRITZ∗ , Q.-G. ZONG∗ and P. W. DALY† ∗Center for Space Physics, Boston University 725 Commonwealth Ave, Boston, MA 02215, USA †Planck

Institute for Solar System Research Katlenburg-Lindau, Germany ‡[email protected]

The high latitude boundaries include boundary between the magnetosheath and cusp, the boundary between the magnetosheath and the High Latitude Trapping Region (HLTR) which is the closed ﬁeld line region on the dayside in the high latitude region, the boundary between cusp and HLTR and the boundary between mantle and cusp. The properties of the high latitude boundaries vary rather dramatically under diﬀerent solar wind conditions. We present statistical results based on four years of data obtained by Cluster when these spacecraft were in the vicinity of the dayside magnetopause. During northward Interplanetary Magnetic Field (IMF), the interfaces between the magnetosheath and cusp are rather clear. The changes of the energetic particle ﬂux, plasma temperature, density, and velocity across the magnetopause under northward IMF were analyzed by superposed epoch analysis. The plasma ﬂow and density decrease, and the proton temperature increases across the magnetopause from the magnetosheath into the cusp. Further, during extreme storm times, the cusp is more turbulent than during quiet times and there is no clear plasma density change across the magnetopause. During low density solar wind conditions, the magnitude of the parallel temperature change across the magnetopause is larger than during high density solar wind conditions. And the magnitude of the plasma density and velocity change is smaller than during high density solar wind conditions.

1. Introduction High latitude magnetospheric boundary layer is a region adjacent to the magnetopause in which magnetosheath plasma has strong inﬂuence. There are four subregions: plasma mantle, entry layer, exterior cusp, Low-Latitude Boundary Layer (LLBL). The plasma mantle, which is located on the ﬁeld lines where the injected magnetosheath plasma continues tailward, was ﬁrst reported by Rosenbauer et al.1 The term “Low-Latitude Boundary Layer”, was 163

May 30, 2006 14:31 WSPC/SPI-B368 Advances in Geosciences Vol. 2

164

ch13

H. Zhang et al.

apparently introduced by Haerendel et al.2 to distinguish the very diﬀerent properties observed at latitudes below about 50–60◦ on the magnetopause surface from the plasma mantle found above this latitude. In addition there are two more boundary regions that are assumed to connect directly to the magnetosheath. The entry layer3 is located on the magnetospheric ﬁeld lines just equatorward of the cusp. It is a region of diﬀusive, turbulent entry of magnetosheath plasma onto ﬁeld lines that map to the low-altitude cusp. It has been indicated that in the entry layer the plasma density is almost as high as the magnetosheath but generally lacking the strong antisunward plasma ﬂow. In fact sunward ﬂow has even been reported by Paschmann et al.3 Lundin4 suggested that a characteristic feature of the entry layer is a strong variability of magnetosheath plasma entry with frequent plasma injection. A systematic study of LLBL was carried out by Phan et al.5,6 using the AMPTE/IRM data. The term “Inner Edge of the Boundary Layer” (IEBL) is adopted from Phan et al.,5,6 who deﬁne the IEBL as the location where the density has dropped to 5% of its magnetosheath value. Whether the IEBL is formed by a further penetration of solar wind plasma into the trapping regions on closed ﬁeld lines via diﬀusion process (e.g., eddy eﬀect2 ) or the appearance of magnetosheath plasma on interconnected ﬁeld lines whose ﬂux tube has not had time to become empty of its ion population, remains an open question. The statistical location of the high-latitude polar cusp and the shape of the surrounding magnetopause have been studied using Hawkeye data.7,8 Lavraud et al.10 investigated the global characteristics of the high altitude cusp and its surrounding regions using three-year Cluster data. They found that the boundary between the cusp and the magnetosheath has a sharp bulk velocity gradient as well as a density decrease and temperature increase as one goes from the magnetosheath to the exterior cusp. The cusp geometry has also been studied by MHD simulation.9 In this paper, we present statistical results of the plasma parameters of the boundary between magnetosheath and cusp (dark gray region shown in Fig. 1) during extreme solar wind conditions based on the data set obtained by Cluster when these spacecraft were crossing the high latitude regions. The extreme solar wind conditions in this paper include quite time (Northward IMF), extreme storm time (Dst < −100 nT), high solar wind density (n > 40 p/cm3 ), and low solar wind density (n < 5 p/cm3 ). The superposed epoch analysis method was used to study the parameter changes across a boundary.

May 30, 2006 14:31 WSPC/SPI-B368 Advances in Geosciences Vol. 2

The High Latitude Boundaries under Extreme Conditions

ch13

165

Fig. 1. Schematic illustration of the high latitude boundaries viewed from dusk direction. Four boundaries are shown: the boundary between mantle and cusp, the boundary between cusp and magnetosheath, the boundary between cusp and HLTR (High Latitude Trapping Region), and the boundary between HLTR and magnetosheath. In this paper, we will focus on the boundary between cusp and magnetosheath (the dark gray one).

2. Boundary and Clock Angle of the IMF The boundary between the magnetosheath and the cusp has been studied by Lavraud et al.10 The identiﬁcation of the cusp was made on the basis of the following criteria: (1) High-density plasma (comparable to that in the sheath) and (2) small or stagnant plasma ﬂow (Vx < 60 km/s). These criteria have been used by Zong et al.,11 Lavraud et al.,12 and Zhang et al.13 We have found that the boundary between the magnetosheath and the cusp is clear sometimes but unclear at other times. The deﬁnition for a clear boundary is: There is a jump in plasma ﬂow (> 30 km/s) (e.g., a ﬂow change from 100 to 0 km/s) and at least two components of the magnetic ﬁeld (> 5 nT) (e.g., the magnetic By and Bz change from −10 to +10 nT). Figure 2 shows an example for a clear and unclear boundary. Figure 2(a) shows a clear boundary observed by Cluster when it travels outbound from the northern magnetosphere into the magnetosheath on March 17, 2001. From this ﬁgure we can see that all the parameters including energetic proton and electron ﬂux, plasma density, velocity, and magnetic ﬁeld have clear boundaries. Figure 2(b) shows an unclear boundary observed by Cluster

May 30, 2006 14:31 WSPC/SPI-B368 Advances in Geosciences Vol. 2

166

ch13

H. Zhang et al.

Fig. 2. Examples for clear and unclear boundary. (a) Clear boundary on March 17, 2001 and (b) unclear boundary on March 19, 2001.

when it travels inbound from the magnetosheath into the southern magnetosphere on March 19, 2001. We can see from this ﬁgure that all the parameters except energetic proton ﬂux have no clear boundaries. We do not know exactly where the boundary is since the parameters change smoothly from the magnetosheath in the magnetosphere if we look at the plot for a longer time which is not shown here, so we put the vertical line which indicates the boundary location at the location where energetic proton ﬂux changes a lot. We have surveyed Cluster data in 2001 and 2002 and found that almost all the boundaries between the HLTR and the magnetosheath are clear (looks like 2(a)). However, the boundary between the cusp and magnetosheath is more complicated. Sometimes it is clear and sometimes it is

May 30, 2006 14:31 WSPC/SPI-B368 Advances in Geosciences Vol. 2

The High Latitude Boundaries under Extreme Conditions

ch13

167

Fig. 3. The IMF clock angle dependence of the boundary between the magnetosheath and cusp.

unclear (looks like 2(a) and 2(b) respectively). It is found that there is some relationship between the boundary and the clock angle of the IMF. In Fig. 3, all the boundaries between the cusp and the magnetosheath in 2001 and 2002 are shown. The arrows indicate the IMF direction projected in the GSE-YZ plane. The red arrows indicate unclear boundaries and green arrows indicate clear boundaries. In the shaded region, all the arrows are green which means when the IMF clock angle is between −65 and 81◦ , the boundary between the magnetosheath and the cusp is clear.

3. Quiet Time Versus Extreme Storm Time The properties of the high latitude boundaries vary rather dramatically under diﬀerent solar wind conditions. In order to study the average variations of key plasma parameters in the vicinity of the magnetopause under diﬀerent conditions, we perform a superposed epoch analysis. We present statistical results based on four years of data obtained by Cluster when these spacecraft were in the vicinity of the dayside magnetopause. Figure 4(a) shows superposed epoch analysis of the energetic particle ﬂux, the plasma temperature, density, and velocity change from cusp region

May 30, 2006 14:31 WSPC/SPI-B368 Advances in Geosciences Vol. 2

168

ch13

H. Zhang et al.

Fig. 4. Superposed epoch analysis of the energetic particle ﬂux, the plasma temperature, density, and velocity change from cusp region across the magnetopause under northward IMF in northern hemisphere.

across the magnetopause under northward IMF in northern hemisphere. The vertical dashed line marked the magnetopause position which is identiﬁed by the jump in plasma parameters including temperature, density and velocity. The x-axis is the minutes after outward magnetopause crossing. The time interval in this plot is 20 min, 10 min before and 10 min after the magnetopause crossing. During northward IMF, the interfaces between the magnetosheath and the cusp are rather clear. The plasma ﬂow and density increase and the proton temperature decreases across the magnetopause from the cusp into the magnetosheath. Figure 4(b) shows superposed epoch analysis of the same parameters as 4(a) but during extreme storm time (Dst < −100 nT). By saying an event is during extreme storm time we mean that the most negative Dst during one storm is < −100 nT and the event is observed during the storm time (initial phase, main phase or recovery phase). In Fig. 4(b) all the events during extreme storm time from 2001 to 2004 are included. If the magnetopause crossing is in the southern hemisphere, we reverse the time

May 30, 2006 14:31 WSPC/SPI-B368 Advances in Geosciences Vol. 2

The High Latitude Boundaries under Extreme Conditions

ch13

169

sequence so that the crossing is still from the magnetosphere into the magnetosheath. Compare Figs. 4(b) and 4(a), we can see that during extreme storm times, the cusp is more turbulent than during quiet times. We also noted that there is no clear plasma density change across the magnetopause during extreme storm time.

4. High Solar Wind Density Versus Low Solar Wind Density The variations of the plasma parameters across the magnetopause when the solar wind density is very high (n > 40 p/cm3 ) and very low (n < 5 p/cm3 ) have also been studied. Figure 5 shows superposed epoch analysis of the energetic particle ﬂux, the plasma temperature, density, and velocity change from cusp region across the magnetopause under northward IMF when solar wind density is low (n < 5 p/cm3 ). The ratio means the 10 min averaged values in the cusp region over that in the magnetosheath. The ratios of the plasma perpendicular and parallel temperature are 2.3 and 3.2, respectively. The ratios of the plasma density and velocity are 0.43 and 0.23, respectively. Figures 6(a) and 6(b) show superposed epoch analysis of the energetic particle ﬂux, the plasma temperature, density, and velocity change from cusp region across the magnetopause under southward IMF when solar 3 wind density is high (n > 40 p/cm3 ) and low (n < 5 p/cm ), respectively. Compare these two ﬁgures, we can see that during low density solar wind conditions, the magnitude of the parallel temperature change across the magnetopause is larger than during high density solar wind conditions, and the magnitude of the plasma density and velocity change is smaller than during high density solar wind conditions.

5. Discussion and Conclusions We have seen from Fig. 4 that during extreme storm time the cusp is more turbulent and the plasma density has no clear change across the boundary. This could be easily understood. During storm time, the whole magnetosphere is very active and the reconnected magnetic ﬁeld lines convect from the subsolar point to the tail passing the cusp region. The magnetopause might be rotational discontinuity during the storm time and the density is constant across the rotational discontinuity, so the density does not change across the magnetopause.

May 30, 2006 14:31 WSPC/SPI-B368 Advances in Geosciences Vol. 2

170

ch13

H. Zhang et al.

Fig. 5. Superposed epoch analysis of the energetic particle ﬂux, the plasma temperature, density, and velocity change from cusp region across the magnetopause under northward IMF when solar wind density is low (n < 5 p/cm3 ).

The plasma characters varies with solar wind density can be understood by the concept of pressure balance. When the solar wind density is very high (low), the plasma density is also very high (low) in the cusp region. To balance the magnetosheath pressure, the plasma temperature must be very low (high). We have studied the properties of the boundary between the magnetosheath and the cusp during extreme solar wind conditions based on four-year Cluster data. Our results can be summarized as follows: When IMF is northward, the interfaces between the magnetosheath and magnetosphere are rather clear. However, this interface will become uncertain when IMF turns southward. The plasma ﬂow and density decrease and the proton temperature increases across the magnetopause from the magnetosheath into the cusp.

May 30, 2006 14:31 WSPC/SPI-B368 Advances in Geosciences Vol. 2

The High Latitude Boundaries under Extreme Conditions

ch13

171

Fig. 6. Superposed epoch analysis of the energetic particle ﬂux, the plasma temperature, density, and velocity change from cusp region across the magnetopause under southward IMF. (a) When solar wind density is high (n > 40 p/cm3 ) and (b) when solar wind density is low (n < 5 p/cm3 ).

During extreme storm time, the cusp is more turbulent than quiet time and there is no clear plasma density change across the magnetopause. During low-density solar wind conditions, the amplitude of the parallel temperature change across the magnetopause is larger than during high-density solar wind conditions and the amplitude of the plasma density and velocity change is smaller than during high-density solar wind conditions. This study shows that solar wind conditions have strong inﬂuence on the cusp characteristics including density, plasma ﬂow, and ion temperature (not only location as shown in previous studies).

Acknowledgments We are grateful to the Cluster RAPID, CIS, and FGM teams for making available the data presented here. This work has been supported at Boston University by NASA grant NAG5-10108.

May 30, 2006 14:31 WSPC/SPI-B368 Advances in Geosciences Vol. 2

172

ch13

H. Zhang et al.

References 1. H. Rosenbauser, H. Gruenwaldt, M. D. Montgomery, G. Paschmann and N. Skopke, Heos 2 plasma observations in the distant polar magnetosphere: The plasma mantle, J. Geophys. Res. 80 (1975) 2723–2737. 2. G. Haerendel, G. Paschmann, N. Sckopke, H. Rosenbauer and P. C. Hedgecock, The frontside boundary layer of the magnetopause and the problem of reconnection, J. Geophys. Res. 83 (1978) 3195–3216. 3. G. Paschmann, G. Haerendel, N. Sckopke and H. Rosenbauer, Plasma and ﬁeld characteristics of the distant polar cusp near local noon: The entry layer, J. Geophys. Res. 81 (1976) 2883–2899. 4. R. Lundin, Plasma composition and ﬂow characteristic in the magnetospheric boundary layers connected to the polar cusp, The Polar Cusp, eds. J. A. Holtet and A. Egeland (Kluwer Academic Publishers, Dordrecht, The Netherlands, 1985), pp. 9–32. 5. T. D. Phan and G. Paschman, Low-latitude dayside magnetopause and boundary layer for high magnetic sheath 1. Structure and motion, J. Geophys. Res. 2101 (1996) 7801–7815. ¨ Sonnerup, Low-latitude dayside 6. T. D. Phan, G. Paschman and B. U. O. magnetopause and boundary layer for high magnetic sheath 2. Occurrence of magnetic reconnection, J. Geophys. Res. 2101 (1996) 7817–7828. 7. X.-W. Zhou and C. T. Russell, The location of the high-latitude polar cusp and the shape of the surrounding magnetopause, J. Geophys. Res. 102 (1997) 105–110. 8. T. E. Eastman, S. A. Boardsen, S.-H. Chen, S. F. Fung and R. L. Kessel, Conﬁguration of high-latitude and high-altitude boundary layers, J. Geophys. Res. 105 (2000) 23221–23238. 9. G. Siscoe and N. Crooker, Cusp geometry in MHD simulations, Surveys in Geophys. 26 (2005) 387–407. 10. B. Lavraud, A. Fedorov, E. Budnik, A. Grigoriev, P. J. Cargill, M. Dunlop, H. R`eme, I. Dandouras and A. Balogh, Cluster survey of the high-altitude cusp properties: A three-year statistical study, Ann. Geophys. 22 (2004) 3009– 3019. 11. Q.-G. Zong, T. A. Fritz, H. Zhang, A. Korth, P. W. Daly, M. W. Dunlop, K.-H. Glassmeier, H. Reme and A. Balogh, Triple cusps observed by cluster — Temporal or spatial eﬀect?, Geophys. Res. Lett. 31 (2004) L09810, doi:10.1029/2003GL019128. 12. B. Lavraud et al., Cluster observations of the exterior cusp and its surrounding boundaries under northward IMF, Geophys. Res. Lett. 29 (2002) 1995, doi:10.1029/2002GL015464. 13. H. Zhang, T. A. Fritz, Q.-G. Zong and P. W. Daly, Stagnant exterior cusp region as viewed by energetic electrons and ions: A statistical study using Cluster Research with Adaptive Particle Imaging Detectors (RAPID) data, J. Geophys. Res. 110, A5 (2005) A05211, doi:10.1029/2004JA010562.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch14

THE MAGNETOSPHERIC CUSP: STRUCTURE AND DYNAMICS Q.-G. ZONG∗,† , T. A. FRITZ∗ , H. ZHANG∗ , S. Y. FU‡ , X. Z. ZHOU‡ , M. L. GOLDSTEIN§ , P. W. DALY¶ , H. REME , A. BALOGH∗∗ and A. N. FAZAKERLEY†† ∗Center for Atmospheric Research, University of Massachusetts Lowell 600 Suﬀolk Street, Lowell, MA 01854-3629, USA †Laboratory for Space Weather, Chinese Academy of Sciences, Beijing, China ‡Institute of Space Physics and Applied Technology Peking University, Beijing 100871, China §NASA Goddard Space Flight Center, Greenbelt, MD, USA ¶Max-Planck-Institut fuer Sonnersystemfunsong, D-37191 Katlenburg-Lindau Germany Centre d’Etude Spatiale des Rayonnements, 9, avenue du Colonel Roche Toulouse Cedex 4, France ∗∗Space and Atmospheric Physics Group, Imperial College, London, UK ††Mullard Space Science Laboratory, Holmbury St. Mary Dorking, Surrey, RH5 6NT, UK

Understanding the polar cusps is essential for a thorough understanding of the entire physics of the magnetosphere, and of the dynamical interaction between the solar wind and any planetary magnetosphere. Energetic electrons are unique to fully assess magnetic ﬁeld-line topology and thus should be able to clearly delineate regions of open and closed magnetic ﬁeld lines in the high-latitude regions and contributed crucially to understanding and resolving an internal debate going on between groups measuring only the lower energy (< 20 keV) plasma. Energetic electrons with high and stable ﬂux were observed in the high latitude boundary/cusp region when the Interplanetary Magnetic Field (IMF) had a predominate positive Bz component. With measurements at larger separations and more coordination of multiple satellite measurements for particular cusp crossings it will become more evident what the true nature of the cusp is and what roles the cusp plays. The boundary normal, velocity and timing analysis obtained by all four Cluster spacecraft indicates that the multiple cusp phenomena are most likely caused by the oscillation of a single northern cusp which was shifted back and forth. Cusp oscillations with a period of ∼ 20 min are observed by Cluster in the high-latitude region, whilst the cold-dense plasma with ﬂuctuations (20-min period) are observed in the dusk-side of the tail plasma sheet by Geotail. This is consistent with the idea that the high latitude reconnection during northward IMF is the responsible mechanism for the formation of the cold-dense plasma sheet.

173

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch14

174

Q.-G. Zong et al.

1. Brief History on the Cusp • ∼Qin dynasty (221–206 BC) Magnetic compass discovered in China. The ﬁrst person recorded to have used the compass as a navigational aid was Zheng He (1371–1435), from the Yunnan province in China, who made seven ocean voyages between 1405 and 1433. • 1600 William Gilbert publishes in London “De Magnete” (“on the magnet”). His explanation of the compass: the Earth is a giant magnet. • Maxwell (∼1880) showed that a perfect conductor adjacent to a dipole formed an image dipole. • Chapman and Ferraro (1931) ﬁrst induced the basic nature of the Earth’s magnetosphere, its two-dimensional and three-dimensional topology has indicated the existence of a dayside magnetic cusp. • Spreiter and Summers (1962) predicted a stagnation ﬂow in the cusp region by using a gas dynamics model. • Heikkila and Winningham (1971) and Frank (1971) showed a highlatitude band of low-energy particle precipitation with magnetosheath (MS) like properties on the dayside at low altitudes which have been accepted as the ﬁrst evidence to discover the magnetospheric cusp. 2. Introduction to the Cusp The boundaries of the magnetosphere including the polar cusp are key regions for the transfer of mass, momentum and energy from the solar wind into the magnetosphere whether the Interplanetary Magnetic Field (IMF) is southward or northward. The ﬁrst identiﬁcation of a thin layer of MS plasma located immediately inside the magnetopause was made by Hones et al.,1 who also introduced the term “Boundary Layer.” Since then, the morphological characteristics as well as plasma properties of the magnetospheric boundary layer have been studied rather intensively.2–8 The term “low-latitude boundary layer (LLBL)” was apparently introduced by Haerendel et al.4 to distinguish the very diﬀerent properties observed at latitudes below about 50–60◦ on the magnetosphere side of the magnetopause surface. In addition there are three more boundary regions (at high latitude) shown in Fig. 1 which are assumed to connect directly to the MS: the plasma mantle (PM); the entry layer (EL); and the exterior cusp or stagnation region. The PM is located on open ﬁeld lines where the injected MS plasma continues tailward as ﬁrst reported by Rosenbauer.2 The plasma density in this region is toward sheath density level and the β 1. As shown in Fig. 1,

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch14

Magnetospheric Cusp

175

Fig. 1. Sketch of the dayside boundary regions related to the polar cusp ﬁeld lines during southward (after Haerendel et al.4 ) and northward IMF. MS: magnetosheath; PM: plasma mantle; LLBL: low-latitude boundary layer; EL: entry layer.

the EL9 is located on the magnetospheric ﬁeld lines just equatorward of the cusp. It is a region of diﬀusive, turbulent entry of MS plasma onto ﬁeld lines that map to the low-altitude cusp. It has been so termed because it appears to be the region of dominant plasma entry into the magnetosphere. The transport mechanism is likely to be achieved through eddy convection which manifests itself in the irregular, low speed plasma ﬂow, and may be incited by the turbulence in the adjacent exterior cusp.4 The exterior cusp/stagnation region is bounded inside by the cusp-like indentation of the magnetopause and outside by the free MS ﬂow which Sckopke et al.10,11 describe as a pocket of hot and “stagnant,” possibly turbulent plasma. In

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch14

176

Q.-G. Zong et al.

fact, as early as 1960s, the stagnation region is already predicted by gas dynamics models.12,13 Furthermore, this picture has been corroborated by HEOS-2 measurements.10,11 The stagnation region cannot be linked to the plasma mantle or LLBL in a simple way. A qualitative explanation was given by Haerendel et al.4 who noted the similarity of the situation near the cusp to hydrodynamic ﬂow around a corner, in which vortex formation and separation are known to occur and to initiate some level of turbulence (Fig. 1). The exterior cusp region appears to be a steady high pressure center of “stagnant” MS plasma, the ﬂow in this region is rather turbulent, both in magnitude and direction. The mantle is generally thicker for southward than northward IMF Bz.10 These ﬁrst researchers believed that the PM is open, and the LLBL is closed. It has been indicated that in the HLBL (entry layer) the plasma density is almost as high as the MS but generally lacks the strong antisunward plasma ﬂow. In fact, even sunward ﬂow has been reported.9 Lundin5,6 suggested that one of the characteristic features of the EL is the strong variability of MS plasma entry with frequent plasma injections. Only a few satellites such as HEOS-2, Prognoz-7, Hawkeye, and Polar have made in situ observation in the high-latitude boundary layer regions, where the proposed entry for the northward IMF takes place. The highlatitude boundary layer has scarcely been studied compared to the region around the subsolar point. 2.1. The deﬁnition of the magnetospheric cusp In text books, the polar cusps are usually deﬁned as funnel-shaped areas in the high latitudes of both hemispheres with near zero magnetic ﬁeld magnetitude. They provide a direct entry for the MS plasma into the magnetosphere.14,15 However, the deﬁnition of the cusp used by MHD simulation16 is “a weakening of the magnetic ﬁeld owing to a pool of MS plasma within the magnetosphere — since the current is the plasma’s diamagnetic current associated with the ﬁeld weakening” or in concept as “a region of open ﬁeld lines extending poleward from the open/closed boundary (which is tied to the dayside merging region on the magnetopause) to where particles are no longer are able to directly enter”. However, there is not always a clear distinction between such a conceptual deﬁnition. Some observational identiﬁcations, and actual determinations of when the cusp is really being observed by a given spacecraft.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch14

Magnetospheric Cusp

177

The primary and the most widely used method of identifying the cusp is by means of a combination of plasma and magnetic ﬁeld observations, although just plasma or magnetic ﬁeld measurements have been used in the past in some cases. Using both sets of observational data the cusp has been deﬁned as a high latitude region with a population of particles of shocked solar wind energies and density somewhere within or near the local noon sector, the criteria are: (1) (2) (3) (4)

turbulent and depressed magnetic ﬁelds, high-density plasma (∼ sheath level), Stagnant plasma ﬂow (V x ∼ 0), The clock angle criterion (the Cusp clock angle should be diﬀerent with the IMF’s), an example of observed cusp is shown in Fig. 2.

A number of questions now need to be investigated in detail: What is the importance of the cusp to the physics of the magnetosphere and the topology of the front side high-latitude magnetopause? What is the nature of the boundaries between diﬀerent regions? What is the plasma transport mechanism through the cusp and the boundary layers? Are the observed double or triple cusps temporal or spatial eﬀects? How are they formed? What is the role of the cusps in supplying plasma to the plasma

km/s

Vx(HIA)

nT

N(HIA)

Bt

Degree

(cm -3)

Cluster/C1 Mar. 4, 2002

Clock Angle(IMF) Clock Angle(CLUSTER)

Fig. 2. A typical cusp observed by Cluster satellite: CIS plasma moment and magnetic ﬁeld obtained by Cluster (Rumba) from 08:00 to 11:00 UT, March 4, 2002. The plasma density and plasma velocity V x are given in the ﬁrst and second panels. The magnetic ﬁeld magnitude (in nT) is shown in the third panel. The magnetic ﬁeld clock angle obtained by ACE (IMF) and Cluster satellites (local) are plotted in the bottom panel.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch14

178

Q.-G. Zong et al.

sheet? Are there multiple cusps active at a given time or does a single region move around more and faster than its low altitude counterpart? Only with measurements at larger separations and with coordination of multiple satellite measurements for particular cusp crossings will it become evident what the true nature of the cusp region is and what role or roles the cusps play. 3. Energetic Particles in the High-Latitude Boundary/Cusp Region 3.1. Observations Figure 3 shows energetic particle spectra in diﬀerent regions — southern HLTR/cusp, radiation belt and northern HLTR/cusp as obtained by Cluster/RAPID during a consecutive orbit. The two cusp crossings happened during a rather quiet time period. As we can see from Fig. 3, the highlatitude magnetosphere (both northern and southern HLTR/cusp) regions are two of three locations that energetic particle are encountered. Both southern and northern high-latitude boundary and/or cusp regions can be distinguished easily by the fact that the ion ﬂux increased sharply in all energy channels from 30 to 400 keV. During the above two quiet time high-latitude boundary/cusp crossings, there were pronounced ﬂuxes of electrons in the high-latitude boundary/cusp region, indicating either a closed ﬁeld line geometry in the cusp region, or a special open ﬁeld line conﬁguration that could trap electrons very eﬃciently for a long time. These electrons lasted from 1706 to 1907 UT on March 16, 2001 and 0600–0920 UT on March 17, 2001, respectively. Furthermore, no obvious substorm injections were observed by the Los Almos satellites for both of the above quiet time high-latitude boundary/cusp crossings. Further, there were no energetic electron events observed by ACE in the upstream interplanetary space during March 16–19, 2001. Thus, these observed electrons should not be solar energetic electrons as described by Lin17 and Klassen et al.18 The lack of substorm activity and high ﬂuxes of electrons upstream at ACE indicates that the observed electrons are locally trapped rather than substorm injected electrons drifting to the highlatitude region or solar ﬂare related electrons. In contrast, two consecutive HLBL/cusp crossings during geomagnetically disturbed times are. 3.2. Particle motion in the cusp region We have seen in Sec. 3.1 that there are energetic particles in the high-latitude boundary/cusp region during northward IMF. Statistical

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch14

Magnetospheric Cusp

179

Fig. 3. An overview of RAPID data from 17:00 UT, March 16 to 10:00 UT, March 17, 2001 together with geomagnetic activity Dst index. From the top the panels show: electron spectra from 20 to 400 keV; and proton spectra from 30 to 2000 keV. The marks indicate the diﬀerent regions (the south cusp, the radiation belt and the north cusp) which Cluster experienced during a full orbit. The “cusp” here refers to all the highlatitude magnetospheric regions, the “radiation belt” here refers to all inner magnetospheric region, the data gap is marked in white.

study showed that energetic ions have been observed in the cusp region during most of the crossings (80%).19 This indicates that those particles are trapped in the high-latitude region. In this section, we will interpret this observation with single particle simulation. According to the traditional dipole ﬁeld model, the dayside high latitude or cusp region cannot trap particles.20,21 The cusp region of the ideal dipole ﬁeld is not an “excluded region” in the St¨ omer theory.22 This means that, in the high latitude region, the particles cannot be trapped for much longer than the bouncing time; the E × B drift will take the particles away. However, the dipole ﬁeld can be modiﬁed fundamentally by interaction with the solar wind. The outer cusp regions where the magnetic ﬁeld lines either

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch14

180

Q.-G. Zong et al.

close in the dayside sector or extend into the nightside sector over the polar cap could be caused by reconnection. This is a region of weak magnetic ﬁeld, which arises directly from the interaction of the solar wind with the geomagnetic ﬁeld as predicted by Chapman and Ferraro23 using a simple image dipole, and demonstrated by the magnetic ﬁeld model of Antonova and Shabansky.24 Instead of a dipolar ﬁeld, the cusp region appears to be quadrupolar. However, the importance of the existence of an oﬀ-equatorial B-minimum in the outer cusp has been underestimated for a long time, although it could be of extreme importance for understanding the behavior of energetic particles in the magnetosphere. Antonova and Shabansky24 and Shabansky25 noted that, with a minimum magnetic ﬁeld existing oﬀ the equator in the outer cusp region, charged particles would not drift but rather branch oﬀ toward the magnetic ﬁeld minimum at high latitudes. This has also been supported by in situ magnetic ﬁeld measurements.26 Shabansky,27 Antonova and Shabansky28 provided observational evidence for the trapping of energetic particles (of several tens of keV, up to a few hundreds of keV) in the high-latitude region. Sheldon et al.29 pointed out that an energetic electron will drift on a closed path around the front of the magnetosphere, and found that electrons could be trapped in the outer cusp. In fact, a temporary trapping in the cusp ﬁeld minimum was ﬁrst examined by Delcourt et al.30 Further, Delcourt and Sauvaud31,32 pointed out that, under the eﬀect of the cuspward mirror force near the dayside magnetopause, the energetic plasma sheet particles initially mirroring near the equator are expelled from low latitudes and subsequently swept into the boundary layer at high latitudes. Figure 4 shows trajectories of test protons (1, 10, 100 keV) launched with 90◦ pitch angle from the cusp region. The trajectory tracing was performed using the Tsyganenko 96 model.31 In this calculation, the full particle dynamics have been considered, not just the guiding center computation; the calculation was performed using a fourth-order Runge–Kutta technique with a time step adjusted to some fraction of the particle gyration periods. It can be seen from Fig. 4 that the test protons launched from the local minimum magnetic ﬁeld region encircle the outer cusp region; all of the protons experience a pronounced bouncing motion in the high-latitude region which diﬀers from mirroring motion on either side of the equator (as ring current ions). Figure 4 shows that the ion trajectories in the outer cusp region are somewhat similar to those on L shells of a dipolar magnetic ﬁeld. The limiting second invariant of these trapped orbits occurs when the

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch14

Magnetospheric Cusp

181

Fig. 4. The trajectory of protons (1, 10 and 100 keV) with 90◦ pitch angle in the Tsyganenko 96 model.

mirror point Bmin approaches the dayside equatorial ﬁeld strength; in the local gradient ﬁeld they drift away from the cusp. These ion trajectories exist both on the dayside (equatorward, with closed magnetic ﬁeld lines) and the mantel region (poleward, with open magnetic ﬁeld lines). This behavior follows from the existence of a local B minimum during the drift path from the closed ﬁeld lines region to the open ﬁeld line region in the frontside magnetosphere. It should be pointed out that large-scale magnetospheric convection is not considered in the present modeling results in the trajectory computation. If convection is considered, as pointed out by Delcourt and Sauvaud,32 these closed drift paths in the outer cusp may be opened (see Fig. 13 of Ref. 32). However, the present modeling results could apply to quiet times when magnetospheric convection is reduced and the convection electric ﬁeld may be only 5% of its value during active times. The observed energetic electrons in the high latitude boundary region may be provided by tail plasma sheet particles because of a minimum magnetic ﬁeld existing oﬀ equator in the high-latitude region of the

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch14

182

Q.-G. Zong et al.

magnetosphere. Delcourt and Sauvaud32,33 pointed out that under the eﬀect of the cuspward mirror force near the dayside magnetopause, energetic plasma sheet particles initially mirroring near the equator are expelled from low latitudes and subsequently swept into the boundary layer at high latitudes. Both electrons and ions can be stably trapped in the high-latitude region during quiet periods. This conclusion is supported by both the observations and the modeling results mentioned above. As magnetospheric convection is enhanced, the electrons initially trapped in the high-latitude region could be de-trapped.31,32 In fact, no stable trapped electrons were observed during active times. These de-trapped electrons could further form an electron layer just outside the magnetopause as observed.34,35

4. The Cusp Dynamics Unraveling the structure of the dayside cusp is a major Cluster objective. Four cusp-like regions were observed consecutively in about ﬁve hours on March 21–22, 2001 by all four Cluster spacecraft when the IMF was northward with a signiﬁcant By component. The four-fold cusp (Fig. 5) was surrounded by the dayside magnetosphere rather than the magnetopause. Evidences were presented indicating that the multiple cusps were probably a temporal sequence. The boundary normal, velocity and timing analysis for six clear boundaries of the cusps indicated that the observed cusps are mostly caused by oscillation of a single cusp which was shifted back and forth between the dayside magnetosphere/trapping region and the cusp region. The normal velocities at boundary interfaces for exit from the cusp were found to be almost three times as large as that for entry into the cusp (Fig. 6). These observations suggest that the shape and location of the cusp is often changing as a result of dynamic processes in the high-latitude regions. Furthermore, by combining the four-Cluster spacecraft positions and crossing times at the interfaces, we are able to determine the normal velocity and direction of the discontinuity based on the triangulation method.36 Assuming planar discontinuity and the speed of the discontinuity was constant in time and space over the Cluster separation distance, the equation is simply given by R1 • n = V n • t1 . Here V n is the normal velocity of the discontinuity, R1 = (r12, r13, r14) is a tensor with r{ij} being the spacecraft separation vectors and

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch14

Magnetospheric Cusp

183

Fig. 5. Four-fold cusp observed by Cluster. Plasma ion and electron spectra are overplotted with energetic particle in the ﬁrst two panels. Last panel shows the total magnetic ﬁeld.

t1 = (t{12}, t{13}, t{14}) consists of the diﬀerences between the crossing times of the corresponding spacecraft. The obtained normal speeds for the six interfaces are given in Fig. 6 (between panels 2 and 3). As we can see from panels 3 and 4, the polar θ angles shows little change in the ﬁrst four interfaces, however, they change signs for the last two interfaces. The wavelike motion can be more clearly seen in the φ angle, the direction of exiting the cusp at 2, 4, 6 is opposite to the entering direction based on the azimuthal angle (Fig. 7). Although Cluster observed four cusp regions, the analysis shows that three of these cusps (later three cusps), which are located in the lowerlatitude, are a temporal feature which can be attributed to cusp boundary movements or possibly wave activities. During this event, the IMF was steady with IMF Bz > 0, and IMF By is slightly larger or comparable with IMF Bz. Between 0100 and 0330 UT on March 22, 2001, the solar wind speed was 300 km/s and the radial

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch14

184

Q.-G. Zong et al.

Fig. 6. Electron ﬂux, magnetic ﬁeld By component measured by the four-Cluster spacecraft during the time period from 01:00 to 03:30 UT on March 22, 2001. The boundary normals were determined by MVA for all available spacecraft during cusp crossings. One of the three cusps is expanded to see the spacecraft crossing order more clearly.

dynamic pressure was around 3 nPa obtained by Wind satellite located at GSM (−11.7, −183.6, −109.6) Re (see Fig. 7). The time lag between Cluster and Wind is about 6 min with Cluster in advance. A solar wind pressure pulse encountered the Earth at ∼ 0058 UT, March 22, 2001, shortly before the second cusp was observed by Cluster while there was a change of solar wind azimuthal ﬂow (shown in Fig. 8). The solar wind East/West ﬂow changed from 6◦ to about 2◦ , and the V x change from −280 to −310 km/s. The Earth’s magnetosphere bears an analogy to the windsock response to changes of the solar wind component ﬂow as suggested by Zong et al.37 When the solar wind azimuthal ﬂow encountered the Earth, the ratio of the Y and Z components of the solar wind dynamic pressure to solar wind thermal pressure Pdy/Pth became 29%. Therefore, the position of the cusp will be changed. Cluster spacecraft entered the cusp, then the solar wind dynamic pressure or other eﬀects shift the cusp back and forth three times as if Cluster ﬂew through three cusps.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch14

Magnetospheric Cusp

185

Fig. 7. Solar wind and IMF observed by the Wind satellite. The Heliospheric Current Sheet (HCS) which is identiﬁed by the high plasma density and sign change of the IMF Bx is marked. The time of this ﬁgure is shifted according to the solar wind velocity.

On the other hand the second and third observed cusps do not seem to have clear links with the solar dynamic pressure or ﬂow components. Nevertheless, the mentioned solar wind pressure pulse could trigger the magnetospheric boundary wave.38 When the IMF is northward, a cold and dense plasma sheet is often observed.39–42 The MS plasma near the cusp region is relative cold compared to magnetospheric plasma, dense and almost stagnant. When cusp reconnection occurs, the newly reconnected ﬂux tubes of both hemispheres sink and contract into the magnetosphere. Subsequently it sweeps around the ﬂank, and is convected tailward. As the plasma is captured and transported to the tail, it is moderately heated near the reconnection site, and the temperature is just below 1 keV.43 The density in the captured ﬂux tubes is characteristic of cold, dense and almost stagnant plasma. The values of temperature, density and low ﬂow speed for the captured plasma are in good agreement with Geotail observations in the dusk ﬂank as shown in Fig. 8. It should be noted, however, the cold-dense plasma sheet is also waving. The period is about 20 min. This agrees with the cusp oscillating period

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch14

186

Q.-G. Zong et al.

Fig. 8. Geotail observations at the duskside tail-ﬂank. From top, the ion density, average energy (temperature, in keV), and the ion bulk ﬂow V x in km/s. Bursty-like cold-dense plasma is observed.

(a rough period of 22 min) very well. This is an additional evidence to support the idea that the cold-dense plasma sheet observed by Geotail is closely related the oscillating cusps which are probably formed by highlatitude reconnection during the extended northward IMF period.

5. Conclusions Energetic ions and electrons have been observed during quiet time (IMF Bz has a predominate positive component). Single particle simulation shows that energetic ions could be temporally trapped in the high-latitude/cusp region whereas electron could not be. Multiple Cusps are observed by Cluster satellite in the high latitude region, in the meantime, the cold-dense plasma with ﬂuctuations are observed in the dusk-side of the tail plasma sheet by the Geotail satellite. This is consistent with the idea that the high-latitude reconnection during northward IMF is the responsible mechanism for the formation of the cold-dense plasma sheet. The observed multiple cusps are very possibly

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch14

Magnetospheric Cusp

187

temporal sequence. The cusp was shifted by the solar azimuthal dynamic pressure or wave back and forth three times in about 5 h interval as if Cluster ﬂew through the cusp four times. Further, we conﬁrm that the solar wind azimuthal ﬂow is the controlling factor of the cusp position and is as strong as, potentially even stronger than, that of the IMF By/Bz component. The full impact of the cusp is going to be evident when we have multiple satellites displaced from one another by large distances (from a large fraction of an Earth’s radius to a few Earth radii), as well as satellites located within 100 km of one another and observing using interferometric techniques. New Cluster orbits will make possible study of the nature of particle boundaries within the cusp and high-latitude regions to resolve the mechanisms that transport, and possibly accelerate, the thermalized plasma and energetic particles in and through the cusp and boundary layers. The multiscale Cluster separations will help to resolve the issue of whether the observed multiple cusps are a temporal or spatial eﬀect, as well as determining the size of the cusp. The question of how multiple cusps are formed will be addressed. An outstanding question is the role and importance of the cusp in supplying plasma to the plasma sheet and the relationship between high-latitude reconnection and the appearance of cold-dense plasma within the plasma sheet. Acknowledgments The present study has been supported by CNSF under projects 40528005, 40374060. This research was supported at Boston University by NASA grant NAG 5-10108 and the International Collaboration Research Team program of the Chinese Academy of Sciences. We are grateful to the Cluster CIS, FGM and RAPID team. One of authors QGZ thanks the Editor-inChief, Dr. Marc Duldig, for his careful reading and editing of this paper. References 1. E. W. Hones, S. I. Akasofu, S. J. Bame and S. Singer, J. Geophys. Res. 77 (1972) 6688. 2. H. Rosenbauer, J. Geophys. Res. 80 (1975) 2723. 3. T. E. Eastman, Geophys. Res. Lett. 3 (1976) 685. 4. G. Haerendel, G. Paschmann, N. Sckopke, H. Rosenbauer and P. C. Hedgecock, J. Geophys. Res. 83 (1978) 3195. 5. R. Lundin, The Polar Cusp, eds. J. A. Holtet and A. Egeland (Kluwer Academic Publishers, Dordrecht, 1985), p. 9. 6. R. Lundin and E. M. Dubinn, Planet. Space Sci. 33 (1985) 891.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch14

188

Q.-G. Zong et al.

7. P. T. Newell and C. I. Meng, J. Geophys. Res. 93 (1988) 14549. 8. P. T. Newell and C. I. Meng, Polar Cap Boundary Phenomena, eds. J. Moen, A. Egeland and M. Lockwood (Kluwer Academic Publishers, Dordrecht, 1998), p. 91. 9. G. Paschmann, G. Haerendel, N. Sckopke and H. Rosenbauer, J. Geophys. Res. 81 (1976) 2883. 10. N. Sckopke, G. Paschmann, H. Resenbauer and D. H. Fairﬁeld, J. Geophys. Res. 81 (1976) 2687. 11. N. Sckopke, G. Paschmann, G. Haerendel, B. U. Sonnerup, S. J. Bame, T. G. Forbes, J. E. W. Hones and C. T. Russell, J. Geophys. Res. 86 (1981) 2099. 12. J. R. Spreiter and A. L. Summers, Planet. Space Sci. 15 (1967) 787. 13. J. R. Spreiter and S. S. Stahara, J. Geophys. Res. 85 (1980) 6769. 14. P. H. Reiﬀ, T. W. Hill and J. L. Burch, J. Geophys. Res. 82 (1977) 479. 15. G. T. Marklund, L. G. Blomberg, C.-G. F¨althammar, R. E. Erlandson and T. A. Potemra, J. Geophys. Res. 95 (1990) 5767. 16. G. Siscoe, N. Crooker, K. Siebert, N. Maynard, D. Weimer and W. White, Surveys in Geophysics 26 (2005) 387. 17. R. P. Lin, Sol. Phys. 100 (1985) 537. 18. A. Klassen, V. Bothmer, G. Mann, M. J. Reiner, S. Krucker, A. Vourlidas and H. Kunow, Astron. Astrophys. 385 (2002) 1078. 19. H. Zhang, T. A. Fritz, Q.-G. Zong and P. W. Daly, J. Geophys. Res. 110 (2004), A05211, doi:10.1029/2004JA010562. 20. J. G. Roederer, Dynamics of Geomagnetically Trapped Radiation (SpringerVerlag, New York, 1970). 21. J. G. Roederer, Space Sci. Rev. 21 (1977) 23. 22. C. St¨ omer, Arch. Sci Phys. Nat. Ser. 4 32 (1911) 117. 23. S. Chapman and V. C. A. Ferraro, Terr. Magn. Atmos. Elect. 36 (1931) 77. 24. A. E. Antonova and V. P. Shabansky, Geomagn. Aeron. 8 (1968) 801. 25. V. P. Shabansky, Space Sci. Rev. 8 (1968) 366. 26. X.-W. Zhou, C. T. Russell, G. Le and N. Tsyganenko, Geophys. Res. Lett. 24 (1997) 1451. 27. V. P. Shabansky, Space Sci. Rev. 12 (1971) 299. 28. A. E. Antonova and V. P. Shabansky, Geomagn. Aeron. 15 (1975) 243. 29. R. B. Sheldon, H. E. Spence, J. D. Sullivan, T. A. Fritz and J. Chen, Geophys. Res. Lett. 25 (1998) 1825. 30. D. C. Delcourt, T. E. Moore, J. A. Sauvaud and C. R. Chappell, J. Geophys. Res. 97 (1992) 16833. 31. N. A. Tsyganenko and D. P. Stern, Modeling the global magnetic ﬁeld of the large-scale Birkeland current systems, J. Geophys. Res. 101 (1996) 27187– 27198. 32. D. C. Delcourt and J.-A. Sauvaud, J. Geophys. Res. 103 (1998) 26521. 33. D. C. Delcourt and J.-A. Sauvaud, J. Geophys. Res. 104 (1999) 22635. 34. C. I. Meng and K. A. Anderson, J. Geophys. Res. 75 (1970) 1827. 35. D. N. Baker and E. C. Stone, Geophys. Res. Lett. 4 (1977) 395. 36. C. T. Russell, M. M. Mellot, E. J. Smith and J. H. King, J. Geophys. Res. 88 (1983) 4739.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch14

Magnetospheric Cusp

189

37. Q.-G. Zong, T. A. Fritz, H. Zhang, A. Korth, P. W. Daly, M. W. Dunlop, K.-H. Glassmeier, H. Reme and A. Balogh, Geophys. Res. Lett. 31 (2004) L09810, doi:10.1029/2003GL019128. 38. D. G. Sibeck, N. L. Borodkova, G. N. Zastenker, S. A. Romanov and J.-A. Sauvaud, Geophys. Res. Lett. 25 (1998) 453. 39. W. Baumjohann, G. Paschmann and C. A. Cattell, J. Geophys. Res. 94 (1989) 6597. 40. M. Fujimoto, A. Nishida, T. Mukai, Y. Saito, T. Yamamoto and S. Kokubun, J. Geomagn. Geoelectr. 48 (1996) 711. 41. M. Fujimoto, T. Terasawa, T. Mukai, Y. Saito, T. Yamamoto and S. Kokubun, J. Geophys. Res. 103 (1998) 4391. 42. T. Terasawa et al., Geophys. Res. Lett. 24 (1997) 935. 43. P. Song and C. T. Russell, J. Geophys. Res. 97 (1992) 1411.

