Seasonal Forecasts, Climatic Change and Human Health
Madeleine C. Thomson • Ricardo Garcia-Herrera Martin Beniston Editors
Seasonal Forecasts, Climatic Change and Human Health Health and Climate
Madeleine C. Thomson International Research Institute for Climate and Society Columbia University LDEO, Palisades New York 10964 USA
Ricardo Garcia-Herrera Departamento Fisica de la Terra II Facultad de Ciencias Fiscas Universidad Complutense de Madrid Ciudad Universitaria 28040 Madrid, Spain
Martin Beniston University of Geneva Site de Battelle/D7 chemin de Drize CH-1227 Carouge Geneva, Switzerland
ISBN 978-1-4020-6876-8 e-ISBN 978-1-4020-6877-5 DOI 10.1007/978-1-4020-6877-5 Library of Congress Control Number: 2007942723 © 2008 Springer Science + Business Media B.V. No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Printed on acid-free paper 9 8 7 6 5 4 3 2 1 springer.com
Acknowledgements
We would like to thank the following for their contributions as presenters, rapporteurs, panelists and discussants who participated in the Climate, Climatic Change and Human Health workshop1 (Wengen Switzerland, September 12-14, 2005), co-sponsored by the Graduate School of International Studies (Geneva, Switzerland), NOAA-OAR (Boulder, Colorado, United States), the University of Acre (Acre, Brazil), WHO (Rome, Italy), IRI, Columbia University (New York, United States), and the EU-ENSEMBLES Project (coordinated by the Hadley Center, UK Met Office, Exeter, United Kingdom). Joan Aron, Paul J. Beggs, Martin Beniston, Raquel R. Cesario, Manuel Cesario, Andrew Comrie, Stephen J. Connor, Charles Delacollette, Henry Diaz, Diane P. Dupont, Kris Ebi, Roger Few, Regula Gehrig, Joelle Goyette-Pernot, Renate Hagedorn, Chris Hewitt, Andre Kamga, Kim Knowlton, Patrick Kinney, Panagiotis Nastos, Gilma C. Mantilla, Vincent Martin, Simon Mason, Andreas Matzarakis, Glen McGregor, Bettina Menne, Ana Rosa Moreno, Andy Morse, Eduardo R. Palenque, Xavier Rodo, Jacinthe Seguin, Lennie Smith, Michel Thibaudon, Yves M. Tourre, Guojing Yang and Ksenija Zaninovic. The meeting was divided into two sessions: Seasonal climate forecasting for health (ENSEMBLES/IRI/WHO) for which special thanks go to Andy Morse for organizing the funding for participation in this session through the ENSEMBLES project. Climate change and health (GSIS and NOAA-OAR, ACRE session) which was broken up into sections covering: Pollen, allergies and climate; Extreme events and health; Water and vector-borne diseases and Policy and adaptation. We would also like to thank Molly Hellmuth, Sylvie Bovel-Yerly, and Maria Salgado for editorial support and David Rogers and Mohammed Boulahya for valuable suggestions with regard to ‘recent developments’ as well as the anonymous reviewers of the manuscripts for their helpful comments. Madeleine C. Thomson Ricardo Garcia-Herrera Martin Beniston 1
http://www.unige.ch/climate/Workshops/wengen.html
v
Contents
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
v
Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ix
Part I
Seasonal Climate Forecasting for Health . . . . . . . . . . . . . . . . Stephen J. Connor, Madeleine C. Thomson, and Bettina Menne
Chapter 1
The Impact on Human Health of Climate and Climate Change: Research in the ENSEMBLES Project from Seasonal to Centennial Timescales. . . . . . . . . . . . . . . . . Chris D. Hewitt
Chapter 2
“Flowering Walnuts in the Wood” and Other Bases for Seasonal Climate Forecasting. . . . . . . . . . . . . . . . . . . . . . . Simon J. Mason
Chapter 3
Climate and Infectious Diseases. . . . . . . . . . . . . . . . . . . . . . . . Louise Kelly-Hope and Madeleine C. Thomson
Chapter 4
Integration of Seasonal Forecasts into Early Warning Systems for Climate-Sensitive Diseases such as Malaria and Dengue. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Stephen J. Connor and Gilma C. Mantilla
Chapter 5
Chapter 6
Climate and Environmental Information and the Control of Epidemic Meningitis in Africa. . . . . . . . . . . . . . . . . . . . . . . Luis E. Cuevas, C. Anthony Hart, and Madeleine C. Thomson
1
5
13
31
71
85
A Multimodel Framework in Support of Malaria Surveillance and Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 Daniel Ruiz, Stephen J. Connor, and Madeleine C. Thomson
vii
viii
Contents
Recent Developments and Next Steps in Seasonal Forecasting and Health. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 Madeleine C. Thomson, Stephen J. Connor, and Bettina Menne Part II
Climate Change and Health . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 Martin Beniston and Ricardo García-Herrera
Chapter 7 The RedGems Information System: Climate, Environment and Public Health . . . . . . . . . . . . . . . . . . . . . . . 135 Yves M. Tourre, Jean P. Lacaux, Antonio Güell, M. LaFaye, Philippe Sabatier, Eliane Coëffier, and Laurent Braak Chapter 8 Impacts of Heat and Ozone on Mortality Risk in the New York City Metropolitan Region Under a Changing Climate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 Kim Knowlton, Christian Hogrefe, Barry Lynn, Cynthia Rosenzweig, Joyce Rosenthal, and Patrick L. Kinney Chapter 9 Physiological Equivalent Temperature as Indicator for Impacts of Climate Change on Thermal Comfort of Humans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 Andreas Matzarakis and Bas Amelung Chapter 10 Weather, Ambient Air Pollution and Bronchial Asthma in Athens, Greece . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 Panagiotis Th. Nastos Chapter 11 Outreach from the IAI Collaborative Research Network on Climate Variability and Human Health Impacts in the Tropical Americas. . . . . . . . . . . . . . . . 189 Joan L. Aron Recent Developments and Next Steps in Climate Change and Health . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 Patrick L. Kinney and Ricardo García-Herrera, and Stephen J. Connor Colour Plates: Appendix 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207 Colour Plates: Appendix 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213 Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223 Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
Contributors
Bas Amelung International Centre for Integrative Studies, University of Maastricht Joan L. Aron Science Communication Studies, Columbia, MA 21045, USA Martin Beniston University of Geneva, Climate Change and Climate Impact – Battelle, Building D, 7 route de Drize, CH 1227 Carouge, Switzerland Laurent Braak MEDES, Toulouse, France Eliane Coëffier Institut Pasteur, Paris, France Stephen J. Connor International Research Institute for Climate and Society, The Earth Institute at Columbia University, Lamont Campus, Palisades, NY 10964, USA Luis E. Cuevas Liverpool School of Tropical Medicine, Pembroke Place, Liverpool, L3 5QA, UK Antonio Güell CNES, Toulouse, France C. Anthony Hart Department of Medical Microbiology and GU Medicine, University of Liverpool, Liverpool, UK Chris D. Hewitt Met Office, Hadley Centre for Climate Prediction and Research, Fitzroy Road, Exeter EX1 3PB, UK Christian Hogrefe State University of New York, Albany, NY, USA
ix
x
Contributors
Louise Kelly-Hope Liverpool School of Tropical Medicine, Pembroke Place, Liverpool, L3 5QA UK Patrick Kinney Department of Environmental Health Sciences, Mailman School of Public Health, Columbia University, New York, USA Jean Pierre Lacaux MEDIAS-France, 18 Avenue Edouard Belin, B.P. 2102, FR-31401 Toulouse Cedex 9, France Barry Lynn Goddard Institute for Space Studies, New York, USA Gilma C. Mantilla National Institute of Health, Ministry of Social Protection, Bogota, Colombia Simon J. Mason International Research Institute for Climate and Society, The Earth Institute at Columbia University, Lamont Campus, Palisades, NY 10964, USA Andreas Matzarakis Meteorological Institute, University of Freiburg, Freiburg, Germany Panagiotis Th. Nastos Laboratory of Climatology and Atmospheric Environment, University of Athens, Greece Joyce Rosenthal Graduate School of Architecture, Planning & Preservation, Columbia University, New York, USA Cynthia Rozenzweig Goddard Institute for Space Studies, New York, USA Daniel Ruiz Escuela de Ingeniería de Antioquia, Calle 25 Sur No. 42–73, Envigado, Antioquia, Colombia Philippe Sabatier INRA/Veterinary School of Lyon, France Madeleine C. Thomson International Research Institute for Climate and Society, The Earth Institute at Columbia University, Lamont Campus, Palisades, NY 10964, USA Yves M. Tourre MEDIAS-France, 18 Avenue Edouard Belin, B.P. 2102, FR-31401 Toulouse Cedex 9, France
Part I
Seasonal Climate Forecasting for Health Stephen J. Connor1, Madeleine C. Thomson1, and Bettina E. Menne2
Introduction At the Ministerial Conference on Epidemic Prevention and Control in the Great Lakes Region, Kampala, June 2003; the Budapest Ministerial Conference for Environment and Health, in 2004; and the World Conference on Disaster Reduction, in Kobe in 2005, the international community requested a programme with specific priorities to assist countries in building early warning systems that would truly reach and serve those at risk. Collaborative efforts between the health community and the climate research community through for example, the European Union funded ENSEMBLE1 and AMMA2 projects have provided a major impetus to developing the scientific basis for integrated seasonal climate based health early warning systems.
Seasonal Climate Forecasts and Their Application to Health This section of the workshop3 focussed on the fledgling experimental and operational use of seasonal forecasts in health early warning systems (both for infectious disease and for extreme weather events) and discussed scientific, policy and practical opportunities and obstacles presented by this new technology. Some seasons are unusually hot or cold or unusually wet or dry. Such seasonal anomalies in the climate may have significant implications for seasonal variations 1
The International Research Institute for Climate and Society, Earth Institute at Columbia University, New York 10964, USA
2
The European Centre for Environment and Health, WHO-Europe, via Francesco
1
http://www.ensembles-eu.org/ http://amma.mediasfrance.org/index 3 Climate, Climatic Change and Human Health Workshop (Wengen Switzerland, September 12-14, 2005) 2
M.C. Thomson et al. (eds.), Seasonal Forecasts, Climatic Change and Human Health. © Springer Science + Business Media B.V. 2008
1
2
S.J. Connor et al.
in climate sensitive diseases including the occurrence of epidemics (WHO 2004) and in the occurrence of extreme climate events (e.g. hurricanes or heatwaves). We now know that these seasonal climate anomalies result from complex interactions between the atmosphere and the underlying surfaces: that is, the world oceans and land surfaces. The atmosphere is particularly sensitive to tropical sea surface temperature (SST) anomalies such as those that occur in association with the El Niño/ Southern Oscillation (ENSO). From the early discovery that seasonal climate could be modeled from sea surface temperature our capacity to predict rainfall and temperature anomalies through seasonal climate forecasts has improved substantially, largely in response to increased understanding of the El Niño/Southern Oscillation (ENSO) phenomenon and improvements in the way we deal with model uncertainty. Routine, scientifically based, and skillful, seasonal climate forecasts are now possible for specific parts of the world, for specific seasons. In contemporary state-of-the-art seasonal climate predictions, computer models of the atmosphere are generally run in groups, or ensembles, of 10 or more. An ensemble climate forecast system predicts not only the most likely evolution of climate, but also the uncertainty in such a prediction. For each member of the ensemble, the model is run with slightly different “initial” conditions. The differences among the ensemble members give forecasters some measure of the likelihood that a particular seasonal climate state will be above, near, or below normal. Seasonal temperature and precipitation forecasts are couched in terms of shifts in these equal probabilities to favor one or two of the categories (Barnston et al. 2003). It is apparent that these generalised forecasts of rainfall and temperature based on terciles (probabilities of below normal, normal, or above normal) are difficult and sometimes impossible to use within the context of specific health early warning systems and that forecasts tailored to specific disease control problems are required. Recent advances in tailored forecasts for Rift Valley Fever (Normalized Difference Vegetation Index: NDVI), malaria (rainfall) and meningococcal meningitis (dust) were presented and the current status of early warning system developments in malaria, rift valley fever and cholera were discussed. While the latter is clearly a climate sensitive disease and epidemics have been associated with large scale climate and hydrological processes, as for many other climate-sensitive diseases there is an important interplay between the extrinsic and intrinsic factors in determining epidemic outbreaks. The fledgling operational use of SCF in Africa for both human health, e.g. malaria in Botswana (Thomson et al. 2005, 2006) and veterinary sector, e.g. rift valley fever in Senegal (Thiongane and Martin 2003) were presented and the importance of using SCF in an integrated early warning system involving vulnerability assessment, environmental/climate monitoring (based on satellite and meteorological station data) and routine surveillance was stressed as was the critical aspect of developing early warning systems around response capability (DaSilva et al. 2004). Such forecast systems require statistical or biological models which capture the impact of climate variables on disease transmission. It was noted that very little research has been done so far on predicting extreme events on the seasonal timescale. For example, neither a precise definition of
Seasonal Climate Forecasting for Health
3
extreme events on the seasonal timescale has been established yet, nor a general consensus has been found on what type of extreme events our attention should be focused on. Another important obstacle for scientific work in this area is the limited availability of long enough data and forecast records for seasonal forecasts. Despite these limitations there is some evidence that skill could be identified in certain regions for certain events. For example in most years, South Indian Ocean tropical cyclones do not make landfall in Mozambique. The main reason is that the tropical cyclone tracks in the South Indian Ocean (SIO) usually have a southward recurvature before reaching the African coast. However, in 2000, two tropical cyclones made landfall in Mozambique with hurricane intensity. In a recent paper (Vitart et al. 2003) it was shown that the main reason for the landfalls in 2000 was the associated characteristics of SIO, ENSO and local sea surface temperatures indicating the potential predictability of cyclones in the region. Skill of climate predictability in the extra-tropics is known to be more limited than that found in the tropics. Results from a study on winter mortality in the UK indicted the importance of seasonal climate descriptors such as such as minimum temperatures and duration of cold spells as determining factors. However the potential value of seasonal climate forecasts for the health sector is constrained by limitations in the predictability of the climate in the northern hemisphere. One important area of the workshop discussion was the potential role that health early warning systems may play in adaptation strategies to climate change. This was highlighted in a presentation of a new project in Colombia (part funded through the World Bank by the Global Environmental Facility on Adaptation to Climate Change). The project aims to strengthen the national disease surveillance system by incorporating SCF, climate monitoring, disease modeling and surveillance in an integrated system which guides control services in more proactive interventions against malaria and dengue. Finally the potential economic benefits of seasonal climate forecasting to the health sector were discussed along with statistical issues surrounding the probabilistic nature of seasonal forecasting and the methods in which this information is captured, interpreted and conveyed to decision makers.
Observations and Recommendations A panel discussion centred on the question ‘can seasonal climate forecasting be useful for health?’ was led by an international panel of: climate scientists, health scientists and policy-makers. It was noted that seasonal climate forecasts were already an operational component of early warning systems for health issues in some countries, e.g. rift valley fever in Senegal and malaria in Botswana. It was recognized that to be effective these SCF need to be part of an integrated early warning and surveillance system with the lead-times and opportunities available for control options to be explicit. Knowledge of the cost effectiveness of earlier intervention options was considered essential.
4
S.J. Connor et al.
The following recommendations emerged from the discussions: ●
●
●
●
The partnership between the health and climate communities should be strengthened and supported at local, national and international level. This partnership be established from the conceptualization stage whenever the development of early warning systems are considered, both at the research and implementation stage. The accountability and capacity of the institutions involved needs to be ensured in order to develop sustainable systems which incorporate seasonal forecasts into a functioning health information and response structure. The rapidly developing knowledge on health and climate linkages and the opportunities for using this information needs to be shared effectively through improved training opportunities. There is a pressing need to develop training curricula, and programmes for short diploma courses, post-graduate degrees, and additional modules to support current public health training programmes, e.g. Tropical Medicine and International Health courses.