This page intentionally left blank

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch15

INITIAL RESULTS FROM THE SIMULATION OF THE HALLOWEEN 2003 STORMS RAMON E. LOPEZ∗,¶ , SALVADOR HERNANDEZ∗ , MICHAEL WILTBERGER† , JOHN LYON‡ and CHARLES GOODRICH§ of Physics and Space Sciences, Florida Institute of Technology 150 W. University Boulevard, Melbourne, FL 32901, USA

∗Department

†NCAR/HAO, ‡Department

3450 Mitchel Lane, Boulder, CO 80301, USA

of Physics and Astronomy, Dartmouth College Hannover, NH 03755, USA

§Department

of Astronomy, Boston University Boston, MA 02115, USA ¶relopez@ﬁt.edu

In this paper, we present initial results from a global MHD simulation of the Halloween 2003 storms. The simulation seems to be representing well the global conﬁguration of the magnetosphere based on comparison to geosynchronous observations. The global conﬁguration showed a highly variable magnetospheric size, at times 40% the size of the regular magnetosphere. The simulation also reproduces the saturation of the polar cap potential, albeit at larger values than are observed, and there is evidence that the Chapman–Ferraro current is not exerting the primary force balance with the solar wind when the polar cap potential saturates. Moreover, the simulation develops an asymmetry in the polar cap potential during the saturation periods, with the higher conductivity polar cap having a lower saturation potential.

1. The Halloween Storms and Their Simulation The storms of late October and early November 2003, known as the “Halloween” storms,1,2 have received an increasing amount of attention from researchers because of the extreme kind of activity they represent. The ICMEs produced by the Sun were among the fastest ever seen, one with a transit time of only 19 h from Sun to Earth.1 Such an event is a valuable opportunity to test simulations of the magnetosphere and understand how well they can reproduce magnetospheric behavior during extreme conditions. The simulation code to be used in this study is the Lyon–Fedder– Mobarry code, which solves the full time-dependent MHD equations over the whole magnetosphere.3 The simulation is not a model in that there are no a priori assumptions made about the structure of the magnetosphere. 191

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch15

192

R. E. Lopez et al.

The only free parameters in the simulations are the solar wind input and the ionospheric conductivity. The code solves the MHD equations on a computationally rectangular but nonorthogonal mesh that is adapted to the magnetospheric problem. The outer boundary considerations for the simulation domain are the solar wind data and supersonic outﬂow. Thus the simulation can be driven with real solar wind input. In the case of the Halloween storms there is a problem, namely that the plasma instrument on ACE was disabled by the event and did not produce reliable densities.1 However, we have been able to reconstruct a nominal solar wind data ﬁle using densities derived from the Geotail Plasma Wave data, projecting the densities back to L1 ballistically using ACE plasma velocities, then merging the resulting plasma data with the ACE magnetometer data to make a merged ﬁle. That ﬁle, plotted in Fig. 1, was used to provide the solar wind boundary condition for the LFM simulation run. The self-consistent ionospheric boundary condition is another important feature of the code that is essential to accurate modeling of the magnetosphere. The code deﬁnes an inner boundary to the MHD calculation, generally set at about 3 RE . The ﬁeld-aligned currents are calculated at that altitude from the curl of B, and mapped to ionospheric altitudes where the height-integrated electrostatic equation is solved for the ionospheric potential. The ionospheric electric ﬁeld is then mapped back to the MHD grid to provide a boundary condition for Faraday’s Law and for the perpendicular velocity. The height-integrated ionospheric conductivity is based on an empirical model, modiﬁed by the ﬁeld-aligned currents.4 The LFM allows us to include the dipole tilt into the simulation, and this has a strong eﬀect on the ionospheric conductivities, with the summer polar cap having much higher conductivity than the winter polar cap. To gain a general appreciation of the event and the magnetospheric response, an overview movie of the simulation run was generated. Several frames of this movie from October 29 are presented in Fig. 2. Features such as the bow shock and magnetopause are quite evident in the density. The 0216 UT frame (top, left) shows a close to average magnetosphere with the magnetopause along the Earth–Sun line located at 9 RE , in line with the solar wind values at that time. At 0705 UT the magnetosphere was extremely compressed by the solar wind, which had a speed of about 1700 km/s and a density around 16 cm−3 . The magnetopause along the Earth–Sun line was located at about 4 RE , which is close to what one would expect from pressure balance.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch15

Simulation of the Halloween 2003 Storms

193

Fig. 1. The reconstructed solar wind data for October 29–30, 2003, described in more detail in the text, that was used to drive the LFM simulation of the event.

At 1347 UT the magnetosphere was also very compressed, and the position of GOES 12 was outside the magnetosphere, indicating that the spacecraft should have made a magnetopause crossing. However, the IMF and magnetosheath ﬁeld pointed northward at that time, as can be seen by

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch15

194

R. E. Lopez et al.

Fig. 2. Four frames from the simulation overview movie on October 29 (from top, left to right) at 0216, 0705, 1347, and 1852 UT. The view is a perspective view of the equatorial plane in SM coordinates. The natural log of the density is in gray-scale and the solar wind magnetic ﬁeld magnitude and direction in the Y –Z plane are included. The tic marks are at 5 RE intervals. The black circle at the origin marks the inner boundary of the MHD simulation. Geosynchronous orbit is plotted as a circle, distorted because of the perspective view in SM coordinates, with the dark part of the trace indicating the part of the orbit that is below the SM equatorial plane. The white dots indicate the position of the GOES 10 and 12 spacecraft (GOES 10 is at later local time than GOES 12), and the arrow indicates the Z-component of the simulation ﬁeld interpolated to the spacecraft position.

the interpolated ﬁeld at the GOES 12 position. A geosynchronous magnetopause crossing under northward IMF is an exceedingly rare event5,6 and testimony to the extreme nature of this storm. At 1852 UT, both simulated GOES spacecraft found themselves in the magnetosheath and observed negative BZ magnetic ﬁelds, due to a combination erosion of the magnetopause6 and large dynamic pressure. Overall, the magnetosphere scale was at times half of the normal size. Another broad overview of the event is presented in Fig. 3, which presents the polar cap potentials and the integrated ionospheric joule heating from the simulation. As expected, during the periods of large southward IMF the polar cap potentials were large, as was the ionospheric joule heating.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch15

Simulation of the Halloween 2003 Storms

195

Fig. 3. Simulation results for integrated joule heating (top panel), polar cap potential (middle panel) and the potential diﬀerence between the northern and southern polar caps. The top two panels show each result for each polar cap, north, and south.

2. Simulation Field at Geosynchronous Orbit To determine if the simulation accurately represented the magnetosphere during this event, we can make a quantitative comparison between the geosynchronous magnetic ﬁeld interpolated to the GOES positions and the actual magnetic ﬁeld observed by the GOES spacecraft. Figure 4 presents that comparison for the Bz -component of the ﬁeld. Inspecting Fig. 4 we see that the simulation reproduced to a signiﬁcant degree the variation of the Bz -component of the magnetic ﬁeld. In particular, one can see that simulation did a good job of representing the magnetopause position. The simulation also reproduced with reasonable ﬁdelity the other components of the magnetic ﬁeld observed by the GOES spacecraft were. This suggests that the simulation did a fairly good job of representing the magnetosphere as a whole, and it also suggests that the reconstructed solar wind ﬁle was a faithful representation of the real solar wind input during this event.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch15

196

R. E. Lopez et al.

Fig. 4. The Bz -component of the magnetic ﬁeld observed by GOES 10 and 12 on October 29 and 30, along with the Bz -component of the magnetic ﬁeld from the simulation interpolated to the GOES positions.

3. Saturation of the Polar Cap Potential It has long been recognized that the polar cap potential is controlled by the solar wind electric ﬁeld under southward IMF conditions.7 Boyle et al.8 provided an empirical relationship between the solar wind speed, magnetic ﬁeld (magnitude and direction), and the polar cap potential. This empirical relationship provided a continuously increasing potential as the solar

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch15

Simulation of the Halloween 2003 Storms

197

wind electric ﬁeld, VB s , increased. That study also found no evidence of a saturation of the polar cap potential for large values of VB s , as had been suggested earlier.9 However, the number of samples at large values of VB s in the data were limited, as were periods of large solar wind density during strongly southward IMF.10,11 Without signiﬁcant samples of magnetospheric response during periods of large VB s , there was no statistical evidence either for polar cap saturation or for a density dependence in the potential.8 However, observations of the polar cap potential rarely exceed 200 kV,7 whereas the Boyle et al. relationship8 would predict a potential of 384 kV for a solar wind with Bx = 5 nT, By = 10 nT, Bz = −30 nT, and V = 700 km/s. Such solar wind values were observed for several hours on March 31, 2001. DMSP F-13 was located such that it was able to make reliable measurements of the potential during this period, yet the measured potential did not exceed about 250 kV. Russell et al.7 also presented evidence from several events to demonstrate that the saturation eﬀect is real and not an artifact of sparse observations. It has been argued that the saturation of the polar cap potential occurs because there is a limit to the size of the Region 1 current system.12 The Region 1 current system reduces the dayside magnetic ﬁeld, leading to magnetopause erosion.6 But there must be a limit to the reduction of the dayside ﬁeld. When this limit is reached, it may be that magnetosphere enters a fundamentally diﬀerent state in which the Chapman–Ferrao current disappears and force balance with the solar wind is accomplished via the Region 1 currents.12 Siscoe et al.12 deﬁne the saturation regime as times when the solar wind electric ﬁeld (due to Bs ) is about 5 mV/m for a nominal 2 nPa dynamic pressure. Siscoe et al.12 also derive a form of the saturation potential that (for large solar wind electric ﬁelds) is inversely dependent on the inospheric conductivity and proportional to the cube root of the solar wind dynamic pressure. We have previously studied the how pressure increases in the solar wind are linked to an increase the magnetospheric driving.11 However, the mechanism we discussed in that paper was not related to the issue of saturation — we discussed how modifying the upstream Mach number would change the magnetosheath conditions and so aﬀect the transfer of energy to the magnetosphere. Thus it is important to separate both eﬀects, which both depend on the solar wind pressure, but very diﬀerent physics. One point of interest is that as the GOES spacecraft crossed the magnetopause (such as around 1800 UT on October 30), the Bz magnetic ﬁeld

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch15

198

R. E. Lopez et al.

at the GOES spacecraft positions in both the simulation and reality was not far from the dipole value of 109 nT. This suggests that the Chapman– Ferraro current, which increases the magnetic ﬁeld in the magnetosphere, was not extremely strong (certainly not strong enough to overwhelm the southward perturbations created by the much more distant Region 1 and cross-tail currents), yet this was at a time when there was a huge force exerted on the magnetosphere. Such an observation is evidence in support of a greatly diminished force on the solar wind from the Chapman–Ferraro current, which would leave only the Region 1 current as a large-scale current capable of exerting a force on the solar wind. And the magnetopause crossings occurred at a time when the IMF was southward and the solar wind electric ﬁeld was large — precisely the conditions for a saturation of the polar cap potential. So the magnetospheric behavior at the magnetopause in the simulation is consistent with the Siscoe et al.12 model for saturation, but did the polar cap potential in fact saturate? The answer lies in Fig. 3. One thing that can be seen immediately is that the potentials are much larger than are observed. However, it is known that the LFM potentials are about a factor of 2–3 too large,4 and, in fact, DMSP recorded peak potentials of 240 kV during this event (M. Hairston, personal communication, 2005). During the ﬁrst hour of October 29, the solar wind speed was about 500 km/s and Bz was −8 nT. Around 2000 UT on the same day the solar wind electric ﬁeld was around 6–7 times greater, however, the potentials at that time were only a bit more than two times the potential at the start of the day. Similar results have been found by others13 — the response of the simulation is not linear and that there is a saturation eﬀect in the simulated potential. Inspecting the top panel of Fig. 3, one can see that southern polar cap had signiﬁcantly more integrated joule heating than did the northern polar cap. This is an ionospheric conductivity eﬀect since the time of year means that the southern hemisphere is more illuminated than the northern hemisphere. This results in a higher conductivity in the southern polar cap. On the other hand, there is an asymmetry in the polar cap potential, with a higher potential in the less conductive polar cap. This potential asymmetry is actually consistent with the theoretical formulation of the polar cap saturation potential, which depends inversely on ionospheric conductivity.12 Thus an ionosphere with a lower conductivity should have a higher saturation potential. This is what we ﬁnd in the simulation, but it leads to a potential diﬀerence between the hemispheres. We

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch15

Simulation of the Halloween 2003 Storms

199

are currently working to determine if this asymmetry is real (through comparison to observations), and if so, what is the voltage source that maintains the asymmetry.

4. Conclusions In this paper, we report initial results from the global MHD simulation of the Halloween 2003 storms. We ﬁnd that the simulation provides a good representation of large-scale magnetospheric structure and dynamics based on a comparison of simulation results to geosynchronous data. The simulation showed a highly dynamic magnetosphere that at times was compressed to less than half its regular size. The simulation also exhibited saturation of the polar cap potential during strong driving by the solar wind, and this potential saturation was attended by a diminished role of the Chapman–Ferraro current in maintaining force balance with the solar wind. The simulation also developed a pronounced asymmetry in the polar cap potential that is consistent with the dependence of the saturation potential on ionospheric conductivity.

Acknowledgments This work was supported by CISM, which is funded by the STC Program of the National Science Foundation under Agreement Number ATM-0120950, and NASA grant NAG5-1057.

References 1. R. E. Lopez, D. Baker and J. Allen, EOS 85 (2004) 105. 2. T. G. Onsager, B. Poppe and W. Murtagh, Space Weather 2, 8 (2004), doi:10.1029/2004SW000099. 3. J. G. Lyon, J. A. Fedder and C. M. Mobarry, J. Atmos. Solar Terr. Phys. 66 (2004) 1333, doi:10.1016/j.jastp.2004.03.020. 4. S. Slinker et al., J. Geophys. Res. 104, A12 (1999) 28379. 5. C. L. Rufenach, R. F. Martin, Jr. and H. H. Sauer, J. Geophys. Res. 94, A11 (1989) 15125. 6. M. Wiltberger, R. E. Lopez and J. G. Lyon, J. Geophys. Res. 108, A6 (2003) 1235, doi:10.1029/2002JA009564. 7. C. T. Russell, J. G. Luhmann and G. Lu, J. Geophys. Res. 106, A9 (2001) 18495. 8. C. P. Boyle, P. H. Reiﬀ and M. Hairston, Empirical polar cap potentials, J. Geophys. Res. 102, A1 (1997) 111–125.

May 30, 2006 9:24 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch15

200

R. E. Lopez et al.

9. P. H. Reiﬀ and J. G. Luhmann, Solar wind control of the polar-cap voltage, Solar Wind-Magnetosphere Coupling, eds. Y. Kamide and J. A. Slavin (Terra Sci., Tokyo, 1986), pp. 453–476. 10. M. R. Hairston, T. W. Hill and R. A. Heelis, Observed saturation of the ionospheric polar cap potential during the 31 March 2001 storm, Geophys. Res. Lett. 30 (2003) 1325, doi:10.1029/2002GL015894. 11. R. E. Lopez, M. Wiltberger, S. Hernandez and J. G. Lyon, Solar wind density control of energy transfer to the magnetosphere, Geophys. Res. Lett. 31 (2004) L08804, doi:10.1029/2003GL018780. 12. G. L. Siscoe, N. U. Crooker and K. D. Siebert, Transpolar potential saturation: Roles of region 1 current system and solar wind ram pressure, J. Geophys. Res. 107, A10 (2002) 1321, doi:10.1029/2001JA009176. 13. V. G. Merkin, A. S. Sharma, K. Papadopoulos, G. Milikh, J. Lyon and C. Goodrich, Global MHD simulations of the strongly driven magnetosphere: Modeling of the transpolar potential saturation, J. Geophys. Res. 110 (2005) A09203, doi:10.1029/2004JA010993.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch16

ROUND-THE-CLOCK GROUND-BASED IMAGING SPECTROSCOPY OF SPACE WEATHER EFFECTS ON THE THERMOSPHERE AND IONOSPHERE SUPRIYA CHAKRABARTI∗ and D. PALLAMRAJU† Center for Space Physics, Boston University 725 Commonwealth Avenue, Boston, MA 02215, USA ∗[email protected] †[email protected]

Visible optical emissions from the Earth’s upper atmosphere carry signatures of the complex manifestation of the Sun–Earth connection. For a comprehensive understanding of the Space Weather eﬀects on the thermosphere and ionosphere, we need to extend ground based optical observations to daytime (sunlit hours). We have developed two versions of an imaging spectrograph that is capable of routinely measuring faint airglow/auroral emissions buried in the bright solar background continuum of the daytime sky in the visible wavelength range — a multiwavelength instrument, called High Throughput Imaging Echelle Spectrograph (HiTIES) with moderate (0.03 nm) resolution and another, the High Resolution Imaging Spectrograph using Echelle grating (HIRISE), with slightly higher (0.01 nm) resolution. These rugged instruments have been used to investigate such wide-ranging solar-terrestrial problems as the forecasting of Equatorial Spread-F development using 630.0 nm dayglow, sunlit cusp, proton aurora as well as a daytime aurora over Boston. These proof-of-concept experiments have been validated by simultaneous, independent observations and have demonstrated their value for studies in space physics and aeronomy. We are presently incorporating improved capabilities and plan to deploy these instruments for various Space Weather studies. In this paper, we review the scientiﬁc contributions we have already made and our future plans.

1. Introduction The Earth’s upper atmosphere is replete with various intricately interwoven and complex phenomena. The upper atmosphere is highly dynamic as it is constantly perturbed by not only the incident energy in the form of solar radiation, particle precipitation, etc., but also by gravity waves from below. Measurements of optical airglow/auroral emissions from atomic and molecular constituents present one of the important and useful means of investigating the atmospheric dynamics. As diﬀerent emissions originate at diﬀerent altitudes, simultaneous measurements of optical emissions at 201

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

202

ch16

S. Chakrabarti and D. Pallamraju

diﬀerent wavelengths yield critical insights into the dynamics of the atmosphere (such as the neutral temperature, winds, gravity waves, etc.) at the altitude of their origin. In spite of decades of observational and theoretical studies, there are many science issues that are not completely understood. They include the production mechanism of irregularities in the low- and equatorial-latitudes, the occurrence and movement of polar cap patches, and the generation of mid-latitude irregularities. These irregularities span several scale sizes and aﬀect radio communication and navigational geolocation. With increasing dependence on technology, both commercial and defense assets are aﬀected by such irregularities produced by geomagnetic storms and neutral dynamics. It is therefore essential that we have a complete understanding of the necessary background conditions and the physical processes by which these irregularities occur to be able to predict them consistently and thereby provide reliable advance warnings on the occurrence of such events. Most of the ground-based optical observations have been limited to night time (solar depression angle > 20◦ ). For most of the Earth, this observational constraint implies that ground based observations are limited to only 30–40% of the time. The observations are further limited to a typical period of two weeks out of a month centered around the new moon. Optical aeronomers have been looking for ways to overcome these observing limitations so that several important questions such as the conjugate aurora, sunlit cusp, Equatorial Spread F (ESF) triggers, etc., could be studied continuously round-the-clock much like radars. Note that these optical instruments, while powerful, will still be at the mercy of favorable weather conditions for the observations. The potential science yield from daytime optical measurements have been discussed extensively during the two CEDAR Workshops in 1997 and in 2002. Moreover, the topic of “Daytime airglow studies of ionospheric structures” has been recognized as being among the top seven highlighted topics of research by the CEDAR Steering committee in 2003 (http://cedarweb.hao.ucar.edu/community/POST−47.pdf). Daytime optical emission measurement from the ground is a signiﬁcant technical advancement in its own right (see the review by Chakrabarti1). Imaging spectrographs can make meaningful scientiﬁc contribution if they have high (< 0.05 nm) spectral resolution. This allows emissions from the atmosphere, which are typically line emissions, to overcome the scattered solar continuum. In addition, if these spectrographs have a large (∼ 180◦ ) ﬁeld of view (FOV) foreoptic, a wide spatial region can be sampled

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch16

Round-the-Clock Ground-Based Imaging Spectroscopy of Space Weather Eﬀects 203

simultaneously. Lastly, most studies of aeronomic interest require simultaneous observation of multiple spectral features, irrespective of their wavelength of occurrence. Conventional imaging spectrographs cannot meet all these requirements. At Boston University, we have designed an instrument by employing an echelle grating as the dispersing element. Echelle-based spectrographs are generally used in astronomical applications. They typically operate in high diﬀraction orders (50–100) and use a low ruling density (around 100 grooves/mm). These spectrographs are compact and optically fast, thereby delivering high optical throughput. However, since they operate at high spectral orders, spectra from diﬀerent orders tend to occupy same space on the focal plane. In astronomy applications, where the sources are typically point-like, a cross-dispersing element separate these spectra in the direction perpendicular to the dispersion direction. This is not a suitable solution for extended sources such as airglow or auroral emissions. Our implementation of two versions of high-throughput and high-resolution imaging spectrographs (Refs. 2 and 3) use interference ﬁlters for order sorting. These instruments have demonstrated their eﬀectiveness by recording atmospheric emissions from several geophysical processes that were buried in the strong solar scattered background continuum.4–8 In this review, we ﬁrst describe the two instruments and provide examples of some of the scientiﬁc studies. We then describe a study involving geomagnetic storm eﬀects on the OI 630.0 nm airglow emissions over lowgeomagnetic latitudes. Finally, we describe our plans for a new instrument that will extend our present capabilities.

2. The Instruments We have developed two instruments — a High Throughput Imaging Echelle Spectrograph (HiTIES)2 and the High Resolution Imaging Spectrograph using Echelle grating (HIRISE).3 They are similar in concept — the primary diﬀerence being their spectral resolution. HIRISE, shown in Fig. 1, has a 0.01 nm resolution compared to the 0.03 nm resolution of HiTIES. Another key diﬀerence is that in the current implementation of HIRISE, it is used only in studies involving a single spectral feature, while HiTIES is used when multiple spectral features, located anywhere in the visible range, are observed simultaneously. Here we brieﬂy describe the components of HIRISE instrument (shown in Fig. 1). The fore optics (1) consists of objective and ﬁeld lenses that

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

204

ch16

S. Chakrabarti and D. Pallamraju

Fig. 1. A schematic drawing of HIRISE. Incident light from the top encounters the ﬁlter which is chosen to approximately deﬁne the pass band of the instrument. The echelle grating is located near the top. Other optical components are used for packaging the instrument.

determine the FOV of the instrument. The light allowed through the slit assembly (2) is isolated over a small spectral range (∼ 10 nm) by the interference ﬁlter (3). Such light is collimated (4) and diﬀracted (5) by an Echelle grating and the imaging lens (6) focuses this diﬀracted beam of light on to a 1 k × 1 k CCD (7). The lenses (4) and (6) are apochromats and their diﬀraction limit is about a factor of three better than the pixel size of the CCD (24 µ). As a test of the performance of the instrument, we ﬁrst compared the theoretical performance, described by the Monte-Carlo ray-trace

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch16

Round-the-Clock Ground-Based Imaging Spectroscopy of Space Weather Eﬀects 205

Fig. 2. Performance summary of the echelle-based imaging spectrograph. (a) Ray-trace results showing the location of the two lines of sodium doublet at 589.0 and 589.6 nm on the focal plane. (b) Spectrum recorded in the laboratory with a sodium lamp. (c) Solar spectrum recorded with the instrument pointed at the sky shows copious Fraunhofer absorption lines including the sodium doublet. (d) The response of the instrument with, as input, both the solar spectrum as well as the laboratory lamp, showing the line emissions ﬁlling in the Fraunhofer absorption doublet.

calculations to the actual performance of the spectrograph. The results, shown in Fig. 2, indicate that the instrument worked as expected both while being illuminated by a laboratory calibration lamp as well as when it pointed at the Sun. The two recorded spectra (panels (c) and (d)) near the sodium doublet are presented in Fig. 3. For the actual measurements, we place a hood/light baﬄe that is painted black, on top of the objective lens to reduce the scattered solar glare into the spectrograph. HIRISE obtains high spectral resolution images and different regions orthogonal to wavelength in these images can be mapped to diﬀerent regions in space due to the imaging property of the instrument. Each of the spectral images is divided into 5–7 regions in order to obtain spatial structures in emissions. The photons from each region of the spectral image are then added to increase the signal to noise ratio (SNR) of the measurement. This spectrum is then compared to the reference solar spectrum to match in wavelength as well as to match the continuum regions through intensity scaling. The reference solar spectrum is obtained by convolving

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

206

ch16

S. Chakrabarti and D. Pallamraju

Intensity (in DN/s)

8000 6000

4000

2000

(a) Solar spec. showing sodium abs. lines 0 5870

5880 5890 Wavelength (in Angstrom)

5900

Intensity (in DN/s)

8000 6000

(b) Solar spec. with lab. sodium lamp

4000

2000 0 5870

5880 5890 Wavelength (in Angstrom)

5900

Intensity (in DN/s)

8000 6000

(c)

4000

2000 0 5888

5890

5892 5894 Wavelength (in Angstrom)

5896

5898

Fig. 3. Spectra recorder near 589 nm by the imaging spectrograph. (a) The solar spectrum clearly shows two deep absorption features due to sodium (see Panel (c) of Fig. 2). (b) The spectrum recorded when the instrument was pointed at the Sun with the line emissions from the sodium lamp in the ﬁeld of view (extracted from Panel (d) of Fig. 2). (c) Same as (b) but with the wavelength scale expanded to demonstrate the high spectral resolution of the instrument that clearly show the narrow emissions lines from the sodium lamp above the solar Fraunhofer absorption lines.

the solar spectrum obtained at the Kit Peak solar observatory with the instrument function of HIRISE. From such comparison the contribution of daytime atmospheric emissions are obtained. It is a challenge to detect the dayglow signal which could be as small as 0.1% of the solar background. Several parameters such as, the signal brightness, background levels, ﬁlter transmission, detector quantum eﬃciency, dark noise, binning of pixels, co-adding of images, optical eﬃciency of the instrument, etc., need to be considered to calculate the

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch16

Round-the-Clock Ground-Based Imaging Spectroscopy of Space Weather Eﬀects 207

exact SNR. Refer to the CEDAR tutorial on “Errors in airglow and auroral measurements” for further details on this issue (http://cedarweb.hao. ucar.edu/workshop/tutorials/2003/pallamraju03.pdf). Here, we present a short, order of magnitude estimate to show that although the background is very bright it is indeed possible to make airglow intensity measurements during daytime. The solar scattered background changes from around 100 to 1 MR/nm for solar zenith angles of 20◦ –90◦ . HIRISE measured OI 630.0 nm dayglow emission brightness varies from 8 to 10 kR during local noon which is consistent with the WINDII measurements carried onboard UARS.9 For our estimation, considering the solar scattered background of 100 MR/nm at 20◦ SZA, for a pass band of 0.01 nm yields a background of 1 MR. For a 8 kR signal the required measurement is 8 kR/1MR (or 0.8%). We usually operate the CCD detector at around 100 000 e− /pixel. So binning three pixels after reading out the CCD to obtain a 0.01 nm resolution (dispersion of HIRISE is 0.004 nm/pixel) yields a total electrons of 300 000. Considering only the poisson statistics the uncertainty in the detection of electrons is (300 000)(0.5) = 547; which is 547/300 000 (or 1 in 1000). Hence, the detection limit of the detector is better than that required to make the daytime airglow measurement. In this 300 000 e− , the electrons due to dayglow are 0.8% = 2400 e− , which gives a SNR of the measurement to be 2400/547 = 4.3. This is for one 0.01 nm spectral element. Data from larger spatial extent (“regions” described above) in such spectral segments are co-added to further increase the SNR by a factor of the square root of the number of the 0.01 nm spectral elements that are co-added. For the data analysis described in Sec. 3, varying number of rows (from 7–26) were added for each region. This varying number of pixel rows on the CCD map to approximately equal spatial extents in latitudes. Hence the SNR of the HIRISE measured dayglow presented here varies from 11 to 22, respectively. Other factors such as the dark noise, noise associated with the A/D conversion, readout noise, etc., are negligible compared to the poisson statistics and do not signiﬁcantly aﬀect the SNR of the daytime emission measurement. Thus, HIRISE is capable of making high spectral resolution (λ/∆λ = 50 000 approx) measurements at a single emission wavelength during daytime.3 This new technique has already been proven with many ﬁrst of their kind results in the ﬁeld of Upper atmospheric research. Examples of these include observation of daytime aurora over Boston,8 groundbased detection of magnetospheric cusps from Sondrestromfjord,6 obtaining experimental evidence of the existence of circulation cells in the neutrals during geomagnetic storms,7 measurement of auroral arcs in the daytime from Sondrestromfjord4 and solving the mystery surrounding the variability

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

208

ch16

S. Chakrabarti and D. Pallamraju

of Ring eﬀect (atmospheric scattering eﬀect) in the daytime sky.10 The various capabilities and possibilities of achievements by the HIRISE technique have been summarized in Ref. 11. In the following section, we describe a study conducted using the HIRISE instrument on the compositional changes due to a geomagnetic storm. The observations were made in Carmen Alto, Chile (23.16S; 70.66W, 10.2◦ S dip lat) in the autumn of 2001.

3. A Study of the Eﬀect of a Geomagnetic Storm on Thermospheric Composition in Low-Latitude In this study, we observed the daytime OI 630.0 nm red line airglow emission enhancements caused by the geomagnetic-storm-induced changes in atmospheric composition over low latitudes. Unusual emission enhancements were observed in the morning hours when the equatorial electrodynamics is not expected to show any appreciable variations. Simultaneous GPS-derived Total Electron Content (TEC) measurements do not show increase, clearly implying no role of electron densities in the observed emission enhancement. Using this information and with the understanding of the storm time behavior of low-latitude upper atmosphere, we interpret that neutral density enhancements are responsible for the observed dayglow emission enhancements. The FOV of HIRISE can be varied from 8◦ to 180◦ by changing its objective lens. Typical spectral images that were obtained with a 140◦ FOV is shown in Fig. 4. The x-axis of this image is wavelength, and the orthogonal direction indicates the slit orientation. The dark bands in this image are the Fraunhofer absorption lines. Due to the imaging property of HIRISE different regions of the spatial direction can be mapped onto diﬀerent regions in the sky. A spatial resolution of 0.5◦ in latitude along the orientation of the slit is possible by HIRISE. Note the 5:45 and 18:58 LT panels show twilight conditions, where the OI 630.0 nm emission line is visible above the continuum. 3.1. Observations The Dst variation for three days starting with November 5, 2001 (see Fig. 5) indicates that the magnetic disturbance eﬀects started during the second half of November 5. By the beginning of November 6, the Dst value changes from −50 to −300 nT in approximately 4–5 h. November 6 was a disturbed

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch16

Round-the-Clock Ground-Based Imaging Spectroscopy of Space Weather Eﬀects 209

Fig. 4. Spectral images obtained on October 22, 2001 from Carmen Alto at four representative local times. The presence of Fraunhofer absorption lines in all four panels indicate that the upper atmosphere was illuminated by sunlight. At lower solar depression angles (twilight times), the OI 630.0 nm emission line intensity clearly shows up as the emission line. For daytime images, the solar scattered background is signiﬁcantly high and so the emissions are not seen in these images.

100

NT

0

DST

-100

NOV 2001 -200

-300 0

3

6

9

12

15

18

21

24

27

30

Day of the month

Fig. 5. Dst index for three days November 5–7, 2001. Storm sudden commencement occurred on the early hours of the 6th with the Dst reaching a value of −300 nT indicating a severe storm occurrence. The Kp and Ap indices peaked at 9− and 300, respectively, on this day.

day with severe disturbances of Kp = 9− for six hours (the corresponding Dst was about −300 nT). November 7 shows a gradual recovery with Kp receding to 2. The Ap index, which most closely describes the disturbance eﬀects at low- and equatorial-latitudes, shows only November 6 to be a

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

210

ch16

S. Chakrabarti and D. Pallamraju

severely disturbed day with a peak index of 300 (compared to the November 5 and 7 with peak Ap around 50). The daily average values ( Ap ) for the three days were 21, 142 and 19, respectively. The F10.7 cm solar ﬂuxes for the three days were similar (230.6, 230.2, and 263.9, respectively). As HIRISE is an imaging spectrograph, light incident from diﬀerent view directions is recorded on diﬀerent pixels on the detector as shown in Fig. 4. For this campaign, observations were carried out over 140◦ FOV. Assuming a constant emission altitude of 230 km for the redline dayglow emission,12–14 these view angles can be approximated to diﬀerent latitudes. They correspond to approximately 7◦ in latitude from 6.5◦ to 13.5◦ magnetic latitude (south). Assuming a “uniform” thickness of the emission layer, we corrected the data in diﬀerent view directions for van Rhijn eﬀect. Figures 6–8 show two-dimensional plots of the brightness of OI 630.0 nm dayglow emissions for November 5–7, respectively. The x-axis indicates the local time, y-axis shows magnetic latitude in the southern hemisphere and the color scale represents OI 630.0 nm emission brightness. Note that the emissions from any view angle other than for zenith “cuts across” diﬀerent latitude locations and as such the magnetic latitudes (shown on the y-axis) should be treated as a representative value only. First, we will discuss the behavior of emissions on the quiet days, November 5 and 7 (Figs. 6 and 8). The ﬁfth data do not show any brightness enhancement in the morning in any latitude. The peak in emissions, seen during 1100–1500 LT at around 9◦ magnetic latitude, is most likely due to the variations in the electron densities, which develop in response to the variations in the equatorial electrodynamical drifts and electric ﬁelds. Indeed the TEC measurements on November 5 and 7 show equatorial ionization anomaly (EIA) crests at 1800 LT, albeit with in the emissions on these two days (Figs. 6 and 8). These diﬀerences are believed to be due to the variation in the strengths of equatorial electrodynamics. Although the dynamical component in the afternoon on both these days shows diﬀerences in terms of EIA development, the morning time behavior is similar. In comparison to the quiet days the magnetically disturbed day, November 6 (Fig. 7), shows emission enhancements by a factor of 2–3 in the morning from 0530 to 0830 LT over all observed latitudes. The temporal extent of the enhancement is larger in the 9◦ –13.5◦ magnetic latitude sector as compared to the 7◦ –9◦ magnetic latitude sector. The afternoon emission enhancements are most likely due to the neutral winds and tides propagating from high-latitudes toward low latitudes that are triggered by the geomagnetic storm. Compared to the quiet days, diﬀerent strengths that

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch16

Round-the-Clock Ground-Based Imaging Spectroscopy of Space Weather Eﬀects 211

Fig. 6. 2D plots of the OI 630.0 nm dayglow for November 5, 2001. The x-axis represents local time and y-axis indicates the approximate magnetic latitudes in the southern hemisphere from assuming a constant emission altitude of 230 km. A peak is seen centered at around 1400 LT at 9◦ magnetic latitude, which is most likely the response to the EIA development. This weaker development of EIA as seen in the dayglow emissions is an indication of weaker equatorial electrodynamics and hence no ESF occurrence is expected for this night. Compare this day with the 7th (Fig. 8) when stronger EIA development occurred.

correspond to the diﬀerences in the peak location the emissions measured on this day are enhanced in almost all view directions at all times, presumably caused by the redistribution in the neutral densities. The zenith measurement (10◦ Mag. Lat.) on November 6 show a decrease in emissions after around 9 LT and then a typical increase later in afternoon. The decrease in emissions is due to the cessation of the redistribution from high latitudes in concert with reduction in Kp values from 9 to 7 (not shown here) on that day. The observed morning emission enhancement can be caused by either an increase in electron and oxygen densities, or a decrease in N2 density, or changes in the neutral and electron temperatures, or an enhancement in photoelectrons ﬂux or solar UV photons. At any given latitude, the solar photons vary only with the sunspot number and solar cycle. Hence one would not expect any short jump in them to cause dayglow emission enhancements through solar photodissociation mechanism, which is one of the source for OI 630.0 nm photons. Similarly, photoelectrons are not

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

212

ch16

S. Chakrabarti and D. Pallamraju

Fig. 7. Same as in Fig. 6 but for November 6, 2001. Notice the sharp rise in emissions during morning hours (0530–0830 LT), which are 2–3 times the quiet time emissions as can be seen in Figs. 6 and 8 (green instead of dark blue). The time duration of morning rise is larger in 9◦ –13.5◦ magnetic latitudes compared with the duration of increase in 7◦ – 9◦ magnetic latitudes. Emission enhancements are also seen around 12◦ and 8◦ magnetic latitudes around 1300 LT and 1400 LT, respectively, most likely indicating propagation of neutral species from high- to low-latitudes. The contribution of the dynamical component is largest on this day compared with the quiet days shown in Figs. 6 and 8, possibly due to the presence of comparatively larger neutral densities that are redistributed by the magnetic storm eﬀects.

expected to vary abruptly during the course of the day to cause a similar increase in the dayglow brightness. The parameter that usually shows most short-term variations is the electron density. However, comparison with simultaneous GPS based TEC measurements on all three days ruled out the possibility of electron density variation contributing to the observed dayglow emission enhancement. Therefore, the only other parameter that can produce the observed dayglow emission enhancement is the short-term variation in the neutrals (densities and temperature). From satellite measurements it is known that neutral densities increase in low-latitudes during geomagnetic storms (for example, Refs. 15–18, 20). This redistribution is caused by transport of neutrals from hot polar latitudes to colder low- and equatorial-latitudes18,19 by TADs.20–22 An essential feature of such a TAD is that it carries equatorward-directed winds along with it. Unfortunately, simultaneous neutral wind data from low-latitudes are not available for comparing the

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch16

Round-the-Clock Ground-Based Imaging Spectroscopy of Space Weather Eﬀects 213

Fig. 8. Same as in Fig. 6 but for November 7, 2001. Notice the strong emissions at around 12◦ –13◦ magnetic latitude at 1300 LT. This represents the optical signature of the development of EIA in the daytime. This day shows relatively stronger EIA development compared with the 5th (see Fig. 6). Such strong EIA development is conducive for post-sunset ESF occurrence. Jicamarca digisonde data conﬁrmed the ESF occurrence on this night. Enhanced 630.0 nm emissions in the afternoon may prove to be the crucial precursors to the post-sunset ESF occurrence.

neutral dynamical variations on the quiet and disturbed periods as the days being discussed are close to the full-moon phase (full moon night was November 1). However, the inter-hemispheric asymmetry in the TEC data on November 06 suggests the existence of strong equatorward winds on this day. We believe that the dayglow emission enhancement seen in HIRISE from 0530–0830 LT (1015–1315 UT) is a direct result of such an increase in neutral densities over low-latitudes. Melendez-Alvira et al.,23 and Witasse et al.,13 respectively, studied the sensitivity of 630.0 nm twilight airglow and dayglow to [O2 ], [N2 ], and [O]. They showed that an increase in oxygen density by a factor of two increases the peak volume emission rate by around 20% through the Photoelectron Impact mechanism.13 Similarly, doubling [O2 ] can increase the total O (1 D) production by at least 60% through photodissociation and dissociative recombination mechanisms.23 Decrease in the N2 density by a factor of two increases the red line volume emission rate due to reduction in quenching by about 15%.12,23 Thus, the sensitivity of the dayglow emission rates to changes in various neutral species is interrelated and is very complex.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

214

ch16

S. Chakrabarti and D. Pallamraju

Hence, in the absence of simultaneous observations of the altitude distribution of the neutral composition during this storm, it can only be inferred that the increase in observed emissions are most likely due to signiﬁcant increase in neutral densities over low-latitudes during November 6 storm. With the capability of HIRISE and HiTIES demonstrated in several proof-of-concept experiments, we believe that the strengths of these two instruments are high spectral resolution, imaging capability, coverage of multiple wavelength ranges simultaneously and high throughput. In the next section we describe our plans for the future.

4. Multiple Wavelength Daytime Emission Measurements Presently the HIRISE spectrograph is operated in a “single-wavelength” mode, where we focus on one spectral feature and record the spectrum around it. However, it is capable of making measurements at multiple wavelengths, albeit over only a small FOV. In the present set-up we can place a “mosaic” ﬁlter (see Fig. 9) orthogonal to the slit eﬀectively making many small slits stacked on top of one another so that diﬀerent sections of the image correspond to diﬀerent wavelength ranges. This arrangement was used to carry out investigations on the Ring eﬀect.10 When we increase the FOV this image will correspond to diﬀerent spatial locations in the sky. In

Fig. 9. A ﬁve-panel mosaic ﬁlter developed for use with the HiTIES spectrograph. Each panel covers roughly 15 nm — suﬃcient to examine an atmospheric spectral feature in great detail. The spectrograph, as it exists now, is limited to 0.03 nm resolution, but can cover 5–10 15-nm bands located anywhere in the visible spectral range. We have a new design of the higher resolution HIRISE instrument, which will extend its capability to such multiple bandpasses at 0.01 nm resolution.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch16

Round-the-Clock Ground-Based Imaging Spectroscopy of Space Weather Eﬀects 215

300 km

230 km

100 km

HIRISE FOV

Fig. 10. Schematic representation of the sky coverage of a three-color oxygen experiment at 557.7 (green), 630.0 (red) and 777.4 nm (burgundy). The multiwavelength HIRISE will be able to simultaneously observe these three emissions in the day time, which will be able to resolve a key issue on the generation of ESF trigger by gravity waves.

order to obtain multiple wavelength information from a wide spatial location (over a large FOV), we have designed a new optical arrangement to HIRISE that is similar to the HiTIES instrument. For the daytime studies the new multiwavelength spectrograph will be capable of simultaneously measuring OI 557.7, OI 777.4 and OI 630.0 nm emissions. In the daytime, compared to the altitude of origin for the 6300 emission, the 557.7 and 777.4 nm emissions originate from lower (∼ 100 km) and higher (∼ 300 km) altitudes (see Fig. 10). Therefore, such a spectrograph will be capable of providing information on the vertical coupling of waves in the upper atmosphere during daytime, which has enormous applications at all latitudes. Note that the instrument can be used to measure other wavelength ranges by simply employing a diﬀerent set of ﬁlters. In the following section, we describe two other representative scientiﬁc studies that will beneﬁt from the new experimental capability. 4.1. All-sky maps of daytime volume emission rates Presently HIRISE obtains all-sky information along the slit orientation. We plan to rotate the instrument to obtain information from diﬀerent spatial

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

216

ch16

S. Chakrabarti and D. Pallamraju

locations. Such information could be employed to obtain 2D spatial maps of daytime optical emissions to investigate the wave propagations in diﬀerent directions during both quiet and magnetically disturbed conditions from diﬀerent latitudes. Once the multiwavelength instrument is developed, we plan on augmenting its capability to obtain 2D all-sky maps. We expect a 2D all-sky image to be obtained in around 30 min. The scientiﬁc yield of such a capability will greatly advance our current understanding of the atmospheric dynamics at multiple altitudes. 4.2. Simultaneous measurements from a HIRISE chain As HIRISE is capable of making round-the-clock all-sky measurements, a chain of such spectrographs can yield information on the behavior of the neutral atmosphere under varying magnetic conditions. Such a chain, called the CEDAR Optical Tomographic Imaging Facility (COTIF), was established earlier for the investigations of various nighttime phenomena. We can now establish a similar chain of HIRISE spectrographs to make round-the-clock investigations of various upper atmospheric phenomena. Information from such experiments could be used in assimilative models to better understand the atmospheric behavior in order to make accurate predictions of both the quiet time and the magnetic disturbance eﬀects on the atmospheric dynamics.

5. Summary The ability to observe airglow and auroral emissions during sunlit hours is one of the last frontiers of observational aeronomy. We have described the tremendous progress that has been made in recent years. With the successful execution of several proof-of-concept experiments, the instruments appear to be ready for routine round-the-clock studies of space weather and other dynamical phenomena.

Acknowledgments This work was supported by the National Science Foundation grant ATM0209796 to Boston University. The authors thank Jeﬀrey Baumgardner for helpful discussions. We would also like to acknowledge T. R. Pedersen at AFRL and J. Araya at UCN, Antofagasta, Chile for their logistic support in carrying out experiments from Carmen Alto, Chile.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch16

Round-the-Clock Ground-Based Imaging Spectroscopy of Space Weather Eﬀects 217

References 1. S. Chakrabarti, J. Atmos. Solar-Terr. Phys. 60 (1998) 1403–1423. 2. S. Chakrabarti, D. Pallamraju, J. Baumgardner and J. Vaillancourt, J. Geophys. Res. 106 (2001) 30337–30348. 3. D. Pallamraju, J. Baumgardner and S. Chakrabarti, J. Atmos. Solar-Terr. Phys. 64 (2002) 1581–1587. 4. D. Pallamraju, J. Baumgardner, S. Chakrabarti and T. R. Pedersen, J. Geophys. Res. 106 (2001) 5543–5549. 5. M. Galand, J. Baumgardner, D. Pallamraju, S. Chakrabarti, U. P. Løvhaug, D. Lummerzheim, B. S. Lanchester and M. H. Rees, J. Geophys. Res. 109 (2004) A07305, doi:10.1029/2003JA010033. 6. D. Pallamraju, S. Chakrabarti, R. Doe and T. Pedersen, Geophys. Res. Lett. 31 (2004) L08807, doi:10.1029/2003GL019173. 7. D. Pallamraju, S. Chakrabarti and C. E. Valladares, Ann. Geophys. 22 (2004) 3241–3250. 8. D. Pallamraju and S. Chakrabarti, Geophys. Res. Lett. 32 (2005) L03S10, doi:10.1029/2004GL021417. 9. S. P. Zhang and G. G. Shepherd, Solar inﬂuence on the O(1D) dayglow emission rate: Global-scale measurements by WINDII on UARS, Geophys. Res. Lett. 31 (2004) L07804, doi:10.1029/2004GL019447. 10. D. Pallamraju, J. Baumgardner and S. Chakrabarti, Geophys. Res. Lett. 27 (2000) 1875–1878. 11. D. Pallamraju and S. Chakrabarti, J. Atmos. Sol-Terr. Phys. (2006), in press. 12. S. C. Solomon and V. Abreu, J. Geophys. Res. 94 (1989) 6817–6824. 13. O. Witasse, J. Lilensten, C. Lathillere and P.-L. Blelly, J. Geophys. Res. 104 (1999) 24639–24655. 14. P. B. Hays, D. W. Rusch, R. G. Roble and J. C. G. Walker, Rev. Geophys. 16 (1978) 225–232. 15. A. G. Burns, T. L. Killeen, W. Deng, G. R. Carignan and R. G. Roble, J. Geophys. Res. 100 (1995) 14673–14691. 16. M. D. Burrage, V. J. Abreu, N. Orsini, C. G. Fesen and R. G. Roble, J. Geophys. Res. 97 (1992) 4177–4187. 17. A. E. Hedin, P. Bauer, H. G. Mayr, G. R. Carignan, L. H. Brace, H. C. Brinton, A. D. Parks and D. T. Pelz, J. Geophys. Res. 82 (1977) 3183–3189. 18. H. G. Mayr, I. Harris and N. W. Spencer, Rev. Geophys. 16 (1978) 539–565. 19. G. W. Prolss, Rev. Geophys. 18 (1980) 183–202. 20. G. W. Prolss, J. Geophys. Res. 98 (1993) 5981–5991. 21. H. Fujiwara, S. Maeda, H. Fukunishi, T. J. Fuller-Rowell and D. S. Evans, J. Geophys. Res. 101 (1996) 225–239. 22. R. L. Balthazor and R. J. Moﬀet, Ann. Geophys. 15 (1997) 1048–1056. 23. D. J. Melendez-Alvira, D. G. Torr, P. G. Richards, W. R. Swift, M. R. Torr, T. Baldridge and H. Rassoul, J. Geophys. Res. 100 (1995) 7839–7853.

This page intentionally left blank

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch17

AURORAL EQUATORWARD BOUNDARY OBSERVED BY THE NOAA-17 SATELLITE L. XIE∗,§ , T. A. FRITZ† , Q.-G. ZONG† , Z. Y. PU∗ , ∗Institute

X. Z. ZHOU∗ and X. LI‡ of the Space Physics and Applied Technology, Peking University Beijing 100871, China †Center

‡Laboratory

for Space Physics, Boston University Boston 02215, USA

for Atmospheric and Space Physics, University of Colorado Boulder, CO, USA § [email protected]

Precipitating energetic electrons and protons measured by the total energy detecter and medium energy proton and electron detector instruments on the NOAA-17 satellite have been investigated. The equatorward boundary of the aurora, deﬁned by the cutoﬀ in the energy ﬂuxes of precipitating particles, is highly coincident with the last closed equipotential line (LCE), which closely represents the plasmapause location (PPL) when the plasmapause moves toward the Earth. This coincidence exists for both low-energy and high-energy (> 300 keV) ions and electrons. The correlation of the equatorward boundary of the aurora with the LCE suggests that the depth which plasmasheet particles can be transported into the inner magnetosphere is well determined by the PPL. The pitch angle scattering which leads to the precipitation of energetic particles is most eﬀective outside the PPL where strong chorus waves are generated by plasmasheet electrons. Further, the observations indicate that the movement of the auroral equatorward boundary and auroral activity are strongly related to the geomagnetic Kp index, which is determined by the solar wind conditions. During the period of intense geomagnetic activity, such as during the October–November 2003 magnetic storm, the auroral equatorward boundary can move to magnetic latitudes as low as 40–45◦ .

1. Introduction Auroral boundaries as a function of magnetic latitude and local time can be deﬁned by using various types of observation data including all-sky images,1 magnetograms,2 satellite images,3 and the intensity of precipitating particles.4 The most equatorward boundary of the auroral oval is the location where both ions and electrons are no longer observed to be

219

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

220

ch17

L. Xie et al.

precipitating (i.e., the transition between bounce trapping and strong pitchangle scattering). Linscott et al. ﬁrst noticed that the plasmapause coincided with the equatorward boundary of the diﬀuse aurora to within δL ∼ 0.25 in the duskside.5 Foster et al. concluded that the equatorward boundary of the trapped electron ﬂux agreed well with the whistler-deduced plasmapause at both dawn and dusk.6 The plasmapause is an important interaction region between the hot plasma sheet and the cool plasmasphere, and as such, it is a signiﬁcant source region of plasma wave generation in the inner magnetosphere. During geomagnetic storms, the ﬂux of energetic particles with energy of > 10 keV is signiﬁcantly enhanced and the precipitation of these highenergy particles, especially the MeV electrons, has been observed and investigated by Baker et al.7 and Blake et al.8 These observations indicate that high-energy electron precipitation frequently results from a strong scattering process by plasma waves, rather than from the weak diﬀusion of stably trapped electrons into the drift loss cone.9 However, the correlation of the auroral equatorward boundary with the precipitation of high-energy particles has not been clearly addressed yet. The study of the equatorward boundary of the auroral zone is useful for the qualitative understanding of the response of the aurora to magnetospheric dynamics. In this paper, the precipitation of auroral particles in a rather wide energy range (50 eV–1 MeV) for both electrons and ions measured on the NOAA-17 satellite has been systematically studied. Some features of the auroral equatorward boundary are reported here. The coincidence between the equatorward boundary of the precipitating auroral particles and the theoretical location of the plasmapause is found to be remarkably well correlated. The precipitation of energetic particles could be caused by pitch-angle scattering generated by wave–particle interactions near the plasmapause.

2. Instrumentation and Data Analysis NOAA-17 was launched into a nearly polar orbit with an altitude of about 850 km and an inclination of 98◦ . It has a period of 101.2 min and nearly covers the full range of L values four times during each orbit. The orbit of NOAA-17 is sun synchronous, lying in the 1000–2200 magnetic local time (MLT) plane. The present study examines the observations of electron and proton ﬂuxes using 16-point spectra at 8-s resolution from the

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch17

Auroral Equatorward Boundary Observed by the NOAA-17 Satellite

221

Total Energy Detector (TED) and the Medium Energy Proton and Electron Detector (MEPED) on board NOAA-17. Two independent measurements are made by TED, one looking outwards at 0◦ and the other 30◦ from the local vertical, which is well within the atmosphere loss cone at mid and highlatitudes. Therefore, the auroral particles within the loss cone were sampled by the TED detector. The energy range of the electrons and protons for TED is 50 eV–20 keV. The MEPED sensor also detects particles from two directions: one points toward the zenith to detect particles precipitating into the auroral ionosphere, and the other points at 90◦ to the zenith to detect particles whose mirror points lie just above atmosphere. The energy ranges for electrons are > 30 keV, > 100 keV, > 300 keV, and for protons are 30–80 keV, 80–240 keV, 240–800 keV, 800–2500 keV, and > 2500 keV. The data measured by the 0◦ MEPED, which looks toward the zenith, is used in this paper. The position of the plasmapause and Alfven layer were determined by theoretical formulations. The plasmapause location (PPL) can be identiﬁed as the last closed equipotential line (LCE) of a steady-state magnetospheric electric ﬁeld distribution.10,11 Therefore, we approximately regard the LCE as the theoretical position of plasmapause in this paper. The Alfven boundary is determined by the break between open and closed drift trajectories whose location is dependent on the energy of the particle. The theoretical position of Alfven layer is calculated by the (U, B, K) coordinate transformation introduced by Whipple.12

3. Observation The NOAA-17 data presented in this paper are L-sorted and plotted versus time. Figure 1 is a typical observation during the period of a high-speed solar wind stream, and contains data from both the MEPED and the TED detectors plotted in intensity versus time from September 10 to November 10, 2003. Panels 1–6 give a comprehensive view of the dynamics of highenergy ﬂux of electrons and protons. The data in these panels are obtained from the 0◦ MEPED. Panels 7 and 8 show the energy ﬂux (in units of mW/m2 ) of low-energy (50 eV–20 keV) electrons and protons measured by TED. The three bottom panels are the hourly Dst, solar wind velocity and 3-hr Kp index, respectively. Between September 10 and November 10, 2003, the Earth’s magnetosphere encountered two recurrent corotating interaction regions (CIRs) and a transient coronal mass ejection (CME) marked by dashed lines in Fig. 1.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

222

ch17

L. Xie et al.

Fig. 1. L-sorted data from September 10 to November 10, 2003. From top to bottom: proton ﬂuxes in three energy ranges (30–80 keV, 80–240 keV, and 240–800 keV), electron ﬂuxes in three energy ranges (30–100 keV, 100–300 keV, and > 300 keV), proton and electron energy ﬂuxs in the 50 eV–20 keV energy range, Dst, solar wind velocity and Kp index.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch17

Auroral Equatorward Boundary Observed by the NOAA-17 Satellite

223

An outstanding feature can be seen in Fig. 1 directly. A clear equatorward boundary can be seen in the observed ﬂux of electrons and protons in all energy ranges. Based on the deﬁnition of the auroral equatorward boundary, the boundaries of these precipitating particles can be considered as the approximate equatorward boundary of the auroral oval. These boundaries shift inward and outward in L-value in response to the changing values of Kp, Dst and solar wind velocity. At the same time the auroral oval also expands with increasing geomagnetic activity. This behavior is consistent with the observations reported from Ref. 13. The most remarkable feature in Fig. 1 is that precipitating particles of diﬀerent energies seem to have the same equatorward boundary. The superposed white curves in Fig. 1 represent the nightside PPL as a function of Kp index. For simplicity the radial distance to the seperatrix equipotential at midnight is taken as the position of plasmapause in this work.14 It can be seen that the stronger cross-tail electric ﬁeld pushes the plasmapause closer to the Earth, and peels oﬀ the outer layer of the plasmasphere. The distance of the plasmapause from the Earth decreases with increasing geomagnetic activity. At the distinct equatorward edge shown in Fig. 1, the clear drop in ﬂux is the result of the plasma sheet plasma being stopped from accessing the inner magnetosphere. Another noticeable feature can be determined further from Fig. 1. The PPL (or LCE) evidently coincides very well with the equatorward boundary of the auroral oval. The very good match indicates that not only low-energy plasma but also high-energy particles are well organized along the approximate theoretical location of the plasmapause. The observations from the 0◦ MEPED detector imply that the plasmapause can be seen as a boundary where ring current particles are scattered. The Alfven layer is another important boundary in the magnetosphere, and it varies with the energy of electrons and protons. Horwitz et al. suggested that the inner edge of the plasma sheet particles could be regarded as the Alfven layer at those energies due to the characteristic plasma sheet energy dispersion.15 The minimum locations of the Alfven layer of 30-kev protons and electrons are plotted in Fig. 2 as black curves. The superposed white curves denote the PPL based on the LCE. It is obvious that the auroral equatorward boundary and the inner edge of the plasma sheet match well with the PPL but not the Alfven layer. This implies that the plasmapause would be a more important boundary than we have known earlier. The isolated intense (−400 nT) magnetic storm on October 28, 2003 (the Halloween storm), which was driven by a large CME event, is represented

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

224

ch17

L. Xie et al.

Fig. 2. L-sorted data for the observed ﬂuxes of 30–80 keV protons and 30–100 keV electrons measured by 0◦ MEPED from September 10 to November 10, 2003.

in Fig. 2. The intense storm has a very clear signature: the boundary of the precipitation particles extends to lower L-shell than the calculated position of the plasmapause and to as low as L ∼ 2. From the invariant latitude (ILAT) data, the auroral equatorward boundary can be located at very low magnetic latitude (Φ ∼ 40–45◦ ). In such a case, the observations indicate that the real PPL could be eroded to lower L values, in agreement with the observations on IMAGE.16 Figure 3 shows the integral energy ﬂux of the precipitating protons and electrons in the northern hemisphere along the ILAT during the time period of Fig. 1. The integral energy ﬂux of electrons and protons can be used to represent the intensity of auroral particles. This ﬁgure shows a remarkable coincidence between the intensity of auroral particles and the Kp index. An examination of solar wind data from ACE (shown in the Fig. 1) for this period shows a correlation of the energy ﬂux enhancement with high solar wind streams and modulations of the CIR.

4. Discussion and Conclusion Previous studies of the auroral equatorward boundary were focused on lower energy (1–10 keV) particles. Those studies indicate that the equatorward boundary of the auroral oval coincides with both the ﬁeld line threading the PPL and the plasma sheet inner boundary.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch17

Auroral Equatorward Boundary Observed by the NOAA-17 Satellite

225

Fig. 3. Integral energy ﬂux of protons and electrons at 50 eV–20 keV along the auroral equatorward boundary in the northern hemisphere.

The observations reported here have shown that the PPL agrees very well with the auroral equatorward boundary of both low-energy and highenergy particles. Because of the changes in the characteristics of the plasma around the PPL, electromagnetic ﬁeld and wave change abruptly at the plasmapause. Therefore, the plasmapause tends to be the site of signiﬁcant wave–particle energy and momentum exchange, cross boundary energy transfer, and particle precipitation into the Earth’s ionosphere and atmosphere.17,18 A schematic in Fig. 4 displays the possible links between the auroral equatorward boundary and the PPL. The precipitation of energetic particles driven by wave–particle interactions around the plasmapause signiﬁcantly contributes to the formation of the auroral equatorward boundary. The abundant waves, i.e., EMIC, ULF, and VLF, near the plasmapause could scatter large pitch-angle particles into the loss cone and cause particle precipitation. Inside the plasmapause, wave activity decreases quickly

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

226

ch17

L. Xie et al.

Fig. 4. Schematic diagram for the formation of the auroral equatorward boundary and its relation with the plasmapause.

and wave–particle interactions become weak. As a result, the boundary of precipitating particles is formed along the plasmapause. Because the Alfven layer is deﬁned as a separatrix located between open and closed drift trajectories, it is not the site where abundant plasma waves are generated. The precipitation mechanism caused by corresponding wave– particle interactions is not eﬀective along the Alfven layer. Therefore, the correlation between the equatorward boundary of the precipitating particles and the Alfven layer is not prominent. It should be noted that the PPL in this paper is determined by the LCE of the Volland–Stern electric ﬁeld model. Here a minimum PPL based on the model by O’Brien et al.19 is used to compare with the auroral equatorward boundary. The red curves in Fig. 2 show the PPL based on an empirical model. The correlation of the auroral equatorward boundary with the PPL based on the empirical model is not as impressive as that with the PPL based on the LCE (white curves). The diﬀerence between the two PPL locations shows that during the geomagnetically active times, both these boundaries move inward. Thereafter, the LCE moves outward faster and to much higher L-shell because it responds to solar wind changes simultaneously, while the PPL based on an empirical model moves more slowly because plasmasphere reﬁlling takes much more time to ﬁll the large Lshells. The LCE has a ﬁner time resolution because it is calculated from solar wind input. The main results of this study can be summarized as follows. A remarkable correlation between the equatorward boundary of auroral particles and the PPL has been found for both low-energy and high-energy particles. However, the Alfven layer does not correlate with the equatorward

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch17

Auroral Equatorward Boundary Observed by the NOAA-17 Satellite

227

boundary of auroral particles as well. This suggests that the plasmapause could be regarded as an indicator of the auroral equatorward boundary, which can directly inﬂuence the pitch-angle distributions of energetic particles in the inner magnetosphere. The movement of the auroral equatorward boundary is strongly related to the Kp index and auroral activity and is modulated by the solar wind velocity. During periods of intense geomagnetic activity (such as the great storm driven by a large CME), the auroral equatorward boundary can shift to lower magnetic latitudes as low as 40–45◦ . The real plasmapause location erodes to lower L than the PPL based on LCE during times of intense storms. This is because the LCE saturates for large Kp.20 At times of extreme geomagnetic activity when Kp reachs a large value, the LCE no longer usefully describes the PPL.

Acknowledgments We are grateful to D. Evans for discussions of the NOAA satellite data and the particle measurements from the NOAA-17 satellite. This work was supported by NSFC grant (40504017) and Education Committee of Beijing (XK100010404).

References 1. Y. Feidstein and Y. Galperin, The auroral luminosity structure in the highlatitude upper atomopsphere: Its dynamic and relationship to the large-scale structure of the earth’s magnetosphere, Rev. Geophys. 23 (1985) 217–275. 2. T. Iijima and T. A. Potemra, Large-scale characteristics of the ﬁled-aligned current associated with substorms, J. Geophys. Res. 83 (1978) 599–615. 3. A. T. Y. Lui, C. D. Anger and S. I. Akasofu, The equatorward boundary of the diﬀusion auroral and auroral substorms as seen by the isis-2 auroral scanning photometer, J. Geophys. Res. 80 (1975) 3603–3614. 4. P. T. Newell, W. J. Burke, E. R. Sanchez, C. I. Meng, M. E. Greenspan and C. R. Clauer, The low-latitude boundary layer and boundary plasma sheet at low altitude: Prenoon precipitation regions and convection reversal boundary, J. Geophys. Res. 96 (1991) 21013–21023. 5. G. R. Linscott and M. W. J. Scourﬁeld, Observations of the plasmapause and diﬀuse aurora, Planetary and Space Science 24 (1976) 299–300. 6. J. C. Foster, C. G. Park, L. H. Brace, E. J. Maier, J. R. Burrows, J. H. Hoﬀman and J. H. Whitteker, Plasmapause signatures in the ionosphere and magnetosphere, J. Geophys. Res. 83 (1978) 1175–1182. 7. D. N. Baker, R. C. Anderson, R. D. Zwickl and J. A. Slavin, Average plasma and magnetic ﬁeld variations in the distant magnetotail associated with nearEarth substorm eﬀects, J. Geophys. Res. 92 (1987) 71–83.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

228

ch17

L. Xie et al.

8. J. Blake, U. S. Inan, M. Walt, T. F. Bell, J. Bortnik, D. L. Chenette and H. J. Christian, Lightning-induced energetic electron ﬂux enhancements in the drift loss cone, J. Geophys. Res. 106 (2001) 29733–29744. 9. R. Horne and R. M. Thorne, Relativistic electron acceleration and precipitationg during resonant interaction with whistler-mode chorus, Geophys. Res. Lett. 30 (2003) 34–37. 10. A. Nishida, Formation of a plasmapause or magnetospheric plasma knee by combined action of magnetospheric convection and plasma escape from the tail, J. Geophys. Res. 71 (1966) 5669. 11. R. A. Wolf, Magnetospheric conﬁguration, Introduction to Space Physics, eds. M. G. Klvelson and C. T. Russell (Cambridge University Press, Cambidge, 1995), pp. 288–329. 12. E. C. Whipple, /u,b,k/ coordinates — A natural system for studying magnetospheric convection, J. Geophys. Res. 83 (1978) 4318–4326. 13. M. S. Gussenhoven, N. Heinemann and D. A. Hardy, Systematics of the equatorward diﬀuse auroral boundary, J. Geophys. Res. 88 (1983) 5692–5708. 14. H. Korth, M. F. Thomsen, J. E. Borovsky and D. J. McComas, Plasma sheet access to geosynchronous orbit, J. Geophys. Res. 104 (1999) 25047–25062. 15. J. L. Horwitz, S. Menteer, J. Turnley, J. L. Burch, J. D. Winningham, C. R. Chappell, J. D. Craven, L. A. Frank and D. W. Slater, Plasma boundaries in the inner magnetosphere, J. Geophys. Res. 91 (1986) 8861–8882. 16. D. N. Baker, S. G. Kanekal, X. Li, S. P. Monk, J. Goldstein and J. L. Burch, An extreme distortion of the van allen belt arising from the ‘hallowe’en’ solar storm in 2003, Nature 432 (2004) 878–881. 17. K. R. Lorentzen, J. B. Blake, U. S. Inan and J. Bortnik, Observations of relativistic electron microbursts in association with vlf chorus, J. Geophys. Res. 106 (2001) 6017–6028. 18. D. Carpenter and J. Lemaire, The plasmasphere boundary layer, Ann. Geophys. 22 (2004) 4291–4298. 19. T. P. O’Brien and M. B. Moldwin, Empirical plasmapause models from magnetic indices, Geophys. Res. Lett. 30 (2003) 1–4. 20. N. C. Maynard and A. J. Chen, Isolated cold plasma regions: Observatrions and their relation to possible production mechenisms, J. Geophys. Res. 80 (1975) 1009–1108.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch18

AURORA-ASSOCIATED PHENOMENA AND THE ePOP MISSION LUDMILA M. KAGAN Department of Physics and Astronomy, the University of Western Ontario London ON, Canada N6A 3K7 [email protected]

The enhanced polar outﬂow probe (ePOP) is the scientiﬁc part of the Canadian Space Environment small satellite “Cassiope” to be launched in the last quarter of 2007. The mission aims to study plasma and neutral outﬂows in the topside polar ionosphere (300–1500 km altitude), wave generation and particle interaction associated with these outﬂows and their eﬀects on radio wave propagation. When compared to the energy involved in natural auroral phenomena, reaching sometimes up to hundreds of keV, the energy input into the ionosphere from ground-based powerful radio transmitters is much more modest, not exceeding 20 eV at the greatest. Even so, by transmitting powerful radiowaves into the ionosphere it is possible to reproduce some of the natural auroral phenomena, for example generation of artiﬁcial aurora,3–5 Alfven waves,6–8 plasma turbulence,9,10 and ion outﬂow8 from the auroral F region. Two data sets presented show the similarity of some natural and radiowave-induced ion outﬂows from the auroral F region, accompanied by a strong electron (but not ion) temperature increase. Interpretive models supported by observations are given for red hydroxyl5 and green atomic oxygen4 radiowave-induced airglow (named artiﬁcial auroras4 ) recently discovered in the E-region.

1. Introduction The enhanced Polar Outﬂow Probe, or ePOP, is the scientiﬁc part of a small Canadian hybrid satellite “Cassiope” to be launched in December 2007. The ePOP is aimed at studying the eﬀects of space weather on communication and navigation and Sun–Earth energy exchange. Auroral lights are one of the most important, and deﬁnitely the most spectacular, manifestation of both space weather and Solar-Terrestrial interactions. The ePOP orbit, being 300 km in perigee and reaching 1500 km in apogee, will provide a unique opportunity for studying the aurora and its associated phenomena, such as outﬂows of ions and neutrals from the Earth’s atmosphere, plasma turbulence and wave activity (Alfven waves in particular) both in situ and remotely (for more details on the ePOP mission and instrumentation, see Ref. 1). The main obstacle to understanding the physical processes behind 229

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

230

ch18

L. M. Kagan

the curtain of a colorful and very dynamical auroral display is the huge variety of sources. Those include, but are not limited to, precipitating particles, electric ﬁelds and currents, density and temperature perturbations, heat ﬂows and waves. A complementary approach to understanding auroral phenomena lies in producing and studying manmade analoges of the aurora-associated phenomena with some of their properties (e.g., plasma density) known using ground-based transmitters of high-power radiowaves.2

2. Type-2 Ion Outﬂows Some of these aurora-associated phenomena, as for example the frequentlyobserved ion outﬂow from the ionospheric F region that occurs over auroral arcs and is accompanied by a strong electron temperature increase and unchanged or almost unchanged ion temperature (the so-called type-2 ion outﬂows)11 (see Fig. 1(a)), have remained unexplained until very recently. Three important ﬁndings12 constitute signiﬁcant progress in understanding this phenomenon. They are: (1) the conclusion that a ﬁeld-aligned electric ﬁeld is indeed responsible for the observed electron temperature increases, based on the investigation of the electron energy balance (a key element of the solution of the problem); (2) association of the ﬁeld-aligned electric ﬁeld with the auroral arc itself and the shear Alfven waves, conﬁrmed by reanalysis of observations11; (3) the hypothesis that the presence of highfrequency turbulence slows the electrons down, which in turn enhances the ion motion. The scenario12 develops as follows. Shear Alfven waves, associated with a southward moving auroral arc, produce a ﬁeld-aligned electric ﬁeld due to electron inertia. In the ionospheric F region where Coulomb collisions become very strong this ﬁeld-aligned electric ﬁeld plays a double role. First, it generates plasma turbulence with a frequency of about twice that of the Coulomb frequency, which slows down electrons but does not aﬀect the ion’s upward motion. Second, this ﬁeld-aligned electric ﬁeld causes electron heating due to counterstreaming of electrons and ions, in which electrons are moving down and ions up. Modeling of the electron trapping by shear Alfven waves in the magnetosphere that results in electron precipitation producing a moving auroral arc13 nicely conﬁrms the above-mentioned scenario.12 We based our scenario for the natural type-2 ion outﬂow on the case study observed with the EISCAT UHF and VHF radars when an auroral arc was moving through the radar ﬁeld-of-view.11 This scenario seems to

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Aurora-Associated Phenomena and the ePOP Mission

ch18

231

Fig. 1. Two examples of type-2 ion outﬂow events observed with the EISCAT UHF radar in Tromso, Norway. (a) Under natural conditions (reproduced from Ref. 11) and (b) induced by high-power radiowaves (reproduced from Ref. 8). In both cases ion outﬂows were accompanied by a strong increase in electron temperature. However, in the case of the natural ionosphere the phenomenon was observed when an auroral arc was moving across the UHF radar ﬁeld-of-view. In the case of radiowave-induced ion outﬂows, the UHR radar was targeting the radiowave-highlighted volume and the ion outﬂow was observed during radiowave transmissions marked with bars along the time axis in Fig. 1(b).

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

232

ch18

L. M. Kagan

be relevant to the case in which high-power radiowaves induce ion outﬂow accompanied by electron temperature enhancement.8 Here the EISCAT UHF radar was targeting the ionospheric volume highlighted by high-power radiowaves. Ion outﬂows accompanied by electron temperature enhancements were observed during transmissions when high-power radiowaves were reﬂected from the F2 region. I reproduce the two above-mentioned data sets for natural and radiowave-induced ion outﬂows in Figs. 1(a) and 1(b), respectively. One can easily see that in both cases the ion outﬂow was accompanied by electron temperature enhancements. The elevated electron temperature suggests a radiowave-generated ﬁeld-aligned electric ﬁeld. Since during the ionosphere modiﬁcation experiment shown in panel b, induced Alfven waves’ auroral activations were observed as well8 this electric ﬁeld is most probably associated with generated Alfven waves that in collisional plasma have a ﬁeld-aligned component (similar to the natural event shown in panel a).

3. Artiﬁcial Aurora Among the radiowave-induced phenomena the so-called artiﬁcial aurora is worthy of special attention due to its importance for ionospheric studies. Until now radiowave-induced aurora has been observed from the ground only. The Fast Auroral Imager (FAI) on board the ePOP mission is designed for fast imaging of auroral and airglow emissions at 630.0 ± 1 nm and in the near infrared (NIR) at 650–850 nm, including those above the high-power radiowave transmitters, both remotely and in situ. Both ﬁlters will register total emission in the two spectral ranges above. Since even the 1-nm vicinity of 630.0 nm (not mentioning the NIR) includes several emission lines: red-line emission from atomic oxygen and Fraunhofer band-emission from molecular oxygen (both observed in natural aurora) as well as natural airglow from hydroxyl Meinel vibrational bands, we are going to support FAI measurements with ground-based optical observations allowing calibration of the FAI data. With the ePOP Canadian mission we also expect signiﬁcant progress in understanding both natural and radiowaveinduced airglow. Artiﬁcial airglow occurs when, due to the interaction of a powerful electromagnetic wave with the ionospheric plasma, electrons acquire enough energy for collisional excitation of neutral species. Since the process of excitation is similar to that producing natural auroral lights, we named this induced airglow “artiﬁcial aurora”.4 Artiﬁcial aurora has been developed

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Aurora-Associated Phenomena and the ePOP Mission

ch18

233

into a very eﬀective tool for studying the structure and dynamics of the atmosphere’s ionized and neutral components. F -region artiﬁcial airglow has been observed for more than 30 years14 and has been well studied and understood.15 It mainly comes from relaxation of excited atomic oxygen from the O(1 D) to the ground state, emitting a photon with a wavelength of 630.03 nm (the excitation energy is 1.97 eV). Observed intensities reached several hundred Rayleighs. On rare occasions this emission may be accompanied by signiﬁcantly less intense (compared to the 630.0-nm airglow) green-line emission up to 60 R, resulting from relaxation of the O(1 S) state of atomic oxygen and emission of a photon with a wavelength of 557.7 nm (the excitation energy is 4.19 eV). F -region airglow has been used for determining plasma drifts, neutral winds, diﬀusion coefﬁcients, and collisional quenching times and tracking the source region of super-energetic electrons (see, e.g., Ref. 3). Artiﬁcial E-region emissions4,5,16 were observed at altitudes between 80 and 125 km and are closely related to sporadic ionization, which is best known for its possible eﬀects on broadcasting, radio communications and navigation. The green line artiﬁcial airglow from atomic oxygen was ﬁrst observed in January 1998 in Arecibo.4,16 There were only two more campaigns that succeeded in generating E-region artiﬁcial aurora at 557.7-nm. One was in August 2004 at the Sura facility in Russia17 and the other in March 2004 at the HAARP facility in Alaska.18 In all three cases the powerful radiowaves were reﬂected by a dense sporadic ionization with a critical frequency more than the transmitted frequency. The induced airglow is due to collisional excitation of atomic oxygen by super-energetic electrons resulting from nonlinear interaction of high-power radiowaves with ionospheric plasma.19 The observed intensities reached up to 100 R at Arecibo, up to 20 R at Sura and about 4 kR at HAARP. Because (1) the energetic electron’s free path at these altitudes does not exceed several hundred meters and (2) the O(1 S) excitation and emission together take only a fraction of 1 s, the induced airglow presents a real-time footprint of sporadic ionization clouds, thus giving the ﬁrst method of measuring the horizontal structure of sporadic ionization.4 I give a schematic illustration of this scenario on the left-hand side of Fig. 2, supporting it with optical data from the 1998 campaign in Arecibo (right-hand side of Fig. 2). Similar to the case in the cartoon, the thin strip of 557.7-nm emission footprints the sporadic ionization cloud at 120 km altitude (lower image) that splits the radiowave transmitter beam in two, each of which induces 630.0-nm emission in the F region (seen as two

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

234

ch18

L. M. Kagan

Fig. 2. Sporadic ionization clouds with critical frequency exceeding the transmitted frequency reﬂect radiowaves impinging on the Es clouds (seen as 557.7-nm airglow shown in light gray in the left panel) while radiowaves propagating through the holes in sporadic ionization are reﬂected in the F region, resulting in 630.0-nm emission from atomic oxygen in the F region (dark gray clouds in the left panel). The two images in the right-hand side are taken with 630.0 (upper right) and 557.7-nm (lower right) ﬁlters on January 22 LT, 1998 in Arecibo. The 557.7-nm cloud in the lower image is located in between two 630.0-nm clouds in the upper images in correspondence with the schematic diagram to the left.

airglow clouds in the upper image). The 557.7-nm artiﬁcial aurora may be used for studying the structure and dynamics of sporadic ionization clouds, tracking the source region of energetic electrons, and determining neutral winds. During the heating-optical-backscatter campaign at the Sura facility, Russia in August 2004, for the ﬁrst time in heating experiments we induced airglow that is currently identiﬁed as vibrationally excited OH emissions.5 The key in these observations was that the light detected in a 2 nm wide ﬁlter centered on 630 nm was seen 1–2 s after launching radiowaves. This short response time ruled out the 630 nm emission from atomic oxygen since it has a rise time of about 30 s.15 In this case sporadic ionization was weak (the Es critical frequency did not exceed 1.5 MHz) and located at much

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Aurora-Associated Phenomena and the ePOP Mission

ch18

235

lower altitudes of 80–85 km. Thus the power necessary for chemical reactions to proceed and to produce vibration-rotationally excited OH Meinel bands was due to radiowave focusing by the lower-altitude underdense ionization clouds. The backscatter signature from this sporadic ionization appears as summer mesosphere echoes (MSE), which have been found to be in good correlation with a visible scattering layer called noctilucent or mesosphere clouds.20 The true mechanism of production of vibrationally excited OH is still to be identiﬁed, although it is clearly related to the presence of water molecules. Note that the occurrence of both sporadic ionization and MSE has a maximum during the summer months of June–August. On the left-hand side of Fig. 3, I draw a schematic illustration of the physical mechanism of radiowave-induced hydroxyl airglow. Radiowave focusing by underdense ionization clouds results in excitation of OH

Fig. 3. Sporadic ionization clouds with critical frequency less than the transmitted frequency change the radiowave’s path inside the Es clouds, resulting in radiowave focusing above holes in the sporadic ionization. This focusing gives enough energy to vibrationally excite OH (shown by the two lower spots in the diagram to the left and seen in the upper image to the right) due to nonlinear interactions of HF waves with lower E region plasma. Some radiowaves propagating through the holes in the sporadic ionization are reﬂected in the F region, resulting in 630.0-nm emission from atomic oxygen in the F region (clouds below the reﬂection level in the diagram to the left are observed in the lower image to the right). The two images on the right-hand side are two subsequent images taken with a 630.0 ± 1 nm ﬁlter on August 16, 2004 at the Sura facility in Russia. The O(1 D) airglow cloud in the image at 00:44:57 LT develops on top of the rightmost hydroxyl red airglow cloud in the image at 00:44:03 LT.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

236

ch18

L. M. Kagan

vibration-rotational bands in the E region above the holes in the sporadic ionization almost instantaneously after launching radiowaves (upper image to the right). The next image (the lower one on the right-hand side of Fig. 3) shows that where the hole in the sporadic ionization layer was big enough to let enough radiowaves through and to be reﬂected from the F region, the 630.0-nm emission from O(1 D) state of atomic oxygen developed in the F region. Since the camera registers total emission along the line-of-sight, the oxygen red line emission was seen on top of the OH airglow cloud. Besides the information on structure and dynamics of sporadic ionization clouds the radiowave-induced OH aurora may give information on local temperature and water content in the atmosphere. 4. Conclusions and Outlook This article has described the radiowave-induced analogues of auroraassociated phenomena, giving more detailed examples for two of them, type-2 ion outﬂow from the ionospheric F region and artiﬁcial aurora. While being very useful for understanding natural phenomena, their radiowaveinduced analogues have puzzles of their own, as for example the mechanism of generation of accelerated (super-energetic) electrons that excite neutral species, producing optical emissions. The Cassiope/ePOP mission payload and orbit provides fantastic possibilities for remote and in situ (F region while in perigee) studies of both natural and radiowave-induced auroral phenomena. The ePOP/FAI orbiting will allow observations of the induced airglow over all operating high-power radiowave transmitters. We are going to calculate the Cassiope passes over the heating facilities for ﬁnding the best observational conditions and experimental setup. The radiowave-induced airglow will give information on plasma drifts, neutral winds, diﬀusion coeﬃcients, and collisional quenching times in the F region; horizontal structure, dynamics of E-region sporadic ionization clouds (plasma irregularities), the airglow-source region, energy of superenergetic electrons, local atmosphere temperature and perhaps water vapor content at 80–90 km in the atmosphere highlighted by a heater beam. We expect this information to complement ePOP observations and the two together to signiﬁcantly improve our understanding of magnetosphere– ionosphere–thermosphere coupling. Acknowledgments I gratefully acknowledge funding and technical support from the Canadian Space Agency for the CASSIOPE/ePOP project, and from the Natural

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Aurora-Associated Phenomena and the ePOP Mission

ch18

237

Sciences and Engineering Research Council of Canada Collaborative Research Opportunities Program.

References 1. A. W. Yau, H. G. James and W. Liu, Adv. Space Res. (2006), in press. 2. L. M. Kagan, Proc. the 5th European Heating Seminar, Finland, March 17–20, 22–24, 1997. 3. P. A. Bernhardt et al., J. Geophys. Res. 105 (2000) 10657. 4. L. M. Kagan et al., Phys. Rev. Lett. 85 (2000) 218. 5. L. M. Kagan et al., Phys. Rev. Lett. 94 (2005) 095004. 6. Yu. M. Yampolski, Radiophys. Quantum. Electron. 32, 6 (1989). 7. Yu. M. Yampolski et al., J. Geophys. Res. 102, A4 (1997) 7461. 8. N. F. Blagoveshchenskaya et al., Ann. Geophys. 23, 1 (2005) 81–100. 9. L. M. Kagan, Proc. the 5th European Heating Seminar, Finland, March 17–20, 19–21, 1997. 10. N. F. Blagoveshchenskaya et al., Artiﬁcial ﬁeld-aligned irregularities in the nightside auroral ionosphere, Adv. Space Res. 37 (2006), in press. 11. J. E. Wahlund et al., J. Geophys. Res. 97 (1992) 3019–3037. 12. L. M. Kagan and J.-P. St.-Maurice, Ann. Geophys. 23, 1 (2005) 13–24. 13. C. E. J. Watt, R. Rankin, I. J. Rae and D. M. Wright, Self-consistent electron acceleration due to inertial Alfven wave pulses, J. Geophys. Res. 110 (2005) A10507, doi:10.1029/2005JA011007. 14. J. C. Haslett and L. R. Megill, Radio Sci. 9 (1974) 1005. 15. G. P. Mantas and H. C. Carlson, J. Geophys. Res. 101 (1996) 195. 16. F. T. Djuth, P. A. Bernhardt, C. A. Tepley, J. A. Gardner, M. C. Kelley, A. L. Broadfoot, L. M. Kagan, M. P. Sulzer, J. H. Elder, C. Selcher, B. Isham, C. Brown and H. C. Carlson, Geophys. Res. Lett. 26 (1999) 1557. 17. N. V. Bakhmet’eva, V. V. Belikovich, L. M. Kagan, A. A. Poniatov, A. V. Tolmacheva, M. C. Kelley and M. J. Nicolls, New results of the lower ionosphere studies by the method of resonant backscatter of radiowaves from artiﬁcial ionospheric irregularities, Radiophysics and Quantum Electronics 48, 9 (2005), in press. 18. T. R. Pedersen and E. A. Gerken, Nature 433 (2005) 498–500. 19. D. L. Newman, M. V. Goldman, F. T. Djuth and P. A. Bernhardt, Phys. Space Plasmas 15 (1998) 259–264. 20. J. Y. Cho and M. C. Kelley, Rev. Geophys. 31 (1993) 243–253.

This page intentionally left blank

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch19

ON THE IMPORTANCE OF THE CROSS-BODY APPROACH IN PLANETARY AERONOMY MARINA GALAND∗,†,¶ , ANIL BHARDWAJ‡,§ and SUPRIYA CHAKRABARTI† ∗Space and Atmospheric Physics Group, Imperial College Prince Consort Road, London, SW7 2BW, UK †Center

for Space Physics, Boston University, 725 Commonwealth Avenue Boston, MA 02215, USA ‡Space

Physics Laboratory, Vikram Sarabhai Space Centre Trivandrum 695022, India ¶[email protected]

Cross-disciplinary and cross-body approaches can be applied to study universal processes occurring in the heliosphere. Magnetospheric, interplanetary, and heliospheric plasmas, all of which are low density plasmas, host similar processes. A cross-disciplinary approach is thus of great relevance for a universal understanding of processes occurring within these various plasmas. On the other hand, the upper atmosphere of planets and moons are a highly collisional medium acting diﬀerently compared to a collisionless plasma. Therefore, the comparative study between solar system bodies hosting atmospheres under diﬀerent settings is a more suitable approach for assessing universal processes in aeronomy. For the past several years the aeronomy community has undertaken many initiatives in comparative studies of solar system atmospheres. We highlight the maturity of this ﬁeld and illustrate its relevance by applying the comparative approach to key scientiﬁc topics. We would like to encourage aeronomers interested in comparative studies to consider participating to International Heliophysical Year (IHY) focused activities. More information on the comparative initiative can be found at the IHY website (http://ihy.gsfc.nasa.gov/) as well as at: http://www.bu.edu/csp/uv/ cp-aeronomy/aeronomy-sol-sys.html.

1. Introduction A cross-body approach is suitable for assessing universal processes in aeronomy, an interdisciplinary ﬁeld aimed to study the upper atmospheric regions (of Earth, planet, moon, comet) where ionization and photodissociation processes play a role.1 In other words, such a discipline focuses on the physics and chemistry occurring in an upper atmosphere, divided — for §Earlier at NASA Marshall Space Flight Center, NSSTC, Huntsville, AL 35805, USA, as NRC Senior Research Associate.

239

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

240

ch19

M. Galand, A. Bhardwaj and S. Chakrabarti

dense atmosphere — into a neutral part (mesosphere, thermosphere, exosphere) and an ionized part (ionosphere).2 Thus, essentially the ﬁeld of “planetary aeronomy” deals with the composition, dynamics, and energetics of the thermosphere–ionosphere system of a planetary body that has an atmosphere. The upper atmospheres encountered in the solar system are extremely diverse3 due to diﬀerences in atmospheric constituents and densities, distance from the Sun, topology and magnitude of the magnetic environment, gravity, rotation rate, and gravity wave forcing, among others. This diversity in setting makes comparative aeronomy an exciting and enriching ﬁeld of research.4 The solar system bodies to whom this approach applies include those with a thick, permanent atmosphere (Venus, Earth, Mars, the giant planets Jupiter, Saturn, Uranus, and Neptune, and Saturn’s moon Titan) and those with a thin or transient atmosphere or with a coma — deﬁned as a gas envelop not restricted by gravity — (Mercury, Galilean moons (Io, Europa, Ganymede, Callisto), Triton (Neptune’s moon), Saturn’s inner, icy moons, Pluto, and comets). Comparative aeronomy is becoming increasingly fruitful as spacecraft mission and Earth-based datasets are assimilated and interpreted using state-of-the-art multidimensional methods. The International Heliophysical Year (IHY) in 2007 which is fully encompassing the comparative aeronomy eﬀort5 emphasizes the timeliness of this initiative and provides a platform for such studies. In this paper we ﬁrst review key scientiﬁc topics in comparative aeronomy. Next, we present some of the actions initiated from the community since 2000 for promoting comparative aeronomy and for sharing scientiﬁc ﬁndings. Finally, we conclude on the importance of the comparative approach for aeronomy and discuss future directions for this discipline.