With the growing international focus on climate following the G8 summit in Gleneagles, Scotland, and the pressing requirement of international organizations for the delivery of the Millennium Development Goals, there is an important opportunity for engagement between the health and climate community. As many of the MDGs and associated development targets are climate sensitive, and have a direct or indirect health focus, now is a good time to nurture common responsibility towards achieving these goals.
References Barnston A.G., Mason S.J., Goddard L., DeWitt D.E., Zebiak S.E. (2003) Multimodel ensembling in seasonal climate forecasting at IRI. Bulletin of the American Meteorological Society 84:1783–1796. DaSilva J., Garanganga B., Teveredzi V., Marx S., Mason S.J., Connor S.J. (2004) Improving epidemic malaria planning, preparedness and response in Southern Africa. Malaria Journal 3(1):37. Thiongane Y., Martin V. (2003) Systeme sous regional d’alerte et de controle de la Fievre de la Valley du Rift (FVR) en Afrique de l’Ouest. EMPRES/FAO Bulletin 7. Thomson M.C., Mason S.J., Phindela T., Connor S.J. (2005) Use of rainfall and sea surface temperature monitoring for malaria early warning in Botswana. American Journal of Tropical Medicine and Hygiene 73:214–221. Thomson M.C., Doblas-Reyes F.J., Mason S.J., Hagedorn R., Connor S.J., Phindela T., Morse A.P., Palmer T.N. (2006) Malaria early warnings based on seasonal climate forecasts from multi-model ensembles. Nature 439:576–579. Vitart F., Anderson D., Stockdale T. (2003) Seasonal forecasting of tropical cyclone landfall over Mozambique. Journal of Climate 16 (23):3932–3945. WHO (2004) Using Climate to Predict Infectious Disease Epidemics: A Review. WHO-Geneva, pp54.
Chapter 1
The Impact on Human Health of Climate and Climate Change: Research in the ENSEMBLES Project from Seasonal to Centennial Timescales Chris D. Hewitt
Abstract Human health is affected by climatic conditions, and our understanding of climate–health relationships is improving, as is our ability to predict the weather and climate. Predicting natural climate variability and the future climate is a major challenge because of complicated processes and interactions in the Earth system. The best tools for this task are physically based climate models. However, the predictions are inherently probabilistic because of uncertainties in the models and the experimental design. The uncertainties can be explored by carrying out a well-designed set of integrations of climate models, to produce an ensemble of results. An ensemble climate forecast system is being developed within the EU ENSEMBLES project for use across a range of timescales (seasonal, decadal and longer) and spatial scales (global, regional and local). The model system will be used to construct probabilistic scenarios of future climate change and climate variability. The outputs of the ensemble prediction system will be used to drive a wide range of applications including health. In the past, assessments of the impacts of climate change have often used climate means, but in numerous applications, particularly health, it is the climatic extremes that are more important. The focus on extreme events in the ENSEMBLES project will be particularly useful in this regard. The workshop on “Climate, Climatic Change and its Impacts on Human Health” took place at the beginning of the second year of the ENSEMBLES project, which was timely: methodologies and techniques for probabilistic predictions were being developed, which could be useful for climate–health studies, and climate model simulations had been carried out.
Keywords ENSEMBLES, climate, climate change, health impacts
Met Office, Hadley Centre for Climate Prediction and Research, Fitzroy Road, Exeter EX1 3PB, UK, Tel: +44 (0)1392 884520; fax: +44 (0)1392 885681; e-mail:
[email protected]
M.C. Thomson et al. (eds.), Seasonal Forecasts, Climatic Change and Human Health. © British Crown Copyright 2007, the Met Office, UK
5
6
1.1
C.D. Hewitt
Climatic Impacts on Human Health
The effects of weather events and climatic conditions on human health are becoming increasingly well recognised and understood; for example, the Intergovernmental Panel on Climate Change (IPCC) devotes an entire chapter of its previous assessment report to the subject (McMichael et al. 2001); the Fourth Assessment Report will also contain an entire chapter devoted to human health, and articles in this book present more recent work. Human health is affected directly by climatic conditions, particularly extreme events such as cold snaps and heat waves (Koppe et al. 2004). Climate also affects health indirectly by modifying the transmission dynamics of some infectious diseases, such as those transmitted by insects or water (WHO 2004). Determining such impacts on human health is complicated, and different societies have different vulnerabilities to climate and susceptibility to disease. However, our understanding of climate–health relationships is improving, as is our ability to predict the weather and climate. For example, seasonal climate forecasting is now considered skilful enough in some regions to provide useful warnings of epidemics a season ahead, as demonstrated in the development of an early warning system for malaria (Thomson et al. 2006). The basis of seasonal forecasting rests in the ability of climate models to capture the atmospheric response to slow changes in conditions at the Earth’s surface, particularly in large-scale sea surface temperature (SST) patterns. The widespread impacts of tropical Pacific SST anomalies associated with El Niño/La Niña events are the most notable examples, and recent studies have shown that a multi-model ensemble system can make successful predictions of such events (Palmer et al. 2004).
1.1.1
Climate Change Impacts
Although we are developing a better understanding of climate–health relationships, we are faced with the complication of a changing climate. The climate has warmed over the 20th century (Folland et al. 2001), with a global average increase in surface temperature of about 0.6°C. There have been regional changes in precipitation, as well as some changes in extreme rainfall and surface temperature. The IPCC state that it is likely that most of the observed warming over the last 50 years is attributable to human activities (Mitchell et al. 2001), and that human influence will continue to change atmospheric composition throughout the 21st century. This will lead to further changes to temperature and precipitation, as well as changes in the year to year variations of such climatic quantities (Cubasch et al. 2001). Human activities also have the potential to affect more wide-ranging environmental and socio-economic conditions, such as air pollution, and food and water availability and quality, which are important for human health. Therefore, as well as needing to improve our understanding of how the climate is likely to change, we need to improve our understanding of how a change in climate will affect health.
1 The Impact on Human Health of Climate and Climate Change
7
Such future changes in climate have the potential to produce both positive and negative impacts on health. For example, current projections from climate models consistently show a future tendency towards reduced cold stress in winter, but increased heat stress in summer (Tebaldi et al. 2006) which may be exacerbated in urban areas by poorer air quality associated with the warmer conditions. The projections also show a greater frequency of heavy precipitation events, but with substantial geographical variability, and high inter-annual and inter-model variability, which makes it more difficult to determine the impact of these events. Changes in climate may also indirectly create health problems. For example, an increase in the intensity and frequency of extreme events, such as heat waves, storms, floods (arising from heavy-precipitation events as well as coastal flooding which may be exacerbated by sea level rise), or drought, may create humanitarian disasters. Any resettlement into already densely populated areas could worsen health-related problems. In addition, the effect of a change in climate may be difficult to determine or respond to for some diseases, for example vector-borne pathogens could spread into areas where disease currently does not exist or was eradicated in the past, and the effects of such pathogens could change in a different climate regime.
1.1.2
Climate Change Projections
Predicting natural climate variability and the future climate is a major challenge because of complicated processes and interactions in the Earth system operating on a range of spatial and temporal scales. The best tools for this task are physically based climate models of the key components of the Earth system. However, such predictions are inherently probabilistic (Collins et al. 2006), because there are uncertainties in the representation of key processes within the models, uncertainties in how the factors that affect climate (such as atmospheric trace gases) will vary in the future, and uncertainties arising from the initial conditions used for the predictions. Such uncertainties can be explored by making multiple integrations of climate models, i.e. integrations using different representations of processes within a particular model, different models, integrations exploring different scenarios of climate forcing, and integrations using different initial conditions. This approach produces an ensemble of results (see Fig. 1.1 for an illustration of a small ensemble) which allows us to quantify the uncertainty in the climate projections by using statistical techniques.
1.2
The ENSEMBLES Project
The ENSEMBLES project (Hewitt and Griggs 2004) will develop an ensemble climate forecast system for use across a range of timescales (seasonal, decadal and longer) and spatial scales (global, regional and local). The model system will be
8
C.D. Hewitt
Fig. 1.1 An illustrative example of the predicted change, due to a doubling of CO2, in summeraverage precipitation (mm/day) over Europe from a small ensemble of 12 slab model simulations (see Appendix 1)
evaluated by making hindcasts, i.e. simulations of the 20th-century climate, which will be compared against observations and analyses, including a new quality-controlled, high-resolution gridded observational dataset for Europe. The model system will then be used to construct probabilistic scenarios of future climate change and climate variability for quantitative risk assessments, to provide policy-relevant information on climate change and its interactions with society. Emphasis will be placed on changes in extreme events (such as the severity and frequency of heat waves, storms, floods and drought), and the effects of high-impact but low-probability events, like a shutdown of the thermohaline circulation in the North Atlantic. The 5-year project started in 2004 and is a major undertaking by more than 60 institutes from 20 countries, mainly in Europe. It is coordinated by the Hadley Centre at the Met Office in the UK, and is funded under the European Commission’s 6th Framework Programme. The project directly addresses the following key objectives of the IPCC (http:// www.ipcc.ch) and the United Nations Framework Convention on Climate Change (UNFCCC, http://unfccc.int), which includes the Kyoto Protocol: to provide the best available scientific information and assessment on climate change and its impacts; to provide input for policy makers concerning the assessment of dangerous anthropogenic interference with the climate system; to reduce uncertainties in knowledge of the climate system and the adverse impacts of climate change; to promote the development and implementation of education and training programmes; to increase the awareness and public access to information on climate change. The IPCC and the UNFCCC are two of the most important international agencies formulating climate change policy. The findings of the ENSEMBLES project will be
1 The Impact on Human Health of Climate and Climate Change
9
of great practical value to policy makers, stakeholders and the public. The results will be disseminated through web sites and informative leaflets intended to improve the understanding of climate change for wide-ranging audiences, along with the production of policy-oriented publications for bodies such as the UNFCCC and IPCC, and peer-reviewed publications, such as contained within the chapters of this book.
1.2.1
Health Applications Within ENSEMBLES
The outputs of the ensemble prediction system will be used to drive a wide range of applications, such as health, agriculture, energy use, and water resource management. The climatic impacts on some of these applications are linked, for example, adverse climatic impacts on agriculture could subsequently affect health through malnutrition, particularly in areas already vulnerable such as sub-Saharan Africa. The complicated feedbacks to the climate system from some of these impact areas will also be addressed, for example, the economic effect of health impacts and subsequent implications for emissions scenarios will be estimated.
Fig. 1.2 Common domain for regional climate model simulations. The colours show the topographic heights (see Appendix 1)
10
C.D. Hewitt
ENSEMBLES will provide climate and climate change information globally using general circulation models (GCMs). GCMs have a typical horizontal resolution of 1–2° (approximately 110–220 km), which may be too coarse for some studies, so ENSEMBLES will also provide climate and climate change information at higher resolution for a common European domain (Fig. 1.2) using regional climate models (RCMs). The RCMs currently provide information at a horizontal resolution of 0.44° (approximately 50 km), but later in the project the resolution will increase to 0.22°. Towards the end of the project, in 2009, additional RCM simulations will be carried out for part of Africa, Which will be a very useful study region for the effect of climate change on human health.
1.3
The “Climate, Climatic Change and Its Impacts on Human Health” Workshop
The workshop on “Climate, Climatic Change and Its Impacts on Human Health” took place at the beginning of the second year of the project, which was timely: methodologies and techniques for probabilistic predictions were being developed, which could be useful for climate–health studies; GCM simulations for the historical period 1860–2000 had been carried out, along with climate change projections for the 21st century using the IPCC SRES (Special Report on Emissions Scenarios) forcing, the so-called A2, A1B, and B1 scenarios; RCM hindcast simulations for 1961–2000 had begun for the common European domain at 50 km resolution, taking boundary conditions from the 40-year European Re-Analysis (ERA40) dataset. The purpose of the historical hindcast simulations is to create an ensemble for detection and attribution of regional climate change and assessment of regional model performance on interannual and shorter time scales over several decades. The multi-decade period is important so that the capability of the models to simulate variability and extremes can be assessed in different phases of longer-term climate variability. The effect on human health of climate and climate change is clearly of great importance to societies everywhere, particularly to vulnerable low-income populations. The ENSEMBLES project will make an important contribution to ongoing research into the effects, primarily through the use of global and regional model simulations on seasonal-decadal timescales. In the past, assessments of the impacts of climate change have often used climate means, but in numerous applications, particularly health, it is the climatic extremes that are more important. The focus on extreme events in the ENSEMBLES project will be particularly useful in this regard. Acknowledgements ENSEMBLES is a research project (contract GOCE-CT-2003-505539) supported by the European Commission under the 6th Framework Programme 2002–2006. I am grateful for comments from Clare Bryden, Richard Graham, Dave Griggs, Andy Morse and an anonymous reviewer. Figure 2.2 is courtesy of Burkhardt Rockel. For more information about the ENSEMBLES project see the website http://www.ensembles-eu.org.
1 The Impact on Human Health of Climate and Climate Change
11
References Collins, M., B. B. Booth, G. R. Harris, J. M. Murphy, D. M. H. Sexton and M. J. Webb (2006). Towards quantifying uncertainty in transient climate change. Climate Dynamics, in press, doi:10.1007/s00382-006-0121-0. Cubasch, U., G. Meehl, et al. (2001). Projections of future climate change. In: Climate Change 2001: The Scientific Basis, J. T. Houghton et al. (eds). Cambridge University Press, Cambridge, pp. 525–582. Folland, C. K., T. R. Karl, et al. (2001). Observed climate variability and change. In: Climate Change 2001: The Scientific Basis, J. T. Houghton et al. (eds). Cambridge University Press, Cambridge, pp. 99–181. Hewitt, C. D. and D. J. Griggs (2004). Ensembles-based predictions of climate changes and their impacts. Eos 85:566. Koppe C., R. S. Kovats, G. Jendritzky and B. Menne (2004). Heat-waves: risks and responses. WHO Regional Office for Europe, Copenhagen. Health and Global Environmental Change Series, No. 2, http://www.euro.who.int/document/e82629R.pdf McMichael A. J., A. Githeko, et al. (2001). Human health. In: Climate Change 2001: Impacts, Adaptation, and Vulnerability, J. McCarthy et al. (eds). Cambridge University Press, Cambridge, pp. 451–485. Mitchell, J. F. B., D. J. Karoly, et al. (2001). Detection of climate change and attribution of causes. In: Climate Change 2001: The Scientific Basis, J. T. Houghton et al. (eds). Cambridge University Press, Cambridge, pp. 695–738. Palmer, T. N., U. Andersen, P. Cantelaube, M. Deque, F. J. Doblas-Reyes, H. Feddersen, R. Graham, S. Gualdi, J.-F Gueremy, R.Hagedorn, M. Hoshen, N. Keenlyside, A. Lazar, V. Marletto, A. P. Morse, B. Orfila, P. Rogel, J.-M. Terres and M. C. Thomson. (2004). Development of a European ensemble system for seasonal to inter-annual prediction. Bulletin of the American Meteorology Society 85: 853–872. Tebaldi, C., J. Arblaster, K. Hayhoe and G. Meehl (2006). Going to the extremes: an intercomparison of model-simulated historical and future changes in extreme events. Climatic Change 79: 185–211. Thomson, M. C., F. J. Doblas-Reyes, S. J. Mason, R. Hagedorn, S. J. Connor, T. Phindela, A. P. Morse and T. N. Palmer (2006). Malaria early warnings based on seasonal climate forecasts from multi-model ensembles. Nature 439: 576–579. WHO (2004). Using climate to predict disease outbreaks: a review. World Health Organization WHO/SDE/OEH/04.01, http://www.who.int/globalchange/publications/oeh0401/en/index.html
Chapter 2
“Flowering Walnuts in the Wood” and Other Bases for Seasonal Climate Forecasting Simon J. Mason
Abstract Although it is impossible to forecast the weather more than a few days in advance, the science of seasonal climate forecasting is premised upon an ability to predict the general weather conditions over a prolonged period of time, without trying to predict the precise weather at any specific time during that period. The forecasting is possible only because sometimes, and primarily within tropical latitudes, the atmosphere is sensitive to unusual conditions at the earth’s surface, and especially at the sea surface. El Niño, and its counterpart La Niña, are the primary examples of such forcing conditions: during El Niño events, much of the equatorial Pacific Ocean is unusually hot (cold during La Niña), and the consequent changes to the heat and moisture supplied to the atmosphere can disrupt weather conditions in many parts of the globe. However, all seasonal climate forecasts involve a great deal of uncertainty, and a key aspect of forecasting at such time scales is to estimate the uncertainty in the prediction reliably. There are two sources of uncertainty in seasonal climate forecasting: the atmosphere is nowhere completely forced by conditions at the surface, but is free to vary according to its own internal dynamics; and the models used to predict the climate system are imperfect. These two sources of uncertainty are addressed by producing a set of model predictions: different initial weather conditions are used to represent the uncertainty from the internal dynamics, and different models to account for the uncertainties arising from imperfect model physics.