2. Key Scientiﬁc Topics in Comparative Aeronomy A cross-body comparison applied to aeronomical quantities highlights the range and complexity of solar system environments4,6–8 and illustrates how comparative aeronomy contributes to a true synthesis of the solar system. For instance, while the observed average exospheric temperature of the upper atmosphere at Mercury, Earth, Jupiter, and Uranus, decreases linearly as a function of distance from the Sun, at Venus, Mars, Saturn, and Neptune, the observed values are lower than those predicted from this linear trend (see Fig. 1). The planets closer to the Sun are not always hotter. Local conditions, such as composition and heating and cooling sources,

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch19

On the Importance of the Cross-Body Approach in Planetary Aeronomy

241

Fig. 1. Comparison of the neutral exospheric temperature of the upper atmospheres of the planets (adapted from Ref. 4, Fig. 1, p. 2). The circles represent the average, observed values and the vertical bars, the diurnal, seasonal, and solar cycle range estimates: Mercury,9 Venus, Earth, and Mars,10 outer planets,11 Pluto.12 The triangles represent the modeled values derived with solar heating alone.11 The dashed line is a ﬁt to the average exospheric temperature observed at Mercury, Earth, Jupiter, and Uranus.

need to be taken into account. A list of key topics and related outstanding questions is given in Table 1. This list, far from being exhaustive, provides an illustration of scientiﬁc challenges in comparative aeronomy. A model is a key tool to assess the contribution of various processes individually to a given physical quantity. For instance, thermospheric heating associated with solar irradiance has been modeled for assessing the contribution of the solar source to the exospheric temperature of upper atmospheres. Such an approach has shown that solar heating is not suﬃcient for explaining the observed exospheric temperature at the giant planets — as illustrated in Fig. 1 — yielding an energy crisis which is still under debate.11 It is also very valuable to adapt aeronomical models developed for Earth to other Solar System bodies. Such a challenge allows us to test a given model under diﬀerent conditions and to assess its robustness regarding the included physical processes. As an illustration, terrestrial thermosphere (/ionosphere) general circulation models (GCM)22,23 have been adapted to: Venus and Mars,10,24–26 Jupiter,27,28 Titan and Triton,29 and Saturn.30 Such an experience is also crucial for application to exoplanets and will provide the only constrain to the future observations of the atmosphere of other worlds.31,32

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

242

ch19

M. Galand, A. Bhardwaj and S. Chakrabarti

Table 1. Key scientiﬁc topics of relevance in comparative aeronomy and related outstanding questions. Key topic

Sample outstanding questions

Ionospheric structures and dynamics

Is a part or the entire ionosphere under photochemical equilibrium? What are the sources of night time ionization, if any?7,8,13

Energy budget

Is solar heating the major heating source of a thermosphere? What are the other signiﬁcant heating and cooling sources? What is the resultant thermal structure?11

Magnetosphere– Ionosphere coupling

How important is external control (solar wind) versus internal control (planetary rotation, satellite) of the auroral activity? What is the role of the ionosphere as a source of magnetospheric plasma and what is its inﬂuence on magnetospheric processes?16–19

Laboratory experiments

What are the rates and cross-sections of atomic and molecular processes which are critically needed and which require to be (re-) determined?20,21

The atmospheric modeling tools are critically dependent on the knowledge of cross section and reaction rates, quantities which can be derived from laboratory experiments20 or from the analysis of atmospheric emissions.33 Other critical physical quantities which drive aeronomical models include solar irradiance, which is still largely unknown, especially in the soft X-ray range responsible for the ionization of the lower part of upper atmospheres. Such uncertainties largely limit the modeling eﬀort. It is crucial to improve the assessment of the solar irradiance spectral proﬁles in soft X-rays and extreme ultraviolet (EUV), as recently discussed at a comparative aeronomy special session at 2005 Spring AGU.34–37 It is important to assess not only the typical solar spectrum in the soft X-ray region (0.1– few keV), but also its large variability ranging from a factor 10 between minimum and maximum solar conditions, to 10 000 during strong solar ﬂares. This is particularly critical for the modeling of the solar soft X-ray scattered from planetary atmospheres.38,39 Complementary to the modeling eﬀort, observations of the upper atmosphere of solar system objects have been very productive during the past 25 years. In situ aeronomical measurements are scarce. Beside the Earth, for which the dataset is acquired onboard rockets or low-altitude satellites, in situ aeronomical measurements include the on-going, fascinating case of Titan. Its upper atmosphere has been directly probed through ﬂybys of the Cassini spacecraft, which arrived in July 2004 at Saturn. A

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch19

On the Importance of the Cross-Body Approach in Planetary Aeronomy

243

total of 44 ﬂybys will be achieved over the four-year nominal mission. The most common observations of planetary upper atmospheres are performed remotely, including optical remote-sensing as well as occultations of a star and of a radio source. Analysis of occultation data provides neutral and electron density proﬁles along the ﬁeld of view. Optical remote-sensing from infrared11,14 and visible40 to ultraviolet41,42 and X-ray43 yields the assessment of atmospheric temperature, drift, and composition, chemical processes, and plasma interactions. Their origin is very diverse, ranging from airglow44–46 to aurora14,15 to reﬂected or ﬂuorescent sunlight.38,39,52 Figure 2 shows a sample of X-ray emissions observed at various solar system bodies. The origin is auroral at Earth through bremsstrahlung continuum produced by precipitating energetic electrons (Fig. 2(a)), in

Fig. 2. X-ray emissions observed from a few objects in the solar system: (a) at Earth by Polar/PIXIE (credit to: N. Ostgaard and NASA); (b) at Jupiter by Chandra47 ; (c) at comet Linear by Chandra48 ; (d) at Moon by Rosat43,49 ; (e) at rings of Saturn by Chandra50 ; (f) at Mars by Chandra.51

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

244

ch19

M. Galand, A. Bhardwaj and S. Chakrabarti

the polar regions of Jupiter through charge exchange of highly ionized heavy ion precipitation (Fig. 2(b)) and at comet Linear through charge exchange of highly ionized heavy ions from the solar wind interacting with the cometary neutral gas (Fig. 2(c)). In the equatorial region of Jupiter the emission is dominantly through resonant and ﬂuorescent scattering of solar X-rays.52 At the moon, the dayside lunar soft X-rays are ﬂuoresced sunlight scattered by elements present in lunar regolith (Fig. 2(d)), while the X-ray emission recently detected from the rings of Saturn is the result of ﬂuorescent scattering of solar X-rays from oxygen atoms in the H2 O icy ring material (Fig. 2(e)). At Mars the emission are due to ﬂuorescent scattering of solar X-rays from O, C, and N present in the atmospheric gases (CO2 , O2 , N2 ) in the Martian upper atmosphere (Fig. 2(f)). However, an X-ray halo around the planet extending up to three Mars radii with an origin similar to that of comets has also been observed.53 Analysis of auroral emissions has been used to assess magnetic ﬁeld conﬁguration, to trace plasma interactions, to identify the energetic particle sources and atmospheric constituents.14,15 For instance, the analysis of auroral X-ray emission identiﬁes the type and characteristics — thus the origin — of the energetic particle population (Figs. 2(a)–2(c)) and is used to derive the time variability of this energy source. Multispectral analysis provides further constraints for identifying processes occurring at a given body.47,53,54 Due to the large variety in magnetic environment, energetic particle source, atmospheric species, and chemistry, the comparative approach applied to atmospheric emissions, such as airglow and aurora, is of great relevance. It yields a synthetic — thus, more critical — view of interactions taking place at diﬀerent solar system bodies, including solar– wind–magnetosphere–ionosphere coupling.16,17,19,55,56

3. Past and Present Actions from the Community Year 2000 marked a renewed interest in comparative studies of solar system atmospheres, as attested by the organization of a Yosemite Conference, a workshop at the Coupling, Energetic, and Dynamics of Atmospheric Regions (CEDAR) meeting, and the creation of a discussion group dedicated to this subject. These initiatives were followed by additional special sessions at international meetings and by the publications of community paper,57 special issue,58 and monograph.4 More detailed information regarding current actions, special sessions, and publications can be found

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch19

On the Importance of the Cross-Body Approach in Planetary Aeronomy

245

at the comparative aeronomy website: http://www.bu.edu/csp/uv/cpaeronomy/aeronomy-sol-sys.html. The information posted does not pretend to be exhaustive. Any member of the community is free to send input, sign up for the mailing list as well as consult past newsletters. It should be noted that at the initiative of the Space Physics and Aeronomy (SPA) president, Prof. M. Mendillo, comparative aeronomy is a full part of Spring AGU meetings, since 2004, through the organization of dedicated special sessions.

4. Future Directions in Comparative Aeronomy Aeronomy is an ideal ﬁeld for comparative studies of solar system bodies due to the diversity and complexity of their environments. Comparative aeronomy contributes to a better understanding and true synthesis of our solar system.4,16,58 It provides a challenge to models and opens new horizons which allow aeronomers to be more critical toward their one-body research. The comparative aeronomy community has been strengthening its eﬀort over the past ﬁve years through the organization of special sessions and through topical publications. The IHY initiative would beneﬁt from the momentum of this community, while at the same time the IHY constitutes a true platform for comparative aeronomy. Among the ﬁve universal process science themes identiﬁed (http://ihy.gsfc.nasa.gov/science themes.shtml), theme two focusing on energy transfer and coupling processes encompasses aeronomical topics. The IHY initiative is providing a web infrastructure to carry out focused activities and facilitate international collaborations. The on-going Cassini and Mars orbiting missions and the upcoming Venus Express are going to enrich the aeronomical database, providing dataset more comprehensive than single ﬂybys can oﬀer and requires more complex models to describe the observed environment. At the same time, comprehensive, 3D models have matured enough the past years for carrying out quantitative cross-body comparison of physical and chemical processes. For easy access to laboratory measurements supporting the modeling and data analysis eﬀort, a central database of cross-sections and reaction rates — including recommended values, references, and error bars — would be of great relevance to planetary aeronomers.21 Comparative aeronomy is now ready to move from a discovery, assessment phase — based on the identiﬁcation of main diﬀerences between solar system bodies — to a more mature phase addressing key outstanding issues (Table 1) quantitatively. Comparative aeronomy also constitutes an

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

246

ch19

M. Galand, A. Bhardwaj and S. Chakrabarti

excellent step towards the aeronomy of exoplanets,31,59 which has already begun32,60 and is expected to play an increasing role in the future.

References 1. S. Chapman, General Assembly of the International Union of Geodesy and Geophysics (1954). 2. M. H. Rees, Physics and Chemistry of the Upper Atmosphere (Cambridge University Press, Cambridge, 1989). 3. I. C. F. M¨ uller-Wodarg, Advances in Astronomy: From the Big Bang to the Solar System, ed. M. Thompson (Imperial College Press, London, UK, 2005), pp. 331–352. 4. M. Mendillo, A. Nagy and J. H. Waite (eds.), Atmospheres in the Solar System: Comparative Aeronomy, Geophys. Monogr. Ser., Vol. 130 (American Geophysical Union, Washington, DC, 2002). 5. N. U. Crooker, EOS Transactions 85 (2004) 351. 6. D. F. Strobel, in Ref. 4, p. 7. 7. M. Mendillo, S. Smith, J. Wroten and H. Rishbeth, J. Geophys. Res. 108 (2003) SIA 6-1, doi:10.1029/2003JA009961. 8. T. Majeed, J. H. Waite, S. W. Bougher, R. V. Yelle, G. R. Gladstone, J. C. McConnell and A. Bhardwaj, Advanc. Space Res. 33 (2004) 197. 9. R. M. Killen, A. Potter, A. Fitzsimmons and T. H. Morgan, Planet. Space Sci. 47 (1999) 1449. 10. S. W. Bougher, R. G. Roble and T. Fuller-Rowell, in Ref. 4, p. 261. 11. R. Yelle and S. Miller, Jupiter : Planet, Satellites, & Magnetosphere, eds. F. Bagenal, W. McKinnon and T. Dowling (Cambridge University Press, 2004), p. 185. 12. J. L. Elliot et al., Nature 424 (2003) 165. 13. A. F. Nagy and T. E. Cravens, in Ref. 4, p. 39. 14. A. Bhardwaj and G. R. Gladstone, Review of Geophysics 38 (2000) 295. 15. M. Galand and S. Chakrabarti, in Ref. 4, p. 55. 16. R. Prang´e et al., Nature 432 (2004) 78. 17. M. G. Kivelson, Space Science Reviews 116 (2005) 299. 18. M. Blanc, R. Kallenbach and N. V. Erkaev, Space Science Reviews 116 (2005) 227. 19. M. Galand, R. Prang´e, D. Lummerzheim, A. Bhardwaj, S. Chakrabarti and R. Gladstone, Eos Trans. AGU, Fall Meeting Suppl., 2005AGUFMSA52A04G (2005). 20. D. L. Huestis, in Ref. 4, p. 245. 21. D. L. Huestis, Eos Trans. AGU, Spring Meeting Suppl., 2005AGUSMS A34A.05H (2005). 22. R. Roble, E. C. Ridley, A. D. Richmond and R. E. Dickinson, Geophys. Res. Lett. 15 (1988) 1325.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch19

On the Importance of the Cross-Body Approach in Planetary Aeronomy

247

23. T. J. Fuller-Rowell et al., Handbook of Ionospheric Models, ed. R. W. Schunk (Scientiﬁc Committee on Solar-Terrestrial Physics, 1996), p. 217. 24. S. W. Bougher, S. Engel, D. P. Hinson and J. R. Murphy, J. Geophys. Res. 109 (2004) E03010. 25. Y. Moudden and J. C. McConnell, J. Geophys. Res. 110 (2005) E04001. 26. M. Angelats i Coll, F. Forget, M. A. L´ opez-Valverde and F. Gonz´ alezGalindo, Geophys. Res. Lett. 32 (2005) L04201. 27. N. Achilleos, S. Miller, J. Tennyson, A. D. Aylward, I. M¨ uller-Wodarg and D. Rees, J. Geophys. Res. 103 (1998) 20089. 28. S. Bougher, J. H. Waite, T. Majeed and G. R. Gladstone, J. Geophys. Res. 110 (2005) E04008. 29. I. C. F. M¨ uller-Wodarg, in Ref. 4, p. 307. 30. I. C. F. M¨ uller-Wodarg, M. Mendillo, R. V. Yelle and A. D. Aylward, Icarus 180 (2006) 147. 31. W. A. Traub and K. W. Jucks, in Ref. 4, p. 369. 32. R. V. Yelle, Icarus 170 (2004) 167. 33. N. Ivchenko, M. Galand, B. S. Lanchester, M. H. Rees, D. Lummerzheim, I. Furniss and J. Fordham, Geophys. Res. Lett. 31 (2004) CiteID L10807, doi:10.1029/2003GL019313. 34. M. Galand and R. V. Yelle, Eos Trans. AGU, Spring Meeting Suppl., 2005AGUSM.SA31A.05G (2005). 35. P. G. Richards, Eos Trans. AGU, Spring Meeting Suppl., 2005AGUSM. SA31A.01G (2005). 36. S. C. Solomon, L. Qian and S. M. Bailey, Eos Trans. AGU, Spring Meeting Suppl., 2005AGUSM.SA32A.04G (2005). 37. T. N. Woods, F. G. Eparvier and W. McClintock, Eos Trans. AGU, Spring Meeting Suppl., 2005AGUSM.SA31A.02G (2005). 38. A. Bhardwaj et al., Eos Trans. AGU, Spring Meeting Suppl., 2005AGUSM. P43A.05B (2005). 39. T. E. Cravens et al., Eos Trans. AGU, Spring Meeting Suppl., 2005AGUSM. P44A.05C (2005). 40. M. Mendillo, F. Roesler, C. Gardner and M. Sulzer, in Ref. 4, p. 329. 41. P. D. Feldman, E. B. Burgh, S. T. Durrance and A. F. Davidsen, Astrophys. J. 538 (2000) 395. 42. J. T. Clarke and L. Paxton, in Ref. 4, p. 339. 43. A. Bhardwaj et al., Proceedings of the 36th ESLAB Symposium, 3–8 June 2002, ESTEC, Noordwijk, The Netherlands, eds. B. Foing and B. Battrick, (ESA Publications Division, Noordwijk, 2002), p. 215. 44. D. T. Hall, P. D. Feldman, M. A. McGrath and D. F. Strobel, Astrophys. J. 499 (1998) 475. 45. T. G. Slanger and B. C. Wolven, in Ref. 4, p. 77. 46. M. H. Stevens, in Ref. 4, p. 319. 47. R. F. Elsner et al., J. Geophys. Res. 110 (2005) CiteID A01207. 48. C. M. Lisse et al., Bull. Am. Astron. Soc./Div. Planet. Sci. Meet. 32, #40.02 (2000) 1070.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

248

ch19

M. Galand, A. Bhardwaj and S. Chakrabarti

49. J. H. M. M. Schmitt, S. L. Snowden, B. Aschenbach, G. Hasinger, E. Pfeﬀermann, P. Predehl and J. Trumper, Nature 349 (1991) 583. 50. A. Bhardwaj, R. F. Elsner, J. H. Waite, G. R. Gladstone, T. E. Cravens and P. G. Ford, Astrophys. J. 627 (2005) L73. 51. K. Dennerl, Astron. Astrophys. 394 (2002) 1119. 52. A. Bhardwaj et al., Geophys. Res. Lett. 32 (2005) CiteID L03S08. 53. J. H. Waite et al., Adv. Space Res. 26 (2000) 1453. 54. J. T. Clarke et al., Jupiter : Planet, Satellites, & Magnetosphere, eds. F. Bagenal, W. McKinnon and T. Dowling (Cambridge University Press, Cambridge, 2004), p. 639. 55. B. H. Mauk, B. J. Anderson and R. M. Thorne, in Ref. 4, p. 97. 56. J. H. Waite and D. Lummerzheim, in Ref. 4, p. 115. 57. D. L. Huestis et al., Planetary atmospheres, planetary decadal study community white paper, Solar System Exploration: Priorities for 2003-2013 (First Decadal Study) (2001), http://www-mpl.sri.com/decadal/patm-home.html. 58. E. Kallio and H. Shinagawa (eds.), Planetary Atmospheres, Ionospheres, and Plasmas Interactions, Advances in Space Res., Vol. 33, Issue 2 (Elsevier Science, Ltd., 2004). 59. H. U. Frey and D. Lummerzheim, in Ref. 4, p. 381. 60. J. J. Fortney, M. S. Marley, K. Loddars, D. Saumon and R. Freedman, Astrophys. J. 627 (2005) L69.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch20

SOLAR TERRESTRIAL AND PLANETARY SCIENCE MISSIONS IN ASIA−OCEANIA: OPPORTUNITIES FOR COLLABORATIVE RESEARCH ANDREW W. YAU Department of Physics and Astronomy, University of Calgary Calgary, Alberta, Canada T2N1N4 [email protected] ANIL BHARDWAJ Space Physics Laboratory, Vikram Sarabhai Space Centre, ISRO Trivandrum 695022, India anil [email protected] IVER H. CAIRNS School of Physics, University of Sydney, Sydney New South Wales 2006, Australia [email protected] C. Z. CHENG National Space Organization, Hsin-Chu City, Taiwan [email protected] WING H. IP Faculty of Science, National Central University, Chungli, Taiwan [email protected] YASUMASA KASABA∗,‡ , KYOUNG W. MIN†,§ , MASATO NAKAMURA∗,¶ and YOSHIFUMI SAITO∗, Aerospace and Exploration Agency, Institute of Space and Astronautical Science, Sagamihara, Kanagawa, Japan 229-8510

∗ Japan

†Korea

Advanced Institute of Science and Technology (KAIST) Daejeon 305-701, Korea ‡[email protected] §[email protected] ¶[email protected] [email protected]

Geoscientists in the Asia–Oceanian Region and elsewhere increasingly share and articulate a strong vision in advancing geosciences through closer cooperation; a vision that provided the impetus for the open Forum on collaborative research opportunities in solar terrestrial and planetary science (ST–PS) 249

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

250

ch20

A. W. Yau et al. missions. This report captures the information presented and discussed at the Forum, including a synopsis of current, planned, and proposed ST–PS missions led by various Asia–Oceanian countries, in the context of the potential opportunities that these missions oﬀer from a geoscience perspective.

1. Introduction As solar terrestrial (ST) and planetary science (PS) research in the Asia– Oceania Region becomes increasingly international in its scope, more countries in the Region are embarking on national space programs of small scientiﬁc satellites, sounding rockets, balloons, and ground-based observations. Geoscientists in the Region and elsewhere increasingly share and articulate a strong vision in advancing geoscience through closer cooperation. Indeed, it is this collective vision that has guided the founding of the Asia Oceania Geosciences Society (AOGS), and provided the impetus for the open Forum on collaborative opportunities in solar terrestrial and planetary research in the Second AOGS Meeting in June 2005 in Singapore. The primary aim of this open Forum was to provide Asia–Oceanian geoscientists an opportunity to discuss and share information on their respective national ST–PS and related space programs, as a step towards promoting and eventually developing at the grass-root level collaborative satellite missions and related research programs among geoscientists in the Region. The aim of this report is to capture and synthesize the information presented and discussed at the Forum. Its intent is to serve as an information resource for researchers in the Region, by drawing attention to areas of shared or common scientiﬁc interests, and potential collaborative spacebased mission and research opportunities. Section 2 provides a synopsis of current, planned, and proposed ST–PS missions, as well as a few astronomy missions of possible interests to the ST–PS community, led by various countries represented in the forum; the synopsis of missions led by China is based on information presented elsewhere in the AOGS meeting. Section 3 describes some of the current opportunities and near-future plans raised at the Forum. Section 4 summarizes the characteristics and development status of the respective missions in the context of potential collaborative opportunities that these missions oﬀer from a geoscience perspective. 2. Solar Terrestrial and Planetary Science Programs in Asia−Oceania A number of Asia–Oceanian countries currently engage in scientiﬁc satellite missions and ground-based programs in ST–PS research, including

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Solar Terrestrial and Planetary Science Missions in Asia–Oceania

ch20

251

Australia, China, India, Japan, Korea, and Taiwan. A few other countries in the Region such as Indonesia, New Zealand, and Thailand are engaged in planning similar satellite programs. 2.1. Australia The Australian Federation Satellite (FedSat) is a three-axis stabilized micro-satellite that weighs under 60 kg and measures (0.5 m)3 in size. FedSat is dedicated to solar terrestrial studies in the topside ionosphere. It is in a DMSP-like Sun-synchronous orbit (98.7◦ inclination, 10:30 LT, and about 800 km altitude). It was launched in mid-December 2002, and has operated since 2003. It carries a ﬂuxgate magnetometer, a star camera, and a diﬀerential Global Position System (GPS) receiver. FedSat is a good example of what international collaboration can accomplish: it was launched by the National Space Development Agency of Japan (NASDA, now part of the Japan Aerospace Exploration Agency, JAXA) and it uses a Canadian micro-reaction wheel to achieve the required attitude stability for its magnetic ﬁeld measurements on a micro-satellite. The magnetometer is a collaborative experiment between the Cooperative Research Center for Satellite Systems (CRCSS) in Australia and the University of California at Los Angeles (UCLA). It has a measurement precision of 0.2 nT, and samples data at up to 100 samples/s, making it possible to study small scale ﬁeld-aligned currents (FAC) down to 100-m scale. The GPS receiver is a Black Jack receiver contributed by the Jet Propulsion Laboratory (JPL) and is designed for ionospheric total electron content (TEC) measurements using the L1 and L2 signals from occulting GPS satellites. The Australian space physics community is starting its ﬁrst decadal plan process, seeking to improve on FedSat and leverage recent work on the Tasman International Geospace Environment Radars (TIGER) SuperDARN radars in Tasmania and New Zealand, optical interferometry, rocketry, and radio receivers, with individual collaborations on international missions such as NASA’s STEREO and Canada’s ORBITALS missions. A new space mission, ideally joint with Asia–Oceanian partners, is one of the desired outcomes. 2.2. China China launched the pair of Double Star (TC-1 and -2) spacecraft in December 2003 and July 2004, respectively. The scientiﬁc objectives of the joint Chinese–ESA Double Star mission are to study magnetic reconnection, the trigger mechanism for magnetic storms, and particle acceleration, diﬀusion, injection, and outﬂow processes during the storms.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

252

ch20

A. W. Yau et al.

Each Double Star spacecraft carries a full complement of plasma, ﬁeld, and wave instruments of CLUSTER heritage. TC-1 is in a near-equatorial (570 × 78 970 km, 28.5◦ inclination) orbit, while TC-2 is in a polar (700 × 39 000 km, 90◦ inclination) orbit. Both spacecraft carry a ﬂuxgate magnetometer, a plasma electron and current experiment, high-energy electron and ion detectors, and a heavy ion detector. In addition, TC-1 carries an active spacecraft potential control device, a hot ion analyzer, and a digital wave processor, while TC-2 carries a neutral atom imager, a low-energy ion detector, and a low-frequency electromagnetic wave detector. A number of future space missions for magnetospheric and heliospheric studies are now in the planning phase. These include a Chinese–French collaborative project, SMESE, to explore solar eruptions by a small satellite. The scientiﬁc payload of SMESE includes a Lyman-α heliograph, an extreme ultraviolet (EUV) heliograph, a Lyman-α coronagraph, a full-sun infrared (IR) telescope, an X-ray spectrograph, and a γ-ray spectrograph, to study solar ﬂares and coronal mass ejection (CME) events. The Phase-A study for SMESE will be carried out in 2006, with the launch targeted in 2010. Other space missions awaiting governmental approval are the KuaFu and the Space Wind and Storms Explorations (SWISE) projects (see Ref. 1). The Kua-Fu project will consist of three spacecraft: Kua-Fu A will hover in the vicinity of the L1 Lagrangian point between the Earth and the Sun and will be used to observe solar ﬂares, CMEs and their interplanetary propagation. The Kua-Fu B1 and B2 spacecraft will observe the geoeﬀectiveness of solar activities at highly elliptical polar orbits (1.8 × 7.0 RE geocentric, 90◦ inclination). The proposed launch of the Kua-Fu mission is between 2012 and 2015. The SWISE project will be made up of three satellites in diﬀerent orbits. SWISE-1 will be in a low-altitude, elliptical polar orbit, and will address important issues in ionosphere–thermosphere coupling. SWISE-2 will operate between 700 km and 7.5 RE , to study the responses of the inner magnetosphere to solar events and interplanetary disturbances. SWISE-3 will cover the region of 2–22 RE in the solar wind and the magnetopause, to study magnetospheric interaction with the solar wind. The proposed launch date is between 2010 and 2012. China has also begun its activities in the ﬁeld of planetary exploration. The lunar program will consist of a lunar orbiter (Chang’e-1), which is to be launched in 2007, a rover mission, and a sample return mission within a time interval of about 15 years. Some initial steps have also been made in the planning of a Mars mission.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Solar Terrestrial and Planetary Science Missions in Asia–Oceania

ch20

253

2.3. India In India, solar terrestrial and planetary science studies often go side-by-side with astronomy and astrophysics studies. Therefore, many of the currently operating missions and those being developed fall in all of these disciplines. India’s ﬁrst space-borne solar astronomy experiment, the Solar X-ray Spectrometer (SOXS), and its companion solar-terrestrial experiment, the Coherent Radio Beacon Experiment (CRABEX), were launched onboard the Indian geostationary satellite GSAT-2 in May 2003. The SOXS consists of two instruments, the Low Energy Detector (4–60 keV) and the HighEnergy Detector (25 keV–10 MeV), which provide full-disk integrated solar X-ray emission in the 4 keV–10 MeV range. Its main objective is to study solar ﬂares at high spectral and temporal resolution. CRABEX is aimed at the study of the spatial structure, dynamic and temporal variations of the ionosphere and several aspects of equatorial electrodynamics. Chardrayaan-1 is India’s ﬁrst mission to the Moon. It is devoted to simultaneous high-resolution mineralogical, elemental, and photo-geological mapping of the lunar surface in the visible, IR, X-ray and low energy γ-ray wavelength regions. Chardrayaan-1 will be launched in late 2007 using an indigenous spacecraft and the Polar Satellite Launch Vehicle (PSLV) of the Indian Space Research Organisation (ISRO), and will be placed in a 100-km lunar polar orbit. It will have an operational life of two years. A truly international mission, it will carry four experiments from Germany, Sweden, UK, and Bulgaria, respectively, and two experiments from USA, and involve participation from Japan, Switzerland, and a few other nations. Its complement of instruments include the Terrain Mapping stereo Camera (TMC), Hyper Spectral Imager (HYSI), Lunar Laser Ranging Instrument (LLRI), Low Energy (0.5–10 keV) X-ray spectrometer (LEX), Solar X-ray Monitor (SXM) in the 2–10 keV energy range, High Energy (10– 200 keV) X-ray/γ-ray spectrometer (HEX), Miniature Synthetic Aperture Radar (Mini-SAR), Near-Infrared Spectrometer (SIR-2), Sub-KeV Atom Reﬂecting Analyzer (SARA), Solar Wind monitor (SWIM), Moon Mineral Mapper (MMM), and Radiation Dose Monitor (RADOM). In addition to the above missions, which fall in ST and PS as well as astrophysics and astronomy, two dedicated Indian astronomy satellites of possible ST–PS interest are currently under development: ASTROSAT is a multiwavelength astronomy mission in low-Earth (∼ 650 km) circular equatorial (∼ 8◦ inclination) orbit in which the instrument complement will cover the UV (1000–3000 ˚ A) and soft and hard X-ray regimes (0.3–8 keV; 2–80 keV),

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

254

ch20

A. W. Yau et al.

and TAUVEX is an Indo-Israeli UV imaging experiment to image large parts of the sky in the wavelength region between 140 and 320 nm. In the future, India is planning other planetary and ST missions that will utilize the expertise available within the country as well as harness international collaborations.

2.4. Japan In Japan, solar system science missions for ST and PS research have been developed under the Institute of Space and Astronautical Science (ISAS) since the launch of the ﬁrst Japanese satellite “Ohsumi” in 1970. In October 2003, JAXA was formed with the uniﬁcation of ISAS, National Aerospace Laboratory (NAL), and NASDA. Currently, solar system science missions constitute one of the major activities in the new agency. In March 2005, a document entitled “JAXA Vision — JAXA 2025” was released as the proposal for future Japanese space activities. Nakamura et al. (2005) discuss the details of proposed solar system science missions included in the Vision (see Ref. 2). Japan has had many successful collaborative missions and exchanges of payloads with USA, Canada, Russia, ESA, and European countries, including collaborations in the mission to Comet Halley and the International Solar Terrestrial Physics (ISTP) project. On the ST side, the Akebono mission is a collaborative mission with Canada, and GEOTAIL is a joint mission with USA. These two missions are currently in their 17 and 14 year of continuous operation, respectively. The science objective of Akebono is to study auroral acceleration processes. GEOTAIL is the ﬁrst spacecraft to make detailed observations of the magnetotail, and in its current orbit is a unique spacecraft for surveying the equatorial magnetosphere. In conjunction with ground-based observations and those from other spacecraft, observations from Akebono and GEOTAIL have signiﬁcantly deepened our knowledge in auroral and magnetotail processes and magnetosphere–ionosphere coupling, from thermal-energy auroral ion outﬂow, to near-tail reconnection, to cold dense tail ion beams in the tail lobe, to name but a few examples. Summary data are available on the JAXA web and the ISTP Coordinated Data Analysis Web (CDAWeb). Any requests for detailed collaborations are welcomed by the projects and their Principal Investigators (PIs). In August 2005, the micro-satellite Reimei (INDEX) was launched. The scientiﬁc objectives of Reimei are to study fast and ﬁne-scale auroral structures. Reimei will undertake comprehensive collaborative studies in

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Solar Terrestrial and Planetary Science Missions in Asia–Oceania

ch20

255

conjunction with international ground observatories and other satellites. In addition to Reimei, a small satellite, “SmartSat”, is currently under development, for space weather research and operational forecasting experiment, by the National Institute of Information and Communications Technology (NICT) for launch in 2008. “SOLAR-B”, a joint mission with NASA, ESA, and the UK Particle Physics and Astronomy Research Council (PPARC) based on the successful Yohkoh mission, will be launched in 2006. SOLAR-B will, for the ﬁrst time, provide quantitative measurements of the full vector magnetic ﬁeld on suﬃciently small scales to resolve elemental ﬂux tubes. Two new ST missions are also being proposed: SCOPE, a mission to study the cross-scale coupling in the plasma universe, and Energization and Radiation in Geospace (ERG), a mission to investigate energization and radiation in geospace. The SCOPE mission will have a number of full-scale post-Magnetospheric Multi Scale/SMART (MMS/SMART) formation-ﬂying satellites, aimed at probing processes of a variety of spatial and temporal scales, from the scales of electron and ion gyro-radius, inertial lengths, to MHD scales. The ERG is planned as a small satellite mission, and aims to study the acceleration and loss mechanisms of relativistic particles in the radiation belts during magnetic storms. This project is exploring possible collaborations with the Region, and with other projects in the International Living with a Star (ILWS) program (http://ilws.gsfc.nasa.gov). Payload-level collaborations are also planned with the NASA MMS-SMART mission and Radiation Belt Storm Probes (RBSP) satellites program. On the planetary side, the Hayabusa (MUSES-C) spacecraft rendezvoused asteroid Itokawa successfully in November 2005. Its goal is to undertake the ﬁrst sample return from asteroids: to encounter an asteroid, rendezvous, and dock with it, and then take samples from the asteroid’s surface and return them to Earth. An international research team will be formed to analyze the returned samples in 2007. Two lunar missions are under development: SELENE and LUNAR-A. SELENE will be launched in 2007, and will study the origin, evolution, and environment of the Moon. LUNAR-A is a lunar penetrator mission. These two missions are the follow on to Japan’s ﬁrst lunar orbiter, Hiten, which was launched in 1990, and will be the basis of its future international lunar exploration program. Following the unfortunate mishap (failure of Mars-orbit insertion) of Nozomi (1998–2003), a collaborative mission with Canada, Sweden, Germany, USA, and France, JAXA is targeting its next two planetary

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

256

ch20

A. W. Yau et al.

orbiters at the terrestrial planets Venus and Mercury. PLANET-C, or Venus Climate Orbiter (VCO), is scheduled for launch in 2010 to Venus, and will focus on the exploration of Venus atmospheric circulation. BepiColombo is a joint ESA–JAXA mission to Mercury, and is scheduled for launch in 2012. BepiColombo consists of two spacecraft: the Mercury Magnetospheric Orbiter (MMO) and the Mercury Planetary Orbiter (MPO). JAXA will provide the MMO, which will survey the magnetic ﬁeld, magnetosphere, and exosphere of Mercury, as well as the inner Heliosphere. Instruments for both spacecraft will be provided by international teams. Data from both will be open to the scientiﬁc community, and the Japanese community will welcome participation from other communities in the Region at the grassroot level. In the future, the Japanese community expects to have increased collaboration with other communities in the Region, similar to its current collaborations with the US, Canada, and European countries. It is envisioned that discussions at the grass-root level initiated in the AOGS will evolve into discussions at the oﬃcial level and to the establishment of more concrete frameworks for collaborations.

2.5. Korea Korea presently has two satellite development programs: KOMPSAT and STSAT. The Korea Multipurpose Satellite (KOMPSAT) program entails the development of remote-sensing small satellites in the 500–1000 kg class, while the Science and Technology Satellite (STSAT) program is directed at the development of micro-satellites in the 100–150 kg class, for the purpose of satellite technology development and basic space science. The ﬁrst STSAT, STSAT-1, is Korea’s fourth micro-satellite developed at the Korea Advanced Institute of Science and Technology (KAIST), and is therefore also known as KITSAT-4. It was launched in September 2003. It weighs 137 kg, measures 60 cm × 60 cm × 80 cm, and is in a 690-km Sunsynchronous orbit. STSAT-1 is a dual, astrophysics and auroral physics micro-satellite, and it carries an UV imaging spectrograph and a suite of energetic auroral particle detectors. The UV imaging spectrograph is designed to study the far UV spectroscopy of hot and glowing gas in the Milky Way galaxy that has been heated by supernova blast waves and colliding interstellar clouds, when the spectrograph is viewing upward, and to study auroral UV emissions, in conjunction with the particle detectors, when it is viewing downward.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Solar Terrestrial and Planetary Science Missions in Asia–Oceania

ch20

257

STSAT-2 is planned for launch in 2007. It will carry a radiometer for meteorological purposes. A Langmuir probe and a dosimeter will also be onboard to monitor the space environment from an elliptical polar orbit (80◦ inclination, 300 km perigee and 1300 km apogee) STSAT-3 is planned for 2010; it will be a three-axis stabilized micro-satellite, and its mission purpose is to be determined in a feasibility study in 2006.

2.6. Taiwan The space program of Taiwan’s National Space Organization (NSPO) is in transition, from its ﬁrst 15-year program (1991–2006) to its second 15year plan (2004–2018). During the ﬁrst 15-year period, NSPO established the necessary infrastructure and capability, such as a space-qualiﬁed satellite integration and test facility, satellite tracking and receiving stations, and command center, etc. NSPO also carried out three satellite missions: FORMOSAT-1, 2 (formerly ROCSAT-1 and 2), and 3. FORMOSAT-1 was launched in January 1999 into a low-inclination, circular orbit, and carried two instruments: the Ionospheric Plasma Electrodynamics Instrument (IPEI), which is an ion drift meter for ionospheric plasma electrodynamics studies, and an ocean color imager in the 440–870 nm wavelength range. A major science accomplishment of IPEI is the observation of large ion density holes in the equatorial region during magnetic storms. The FORMOSAT-1 mission ended in June 2004. FORMOSAT-2, which weighs about 750 kg, was launched in May, 2004 into a Sun-synchronous orbit (99.14◦ inclination, 891-km altitude). It carries a Remote Sensing Instrument (RSI), a color (ﬁve-wavelength) imager and a black-and-white imager, and an Imager of Sprite and Upper Atmosphere Lightning (ISUAL) that uses six wavelength bands in the 100– 900 nm range. The RSI has a 2-m resolution and it monitors terrestrial and marine environment and resources in real time with applications mainly focused on agriculture, forestry, land use, natural disaster assessment, environment monitoring, as well as academic research, and public education. The ISUAL observes airglow, aurora, and transient luminous events (TLE) such as sprites, which are natural upward lightning discharge phenomena toward the ionosphere from the tropopause. The ISUAL is a joint research program by NSPO, University of California at Berkeley, National Cheng Kung University in Taiwan, and Tohoku University in Japan. The ﬁrst sprite image was obtained by ISUAL in July 4, 2004, and is the ﬁrst observation of sprites from space. Since then ISUAL has made a survey of the

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

258

ch20

A. W. Yau et al.

global distribution of more than 1000 TLEs, which occur mainly in the low latitude regions between −50◦ and 50◦ . The FORMOSAT-3 mission consists of a constellation of six identical micro-satellites, and is scheduled for launch in March 2006. It is a collaborative mission between NSPO and the University Corporation for Atmospheric Research (UCAR), and is also called COSMIC in the USA. The satellite design life is ﬁve years. Each satellite has a disk shape of 103 cm diameter and 16 cm thickness, and weighs about 70 kg. All six satellites are to be launched by a single launch vehicle into a temporary parking orbit of 72◦ inclination and 500 km altitude. It will take about 13 months for all six satellites to maneuver into their respective orbit planes, which will be nominally at 24◦ apart in ascending node. Each satellite will have a circular polar orbit at 72◦ inclination and 700–800 km altitude, and the orbit period will be about 100 min. Each FORMOSAT-3 satellite carries three science instruments: a Black Jack space-borne GPS receiver, which measures the amplitude and phase of GPS signals, a Tri-Band Beacon (TBB) transmitter, which emits three coherent radio signals at 150, 400, and 1066.7 MHz, and a Tiny Ionospheric Photometer (TIP), which measures photon emission at 135.6 nm wavelength. The GPS receivers will be employed for GPS limb occultation sensing of the atmosphere (vertical proﬁles of air pressure, temperature, and water vapor) for meteorological and climate research as well as for weather forecasting. The received GPS signals will also be used to perform geodetic research of the Earth’s gravity ﬁeld. Combined with data from TIP and ground TBB receivers, global three-dimensional (3D) distribution of ionospheric electron density and scintillation can be obtained for space weather monitoring and modeling. The FORMOSAT-3 satellites should provide at least 2500 vertical proﬁles of near-real time atmospheric and ionospheric data with almost uniform global distribution per day. The data will be employed for the International Polar Year (IPY) campaign for studying environment in the polar region, where ground-based observations are limited. They will also be used for the ILWS campaign for studying space weather related science. As NSPO enters its second 15-year plan, a key emphasis will be to establish the building and development capability for satellite bus and scientiﬁc payloads in Taiwan. The NSPO is developing a master space science plan. A Low-Earth-Orbit (LEO) satellite, ARGO, is planned, for launch in late 2008. It will weigh about 450 kg and have an orbit of 98◦ inclination and 620 km altitude. Besides carrying a remote sensing camera, ARGO will have scientiﬁc instruments and the science goal will be to study the physics of

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Solar Terrestrial and Planetary Science Missions in Asia–Oceania

ch20

259

magnetosphere–ionosphere coupling, space weather monitoring, and electromagnetic wave phenomena. One to two scientiﬁc micro-satellites (about 50 kg in weight) are also planned for launch in 2008–2009. In summary, NSPO plans to launch a series of scientiﬁc satellites during the second 15year plan period and will actively pursue international collaboration on scientiﬁc satellite missions.

2.7. International collaboration International collaboration between research groups in Asia–Oceania and those elsewhere is a signiﬁcant part of many satellite missions and research programs in the preceding subsections, and in many other non-Asia– Oceania-led missions. The latter include a number of collaborative opportunities in current or planned ESA, NASA, Canadian, and other national programs. In the case of the ESA program, these opportunities include Ulysses, SOHO, Cluster II, and Double Star in the ST program, and Cassini Huygens, Mars Express, BepiColombo 2012, and Venus Express 2005 in the PS program; as noted in Sec. 2, Double Star and BepiColombo are joint ESA/China and ESA/Japan missions, respectively. In the case of the NASA program, these collaborative opportunities include the Solar Terrestrial Probes, which includes STEREO (April 2006 launch), MMS (2013 launch), and the Geo-Electrodynamic Connections (GEC) mission, and the Explorer Program, which includes THEMIS, TWINS, and the Interstellar Boundary (IBEX) mission. In the case of the Canadian program, these collaborative opportunities include the CASSIOPE Enhanced Polar Outﬂow Probe (e-POP) and Atmospheric Composition Explorer (ACE) small satellites, the Canadian Geospace Monitoring (CGSM) program, and the proposed Outer Radiation Belts Injection, Transport and Loss Satellite (ORBITALS), Ravens, and Chinook missions. The e-POP is a polar orbiter for the study of ionospheric ion outﬂow and radio wave propagation, and ACE for the study of ozone depletion using Fourier transform spectroscopy and aerosol extinction using solar occultation. The ground CGSM network comprises of an array of all sky imagers, magnetometer chains, solar radio telescopes, SuperDARN radars, and digital ionosondes, and an integrated facility of data assimilation and modeling. Japan is an international partner in e-POP. In addition to the above programs, there exist opportunities associated with the current national programs of Sweden, Italy, and the United Kingdom, including the Swedish ODIN and SMART-1 missions.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

260

ch20

A. W. Yau et al.

3. Current Opportunities and Near-Future Plans Several opportunities currently exist for collaborative small or microsatellite solar terrestrial and planetary science missions among Asia– Oceania countries. First, Ariane-Space is presently oﬀering a piggyback launch of up to a 100 kg spacecraft for approximately one million euros, with up to eight spacecraft possibly being launched at one time. Second, ESA is expected to shortly issue a “Call for Missions”. This will provide a chance for AOGS members to propose both missions and speciﬁc instruments in collaboration with teams from European Union countries. Third, ISRO has in the past launched several small satellites (∼ 50– 150 kg) for other countries and agencies. Riding as a piggyback, the launch of such satellites is relatively cheap. The AOGS members and all Asia– Oceania countries and space agencies can use this mode of launch to send their small missions to space in polar or geosynchronous orbits. Moreover, the launch capabilities of ISRO can also be utilized for launching larger satellite-based mission with 1000–2000 kg in near-Earth polar or geosynchronous orbits. A possible funding source for such collaborative missions is the Association of South East Asian Nations (ASEAN). Other possible sources include the Asia–Paciﬁc Regional Space Agency Forum and the United Nations Economic and Social Commission for Asia and the Paciﬁc (UNESCAP; http://www.unescap.org/icstd/SPACE/index.asp). An attractive idea is that the AOGS could ask ASEAN for funds for a pan-Asia–Oceania mission. With ASEAN funding, instruments contributed by multiple nations, and a cheap launch possible from Ariane, ISRO, and perhaps other space programs, it may prove possible to have an aﬀordable joint mission for many groups in the AOGS region. The Forum agreed that a Working Group should be set up to discuss pan-AOGS missions and develop ways to collaborate and work together. This will link research groups and nations, leading to new ideas and collaborations and perhaps a joint space mission. The Working Group will also organize a website, with links to national agencies and space committees.

4. Summary and Conclusion Table 1 lists the current or planned ST–PS satellite missions in Asia– Oceania that are identiﬁed as potential opportunities for collaboration. This table summarizes the science objectives, characteristics, and current status

URL/Mission

Australia www.crcss.csiro.au FedSat China

India

Launch 2002

Orbit

Status

Sun-sync, 800 km

Ionosphere — FAC, TEC, communications

O

Equatorial, 1.1 × 13.3 Re, 28.5◦ Polar, 1.1 × 7.1 Re, 90◦

Magnetic reconnection, storm mechanism, particle acceleration

O

Solar eruptions Solar ﬂare, CME, and their eﬀects on geospace

P P

Response of inner magnetosphere to solar and interplanetary events, and ionosphere– thermosphere coupling

P

ST, a, b

Solar X-rays Ionospheric tomography

O O

ST ST

www.cnsa.gov.cn Double Star TC-1

2003/12

Double Star TC-2

2004/07

SMESE Kua-Fu

(2010) (2012+)

SWISE

(2010+)

To be determined Kua-Fu A: L1 Lagrangrian; Kua-Fu B1, B2: polar, 1.8 × 7.0 Re, 90◦ SWISE-1: polar; SWISE-2: 700 km × 7.5 Re; SWISE-3: 2 × 22 Re, detailed orbits to be determined

www.isro.org SOXS on GSAT-2 CRABEX on GSAT-2

2003/05 2003/05

Geostationary, 6.6 Re Geostationary, 6.6 Re

International

Notes ST, µ

ESA

ST, a ST, a

France Belgium, Canada, UK

ST, a ST, a

ch20

Science objectives

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Country

Solar Terrestrial and Planetary Science Missions in Asia–Oceania

Table 1. Characteristics and status of ST/PS satellite missions presented in the open forum and identiﬁed as potential opportunities for collaboration in Asia–Oceania.

261

Country

Japan

URL/Mission

Launch

Chardrayaan-1

(2007)

www.isas.jaxa.jp Akebono 1989/02 1992/07

Reimei

2005/08

PS

Polar, 275 × 10 500 km, 75◦ Equatorial, 8 × 210 Re, −7◦ ; 9 × 30 Re (c) Sun-sync, 610 km

Auroral acceleration

O

Canada

ST

Magnetotail structure and dynamics Small-scale auroral structure Space weather and forecast experiment Solar magnetic ﬁeld etc. by opt/X/EUV Multi-scale reconnection, shock Radiation belt relativistic particles Asteroid rendezvous, sample return Lunar origin, evolution, environment

O

USA

ST

(2006)

Sun-sync

SCOPE

(2015)

ERG

(N/A)

Equatorial, 2.1 × 30 Re (p) Equatorial, 1.04 × 6.6 Re (p) Asteroid Itokawa Lunar polar 100 km (main s/c)

O

ST, µ

D

ST

L

USA, UK

ST

P

ST, c

P

ST, µ

O

PS, d

D

PS/ST

ch20

Sweden, UK, Germany, USA, Japan, Bulgaria, Switzerland

Solar-B

(2007)

Notes

D

N/A

SELENE

International

Mineralogical, elemental, and photo-geological mapping of lunar surface; solar wind, and radiation

(2008)

2003/05

Status

100 km lunar polar orbit

SmartSat

HAYABUSA

Science objectives

A. W. Yau et al.

Geotail

Orbit

(Continued )

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

262

Table 1.

Korea

Taiwan

URL/Mission

Launch

Orbit

Planet-C

(2010) (2012)

Lunar-A

(N/A)

Venus climate and circulation Mercury B-ﬁeld and magnetosphere Lunar interior seismology

D

BepiColombo/MMO

Venus 300 km × 10 Rv, 172◦ Mercury 400 × 11 800 km Lunar surface: near and far side

D

PS, f

Sun-sync, 690 km

Far UV in Milky Way; auroral UV Remote sensing; space environment To be determined

O

ST, µ, g

D

ST, µ

P

ST, µ

satrec.kaist.ac.kr STSAT-1

2003/09

STSAT-2

(2007)

STSAT-3

(2010)

Polar, 300 × 1500 km, 80◦ (p) Polar (p)

2004/05

Sun-sync, 890 km, 99◦

www.nspo.org.tw FORMOSAT-2 FORMOSAT-3 ARGO

(2006)

Polar, 600 km, 90◦

(2008/09)

Sun-sync, 620 km

Science objectives

Remote sensing; lightning and sprites GPS occultation and TEC Ionospheric eﬀect of space weather

Status

D

International

Notes PS

ESA(MPO)

PS/ST, e

O

USA

ST

L

USA

ST, µ, h

D

TBA

ST

ch20

263

Launch: Launch year and month; ( ): scheduled, planned, or proposed launch date only; Orbit: Initial orbit perigee and apogee in altitude (km) or geocentric distance (Re) and inclination; (p): planned; (c): current; Status: Mission status; O: operating; L: launch ready; D: under development; A: approved; P: planned/proposed; International: International partners. Notes: ST: solar terrestrial; PS: planetary science; AT: Astronomy; µ: micro-satellite; a: based on information presented elsewhere in the AOGS meeting, see text; b: Chang’e-1 and other planetary missions not listed due to lack of detailed information; c: 5 spacecraft constellation; d: November 2005 arrival at asteroid; e: one of two BepiColumbo spacecraft; f: penetrator mission; g: also astrophysics payload, see text; h: 6-micro-satellite constellation.

Solar Terrestrial and Planetary Science Missions in Asia–Oceania

Country

(Continued )

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Table 1.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

264

ch20

A. W. Yau et al.

of the respective missions. The list is not meant to be exhaustive, and covers primarily missions that were discussed in some details in the Open Forum. In conclusion, the discussion in the Forum underscores the beneﬁts of sharing and discussing information, at the grass-roots level, among Asian and Oceanian geoscientists on their respective national ST–PS and related space programs. The need for doing so arises because researchers are best able to identify potential scientiﬁc collaborative opportunities in a proposed or planned mission, and the best time to do so is the formative stage of a mission, when it is easier or more possible to accommodate relevant mission or instrument design requirements. This is particularly true in multipurpose missions such as STSAT and FORMOSAT, in which ST–PS science is not the primary or only mission objective. The discussion also attests to the overwhelming interest of researchers in the Region to pursue collaborative research programs, and the signiﬁcant long-term scientiﬁc beneﬁts that such collaborations engender.

Acknowledgments We wish to acknowledge the contributions of Dan Baker, Lars Blomberg, Bernard Foing, Stefano Orsini, Craig Pollock, and many other colleagues to the Open Forum discussions.

References 1. X. H. Feng, F. S. Wei, G. L. Huang, M. Zhang and Y. H. Yan, Solar and space progress in China associated with IHY, IAGA Scientiﬁc Assembly, IHY International Planning Workshop, 18–29, July 2005, Toulouse, France. 2. M. Nakamura, M. Kato and Y. Kasaba, JAXA future program for solar system sciences, Advances in Geosciences (2006), in press.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch21

INCOHERENT SCATTER RADAR IN IONOSPHERIC RESEARCH: PAST CONTRIBUTION AND FUTURE PROMISE T. HAGFORS Max Planck Institute for Solar System Research Max Planck Strasse 2 D-37191 Katlenburg-Lindau, Germany

A very brief review is given of the fundamentals of Incoherent Scatter Radar (ISR) including history, theoretical development and observatories. A few examples of the observational results pertaining to the ionosphere are given, and some new technical and observational innovations are touched upon.

1. Introduction The observation of scattering from thermally induced plasma density ﬂuctuations in the ionosphere has developed into the most important ground based diagnostic tool for the determination of the state of the ionospheric plasma. In this paper this development will be traced from the beginning with Gordon’s1 original proposal; through the ﬁrst conﬁrmation of his prediction by Bowles,2 through the race to build the ﬁrst observatories, the conﬁrmation of the linear plasma theory, the spreading of what came to be called Incoherent Scatter Radar (ISR) and through some predictions which, after 45 years, may lay ahead for these types of observations. But before embarking on this exposition it is well to remember the historical context that immediately made ISR so exciting, causing it to spread so rapidly to the research communities of many countries. In 1957 the USSR launching of Sputnik created shock and consternation in the US because of the fear that the USA was about to be bypassed in scientiﬁc and technological development. In 1958 USA and the USSR each had massive arsenals of Intercontinental Ballistic Missiles (ICBM) and Nuclear weapons. USSR conducted 34 atomic bomb-tests; the US 77. In North America Ballistic Missile Early Warning Systems (BMEWS) were in operation at three sites. Gordon’s postulation of incoherent scattering was seen as a method to routinely monitor the ionosphere, of great interest in the early warning (BMEWS) detection schemes. 265

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

266

ch21

T. Hagfors

2. What is Incoherent Scatter? The proposal of Ref. 1 involved the scattering from individual electrons independently of the surrounding medium. This scattering process is depicted in Fig. 1. Here Einc is the electric ﬁeld incident on the electron, ue is the oscillating velocity amplitude of the electron, χ is the angle between the velocity vector and the radius vector to the receiver and Escut is the scattered electric ﬁeld from the electron. In Gordon’s original work the power from all the electrons in the scattering volume were added, and a truly incoherent scatter signal resulted. This turned out to be too simplistic as shown by the ﬁrst experiments by Bowles.2 In these initial experiments it appeared as if the scattering came from the ions rather than from the electrons, an impossibility in view of the large mass of the ions compared with that of the electrons. The explanation was soon found and stems from the electrostatic interaction of ions and electrons preventing the electrons from traveling freely and independently of the ions. It is as if the ions surround themselves with a cloud of electrons. These clouds travel with the velocity of the ions. Rather than reﬂecting the electron motion and temperature the scattered waves come from the electron clouds traveling with the ions. The number of parameters controlling the scattering therefore turned out to be much larger than originally foreseen, and this opened up the possibility of determining a much larger number of plasma parameters. Due to the collective eﬀects in the plasma it appears that the scattering is determined by plasma waves with wave vector equal to the diﬀerence of the wave vectors of the transmitted and the received waves (see Fig. 2).

Fig. 1.

Scattering single electron.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch21

ISR in Ionospheric Research: Past Contribution and Future Promise

267

Fig. 2. Bistatic scattering from a volume containing the ionospheric plasma. The volume is determined by the polar diagrams of the antennas and by the pulse delay resolution of the system.

In terms of the Fourier components of the density ﬂuctuation in the plasma the radar cross section per unit of volume is: σradar = 4πσtotal = 4πre2 V −1 sin2 χ |N (k)|2 = 4πre2 sin2 χN (k) , where re is the classical radius of the electron and where 4πre2 = 0.998 × 10−28 ≈ 1.00 × 10−28 m2 a nice round number to remember.

3. How are Physical Parameters Determined? The proposal of Ref. 1 envisaged the determination of electron temperature and electron density throughout the ionosphere within the range of the scattering. The power spectra of the received signals were much more complicated than at ﬁrst thought and depended on electron and ion temperatures, density of electrons and several species of ions and on collision frequencies. The plasma theory developed for the calculation of spectra for the many parameters of the plasma is one of the most accurate and

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

268

ch21

T. Hagfors

successful theories in plasma physics. The procedure for the extraction of data on the ionospheric plasma is to ﬁt the theoretical spectra as closely as possible to the observed spectra, and adopting as parameters the ones giving the best ﬁt. Figure 3 shows an early version of power spectra plotted against normalized frequency oﬀset for a number of diﬀerent ratios of plasma wavelength to Debye length. Here the Debye length is given by ωp /ve where ωp is the angular plasma frequency and where ve is the thermal velocity of the electrons. In Fig. 3 the spectral distribution assumed a single ion species and an equal temperature of ions and electrons. There are many cases where the electrons and the ions are at diﬀerent temperatures, particularly when the production of electrons is high. Then the temperature of the electron gas tends to be higher than that of the ions. This has a profound eﬀect on the shape of the spectra as shown in Fig. 4. Clearly the ﬁtting of observational data to the shape of these theoretical spectra can be used to derive the temperature ratio.

Fig. 3. Theoretical spectra used to determine electron temperature, densities, and plasma density.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch21

ISR in Ionospheric Research: Past Contribution and Future Promise

269

Fig. 4. Plots of Ni (k, ω)/n0 with Te /Ti as parameter for a value of 1/kλD = 0.1. The spectra for oxygen (fully drawn) and for hydrogen (stippled) show slightly diﬀerent behavior even though plotted against normalized frequency θi .

Mixtures of ions of diﬀerent mass also have characteristic spectral features which can be used in plasma diagnostics. Figure 5 shows an example of spectra of a plasma with oxygen and hydrogen ions. The collisions between the plasma particles and the mostly much denser neutral background also aﬀect the spectral distribution in characteristic ways which can be used to derive density of the neutral gas, at least in principle.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

270

ch21

T. Hagfors

Fig. 5. Spectra for various mixing ratios of hydrogen and oxygen, in this particular case for a ratio of Te /Ti = 2.

Particularly fruitful investigations of the plasma dynamics in the ionosphere have been made through studies of the mean frequency shifts of the spectra, from which plasma motion induced either by electric ﬁelds or by neutral wind interacting with the plasma can be educed.

4. Observatories Built to Exploit the Technique As the possibilities for scientiﬁc investigations by ISR became obvious the pressure to build observatories to exploit the technique for scientiﬁc investigations became strong and several observational facilities were constructed, in the US, Peru, France, UK, Japan, and Russia. Most of these were located in low or middle latitudes. It became clear quite early that

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch21

ISR in Ionospheric Research: Past Contribution and Future Promise

271

the technique could also ﬁnd fruitful application at high latitudes to study auroral phenomena, and to establish a link between the ionosphere and the magnetosphere. In order to extend the technique to higher latitude investigations a radar system was moved from Stanford, California to Chatanika, Alaska, and then on to Søndre Strømfjord, Greenland. A high latitude state of the art observatory, EISCAT, supported by a consortium of European research councils, was subsequently installed in Northern Scandinavia. Figure 6 shows the two observatories built shortly after the original proposal was made. The second generation observatory, EISCAT, relied on the experience gained in the operation of the previous observatories, and extended the capabilities of ISR in many ways. The transmitter site near Tromsø, Norway operated at two diﬀerent wavelengths, 1.3 m and at 0.3 m. At the shorter wavelength receivers were placed in Sodankyl¨ a, Finland and Kiruna, Sweden in addition to one at the transmitting site. This allowed for the measurement of three-dimensional plasma velocities, as sketched in Fig. 7. The system was extended to Svalbard to allow studies within the polar cap and to collaborate with the Cluster II satellite mission. In addition this extension has led to some exciting and unexpected new studies of the ﬁne scale structure of auroral arcs.

5. A Few Examples of Scientiﬁc Results It is not possible in this brief talk to do justice to the large number of investigations which have been carried out. Somewhere in the range of 2000 papers based on ISR have been written. Clearly one cannot do justice to this work. The few examples I can describe here reﬂect my personal interest and it should in no way be construed that the papers not mentioned are of inferior quality and less interest. Extremely strong echoes have been observed near 80 km altitude indicating the presence of charged ice crystal embedded in the weakly ionized plasma. It is quite likely that the behavior of this layer, if not having a controlling inﬂuence on climate and weather, at least is closely related to both as a tracer. These echoes are referred to as Polar Mesospheric Summer Echoes, or PMSE. and have been the subject of extensive studies by ISR systems. Electron density, electron temperature, ion temperature, and ion velocity over a large height range are now routinely obtained in the observatories mentioned above and others, and such data can be exploited to reveal the

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

272

ch21

T. Hagfors

(a)

(b)

Fig. 6. (a) The Jicamarca observatory, near the magnetic equator close to Lima, Peru. The observatory is operating at 50 MHz with a square array antenna made up of 18 000 crossed dipoles. The observatory has been in steady development, and is still contributing valuable scientiﬁc data on equatorial irregularities and currents, meteors and mesospheric phenomena and (b) the Arecibo observatory, a spherical reﬂecting telescope, 300 m in diameter with more than 40 000 individually reﬂecting panels. For ISR purposes the observatory is operating at 430 MHz with a beam steerable as much as 20◦ from the zenith. The observatory originally built for ionospheric research has proved extremely useful in radar astronomy of our solar system, and in radio astronomy in pulsar research, in VLBI research and in studies of cosmology by studies of galaxies at the greatest distances in the universe.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch21

ISR in Ionospheric Research: Past Contribution and Future Promise

273

Fig. 7. The EISCAT tristatic UHF system with the VHF monostatic system and the HF ionospheric modiﬁcation radar. This system has been expanded to Longyearbyen, Svalbard.

chemical and dynamical properties of the upper atmosphere and to support space experiments. Two examples of observational results are shown in Figs. 8 and 9.

6. What Lies Ahead in ISR? The most exciting new development is the planned introduction of phased array radar systems with distributed transmission elements and the ability to steer the beam from transmitter pulse to transmitter pulse. Furthermore the antenna array will be made in transportable sections allowing the IRS

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

274

ch21

T. Hagfors

Fig. 8. Very rapid electron density variation related to the aurora. They are caused by magnetospheric phenomena directly caused by burst-like electron density precipitation.

Fig. 9. Waves and tides in the atmosphere as observed through plasma motion. Shown are the plasma density variation as a function of height for diﬀerent spatial wavelengths. The variation of the phase of these waves with height is characteristic of internal.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch21

ISR in Ionospheric Research: Past Contribution and Future Promise

275

Fig. 10. Two elements of the array antenna under development at SRI (Stanford Research International) in Menlo Park, California.

radar to operate at magnetically or geographically interesting places. A sketch of a two panel system of this antenna is shown in Fig. 10. Each square panel is about 2.6 m on the side, and a typical application such as in Poker Flat Alaska will employ 128 panels. Another interesting and promising application of ISR is in the study of extremely ﬁne spatial structure (down to the 10 m level) with high time resolution. This has been tried out successfully in Svalbard in the study of the ﬁne structure of auroral arcs.

Acknowledgments I have beneﬁtted greatly from discussions and collaboration with a number of colleagues, D. T. Farley, C. Heinselman, C. LaHoz, M. Sulzer, R. Woodman and a number of others, too numerous to be listed because of the limited space allowed by the editor. For the same reason I only list the two original papers as references.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

276

ch21

T. Hagfors

References 1. W. E. Gordon, Incoherent scattering of radio waves by free electrons with applications to space exploration by radar, Proc. IRE 46 (1958) 1824–1829. 2. K. L. Bowles, Observations of vertical incidence scatter from the ionosphere at 41 Mc/s, Phys. Rev. Lett. 1 (1958) 454–455.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch22

VERTICAL GEOMAGNETIC CUTOFF RIGIDITIES FOR EPOCH 2000 — DEVIATIONS FROM EXPECTED LATITUDE CURVES D. F. SMART and M. A. SHEA Air Force Research Laboratory (VSBX) 29 Randolph Road, Hanscom AFB, Bedford, MA 01731, USA

The Earth’s geomagnetic ﬁeld is evolving rapidly (in geological time) and as a consequence, the amount of geomagnetic shielding at a speciﬁc location is also changing. Geomagnetic cutoﬀ rigidities derived from the International Geomagnetic Reference Fields (IGRF) are a basic quantity necessary to compute the radiation dose due to cosmic radiation experienced along aircraft routes. Aircraft measurements of the radiation dose experienced along speciﬁc ﬂight paths are suﬃciently precise that the secular variation of the geomagnetic ﬁeld is observable. The ninth generation of the IGRF describes the Earth’s magnetic ﬁeld to previously unattainable precision. An updated set of geomagnetic cutoﬀ rigidities for Epoch 2000 utilizing the high-precision geomagnetic ﬁeld coeﬃcients resulted in an updated and more precise world grid of geomagnetic cutoﬀ rigidity values. We describe our computations and examine some of the discontinuities in the eﬀective geomagnetic cutoﬀ rigidities that may result in limitations in the applications of vertical geomagnetic cutoﬀ rigidities.

1. Geomagnetic Cutoﬀ Rigidities Geomagnetic cutoﬀ rigidities (rigidity is momentum per unit charge) are used as a tool for the following: (1) To quantitatively measure the shielding provided by the earth’s magnetic ﬁeld. (2) To predict the energetic charged particle transmission through the magnetosphere to a speciﬁc location as a function of direction. (3) To order measurements of charged particle data, or the charged particle eﬀects such as radiation dose. Determination of geomagnetic cutoﬀ rigidity values involves one of the problems that has “no mathematical solution in closed form”, except for the special case of a dipole magnetic ﬁeld. The generally accepted manner for determining cutoﬀ rigidities is by tracing particle trajectories in a high order simulation model of the magnetic ﬁeld (see Ref. 1 for a 277

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

278

ch22

D. F. Smart and M. A. Shea

review). The geomagnetic cutoﬀ rigidity determination procedure is very computer intensive, involving the evaluation of millions of individual trajectory calculations. Geomagnetic cutoﬀs are directional. Because of the extremely computer intensive procedures necessary to determine angular geomagnetic cutoﬀs (vertical is only one direction), empirical practices have been adopted. The usual procedure is to use the calculated vertical geomagnetic cutoﬀ rigidity value and estimate the geomagnetic cutoﬀ in other directions by application of St¨ ormer2 theory normalized to the vertical geomagnetic cutoﬀ value. Charged particles traversing the earth’s magnetic ﬁeld undergo a force that results in a curved path. The presence of nondipole terms and the oﬀset of the magnetic center with respect to the geocenter greatly complicates the geomagnetic cutoﬀ problem. Trajectories that would normally be allowed in a pure dipole magnetic ﬁeld become forbidden due the presence of a solid object (the earth). High rigidity cosmic rays (such as the trajectory labeled as 1 in Fig. 1), travel relatively simple orbits. As the rigidity of the particle decreases the amount of geomagnetic bending increases (examine the trajectories labeled 4 and 5 in Fig. 1) and the orbits become more complicated, forming intermediate loops. When these loops intersect the

Fig. 1. Illustration of the cosmic ray trajectory-tracing process. The highest rigidity (most resistant to geomagnetic bending) is labeled 1 and the lowest rigidity is labeled 15.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Vertical Geomagnetic Cutoﬀ Rigidities for Epoch 2000

ch22

279

solid earth, the orbits are forbidden. At still lower rigidities, there may be allowed trajectories; trajectory 15 in Fig. 1 is an allowed orbit. A common misconception is that geomagnetic cutoﬀs are “sharp”. The geomagnetic cutoﬀ is “diﬀuse” except for vertical geomagnetic cutoﬀs near the magnetic equator. There are rigidities (energies) above which all charged particles are allowed [the upper (RU ) geomagnetic cutoﬀ], and there are rigidities (energies) below which all charged particles are forbidden [the lower (RL ) geomagnetic cutoﬀ] (see Ref. 3 for detailed deﬁnitions of the various cutoﬀ rigidity parameters). In most cases, the transmission of charged particles decreases from fully allowed to totally forbidden over a surprisingly large range of charged particle rigidities. The region between full access and totally forbidden access is called the cosmic ray penumbra (a term adapted from optics). The cosmic ray penumbra is a chaotic region of allowed and forbidden trajectories lying between RU and RL . The penumbra frustrates the desire for a simple number for the geomagnetic cutoﬀ. Eﬀective geomagnetic cutoﬀs (RC ) are obtained by a linear average of the allowed bands in the penumbra.4 The International Geomagnetic Reference Field (IGRF) models are generated under the auspices of the International Association of Geomagnetism and Aeronomy (IAGA).5 These models are current state-of-the-art representations of the earth’s internal geomagnetic ﬁeld. The Earth’s magnetic ﬁeld is changing very rapidly (see Fig. 2) and IAGA updates the IGRF models every ﬁve years. The secular change in the geomagnetic ﬁeld is suﬃciently rapid that on a decadal scale, diﬀerences in the cosmic ray exposure and hence the radiation dose due to cosmic rays are measurable. These changes are most rapid in the North and South Atlantic region as illustrated in Fig. 3. Cosmic radiation measurements along speciﬁc ﬂight paths can detect the eﬀects of these secular variations in the geomagnetic ﬁeld.6 Since there is a signiﬁcant change in the earth’s magnetic ﬁeld the United States Federal Aviation Agency has commissioned a new set of vertical geomagnetic cutoﬀ rigidities using the IGRF 2000 ninth generation magnetic ﬁeld (degree and order 13).5 This magnetic ﬁeld model has unprecedented precision; older models were restricted to degree and order 10. Utilizing this updated magnetic ﬁeld model, we have calculated a new set of vertical geomagnetic cutoﬀ rigidities for every degree in latitude and 5◦ in longitude over the Earth. (Previous world grids of vertical cutoﬀ rigidities were calculated for much coarser world grids: 5◦ in latitude by 15◦ in longitude for the 1955, 1965, and 1975 epochs, and 5◦ in latitude

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

280

Fig. 2.

ch22

D. F. Smart and M. A. Shea

Change in magnitude of the Earth’s dipole as represented by the G(1,0) term.

by 5◦ in longitude for 1980 and 1990.) This unprecedented precision and resolution should provide a “benchmark” for the future and also resolve any ambiguity in the behavior of the calculated vertical geomagnetic cutoﬀ rigidities. Figure 4 is a world map displaying the iso-rigidity contours derived employing the IGRF 2000 model.

2. Deviations from the Expected Latitude Curve When vertical geomagnetic cutoﬀ rigidities are plotted as a function of latitude along a longitudinal meridian, the normal expectation is a smooth curve similar to that expected from St¨ ormer2 theory. When the trajectoryderived vertical cutoﬀ rigidities are plotted along a longitudinal meridian, there are signiﬁcant deviations from an idealized smooth curve, especially in the RL geomagnetic cutoﬀ values. The vertical cutoﬀ rigidity penumbral width has signiﬁcant and abrupt variations as the rigidity changes along any speciﬁc longitude. There is an oﬀset of the eﬀective magnetic center from the geocenter. This oﬀset generates a longitude dependence eﬀect in the penumbral structure that results in additional discontinuities in the eﬀective geomagnetic cutoﬀ rigidity.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Vertical Geomagnetic Cutoﬀ Rigidities for Epoch 2000

ch22

281

Fig. 3. Change in vertical cutoﬀ rigidity (in units of GV) between 1950 and 2000 (black indicates increase, red indicates decrease).

Fig. 4. Iso-rigidity contours for vertical geomagnetic cutoﬀ rigidities (in units of GV) for Epoch 2000.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

282

ch22

D. F. Smart and M. A. Shea

This deviation from an idealized latitude curve has been noted ever since precision cosmic ray measurements have been organized by the calculated eﬀective vertical cutoﬀ rigidity. Carmichael et al.7 ﬁrst published this eﬀect when attempting to order his careful land-based latitude survey (having 0.1% statistical precision in the cosmic ray neutron monitor counting rate data). The most extreme discontinuities in the eﬀective vertical geomagnetic cutoﬀ rigidity with latitude occur in the tropical latitudes near the North American continent. The reason for these large penumbral discontinuities is that the “mirror point” of the cosmic ray trajectories occurs in the southern hemisphere where the relative oﬀset of the eﬀective magnetic dipole center is at it’s maximum distance, and some trajectories that would normally re-enter at other longitudes, have an additional “bounce” in their orbit and become allowed. The discontinuous features are measurable with cosmic ray instrumentation having a statistical accuracy of 1% or better. We are not aware of any cosmic radiation dose data acquired on aircraft that has the statistical precision to show this eﬀect. However, the penumbral eﬀect has been noted by modern cosmic ray latitude surveys conducted by cosmic ray neutron monitors on ocean voyages to the Antarctic.8 In the following plots, all three vertical geomagnetic cutoﬀ values (RU , RL , and RC ) have been plotted in the rigidity region where the penumbra is signiﬁcant. Examination of these plots allows a visualization of the extent of the cosmic ray penumbra and the discontinuities in the eﬀective geomagnetic cutoﬀs (RC ) with latitude. In these plots, the top line displays the upper cutoﬀ rigidity, the bottom line displays the lower cutoﬀ rigidity and the center line displays the eﬀective cutoﬀ rigidity. Figure 5 illustrates an eﬀective vertical cutoﬀ rigidity latitude curve in the European–African sector of the world. The deviations of the vertical cutoﬀ rigidities can be noted in the northern latitudes between 25 and 45◦ . Figure 6 illustrates an eﬀective vertical cutoﬀ rigidity latitude curve in the Asian region. The slightly larger deviations in the Australian and Japanese latitudes are apparent in this ﬁgure. It is possible that with careful monitoring along the commercial air routes in this sector, that these eﬀects might be found in the radiation dose data. Figure 7 illustrates the eﬀective vertical cutoﬀ rigidity discontinuities in the latitude curve in the North American sector of the world. These eﬀects are repeatedly observed in surface cosmic ray latitude curves7,8; however, we know of no systematic eﬀort to ﬁnd this eﬀect in aircraft cosmic ray radiation dose data.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Vertical Geomagnetic Cutoﬀ Rigidities for Epoch 2000

ch22

283

Fig. 5. Illustration of the cutoﬀ rigidity latitudinal variations in the European (top)– African (bottom) region. In each panel the top trace is RU , the center trace is RC , and the lower trace is RL .

3. Summary Geomagnetic cutoﬀ rigidities derived from the IGRF are a basic quantity necessary to compute the radiation dose due to cosmic radiation experienced along aircraft routes. The ninth generation of the IGRF

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

284

ch22

D. F. Smart and M. A. Shea

Fig. 6. Illustration of the cutoﬀ rigidity latitudinal variations in the Asian region for the Japanese region (top) to the Australian region (bottom). In each panel the top trace is RU , the center trace is RC , and the lower trace is RL .

describes the Earth’s magnetic ﬁeld to previously unattainable precision. An updated set of geomagnetic cutoﬀ rigidities for Epoch 2000 utilizing these high precision geomagnetic ﬁeld coeﬃcients has resulted in a more precise world grid of geomagnetic cutoﬀ rigidity values. Examination of the vertical cutoﬀ rigidity latitude curves reveals computational artifacts and

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Vertical Geomagnetic Cutoﬀ Rigidities for Epoch 2000

ch22

285

Fig. 7. Illustration of the cutoﬀ rigidity latitudinal variation extremes in the region of the North American continent. The top trace is RU, the center trace is RC , and the lower trace is RL .

discontinuities that may result in limitations in the applications of vertical geomagnetic cutoﬀ rigidities.

References 1. D. F. Smart, M. A. Shea and E. O. Fl¨ uckiger, Space Science Reviews 93 (2000) 271. 2. C. St¨ ormer, The Polar Aurora (Clarendon Press, Oxford, 1955). 3. D. J. Cooke, J. E. Humble, M. A. Shea, D. F. Smart, N. Lund, I. L. Rasmussen, B. Byrnak, P. Goret and N. Petrou, Il Nuovo Cimento C 14 (1991) 213. 4. M. A. Shea, D. F. Smart and K. G. McCracken, J. Geophys. Res. 70 (1965) 4117. 5. IAGA, Division 4, Working Group 8, Geophysical Journal International 155 (2003) 1051, doi:10.1111/j1265-246x.2003.0212.x. 6. P. J. K¨ onig and P. H. Stoker, J. Geophys. Res. 68 (1981) 219. 7. H. Carmichael, M. A. Shea, D. F. Smart and J. R. McCall, Can. J. Phys. 47 (1969) 2067. 8. J. M. Clem, J. W. Bieber, M. Duldig, P. Evenson, D. Hall and J. Humble, J. Geophys. Res. 102 (1997) 26919.

This page intentionally left blank

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch23

ULTRA LONG RANGE AIRCRAFT OPERATIONS AND SPACE WEATHER I. L. GETLEY Department of Aviation, University of New South Wales Sydney, NSW 2052, Australia [email protected] M. L. DULDIG Department of the Environment and Heritage, Australian Antarctic Division Kingston, Tas. 7050, Australia [email protected]

The advent of ultra long range modern jet aircraft brings into focus issues concerning cosmic and solar radiation exposure to aircrew and passengers alike. Recent development of computer models to help predict these levels of exposure have gone some way to allow aircraft operators to meet various legislative requirements for monitoring aircrew and passengers. These models do not allow for solar particle events and the proposed longer duration ﬂights will increase time at altitude and hence the levels of exposure, particularly during such events. This paper considers the issues involved, in particular the use of exposure codes versus direct measurement.

1. Introduction Exposure of ﬂight crews of jet aircraft to cosmic radiation has generated considerable debate in aviation, the scientiﬁc community, and by health authorities responsible for regulation of dose limits. The International Commission on Radiological Protection introduced ICRP 60 in 1991.1 This designated ﬂight crews as radiation workers and set the maximum yearly exposure per ﬂight crew member to not more than 20 mSv. The public dose limit is set at 1 mSv as is the limit for pregnant women including female crew, thus giving the conceptus the same protection as the general public. These dose rates are eﬀective dose rates, and reﬂect the diﬀerent sensitivities that individual organs and body tissue have to radiation doses. These are therefore an expression of this overall risk. In March 2000, the Joint Aviation Authority (JAA), the regulatory authority of member states of the European Union, ﬁled a notice of 287

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

288

ch23

I. L. Getley and M. L. Duldig

proposed amendment NPA OPS-23 to Joint Aviation Regulations JAROPS Pt1. It proposed to amend JAR-OPS 1.390 Cosmic Radiation, requiring aircraft operating above 49 000 ft to have serviceable radiation monitoring equipment. The amendment noted that crew subjected to greater than 1 mSv per year should be monitored. For this purpose, quarterly radiation sampling under AMC OPS 1.680(a)(2) that provides for an actual periodic measurement, is deemed an acceptable means of compliance when compared against predictive ﬁgures (by accepted computer methodology). JAR-OPS 1.390 Cosmic Radiation (b) Passive Monitoring, states: “An operator shall take account of the in ﬂight exposure to cosmic radiation of their ﬂight crew and cabin crew while on duty and shall take the following measures for those crew liable to be subject to exposure of more than 1 mSv per year: (1) assess their exposure; (2) arrange work schedules where practicable, to keep exposure below 6 mSv per year; (3) inform them of the health risk associated with likely exposure; (4) ensure that the working schedules for female crew, once they have notiﬁed the operator that they are pregnant, to keep the equivalent dose to the fetus as low as can be achieved and in any case ensure that the dose does not exceed 1 mSv for the remainder of the pregnancy; (5) ensure that where exposure is considered likely to exceed 6 mSv per year, records are kept for each ﬂight or cabin crew member aﬀected, and that the appropriate medical surveillance is applied”. “In addition, no further controls are necessary for aircrew whose annual dose can be shown to be less than 1 mSv. For aircrew whose annual dose falls in the range 1–6 mSv there should be individual estimates of dose. These estimates of dose should be made available to the individual concerned. For ﬂights below 49 000 ft these may be carried out using an appropriate computer program and internationally agreed information on radiation levels for various routes and altitudes ﬂown”. These computer-derived dose estimates will generally be moderately cautious over estimates of long-term mean doses. That this is the case should be conﬁrmed by occasional measurements using either active instruments on speciﬁc ﬂights or passive measuring devices for a number of ﬂights on an individual route. It is against this background that the concept of As Low As Reasonably Achievable (ALARA) was promoted with regard to occupational radiation exposure of ﬂight and cabin crew, in fact for any at-risk groups involved in frequent ﬂying.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Ultra Long Range Aircraft Operations and Space Weather

ch23

289

2. Radiation Eﬀects 2.1. Ionizing radiation The principal ionizing radiation is galactic cosmic radiation or simply cosmic radiation. The main source of this radiation is thought to be supernovae and by the 1940s, it was clear that the main constituent of cosmic rays (CR) were relativistic protons with smaller fractions of heavier particles. The distribution being approximately; 90% protons (hydrogen nuclei), 9% alpha particles (helium nuclei) and 1% made up of heavier nuclei. The Latitude Eﬀect, where the intensity of the radiation was found to increase toward the magnetic poles at the Earth’s surface, was detected in the 1930s and was ascribed to the fact that CR were charged particles that were being deﬂected by the Earth’s magnetic ﬁeld. At jet cruising altitudes, in mid-latitudes, there is a 2.5–5-fold increase compared to radiation intensity at the Earth’s surface. Near the poles, the particles arrive parallel to the ﬁeld lines and follow trajectories into the atmosphere where they interact to produce secondary radiation, some of which reaches the Earth’s surface. Another source of ionizing radiation is photons. The kinetic energies of both the charged particles and photons are distributed over a very wide spectrum, with the intensity of the energy spectrum of CR peaking at around 1 GeV (109 eV). At these levels a proton with energy 1 GeV is traveling at about 87% of the speed of light. Protons with energies up to around 20 GeV can be produced by the Sun during solar ﬂare activity but the bulk of the particles from ﬂares have energies in the MeV range. The intensity of CR is not constant but is subject to solar modulation in which variations in the Sun’s general activity changes the interplanetary magnetic ﬁeld. Cosmic radiation with energies below 1 GeV is most aﬀected by this modulation and can vary in intensity by a factor of 10. As a result, when the Sun is at a high point of its activity cycle the average cosmic radiation intensity above a few hundred MeV is reduced. This occurs because the solar wind, carrying magnetic ﬁelds into interplanetary space, is more eﬀective at preventing the interstellar ﬂux of CR from reaching the inner solar system around the Earth during periods of high solar activity. This increased shielding by magnetic ﬁeld scattering of the incoming higher energy galactic cosmic radiation during high solar activity reduces the intensity of cosmic radiation at the Earth, with a corresponding decrease

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

290

ch23

I. L. Getley and M. L. Duldig

in dose rates. The lower energy particles that comprise the solar wind (and indirectly cause the modulation of the CR) are themselves too low in energy to generate ionizing radiation at aircraft cruising altitudes. However, as mentioned above, on rare occasions the Sun can be a source of intense bursts of very high energy particles associated with solar ﬂares and coronal mass ejection driven shocks. The mechanism that accelerates these particles to such high energies is hotly debated in the literature. Shock acceleration and stochastic processes both probably play a part. Although the solar activity shows a clear approximately regular 11-year cycle (to which the galactic cosmic ray ﬂux responds inversely) the solar ﬂare related particles are only weakly correlated with solar activity and can occur at any time in the solar cycle. As previously mentioned the Earth’s magnetic ﬁeld helps to deﬂect cosmic rays away from the Earth. The highest protection is therefore aﬀorded at the magnetic equator or lower magnetic latitudes and the intensity increases as one moves towards the magnetic poles. The very energetic primary cosmic radiation particles (mostly protons and alpha particles) undergo a variety of nuclear reactions as they enter the atmosphere and collide with and break apart the nuclei of nitrogen, oxygen and other atoms. These collisions release numerous secondary particles (the actual numbers and types depend on the mass and energy of the initiating cosmic ray) including protons, neutrons, electrons, positrons, and pions. In addition, γ-rays are generated by energy–mass transformations when the positrons collide and annihilate with atmospheric electrons. The pions are short lived and decay into highly penetrating muons which typically survive to the Earth’s surface and can be observed deep underground. Following a collision, the secondary particles may in turn have enough energy to produce still more particles. This results in an atmospheric cascade, commencing at about 20 km above the Earth’s surface, and moving down at speeds close to the speed of light. At any given instant the cascade is shaped like a disk with the main concentration of secondary particles at the center. The maximum intensity of the secondary radiation occurs at about 15 km above the Earth’s surface. Higher energy cosmic ray showers produce peak radiation levels as low as 2 km altitude but the ﬂux of these cosmic rays is vastly smaller than the 1–20 GeV CR and they comprise only a small fraction of the overall dose. Thus, when galactic cosmic radiation particles enter the atmosphere, the number of ionizing particles initially increases with decreasing altitude (due to these secondary collisions) then decreases with further altitude decrease

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Ultra Long Range Aircraft Operations and Space Weather

ch23

291

as the particles lose energy. The particles entering the atmosphere (primary particles) plus the particles produced by impact (secondary particles) are referred to collectively as galactic cosmic radiation. A single primary particle may have suﬃcient energy to generate millions of secondary particles. Individual particles of energy 3 × 1020 eV (∼ 5 J — equivalent to a tennis ball traveling at 300 km/h or lifting one kilogram 0.5 m against gravity) have been detected.2 The CR follow spiral trajectories around the magnetic ﬁelds in space that they encounter and around the Earth’s own magnet ﬁeld. Consequently it is generally not possible to determine the particle origins. In fact, cosmic rays arrive from all directions in space, with approximately equal intensity both day and night. Some directionality (or anisotropy) is present at extremely low levels of less than one percent but during solar ﬂare related events quite strong anisotropy is possible resulting in high radiation levels at one location and much lower levels at another location with the same altitude and geomagnetic latitude but diﬀerent longitude or in the opposite hemisphere. The eﬀective dose rate,1,3 a measure of the biological harmfulness of ionizing radiation, follows the same general pattern as the number of ionizing particles. That is, the dose rate initially increases with decreasing altitude, then decreases with further decrease in altitude. The maximum dose rate varies with the approximate 11 years solar-activity cycle for any given altitude and geographic location on the Earth. The secondary particles produced can be divided into two categories, high and low linear energy transfers (i.e., high LET and low LET). High LETs give up their energy quickly in tissue and have short tracks along which there is intense energy dissipation. Low LETs give up their energy slowly and produce long tracks with sparse energy dissipation. The less energy a primary particle carries the fewer secondary particles it generates and consequently the higher the altitude at which ionizing radiation ceases. At high latitudes when solar activity is at a minimum, the number of low-energy particles entering the atmosphere is at a maximum, causing radiation levels to peak at high altitudes. With decreasing latitude and/or increasing solar activity, fewer low-energy particles enter the atmosphere so that maximum radiation levels occur at lower levels (a result of the previously mentioned cascade created by higherenergy particle impacts). At the low latitudes, i.e., equatorial regions, the altitudes at which radiation levels peak are largely unaﬀected by solar activity.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

292

ch23

I. L. Getley and M. L. Duldig

2.2. Ionizing radiation from the Sun As mentioned earlier, eruptions at the Sun associated with ﬂares can lead to explosive emission of fully ionized, low energy particles. Through shock and/or stochastic acceleration some of these particles are accelerated to very high energies. These increases are referred to by various terms such as: solar–particle events, solar energetic–particle events, solar–proton events (SPEs) and cosmic ray ground level enhancements. As with galactic ionizing radiation, the particles interact with air atoms creating the same cascade of particles but sometimes at much greater intensities, which is a cause for concern for aircrew exposure. However, only on rare occasions does a solar–proton event lead to a substantial increase in the ionizing radiation levels at normal jet aircraft cruise altitudes. The most energetic of these particles can reach the Earth’s atmosphere within 10 min after ejection from the Sun.4 These particles arrive from approximately the direction of the local interplanetary magnetic ﬁeld connecting the Sun and the Earth. Due to scattering eﬀects of the interplanetary magnetic ﬁeld, this anisotropy diminishes and particles are found to be arriving from all directions.5 Thus within a half hour to a few hours after the start of the event, radiation levels in the atmosphere on both the sunward (light) and dark hemispheres of the Earth come close to being equal.6 Radiation levels at ﬂight altitudes during the solar–proton event of September 29–30, 1989, were amongst the highest recorded in over 30 years to that time. During this event, the highest eﬀective dose rates of ionizing radiation from the Sun, at 40 000 ft and at various locations above 60◦ geomagnetic latitude, ranged from about 0.025–0.12 mSv/h.7 Immediately prior to this event, the eﬀective dose rate from galactic cosmic radiation at the same locations and altitudes were ∼ 0.007 mSv/h. During this event, the recommended radiation limits for nonpregnant and also pregnant (max 1 mSv) crew members would probably not have been exceeded on a high-latitude ﬂight at 40 000 ft. However, the cumulative dose from galactic cosmic radiation received on ﬂights before and after the event and radiation received during the event could have exceeded recommended limits for both pregnant and nonpregnant crew members.8 In particular, crew continually scheduled on high latitude/altitude ﬂights would have been at greater risk of exceeding exposure limits. Theoretical Monte Carlo simulations of the additional exposure resulting from SPEs have been undertaken recently9,10 whilst exposure measurements during SPEs have also been made.11

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Ultra Long Range Aircraft Operations and Space Weather

ch23

293

2.3. Exposure limits ionizing radiation As of April 2002, limitations for occupational exposure to ionizing radiation recommended by the International Commission on Radiological Protection (ICRP)1 as shown in Table 1 are also recommended by the US Federal Aviation Administration (FAA).12 The ICRP recommended limit for aircrew of jet aircraft is a ﬁve years mean eﬀective dose of 20 mSv/year, with no more than 50 mSv in a single year. For a pregnant crew member, starting when she reports her pregnancy to management, the ICRP recommends that her working conditions be such as to make it “unlikely that the additional equivalent dose to the conceptus will exceed about 1 mSv”. Further, the FAA advises that pregnant crew members follow the recommendation by the National Council on Radiation Protection and Measurements and ensure that equivalent dose to the conceptus not exceed 0.5 mSv in any month.13 A Commission of the European Communities directive13 and an associated document eﬀecting its implementation,14 state: “regular assessments of occupational radiation exposure should be made for crew members likely to be occupationally exposed to more than 1 mSv in a year”. It further stated “These schedules should include radiation received on the job from natural sources. Work schedules for crew members should be arranged to keep annual exposures below 6 mSv. For those workers whose annual exposure exceeds 6 mSv, medical surveillance (Directive Article 31) and record keeping (Directive Articles 28 and 34) are recommended”. For a pregnant crew member, Article 10 of the Directive states, “starting when she reports her pregnancy to management, her work schedule should be such that the equivalent dose to the child to be born (the conceptus), will be as low as Table 1.

Radiation exposure limit recommendations.

General public (annual) Infrequent exposure for GP Occupational annual exposure Cumulative lifetime limit During pregnancy † With

ICRP (mSv)

NCRP (mSv)

1 – 20 – 2

1 – 50 10 × age 5†

no single monthly exposure to exceed 0.5 mSv. Note: That whilst a limit of 2 mSv/year applies to the conceptus as it does to the general public, the additional shielding provided by the mother’s abdomen reduces the dose to conceptus to 1 mSv/year.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

294

ch23

I. L. Getley and M. L. Duldig

reasonably achievable and unlikely to exceed 1 mSv. This limit should not be exceeded either for the remainder of the pregnancy or for the whole pregnancy, according to how Article 10 is implemented in National Legislation”. Monte Carlo techniques have also been employed to assess the exposure of the embryo and fetus arising from cosmic ray neutrons.15 From the above table we can see that the ICRP occupational exposure limit of 20 mSv/year is substantially below the NCRP limit of 50 mSv/year (approximately 556 chest X-rays). The extremely low value set for pregnancy result from the very much lower number of cells of the fetus and child, whose cells are dividing more rapidly during growth and as such are far more susceptible to cell-destructive radiation, particularly during the ﬁrst trimester when major organs of the body are developing.