Keywords Seasonal climate forecasting, El Niño, La Niña, uncertainty
International Research Institute for Climate and Society, The Earth Institute at Columbia University, 232 Monell, Lamont Campus, 61 Route 9W, Palisades, New York 10964-8000, USA Tel: +1 845 680 4514; fax: +1 845 680 4865; e-mail:
[email protected]
M.C. Thomson et al. (eds.), Seasonal Forecasts, Climatic Change and Human Health. © Springer Science + Business Media B.V. 2008
13
14
2.1
S.J. Mason
Introduction
The poem Wind on the Hill by A. A. Milne ends with the claim “But where the wind comes from/Nobody knows”, pointing to the perceived impossibility of forecasting the weather. While most people will acknowledge that there have been some advances in weather forecasting over the last half century or so, it is eminently reasonable to ask why scientists can presume to forecast how the atmosphere is likely to behave over the next few months, as they apparently claim to do when issuing seasonal climate forecasts, when it is evident that there is still great inaccuracy in forecasting what is going to happen over the next few days, or even hours. Surely it is much more difficult to forecast the near future than the more distant future. This paradox hinges upon a distinction between “weather” and “climate”, and the fact that there are fundamentally different reasons why each may be predictable. “Weather” can be understood as the state of the atmosphere as it is experienced at any given time, whereas “climate” is some kind of summary (often expressed as a simple average, although involving much more than that) of weather over a longer period of time. Weather forecasting involves making claims about the state of the atmosphere at specific times in the future, whereas seasonal climate forecasting involves making claims only about the general state of the atmosphere over the next few months without having to worry about the precise weather at any time during those months. An analogy can be drawn by considering the problem of forecasting the outcome of a sporting event, such as a soccer match: forecasting the position of the ball more than a few seconds in advance is virtually impossible, but forecasting the final outcome of the match is much more viable. When forecasting the final score, a statement is made only about the general run of play, and no attempt is made to claim knowledge of where the ball will be at any specific time during the match – it is much harder to forecast exactly when goals are going to be scored than what the final score will be. Similarly, with seasonal climate forecasting, it is only the general weather conditions throughout the season that are being forecast, and no attempt is made to forecast the exact weather on any given day within the season. Thus seasonal climate forecasting works on the basis that, although it is impossible to forecast the weather on each of the next 90 days, say, it may be possible to predict which kind of weather patterns may occur unusually frequently or infrequently, or may persist for unusually long periods of time, or be unusually intense. Just as it is possible for the weaker team to dominate the game for a short period, the length of the season for the climate forecast needs to be sufficiently long to ensure that a short spell of otherwise uncharacteristic weather for the season does not negate the forecast. At the same time, if the target season (the period for which the seasonal climate prediction is made) is too long, the end of the season may be too far in the future for the prediction to be reliable, and/or times of the year with fundamentally different climate characteristics, and hence that may require distinct predictions, may be combined. In most cases, a 3-month period is optimal for a seasonal forecast.
2 “Flowering Walnuts in the Wood” and Other Bases
2.2 2.2.1
15
Why the Seasonal Climate Can Be Forecast Historical Examples of Seasonal Forecasting
Even if a distinction between weather and climate forecasting is granted, the question remains of how it is possible to forecast the climate when the weather cannot be forecast accurately – climate, after all, is defined by weather. Although a firm scientific basis for forecasting the climate for the next few months has emerged only over the last few decades, it is a discipline that has been practiced for millennia. In fact, the world’s earliest surviving work of literature, the Epic of Gilgamesh, contains an example of a seasonal climate forecast: Utnapishtim is warned by the god Ea of forthcoming prolonged excessive rainfall, and is told to prepare for flooding. This tale has strong parallels with the story of Noah in the Book of Genesis, and there are numerous other examples of warnings of either excessive rainfall or periods of agricultural drought in the Hebrew Scriptures. In all these examples, warnings are provided by divine revelation, but a much more widespread basis for similar predictions has been inference: specifically, various observations of nature have been interpreted as harbingers of coming weather and climate conditions for centuries (Marriott 1981; Inwards 1994). In Vergil’s Georgics I, for example, a quotation from which forms the title of this chapter, it is observed that heavy blossoming of walnut trees (or sometimes translated as almond trees) typically is followed by weather conditions conducive to bumper harvests. Examples similar to those of Vergil’s, in which weather conditions over the coming season are predicted on the basis of some unusual observation, most frequently in the plant kingdom,1 can be found from virtually all cultures, some of which still survive in popular folklore. In at least a few of these cases, the observations of nature have been shown to have a sound inferential basis (e.g., Orlove et al. 2000, 2002), and cannot simply be written off as the inevitable few cases of Type I statistical errors to be expected given a large number of spurious forecasting methods. These few examples work because in some places, and at some times of the year, weather patterns can persist for long periods of time, or recur with unusual frequency. That some aspects of nature, such as plants, may be sensitive, or that some observation of the atmosphere itself may be made, in the early stages of these prolonged weather anomalies2 is perfectly reasonable. Some modern scientific methods of seasonal climate forecasting work on a similar basis to forecasts based on observations of nature: specifically, various statistical methods of forecasting the climate involve identifying occasions in the past in 1 Observations of animal behaviour are used more frequently for weather than for climate forecasting. 2 “Anomaly” is a technical term used in seasonal climate forecasting to define the difference between the observed weather conditions over a period of time (the “seasonal climate”), and the average weather conditions for that time of year as measured over a period of a few years (typically about 30 years – the “climatology”).
16
S.J. Mason
which climate anomalies at one time of the year have been frequently followed by climate anomalies at a later time of the year. These “analogue” procedures assume that climate anomalies may persist or evolve in predictable ways (e.g., van den Dool 1994; van den Dool et al. 2003). Analogue procedures are familiar to most of us: we often ask questions such as whether the recent hot spell, for example, is likely to continue for the next few weeks, or whether the coming winter is going to be unusually cold given the wet summer that we may have just endured. Unfortunately, many of the methods based on observations of nature cannot be tested scientifically because of inadequate historical records of the relevant observations either because they have not been recorded systematically, or because the observations are inherently subjective and thus difficult, if not impossible, to measure consistently. How does one measure how “heavy” Vergil’s walnut tree blossoming is in any given season, for example? Even given sufficient data, any forecasting system has to be supported by sound causal explanations linking the climate to be predicted with the antecedents because of the possibility of identifying relationships that are purely accidental. Thus, despite a long history of practice, seasonal forecasting became a scientific discipline only in the last few decades, after important advances in understanding of the variability of the atmosphere as described below.
2.2.2
Climate Patterns Around the Globe May Be Related
Given that the atmosphere itself, if left to vary freely, can change markedly in a matter of days, for prolonged weather anomalies to be predictable the atmosphere would have to be constrained or forced somehow (Harrison 2005). Further, if such forcing does occur, it is likely that climate anomalies would occur at a number of locations at the same time. Evidence in support of large-scale forcing of the atmosphere was found before the forcing mechanism itself was identified. In the early 20th century, Gilbert Walker, building upon work initiated in the late-1800s (see Allan et al. (1996) for a review of the early history of seasonal climate forecasting), indicated that unusual climate conditions in one part of the globe are frequently associated with unusual conditions in distant locations, either synchronously or at some lag. Climate anomalies in one part of the globe that are associated with anomalies in another part because of related disruptions in the atmosphere are known as “teleconnections”. The most important teleconnection pattern identified by Walker involves opposite changes in atmospheric pressure between the western and eastern Pacific Ocean (Glantz et al. 1991). This pattern is known as the Southern Oscillation, and is monitored by comparing the atmospheric pressure at Darwin, in northern Australia, with that at Tahiti, in French Polynesia. Seasonally averaged sea-level pressure values for these two locations are compared in Fig. 2.1 (results are shown only for the January–March season), where a tendency is evident for increased (decreased) pressure to occur in Darwin when decreased (increased) pressure occurs in Tahiti. These pressure changes are important because they
2 “Flowering Walnuts in the Wood” and Other Bases
17
Fig. 2.1 January–March averaged sea-level pressure (hPa) at Darwin (grey line) and Tahiti (black line) from 1951 to 2000
involve major disruptions to the trade winds across the southern Pacific, involve large-scale shifts in the location of areas of heavy rainfall, and, in turn, can affect climate conditions in other parts of the globe.
2.2.3
Causes of Seasonal Climate Variability
Only since about the mid-1960s have mechanisms been identified for explaining why certain weather patterns can be so persistent or frequent, and can occur over distant parts of the globe at the same time. Specifically, it was identified that the behaviour of the atmosphere is closely related to the state of the equatorial Pacific Ocean. About every 3–10 years large-scale warming of the equatorial Pacific Ocean occurs, typically lasting about 9–12 months, and most frequently starting in the spring season of the northern hemisphere (Zebiak 1999). These warming episodes became known as El Niño (after an annually occurring much smaller scale warming off the coast of Peru), and are most commonly measured by averaging sea-surface temperatures over core regions of the warming, the most important of which is known as the Niño3.4 region (5° N–5° S, 170° W–120° W). The occurrence of El Niño is closely related to the Southern Oscillation as illustrated in Fig. 2.2. January–March seasonally averaged sea-surface temperatures for
18
S.J. Mason
Fig. 2.2 January–March sea-surface temperatures (°C) averaged over the Niño3.4 region (5° N–5° S, 170° W–120° W) (grey line), and of the Southern Oscillation Index (black line) for 1951–2000
the Niño3.4 region (grey) are shown together with the Southern Oscillation Index (black) for the same period. The y-axis for the Southern Oscillation Index is inverted to emphasise the similarity between the two series. Figure 2.2 shows that when the Pacific Ocean warms the Southern Oscillation Index almost invariably declines – i.e. the difference between the atmospheric pressure in Tahiti and in Darwin decreases, and may even change sign if the warming becomes unusually strong (as in 1983, for example; see Fig. 2.1). Such a strong relationship is highly suggestive of a causal relationship, and indeed the major disruptions in the trade winds over the Pacific Ocean that were mentioned above as being part of the Southern Oscillation are not only required for El Niño events to develop, but the El Niño events themselves cause such prolonged disruptions to the trade winds (Bjerknes 1966, 1969, 1972). The mechanisms involved include the effects of the trade winds on ocean waves and currents, and the effects of the resultant changes in sea temperatures on atmospheric pressure, and hence on the trade winds themselves. A detailed description of these mechanisms is beyond the scope of this chapter, suffice it to say that an interaction between the ocean (El Niño) and the atmosphere (Southern Oscillation) occurs over the equatorial Pacific Ocean. Because the ocean and atmosphere are intricately related in the equatorial Pacific, this coupled system is often referred to as the El Niño – Southern Oscillation (ENSO) phenomenon. The disruption to the atmosphere that is associated with the Southern Oscillation can be manifest downstream of the equatorial Pacific, and so the impacts of El Niño
2 “Flowering Walnuts in the Wood” and Other Bases
19
Fig. 2.3 Correlations between January–March 1961–2000 mean sea-level pressure and the Southern Oscillation Index for the same period. Correlations stronger than ± 0.30 (covering about 40% of the globe) are statistically significant at a 1% level of significance. The mean sea-level pressure data are from the European Centre for Medium-Range Weather Forecasts (ECMWF) Reanalysis (Uppala et al. 2005), which is a model’s best estimate of the state of the global atmosphere given scattered observations (see Appendix 1)
(and La Niña) can occur in disparate parts of the globe. To illustrate such teleconnections, Fig. 2.3 shows correlations between the January–March mean sea-level pressure and the Southern Oscillation Index for the same period. The polar areas of the Oscillation are clearly evident with strong negative correlations over the western Pacific and eastern Indian oceans, and positive correlations in the central Pacific; but correlations are strong elsewhere too, and sea level pressure over about 40% of the globe is notably affected by swings in the Southern Oscillation at this time of year. Some of the teleconnections in Fig. 2.3 are a result of a reorganization of the general circulation of the atmosphere itself (for example, over southern parts of North America), but others may be fuelled or modified by ocean–atmosphere interactions in other areas (for example, over eastern Africa). The tropical Indian Ocean, for example, typically becomes unusually warm during El Niño events because of changes in the trade winds over the basin, and these changes help to effect an influence of ENSO on Africa. However, ocean–atmosphere interaction can occur independently of ENSO; the tropical Atlantic Ocean, for example, is thought to have an important effect on the overlying atmosphere sufficient to affect rainfall over north-east Brazil and over large parts of West Africa. Where and when the interaction between the atmosphere and the ocean is sufficiently strong, the relatively slow timescale of variability of the ocean can have a prolonged influence on the atmosphere, thus making the climate predictable at seasonal timescales (Barnston et al. 1994; Palmer and Anderson 1994; Shukla 1998).
20
S.J. Mason
However, the strength of this interaction permits seasonal climate forecasting only at certain times of the year and in certain parts of the globe (Mason and Goddard 2001). As a rather crude generalization, the atmosphere is most sensitive to changes in sea-surface temperatures in areas where sea temperatures are high, and so the tropical atmosphere tends to be more strongly affected than the atmosphere in higher latitudes. For this reason seasonal predictions of the atmosphere generally are most skilful for the tropics, whereas weather forecasts tend to be more skilful in the extra-tropics where the passage of warm and cold fronts is easier to predict than the vagaries of convection. As mentioned, this tropical – extra-tropical distinction is a crude generalization, and there is considerable spatial and temporal variability in the predictability of seasonal climate, even within the tropics. To illustrate, Fig. 2.4 provides a crude indication of areas within Africa where seasonal rainfall totals are considered predictable. These areas have been identified by isolating areas where
Fig. 2.4 Areas in Africa and seasons for which seasonal climate anomalies are strongly correlated with preceding sea-surface temperature anomalies, and thus which are thought to be predictable
2 “Flowering Walnuts in the Wood” and Other Bases
21
seasonal rainfall is strongly correlated with preceding sea-surface temperatures in at least part of the tropical oceans. The map has been simplified by indicating only those areas in which the strong correlations are spatially coherent. The boundaries of the regions should not be considered precisely marked, and the results are subject to data availability and quality, so other regions and seasons that are not marked may be predictable.