2.4. Biological eﬀects of ionizing radiation To date, quantifying the increased risk of cancers to normal workers (nonpregnant) due to low-level radiation such as experienced by aircrew over a long working career have proved diﬃcult. For example 20 000 air crew exposed to no more than 2 mSv/year, statistically would result in two additional cancer related deaths yearly above the normal population sample, i.e., 60 additional over 30 years.8 Also because relatively few people have been exposed to suﬃciently high doses of ionizing radiation, it has been next to impossible to study or use epidemiology to measure the risks. Remembering that atomic bomb survivors who developed fatal cancers were subjected to low LETs, LET originating from high energy photons and γ-rays, usually resulting in massive doses and subsequent radiation burns and massive cell destruction, these exposures were usually far in excess of 3 Sv (3000 mSv). So what do we know of the eﬀects of ionizing radiation on the body cells? Firstly most cancers arise from a single cell that has accumulated multistage genetic changes that allow abnormal growth. The radiation tracks or patterns of diﬀusion within the cell can cause DNA damage, mutation, and chromosome aberrations in a single cell. Hence radiation contributes as one of the steps in the multi-stage cancer development, increasing cancer risk many years later. It is also recognized that radiation can induce genome instability, raising the chances of many genetic changes and possibly increased cancer risk from these genetic changes. Insult from ionizing radiation is always in the form of structured tracks from charged particles.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Ultra Long Range Aircraft Operations and Space Weather

ch23

295

The fundamental measurement of dosimetric quantity in radiation protection is referred to as “absorbed dose”, which is the amount of energy absorbed per unit mass, or Energy (gray) Absorbed dose = Mass Ionization ≈ (gray) . Mass The gray is the unit for absorbed dose and equals 1 J of energy released per 1 kg of exposed tissue. Absorption of 3 Sv (3000 mSv) of γ-rays would be suﬃcient to cause death in approximately 50% of the population. This equates to only 3 J/kg in the body, a relatively small measure of energy, which if absorbed in water would cause a temperature change of 0.0007◦C. The biological eﬀects that may occur depend not only on the “absorbed dose” but also on the type and energy of the radiation and the tissues involved. This “radiation weighting factor” when applied to the “absorbed dose” provides us with the “equivalent dose”. 2.5. Cosmic ray eﬀects on cells The CR, like other forms of low level ionizing radiation, when penetrating living tissue, can damage the DNA molecules and strands. A cosmic ray, ionizing an atom comprising part of a DNA strand can cause a double strand-break, the deletion of a base from one base pair in the chain, or chemical cross-linking, in the two strands. It is also possible for an indirect process to cause this type of damage, where a free radical created by the radiation attacks a DNA molecule.16 The LET, linear energy transfer approximates the ionization density along the track.16 The probability of change is very low, most changes are harmless or self repair. A single electron track interaction in a cell from a γ-ray has16 : Chance of chromosome aberration, 1 in 104 , Chance of mutation of a particular gene, 1 in 108 . There is a risk at any dose (even for one track) but the risk is small at small doses (i.e., few tracks). The natural background radiation would approximate one track per year through most body cells. Hence the probability of mutation depends directly on the number of tracks (i.e., dose). Neutrons and α-particles are more damaging for the same absorbed dose, hence the need for weighting factors.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

296

ch23

I. L. Getley and M. L. Duldig

A DNA molecule that has not been repaired or has been mis-repaired, leads to a damaged chromosome that can lead to cancer. Precise outcomes associated with a given exposure cannot be deﬁned accurately for any isolated individual, as the eﬃciency of our individual repair mechanisms vary. The long-term eﬀects of continuing low doses of radiation have not been fully researched and there are currently moves in Europe to further examine the eﬀects of neutron absorption and neutron weighting factors at jet aircraft cruising altitudes. Chromosome aberration analysis in blood samples of aircrew members have shown elevated levels of dicentric chromosomes and ring chromosomes17 which are known to be sensitive indicators of radiation exposure.

3. Epidemiological Studies Two major studies of aircrew quantify possible radiological eﬀects on aircrew. The ﬁrst, shown in Table 2, involved 2740 Air Canada pilots from 1950 to 1992 and was compared against the Canadian general population as a base line. Table 3 shows the results of the second study of 1577 female Finnish Airlines Cabin crew, until 1992 compared with the Finnish female population. Table 2.

Air Canada pilot cancer incidence, 1950–1992. Observed

Decrease in “all cancers” Decrease in lung cancers Increase in acute myeloid leukemia Increase in prostrate cancer

125 11 6 34

Expected 176 39 1.3 18

Note: The decreases were dominated by diﬀerent smoking habits from the general population. Source: Ref. 19. Table 3.

Finnish Airlines female crew breast cancer incidence.

Increase in breast cancer

Observed

Expected

20

11

Note: Whilst there was an increase above expectations, social class, reproductive issues and the small sample could have skewed these results (i.e., chance). Source: Refs. 20 and 21.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Ultra Long Range Aircraft Operations and Space Weather

ch23

297

A Norwegian Airline study of 3701 male airline pilots found a signiﬁcant increase in the levels of melanoma and of nonmelanoma skin cancers, compared to the general population. They reported a signiﬁcant increasing trend of malignant melanoma associated with cumulative block hours and with cumulative ﬂight dose of galactic cosmic radiation.18 Nicholas et al.,22 investigated self-reported disease incidence rates in 6533 active and retired male airline pilots in the US and Canada, using the US general population as the control group. The investigation showed pilots had a signiﬁcantly higher incidence of melanoma but had signiﬁcantly lower incidences of prostate cancer, lymphoma, leukemia, bladder cancer, colon cancer, lung cancer, kidney cancer, cancer of the brain and nervous system, and of oral cancer. Reynolds et al.,23 compared cancer incidence in members of the Association of Flight Attendants who resided in California with the incidence of cancer in the general population of the state. Their study indicated a signiﬁcant increase in invasive breast cancer in ﬂight attendants compared with the non-Hispanic white women and with women of all races. They also noted signiﬁcant increases in skin melanoma of the ﬂight attendants. Nicholas et al.,22 in research conducted between 1984 and 1991 found that Proportional Mortality Ratio (PMR) studies showed increase cancer of the kidneys and pelvis. Increased mortality was suggested for prostate cancer, brain cancer, colon cancer and cancer of the lip, buccal cavity and pharynx. Decreased cancer mortality was suggested for stomach cancer and cancer of the trachea, bronchus and lung. In D. Irvine and D. Davies “British Airways Flightdeck Mortality Study 1950–1992”,24 the “all causes” Standardized Mortality Ratio (SMR) for aircrew conﬁrmed the expected “healthy worker eﬀect”. In pilots, apart from aircraft accidental deaths, most SMRs showed signiﬁcant reduction in all causes of mortality. The SMRs for brain cancer and colon cancer were no longer signiﬁcant. The SMR for melanoma was raised in pilots but not for ﬂight engineers and overall life expectancy for aircrew was 4–5 years longer for long haul aircrew above the general British population. Among US commercial pilots and navigators 1984–1991, PMRs suggested increased mortality for prostate, brain and colon cancer but decreased for stomach, trachea, bronchus and lung. The latter could in part be due to changed smoking habits in the general ﬂying population. It was noted that noncancer causes showed signiﬁcantly increased motor neuron diseases and diseases of the nervous and sense organs.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

298

ch23

I. L. Getley and M. L. Duldig

Before moving on to the relative merits of predictive codes or direct monitoring readers are referred to two excellent review articles. Blakely25 reviews the biological eﬀects of cosmic radiation whilst Boice et al.26 review the epidemiological studies published to that date.

4. Predictive Codes Versus Direct Monitoring Various computer models have been developed over the past decade by groups from diﬀerent regions and legislative frameworks, in an attempt to accurately model measured dosimetry at aircraft cruise altitudes. In Europe, the EPCARD 3.1 code has been developed using the Monte Carlo FLUKA transport code.27 In the USA, the FAA developed CARI (currently CARI 6M), based on its own LUIN transport code, that allows for multiple waypoints rather than simple great circle track between departure point and destination.28 More recently a measurement based code entitled PCAIRE (Predictive Code for Aircrew Radiation Exposure) has been developed at the Royal Military College, Canada.29 This code was based on ﬂight dose rate measurements using a TEPC (Tissue Equivalent Proportional Counter) to determine a semi-empirical model that describes absorbed dose rate as a function of position (vertical cut-oﬀ rigidity), altitude (atmospheric depth) and date (heliocentric potential and solar modulation). Similar to PCAIRE, CARI-6M requires minimal input of ﬂight data, namely date of ﬂight, the origin and destination airports, the altitudes and time intervals at those altitudes to give the heliocentric potential and magnetic shielding value.28 Continuing validation of these codes is essential. This will ensure maximum accuracy in all conditions and will be essential at regular, three to six monthly intervals on some routes, particularly with next generation higher altitude ﬂight. It will also indicate where the present codes are inadequate and facilitate further code development. The advent of the TEPC, that simultaneously measures the low LET (ionizing) and high LET (neutron) components, gives a close approximation to the Total Dose equivalent of the mixed radiation environment within an aircraft at cruising altitudes and is ideal for this validation process. Instruments such as the Hawk TEPC, have inbuilt Global Positioning Satellite (GPS) systems to allow time-correlated latitude, longitude and altitude information for comparison against the predictive codes. This instrument takes a measurement every ﬁve minutes and can be downloaded to a laptop for comparison in validation against code.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Ultra Long Range Aircraft Operations and Space Weather

ch23

299

All predictive codes are, by design, able to estimate the exposure to cosmic radiation over a particular ﬂight by evaluation of the latitudes, altitudes and cut-oﬀ rigidities for the time and date of the ﬂight. Unfortunately, it is impossible to predict the time or magnitude of a solar particle event which can increase the radiation dose rate substantially. In October and November 2003, some three years after solar maximum, three signiﬁcant solar particle events over a week included the fourth largest proton event since 1976. A solar particle event in February 1956 produced an increase in background radiation 45 times the normal level at sea level at moderate magnetic latitudes and in September 1989 another event resulted in a ground level event registering a 4.5-fold increase above background. On January 20, 2005, a solar particle event produced a 56-fold increase at South Pole (altitude 2800 m) and 33-fold at McMurdo (sea level) which was the largest event in the space era and occurring at a time approaching the minimum in the 11 years solar-cycle. These events can potentially increase radiation doses signiﬁcantly at jet aircraft cruising altitudes particularly on extended long range ﬂights.

5. Ultra Long Range Flight Generally it can be concluded that whilst background cosmic radiation can add a sizeable increase in ambient dose equivalent in the course of a year to an aircrew member or frequent ﬂyer (approximately 3–5 times that of normal public exposure), there does not appear to be any deﬁnitive results of a noticeable increase in any particular cancers from studies to date. The biggest issue is for that of the pregnant ﬂyer who will be exposed to a signiﬁcant increase in exposure from both cosmic and solar radiation during longer ﬂights. Increasing ﬂight times from 12 to 18+ h will conservatively increase this exposure by 20–30% for cross equatorial ﬂights or mid-latitude ﬂights, i.e., between 35◦ N and 35◦ S. Flights at higher latitude can expect signiﬁcantly higher dose increases due to both the latitude eﬀect and the in-ﬂight time spent at higher altitudes. At geomagnetic latitudes above 40◦ N and 40◦ S, for every 4000 ft increase in altitude above 30 000 ft, the dose rate increases by approximately 30%. Solar events can have signiﬁcant short-term eﬀects that will aﬀect ﬂight operations, communications, navigation (satellites) and biological exposure. Since forecasts for signiﬁcant solar events are inaccurate at present, warnings of such events usually come with or after the onset of such an

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

300

ch23

I. L. Getley and M. L. Duldig

event. This allows a response to the resulting increase from photons and X-rays that aﬀects radio transmissions through ionospheric disturbance as well as an increase in potential biological eﬀects. The energetic particles associated with such events usually arrive within 15–40 min and can last up to a day and put at risk satellites, astronauts and polar ﬂights. Coronal mass ejections (CMEs) are fast plasma streams ejected from the Sun’s surface that impact Earth a minimum of 18 h after eruption but they can take up to several days to arrive. They energize the magnetosphere and ionosphere, aﬀecting electric power grids, navigation systems and radio communications (HF). Due to the location of detecting satellites, warning times of 30 min before CME impact are typical although forecasting of CME arrival is rapidly improving. There is currently little or no warning of a solar particle (proton) event. Communications on polar ﬂight routes between North America and Asia and North America and Europe are susceptible during such solar events. Radio communications on frequencies between 30–300 MHz (VHF) give way to high frequency communications (HF) of 3–30 MHz. These frequencies are highly susceptible to solar interference and whilst SATCOM, satellite communications are considered as the backup, they are rarely available above 82◦ N latitude. GPS satellites can also be aﬀected.

6. Polar Route Flights Of particular concern in long haul ﬂights are the polar ﬂight routes. It is now generally recognized that limits need to be placed on aircrew ﬂying these routes during the course of a year. The unpredictability of solar events and the lack of warning on solar proton events mean that any signiﬁcant event could result in pregnant crew or passengers exceeding the recommended limit of 1 mSv during one of these ﬂights. Airlines ﬂying these routes generally require 4 h to re-plan for a lower latitude ﬂight, as limitations on available freight and passenger loads require oﬀ-load to enable higher fuel uplifts for the longer, lower latitude ﬂight routes. Space weather alerts are now provided by the National Oceanic and Atmospheric Administration, and are provided to airspace users through a website and email alert facility at [email protected].

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Ultra Long Range Aircraft Operations and Space Weather

ch23

301

7. Conclusion The advent of ultra long haul ﬂight operations will reduce travel times by air for passengers but will increase exposure levels to both cosmic and solar radiation, due to time at altitude, over the same ﬂight distances. The increase in dose will vary from 20 to 30% for ﬂights conducted in the lower latitudes and signiﬁcantly higher on polar routes. Above 30 000 ft (9500 m) at high magnetic latitudes, dose rates increase by approximately 30% for each 4000 ft increase in cruise altitude. It has been estimated that a return transpolar ﬂight between New York and South East Asia would result in 0.10 mSv each way and that ﬁve such trips would result in the general population limit of 1 mSv in a year being achieved. Signiﬁcant solar activity would increase this dose.

Acknowledgments The authors acknowledge the assistance of Dr. Wallace Friedberg, Federal Aviation Administration, Civil Aerospace Medical Institute for his help in the section on ionizing radiation; Mr Joseph Kunches and William Murtag for their assistance in provision of data on polar route ﬂights and ﬂight communication limitations; and Qantas Airways, for continued support in the collection of data on high latitude ﬂights.

References 1. 1990 Recommendations of the International Commission on Radiological Protection, Annals of the ICRP 21, Issues 1–3 (1991). 2. D. J. Bird, S. C. Corbat´ o, H. Y. Dai, B. R. Dawson, J. W. Elbert, T. K. Gaisser, K. D. Green, M. A. Huang, D. B. Kieda, S. Ko, C. G. Larsen, E. C. Loh, M. Luo, M. H. Salamon, D. Smith, P. Sokolsky, P. Sommers, T. Stanev, J. K. K. Tang, S. B. Thomas and S. Tilav, Phys. Rev. Lett. 71 (1993) 3401. 3. J. S. Nicholas, K. Copeland, F. E. Duke, W. Friedberg and K. O’Brien, III, Galactic cosmic radiation exposure of pregnant aircrew members II, Federal Aviation Administration, Oﬃce of Aviation Medicine Report DOT/FAA/AM-00/33 (2000), p. 3. 4. IFALPA Human Performance (HUPER) Committee Meeting Texas, October 27–29, 1998. 99HUP047/12 October 1998 — Discussion Paper on Cosmic Radiation (1998). 5. J. G. Wilson, Cosmic Rays (Springer-Verlag, New York, 1976). 6. T. Foelsche, R. B. Mendell, J. W. Wilson and R. R. Adams, Measured and calculated neutron spectra and dose equivalent rates at high altitudes; relevance to SST operations and space research. NASA Report No NASA

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

302

7. 8. 9. 10. 11. 12. 13.

14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29.

ch23

I. L. Getley and M. L. Duldig

TN D-7715, Washington, DC, National Aeronautics and Space Administration (1974), p. 4. K. O’Brien and H. H. Sauer, Technology 7 (2000) 449. W. Friedburg, K. Copeland, F. E. Duke, J. Nicholas, E. B. Darden, Jr. and K. O’Brien, III, Occup. Med. 17 (2002) 293. K. Copeland, H. Sauer and W. Friedburg, Solar radiation alert system, Federal Aviation Administration Report DOT/FAA/AM-05/14 (2005). C. S. Dyer, F. Lei, S. N. Clucas, D. F. Smart and M. A. Shea, Adv. Space Res. 32 (2003) 81. P. Beck, M. Latocha, S. Rollet and G. Stehno, Adv. Space Res. 36 (2005) 1627. W. Friedburg, K. Copeland, F. E. Duke, K. O’Brien, III, and E. B. Darden, Jr., Health Phys. 79 (2000) 591. Exposure of population in the United States and Canada from Natural Background Radiation, NCRP Report No 94, National Council on Radiation Protection and Measurements (1987), p. 8. Limitations of exposure to ionizing radiation, NCRP Report No 116, National Council on Radiation Protection and Measurements (1993), p. 37. J. Chen, B. J. Lewis, L. G. I. Bennett, A. R. Green and B. L. Tracy, Rad. Prot. Dos. 114 (2005) 475. Aviation Health Institute — Cosmic Radiation Seminar London, June 29, 1999, 00HUP029 attachments 1, 2, 3, 4, 5, 6 and 7. A. Heimers, Int. J. Radiat. Biol. 75 (1999) 691. T. Haldorsen, J. B. Reitan and U. Tveten, Scand. J. Work Env. Health 26 (2000) 106. P. R. Band, N. D. Le, R. Fang, M. Deschamps, A. J. Coldman, R. P. Gallagher and J. Moody, Am. J. Epidemiol. 143 (1996) 137. E. Pukkala, A. Auvinen and G. Wahlberg, Brit. Med. J. 311 (1995) 649. E. Lynge, Brit. Med. J. 312 (1996) 253. J. S. Nicholas, G. C. Butler, D. T. Lackland, G. S. Tessier, L. C. Mohr, Jr. and D. G. Hoel, Aviat. Space Environ. Med. 72 (2001) 821. P. Reynolds, J. Cone, M. Layefsky, D. Goldberg and S. Hurley, Cancer Causes & Control 13 (2002) 317. D. Irvine and D. Davies, Aviat. Space Env. Med. 63 (1992) 27679. E. A. Blakely, Health Phys. 79 (2000) 495. J. D. Boice Jr., M. Blettner and A. Auvinen, Health Phys. 79 (2000) 576. S. Roesler, W. Heinrich and H. Schraube, Rad. Prot. Dos. 98 (2002) 367. Radiobiology Research Team http://www.cami.jccbi.gov-CARI 6M program. B. J. Lewis, M. J. McCall, A. R. Green, L. G. I. Bennett, M. Pierre, U. J. Schrewe, K. O’Brien and E. Felsberger, Aircrew exposure from cosmic radiation on commercial airline ﬂights, Rad. Prot. Dos. 93 (2001) 293.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch24

MEASUREMENTS AND PCAIRE FOR MONITORING THE COSMIC RADIATION EXPOSURE OF AIRCREW L. G. I. BENNETT∗ , B. J. LEWIS, A. R. GREEN, A. BUTLER and M. TAKADA Department of Chemistry and Chemical Engineering Royal Military College of Canada, P. O. Box 17000, Stn Forces Kingston, Ont, Canada K7K 7B4 ∗ [email protected] I. L. GETLEY Department of Aviation, University of New South Wales Sydney, 2052, Australia [email protected]

With the fact that aircrew are constantly exposed to galactic cosmic radiation along with recommendations made by the International Commission on Radiological Protection in 1990, aircrew in many countries have been classiﬁed as occupationally exposed to cosmic radiation. Over the past decade, the authors have measured aircrew radiation exposure at jet altitudes with various passive and active radiation monitors. The data have been encapsulated into a program, Predictive Code for Aircrew Radiation Exposure. In both the European Union and Canada, the preferred method for routine assessment of aircrew radiation exposure is by computations supported by periodic measurements. On recent measurement ﬂights over the Paciﬁc basin, various radiation monitors have been used in this manner. A small spectrometer (Liulin) and dosimeter-type detectors (FH 41B and B-10) have been compared to the tissue equivalent proportional counter and a low and high linear energy transfer combination (ionization chamber + smart wide energy neutron detection instrument) with comparable results. Our conversion from absorbed dose to ambient dose equivalent for the Liulin has been validated on these ﬂights and a regional function has been used to correct the small FH 41B-10 dosimeter.

1. Introduction Canada has adopted the 1990 recommendations of International Commission on Radiological Protection (ICRP)-60,1 which include the recognition of aircrew as being occupationally exposed to natural radiation and, in 2001, Transport Canada issued a Commercial and Business Aviation Advisory Circular to Canadian airlines.2 The latter document is consistent with legislation in most European Union (EU) Member States.3 The main points 303

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

304

ch24

L. G. I. Bennett et al.

are that the intervention level is to be 6 mSv/year and that predictive codes may be used supported by periodic monitoring in lieu of personal or aircraft dosimetry. From 1991 to 1998, surveys were conducted by the Royal Military College of Canada (RMC) — ﬁrst on Canadian Forces pilots and then on ﬂight deck and cabin crew of six commercial carriers. The detectors used were neutron-sensitive bubble detectors (BDs), which were carried by volunteer personnel. Simultaneously, several scientiﬁc ﬂights were carried out to measure the mixed radiation ﬁeld at altitude and to determine appropriate conversion factors to estimate the total route ambient dose equivalent from only the neutron dose measured by the bubble detectors. Initially, laboratory-based equipment was brought on board for measurements at altitude in combined passenger/cargo (combi) conﬁgurations of both military and commercial aircraft.4 From 1999 to the present, more portable equipment was procured and ﬂown on both military and commercial passenger aircraft in the cabin. The equipment suite consisted of thermoluminescent detectors (TLDs), bubble detectors, an ionization chamber (IC), a smart wide energy neutron detection instrument (SWENDI), and a tissue equivalent proportional counter (TEPC). The TLDs and IC measure the low linear energy transfer (LET) radiation; the BDs and SWENDI measure the high LET component, while the TEPC, considered the reference instrument,5 measures both. The TLDs and BDs give an estimate of the total accumulated dose for the ﬂight, while the IC, SWENDI, and the TEPC measure every several minutes throughout the ﬂight. Totaling the low and high LET contributions of the appropriate combinations of detectors gave consistent total route ambient dose equivalents and gave conﬁdence in the conversion factors used for the BDs in the surveys. In addition, the TEPC and IC + SWENDI data obtained on these 150 ﬂights around the world were used to establish the Predictive Code for AIrcrew Radiation Exposure (PCAIRE).6 Recently, with the potential employment of a predictive code by air carriers, the suitability of smaller monitors for the periodic conﬁrmation of the code has been studied by the authors. The Liulin 4N spectrometer and the Eberline FH 41B and B-10 Geiger tube based detectors have been ﬂown, speciﬁcally, on routes over the Paciﬁc basin. In this paper, the response of these detectors, which varies with the type of radiation, is compared at various altitudes and latitudes and at diﬀerent periods in the solar cycle to the TEPC and to the IC plus SWENDI combination as well as to three predictive codes, PCAIRE, CARI,7 and EPCARD.8

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch24

Measurements and PCAIRE for Monitoring the Cosmic Radiation Exposure

305

2. Measurement Flights Various combinations of equipment were used to measure the radiation levels on the Qantas Airways and Air Canada ﬂights shown in Fig. 1. On each ﬂight, measurements were obtained every several minutes throughout the ﬂight from the various detectors. For example, the measurement results obtained on an Air Canada polar ﬂight direct from Toronto to Hong Kong with its more southerly return are depicted in Fig. 2, along with the ﬂight altitudes. The short ﬂights at each end of the trip were the approximate one-hour Kingston/Toronto connecting ﬂights, while the interval between the Toronto/Hong Kong and return ﬂights reveals the natural radiation readings at ground level for two of the instruments that were left on. The decreasing nature of the results near the end of the outbound ﬂight reveal the decrease in latitude, while the opposite trend can be seen at the beginning of the return ﬂight to Canada. Integrating the measured dose equivalent rates (as shown in Fig. 2) from take-oﬀ to landing will provide an estimate of the radiation exposure received on a given ﬂight in terms of ambient dose equivalent, H ∗ (10).

Fig. 1. Map of recent measurement ﬂights over the Paciﬁc basin.9 Points of departure and arrival are indicated on the map using the following airport codes: AKL: Auckland, New Zealand; HKG: Hong Kong; HNL: Honolulu, Hawaii, US; ICN: Seoul, Korea; LAX: Los Angeles, California, US; LHR: Heathrow Airport, London, UK; MEL: Melbourne, Australia; PEK: Beijing, China; SIN: Singapore; SYD: Sydney, Australia; YVR: Vancouver, British Columbia, Canada; YYZ: Toronto, Ontario, Canada.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

306

L. G. I. Bennett et al.

10

Toronto to Hong Kong

Hong Kong to Toronto 12000 IC+SWENDI HAWK FH41B

10000

LiuLin Flight Altitude

8000

1

6000

Altitude (m)

Ambient Dose Equivalent Rate (µSv/h)

ch24

4000

2000

0.1

4/18/05 9:00

0 4/18/05 21:00

4/19/05 9:00

4/19/05 21:00

4/20/05 9:00

4/20/05 21:00

Date and Time

Fig. 2. Ambient dose equivalent rates for various detectors and altitude for the Air Canada polar ﬂight from Toronto to Hong Kong and the more southerly return ﬂight.

Table 1 provides a summary of these values measured with the various monitors on the ﬂights depicted in Fig. 1. Comparing these values, it can be seen that those ﬂights covering higher latitudes, both north and south, have higher route doses. In addition to the measured results, exposure levels on these ﬂights were calculated using PCAIRE, CARI, and EPCARD in order to support the use of a predictive code to assess aircrew exposure over a career.2 The measurement results and code predictions all agree within 30% of each other, the typical value expected for radiation measurements. The results from these ﬂights are used to conﬁrm the predictive capability of the semi-empirical PCAIRE code, which covers the full solar cycle eﬀect and can account for a changing magnetic ﬁeld due to the occurrence of geomagnetic storms by impressing a uniform storm ﬁeld on the normal quiet ﬁeld. In this case, the IGRF-95 vertical cutoﬀ rigidity is modiﬁed as a function of the Kp-index. Moreover, it is continually being developed over the solar cycle by such combined experimental and theoretical eﬀorts.

3. Comparison with Smaller Monitors As mentioned previously, the measurements described above allow for an assessment of the suitability of smaller monitors (i.e., the Liulin 4N and the Eberline FH 41B) for the periodic conﬁrmation of predictive code results.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch24

Measurements and PCAIRE for Monitoring the Cosmic Radiation Exposure Table 1.

307

Measured and calculated route ambient dose equivalent, H ∗ (10) (µSv).

Flight route

Date

HAWK

YYZ–HNL HNL–SYD

20-12-02 29-12-02

24 15

Antarctic 1 Antarctic 3 SYD–LAX LAX–SYD AKL–LAX LAX–MEL SIN–LHR LHR–SIN YVR–ICN ICN–YVR LAX–SYD SYD–LAX LAX–AKL SYD–LAX YVR–PEK PEK–YVR YYZ–HKG (polar) HKG–YYZ

31-12-02 09-02-03 11-02-03 04-11-03 21-11-03 23-11-03 02-12-03 05-12-03 13-09-04 17-09-04 28-10-04 03-11-04 05-11-04 07-02-05 28-03-05 02-04-05 18-04-05 20-04-05

IC + FH 41Ba SWENDI Liulin unc. cor. PCAIREb CARIc EPCARDd n/a n/a

n/a n/a n/a n/a

26 (gc) 18 (gc)

24 17

29 17

34 33 22 20 16 23 27 18 44 30 22 21 19 20 34 37 60

28 18 (BD + IC) 39 35 28 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a 35 40 62

n/a n/a n/a 20 16 23 34 19 51 30 26 26 22 n/a 34 n/a 65

31 27 n/a 28 21 30 30 21 n/a n/a 29 27 25 27 n/a n/a 49

38 33 n/a 27 20 30 30 22 n/a n/a 28 26 24 26 n/a n/a 62

38 33 29 (gc) 30 21 32 (gc) 36 (gc) 23 (gc) 52 37 31 27 26 28 32 41 66

34 33 28 25 19 28 32 22 46 29 28 26 24 25 28 32 64

n/a n/a 26 24 20 30 35 25 44 37 25 22 22 23 30 36 57

59

60

56

48

60

60

58

56

a Uncorrected

(unc.) results as obtained directly from the FH 41B or FH 41B-10 have been adjusted using Eqs. (1) and (2) to yield corrected (cor.) results (see text). b Visual PCAIRE (Version 8.0f) calculations were made using the IGRF-95 vertical cutoﬀ rigidity model and the Climax count rate solar modulation model. The notation (gc) indicates that the calculations were made assuming a great circle route. c Eﬀective doses predicted by CARI-6M have been divided by 1.25 to provide H ∗ (10).5 d All EPCARD (Version 3.2) calculations were made assuming a great circle route.

The Liulin 4N is similar to the TEPC in that both are LET spectrometers, so that the energy deposited is quantiﬁed by a multi-channel analyzer (MCA). For the TEPC, the absorbed dose, D, is a function of the lineal energy for a given channel and the number of counts in that channel. The dose equivalent, HTEPC , as seen by the detector, is then computed using a quality factor recommended by ICRP-60, Q(LET) versus q(y).1 The ambient dose equivalent, H ∗ (10), is then determined by proper calibration of the instrument at reference ﬁelds.6 For the Liulin, which should be only a low LET spectrometer due to its silicon detector, no conversion to total ambient dose equivalent was available. However, on previous ﬂights, it was determined that the absorbed doses from both instruments were similar, so that an appropriate method of determining H ∗ (10) was investigated.10 By applying an appropriate quality factor, Q(LET), to each

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

308

ch24

L. G. I. Bennett et al.

of its 256 channels, the ambient dose equivalent can be calculated. To do so, a channel width of 0.5 keV/µm was determined by comparing Liulin data to TEPC data obtained on a ﬂight from Sydney to Johannesburg. Using this channel width, the LET for each channel was determined so that the value of Q associated with each channel could be determined using the Q(LET) relationship recommended by ICRP-60.1 As can be seen in Table 1, the H ∗ (10) results so determined from the Liulin are comparable within 30% to those of the TEPC. This good agreement can be considered a validation of our method of determining the ambient dose equivalent as well as the usefulness of the Liulin for in-ﬂight measurements. The Eberline FH 41B and B-10, being Geiger tube based detectors, might appear to be less useful in a mixed radiation ﬁeld that also increases in neutron content with increasing latitude from the equator. Unlike the TEPC and Liulin spectrometers, which measure an energy spectrum, the FH 41 cannot take into account this changing neutron component automatically. Although it does not detect neutrons directly, it may detect the secondary eﬀects of neutrons. When the neutron component of the ﬁeld is relatively low (such as in the equatorial regions), the FH 41B appears to provide an adequate estimation of H ∗ (10); however, at higher latitudes and increasing neutron component, it would be expected to underestimate H ∗ (10).11 Indeed, this result can be observed by examining the uncorrected integral FH 41B values in Table 1. Those from the trans-equatorial ﬂights (between LAX and either SYD, MEL or AKL) and the ﬂights close to Singapore (SIN) agree very well with the other measurements and predictions. In contrast, the uncorrected FH 41B data from higher latitude ﬂights, in either the northern hemisphere (between YYZ and HKG) or the southern hemisphere (ﬂights to Antarctica), are noticeably lower than the other measurements and code predictions. This trend can also be seen in the detailed data in Fig. 2 where the FH 41B-10 values can be seen to be lower than those from the other instruments at higher latitudes, but comparable to the other instruments at lower latitudes (i.e., closer to Hong Kong on each ﬂight). Based on these observations, it should be possible to correct the FH 41B data to compensate for the varying ﬁeld composition, if one knows the ﬂight path. Comparing the IC + SWENDI results to those from the FH 41B-10 on the ﬂight between Hong Kong and Toronto, a linear correction function, f (Rc ) = −0.0352Rc + 1.3666 ,

(1)

has been derived from plotting the ratio of both detectors’ readings against the vertical cutoﬀ rigidity, Rc , in GV. With knowledge of the ﬂight path,

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch24

Measurements and PCAIRE for Monitoring the Cosmic Radiation Exposure

309

each FH 41B dose equivalent rate data point throughout any ﬂight can then be corrected such that H˙ corrected = f (Rc ) · H˙ FH 41B .

(2)

Note that the correction function will be exactly 1 at an Rc value of 10.4 GV and slightly less than 1 for higher Rc values. As a result, when the correction function is applied to trans-equatorial ﬂights (where the Rc can reach a maximum of approximately 17 GV), the corrected number will be slightly lower than the uncorrected value. For ﬂights nearer the poles (where the Rc approaches 0 GV), the corrected value will be a maximum of 37% higher than the uncorrected value. The beneﬁts of this linear function can be seen by comparing the uncorrected and corrected values for the FH 41B-10 shown in Table 1. However, except for those ﬂights encompassing higher latitudes, the correction remains within the error margin of 30% for the measurements. In fact, for the trans-equatorial ﬂights noted above, there is little, if any, diﬀerence between the corrected and uncorrected values. Further ﬂights to cover a wider range of altitudes and Rc values will be undertaken to establish this preliminary function. Use of such a regional function thus could allow use of a small dosimeter-type monitor that is not a proper spectrometer (like the TEPC or Liulin) for normal, periodic measurements.

4. Conclusions Measurements have been conducted on commercial ﬂight routes over the Paciﬁc basin by a variety of radiation monitors, from spectrometers (TEPC and Liulin) and low and high LET combinations (IC + SWENDI) to dosimeter-type detectors (FH 41B and B-10), with comparable results. By comparison to the TEPC and the IC + SWENDI combination, the conversion from the absorbed dose to ambient dose equivalent for the Liulin has been validated on these ﬂights. A regional function has been used to correct the small FH 41B-10 dosimeter, which indicates that such a smaller and readily available dosimeter could also be used to measure aircrew radiation exposure during normal conditions. However, during large solar proton events, further study will be required. These various detectors will be needed to support the use of a predictive code for career assessment of aircrew exposure.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

310

ch24

L. G. I. Bennett et al.

Acknowledgments The support of the Air Canada and Qantas Airways in allowing the measurements to be made on board for this study is gratefully acknowledged. The ﬁnancial support of Transport Canada and the Canadian Air Force is acknowledged.

References 1. International Commission on Radiological Protection, 1990 Recommendations of the International Commission on Radiological Protection, ICRP Publication 60 (Pergamon Press, Oxford, 1991). 2. Transport Canada, Measures for Managing Exposure to Cosmic Radiation of Employees Working on Board Aircraft, Commercial and Business Aviation Advisory Circular No. 0183, April 5, 2001, www.tc.gc.ca/CivilAviation/ commerce/circulars/AC0183.htm. 3. J.-M. Courades, Radiat. Prot. Dosim. 86 (1999) 7. 4. P. T. Tume, B. J. Lewis, L. G. I. Bennett and T. Cousins, NIM Phys. Res. A 406 (1998) 153. 5. G. C. Taylor, R. D. Bentley, T. J. Conroy, R. Hunter, J. B. L. Jones, A. Pond and D. J. Thomas, Radiat. Prot. Dosim. 99 (2002) 435. 6. B. J. Lewis, M. Desormeaux, A. R. Green, L. G. I. Bennett, A. Butler, M. McCall and J. D. Saez Vergara, Radiat. Prot. Dosim. 111 (2004) 151. 7. W. Friedberg, F. E. Duke, L. Snyder, K. Copeland, K. O’Brien, D. E. Parker, M. A. Shea and D. F. Smart, CARI-6M, Civil Aerospace Medical Institute, Federal Aviation Administration, Oklahoma City, OK, USA (2001). 8. H. Schraube, G. Leuthold, W. Heinrich, S. Roesler, V. Mares and G. Schraube, EPCARD (European Program Package for the Calculation of Aviation Route Doses), GSF National Research Centre for Environment and Health (2002). 9. K. L. Schwartz, Great circle mapper, http://gc.kls2.com (accessed June 7, 2005). 10. A. R. Green, L. G. I. Bennett, B. J. Lewis, F. Kitching, M. J. McCall, M. Desormeaux and A. Butler, Adv. Space Res. 36 (2005) 1618. 11. I. L. Getley, A. R. Green, L. G. I. Bennett, B. J. Lewis and M. L. Duldig, Adv. Geosci. 2 (2006).

May 30, 2006 16:12 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch25

MEASUREMENT AND MODELING OF HIGH LATITUDE FLIGHTS IN THE SOUTHERN HEMISPHERE I. L. GETLEY Department of Aviation, University of New South Wales Sydney, NSW 2052, Australia [email protected] A. R. GREEN∗ , L. G. I. BENNETT† and B. J. LEWIS‡ Department of Chemistry and Chemical Engineering Royal Military College of Canada, P. O. Box 17000 Stn Forces, Kingston Ontario, Canada K7K 7B4 ∗ [email protected] † [email protected] ‡ [email protected] M. L. DULDIG Department of the Environment and Heritage Australian Antarctic Division, Kingston, Tas. 7050, Australia [email protected]

During the southern summers of 2003/2004 and 2004/2005, radiation monitoring was carried out on numerous scheduled ﬂights covering high latitudes in the southern hemisphere. Two small, commercially available monitors of diﬀerent design and potential ability to measure the radiation spectrum at altitude were employed to determine their usefulness as small, portable detectors on ﬂights. The Liulin-4N LET spectrometer has a silicon semiconductor-based PIN diode detector while the Eberline FH 41B-10 has a Geiger tube. They were ﬂown with a “HAWK” tissue equivalent proportional counter (TEPC) as a reference instrument for comparison. The two monitors gave results expected within experimental tolerances for monitoring of mixed ﬁelds when compared to the TEPC and to two exposure codes, PCAIRE and CARI.

1. Introduction The principal ionizing radiation present at commercial jet altitudes (nominally, 35 000 feet), results from the interaction between galactic cosmic radiation (GCR) and the Earth’s atmosphere. The GCR consists primarily of charged particles (mainly protons, some alpha particles and a few heavier nuclei).1 The GCR particles are aﬀected ﬁrst by their magnetic rigidity resulting in a latitude and altitude dependant dose equivalent rate 311

May 30, 2006 16:12 WSPC/SPI-B368 Advances in Geosciences Vol. 2

312

ch25

I. L. Getley et al.

that varies over a solar cycle by up to 40% at latitudes above 65◦ and altitudes above 31 000 feet. The GCR is maximum during solar minimum as a result of the decrease in the solar modulation. Second, this primary GCR is aﬀected by the Earth’s magnetic ﬁeld such that particles entering at the poles experience little deﬂection and those entering the atmosphere near the equator experience signiﬁcant deﬂection depending upon their rigidity value compared to the geomagnetic cut-oﬀ rigidity at the equator which can vary with geomagnetic disturbance (e.g., as measured by the Kp index). Third, against the increase in secondary radiation with passage through the atmosphere is the competing eﬀect of atmospheric absorption and production of secondary radiation, such that the dose rate also varies with altitude. Thus, aircrew radiation exposure from GCR is a function of solar cycle period (date), geomagnetic latitude and the ﬂight level or altitude.

2. Research Task, Equipment, and Codes While many measurements of the jet-altitude radiation ﬁeld have been made and compared to exposure software codes, most of the published studies have involved ﬂights at equatorial and northern hemisphere latitudes. This study aims to measure radiation levels (with a “HAWK” TEPC, a Liulin4N, and an Eberline FH 41B) speciﬁcally on southern hemisphere ﬂights and to compare the output of two exposure codes (CARI and PCAIRE) to these measurements. A brief description of the instruments and codes is provided below. Both the “HAWK” TEPC and the Liulin-4N are designed to measure mixed radiation ﬁelds (such as that present at altitude). Both use a multichannel analyzer (MCA) to analyze the energy deposited in their respective detectors to obtain a lineal energy or linear energy transfer (LET) spectrum, respectively. The TEPC has 12.7 cm diameter propane-ﬁlled proportional counter with a 2.13-mm thick polymer skin to simulate tissue, whereas the Liulin-4N has a silicon semiconductor-based PIN diode detector. In both instruments, the absorbed dose, D, is recorded as a function of the energy deposited from both high- and low-LET radiation. The ambient dose equivalent, H*(10), can be obtained from the TEPC data by use of a quality factor, Q(LET), as recommended by ICRP-60,2 and an applied calibration factor.3 In previous research, it was determined that the absorbed dose from both instruments are in agreement, so that methods to convert the Liulin absorbed dose to the dose equivalent were investigated.4 A full spectral analysis similar to that used in the TEPC was employed.

May 30, 2006 16:12 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Measurement and Modeling of High Latitude Flights

ch25

313

In contrast, the Eberline FH 41B and B-10 dosimeters speciﬁcally measure photons and charged particles, due to their Geiger tubes with diﬀerent ﬁlters for the B (Hx photon equivalent) and B-10 (ambient dose equivalent), calibrated to Cs-137 = 662 keV.5 Whilst the FH 41B-10 does not measure the neutron component directly [i.e., the B10 (n, α) reaction], it measures the neutron progeny created in the form of charged particles, but with low eﬃciency. The CARI code6 is a Civil Aerospace Medical Institute (CAMI) program developed by the Federal Aviation Administration (FAA) using the theoretically-based LUIN code7 to provide a deterministic radiation dose over a great circle or waypoint designated track. The CARI code outputs results in terms of an eﬀective dose, E, which can be converted to ambient dose equivalent by using the approximate relationship [E/1.25 = H*(10)].8 This relationship of E/H*(10) is a function of altitude and latitude varying from 1.2 to 1.5.9 The PCAIRE is a semi-empirical model developed using data sets from near solar maximum and solar minimum, normalized to an altitude of 10.7 km.3 These data are plotted against the vertical cut-oﬀ rigidity, Rc , using the International Geomagnetic Reference Field, IGRF-1995, to Fig. 1. Flight routes ﬂown during the course of cosmic radiation measurements at high southern latitudes allow for a correlation between the ambient dose equivalent rate, H*(10) in µSv h−1 , and global position, altitude and period in the solar cycle.

Fig. 1. Flight routes ﬂown during the course of cosmic radiation measurements at high southern latitudes. The color coding shows the ﬂight pairings in both directions with the lower latitudes being the return ﬂights to Australia. The nonsymmetrical ﬂight routes result from ﬂight paths following upper wind currents or jet streams on the return legs.

May 30, 2006 16:12 WSPC/SPI-B368 Advances in Geosciences Vol. 2

314

ch25

I. L. Getley et al.

(a)

(b) Fig. 2. Comparison of measured and predicted H*(10) rates on ﬂights from Sydney to Johannesburg reaching maximum latitudes of (a) 62 and (b) 65◦ S. The Eberline FH 41B data are integrated over six-minute intervals; the TEPC data over 30-minute intervals; and the Liulin data over 60-minute intervals.

May 30, 2006 16:12 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Measurement and Modeling of High Latitude Flights

ch25

315

3. Inﬂight Measurements Through the southern summers of 2003/2004 and 2004/2005, a number of ﬂights were conducted between Sydney, Australia (SYD) and Johannesburg, South Africa (JBG) (Fig. 1). These summertime ﬂights were chosen based on the higher altitudes and more southerly latitudes that could be achieved in the warmer months. Even during these months, ﬂights at 65◦ S latitude and altitudes of 35 000 feet and above reach outside temperatures close to minus 70◦ C, the limiting temperature below which fuel freezing starts to occur in modern jet aircraft. The winter months would require signiﬁcantly lower latitudes, primarily due to the stronger upper jet streams and lower altitudes to avoid freezing temperatures. The highest latitude ﬂights at 65◦ S geographic latitude result in ﬂights achieving geomagnetic latitudes close to 80◦ S magnetic, since the geomagnetic south pole is located to the south of Australia. The H*(10) rates obtained from the three instruments throughout two SYD–JBG ﬂights reaching maximum latitudes of 62 and 65◦ S are compared to calculations made with both PCAIRE and CARI in Fig. 2. All of the instruments and calculations agree within the expected 30% error; however, the data from the FH 41B appear to be lower than the other measurements and code calculations as the ﬂights approach higher latitudes (with the highest latitude reached approximately mid-ﬂight). This discrepancy arises as a result of the varying composition of the radiation ﬁeld with latitude. Based on previous measurements, the neutron contribution to H*(10) will be approximately 45% on high-latitude ﬂights, compared to approximately 35% on ﬂights approaching or crossing the equator.3 On these latter type of ﬂights (lower latitudes), the FH 41B appears to provide an adequate estimation of H*(10); however, since the FH 41B cannot directly detect neutrons, it begins to under-respond as the neutron component begins to increase with higher latitude. Based on a comparison of the FH 41B to other instruments that are capable of measuring the complete low- and high-LET contributions of the mixed ﬁeld over a range of Rc values, a correction factor has been derived as a function of Rc . The function can be used to adjust the FH 41B data for other ﬂights as has been done in Table 1.10 In Table 1 is shown a summary of the results from the various instruments and from the two codes for all four of the return ﬂight pairings. As seen in Table 1, the integrated values of the route dose equivalents from the three instruments and two exposure codes agree within 30%, typical of errors found in other studies such as EURADOS. Along with

Date

Highest latitude (◦ S)

Maximum altitude at highest latitude (ft)

HAWK

Liulin

FH41(a)

PCAIRE(b)

CARI(c)

08-Nov-03 11-Nov-03 03-Jan-04 05-Jan-04 01-Dec-04 05-Dec-04 15-Dec-04 19-Dec-04

62 49 65 54 56 47 65 45

35 000 37 000 35 000 37 000 35 000 37 000 35 000 37 000

42 37 47 40 60 43 n/a 46

40 35 47 42 62 48 60 49

51 43 54 45 57 47 56 47

51 43 52 44 61 51 62 51

43 36 48 40 55 45 54 45

Ambient dose equivalent, H*(10) (µSv)

(a) Data from the FH 41B-10 have been corrected using an Rc -dependent function. (b) Visual PCAIRE (version 8.0f) calculations were made using the IGRF-95 vertical cutoﬀ rigidity model with solar modulation derived from the Climax neutron monitor count rate. (c) Eﬀective doses predicted by CARI-6M have been divided by 1.25 to provide H*(10).

I. L. Getley et al.

SYD–JBG JBG–SYD SYD–JBG JBG–SYD SYD–JBG JBG–SYD SYD–JBG JBG–SYD

Flight data showing highest altitude at highest latitude ﬂown.

May 30, 2006 16:12 WSPC/SPI-B368 Advances in Geosciences Vol. 2

316

Flight route

Table 1.

ch25

May 30, 2006 16:12 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch25

Measurement and Modeling of High Latitude Flights

317

this expected error in measurements, the diﬀerences in the results for each ﬂight are due to the diﬀerent latitudes, altitudes, time aloft, and time in the solar cycle. In general, the route-integrated H*(10) value is higher on the SYD–JBG ﬂight than on the corresponding JBG–SYD return ﬂight due mainly to a much longer ﬂight time (approximately 13 h versus 10 or 11 h). Looking at the earlier data (November 2003 and January 2004) to the later data (December 2004), there is an approximate 15–20% increase in the route-integrated H*(10) values for the same route at similar latitudes and altitudes. As seen in Fig. 3, this increase over one year can be explained by the solar cycle eﬀect. In this ﬁgure, the approximate 11-year periodic variation in the solar cycle can be seen along with the anticoincidence nature of the sun’s activity (expressed as sunspot number) with the radiation on Earth (Climax hourly count rate). The increase in the radiation ﬁeld as solar sunspot activity decreases between the two sets of ﬂights is noted in the ﬁgure. Thus, as solar minimum is approached within the next two years, a reduced shielding eﬀect from cosmic radiation will mean increased radiation. Comparing the three detectors, the results from the two spectrometers (the TEPC and the Liulin) are in good agreement, as are the results from

4500

Sunspot Number

350

4000

300 3500

250 200

03/04 Flights

150

04/05 Flights

3000 2500

100 50

2000 19

20

21

22

23

Climax Hourly Count Rate /100

400

0 1500 1953 1958 1963 1968 1973 1978 1983 1988 1993 1998 2003 2008 Year

Fig. 3. Plot of Climax neutron monitor count rate and sunspot number versus date. The lower curve (black solid line) shows the number of sunspots per month (left-hand axis) over the past ﬁve solar cycles,11 with the dashed line showing the predicted number of sunspots per month until the end of the current solar cycle.12 The upper (gray) curve shows the monthly average of the hourly count rate from the Climax ground-based neutron monitor (right-hand axis).13

May 30, 2006 16:12 WSPC/SPI-B368 Advances in Geosciences Vol. 2

318

ch25

I. L. Getley et al.

the FH 41 detectors, with the applied correction.8 Note that, although the correction function is only based on a set of ﬂights in the same region as these ﬂights, it is expected that it will be valid for other regions. The two code calculations are also in good agreement with the measurements.

4. Conclusions Various radiation monitors were used to measure the cosmic radiation exposure on scheduled routes at high latitudes in the southern hemisphere. As previously observed, the Liulin spectrometer has given results similar to those of the TEPC. With an appropriate correction function, the Geiger tube-based FH 41B-10 has given an adequate representation of the dose equivalent for these ﬂights. Both the PCAIRE and CARI codes also show results in agreement with the measurements. The diﬀerences in the conditions for each ﬂight (latitude, altitude, ﬂight times, and solar cycle) are evident in the results.

Acknowledgments The authors would like to thank Qantas Airways for allowing Capt. Getley to ﬂy the equipment and the Royal Military College for the loan of equipment. The continued support of Jason Middleton, Department of Aviation, University of New South Wales is recognized. The ﬁnancial support of Transport Canada and the Canadian Air Force is acknowledged.

References 1. W. Heinrich, S. Roesler and H. Schraube, Radiat. Prot. Dosim. 86 (1999) 253. 2. International Commission on Radiological Protection, 1990 Recommendations of the International Commission on Radiological Protection, ICRP Publication 60 (Pergamon Press, Oxford, 1991). 3. B. J. Lewis, M. Desormeaux, A. R. Green, L. G. I. Bennett, A. Butler, M. McCall and J. D. Saez Vergara, Radiat. Prot. Dosim. 111 (2004) 151. 4. A. R. Green, L. G. I. Bennett, B. J. Lewis, F. Kitching, M. J. McCall, M. Desormeaux and A. Butler, Adv. Space Res. (2005), in press. 5. I. L. Getley, Space Weather 2 (2004) S05002, doi:10.1029/2003SW000058. 6. W. Friedberg, F. E. Duke, L. Snyder, K. Copeland, K. O’Brien, D. E. Parker, M. A. Shea and D. F. Smart, CARI-6M, Civil Aerospace Medical Institute, Federal Aviation Administration, Oklahoma City, OK, USA (2001).

May 30, 2006 16:12 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Measurement and Modeling of High Latitude Flights

ch25

319

7. K. O’Brien, LUIN, a code for the calculation of cosmic ray propagation in the atmosphere (update of HASL-275), Report EML-338, US Department of Energy, New York (1978). 8. G. C. Taylor, R. D. Bentley, T. J. Conroy, R. Hunter, J. B. L. Jones, A. Pond and D. J. Thomas, Radiat. Prot. Dosim. 99 (2002) 435. 9. D. T. Bartlett, Radiat. Prot. Dosim. 86 (1999) 263. 10. L. G. I. Bennett, B. J. Lewis, A. R. Green, M. J. McCall, A. Butler, M. Desormeaux, F. Kitching, I. L. Getley and M. Takada, Adv. Geosci. (2006). 11. Monthly-averages of the International Sunspot Numbers are available at http://www.ssl.msfc.nasa.gov/ssl/pad/solar/greenwch/spot num.txt. 12. Predicted Sunspot Numbers are compiled by the United States National Oceanic and Atmospheric Administration (NOAA), Space Environment Center (SEC) and are available at http://www.sec.noaa.gov/ftpdir/weekly/ Predict.txt, accessed March 9, 2004. 13. Monthly-averaged Climax counting rate data obtained from the Solar and Upper Atmosphere Group, National Geophysical Data Centre, NOAA, Boulder, Colorado at ftp://ftp.ngdc.noaa.gov/STP/SOLAR DATA/ COSMIC RAYS/climax.tab.

This page intentionally left blank

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch26

LINK BETWEEN COSMIC RAYS AND CLOUDS ON DIFFERENT TIME SCALES ILYA G. USOSKIN∗ and GENNADY A. KOVALTSOV† Sodankyl¨ a Geophysical Observatory (Oulu unit) University of Oulu, P. O. Box 3000, FIN-90014, Finland ∗Ilya.Usoskin@oulu.ﬁ

A possible mechanism of solar variability inﬂuence upon the Earth’s climate is related to a link between the cosmic ray ﬂux and cloudiness. Here we review evidences relating terrestrial climate variability to changes of cosmic ray ﬂux in the Earth’s vicinity on diﬀerent time scales. On daily scales, major Forbush decreases and solar energetic particle events can aﬀect the cyclogenesis in sub-polar regions. At inter-annual scales, a signiﬁcant correlation between low clouds and cosmic ray induced ionization has been found. Diﬀerent climate reconstructions depict a correlation with variations of the geomagnetic ﬁeld intensity throughout the last millennia, providing additional support to a systematic eﬀect of cosmic rays. On very long time scales, a close relation was reported between the global climate and variations of cosmic ray ﬂux expected from local galactic environment changes. Although none of these facts alone is conclusive, in the aggregate they strongly support the link between cosmic rays and climate on Earth. These links are based on phenomenological relations, and theoretical development and experimental investigation of this hypothesis is ongoing.

1. Introduction The Earth climate is ultimately driven by solar irradiance received by the terrestrial system. However, the detailed process of long-term climatic changes is not yet understood. The most direct mechanism is related to total solar irradiance (TSI) variations caused by variable solar magnetic activity. However, direct measurements of TSI during the last decades show that, while variations of TSI are closely related to the solar activity, their magnitude is too small to explain the climate variations.1–3 Diﬀerent solutions to the problem are discussed (see, e.g., a review in Ref. 4) such as a long-term trend in the irradiance,5–7 a terrestrial ampliﬁer of the irradiance variations,8,9 or a concurrent mechanism which is also driven by the solar activity. Cosmic rays (CR) are a good candidate for the latter option †Permanently

at Ioﬀe Physical-Technical Institute, St. Petersburg, Russia. 321

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2 ch26

322

I. G. Usoskin and G. A. Kovaltsov

(see, e.g., Refs. 10 and 11). Interaction of CR with the terrestrial atmosphere may aﬀect cloud formation and thus modify the terrestrial energy balance. Even a small change in cloud cover shifts the balance between albedo and transmission of the atmosphere at diﬀerent wavelengths. This strongly aﬀects the amount of absorbed radiation, and therefore, climate, without notable changes in the solar irradiance. The ﬂux of CR is modulated in the heliosphere thus providing a link to the solar magnetic activity. The CR as a possible climate driver is a topic of high interest nowadays, and quite a number of papers have been published recently discussing different aspects of this relation. Signiﬁcant progress has been made during the last few years. Here we aim to review numerous results studying the CR–climate link, trying to highlight those which can be directly associated to CR rather than to solar irradiance variations. We highlight the problem from the point of view of a cosmic ray physicist. In Sec. 2, a brief description of possible mechanisms linking CR to cloud formation is presented. In Sec. 3, we review the empirical relations between CR and climate on diﬀerent time scales. Conclusions are summarized in Sec. 4.

2. Possible Mechanisms The amount of energy brought by CR into the terrestrial system is negligible compared to solar radiation, but their presence in the atmosphere is important since CR form the main source of ionization in the troposphere and lower stratosphere. Thus CR aﬀect the chemical–physical conditions of the atmosphere and may inﬂuence the ability of the terrestrial system to absorb/trap/reﬂect solar radiation through, e.g., cloud cover. Clouds play an important role in the radiation budget of the atmosphere by both trapping outgoing long wave radiation and reﬂecting incoming solar radiation. Although these two processes have opposite signs, the net eﬀect of cloudiness is cooling. Therefore, CR act as a trigger so that even a small input variation can produce a strong eﬀect via controlling the atmospheric transparency. However, the details of this seemingly simple scenario are as yet far from being completely understood. Two main mechanisms of CR aﬀecting clouds are discussed in the literature (see, e.g., reviews in Refs. 12 and 13). One is based on the cosmic ray induced ionization (CRII) of the atmosphere.14–16 Ions created by CR rapidly interact with molecules in the atmosphere and are converted into complex cluster ions (aerosols), which

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch26

Link Between Cosmic Rays and Clouds

323

may grow by ion–ion recombination or ion–aerosol attachment and thus aﬀect the number of aerosols acting as cloud condensation nuclei. Another mechanism proposed by Tinsley17,18 employs interaction between the electric ﬁeld and cloud formation. The CRII controls the atmospheric conductivity, while the same processes which modulate CR (interplanetary magnetic ﬁeld, solar wind, interplanetary shocks, etc.) aﬀect the state of Earth’s magnetosphere. Both mechanisms aﬀect the global electrical circuit, which can modify the precipitation and ice formation in super-cooled water. Also, electroscavenging includes dynamical eﬀects on storm systems. Also alternative mechanisms may aﬀect clouds without a direct inﬂuence of CR. For example, cloud variations can be a result of circulation changes due to stratospheric heating caused by the ozone absorption of solar UV radiation.8,19 Furthermore, such changes may lead to changes in winter circulation patterns that aﬀect middle latitude storm tracks.8

3. Cosmic Rays Versus Climate on Diﬀerent Time Scales Variations of CR are caused by diﬀerent mechanisms on diﬀerent time scales.25 In the following subsections we consider them separately. 3.1. Daily scales Regular variations of CR ﬂux depict a diurnal cycle at the level of a few percent due to the local CR anisotropy. We do not consider it here since this diurnal variation cannot be distinguished in the atmospheric data because of the day–night eﬀect. Other CR variations on daily time scale are sporadic. Interplanetary transient phenomena such as, e.g., interplanetary shocks, can suppress the ﬂux of GCR by tens of percent over a few hours, with the subsequent recovery taking several days. This phenomenon is known as a Forbush decrease. On the other hand, strong solar ﬂares or CME-driven shocks can accelerate solar/interplanetary particles leading to a strong increase of cosmic ray ﬂux at the Earth’s orbit called a solar energetic particle (SEP) event. Typical proﬁles of a Forbush decrease and a SEP event are shown in Figs. 1(a) and 1(b), respectively. Many statistical studies have been performed looking for a relation between sporadic CR variations and atmospheric characteristics. Pudovkin and Veretenenko26 reported some reduction of the mean cloud cover after Forbush decreases at high latitudes (> 60◦ N). Roldugin and Tinsley27

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

I. G. Usoskin and G. A. Kovaltsov

(a)

100 90 80 302

304

306

308

310

312

314

Oulu NM count rate [%]

Oulu NM count rate [%]

324

400

(b) 300 200 100 0

2003, Day of year Oulu NM count rate [%]

ch26

6

12

18

24

20 January 2005, hours

100

(c)

90 80 70 1965

1970

1975

1980

1985

1990

1995

2000

2005

-1

1.8

Q14C [atc m s ]

2.0

-2

Year (d)

1.6 1.4 1.2 -10000

-8000

-6000

-4000

-2000

0

2000

10

1.5 1.6

(e)

(f ) Φ(t)/ Φ0

Be Production [arb.u.]

Years-BC/AD

1.4

1.0

1.2

0.5 1.0 0.8 0

50

100

Age [ky]

150

200

0.0 0

100

200

300

400

500

Age [My]

Fig. 1. Cosmic ray variations on diﬀerent time scales: (a) Forbush decrease recorded by Oulu NM; (b) SEP event recorded by Oulu NM; (c) inter-annual variations recorded by Oulu NM; (d) millennial variations of the 14 C production rate (after Refs. 20 and 21); (e) multi-millennial variations of the 10 Be production rate (after Refs. 22 and 23); and (f) model simulation of the galactic cosmic ray density variations due to the galactic spiral arm crossings (after Ref. 24).

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch26

Link Between Cosmic Rays and Clouds

325

found changes in the atmospheric transparency associated with Forbush decreases at high latitudes (> 55◦ N). Kniveton and Tinsley28,29 and Todd and Kniveton30 found zonal mean total cloud anomalies associated with Forbush decreases, particularly in polar and equatorial regions. Stozhkov et al.31,32 reported observed changes in precipitation related to Forbush decreases and SEP events. On the other hand, Tinsley et al.13,33,34 suggested (later conﬁrmed in Ref. 35) that vorticity in polar/subpolar regions can be aﬀected by CR during the cold season — both reduction33 after Forbush decreases and increase35 during/after solar particle events were reported (Fig. 2). In summary, there are hints for an eﬀect of CR on the cloudiness/transparency/cyclogenesis, particularly in high latitude regions during cold seasons, but the results are so far not robust. The primary eﬀect of CR may be related to vorticity/cyclogenesis.

3.2. Inter-annual/decadal variations

8x10

-9

6x10

-9

4x10

-9

2x10

-9

0 -15

(a)

-10

-5

0

5

Days from key date

10

15

Vorticity Area Index [103 km 2]

Local sum of ζ2

Temporal variations of CR are dominated by the 11-year cyclicity related to the solar magnetic activity cycle (see Fig. 1(c)). Similar 11-year cycle has been reported in the global low cloud coverage, in association with CR ﬂux, by Svensmark and Friis-Christensen36,37 and developed by Marsh and Svensmark in a series of papers.15,38 This result initiated a dispute in the literature with both for39–42 and contra43–46 arguments. However, recent studies47–49 show that not only temporal but also latitudinal distribution of low clouds closely follow CRII variations (Fig. 3). Low cloud

(b)

37 36 35 34 33 32 -10

-5

0

5

10

15

20

Days from key date

Fig. 2. Superposed epoch changes of the vorticity index. (a) Squared relative vorticity in the North Atlantic region associated with solar particle events (key date) (after Ref. 35) and (b) vorticity area index in the Northern hemisphere associated with Forbush decreases (key date) (after Ref. 33).

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

326

ch26

I. G. Usoskin and G. A. Kovaltsov

Normalized variation [%]

10

8

6

4

2

0 -60

-30

0

30

60

Latitude [deg] Fig. 3. Latitudinal proﬁles of the relative variations of the measured low cloud amount (dots) and computed CRII (line) (after Ref. 47).

(a)

(b)

(c)

Fig. 4. Time proﬁles of detrended low cloud amount in percent of the area coverage (solid symbols, left axis) and CRII (open symbols, right axis) for (a) the global average (60◦ S < λ < 70◦ N), (b) tropics (|λ| < 25◦ ), and (c) mid-latitudes (60◦ S < λ < 25◦ S and 25◦ N < λ < 70◦ N) (after Ref. 47).

amount time series can be decomposed into a long-term slow trend and inter-annual variations, the latter depicting very signiﬁcant correlation with CRII over the globe (see Fig. 4). A quantitative regression model has been suggested47 with a nearly one-to-one relation between the relative variations of cloud amount and CRII. These results support the idea that low cloud amount is modulated by CRII at inter-annual timescales between 1984 and 2000. On the other hand, high clouds show anti-correlation with CRII and middle clouds no apparent correlation, while the total cloud cover

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Link Between Cosmic Rays and Clouds

ch26

327

shows only marginal correlation.50 Pall´e50 proposed that low clouds can be partly masked by high clouds in the satellite data set. It seems more likely that other mechanisms, e.g., via the global current system17 or UV heating8 which work in anti-phase with CR variations, may dominate at higher altitudes. A careful study including detailed modeling is needed to disentangle diﬀerent eﬀects at diﬀerent altitudes. In summary, the link between low clouds and CRII looks quite reliable on the inter-annual time scale after 1983, including also the latitudinal/geographical pattern, however, a more detailed study is needed to understand the relation with other cloud types. 3.3. Centennial to millennial time scales Variations of cosmic ray ﬂux are deﬁned mostly by solar activity changes on the centennial time scale. On longer time scales (Fig. 1(d)), geomagnetic ﬁeld changes become increasingly important and start dominating CR variations on time scales longer than several millennia (Fig. 1(e)). A detailed study51 of a possible link between CR and cloudiness was performed using sunshine observations during the 20th century. Although the data are not easy to interpret and analyze, they concluded that a link between total cloud cover and CR is unlikely but the data are in general agreement with the hypothesis of a link between low clouds and CR. There are numerous correlations between solar activity and climatic proxies (e.g., δ 18 O or drift ice debris52 ) during the Holocene, which conﬁrm the link between solar activity and climate. However, such studies cannot distinguish between CR and other solar activity driven eﬀects, e.g., solar irradiance. In order to study explicit CR eﬀects one needs to look for changing CR ﬂux unrelated to solar activity, such as geomagnetic ﬁeld variations and changes of the local galactic environment. On the multi-millennial time scale it was found that periods of geomagnetic ﬁeld reversal roughly correspond to cold episodes of the paleoclimatic reconstructions,53–55 although this correlation is not strong.56 A detailed study57 has revealed a weak but persistent correlation between Northern hemisphere temperature and the geomagnetic ﬁeld intensity during the last millennium, implying that CR play a role in climate variations. 3.4. Geological time scales On the geological time scales (longer than a million years) CR variations are determined by the local galactic environment. It is expected that the

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

328

ch26

I. G. Usoskin and G. A. Kovaltsov

density of CR is higher when the Earth is inside dense galactic spiral arms. Shaviv and Veizer24,58–60 reported a similarity between paleoclimatic reconstructions and variations of CR ﬂux due to the modeled galactic spiral arms crossings, within the uncertainties of the latter (Fig. 5). This result has been both disputed61 and supported62 by other researchers. Note that an interpretation of this correlation is not straightforward. In particular, it assumes the constancy of other drivers and the type of climate throughout millions of years. However, e.g., galactic dust, which is abundant in galactic spiral arms, may lead to cooling of the climate during the spiral arm crossing,63 in synchronization with the CR eﬀect. The rate of geomagnetic ﬁeld reversals also varies on mega-year scale64,65 quite synchronously with the climatic variation (see Fig. 5). This itself modulates the CR ﬂux impinging on the Earth also in synchronization with the CR eﬀect due to spiral arm crossing. The corresponding geological processes, leading, e.g., to dust/smoke loading into the atmosphere or changing its physical–chemical characteristics, may also directly aﬀect the climate.

Φ(t)/ Φ

o

1.5

(a) 1.0

0.0 2 1 0 -1 -2 -3

(b)

-1

Relative rate [My ]

o

∆T[ C]

0.5

4

(c)

3 2 1 0 0

100

200

300

400

500

Age [My] Fig. 5. Very long time variations of: (a) CR ﬂux (after Ref. 24), the same as Fig. 1(f); (b) paleoclimatic reconstruction (after Ref. 24); and (c) the rate of geomagnetic ﬁeld reversal per million years (after Refs. 64 and 65).

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Link Between Cosmic Rays and Clouds

ch26

329

4. Conclusions We have reviewed recent results and evidence linking cosmic ray ﬂux to terrestrial climate. The results can be summarized as follows: There are numerous hints of an instantaneous relation between CR and vorticity/cyclogenesis index at high latitudes during cold seasons at the daily time scale. A signiﬁcant empirical relation was found between temporal interannual and spatial variations of low cloud amount and cosmic ray induced ionization for the period 1983–2000. However, the relation between diﬀerent types of clouds still needs to be understood. Although a link between solar activity and climate seems plausible on the millennial time scale, only a marginal correlation with the geomagnetic ﬁeld variations supports the idea of CR inﬂuence upon climate. Evidence has been presented on a correlation between the mega-year time scale climate proxy series and model variations of CR due to the changes of the galactic surroundings. However, large uncertainties make this result only indicative. In conclusion, a CR-climate link seems a plausible climate driver, but the present correlations favoring it, while numerous, are not solid. However, in the aggregate, they support the existence of a link between CR and the climate on Earth. The need for a quantitative model able to describe the cosmic ray eﬀect on the atmospheric properties is critical. The next step is to proceed from phenomenological statistical studies to quantitative semiempirical and physical models.

Acknowledgments We acknowledge the support from the Academy of Finland, the Finnish Academy of Science and Letters, Vilho, Yrj¨ o and Kalle V¨ ais¨ al¨ a Foundation, and the Russian Academy of Sciences.

References 1. 2. 3. 4.

J. Lean, J. Beer and R. Bradley, Geophys. Res. Lett. 22 (1992) 3195. C. Fr¨ ohlich, Space Sci. Rev. 94 (2000) 15. P. Foukal, Geophys. Res. Lett. 29, 23, (2002) CiteID 2089. J. D. Haigh, M. Lockwood and M. S. Giampapa, The Sun, Solar Analogs and the Climate, Saas-Fee Advanced Course 34 (Springer, 2005).

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

330

5. 6. 7. 8. 9. 10. 11. 12. 13.

14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46.

ch26

I. G. Usoskin and G. A. Kovaltsov

C. Fr¨ ohlich and J. Lean, Astron. Astrophys. Rev. 12 (2004) 273. S. K. Solanki and N. A. Krivova, Solar Phys. 224 (2002) 197. Y.-M. Wang, J. Lean and N. R. Sheeley, Jr., Astrophys. J. 625 (2005) 522. J. D. Haigh, Science 272 (1996) 981. D. Shindell et al., Science 284 (1999) 305. E. R. Ney, Nature 183 (1959) 451. R. E. Dickinson, Bull. Am. Meteorol. Soc. 56 (1975) 1240. R. G. Harrison and K. S. Carslaw, Rev. Geophys. 41 (2003) CiteID 1012. B. A. Tinsley and F. Yu, Solar Variability and its Eﬀects on Climate, Geophys. Monogr., Vol. 141, eds. J. Pap and P. Fox (AGU, Washington, 2004), p. 321. F. Yu and R. P. Turco, Geophys. Res. Lett. 24 (1997) 1927. N. Marsh and H. Svensmark, Phys. Rev. Lett. 85 (2000) 5004. F. Yu, J. Geophys. Res. 107, A7 (2002) CiteID 1118. B. A. Tinsley, Space Sci. Rev. 94 (2000) 231. B. A. Tinsley et al., J. Atm. Sci. 57 (2000) 2118. J. D. Haigh, Nature 370 (1994) 544. S. K. Solanki et al., Nature 431 (2004) 1084. I. G. Usoskin and B. Kromer, Radiocarbon 47, 1 (2005) 31. M. Frank et al., Earth Planet. Sci. Lett. 149 (1997) 121. M. Sharma, Earth Planet. Sci. Lett. 199 (2002) 459. N. Shaviv and J. Veizer, GSA Today 13, 7 (2003) 4. L. I. Dorman, Cosmic Rays. Variations and Space Explorations (NorthHolland, Amsterdam, 1974). M. I. Pudovkin and S. V. Veretenenko, J. Atm. Terr. Phys. 57 (1995) 1349. V. C. Roldugin and B. A. Tinsley, J. Atm. Solar-Terr. Phys. 66 (2004) 1143. D. R. Kniveton, J. Atm. Solar-Terr. Phys. 66 (2004) 1135. D. R. Kniveton and B. A. Tinsley, J. Geophys. Res. 109 (2004) D11201. M. C. Todd and D. R. Kniveton, J. Atm. Solar-Terr. Phys. 66 (2004) 1205. Y. I. Stozhkov, J. Phys. G 29 (2003) 913. Y. I. Stozhkov et al., Geomag. Aeron. 36 (1996) 211. B. A. Tinsley, G. M. Brown and P. H. Scherrer, J. Geophys. Res. 94, D12 (1989) 14783. B. A. Tinsley and G. W. Deen, J. Geophys. Res. 96 (1991) 22283. S. V. Veretenenko and P. Thejll, J. Atm. Solar-Terr. Phys. 66 (2004) 393. H. Svensmark and E. Friis-Christensen, J. Atm. Terr. Phys. 59 (1997) 1225. H. Svensmark, Phys. Rev. Lett. 81 (1998) 5027. N. Marsh and H. Svensmark, Space Sci. Rev. 107 (2003) 317. K. Gierens and M. Ponater, J. Atm. Solar-Terr. Phys. 61 (1999) 795. J. Kirkby and A. Laaksonen, Space Sci. Rev. 94 (2000) 397. K. S. Carslaw, R. G. Harrison and J. Kirkby, Jr., Science 298 (2002) 1732. N. Marsh and H. Svensmark, J. Geophys. Res. 109, D14 (2004) D14205. P. M. Udelhofen and R. D. Cess, Geophys. Res. Lett. 28 (2001) 2617. J. E. Kristj´ ansson et al., Geophys. Res. Lett. 29, 23 (2002) 2107. B. Sun and R. S. Bradley, J. Geophys. Res. 107, D14 (2002) CiteID 4211. P. Laut, J. Atm. Solar-Terr. Phys. 65 (2003) 801.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Link Between Cosmic Rays and Clouds

ch26

331

47. I. G. Usoskin et al., Geophys. Res. Lett. 31 (2004) L16109. 48. I. G. Usoskin, O. G. Gladysheva and G. A. Kovaltsov, J. Atm. Solar-Terr. Phys. 66, 18 (2004) 1791. 49. E. Pall´e, C. Butler and K. O’Brien, J. Atm. Solar-Terr. Phys. 66 (2004) 1779. 50. E. Pall´e, Geophys. Res. Lett. 32 (2005) CiteID L03802. 51. E. Pall´e and C. J. Butler, J. Atm. Solar-Terr. Phys. 64 (2002) 327. 52. G. Bond et al., Science 294 (2001) 2130. 53. C. Doake, Earth Planet. Sci. Lett. 38 (1978) 313. 54. V. Courtillot et al., Nature 297 (1982) 386. 55. M. Christl et al., J. Atm. Solar-Terr. Phys. 66 (2004) 313. 56. G. Wagner et al., J. Geophys. Res. 106, D4 (2001) 3381. 57. I. G. Usoskin et al., J. Geophys. Res. 110 (2005) A10102. 58. N. Shaviv, Phys. Rev. Lett. 89 (2002) Cite ID 051102. 59. N. Shaviv, New Astro. 8 (2003) 39. 60. J. Veizer, Geosci. Canada 23 (2005) 13. 61. S. Rahnstorf et al., EOS Transactions 85, 4 (2004) 38. 62. D. R. Gies and J. W. Helsel, Astrophys. J. 626 (2005) 844. 63. A. A. Pavlov et al., Geophys. Res. Lett. 32 (2005) CiteID L03705. 64. E. Eide and T. H. Torsvik, Earth Planet. Sci. Lett. 144 (1996) 389. 65. J. Wendler, J. Atm. Solar-Terr. Phys. 66 (2004) 1195.

This page intentionally left blank

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch27

THE EFFECT OF SOLAR ACTIVITY ON ANNUAL PRECIPITATION IN DELINGHA REGION, TIBETAN PLATEAU FOR THE LAST 1000 YEARS LEI HUANG∗,§ , XUEMEI SHAO∗,† , HONGBIN LIU‡ , ERYUAN LIANG† and LILY WANG∗,† of Geographic Sciences and Natural Resources Research Chinese Academy of Sciences, Beijing 100101, China

∗Institute †Institute

of Tibetan Plateau Research, Chinese Academy of Sciences Beijing 100085, China ‡National

Climate Center, Beijing 100081, China §[email protected]

The relationship between precipitation variations in Delingha region, Tibetan Plateau, and solar activity for the last 1000 years is investigated by pattern matching and statistical methods. The precipitation variations in the study region derived from tree ring studies show visually matching patterns with that of the solar activity. It is evident that the precipitation decreased signiﬁcantly during the periods of the Spoerer, Maunder, and Dalton minima of solar activity during the Little Ice Age. Power spectral analysis and wavelet analysis showed the presence of statistically signiﬁcant period of 200 years as a possible connection between precipitation and solar forcing. Detailed study with cross wavelet analysis found solar activity can aﬀect long-term variations of precipitation mainly on century-scale of 200 years. Our analysis suggests that solar activity probably played an important role in inﬂuencing precipitation variations in the study region for the last 1000 years.

1. Introduction The characterization of variations in solar activity has been one of the elements in understanding climate change. The issue of solar variability and its inﬂuence on the Earth’s climate has been receiving more and more attention and has been studied widely in recent years.1 The research on this issue requires development of high-resolution proxy records of climate spanning several centuries and millennia. The available proxy records which span millennia and beyond are largely restricted to mid- and high-latitude and some high-altitude regions. The Tibetan Plateau is of paramount climatic importance for its role in modulating global and regional atmospheric circulation, and its climate is also strongly sensitive to global climatic change. The northeastern Tibetan 333

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

334

ch27

L. Huang et al.

Plateau is a region rich in natural records of past climate variability such as tree rings and lake sediments. In particular, this region has a unique concentration of tree-ring records of 1000 years and longer. Tree-ring records are annually resolved and are often in better agreement with instrumental data.2 Thus, tree-ring records that are long enough to preserve lowfrequency variability are particularly important for the investigation of long-term climate changes and the sun–climate connection. Based on the tree ring width chronologies developed in northeastern of Tibetan Plateau, we have reconstructed annual precipitation variations in Delingha region for the last 1000 years.3 In this study, we search for associations between solar activity and the precipitation variability and investigate the possible solar inﬂuence on precipitation variations in northeastern Tibetan Plateau during the last millennium.

2. Data and Methods The study region is located in eastern part of Qaidam Basin, northeastern Tibetan Plateau. This region is characterized by arid climate with mean annual precipitation varying from 200 to 150 mm, declining from east to west. Qilian juniper (Sabina przewalskii Kom.) trees grow at the elevations of 3500–4000 m on the sunny and semi sunny slopes of mountains, where they receive the maximum precipitation. Qilian juniper can attain the age of over 1000 years and therefore provide a remarkable source for studying the past environmental changes in this region. A total of seven tree ring-width chronologies were established in the study area (Fig. 1). The ﬁrst principal component with same sign for all the chronologies was extracted to represent dominant ring-width variations in this region. The regression analysis was used to develop calibration equation between the ﬁrst principal component and climate variables. An annual precipitation calculated from previous July to current June was reconstructed for the last 1000 years.3 Solar forcing calculated for the last 1000 years by Crowley4 was used as solar activity indicator. We chose wavelet transform method5 to analyze the climatic variations and solar eﬀect. The complex Morlet wavelet analysis is used because it is more appropriate for detecting variations in the periodicities of geophysical signals at all timescales. Cross wavelet analysis was also used to investigate the temporal variability of the relationship between precipitation and solar forcing.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch27

Eﬀect of Solar Activity on Precipitation in Delingha for the Last 1000 Years

Fig. 1.

335

Locations of the tree-ring sites in the study region.

3. Results and Discussion We ﬁrst examined the long-term variability of the reconstructed precipitation in Delingha region to investigate any changes occurring in its characteristics. The standard deviation of the interannual variability was 34.6 mm, which is 26% of the mean value (132 mm). Figure 2 displays the annual precipitation variations during the last 1000 years. If the mean annual precipitation from 1971 to 2000 is taken as the baseline, the last 1000 years is characterized by generally lower precipitation. However, if the mean of annual precipitations for the last 1000 years is considered, main wet periods occurred during AD 1520–1633 and 1933–2001, whereas dry intervals occurred during AD 1429–1519, 1634–1741, and 1781–1839. These three dry periods are coincident with solar minima during the Little Ice Age (Fig. 2). The Maunder Minimum centered around late 17th century matches the strongest interval of dry period in the last 400 years. Wide-scale drought is well documented from the early to middle of the 18th century. Sediment records of Qinghai Lake6 and ice cores of Guliya7 also show the early decades of the 1800s as a dry interval in the NE Tibetan

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

336

ch27

L. Huang et al.

Fig. 2.

Variations of the precipitation for the last 1000 years.

Plateau. This record of drought closely coincides with the Dalton solar minimum. Before the Little Ice Age, the Wolf Minimum and the Oort Minimum also corresponding with dry intervals in Delingha region. The Maunder Minimum (1645–1715) is a period in which solar activity decreased notably. In our study area, the Maunder Minimum occurred during one of the driest periods of the past 1000 years. The mean value of precipitation in this period is 110 mm, much lower than that of the entire series; the period before the mega drought, 1574–1644, averaged 151.1 mm, and the period following the drought, 1716–1786, averaged 136.1 mm. The t-test of diﬀerence between the means veriﬁed the existence of statistically signiﬁcant diﬀerences with the periods 1574–1644 and 1716–1786. Therefore, it can be aﬃrmed that the decrease of solar activity during the Maunder Minimum may have a notable inﬂuence on precipitation of Delingha region. The same analysis was carried out for the other four solar minima of the last 1000 years (Table 1). Except for the Oort Minimum, the other three solar minima (Wolf, Spoerer, and Dalton) show similar results as the Maunder Minimum. It is evident that the precipitation decreased signiﬁcantly during the periods of the Wolf, Spoerer, Maunder, and Dalton minima. A Blackman-Tukey power spectra of the series show peaks at periodicities of 300, 200, 150, 120, and 100 years with the 200-year cycle the most signiﬁcant. The 200-year oscillations are thought to be the result of the

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch27

Eﬀect of Solar Activity on Precipitation in Delingha for the Last 1000 Years Table 1. Solar minimum Oort Wolf Spoerer Maunder Dalton

Fig. 3.

337

Solar minimum and precipitation in Delingha.

Preceding period (signiﬁcant of t-test) 137.2 140.8 151.1 143.8

(95%) (99%) (99%) (99%)

Minimum period 132.3 125.9 112.2 110.0 121.7

Afterward period (signiﬁcant of t-test) 136.8 (< 95%) 140.9 (99%) 149.8 (99%) 136.1 (99%) 135.4 (95%)

Morlet wavelet transformation of annual series of precipitation.

inﬂuence of the solar Suess cycle on climatic parameters.1 We then made a complex Morlet wavelet analysis of the annual precipitation series and the real part of wavelet coeﬃcient is presented in Fig. 3. The X-coordinate of Fig. 3 shows the time interval (years) and the Y -coordinate shows the oscillation periods. In Fig. 3, the periodicities of the Suess cycle are the strongest. Climatic variations in the range of periods of the Suess cycle occur during the entire time interval. The Suess cycle of 200-year has been found in other proxies of climatic change on the Tibetan Plateau. Time series analysis of the record from Guliya ice core, located in the west of the Tibetan Plateau, shows the 200-year cycle over the last 2000 years.7 The 200-year cycle has also been found in Qinghai Lake sediment and is attributed to solar forcing of century-scale variability in regional precipitation for the past 800 years.6

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

338

ch27

L. Huang et al.

Fig. 4.

Cross wavelet analysis of precipitation and solar activity.

The cross wavelet analysis gives a measure of the correlation between the two time series as a function of both scale (period) and time. A good correspondence in 200-year scale variations was observed between the spectral amplitude of precipitation and the solar activity by cross wavelet analysis (Fig. 4).

4. Discussion and Conclusions An increasing body of evidence makes clear the existence of a marked inﬂuence of solar variability on climate with semi-periodic cycles at decadal, centennial, and millennial scales.8 On centennial time scales, the documented 200-year solar cycle was found to be compatible with a record of drought derived for the past 2600 years in the Maya Lowlands9 and has been linked to drought in the northern Great Plains over the last 2000 years.10 The association between solar cycle and climate variations has been recognized and some of the mechanisms have been identiﬁed.11,12 The sun–climate relationships have been investigated with atmospheric general circulation models. Shindell et al.13 have found that a 0.1% decrease in solar irradiance related to the 11-year sunspot cycle could alter the atmosphere’s dynamic response to changes in stratospheric ozone and temperature, producing a change in surface temperature. On multidecadal timescales, Shindell et al.13 demonstrated that an estimated 0.25%

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch27

Eﬀect of Solar Activity on Precipitation in Delingha for the Last 1000 Years

339

reduction of solar irradiance during the Maunder Minimum resulted in surface climatic patterns over the Northern Hemisphere similar to those associated with the low-index state of the Arctic Oscillation (AO) and North Atlantic Oscillation (NAO). The NAO variation has been found to be closely related to solar activity during the last two centuries.14 Correlation analysis shows that the precipitation in Delingha was positively correlated with AO index of March–April in the period of 1899–2001 with the correlation coefﬁcient of 0.363, signiﬁcant at 99% conﬁdence level. This suggests that the atmospheric circulation in high AO index period can enhance precipitation in Delingha region. The precipitation variations and solar activity in our study area show visually matching patterns. Power spectra and wavelet analyses revealed the presence of statistically signiﬁcant period of 200 years in precipitation variations suggesting a possible connection with solar forcing which was further conﬁrmed by a cross-wavelet analysis. Our result suggests that the variability of solar activity aﬀected precipitation variations in Delingha region, NE Tibetan Plateau for the last 1000 years. Acknowledgments This research was supported by the Key Project of Knowledge Innovation of the CAS (Grant No. KZCX3-SW-321). The authors thank those who provided helpful comments and suggestions in the preparation of this article and two anonymous referees for their valuable comments and considerable improvements on preliminary drafts of this paper. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

D. Rind, Science 296 (2002) 673. P. D. Jones et al., Holocene 8 (1998) 455. X. M. Shao et al., Sci. in China (D) 48 (2005) 939. T. J. Crowley, Science 289 (2000) 270. C. Torrence and G. Compto, Bull. Am. Met. Soc. 79 (1998) 61. J. W. Zhang et al., Chinese Sci. Bull. 9 (2004) 1451. T. D. Yao et al., Sci. in China (D) 39 (1996) 425. P. Laut, J. Atm. Solar Terr. Phys. 65 (2003) 801. D. A. Hodell et al., Science 292 (2001) 1367. Z. Yu and E. Ito, Geology 27 (1999) 263. G. A. Schmidt et al., Quatern. Sci. Rev. 23 (2004) 2167. M. A. Palmer et al., Adv. Sp. Res. 34 (2004) 343. D. T. Shindell et al., Science 294 (2001) 2149. B. Kirov and K. Georgieva, Phy. Chem. Earth 27 (2002) 441.

This page intentionally left blank

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch28

CLIMATE AND EXTREME EVENTS IN CENTRAL-SOUTHERN REGION OF EASTERN CHINA DURING 1620–1720 JINGYUN ZHENG∗,§ , QUANSHENG GE∗ , XIUQI FANG∗,† and ZHIMIN MAN‡ ∗Institute of Geographic Sciences and Natural Resources Research Chinese Academy of Sciences, Beijing 100101, China †School

of Geography, Beijing Normal University, Beijing 100875, China

‡Institute

of Chinese History and Geography, Fudan University Shanghai 200433, China §[email protected]

Climate and extreme events in central-southern region of eastern China during 1620–1720 are studied based on the reconstructed series of inter-decadal temperature, precipitation variability, together with historical written records. It is suggested that the climate of 1620–1720 was characterized by severe cold and great variability, with more frequent occurrences of severe winters and extreme droughts. We found the 1650s–1670s to be the coldest 30-year in central-southern region of eastern China for the last 2000 years, and the decade of 1634–1644 to experience the most severe sustained drought since 500 AD. These anomalous climate conditions and extremes may be connected to the Maunder Minimum of rare sunspot appearance and reduced solar activity.