2.3 2.3.1
Approaches to Seasonal Climate Forecasting Forecast Models
For areas such as those shown in Fig. 2.4 where the seasonal climate variability is predictable, forecasts can be made either by considering the historical interaction between the atmosphere and the ocean using a statistical model to describe these observed relationships (e.g., Namias 1991; Ward and Folland 1991), or by trying to model the actual physical processes involved in the interaction. Alternative, but consistent, examples of statistical modelling have already been mentioned. Specifically, analogue procedures relate climate anomalies in one season to preceding anomalies, but are premised on some form of persistent, or predictably evolving, oceanic forcing. Although analogue models are still used, a more widespread approach is to construct a statistical model to describe the historical relationship between sea-surface temperatures at one time of the year, and subsequent seasonal climate anomalies, and then to use this statistical model and the latest observations of sea-surface temperatures to make a prediction. When describing the historical relationship, a lag between the observed sea temperatures and the seasonal climate anomalies is incorporated to provide scope for prediction. The time period between the sea temperature measurement and the seasonal climate is known as the lead-time. It is assumed that the sea temperatures either remain similar to their measured values through the target season, or evolve in a predictable manner. As the lead-time increases, the uncertainty as to the actual sea temperatures that will occur during the target period increases, and so the ability to make an accurate prediction diminishes. The second approach to seasonal climate forecasting involves running a model, known as a dynamical model, very similar to those used to make weather forecasts. Apart from the obvious difference in the length of time for which the forecast is made, when producing seasonal climate forecasts it is also necessary to model the dynamics of at least the influence of the ocean on the atmosphere, and ideally the interactions between the two should be considered. In weather forecasting although the influence of the ocean on the atmosphere can be important (Walker and Lindesay 1989; Barsugli et al. 1999) the feedback can generally be ignored because any changes in the ocean over a few days would typically be small. There are a number of fundamentally different ways of modelling these interactions between the atmosphere and the ocean. These approaches range in complexity:
22
S.J. Mason
in the simplest case, only the atmosphere is forecast using dynamical models while the ocean is kept unchanged; in the most complex case the dynamics of the atmosphere and the ocean evolve together. In the simpler case in which it is only the atmosphere that is modelled, accurate seasonal climate forecasts require accurate forecasts of the ocean to be made prior to running the atmospheric model (Bengtsson et al. 1993). A “two-tiered” forecast is thus required: forecasts of the ocean are made first, followed by forecasts of the atmosphere with the ocean conditions prescribed (Barnston et al. 2005). Forecasts of the ocean have involved methods from as simple as assuming that the most recently observed conditions will remain unchanged, or perhaps evolve slowly back towards their long-term average conditions (Mason et al. 1999; Graham et al. 2000, 2005), through more sophisticated statistical forecasts, methods based on dynamical models of the ocean (Stockdale et al. 1998), or some combination of the above (Barnston et al. 2003). These two-tiered approaches allow the ocean to influence the atmospheric variability, but do not permit the atmosphere to feedback to the ocean (Goddard et al. 2001). The most complex method of seasonal climate forecasting is to model the dynamics of the atmosphere and the ocean together. The standard approach is to “couple” separate models for the atmosphere and ocean that are run synchronously and interactively (Stockdale et al. 1998; Palmer et al. 2004; Graham et al. 2005; Guérémy et al. 2005; Saha et al. 2006). Such “fully-coupled” models generate forecasts of the atmosphere and of the ocean iteratively, and so sometimes are referred to as “one-tiered” forecasting systems. These fully coupled systems are considered the state-of-the-art in seasonal climate forecasting since they represent the most comprehensive attempt to model all the components of the climate system thought to be relevant for understanding atmospheric variability at seasonal timescales. However, because of their computational advantages and, in some cases, higher levels of forecast skill, two-tiered systems and the simpler empirical models are still widely used in operational forecasting.
2.4 2.4.1
Uncertainty in Seasonal Climate Forecasting Estimating the Uncertainty in the Forecast
Whether it is a statistical or a dynamical approach that is used to forecast the climate, the weather variability over the season constitutes an inherent source of uncertainty, and it is important that the seasonal forecast provides some form of indication of the degree of uncertainty in the prediction. It is because of the large uncertainty that seasonal forecasts are expressed probabilistically rather than as precise estimates of seasonal total rainfall, for example, or of averaged temperatures. The probabilistic forecasts are commonly presented in the form of a set of probabilities that the seasonal total rainfall, for example, will be within specific
2 “Flowering Walnuts in the Wood” and Other Bases
23
ranges. The definition of these ranges as the terciles of an historical set of rainfall totals for the same season is widespread, although other ranges that are more meaningful to users of the seasonal forecasts are being used increasingly (e.g., Thomson et al. 2006). Reliable estimation of the uncertainty in the forecast is a non-trivial problem. When the forecasts are made using statistical models, the errors in past predictions using the same models can be measured, and the distribution of the sizes of these errors, given a sufficiently large number of past forecasts, provides an indication of the distribution of expected errors for the current forecast (Mason and Mimmack 2002). Although a similar approach could in theory be used for the dynamical model predictions, in practice not only is the availability of past forecasts almost always highly limited, thus precluding any reasonable estimate of the distribution of forecast errors, but uncertainty in any single prediction from a dynamical model is much greater than that from a statistical model, and these additional sources of uncertainty need to be understood before using dynamical models for making a seasonal forecast. Specifically, there are two sources of uncertainty when using dynamical models: the representation of the physics of the real atmosphere within the model is greatly simplified, and there are aspects of the physics of the climate system that are poorly understood; the prediction by the dynamical model is highly sensitive to the initial weather conditions that are specified within the model. This latter problem is considered first.
2.4.2
Sources of Uncertainty in Seasonal Forecasts
2.4.2.1
Uncertainty Resulting from Weather Variability
Dynamical models work by predicting how the state of the atmosphere will evolve given the laws of physics that govern the evolution of the atmosphere. They therefore require the current state of the atmosphere to be specified. However, even very small errors in the estimate of this initial state of the atmosphere (known as the “initial conditions”) can magnify very quickly such that very different weather conditions can be predicted for only a few days in the future given only a very small change in the initial conditions. It is believed that the real atmosphere behaves in the same way: that a minor disruption to the atmosphere (perhaps by the proverbial flutter of a butterfly’s wings) will magnify over a few days so that very different weather conditions are experienced to what would have occurred had the butterfly stayed still. This sensitivity of the atmosphere (both the real one and the model’s) explains why the weather is so poorly predicted beyond a few days: the unpredictability of the exact state of the atmosphere any time in the future is an inherently insurmountable problem. So why is this “chaotic” nature of the dynamical model’s atmosphere a problem when predicting seasonal climate; presumably, if the dynamical model is a good one, the predicted weather over the target season should be influenced by the model’s sea
24
S.J. Mason
temperatures in a similar way to the real oceans’ influences on the real atmosphere? The problem arises from the fact that since even in the most predictable areas and seasons only about two-thirds of the year-to-year variability in rainfall can be attributed to sea-surface temperature forcing, a single model prediction with a similar degree of forced variability is quite likely to be a poor estimate of how the real atmosphere will evolve. While nothing can be done about the fact that the actual evolution of the observed atmosphere represents only one possible outcome, we can at least get an idea of how characteristic the single model prediction is of other possible predictions given similar sea temperature forcing (in a “two-tiered” system) and/or initial conditions (in a “one-tiered” system). Dynamical models therefore are invariably run a number of times using similar initial conditions, so that a set, or “ensemble” of predictions are generated. The objective of generating an ensemble of predictions is to simulate the distribution of possible outcomes that the real atmosphere can take. Assuming that this distribution is estimated reliably, one indication of the uncertainty in the seasonal forecast could be provided by the degree to which the seasonal statistics of these various predictions agree. If the average of an ensemble of predictions indicates a seasonal rainfall total that is unusually high, then presumably we can be more confident that the actual seasonal total will be high if all the predictions indicate a similarly high total than if the totals are more scattered even though the average is unchanged. In practice, however, it has been difficult to demonstrate that year-to-year changes in the spread of the ensemble of predictions from a dynamical model represent meaningful variability in the actual uncertainty in the predicted outcome. Part of the problem is methodological: powerful procedures for identifying a relationship between the ensemble distribution and the uncertainty in the prediction are lacking. However, this problem does not negate the generation of the ensemble: it has been widely demonstrated that the mean of an ensemble of predictions is a more accurate forecast, on average, than the predictions of any of the individual ensemble members (Kumar and Hoerling 1995; Kumar et al. 2001). This principle of improving the forecast by considering the mean of an ensemble of predictions can be illustrated simply by comparing 30 years of model forecasts3 of January–March precipitation for Kalbarri,4 Australia (27°42′43″ S, 114°09′54″ E, 6.0 m altitude). The correlations between the simulated and the observed rainfall are shown in Fig. 2.5 with the forecasts calculated from the mean of an increasing
3 The forecasts are from the ECHAM4.5 atmospheric general circulation model, which was developed by the Max Planck Institute (Roeckner et al. 1996). The model was run in a two-tiered mode, with sea-surface temperatures predicted by persisting observed anomalies for November. 4 General circulation models produce forecasts that represent average conditions over relatively large grids (in this case about 2.8° × 2.8°, or about 60,000 km2) rather than for specific locations. Because of model inaccuracies, resulting in part from the coarseness of the spatial resolution of the model, geographical features in the model can be displaced. Normally a “downscaling” procedure would be applied to try and convert the model output to a prediction for a specific location, but for this example the simulations for Kalbarri are taken simply as the model value for the grid containing 27°42′43″ S, 114°09′54″ E.
2 “Flowering Walnuts in the Wood” and Other Bases
25
Fig. 2.5 Correlations between observed and ensemble-mean simulations of January–March 1971–2000 precipitation for Kalbarri, Australia, for varying ensemble sizes
number of ensemble members. The improvement in the correlation is quite marked, but the correlation stabilizes given an ensemble size of about seven or eight, which is typical for many areas in the tropics (Kumar et al. 2001).
2.4.2.2
Uncertainty Resulting from Imperfect Prediction Models
Moving on to the other source of uncertainty in dynamical model predictions, just as errors in the forecast can arise from imperfect knowledge of the current state of the weather (the “initial conditions”), so also uncertainties arise from the simplification of the real climate system by the models. This problem is addressed in a similar way to the initial condition problem: generate a large number of predictions by adjusting the physics of the model and/or by using a set of completely different models, and construct a forecast by considering this set of predictions. A number of recent studies have demonstrated that an ensemble of predictions from different models affords a better forecast than a similarly sized ensemble of predictions from any single model (Hagedorn et al. 2005; Doblas-Reyes et al. 2005), provided that there are no major differences in the quality of the individual models. These “multimodel” predictions improve upon the single model because of the improved indication of uncertainty in the forecast resulting from imperfections in the model physics. Because of the differences in the physics of two models (or because of the perturbation of the physics in a single model) the predictions from a set of initial
26
S.J. Mason
conditions diverge just as if the initial conditions themselves had been perturbed, and so a distribution of predicted outcomes is generated in the same way as for an ensemble from a single model. Multi-model ensembles are currently considered the preferred approach to operational seasonal climate forecasting. The idea permits the combination of predictions from dynamical and statistical models, although in practice the dynamical and statistical approaches remain largely distinct. Some consideration has been given to optimal ways of combining the predictions from different models by assigning models differing weights depending upon their relative historical performances (e.g., Rajagopalan et al. 2002; Robertson et al. 2004), but since it is difficult to demonstrate robust differences in model performance, equal weighting of models is a hard standard upon which to improve (Mason 2008).
2.5
Summary
Seasonal forecasting is possible because the atmosphere is forced and constrained in some parts of the globe and at certain times of the year by conditions in the oceans, especially in the tropical Pacific. However, because of the vagaries of the atmosphere it is not possible to predict what the weather is going to be like beyond a few days. Because the seasonal forecast is itself sensitive to the details of the weather forecast, a large number of predictions are typically made so that the uncertainties arising from the inherent errors in the weather forecasts are represented. These uncertainties are a result both of the imperfect nature of the models themselves and of imperfect knowledge of the current state of the weather. Given the impossibility of deciding which of the ensemble of predictions is the most likely, probabilistic forecasts are issued instead, and aim to indicate the range of possible climate conditions that may occur. Acknowledgements This paper was funded in part by a grant/cooperative agreement from the National Oceanic and Atmospheric Administration (NOAA), contract number NA07GP0213 with the Trustees of Columbia University. The views expressed herein are those of the authors and do not necessarily reflect the views of NOAA or any of its sub-agencies. The rainfall data for Kalbarri were supplied by A. Watkins, Bureau of Meteorology, and the ECHAM4.5 simulations were generated by D. DeWitt, X. Gong, S. Li, and L. Goddard, International Research Institute for Climate and Society, with computer resources provided in part by the NCAR CSL. Comments on earlier drafts by A. G. Barnston, L. Goddard, M. S. J. Harrison, and M. C. Thomson are gratefully acknowledged.
References Allan, R. J., J. A. Lindesay, and D. E. Parker (1996). El Niño – Southern Oscillation and Climatic Variability. CSIRO Publishing, Collingwood, 405 pp.
2 “Flowering Walnuts in the Wood” and Other Bases
27
Barnston, A. G., H. M. van den Dool, S. E. Zebiak, T. P. Barnett, M. Ji, D. R. Rodenhuis, M. A. Cane, A. Leetmaa, N. E. Graham, C. R. Ropelewski, V. E. Kousky, E. A. O’Lenic, and R. E. Livezey (1994). Long-lead seasonal forecasts – where do we stand? Bulletin of the American Meteorological Society 75: 2097–2114. Barnston, A. G., S. J. Mason, L. Goddard, D. G. DeWitt, and S. E. Zebiak (2003). Multi-model ensembling in seasonal climate forecasting at IRI. Bulletin of the American Meteorological Society 84: 1783–1796. Barnston, A. G., A. Kumar, L. Goddard, and M. P. Hoerling (2005). Improving seasonal prediction practices through attribution of climate variability. Bulletin of the American Meteorological Society 86: 59–72. Barsugli, J., J. S. Whitaker, A. F. Loughe, P. D. Sardeshmukh, and Z. Toth (1999). The effect of the 1997/98 El Niño on individual large-scale weather events. Bulletin of the American Meteorological Society 80: 1399–1411. Bengtsson, L., U. Schlese, E. Roeckner, M. Latif, T. P. Barnett, and N. E. Graham (1993). A twotiered approach to long-range climate forecasting. Science 261: 1026–1209. Bjerknes, J. (1966). A possible response of the atmospheric Hadley circulation to equatorial anomalies of ocean temperature. Tellus 18: 820–829. Bjerknes, J. (1969). Atmospheric teleconnections from the equatorial Pacific. Monthly Weather Review 97: 163–172. Bjerknes, J. (1972). Large-scale atmospheric response to the 1964–65 Pacific equatorial warming. Journal of Physical Oceanography 2: 212–217. Doblas-Reyes, F. J., R. Hagedorn, and T. N. Palmer (2005). The rationale behind the success of multi-model ensembles in seasonal forecasting – II. Calibration and combination. Tellus 57A: 234–252. Glantz, M. H., R. W. Katz, and N. Nicholls (1991). Teleconnections Linking Worldwide Climate Anomalies: Scientific Basis and Societal Impact. Cambridge University Press, Cambridge, 545 pp. Goddard, L., S. J. Mason, S. E. Zebiak, C. F. Ropelewski, R, Basher, and M. A. Cane (2001). Current approaches to seasonal-to-interannual climate predictions. International Journal of Climatology 21: 1111–1152. Graham, R. J., A. D. L. Evans, K. R. Mylne, M. S. J. Harrison, and K. B. Robertson (2000). An assessment of seasonal predictability using atmospheric general circulation models. Quarterly Journal of the Royal Meteorological Society 126: 2211–2240. Graham, R. J., M. Gordon, P. J. McLean, S. Ineson, M. R. Huddleston, M. K. Davey, A. Brookshaw and R. T. H. Barnes (2005). A performance comparison of coupled and uncoupled versions of the Met Office seasonal prediction general circulation model. Tellus 57A: 320–319. Guérémy, J. -F., M. Déqué, A. Braun, and J. -P. Piedelièvre (2005). Actual and potential skill of seasonal predictions using the CNRM contribution to DEMETER: coupled versus uncoupled model. Tellus 57A: 308–319. Hagedorn R., F. J. Doblas-Reyes, and T. N. Palmer (2005). The rationale behind the success of multi-model ensembles in seasonal forecasting – I. Basic concept. Tellus 57A: 219–233. Harrison, M. S. J. (2005). The development of seasonal and inter-annual climate forecasting. Climatic Change 70: 201–220. Inwards, R. (1994). Weather Lore: A Collection of Proverbs, Sayings and Rules Concerning the Weather. Senate, London, 190 pp. Kumar, A. and M. P. Hoerling (1995). Prospects and limitations of seasonal atmospheric GCM predictions. Bulletin of the American Meteorological Society 76: 335–345. Kumar, A., M. P. Hoerling, and A. G. Barnston (2001). Seasonal predictions, probabilistic verifications, and ensemble size. Journal of Climate 14: 1671–1676. Marriott, P. J. (1981). Red Sky at Night Shepherd’s Delight!: Weather Lore of the English Countryside; 1900 Sayings Explained and Tested. Sheba Books, Oxford, 376 pp.