1. Introduction The Maunder Minimum interval, 1645–1715, now commonly recognized to be a period of rare sunspot appearance and reduced solar activity,1,2 has long been suspected to produce detectable climate variations and perhaps extreme conditions both on regional and even hemispheric scales.3–7 According to two recent reconstructed North Hemisphere temperature series, temperatures during this period were the coolest for the last 2000 years.8,9 And the instrumental temperature record showed that the yearly resolved Central England winter temperatures during the Late Maunder Minimum (LMM, 1675–1715) were about 1◦ C lower than that during 1961–1990.10 Numerous related studies of regional climatic reconstruction revealed that climate condition during the LMM was really unusual. In Europe, for example, winters of the LMM were characterized by a higher frequency of severe climatic conditions than those of the twentieth century.11

341

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

342

ch28

J. Zheng et al.

Various climate proxy reconstructions also found minimum or relatively cool temperatures to prevail in that period at other parts of the world, such as Russia,12–14 North America,15–17 Southern America,18 South Africa,19,20 and tropical southwest Paciﬁc.21 Yet, climate during the Maunder Minimum may exhibit details in timing of changes and trends that can be quite diﬀerent from regions to regions.22 In this study, we focus on the climate and extreme events in central-southern region of eastern China (east of 105◦ E and south of 40◦ N approximately, see Fig. 1 for study area) during 1620–1720, a period also described as a time of “prolonged sunspot minimum”.2 The source data used here include the reconstructed series of inter-decadal temperature, precipitation variability,23,24 phenological proxy,25 and numerous descriptions of abnormal freezing conditions of seas, lakes, and rivers extracted from Chinese historical documents over central-southern region of eastern China, as well as the collateral evidences of tree-ring proxy in the Middle Qilian Mountain and Delingha of northeastern Tibetan Plateau.26,27

Fig. 1. The sketch map for studying area. TL = Tai Lake; DtL = Dongting Lake; PyL = Poyang Lake; HpR = Huangpu River. Shading area shows the spatial coverage for the series of winter half-year temperature anomaly reconstructed by Ge et al.23 Stars show the distribution of document records that historiﬁed severe winters during 1620–1720. Triangle shows the site of the Middle Qilian Mountain, northeastern Tibetan Plateau where the proxy-temperature was reconstructed from tree-rings by Liu et al.26

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch28

Climate and Extreme Events in Central-Southern Region of Eastern China

343

2. Temperature and Precipitation Anomalies During 1620–1720 from Long-Term Proxy Series 2.1. Temperature

1.0

30-year mean 10-year mean

(a)

10-year mean of plant phenodate 10-year mean of last date of snowfall 10-year mean of the date of accumulated temperature of 120˚C over 5˚C Mean of referee period (1963-1982)

(b)

0.0 -1.0

0 -10

(c) Annual ring-width index 30-year adjacent averaging Mean of series 1 0

200

400

600

800

10

1.5 1.0 0.5

1000 Year

1200

1400

1600

1800

2000

Phenodate (day) Tree-ring index

Temperature (°C)

Several century-long proxy-temperature series, covering a wide geographical area of China, are shown in Fig. 2. Figure 2(a) shows the series of winter half-year (October–April) temperature anomaly (relative to the mean of 1951–1980) in the middle and lower reaches of Yellow River and Yangtze River during the last 2000 years, which was reconstructed by using the phenological cold/warm events recorded in historical documents and some previously published series.23 The time resolution of this series was 10year for 960s–1100s and 1500s–1990s, as well as 30-year for the others. It is worth noting that this reconstruction has signiﬁcant improvements in the collection and veriﬁcation of the original records, accurate dating while adopting identical time resolution, all of which can be considered as reﬁnements over previous temperature reconstruction eﬀorts by Zhu,28 and Zhang.29 Thus, the conﬁdence of the results presented in Fig. 2(a) is enhanced. According to Fig. 2(a), the period of 1620–1720 was one of the coldest periods in the last 2000 years in the middle and lower reaches of Yellow

Fig. 2. (a) Winter half-year temperature anomaly (relative to the mean of 1951–1980) for the last 2000 years in the middle and lower reaches of Yellow River and Yangtze River (from Ref. 23). (b) Phenological date anomaly during 1580s–1980s in the lowermiddle Yangtze River Valley (from Ref. 25). (c) Tree-ring index during 1000–2000 in the Middle Qilian Mountain of northeastern Tibetan Plateau (from Ref. 26). The interval of 1620–1720 is between the two vertical dash lines.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

344

ch28

J. Zheng et al.

River and Yangtze River, with the mean of winter half-year temperatures during 1620–1720 is about 0.5◦ C lower than those of 1951–1980. The period of 1650s–1670s was the coldest 30-year during the last 2000 years in this region, with anomaly of −1.1◦ C (relative to the mean of 1951–1980). The 1650s and 1660s stood out as two of the coldest decades since 1500, with anomaly of about −1.3◦ C (relative to the mean of 1951–1980). Based on available records of frost hazard of Chinese orange (usually called mandarin orange) and southern boundary of lake and river freeze,30 it was estimated that the maximum amplitude of a southward displacement of climatic zone during the relatively cool 1620–1720 period was about 3◦ latitude than that during 1950s–1970s. This new result can be validated by the collateral evidences of spring phenology records in the lower-middle Yangtze River Valley of eastern China for 1580–1920 (Fig. 2(b)), which shows an evident cold period in 1620–1700 and the coldest decades in 1650s–1690s.23 It is further noteworthy that a new tree-ring record also showing a detectable cold period in 1620–1720 in the Middle Qilian Mountain of northeastern Tibetan Plateau since 1000 (Fig. 2(c)). Such rather coherent proxy-temperature records for both eastern and western China indicate that the cold climate happened not just only in central-southern region of eastern China but also for most of China.

2.2. Precipitation In order to study the inter-decadal precipitation variability, the time series of dry–wet ratio in North China Plain, Jiang-Huai Valley (i.e., the Valley of Huai River and Middle-Lower Yangtze River), Southeastern China (see Fig. 3 for the locations) are presented in Fig. 4. All these proxy series were reconstructed from the dataset of dryness/wetness grades at 48-sites of eastern China since 500 AD, which were derived from Chinese historical drought and ﬂood records,29 after eliminating the eﬀect of uneven temporal and spatial distributions in the source records. The detailed comparison of each regional dry–wet ratio series and instrumental precipitation data, as well as other independently reconstructed precipitation data,31 indicated that the reconstructed regional dry–wet ratio series can serve as a reasonable proxy of precipitation variation.24 Figure 4(a) shows that in North China Plain, 1620–1720 was a transition period from relative dry to relative wet conditions, with great precipitation variability and an extremely dry event in the early part of the period. In Jiang-Huai Valley (Fig. 4(b)), the precipitation variation during

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch28

Climate and Extreme Events in Central-Southern Region of Eastern China

345

Fig. 3. Location of North China Plain, Jiang-Huai Valley and Southeastern China. Dots show the distribution of the dryness/wetness grades dataset (data from Ref. 29). Tree shows the site of Delingha, northeastern Tibetan Plateau where the yearly precipitation reconstructed from tree-ring by Shao et al.31 2.0

(a) North China Plain

Dry-wet ratio

0.0 -2.0 2.0

(b) JiangHuai Valley

0.0 -2.0 2.0

(c) Southeastern China

0.0

P (mm)

-2.0 200 100 500

Annual 30-year adjacent averaging Mean of series 600 700

800

(d) Delingha, Northeastern of Tibetan Plateau

900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000

Year

Fig. 4. (a)–(c) Reconstructed proﬁles of precipitation variability over central-southern region of eastern China during 500–2000 (modiﬁed from Ref. 24). Blue lines: dry–wet ratio; black dash lines: 30-year FFT ﬁlter smoothing; black solid horizontal lines: mean of series; black dash horizontal lines: 1.645 times standard deviation of dry–wet ratio time series; red and blue vertical lines: the intervals of sustained drought (in red) and persistent ﬂood (in blue) during 1620–1720. (d) Precipitation derived from tree-rings in Delingha of northeastern Tibetan Plateau during 1000–2000 (from Ref. 31). The interval of 1620–1720 is marked by the green dash block.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

346

ch28

J. Zheng et al.

1620–1720 was similar to the North China Plain, but with relatively smaller variability. While in Southeastern China (Fig. 4(c)), most of 1620–1720 was relatively dry. In summary, the precipitation variation over centralsouthern region of eastern China during 1620–1720 was characterized by a time of relative dry and great variability, with a higher frequency of sustained droughts. The reconstructed 1000-year precipitation history derived from tree-rings at Delingha of northeastern Tibetan Plateau27 also showed relative dryness and great precipitation variability during 1620–1720, which may implies that this precipitation characteristic is not only restricted to central-southern region of eastern China. Our result of large precipitation variability during a very cold period in this region can be supported by an earlier study, which also found that greater precipitation variability would occur during a cool period in China.32 The physical mechanism behind such a climatic phenomenon has been argued to be related to the expansion of circumpolar vortex and the east–west shift of the ampliﬁed planetary waves in middle latitudes during a cool period that in turn increase climatic variability in certain areas of the hemisphere, especially the eastern part of the continent or near the west coast of the ocean in middle latitudes.

3. Extreme Events During 1620–1720 To identify any potential extremes among severe winters during 1620–1720, the historical records of 16 severe winters, listed in Table 1, are used to compare the severity of individual winter. The source records (see Fig. 1 for spatial coverage) for Table 1 are extracted by us from many Chinese historical documents, such as local gazettes, diary, and scattered historical writings, as well as some previous publications,28,33,34 of which most of records from local gazettes can be found in A compendium of Chinese Meteorological Records of the Last 3000 years 35 independently edited by Zhang recently. The winter (December–February) temperature anomalies (relative to the mean of 1951–1980) for six cold winters during 1951–2000 bracketed in Table 1 were calculated based on the instrumental data from 20 stations used in our previous study.23 According to Table 1, the 16 selected severe winters during 1620–1720 were all colder than those in 1951–2000. Several unusually cold phenomena of 1620–1720, such as freezing of China East Sea, middle-lower rivers of Yangtze and Huangpu, have never been observed in the modern period at least since 1951. The frequency (16/100) of severe winters during 1620–1720 was also higher than that (6/50) of 1951–2000.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch28

Climate and Extreme Events in Central-Southern Region of Eastern China

347

Table 1. Comparison of the severity of 16 cold winters during 1620–1720 versus six cold winters during 1951–2000. Freeze occurred in sea, lake, and river Winter of 1620–1720 1620–21 1636–37 1653–54 1654–55 1655–56 1660–61 1665–66 1670–71 1676–77 1683–84 1689–90 1690–91 1694–95 1700–01 1714–15 1720–21 1951–2000 1954–55 (−1.6◦ C)∗ 1956–57 (−1.8◦ C) 1963–64 (−1.1◦ C) 1968–69 (−0.9◦ C) 1971–72 (−1.1◦ C) 1976–77 (−1.1◦ C)

CES TL DtL PyL HaR HuR HpR PYR SRL F

SF

SF SF

F F

F

F

SF

SF F F

F

SF F

SF

SF

SF SF

F F SF

F

SF

SF

F

F

F

F

SF

SF SF SF SF SF SF SF SF SF SF SF SF SF SF SF F

F

F

F

F

CW- DPOS S&T VS S VS VS S VS VS VS VS VS VS VS VS VS VS VS

VS VS VS VS VS VS VS VS VS VS VS VS VS VS VS VS

S S S S S S

VS

S S S

*The numbers within brackets show temperature anomaly (relative to mean of 1951– 1980) of the middle and lower reaches of Yellow River and Yangtze River for each winter (December–February). Note: CES = China East Sea; TL = Tai Lake; DtL = Dongting Lake; PyL = Poyang Lake; HaR = Lower Han River; HuR = Lower Huai River; HpR = Lower Huangpu River; PYR = Part of the middle-lower Yangtze River; SRL = most of small rivers and lakes located in the south of Yangtze River; CW-OS = Severe snowy and cool overcast weather with long continuous days (at least more than 10 days) in the lower-middle Yangtze Valley, or southern China; DP-S&T = Severe damage on plants of Chinese orange (usually called mandarin orange), bamboo, lichee and other subtropical or tropical plants. Where “F” means freeze; “SF” means severe freeze; “S” means severe; “VS” means very severe.

The extreme precipitation events, either severe drought or heavy ﬂood, are deﬁned as an interval with sustained anomaly for longer than three years and large-amplitude precipitation anomaly beyond 1.645 times standard deviation based on dry–wet ratio series over one region. Because the time resolution of dry–wet ratio series presented in Fig. 4 was based on 10-year

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

348 Table 2.

ch28

J. Zheng et al. The extreme precipitation events during 1620–1720 over eastern China.

When

Where

Max. annual anomaly of dry–wet ratio (sigma unit)

Sustained drought

1627–1629 1634–1644 1671–1679 1687–1697 1719–1723

NC NC, JH, extend to part of SEC SEC SEC NC, extend to part of JH

−1.65 −3.21(NC), −2.1 (JH) −1.65 −1.80 −1.72

Persistent ﬂood

1646–1654 1702–1709

NC JH, extend to part of NC

Event

2.09 1.65

Note: NC, JH, SEC means North China Plain, Jiang-Huai Valley, southeastern China respectively.

running means, we use the original dryness/wetness grade data, together with the historical descriptions of the impacts of severe drought and ﬂood, to better determine the starting and ending years of each event. The results, shown in Table 2, indicate that there were ﬁve sustained droughts and two persistent ﬂoods during 1620–1720. The most severe sustained drought for the last 1500 years occurred in 1634–1644. This unprecedented drought resulted in the crop failure and concomitant famine, which triggered a largescale peasant rebellions and social turbulence further across much of the central-eastern China, and may played important role for the collapse of the Ming Dynasty as argued in previous studies.36,37

4. Conclusion and Discussion This study presents the climatic characteristics in central-southern region of eastern China during the well-known solar Maunder Minimum period of 1620–1720, by analyzing the inter-decadal temperature and precipitation variations, as well as identifying extreme events. The results indicate that the climate of 1620–1720 was characterized by severe cold and great variability, with relatively higher frequency of severe winters and extreme droughts when compared to modern periods (i.e., 1951–2000). The 1650s– 1670s was the coldest 30-year for the last 2000 years in central-southern region of eastern China, and the decade of 1634–1644 experienced the most severe sustained drought since 500 AD. Similar climatic characteristics were also recognized in many other regions of the world during LMM by numerous studies mentioned in Sec. 1. This implies that abnormal solar activity during the Maunder Minimum may yield unusual climate and higher frequency of extreme events that could be coherent world wide as argued in

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch28

Climate and Extreme Events in Central-Southern Region of Eastern China

349

many previous studies.38 Our study for eastern China contributes important regional evidences on this phenomenon. Physical connections responsible for all these observed climate variabilities, either worldwide or restricted to eastern China, should be further pursued in addition to the zero-order descriptive work presented here. Acknowledgments This study was supported by the project of “Extreme climate events during historical time: Reconstruction, impact and adaptation” (KZCX3-SW-321) from Chinese Academy of Sciences. We appreciate Willie Soon for his constructive comments and unselﬁsh help in English editing. References 1. J. A. Eddy, Science 192 (1976) 1189. 2. W. W.-H. Soon and S. H. Yaskell, The Maunder Minimum and the Variable Sun–Earth Connection (World Scientiﬁc, Singapore, 2004). 3. S. Baliunas and W. Soon, Global Warming: The Science and the Politics, ed. L. Jones (The Fraser Institute, Vancouver, British Columbia, Canada, 1997), pp. 77–90. 4. T. Crowley, Science 289 (2000) 270. 5. D. T. Shindell, G. A. Schmidt, M. E. Mann, D. Rind and A. Waple, Science 294 (2001) 2149. 6. D. T. Shindell, G. A. Schmidt, R. L. Miller and M. E. Mann, J. Climate 16 (2003) 4094. 7. C. F. Keller, Advances in Space Research 34 (2004) 315. 8. M. E. Mann and P. D. Jones, Geophys. Res. Lett. 30 (2003) 1820. 9. A. Moberg, D. M. Sonechkin, K. Holmgren, N. M. Datsenko and W. Karl´en, Nature 433 (2005) 613. 10. G. Manley, Q. J. R. Met. Soc. 100 (1974) 389. 11. J. Luterbacher, R. Rickli, E. Xoplaki, C. Tinguely, C. Beck, C. Pﬁster and H. Wanner, Climatic Change 49 (2001) 441. 12. M. M. Naurzbaev and E. A. Vaganov, J. Geophys. Res. 105 (2000) 7317. 13. I. Kalugin, V. Selegei, E. Goldberg and G. Seret, Quatern. Int. 136 (2005) 5. 14. W. Mackay, D. B. Ryves, R. W. Battarbee, R. J. Flower, D. Jewson, P. Rioual and M. Sturm, Global Planet. Change 46 (2005) 281. 15. H. C. Fritts and J. M. Lough, Climatic Change 7 (1985) 203. 16. H. Luckman and R. J. S. Wilson, Climate Dynamics 24 (2005) 131. 17. R. D’Arrigo, E. Mashig, D. Frank, R. Wilson and G. Jacoby, Climate Dynamics 24 (2005) 227. 18. R. Villalba, A. Lara, J. A. Boninsegna, M. Masiokas, S. Delgado, J. C. Aravena, F. A. Roig, A. Schmelter, A. Wolodarsky and A. Ripalta, Climatic Change 59 (2003) 177.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

350

ch28

J. Zheng et al.

19. K. Holmgren, W. Karlen, S. E. Lauritzen, J. A. Leethorp, T. C. Partidge, S. Pikeh, P. Repinski, C. Stevenson, O. Svanered and P. D. Tyson, Holocene 9 (1999) 295. 20. P. E. Tyson, W. Karlen, K. Holmgren and G. A. Heiss, S. Afr. J. Sci. 96 (2000) 121. 21. T. M. Quinn, T. J. Crowley and F. W. Taylor, Paleoceanography 13 (1998) 412. 22. W. Soon, S. Baliunas, C. Idso, S. Idso and D. R. Legates, Energy & Environment 14 (2003) 233. 23. Q. S. Ge, J. Y. Zheng, X. Q. Fang, Z. M. Man, X. Q. Zhang and W.-C. Wang, Holocene 13 (2003) 933. 24. J. Y. Zheng, W.-C. Wang, Q. S. Ge, Z. M. Man and P. Y. Zhang, PAGES 2nd Open Science Meeting: Paleoclimate, Environmental Sustainability and our Future (PAGES International Project Oﬃce, Bern, Switzerland, 2005), p. 43 25. S. Hameed and G. F. Gong, Geophys. Res. Lett. 21 (1994) 2693. 26. X. H. Liu, D. H. Qin, X. M. Shao, T. Chen and J. W. Ren, Sci. China (D) 48 (2005) 521. 27. X. M. Shao, L. Huang, H. B. Liu, E. Y. Liang, X. Q. Fang and L. L. Wang, Sci. China (D) 48 (2005) 939. 28. K. Z. Zhu, Scientia Sinica 16 (1973) 226. 29. P. Y. Zhang, Climate Changes in China during Historical Times (Shandong Science and Technology Press, Ji’nan, China, 1996) (in Chinese). 30. Z. M. Man, Fudan Journal (Social Sciences) 5 (1999) 72 (in Chinese). 31. J. Y. Zheng, Z. X. Hao and Q. S. Ge, Sci. China (D) 48 (2005) 2182. 32. S. Z. Zheng and L. W. Feng, Scientia Sinica (B) 29 (1986) 441. 33. G. F. Gong, P. Y. Zhang and J. R. Zhang, Geographical Symposium (also called Collected Papers in Geography) 18 (1987) 129 (in Chinese). 34. Z. J. Ma (ed.), Severe Natural Disasters and Relief Strategies in China: Chronology for 1949–1990 (China Ocean Press, Beijing, China, 1995) (in Chinese). 35. D. E. Zhang (ed.), A Compendium of Chinese Meteorological Records of the Last 3000 Years (Jiangsu Education Publishing House, Nanjing, China, 2004) (in Chinese). 36. K. J. Hsu, Sci. China (D) 41 (1998) 449. 37. Y. C. Zhang, C. S. Zhang and L. H. Zhang, Earth Sci. Frontiers 10 (2003) 265 (in Chinese). 38. D. Rind, Science 296 (2002) 673.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch29

EFFECTS OF TYPHOON ON THE IONOSPHERE YI-MOU LIU∗,§ , JING-SONG WANG∗,† and YU-CHENG SUO‡ ∗School

of Earth and Space Science, Peking University, Beijing, China † China Meteorological Administration, Beijing, China ‡China Research Institute of Radio Wave Propagation, Beijing, China §[email protected]

In this paper, we analyzed the ionospheric responses over the southern sea of China before and after typhoons’ landing, using the data obtained at GuangZhou and Hai-Kou stations. Several typical typhoon cases were selected by excluding the inﬂuence of geomagnetic disturbance. Results show that after typhoons’ landing, the foF2 begins to decrease and reaches its minimum about two days later. The moving of the turbopause due to the typhoon activities was argued to be the reason of the ionospheric responses to typhoon.

1. Introduction More and more evidences indicate that the ionosphere can be aﬀected by severe meteorological phenomena, especially during the time of thunderstorms1–3 and typhoons.4,5 The existence of the relationship between the ionospheric disturbances and strong winds in the troposphere has been observed and discussed. Bauer6 analyzed the ionospheric responses to the passage of hurricanes and found that with an approaching hurricane, the foF2 began to rise, and reached its maximum when the hurricane passed closest to the sounding station (see Fig. 1). But Shen4 showed another scenario. Using the same method as Bauer’s, Shen studied the correlations between the typhoon and the foF2 of ionosphere and the results show that the foF2 decreases in the area right above the center of the typhoon, but increases in the surrounding area. He suggested that the strong vertical convection in the troposphere caused by typhoon could change the local circulation system in the stratosphere and mesosphere, and then inﬂuence foF2 . In this paper, we used the data obtained at Guang-Zhou (23◦ N, 113◦ E) and Hai-Kou (20◦ N, 110.2◦E) stations to analyze the ionospheric disturbances over the southern sea of China during the period of the three typhoons: Penny (October 1998), Maggie (June 1999), and Durian (July 351

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

352

ch29

Y.-M. Liu, J.-S. Wang and Y.-C. Suo

Fig. 1. Meteorological and geophysical data during the period of hurricane Hazel. D(foF2 ) denotes the critical frequency of the F2 layer in terms of its departure from the monthly mean. KCH is the geomagnetic index. H500 is the height of the 500-mb pressure surface and P is the surface pressure (from Ref. 6).

2000). These typhoons are chosen because of the geomagnetic quiet during these typhoons. The results show that the foF2 decreases after typhoons’ landing, and the magnitude, and the duration of such disturbances are positively correlative to the wind power. We suggest that the moving of the turbopause due to the typhoon activities may be the cause of the ionospheric responses to typhoons. 2. Observations and Results Table 1 lists the landing location, time, maximal velocity, and center pressure when the three investigated typhoons — Penny, Maggie, and Durian are landing on the continent of China. Figure 2 illustrates the locations of the sounding stations, the paths and the landing locations of the three typhoons. Figure 3 presents the critical frequency of the F2 layer in terms of its relative departure from the monthly median, the geomagnetic Dst-index, the height of the 500-mb pressure surface H500 , and the surface pressure during the period of typhoon Durian, which happened in 2001. The red thin curve denotes the relative variations of the foF2 obtained at Guang-Zhou

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch29

Eﬀects of Typhoon on the Ionosphere Table 1.

353

The parameters of typhoon.

Num

Name

Year

Month

Day

Hour

Lat. (◦N)

Long. (◦E)

Velocity (m/s)

Pressure (mb)

9803 9903 0103

Penny Maggie Durian

1998 1999 2001

8 6 7

11 6 2

7 22 2

21.8 23 21.1

112 116.4 110.3

25 35 35

990 970 970

Fig. 2. The locations of the sounding stations and the paths and landing locations of typhoons.

Fig. 3. The foF2 , geomagnetic Dst-index, the height of the 500-mb pressure surface (H500 ) and the surface pressure (P ) during the period of typhoon Durian. The vertical line represents the time of typhoon’s landing and the vertical dashed line denotes the time when foF2 reaches its minimum.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

354

ch29

Y.-M. Liu, J.-S. Wang and Y.-C. Suo

station and the red thick curve is the corresponding daily running average, while the black curves are for that at Hai-Kou station. The vertical solid line denotes landing time of the typhoon, and the vertical dashed line presents the time when the foF2 reaches its minimum. Figures 4 and 5 are for typhoon Maggie and Penny, respectively.

Fig. 4. The foF2 , geomagnetic Dst-index, the height of the 500-mb pressure (H500 ) surface and the surface pressure (P ) during the period of typhoon Maggie.

Fig. 5. The foF2 , geomagnetic Dst-index, the height of the 500-mb pressure (H500 ) surface and the surface pressure (P ) during the period of typhoon Penny.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch29

Eﬀects of Typhoon on the Ionosphere Table 2. Name Penny Maggie Durian

355

The comparisons of the three typhoons.

Velocity (m/s)

Distance

MG

MH

Td (day)

Tdelay

25 35 35

DG = DH DG < D H DG > D H

−11% −16% −20%

−6% −10% −20%

∼2 ∼ 3.5 ∼ 2.7

∼2 ∼2 ∼ 1.8

Note: DG and DH denote the distances between the landing location and the sounding stations, Guang-Zhou and Hai-Kou, respectively. MG and MH represent the maximal perturbation of the ionosphere over the two stations. Td is the duration of the perturbation and Tdelay denotes the interval between landing time and the time of minimal foF2 .

It can be seen that with the approaching of typhoons, the H500 and the surface pressure begin to decrease. After typhoons’ landing, the foF2 begins to decrease and reaches its minimum about two days later, while the H500 and surface pressure rise and reach their maximum. It has been known that the foF2 is sensitive to the daily variation of the magnetic activities. But from Figs. 3–5, we can see that the geomagnetic condition is relatively quiet during the period of the three typhoons. For the Penny case, the foF2 ﬁrst increases shortly after the typhoon’s landing, and then decreases. Let us reconsider Bauer’s hurricane case we showed before (see Fig. 1). It can be seen that although the foF2 increases with the approaching of hurricane Hazel, it also decreases after hurricane Hazel’s passing (note that his Y -axis is inverse), which is consistent with that during the period of the three typhoons. Therefore, for the four cases above, we can see that the foF2 in two cases increase ﬁrst, but not in another two cases, when the typhoons (hurricanes) are approaching. That is to say, whether the foF2 increases or not with an approaching typhoon is uncertain. However, it does decrease after typhoon’s (hurricane’s) landing on the continent. Table 2 shows the comparison of the three typhoons in terms of the distances between the landing location and both sounding stations (DG and DH ), the maximal perturbation of the ionosphere (MG and MH ) over Guang-Zhou and Hai-Kou stations, the duration of the perturbation (Td ) and the interval between the landing time and the time of minimal foF2 (Tdelay ). It appears that the magnitude and the duration of perturbation are positively correlated to the wind power. 3. Discussion For an atmosphere (e.g., that of the earth or Mars), below the turbopause, because of the existence of turbulence, every components of the atmosphere

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

356

ch29

Y.-M. Liu, J.-S. Wang and Y.-C. Suo

distribute with the same scale height, while above the turbopause, each component distributes mainly according to their respective scale height. The altitude of turbopause is deﬁned as the altitude where the value of the molecular (Dm ) and the turbulent diﬀusion (Dt ) coeﬃcients are equal, i.e., Dm = Dt .7 The formula for the turbulent diﬀusion coeﬃcient is8 :

3 H(z0 ) dz exp Dt (z) = Dt (z0 ) , (1) H(z) δ H(z) where H(z) is the scale height of the neutral atmosphere and δ is a constant to be determined. The molecular diﬀusion coeﬃcient (Dm ) is inversely proportional to the neutral density.7 The turbulence is formed mainly via breaking of buoyancy waves, moving from the surface upwards in the atmosphere of a planet.8 Izakov8 has calculated the altitude proﬁles of the turbulent Dt (z) and molecular Dm (z) diﬀusion coeﬃcients for quiet conditions and for a dust storm on Mars. His results show that the homopause altitude rises from 125 km to about 150 km (see Fig. 6) during the dust storm, which agrees well with the observations. He argues that this occurs due to the increase of turbulent diﬀusion coefﬁcient near the surface of the planet (Dt (z0 ) increases) and the expansion

Fig. 6. Vertical proﬁles of turbulent (Dt ) and molecular (Dm ) diﬀusion coeﬃcient in Martian atmosphere. Curves 1 and 3 — Dt and Dm for quiet conditions; curves 2 and 4 — Dt and Dm for a dust storm (from Ref. 8).

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Eﬀects of Typhoon on the Ionosphere

ch29

357

of the atmospheric column during the dust storm. Wang and Nielsen9 analyzed the behavior of the Martian dayside electron density peak during the global dust storm and found that the rising of the turbopause during the dust storm can change the structure of the upper atmosphere and ﬁnally inﬂuence the ionosphere remarkably by means of photochemical reactions. In the same way, we argue that during the period of typhoon, violent air–ocean or air–land interactions will greatly strengthen the turbulence in the lower atmosphere and this will lead to the increase of the turbulent diﬀusion coeﬃcient Dt , just as Izakov’s calculation result shows. Since the molecular diﬀusion coeﬃcient Dm is inversely proportional to the neutral density, the expansion of the lower atmosphere during typhoon will lead to the decrease of Dm . Both Dm and Dt are increasing with altitude, and Dm is less than Dt below the turbopause, but becomes larger than Dt above the turbopause.8 An increase of Dt and a decrease of Dm will both lead to the altitude increase of the turbopause and ﬁnally change the structure of the upper atmosphere. Figure 7 shows distributions of the atmospheric components before and after the increase of turbopause. It can be seen that the neutral densities increase after the rising of the turbopause. Since the upper atmosphere is the background for the process of photochemical reactions in the ionosphere, the change of the structure of the upper atmosphere will inﬂuence the ionosphere correspondingly. Here we use our model10 to calculate such inﬂuence. Figure 8 shows the proﬁles of the electron density obtained from

Fig. 7.

Variations of neutral atmosphere due to the rising of turbopause in altitude.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

358

ch29

Y.-M. Liu, J.-S. Wang and Y.-C. Suo

Fig. 8. Comparison among proﬁles of electron density obtained from our model, IRI2000 and observation.

Fig. 9.

Variation of foF2 due to the rising of turbopause in altitude.

our model and IRI2000. The good agreement of our modeling results to both IRI2000 and observations conﬁrms that our model is reasonable. The modeling results of the foF2 variations due to the moving of the turbopause are shown in Fig. 9. It can be seen that with the increasing of turbopause the foF2 decreases. Rishbeth et al.11 generalized that foF2 ∼ (qm /βm )1/2 , where qm is the production rate and βm is the recombination rate of electron, just as it would be in the absence of diﬀusion. According to our analysis above, the turbulence in the lower atmosphere will be strengthened during the time of typhoon and lead to the increase of the turbopause in altitude. This will, as shown in Fig. 7, cause the increase of the neutral densities in the ionosphere. Because βm is proportional to

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Eﬀects of Typhoon on the Ionosphere

ch29

359

the neutral density, the increase of neutral density will lead to the increase of βm and ﬁnally lead to the decrease of foF2 , and this may explain the observations that the magnitude and the duration of perturbations seem to be positively correlated to the wind power. That is because the stronger the wind is, the stronger the turbulence will be, and this will lead to the higher altitude of the turbopause, and ﬁnally lead to the lower electron density. The air–land interactions will more eﬀectively strengthen the turbulence in the lower atmosphere than air–ocean interactions will. Guang-Zhou station is located on the continent of China, while Hai-Kou station is surrounded by the ocean. So air–land interactions are stronger in Guang-Zhou than that in Hai-Kou. Therefore, in the case of Durian, though DG > DH , the maximal perturbations at Guang-Zhou and that at Hai-Kou are nearly the same. In the case of Maggie, DG < DH , and the turbulence of the lower atmosphere in Guang-Zhou is stronger than that in Hai-Kou, the maximal perturbation in Guang-Zhou is greater than that in Hai-Kou, while in the case of Penny, DG = DH , the maximal perturbation in Guang-Zhou is still greater than that at Hai-Kou. 4. Summary The study of the correlation between the ionosphere and the severe meteorological phenomena in the troposphere is very interesting. We analyze the ionospheric responses to three selected typhoons during which the geomagnetic condition is very quiet. The results show that after typhoon’s landing, the ionospheric foF2 tends to decrease (for typhoon Penny, it increases ﬁrst and then decreases) and the magnitude and the duration of such disturbance seem to be positively correlated to the wind power. Finally, the increase of the turbopause due to the typhoon activities is introduced to explain such phenomena. The modeling results agree well with the observations. The reason why the foF2 increases ﬁrst in the case Penny needs to be further discussed. Acknowledgments The work was supported by the Chinese National Science Foundation under grant 40374058 and the National Key Laboratory of Electromagnetic Environment under grant 51486010103JW0201. The air pressure data are from the Meteorological Data Services of China Meteorological Administration. The Dst data are from Space Physics Interactive Data Resource of National Geophysical Data Center (http://spidr.ngdc.noaa.gov).

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

360

ch29

Y.-M. Liu, J.-S. Wang and Y.-C. Suo

References 1. D. M. Baker and K. Davies, F2 -region acoustic waves from severe weather, J. Atmosph. Terr. Phys. 31 (1969) 1345–1352. 2. K. Davies and J. E. Jones, Ionospheric disturbances in the F2 region associated with severe thunderstorms, J. Atmosph. Sci. 28 (1971) 254–262. 3. K. Davies and J. E. Jones, Acoustic waves in the ionospheric F2 -region produced by severe thunderstorms, J. Atmosph. Sci. 35 (1973) 1737–1744. 4. C.-S. Shen, The correlations between the typhoon and the foF2 of ionosphere, Chinese Journal of Space Science 2, 4 (1982) 335–340. 5. N. V. Isaev, V. M. Sorokin, V. M. Chmyrev and O. N. Serebryakova, Ionospheric electric ﬁelds related to sea storms and typhoons, Geomagnetism and Aeronomy 42, 5 (2002) 670–675. 6. S. J. Bauer, An apparent ionospheric response to the passage of hurricanes, J. Geophys. R 63 (1958) 265–269. 7. M. D. Yamanaka, Homopause control by gravity wave breaking in the planetary atmospheres, Adv. Space. Res. 15, 4 (1995) 47–50. 8. M. N. Izakov, Turbulence and anomalous heat ﬂuxes in the atmospheres of Mars and Venus, Planet. Space Sci. 49 (2001) 47–58. 9. J.-S. Wang and E. Nielsen, Behavior of the Martian dayside electron density peak during global dust storms, Planet. Space Sci. 51 (2003) 329–338. 10. Y. Deng, J. Wang and Z. Xiao, A physical model for ionospheric vertical proﬁle and comparison between it and IRI-90, Chinese Journal of Space Science 20, 2 (2000) 103–112. 11. H. Rishbeth and O. K. Garriott, Introduction to Ionospheric Physics (Academic Press, New York and London, 1969), pp. 151–154.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch30

FORMATION AND OBSERVATIONS OF SHADOW BANDS DURING THE TOTAL SOLAR ECLIPSE OF NOVEMBER 23, 2003 NEAR MAITRI, ANTARCTIC HARI OM VATS∗,§ , S. P. BAGARE†,¶ and S. M. BHANDARI‡,¶ ∗Physical †Indian

Research Laboratory, Ahmedabad 380009, India

Institute of Astrophysics, Bangalore 560034, India

‡Space

Application Centre, Ahmedabad 380015, India §[email protected]

The phenomenon of shadow bands takes place due to the scattering of sunlight by the irregularities in the local atmosphere. Observation of shadow bands were made just before and after the total solar eclipse (TSE) on November 23, 2003 near the Indian Antarctic station Maitri (long. 11◦ 45 E and lat. 70◦ 45 S). The formation and observations of shadow bands during this TSE are presented. The results of correlation analysis show that the correlation (1) falls to half at separation of ∼ 35 cm in horizontal direction; and (2) also at ∼ 100 ms time lag. The power spectrum of the shadow band ﬂuctuations had a power law distribution (spectral index ∼ −2) which is diﬀerent from earlier eclipse observations reported in the literature. However, it closely resembles theoretical predictions of the extended screen scintillation theory.

1. Introduction The phenomenon of shadow bands occurs just before and immediately after a total solar eclipse. The reason for this is the scattering of light from the thin uneclipsed solar arc by air turbulence. The geometry of the solar eclipse plays a very important role in the duration and clear visibility of these bands. In most total solar eclipses this phenomenon last for a very brief period (< 1 min) and it is also very faint. Thus the observations and recording of this have been very rare. The path of totality of the solar eclipse of November 23, 2003 was over the Antarctic region. From the circumstances of this eclipse and on the basis of refractive–diﬀractive scattering approach of Booker and Vats,1 it was realized that shadow bands will be clearer and will last longer than those during other eclipses. The Indian solar eclipse team set up a nonreﬂective screen facing the Sun and recorded ¶ Member

of the Indian Eclipse Expedition Team to Maitri, Antarctic. 361

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

362

ch30

H. O. Vats, S. P. Bagare and S. M. Bhandari

the phenomenon with a fast video camera at Maitri (long. 11◦ 45 E and lat. 70◦ 45 S). It was found that the shadow bands became visible 4 min before totality and lasted for nearly 7 min after totality. It was an extremely exciting event to observe and investigate.2 The bands were easily distinguished by naked eye and were perceived as moving patterns enhanced against a stationary background. These bands were somewhat disorganized, but did tend to form a quasi-linear pattern almost parallel to the tangent to the center of solar crescent. This justiﬁes their common name. In the early days observers have seen these and usually made an artistic picture based on the observations and some have made marks on pre-erected screens on a site of totality. The paucity of real scientiﬁc records of this phenomenon led to the proliferation of exotic proposals to explain the shadow bands. Marschall et al.3 and Codona4 summarized the earlier observations and belief about the shadow bands phenomenon. Now it is known that these patterns are linked to the phenomenon of stellar scintillation usually termed as twinkling of stars5 and the propagation of light through the atmospheric turbulence near the ground (Refs. 1 and 6 and references therein). The work of Booker and Vats1 is based on the propagation of laser light horizontally near the ground at a height of ∼ 1 m. During the total solar eclipse (TSE) of November 23, 2003 the path of totality was over the Antarctic region and the Sun’s elevation was from grazing to a maximum of 15◦ , at the observing site near Maitri, Antarctica, it was ∼ 2◦ .7 only. Thus it was like horizontal propagation of light from visible crescent Sun before and after totality. The approach of Booker and Vats1 describes light propagation similar to this. Here we report the observation and preliminary analysis of the shadow bands of this eclipse. Further details of the circumstances of this eclipse may be found in “Eclipse predictions (maps) courtesy of Fred Espenak, NASA/GSFC” (http://sunearth.gsfc.nasa.gov/eclipse).

2. Shadow Bands Formation and Observations The basic principle of shadow band phenomenon can be easily understood by the constructive and destructive interference of the randomly scattered rays from the solar crescent. Here the scattering is due to the turbulent patches present in the local atmosphere. Constructive and destructive interference gives high and low intensities, respectively. Thus each point source on the solar crescent would form a pattern on the ground. Since the sources are aligned along the visible arc (solar crescent just before and just after

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch30

Formation and Observations of Shadow Bands

363

Intensity in arbitrary units

250

240

230

220

0

5

10

15

20

Time in seconds

Fig. 1. A sample of time sequence of shadow bands intensities at the selected pixel (150, 110) on the screen.

totality), these patterns produce band like structures. In practice the bands are not uniform. Since atmospheric turbulence changes with time, altitude, and local wind velocity, the pattern changes with time and moves in space. On average the bands are aligned almost parallel to line joining the two ends of the crescent Sun. The shadow bands were recorded by video camera and the image sequence was extracted using software. Vats et al.7 presented basic details of this experimental set up and the examples of shadow band structures as seen on November 23, 2003. The recording frames had a digital size of 384 × 288 pixels. A pixel whose (x, y) coordinates on the recorded frames was (150, 110) was selected for the initial temporal analysis of the shadow band pattern. The time intensity series was obtained by sampling intensity values at the selected pixel (150, 110). This typical intensity time sequence was sampled every 40 ms as shown in Fig. 1. The correlation and spectrum analysis are presented in Sec. 3.

3. Correlation and Power Spectrum From the recording set up, it was realized that 5 pixels correspond to ∼ 4 cm along the local horizontal or vertical direction. Using this pixel — distance scaling, eight temporal series separated by ∼ 0–28 cm in the steps of 4 cm

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

364

ch30

H. O. Vats, S. P. Bagare and S. M. Bhandari

were selected for the cross-correlation analysis. These eight time series are formed by intensities at pixels (150, 110), (155, 110), (160, 110), (165, 110), (170, 110), (175, 110), (180, 110), and (185, 110) from the sequence of shadow band images. The cross-correlations were calculated between these time series taking the time series at pixel (150, 110) as the reference. Figure 2 shows these calculations for a time lag −0.25–0.25 s for all eight time series. One of these curves has a maximum value of 1 for 0 time lag and is also symmetric. This is for the time series at pixel (150, 110) with itself and hence it is auto correlation function. From this curve (auto correlation function) one can estimate that correlation would fall from a maximum value of 1–0.5 at time lag of ∼ 92 ms. It is observed that all other correlograms are asymmetric. The curves in Fig. 2 are very interesting. The maximum (peak) value of correlation coeﬃcient decreases and shifts towards more negative time lags as the separation increases. However, there is one exception in that the maximum value in last curve is slightly higher than that in last but one, however, the time lag of maximum correlation still increases with the larger separation. This separation is only along the x-axis of the image which was almost horizontal.

1.0

Cross correlation

0.8

0.6

0.4

0.2

0.0 -0.2

-0.1

0.0

0.1

0.2

Time lag in seconds

Fig. 2. Eight correlograms of time series of the intensity values from the individual frames. The correlograms are for separation in the range 0–28 cm in steps of 4 cm. It may be noted that the maximum cross-correlation decreases and its position shifts toward negative time lags in each curve.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

Formation and Observations of Shadow Bands

ch30

365

The variation of time lag and maximum value of cross-correlation as a function of separation along x-axis or along the horizontal direction are shown in Figs. 3 and 4, respectively. Figure 3 shows that the time lag of maximum cross correlation increases almost linearly with separation of shadow band time series along the x-axis. It is possible to ﬁt a line through the points in 80

Time lag in msec

60

40

20

0 0

5

10

15

20

25

30

Seperation in cm

Fig. 3. Variation of time lag of maximum correlation as function of separation of the time series along the x-axis.

Maximum Correlation value

1.0

0.9

0.8

0.7

0.6

0.5 0

5

10

15

20

25

30

Seperation in cm

Fig. 4. Variation of maximum cross-correlation value with separation of points at which the time series occur. The separation is along the x-axis.

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

366

ch30

H. O. Vats, S. P. Bagare and S. M. Bhandari

Fig. 3 which has slope ∼ 2.86. This would give time lag of ∼ 100 ms for a separation of ∼ 35 cm. The maximum value of cross-correlation with further higher separations would decrease and eventually become negligible. The rate of decrease slows down beyond a separation of ∼ 10 cm. The rate of decrease is ∼ 0.01 cm−1 beyond 10 cm as against 0.074 cm−1 up to 10 cm. Although the values on x-axis in this ﬁgure are from 0 to 30 only, however, from the rate of the decreasing trend it is possible to extrapolate that the value of maximum cross-correlation of 0.5 between the two series would happen if these are separated by a distance of ∼ 35 cm. From Fig. 2, it is estimated that the correlation falls symmetrically to 0.5 at a time lag of ∼ 92 ms and an equivalent decrease in cross correlation is at a distance of ∼ 35 cm. This gives the characteristic pattern drift speed ∼ 3.8 m/s along the x-axis. The characteristic pattern drift of ∼ 3.8 m/s is in agreement with that reported by Jones and Jones8 for their observations of shadow bands during the eclipse of July 11, 1991. The full cross-correlation analysis would give other component of the drift speed and hence would enable one to determine characteristic scale size of the turbulence and the direction of its motion. The samples of time series were obtained from each minute of the shadow band records (4 and 7 min long for just before and just after totality, respectively). Each sample was processed using a fast Fourier transform of the shadow band intensities. This enabled us to get a set of 11 power spectra. An average of these 11 power spectra is shown in Fig. 5. This power spectrum has a low-frequency roll oﬀ ∼ 0.65 Hz. and a power-law shapes up to the highest frequency (12.475 Hz). There appears to be a linear slope of the spectral power beyond the roll oﬀ frequency. The estimated slope is ∼ −2, this is shown by straight line in the high-frequency part of the spectrum. The value of spectral slope is very close to the Kolomogrov spectral slope of −5/3 used in the theoretical work of Booker and Vats1 for a case study of horizontal propagation of laser light over a distance of 1–8 km. The overall shape of the spectrum agrees reasonably with theoretical spectrum.1

4. Conclusions The cross-correlation and spectrum analysis of the shadow band patterns observed and recorded near Maitri during the total solar eclipse on November 23, 2003 are presented here. This eclipse took place, when the Sun was at grazing angle from the local horizon. The duration of shadow band patterns were very much extended before and after totality. The

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

ch30

Formation and Observations of Shadow Bands

367

Log Power (arbitrary units)

2

1

0

-1

-2 0.1

1

10

Frequency (in Hz) Fig. 5. Average of 11 power spectra of shadow band intensity recorded on November 23, 2003 at Maitri.

spectrum of the observed shadow band ﬂuctuation indicate the presence of Kolomogrov or power law (with a spectral slope ∼ −2) type turbulence in the local troposphere. This matches fairly well with theoretical work,1 however, diﬀers signiﬁcantly from the spectra of shadow bands observed at diﬀerent locations usually with higher elevation angle of the Sun (Ref. 9 and references therein). Codona6 deﬁned two regimes for the temporal intensity spectrum slopes as −5/3 and −17/3. The present spectral shape (Fig. 5) is quite in contrast, possibly due to the observing geometry and the location. The preliminary correlation analysis gives a typical characteristic scale size of ∼ 35 cm and pattern drift speed ∼ 3.8 m/s along the x-axis. More detailed analysis and model simulations are needed to understand these observations and their diﬀerences with other shadow band studies reported in the literature.

Acknowledgments The authors acknowledge ﬁnancial and the technical support of the Department of Ocean Development, Government of India and NCOAR, Goa. The encouragement received by us from the inception of this experiment was enormous; no words can suﬃciently express our gratitude. The enthusiastic

May 30, 2006 14:41 WSPC/SPI-B368 Advances in Geosciences Vol. 2

368

ch30

H. O. Vats, S. P. Bagare and S. M. Bhandari

support extended by the members of the wintering Maitri station team made this work possible in diﬃcult conditions.

References 1. H. G. Booker and H. O. Vats, Rad. Science 20 (1985) 833. 2. S. M. Bhandari, S. P. Bagare and H. O. Vats, Observations of unique extended duration shadow band activity during the total solar eclipse of November 23, 2003 near Maitri, Antarctica, National W’shop on Indian Antarctic Research — A Status Review, Goa, India (2004). 3. L. A. Marschall, R. Mohan and R. C. Henry, A. Optics 23 (1984) 4390. 4. J. L. Codona, Sky Tel. 60 (1991) 482. 5. A. T. Young, Sky Tel. 38 (1969) 309. 6. J. L. Codona, Astron. Astrophys. 164 (1986) 415. 7. H. O. Vats, S. M. Bhandari and S. P. Bagare, Proc. URSI Session on “Scattering and Diﬀraction”, New Delhi, India (2005). 8. B. W. Jones and C. A. L. Jones, J. Atmos. Terr. Phys. 15 (1994) 1535. 9. B. W. Jones, J. Atmos. Terr. Phys. 61 (1999) 965.