28
S.J. Mason
Mason, S. J. (2008). From dynamical predictions to seasonal forecasts. In: Understanding and Adapting to Climate Variability, A. Trocolli, M. S. J. Harrison, D. L. T. Anderson, and S. J. Mason (eds.). Springer Academic Publishers, Dordrecht, in press. Mason, S. J. and L. Goddard (2001). Probabilistic precipitation anomalies associated with ENSO. Bulletin of the American Meteorological Society 82: 619–638. Mason, S. J. and G. M. Mimmack (2002). Comparison of some statistical methods of probabilistic forecasting of ENSO. Journal of Climate 15: 8–29. Mason, S. J., L. Goddard, N. E. Graham, E. Yulaeva, L. Sun and P. A. Arkin (1999). The IRI seasonal climate prediction system and the 1997/1998 El Niño event. Bulletin of the American Meteorological Society 80: 1853–1873. Namias, J. (1991). Spring and summer 1988 drought over the contiguous United States – causes and prediction. Journal of Climate 4: 54–65. Orlove, B. S., J. C. H. Chiang, and M. A. Cane (2000). Forecasting Andean rainfall and crop yield from the influence of El Niño on Pleiades visibility. Nature 403: 68–71. Orlove, B. S., J. C. H. Chiang, and M. A. Cane (2002). Ethnoclimatology in the Andes. American Scientist 90: 428–435. Palmer, T. N. and D. L. T. Anderson (1994). The prospects for seasonal forecasting – a review paper. Quarterly Journal of the Royal Meteorological Society 120: 755–793. Palmer, T. N., A. Alessandri, U. Anderson, P. Cantelaube, M. Davey, P. Délécluse, M. Déqué, E. Díez, F. J. Doblas-Reyes, H. Feddersen, R. Graham, S. Gualdi, J. -F. Guérémy, R. Hagedorn, M. Hoshen, N. Keenlyside, M. Latif, A. Lazar, E. Maisonnave, V. Marletto, A. P. Morse, B. Orfila, P. Rogel, J. -M. Terres, and M. C. Thomson (2004). Development of a European ensemble system for seasonal to inter-annual prediction (DEMETER). Bulletin of the American Meteorological Society 85: 853–872. Rajagopalan, B., U. Lall, and S. E. Zebiak (2002). Categorical climate forecasts through regularization and optimal combination of multiple GCM ensembles. Monthly Weather Review 130: 1792–1811. Robertson, A. W., U. Lall, S. E. Zebiak, and L. Goddard (2004). Improved combination of multiple atmospheric GCM ensembles for seasonal prediction. Monthly Weather Review 132: 2732–2744. Roeckner, E., K. Arpe, L. Bengtsson, M. Christoph, M. Claussen, L. Dümenil, M. Esch, M. Giorgetta, U. Schlese, and U. Schulzweida (1996). The atmospheric circulation model ECHAM-4: Model description and simulation of present-day climate. MPI-Rep. 218, MPI für Meteorologie, Hamburg, 90 pp. Saha, S., S. Nadiga, C. Thiaw, J. Wang, W. Wang, Q. Zhang, H. M. van den Dool, H. -L. Pan, S. Moorthi, D. Behringer, D. Stokes, M. Peña, S. Lord, G. White, W. Ebisuzaki, P. Peng and P. Xie (2006). The NCEP Climate Forecast System. Journal of Climate 19: 3483–3517. Shukla, J. (1998). Predictability in the midst of chaos: a scientific basis for climate forecasting. Science 282: 728–731. Stockdale, T. N., D. L. T. Anderson, J. O. S. Alves, and M. Balmaseda (1998). Global seasonal rainfall forecasts using a coupled ocean-atmosphere model. Nature 392: 370–373. Thomson, M. C., F. J. Doblas-Reyes, S. J. Mason, R. Hagedorn, S. J. Connor, T. Phindela, A. P. Morse, and T. N. Palmer (2006). Multi-model ensemble seasonal climate forecasts for malaria early warning. Nature 439: 576–579. Uppala, S. M., P. W. Kållberg, A. J. Simmons, U. Andrae, V. da Costa Bechtold, M. Fiorino, J. K. Gibson, J. Haseler, A. Hernandez, G. A. Kelly, X. Li, K. Onogi, S. Saarinen, N. Sokka, R. P. Allan, E. Andersson, K. Arpe, M. A. Balmaseda, A. C. M. Beljaars, L. van de Berg, J. Bidlot, N. Bormann, S. Caires, F. Chevallier, A. Dethof, M. Dragosavac, M. Fisher, M. Fuentes, S. Hagemann, E. Hólm, B. J. Hoskins, L. Isaksen, P. A. E. M. Janssen, R. Jenne, A. P. McNally, J. -F. Mahfouf, J.- J. Morcrette, N. A. Rayner, R. W. Saunders, P. Simon, A. Sterl, K. E. Trenberth, A. Untch, D. Vasiljevic, P. Viterbo, and J. Woollen (2005). The ERA-40 re-analysis. Quarterly Journal of the Royal Meteorological Society 131, 2961–3012, doi:10.1256/qj.04.176.
2 “Flowering Walnuts in the Wood” and Other Bases
29
van den Dool, H. M. (1994). Searching for analogues: how long must one wait? Tellus 46A: 314–324. van den Dool, H. M., J. Huang, and Y. Fan (2003). Performance and analysis of the constructed analogue method applied to US soil moisture over 1981–2001. Journal of Geophysical Research 108: D08617, doi:10.1029/2002JD003114. Walker, N. D. and J. A. Lindesay (1989). Preliminary observations of oceanic influences on the February–March 1988 floods in central South Africa. South African Journal of Science 85: 164–169. Ward, M. N. and C. K. Folland (1991). Prediction of seasonal rainfall in the north Nordeste of Brazil using eigenvectors of sea-surface temperatures. International Journal of Climatology 11: 711–743. Zebiak, S. E. (1999). El Niño and the science of climate prediction. Consequences 5: 3–15.
Chapter 3
Climate and Infectious Diseases Louise Kelly-Hope1 and Madeleine C. Thomson2,*
Abstract Climate plays an important role in the transmission of many infectious diseases; it not only determines spatial and seasonal distributions, but influences inter-annual variability, including epidemics, and long-term trends. This paper collates published scientific literature on climate and 20 infectious diseases that cause considerable morbidity and mortality worldwide. It highlights what has been done to date, identifies gaps and assesses the role of climate information in improving health system performance, especially in developing countries. Parasitic, viral and bacterial diseases are discussed in the light of climate impacts on classified according to geographic distribution, seasonality, interannual variability, or climatic shifts. Study methods range from simple comparisons in seasonality, to detailed risk analyses, predictive models and early warning systems for epidemics. Malaria and dengue were found to be the most researched diseases with respect to climate, followed by meningococcal meningitis, schistosomiasis, rotavirus, and leishmaniasis. Studies on diseases with long development periods tended to focus on spatial patterns for the creation of risk maps while acute diseases focused interannual variability and the creation of climate-driven early warning systems. An emerging area identified in this review is the potential for climate information to improve the quality of intervention impact assessment where diseases are climate sensitive. We note that despite an extensive literature for some diseases very little research has been done in the countries with the highest number of child deaths and under-five mortality rates. This review provides a platform from which to launch future research and policy development in relation to climate-sensitive disease, and suggests that vulnerable countries should be the priority focus of this effort.
1
Liverpool School of Tropical Medicine, Liverpool, L3 5QA UK
2
International Research Institute for Climate and Society, Columbia University, New York, 10964 USA * Corresponding author: Tel: + 1 845 680 4413; fax: + 1 845 680 4864; e-mail: mthomson@iri. columbia.edu
M.C. Thomson et al. (eds.), Seasonal Forecasts, Climatic Change and Human Health. © Springer Science + Business Media B.V. 2008
31
32
L. Kelly-Hope and M.C. Thomson
Keywords Infectious disease, parasite, bacteria, virus, climate, climate variability, climate change, rainfall, temperature, sea surface temperature, NDVI, geographic distribution
3.1
Introduction
The importance of infectious disease as a determinant and outcome of poverty in developing countries is a prominent argument for international and national investment in controlling patterns of infectious disease transmission (WHO 1999, 2001; Black et al. 2003). This argument is reflected in the United Nations Millennium Development Goals (MDGs) (UN 2002; Sachs 2004; Sachs and McArthur 2005). Although the infectivity and transmission processes unique to many diseases are complex, several factors have been identified as direct drivers of disease risk. Among these are climatic factors such as rainfall and temperature. Indeed, climate plays an important role in the transmission of many of the diseases whose control is important to the achievement of the MDGs to the extent that in some ecological and economic settings climate variability may undermine the potential for achieving these goals. Climate not only determines the spatial and seasonal distribution of many infectious diseases (Burke et al. 2001), but is often a key determinant of interannual variability, including epidemics (Kovats et al. 1999, 2003; WHO 2004; Kuhn et al. 2005), and long-term trends (Haines and Patz 2004; Patz et al. 2005). More specifically, climate information can be used to improve our assessment of interventions for climate-sensitive diseases and human health (McMichael et al. 2003; Hansen et al. 2004; IRI 2005; Connor et al. 2006). While many other factors undoubtedly play critical roles in disease propagation (e.g. immune status, socioeconomic status, etc.), this review focuses specifically on how climate is associated with disease incidence in the human population. There is heightened interest in supporting health systems to improve the management of climate-sensitive diseases. A special report to the third IPCC (Intergovernmental Panel on Climate Change (IPCC) 2001) stated that ‘An effective health system can help to address the adverse health impacts of climate change’ and ‘Thus, in terms of technology transfer there is a need to ensure that technologies are available at national and local levels for coping with any changes in the burden of disease that might be associated with climate change’. Despite a rapidly increasing interest in the use of climate data by the public health sector (Kovats et al. 1999, 2003; Burke et al. 2001; IPCC 2001; McMichael et al. 2003; Haines and Patz 2004; Hansen et al. 2004; WHO 2004; Kuhn et al. 2005; Patz et al. 2005; IRI 2005), a considerable effort is still required to develop policy-relevant evidence for decision-makers involved in controlling climate-sensitive diseases. In order to achieve the MDGs it is crucial that appropriate polices are developed and implemented to improve health system performance (Anon 2004; Travis et al. 2004; Wyss 2004). Climate information services may play a role in this if appropriate tools and analysis can be used effectively to improve the ability of
3 Climate and Infectious Diseases
33
those engaged in promoting, preventing or improving the health of populations to (a) detect and treat diseases, (b) monitor and predict epidemics, (c) implement intervention and control strategies, and (d) monitor the impact of interventions (Connor et al. 2006). This review collates published scientific literature on climate and infectious diseases. It focuses on 20 major diseases, which are influenced by climatic factors and cause considerable morbidity and mortality worldwide. Studies included herein are restricted to those that have quantified the relationship between climate and human infectious disease through statistical association. These range from simple comparisons in seasonality, to detailed risk analyses, predictive models and early warning systems for epidemics. Additionally, these studies are classified according to climates impact on their spatial distribution, seasonality, interannual variability, and trends. This extensive compilation of historical and contemporary literature will highlight what has been done to date, identify gaps and assess the role of climate in improving health system performance, especially in developing countries. It will provide a platform from which to launch future research and policy development in relation to climate-sensitive disease. The diseases presented in the Table 3.1 are grouped as parasitic, viral and bacterial and include those with both short and long development periods. Their sensitivities to climate differ; those with a short development period tend to be highly seasonal or epidemic in nature, with clinical manifestations readily identified (often severe), and usually the basis of epidemiological research. Recently, several diseases were identified as candidates for climate-based early warning systems as a means of improving preparedness for and in response to epidemics (Kuhn et al. 2005) (Table 3.2). In contrast, chronic diseases with long development periods, in which the pathogen may survive for many years in the human host (e.g. lymphatic filariasis), may exhibit little or no seasonal or interannual variability, even though transmission may be driven by climatic factors. In this case, subclinical infections or preliminary disease are detected by other means, such as skin biopsies/snips (e.g. onchocerciasis; Botto et al. 2005), thick blood films (e.g. loa loa; Wanji et al. 2005) and urine or stool samples (e.g. schistosomiasis; Brooker et al. 2001; Kabatereine et al. 2004).
3.2
Geographical/Spatial Distribution
Defining the geographical distribution of a disease within a country or region is a fundamental step to understanding its epidemiology, as it allows health systems to identify epidemic/endemic zones and vulnerable groups at risk. It also allows comparisons among diseases, analysis of temporal trends and identification of climatic and other factors that may influence the spatial heterogeneity of disease. No disease is uniformly distributed, even though there appear to be broad influential spatio-climatic parameters, patterns and trends. For instance, mosquito-borne parasitic and arboviral diseases are commonly found in hot, humid regions of the world, while
34
L. Kelly-Hope and M.C. Thomson
Table 3.1 Disease, transmission mechanism, climate and environmental drivers, country of study and references Disease/Transmission Characteristics Parasitic 1. Malaria Plasmodium sp.
Country
References
Space Africa
Brazil China Rainfall, humidity, temperature, China surface water puddles, river margins, irrigation, altitude, NDVI East Africa Mosquitoes Anopheles sp.
Ecuador Ethiopía India/Pakistan
IndoChina Kenya Peru Peru Philippines South Africa Sri Lanka Tanzania Tanzania Thailand West Africa Zimbabawe Seasonal Cameroon Ethiopia Ghana Kenya Mali Rwanda Sudan
Le Sueur et al. 1997; Craig et al. 1999; Snow et al. 1999; Small et al. 2003 Camargo et al. 1996 Yang et al. 2002 Bi et al. 2003b
Hay et al. 2002b; Omumbo et al. 2005a; Omumbo et al. 2005b Cedeno 1986; Moreira 1986 Belize Hakre et al. 2004 Abeku et al. 2003; Teklehaimanot et al. 2004a, b Christophers 1911; Gill 1921, 1923; Yacob and Swaroop 1945, 1946; Mathur et al. 1992; Bouma and van der Kaay 1994; Akhtar and McMichael 1996; Gupta 1996; Singh and Sharma 2002; Bouma et al. 1996 Nihei et al. 2002 Omumbo et al. 2004 Guthmann et al. 2002 Guthmann et al. 2002 Leonardo et al. 2005 Craig et al. 2004 Gill 1936; Ramasamy et al. 1992; Van Der Hoek et al. 2003 Bodker et al. 2003 Bodker et al. 2003 Rosenberg et al. 1990; Nacher et al. 2004a, b; Kleinschmidt et al. 2001 Siziya et al. 1997; Mabaso et al. 2005 van der Kolk et al. 2003; Akenji et al. 2005 Abeku et al. 2002, 2003; Teklehaimanot et al. 2004a, b Afari et al. 1993; Baird et al. 2002; Koram et al. 2003 Hay et al. 2001; Shanks et al. 2002; Munyekenye et al. 2005 Bouvier et al. 1997; Dicko et al. 2005 Loevinsohn 1994 Giha et al. 2000 (continued)
3 Climate and Infectious Diseases
35
Table 3.1 (continued) Disease/Transmission Characteristics
Country
References
Malaria cont.
The Gambia
Greenwood et al. 1987; Brewster and Greenwood 1993 Kilian et al. 1999; Lindblade et al. 1999; Odongo-Aginya et al. 2005
Uganda Interannual Botswana China Colombia East Africa Ethiopia
Thomson et al. 2005, 2006a Bi et al. 2005 Bouma et al. 1997; Poveda et al. 2001 Zhou et al. 2004, 2005 Abeku et al. 2002; Teklehaimanot et al. 2004a, b India/Pakistan Gill 1923; Swaroop 1946 India/Sri Lanka Bouma and van der Kaay 1996 Indonesia Anon 1999 Kenya Hay et al. 2001 Kenya Hay et al. 2001 Madagascar Bouma 2003 Niger Julvez et al. 1997 Peru Valencia Tellería 1986 Senegal Ndiaye et al. 2001 South America Organization 1998; Gagnon et al. 2002 Southern Africa Anon 2002; DaSilva et al. 2004; GroverKopec et al. 2005; Connor et al. 2007 Tanzania Lindsay et al. 2000 Thailand Hay et al. 2000 Uganda Kilian et al. 1999; Lindblade et al. 1999 Venezuela Bouma and Dye 1997 Trend Africa East Africa Pakistan Rwanda West Africa 2. African Trypanosomiasis / Sleeping Sickness, Ngana e.g. Trypanosoma brucei gambiensis
Small et al. 2003 Hay et al. 2002b Bouma et al. 1996 Loevinsohn 1994 Brewster and Greenwood 1993; Trape 1999
Space Africa
Rogers 1991; Rogers and Williams 1993
Togo Uganda
Hendrickx et al. 1999, 2000 Rogers 2000; Odiit et al. 2005
Tsetse Glossina sp. Seasonal Gallery forests, savannah wood- Africa land, temperature, NDVI Kenya Uganda
Rogers and Williams 1993 Wellde et al. 1989 Rogers 2000 (continued)
36
L. Kelly-Hope and M.C. Thomson
Table 3.1 (continued) Disease/Transmission Characteristics
Country
3. Schistosomiasis / Bilharzias Space Brazil Schistosoma sp. China Snails e.g. Bulinus Africanus Surface water, NDVI, temperature, rainfall, elevation
Raso et al. 2005 Malone et al. 1994
Ethiopia Philippines Tanzania Uganda
Kristensen et al. 2001 Cross et al. 1984; Leonardo et al. 2005 Brooker et al. 2001 Kabatereine et al. 2004; Stensgaard et al. 2005
4. Leishmaniasis genus Leishmania
Space Brazil
e.g. Phlebotomine Sandflies
Colombia Sudan Tunisia Seasonal Brazil French Guiana Turkey Turkmenistan Interannual Bolivia Brazil Costa Rica
5. Lymphatic filariasis e.g. Wuchereria bancrofti in Africa Mosquitoes: Anopheles, Aedes and Culex sp. Rainfall, humidity, temperature, surface water, NDVI 6. Onchocerciasis / River Blindness Onchocerca volvulus Blackflies: Simulium sp.
Bavia et al. 1999, 2001a; 2005c Zhou et al. 2001; Yang et al. 2005a; 2005c
Cote d’Ivoire Egypt
Seasonal Brazil
Rainfall, temperature, NDVI, land cover, elevation
References
Bavia et al. 1999, 2001 Thompson et al. 2002; Werneck and Maguire 2002; Bavia et al. 2005 King et al. 2004 Elnaiem et al. 2003 Thomson et al. 1999 Ben Salah et al. 2000
Thompson et al. 2002; Martins et al. 2004 Nacher et al. 2001, 2002 Uzun et al. 1999 Neronov and Malkhazova 1999 Gomez et al. 2006 Franke et al. 2002a, b Chaves and Pascual 2007
Space Africa
Lindsay and Thomas 2000
Egypt West Africa
Thompson et al. 1996; Hassan et al. 1998a, b Kelly-Hope et al. 2006
Seasonal Ghana
Gyapong et al. 1996
Space Ethiopia Venezuela
Gebre-Michael et al. 2005 Botto et al. 2005
Seasonal Ethiopia
Gebre-Michael et al. 2005 (continued)
3 Climate and Infectious Diseases
37
Table 3.1 (continued) Disease/Transmission Characteristics Rainfall, temperature, NDVI, wind, river discharge
7. African Eye Worm Loa loa
Country
References
Nigeria
Nwoke et al. 1992
Interannual Sierra Leone Space Cameroon
Thomson et al. 1996 Wanji et al. 2003; Thomson et al. 2004a
Chrysops sp. Forest canopy, forest soils, NDVI 8. Guinea worm Dracunculus medinensis
Space Ghana
Belcher et al. 1975; Hunter 1997
Cyclops sp.
Seasonal Burkina Faso Ghana
Steib and Mayer 1988 Belcher et al. 1975
Interannual Ghana
Hunter 1997
Trend Ghana
Hunter 1997
Seasonal Brazil West Africa
Vasconcelos et al. 1997, 2001 Traore-Lamizana et al. 1996
Trend Brazil
Vasconcelos et al. 1997, 2001
Space Africa Kenya Saudi Arabia
Anyamba et al. 2002 Linthicum et al. 1987 Elfadil et al. 2006
Surface water, high and low rainfall
Viral 9. Yellow Fever Flavivirus Mosquitoes Aedes, Haemagogus and Sabethes sp.)
Rainfall, Temperature 10. Rift Valley Fever Phlebovirus Mosquitoes Aedes and Culex sp.
Seasonal Rainfall, humidity, surface water, Kenya temperature, NDVI Nigeria Interannual Kenya Senegal 11. St Louis encephalitis Mosquitoes Culex sp. Low rainfall, drought high temperature, land surface wetness
Davies et al. 1985; Linthicum et al. 1987, 1999 Olaleye et al. 1996 Linthicum et al. 1999; Anyamba et al. 2001 Thonnon et al. 1999’ Thiongane and Martin 2003
Seasonal United States
Anon 1994; Barker et al. 2003
Interannual United States
Shaman et al. 2004
(continued)
38
L. Kelly-Hope and M.C. Thomson
Table 3.1 (continued) Disease/Transmission Characteristics 12. Japanese encephalitis
Country
References
Space China
Okuno et al. 1975
Mosquitoes Culex sp. Seasonal Monsoon, rainfall, temperature, China India Interannual Thailand
Okuno et al. 1975; Bi et al. 2003a Kanojia et al. 2003; Phukan et al. 2004 Suwannee et al. 1997
13. Murray Valley encephalitis/ Space Australian encephalitis Australia Mosquitoes Culex, Anopheles and Mansonia sp. Wet season, rainfall, flooding
Seasonal Australia
14. Ross River virus/Epidemic Space polyarthritis Australia Alphavirus
Mosquitoes mainly Culex, Ochlerotatus, Aedes, Man sonia and Anopheles sp.
15. Hemorrhagic fever with renal syndrome HFRS group
Rodents mainly Rattus, Apodemus and Clethrionomys sp.
Done et al. 2002; Tong et al. 2002; Woodruff et al. 2002; Kelly-Hope et al. 2004a, 2004b, 2004c, Gatton et al. 2005
Seasonal
Australia
Rainfall, flooding temperature, humidity, high tides
Broom et al. 2002, 2003; Whelan et al. 2003; Cordov et al. 2000
Interannual Australia
Tong et al. 1998; Tong and Hu 2001; Done et al. 2002; Kelly-Hope et al. 2002; Tong et al. 2002; Whelan et al. 2003; Hu et al. 2004; Kelly-Hope et al. 2004a, 2004c; Tong et al. 2004, 2005; Gatton et al. 2005 Harley and Weinstein 1996; Maelzer et al. 1999; Done et al. 2002; Woodruff et al. 2002; Kelly-Hope et al. 2004b
Seasonal China Croatia Russia
Chen and Qiu 1993, 1994; Bi et al. 1998, 2002 Mulic and Ropac 2002; Mulic et al. 2003 Nurgaleeva et al. 1988
Interannual China
Bi et al. 2002, 2005; Bi and Parton 2003
Rainfall, temperature, humidity, flooding (continued)
3 Climate and Infectious Diseases
39
Table 3.1 (continued) Disease/Transmission Characteristics 16. Dengue and Dengue Hemorrhagic Fever Flavivirus Mosquitoes Aedes sp. Temperature, rainfall, humidity
Country
References
Space Global Mexico Taiwan Seasonal Bangladesh Barbados India Indonesia Malaysia Mexico Thailand Venezuela
Hales et al. 2002 Peterson et al. 2005 Wu et al. 2006 Amin et al. 1999 Depradine and Lovell 2004 Chakravarti and Kumaria 2005 Corwin et al. 2001 Li et al. 1985 Koopman et al. 1991; Peterson et al. 2005 Nakhapakorn and Tripathi 2005; Thammapalo et al. 2005 Barrera et al. 2002
Interannual Colombia
Gagnon et al. 2001 Surinam Gagnon et al. 2001 French Guiana Gagnon et al. 2001 Indonesia Depradine and Lovell 2004, Kovats 2000; Corwin et al. 2001; Gagnon et al. 2001 Mexico Hurtado-Diaz et al. 2006 Puerto Rico Schreiber 2001 South Pacific Hales et al. 1996, 1999 Thailand Hay et al. 2000; Cazelles et al. 2005
17. Rotavirus Filth flies e.g. Musca sp. via mechanical transmission Humidity, cool/winter, dry months, low rainfall, water shortages, flood
Trend Global
Hales et al. 2002
Seasonal Africa
Cunliffe et al. 1998
Bangladesh Brazil
Ahmed et al. 1991; Fun et al. 1991 Coiro et al. 1983; Bittencourt et al. 2000; da Rosa e Silva et al. 2001
Ghana Global India Indonesia Japan
Armah et al. 1994 Cook et al. 1990 Ram et al. 1990; Phukan et al. 2003 Corwin et al. 2005 Konno et al. 1983 Kuwait Al-Nakib et al. 1980 Kenya Mutanda et al. 1984 Zambia Mpabalwani et al. 1995 Nigeria Gomwalk et al. 1990, 1993 South Africa Steele et al. 1986; Haffejee and Moosa 1990 South America Kane et al. 2004 (continued)
40
L. Kelly-Hope and M.C. Thomson
Table 3.1 (continued) Disease/Transmission Characteristics
Country
References
The Gambia
Hanlon et al. 1987; Brewster and Greenwood 1993 Ijaz et al. 1994
United Arab Emirates Bacterial 18. Meningococcal meningitis Neisseria meningitides
Space Africa
Airborne aerosol
Benin
Absolute humidity, dry, dusty, wind, temperatures
Seasonal Cameroon
West Africa
Cunin et al. 2003 Democratic Rep. of Congo/Zaire Cheesbrough et al. 1995 Girgis et al. 1993 Djibouti Haberberger et al. 1990 Sultan et al. 2005 Skalova 1984 Campagne et al. 1999; Molesworth et al. 2001 Benin Besancenot et al. 1997 Greenwood et al. 1979, 1984 Greenwood et al. 1985; Brewster and Greenwood 1993 Cvjetanovic et al. 1978; Skalova 1984
Interannual Mali West Africa
Sultan et al. 2005 Thomson et al. 2006b
Space Australia India
Tedesco 1980 Gupta and Preobragenski 1964
Kenya Mali
Schwab et al. 1995 Schemann et al. 2002
Sudan
Salim and Sheikh 1975
Egypt Mali Mongolia Niger Nigeria The Gambia
19. Trachoma Chlamydia trachomatis Flies e.g. Musca sorbens via mechanical transmission Aridity, dust, environmental/ dryness, relative humidity
Lapeyssonnie 1963; Cheesbrough et al. 1995; Molesworth et al. 2003 Besancenot et al. 1997
Seasonal India Sudan
Cooper 1964; Gupta and Preobragenski 1964 Salim and Sheikh 1975
20. Cholera Vibrio cholerae
Space Bangladesh
Huq et al. 2005
Fecal/oral route and filth flies e.g. Musca sp. via mechanical transmission
Seasonal
Bangladesh Peru
Huq et al. 2005; Koelle et al. 2005a Franco et al. 1997; Lama et al. 2004 (continued)
3 Climate and Infectious Diseases
41
Table 3.1 (continued) Disease/Transmission Characteristics Water and air temperature, water depth, rainfall and conductivity, algal blooms, flooding, sunlight, SST
Country
References
Mexico
Chavez et al. 2005
Interannual Bangladesh Peru West Africa Ghana
Lobitz et al. 2000; Pascual et al. 2000; Rodo et al. 2002; Koelle et al. 2005b Speelmon et al. 2000; Lama et al. 2004 Constantin et al. 2006 de Magny et al. 2006
Table 3.2 Climate-sensitive diseases, include those with EWS potential (Kuhn et al. 2005) Early warning system potential Malaria African trypanosomiasis Leishmaniasis Yellow fever Rift Valley fever Dengue and dengue hemorrhagic fever St. Louis encephalitis Japanese encephalitis Murray Valley encephalitis Ross River virus Meningococcal meningitis
bacterial infections such as trachoma and epidemic meningococcal meningitis prevail in countries with a prolonged dry season. (Cooper 1964; Sarkies 1967; Schwab et al. 1995). Specifically, high rates of trachoma in northwest India have been associated with low humidity, winds (Arabian) and dust storms (Cooper 1964), and in Kenya and Sudan with climatic aridity (Salim and Sheikh 1975; Schwab et al. 1995). Winds (Harmattan) and dusty conditions have also been linked to meningococcal epidemics the Sahel region of West Africa (Greenwood et al. 1984; Besancenot et al. 1997; Sultan et al. 2005), an area known as the ‘Meningitis Belt’ which coincides with 300–1,100 mm annual rainfall (Lapeyssonnie 1963; Molesworth et al. 2003). Notably, meningococcal outbreaks tend not to occur in humid, forested or coastal region areas, as high continuous humidity appears to inhibit disease (Haberberger et al. 1990; Cheesbrough et al. 1995; Molesworth et al. 2003). Changes to the micro-climate as a result of landuse/cover change is an important driver of changes in transmission in some areas. For example, deforestation in the Amazon has increased the abundance of anopheline vectors which thrive better in sunlight pools rather than jungle breeding sites, thereby increasing the risk of disease (Vittor et al. 2006).
42
L. Kelly-Hope and M.C. Thomson
Defining exclusion zones based on climate is also useful. It provides further insight into a disease’s ecology, and helps allocate human and financial resources to high-risk areas. This is exemplified in Uganda with schistosomiasis, in which a large parasitological survey found no transmission at altitudes >1,400 m, or where total annual rainfall was < 900 mm. Subsequently, this information helped the design and implementation of the national control program currently underway (Kabatereine et al. 2004). The filarial worm Loa loa has recently emerged as a parasite of significant public health importance as a consequence of its impact on the African Programme for Onchocerciasis Control. Severe, sometimes fatal, encephalopathic reactions to ivermectin (the drug of choice for onchocerciasis control) have occurred in some individuals with high Loa loa microfilarial counts in West and Central Africa including Cameroon. A modeled distribution map of the Loa loa prevalence in Cameroon has been created based on epidemiological data, altitude (as a proxy for temperature), satellite derived vegetation indices and knowledge of transmission dynamics. The map indicates areas where the risk of Loa loa is exceeding low and ivermectin distribution may be undertaken safely (Thomson et al. 2004a). Similarly, as part of the global elimination program for lymphatic filariasis, prevalence surveys have been conducted to identify high/low risk locations prior to mass drug administration (Gyapong et al. 2002). In West Africa, filariasis prevalence was found to be high in the Sahel region, and positively correlated with low rainfall and low vegetation greenness (Kelly-Hope et al. 2006). Interestingly, a negative spatial association was found with malaria, which was more prevalent in the humid savanna zone of this region. This suggests that within defined regions, different climate and ecological factors may drive different disease distributions. This is also evident in the Philippines, where the magnitude and distribution of malaria and schistosomiasis differ in two distinct regions (Leonardo et al. 2005). In addition to spatial differences, identifying the locations where diseases overlap is valuable. This can help to identify common risk factors, ecological niches and may facilitate coordinated control and intervention strategies. This is becoming increasingly relevant with the number of disease elimination and control programs currently underway globally (Molyneux 2004; Molyneux et al. 2005). Unfortunately, data on different diseases are not usually collected simultaneously, and are rarely examined in relation to climate variables. As shown in Table 3.1, our comparisons were restricted by the spatial distributions of studies of malaria, Rift Valley fever and trachoma in Kenya, trypanosomiasis and schistosomiasis in Uganda, and leishmaniasis and schistosomiasis in Brazil, which have all been examined using different climate parameters, on different scales, at different times. This disparate data collection process prevents any meaningful comparisons. We posit that the formation of a disease atlas comprising high resolution, subnational data may be useful to national health systems, for policy discussions, guiding interventions and costeffective monitoring at a range of spatial scales. However, standard ways in which to collect and analyze field data are first needed to optimize such an endeavor. Maps are useful visual tools, and have long been used for displaying prevalence and infection intensity data. Mapping the spatial relationship between climate and disease has been described during the early 1900s in India (Punjab) (Christophers 1911; Gill 1921, 1923; Yacob and Swaroop 1945, 1946) and Sri Lanka (Gill 1936),
3 Climate and Infectious Diseases
43
where factors defining malaria outbreaks were compared. Throughout the 1940– 1970s, a series of simple maps, graphs and methodologies were used to examine the climatic impact on diseases such as meningitis (Lapeyssonnie 1963), Japanese encephalitis (Okuno et al. 1975), lymphatic filariasis (Brengues 1975) and trachoma (Cooper 1964; Salim and Sheikh 1975). However, it was not until the introduction and widespread availability of environmental satellite images and computer-based geographic methods in the early 1980s that more sophisticated spatial and statistical methods have been developed and utilized. Geographic information systems (GIS) and remote sensing (RS) were used initially to define the ecological parameters of schistosomiasis in the Philippines (Cross et al. 1984) and Rift Valley fever in Kenya (Linthicum et al. 1987). Since then, scientists have studied many more diseases in different geographical regions, largely due to the increased affordability and accessibility of computers, specialized software and geo-referenced spatial data. Recently, a number of malaria risk maps, based on climatic parameters, have been modeled for Africa (Le Sueur et al. 1997; Craig et al. 1999; Snow et al. 1999; Kleinschmidt et al. 2001; Omumbo et al. 2005b), and an international collaboration of scientists and institutions have developed an Atlas of Malaria for Africa available via the internet(MARA/ ARMA 1998). Researchers have further linked climate with many other infectious diseases, including trypanosomiasis in Uganda (Odiit et al. 2005), schistosomiasis in Philippines (Leonardo et al. 2005), China (Yang et al. 2005a), Cote d’Ivoire (Raso et al. 2005) and Uganda (Stensgaard et al. 2005), dengue and dengue hemorrhagic fever in Thailand (Nakhapakorn and Tripathi 2005), leishmaniasis in Brazil (Bavia et al. 2005), onchocerciasis in Ethiopia (Gebre-Michael et al. 2005), filariasis in West Africa (Kelly-Hope et al. 2006) and Ross River virus disease in Australia (Gatton et al. 2005), using these advances in GIS and RS technologies.
3.3
Seasonality
Climate-sensitive diseases usually have distinct seasonal patterns, especially those with short development periods. Knowing a specific disease’s incubation period and associated climatic influences will help health workers determine when to expect high incidences or outbreaks of the disease, aid diagnosis and direct timely interventions. Clearly, broad temporal climatic patterns exist, and it is useful to recognize that the prevalence of diseases such as malaria, dengue and Rift Valley fever increase during and immediately following the warm, rainy season, while that of meningococcal meningitis and trachoma peak in the dry pre-monsoonal period, and rotavirus prevails during the dry and/or cold months of the year. However, it is also important to be aware that one disease may exhibit different seasonal patterns in different ecological zones. In West Africa, for example, guinea worm transmission occurs during the dry season in the humid savanna zone (southern) (Belcher et al. 1975), but during the wet season in the dry savanna zone (northern) (Steib and Mayer 1988). Similar differences have been noted for malaria in Asia, where an outbreak in the dry region of India (Punjab) occurred with the onset of the monsoonal rainfall (Christophers 1911; Gill 1923; Yacob and Swaroop 1945,
44
L. Kelly-Hope and M.C. Thomson
1946). Conversely, drought and a rise in humidity seemed to be important malarial onset factors in the wet zone of Sri Lanka (Gill 1936). This latter pattern is supported by local knowledge noted by Gill (1936), ‘it has long been known to medical men and laymen alike that a wet year in the Wet Zone is a healthy year, and that in this zone a failure of the monsoon is almost invariably followed by an unusual prevalence of malaria’, and highlights the importance of engaging local people and their knowledge in directing empirical research to quantify climate and disease patterns. Clearly, identifying the specific climatic factors driving high risk disease outbreak and transmission periods is important, as they can vary greatly among diseases and regions. Many studies have compared morbidity and mortality variables among the wet, dry, hot and/or cold seasons, but few have conducted any time series analyses on more than one disease. Developing a calendar, namely a disease calendar, depicting the high-risk months of a few key diseases may useful for national health systems. It may help to determine when and where human and financial resources should be allocated, and help regional health workers detect and treat diseases appropriately (Mabaso et al. 2005, 2007a). For example, in India high rates of Japanese encephalitis (June– October) (Kanojia et al. 2003; Phukan et al. 2004) occur during the rainy season, dengue is prevalent in the post-monsoonal period (October–December) (Chakravarti and Kumaria 2005) and rotavirus peaks during the winter months (November–February) (Ram et al. 1990; Phukan et al. 2003). Furthermore, retrospective analyses of historical disease data may elucidate climate connections, which are now more readily achievable with the recent advances in GIS and RS technology and access to climate datasets. In addition to seasonal differences, a number of studies have quantified associations among specific climatic variables and the time to disease onset or peak period. For example, scientists identified correlations between rainfall, temperature and cholera with a 4–8-week lag time in Bangladesh (Huq et al. 2005), between heavy rainfall and dengue with a 2–3-month lag in Malaysia (Li et al. 1985), and between high rainfall and guinea worm with a year lag period in West Africa (Steib and Mayer 1988). These studies provide some insight into climatic factors that are potentially important for the disease-causing organism’s incubation, development and transmission periods, and may help to develop early warning protocols. Defining low risk periods of disease outbreak and transmission identifies specific factors that may inhibit or halt disease transmission. Though anecdotal, reports have indicated a lack of malaria during drought years in Sudan (Theander 1998; Giha et al. 2000), the cessation of meningococcal disease epidemics with the onset of rains in West Africa (Lapeyssonnie 1963), and a reduced threat of meningococcal meningitis during the hot summer months in Djibouti (Haberberger et al. 1990). Similarly, an area of Honduras was considered too hot for anopheline mosquitoes to survive, which accounted for a drop in malaria cases (Almendares et al. 1993), and in Thailand the transmission of dengue ceased with the onset of the cold dry season (Barbazan et al. 2002).
3 Climate and Infectious Diseases
3.4 3.4.1
45
Inter-Annual Variability Extreme Events
Extreme or adverse weather conditions have the potential to either promote or inhibit disease transmission, and are often linked to long-term, large-scale or cyclic climate phenomena. Furthermore, it has been suggested that climate changes could also affect large-scale weather patterns by increasing the frequency and intensity of extreme events, such as prolonged droughts and heavy rainfall, including severe tropical storms and floods (McMichael et al. 2003). For example, the incidence of hemorrhagic fever with renal syndrome (HFRS; e.g. Korean hemorrhagic fever, epidemic hemorrhagic fever, and nephropathis epidemica) carried and transmitted by rodents, may be negatively associated with heavy rainfall in China which destroys rodent habitats in some instances (Bi et al. 1998). Likewise dengue and/or dengue hemorrhagic fever is negatively associated with anomalously high rainfall in Malaysia (Li et al. 1985), Thailand (Thammapalo et al. 2005) and Barbados (Depradine and Lovell 2004), which may be related to the fact that the Aedes mosquito larvae are washed away from containers during heavy downpours. This has potentially important public health implications for disease incidence forecasting, as extra rainfall does not necessarily imply an immediate risk of dengue or other vector/ rodent borne disease. This type of prediction may be important for other vector-based diseases like malaria and Japanese encephalitis, whose vectors have different environmental requirements. However, it is possible that the risk of dengue will be greater following heavy rainfall when mosquitoes can re-establish breeding sites and the seasonal larval index of Aedes mosquitoes increases (Strickman and Kittayapong 2002), which may have occurred in Delhi, India where an unexpected outbreak followed the wettest monsoon in 25 years (Chakravarti and Kumaria 2005). The potential consequences of extreme events are evident, as torrential rainfall from Hurricane Mitch in 1998 killed 11,000 people in Central America, and Honduras reported 30,000 cholera, 30,000 malaria and 1,000 dengue cases (Epstein 1999). Similarly, severe flooding in southern Mozambique in 2000 killed hundreds of people, displaced thousands and caused the spread of malaria, typhoid and cholera (Kondo et al. 2002; Ahern et al. 2005). During this period, roughly 17,000 cases of cholera were reported. The impact of natural disasters is devastating worldwide (as recently shown by hurricane Katrina), but particularly in developing countries where populations are highly vulnerable, frequently displaced, homes and livelihoods are destroyed and local infrastructure is severely damaged (Allan et al. 1998; Few et al. 2004). Such events also have a severe impact on the health care sector and its ability to respond adequately. This further emphasizes the value of developing early warning systems – including cyclone warning systems – to aid health systems and health workers to be better prepared for impending disaster. This would, in turn, reduce the population’s vulnerability to disasters and improve disaster management. A detailed international disaster database (EM-DAT http://www.em-dat.net/)
46
L. Kelly-Hope and M.C. Thomson
collates data from various sources, and provides some insight into the magnitude of the effect and the economic losses that have occurred due to disasters over the past century (Ahern et al. 2005; EM-DAT 2005). Climate-sensitive diseases frequently exhibit interannual variability associated with unusual weather conditions, which may recur periodically. It is useful for national health systems to be aware of the various interannual climatic patterns that could impact on disease patterns in their country. Gill (1936) was among the first to describe the varying temporal patterns of malaria in different regions of the world. Epidemics were noted to occur every 5 years in Sri Lanka, every 10 years in India (Punjab), every 11 years in Argentina and every 12 years in Algeria, and it was suggested that the 11 year sunspot cycle (or certain phases thereof) may be related to the epidemics in the latter three countries, as well as the great malaria epidemic in Mauritius in 1867. Similar trends were observed in East Africa and elsewhere with yellow fever outbreaks. Seasonal climate anomalies result from complex interactions between the atmosphere and the underlying surfaces: that is, the world oceans and land surfaces. The atmosphere, which fluctuates very rapidly on a day-to-day basis (weather), is tied to the more slowly evolving components of the earth system, which are capable of exerting a sustained influence on climate anomalies extending over a season or longer, far beyond the 1–2-week limit of deterministic predictability of the weather. The atmosphere is particularly sensitive to tropical sea surface temperature (SST) anomalies such as those that occur in association with the El Niño/Southern Oscillation (ENSO) (Goddard et al. 2001) but other phenomena may also be important. A range of indices are used to assess the links between large-scale climate fluctuations and disease incidence.
3.4.2
ENSO
ENSO is the most commonly studied driver of cyclical climate phenomenon with regard to human disease (Kovats et al. 1999, 2003). It is characterized by exceptionally marked and prolonged warm periods of sea surface temperature (SST) that appear in the central and eastern equatorial Pacific Ocean every 3–7 years. Extreme phases of the ENSO phenomenon have been linked to precipitation anomalies in many areas of the world (Ropelewski and Halpert 1987). In some areas precipitation may increase during warm (El Niño) or cold (La Niña) ENSO events, while in others drought may be more likely. For example, in southern Africa droughts and drought disasters tend to occur in the December–March rainy season following the onset of an El Niño event (Thomson and Abayonmi 2003). An El Niño event, or the impact of one a year later, correlates with increased malarial risks in Sri Lanka (1870–1945), India/Pakistan (1867–1943) (Bouma and van der Kaay 1996), Pakistan (1970–1993) (Bouma and van der Kaay 1994), India (Rajasthan) (1982–1992), Venezuela (1975–1990) (Bouma and Dye 1997) and Colombia (1960–1992) (Bouma et al. 1997, 2001). A large El Niño event in 1997/98
3 Climate and Infectious Diseases
47
led to widespread flooding in East Africa, and a sixfold increase in malaria in Kenya during the first 2 months of 1998, compared with 1997, which also coincided with outbreaks of Rift Valley fever and cholera (Allan et al. 1998; McLigeyo 1998; Linthicum et al. 1999). Significant El Niño associations have also been found with dengue outbreaks in Thailand (Cazelles et al. 2005), South Pacific (Hales et al. 1996, 1999) and South America (Gagnon et al. 2001), with Ross River virus disease in Australia (Woodruff et al. 2002) and with cholera in Bangladesh (Pascual et al. 2000; Rodo et al. 2002) and Peru (Speelmon et al. 2000; Lama et al. 2004). In South America, ENSO related flooding in the dry coastal region of north Peru was associated with a malaria epidemic, while drought conditions were important factors for malaria outbreaks in Columbia, Guyana and Venezuela (Gagnon et al. 2002) Rather than relating ENSO events to particular outbreaks it may be more useful to use the underlying indices which determine an ENSO or similar event and thus provide a continuous indication of climate conditions. Numerous indices of the ENSO phenomenon have been derived, but a simple average of sea-surface temperatures (SSTs) over the area 5° N–5° S, 170°–120° W, known as the Niño 3.4 region (Barnston et al. 1997) is widely used because of its conceptual simplicity and ease of calculation. A relationship between the associated NINO 3 index and leishmaniasis in Brazil has been demonstrated (Franke et al. 2002b). Recently in Botswana, the NINO 3.4 index was significantly correlated with standardized malaria incidence anomalies (Thomson et al. 2005). The correlation of malaria anomalies with SST has been shown to be a regional phenomena (Mabaso et al. 2007b). These studies suggest the potential use of SSTs for disease forecasting through their influence on regional climate variability. Acquisition of climate datasets and the use of GIS and RS technology may enable research teams to develop more specific regional disease models, which could be used by the national health systems. While SST patterns in the eastern equatorial pacific drive ENSO events variations in sea surface temperature elsewhere, e.g. the Western Indian Oceans are also important factors that drive climate variability, and are useful guides to predicting regional climate patterns and related diseases. In East Africa, outbreaks of Rift Valley fever have been associated with above-average rainfall and elevated SSTs in the western Indian Ocean and El Niño events in the Pacific (Linthicum et al. 1999). It is well understood that the interannual variability of Rift Valley fever in sheep and goats relates directly to Aedes mosquitoes that emerge from transovarially infected eggs found on the edges of dambos (depressions located in the valley bottoms of many headwater catchments) after periods of excessive rainfall and flooding (Linthicum et al. 1985). Humans may become infected with RVF through mosquito bites but are more commonly infected through contact with blood, body fluids or organs of infected animals or ingestion of raw milk. For some diseases it has been suggested that changes in SST may increase the risk of disease outbreaks directly. High SSTs during the 1991/92 El Niño may have contributed to the reemergence of cholera in Peru, and its subsequent spread across the South American subcontinent, affecting thousands of people (Colwell 1996). Correlations between SSTs in the Bay of Bengal and cholera cases in Dhaka have been observed (Lobitz et al. 2000).
48
L. Kelly-Hope and M.C. Thomson
Indices based on atmospheric phenomena have also be related to climate-sensitive disease outbreaks. One such index is the Southern Oscillation Index (SOI) which represents the atmospheric expression of ENSO, with records dating back to the late 19th century. Specifically, the SOI measures the normalized difference in sea level pressure between Tahiti and Darwin, Australia, and variations indicate changes in the locality of convective rainfall. Negative values of the SOI are associated with warmer than average sea surface temperatures in the central-eastern equatorial Pacific and a shift of convective rainfall from the Austral-Indonesian region to the central tropical Pacific, or El Niño conditions. Positive SOI values are associated with cooler than average sea temperatures and intensification of convective rainfall in the far western Pacific, or La Niña conditions. An SOI analysis may also highlight differences between diseases. Such an analysis was conducted with data from China, where analysis of two decades of data on malaria and hemorrhagic fever with renal syndrome indicated that there were positive and negative associations, respectively, between the SOI and the monthly incidence of each disease (Bi et al. 2005). Other indices of large-scale climate fluctuations include the Quasi-Biennial Oscillation (QBO) index (formally known as Singapore Winds), which was studied with regard to the Ross River virus disease incident in Australia, and was found to account for 77% of the disease variance in cases (Done et al. 2002).
3.5
Early Warning Systems
While nature has conspired to limit our ability to forecast day-to-day weather, there exists a firm scientific basis for the prediction of seasonal mean climate anomalies (i.e., departures from normal of averages and other statistics of weather over a season or longer) largely based on our ability to monitor global SSTs, which, as we note above, constitute the primary forcing of seasonal climate variability (Stockdale et al. 1998; Goddard et al. 2001). The scientific basis for seasonal forecasting is described in detail by Mason (this book). Recently, a system for forecasting anomalously high and low malaria incidence anomalies using dynamically based seasonal-timescale ensemble predictions of climate has been reported (Thomson et al. 2006a). Several European models of the coupled ocean–atmosphere system are combined into a multi-model ensemble forecast system and successfully applied to the prediction of malaria risk in Botswana, where links between malaria and climate variability are well established (Thomson et al. 2005). The practical application of this work in Southern Africa has been documented (Connor and Mantilla, this volume; Connor et al. 2007). What characterizes this approach is that climate forecasting and monitoring is integrated into an early warning system which includes vulnerability assessment and routine case surveillance. The monitoring component is supported by a web-based tool provides timely alerts to control programs and international organizations (i.e. Roll Back Malaria partners) about increased epidemic risks in Africa (Anon 2002; Grover-Kopec et al. 2005).
3 Climate and Infectious Diseases
49
An early warning system based on this approach has recently been developed in Eritrea for operational use (Ceccato et al. 2008). Eritrea has two distinct rainy seasons in different parts of the country. The seasonal forecasting skill from Global Circulation Models was low for both seasons with the exception of the June–July– August season on the Eastern border. For epidemic control, shorter range warning based on remotely sensed rainfall estimates (rain gauge data are too few) and an enhanced epidemic early detection system were recommended. The Highland Malaria Project (HIMAL) is another initiative being developed for the early warning and detection of epidemics, using weather monitoring and a network of sentinel sites in four pilot districts of Kenya and Uganda (Abeku et al. 2004). Although these computer-based tools are currently limited to organizations with advanced technological resources, these approaches will become increasingly more commonplace, as developing countries attain requisite computer hardware and software tools, access to the internet and staff are trained in appropriate methodologies and tools. Other examples of early warning systems include ‘The Global Emerging Infections System’ run by the US Department of Defense (DoD-GEIS 2005). In collaboration with NASA, a system was designed for predicting Rift Valley fever in East Africa, based on a combination of warmer-than-normal equatorial Pacific Ocean temperatures associated with El Niño and rising SSTs in the western equatorial Indian Ocean. Outgoing Longwave Radiation (OLR) anomalies and increases in the normalized vegetation index (NDVI) A recent advance in this area of enquiry has been detailed evidence of the high level of predictability of NDVI following the East African Short rains from global climate models (Indeje et al. 2006). Although this system provides no information on the size of the impending epidemic the spatial risk maps produced by this process, once validated, could help identify high-risk locations, which in turn could lead to domestic animal vaccinations and the implementation of appropriate interventions such as mosquito control programs and public health. Monthly risk maps are available through the internet (DoD-GEIS 2005). A recent survey of the literature (published and grey) (Savory et al. 2006) has shown that recent epidemics of meningococcal meningitis in Africa largely fall within the risk areas identified in a previous study from a climate–landcover derived spatial risk model (Molesworth et al. 2003). One important reason for the limited number of early warning systems is the fact that many epidemic prone countries lack good epidemiological surveillance systems to enable early detection and provide the baseline data to improve the understanding of the drivers of epidemics (Cox and Abeku 2007). Several smaller scale (national and subnational) studies have developed early warning indices or models based on weather factors, seasonal trends and past disease cases. For example, studies have been conducted on malaria in Ethiopia (Abeku et al. 2002; Teklehaimanot et al. 2004a, b), dengue in Thailand (Bartley et al. 2002), Ross River virus disease in Australia (Gatton et al. 2005), meningitis in Mali (Sultan et al. 2005) and on cholera in Bangladesh (Koelle et al. 2005a). The advantage of these smaller studies is that they may provide more specific localized information. These studies use a range of climate indicators and statistical methods,
50
L. Kelly-Hope and M.C. Thomson
and provide a framework for future studies. The scale of the study is secondary to the availability and utility of an appropriate seasonal reference tool to predict any impending disease risk.
3.6
Trends
Longer-term trends in health data may be caused by a number of drivers including ‘climate shifts’ (used here to describe changes in the climate which occur abruptly over the period of a decade or two or as significant trends). The best example is that of the Sahelian drought during which the region lost approximately 30% of its annual rainfall over a 30-year period. Such shifts in climate may have profound effects on the spatial and temporal distribution of climate-sensitive diseases (Thomson et al. 2004b). In the Gambia, several studies have suggested that the epidemiological course of malaria is slowly displaying more epidemic and unstable patterns, possibly owing to the shorter rainy seasons associated with the protracted Sahelian drought, with devastating consequences on the acquisition of childhood immunity (Brewster and Greenwood 1993). In Rwanda, an upsurge in malaria during the late 1980s was linked to enhanced transmission resulting from increased trend in temperature and rainfall (Loevinsohn 1994). In Pakistan, the climate records gathered since 1876 showed mean increases in temperature by 1.5–2°C during November and December, October rainfall, and mean humidity (since 1950) (Bouma et al. 1996). It was suggested that these changes rendered climate conditions more favorable for malaria transmission, and may account for the increase in the number of malaria cases during the mid-1990s. The potential impact of climate changes over longer time-frames on climatesensitive diseases remains uncertain (McMichael et al. 2003). However, it is hypothesized that global warming and associated temperature increases will expand the geographical and temporal ranges, and thus the prevalence of these diseases (Epstein 2001, 2002, 2005). Most vulnerable to these changes are vector-borne diseases such as malaria, dengue, leishmaniasis and schistosomiasis, because of the climate dependencies of the vectors such as mosquitoes, flies and snails (Patz et al. 2005) and in particular the important role of temperature in the disease transmission dynamics. Although there has been much discussion in the literature regarding the importance of climate change as a driver of recent epidemics in the East Africa highlands (Hay et al. 2002a; Patz et al. 2002; Pascual et al. 2006) there is a general consensus that malaria is among the most sensitive diseases to such changes. Indeed, small increases in temperature have a disproportionate effect on malaria infection risk, especially in fringe communities where malaria is unstable or epidemic in nature. Mathematical modeling suggest an increase in the number of people at risk to malaria as a direct consequence of climate change is highly probable (McMichael et al. 2003) although human factors are likely to play a dominant role in many areas of the world. The risk of contracting other diseases will certainly change, and in some cases increase, as suggested in a recent study in China, which determined that an average temperature increase of 1°C over the last 30 years in
3 Climate and Infectious Diseases
51
January would put at risk an additional 20.7 million people for contracting schistosomiasis (Yang et al. 2005b). As pointed out by Cazelles and Hales (2006) a major consideration in the analysis of health and climate data over long time periods (especially in the context of a changing climate) is the potential nonstationarity of the relationship including (a) nonstationarity of the average, leading to a trend in the observed time series (b) nonstationarity of the variance, including changes in dominant periodic components over time, and (c) nonstationarity of the relationships between several observed signals. This has important public health implications as nonstationarity associated with climate change may limit the opportunity to create forecasting models based on robust relationships between climate drivers and health outcomes.
3.7
Other Factors
Much of the impact of climate on disease transmission is mediated through the impact of climatic factors on the environment including creation of breeding sites (flooding) and landcover (vector or animal reservoir habitat). For example, high malaria in India has been attributed to a high water table, soil type, irrigation and water quality (Srivastava et al. 1999). In Belize, high malaria outbreaks have been linked to vegetation and its proximity to rivers (Hakre et al. 2004). In Uganda, land use change may alter malaria transmission rates by modifying temperature in a highland region (Lindblade et al. 2000), while in Asia rice cultivation and pig breeding have increased the spread of Japanese encephalitis. In Africa, livestock movements have been linked to the dissemination of Rift Valley fever (Chevalier et al. 2004) and the prevalence of schistosomiasis to local water development and dam construction (Ernould et al. 1999). These examples highlight the likelihood that climate-sensitive diseases are complex and influenced by a multitude of auxiliary factors that interact to promote disease infectivity and transmission. Other non-climatic factors controlling disease status are also important, and although not highlighted in this chapter, should be considered. Climate is but one aspect of the spatial and temporal variability of climate-sensitive diseases. Infectious diseases typically possess a complex aetiology, and are likely caused and propagated by many different processes. Social, biological and economic factors such as population immunity, local topography, population migration, urbanization, land use patterns, housing conditions, health service provision, drug resistance and mosquito control measures also may act as important drivers in disease transmission.
3.8
Conclusion
This review highlights the range of studies conducted on climate and 20 infectious diseases of major public health importance. The literary gaps are highlighted in the accompanying Table 3.1, and demonstrate that malaria and dengue are the most
52
L. Kelly-Hope and M.C. Thomson
researched of the infectious diseases, with a variety of studies conducted on different continents at different times. Studies on chronic diseases were more focused on spatial patterns, with regard to climatic factors, while acute diseases tended to be studied with respect to seasonality and interannual variability. Few studies examined seasonal climate forecasts or climate shifts, which may be related to the lack of high resolution long-term disease datasets necessary for such studies. The studies cited come from many countries, however, mostly only 1–3 reports of the 20 diseases investigated come from the same country. The countries with the broadest range of diseases studied included India (malaria, dengue, Japanese encephalitis, trachoma, rotavirus), Kenya (malaria, trypanosomiasis, Rift Valley fever, trachoma, rotavirus) and Brazil (malaria, schistosomiasis, leishmaniasis, yellow fever, rotavirus). The 20 diseases chosen highlights those climate-sensitive diseases that are high-level health care concerns, and are highly prevalent in many countries. Prioritizing diseases (listing 1–5 most prevalent/burdensome) may help national health systems to focus on improving data collection, defining diseases and identifying interventions and cures. In general, very little research on climate and disease published in English in peer review journals has been conducted in the countries with the highest numbers of child deaths (India, Nigeria, China, Pakistan and the Democratic Republic of Congo), especially under-five mortality rates (Sierra Leone, Niger, Angola, Afghanistan, Liberia) (Black et al. 2003). If MDGs are to be met, then these countries should be the focus of future research, interventions, health worker training and infrastructure development. For those countries where natural or political disasters have created a state of emergency or regions emerging from periods of political instability, where rapid progress must be made to re-establish health services then predictive studies in neighboring countries, and extrapolations may help to identify high risk locations and times, and facilitate the appropriate health care via NGOs, for example (Thomson et al. 1999). Although this review elucidates many disease-specific gaps, it helps to define areas in need of future research and how climate analysis or tools may aid this endeavor. For example, very little information is documented on cholera climate interactions outside of Asia (Griffith et al. 2006) although recent studies from West Africa indicate that regional climate variability may be a driver (de Magny et al. 2006; Constantin et al. 2006). This disease is of particular concern on the continent, since over 90% of the cases reported to WHO occur in Africa. Whilst it is difficult to collect accurate data, especially when most cases are from epidemics occurring from natural disasters and/or political instability, retrospective analysis of selected outbreaks and meteorological data may highlight putative climatic risk factors. Furthermore, we have restricted this review to 20 climate-sensitive diseases, but other disease such as Chagas disease (American trypanosomiasis), Crimean-Congo hemorrhagic fever, tick-borne encephalitis, shigellosis, typhoid and influenza should also be considered in future climate-related studies of disease. Understanding the effects of weather and climate variability and change on the epidemiology of infectious diseases is important for planning specific disease related interventions and monitoring their impact. Specifically, climate information
3 Climate and Infectious Diseases
53
can help national health systems to define where, when and who is the most at-risk. It is essential, however, that more ‘ground truth’ epidemiological time series data is collected. Without these ‘standardized’ data sets an understanding of basic epidemiology and associated risk factors is limited. As part of international initiatives and elimination programs, baseline and monitoring data are now being collected for diseases such as malaria, schistosomiasis, lymphatic filariasis, onchocerciasis, trypanosomiasis and trachoma in many regions of the world in order to assess the burden of disease and the impact of interventions. Similar datasets are needed to aid in the control of many other diseases. It is also important that while these datasets are utilized for programmatic operational research purposes on the one hand they are available for local decision making on the other (Macfarlane et al. 2007). For this to be possible methodologies and tools should be designed that are appropriate for local and regional use and capacity to use these must be developed throughout the health sector and with appropriate partners. Since the climate and environmental information required for such studies comes from outside of the health sector, it is essential that sustainable multi-sectoral collaborations be established. Overcoming current constraints to such partnerships (IRI 2006) is a pre-requisite for moving forward. Acknowledgements This paper was funded in part by a grant/cooperative agreement from the National Oceanic and Atmospheric Administration (NOAA), contract number NA07GP0213 with the Trustees of Columbia University. The views expressed herein are those of the authors and do not necessarily reflect the views of NOAA or any of its sub-agencies. Comments on earlier drafts by L. Goddard and S.J. Connor and assistance with formatting and proof reading by M. Hellmuth and M. Salgado are gratefully acknowledged.
References Abeku, T. A., S. J. de Vlas, G. Borsboom, A. Teklehaimanot, A. Kebede, D. Olana, G. J. van Oortmarssen and J. D. Habbema (2002). “Forecasting malaria incidence from historical morbidity patterns in epidemic-prone areas of Ethiopia: a simple seasonal adjustment method performs best.” Tropical Medicine and International Health 7(10): 851–857. Abeku, T. A., G. J. van Oortmarssen, G. Borsboom, S. J. de Vlas and J. D. Habbema (2003). “Spatial and temporal variations of malaria epidemic risk in Ethiopia: factors involved and implications.” Acta Tropica 87(3): 331–340. Abeku, T. A., S. I. Hay, S. Ochola, P. Langi, B. Beard, S. J. de Vlas and J. Cox (2004). “Malaria epidemic early warning and detection in African highlands.” Trends in Parasitology 20(9): 400–405. Afari, E. A., T. Nakano, F. Binka, S. Owusu-Agyei and J. Asigbee (1993). “Seasonal characteristics of malaria infection in under-five children of a rural community in southern Ghana.” West African Journal of Medicine 12(1): 39–42. Ahern, M., R. S. Kovats, P. Wilkinson, R. Few and F. Matthies (2005). “Global health impacts of floods: epidemiologic evidence.” Epidemiological Review 27: 36–46. Ahmed, M. U., S. Urasawa, K. Taniguchi, T. Urasawa, N. Kobayashi, F. Wakasugi, A. I. Islam and H. A. Sahikh (1991). “Analysis of human rotavirus strains prevailing in Bangladesh in relation to nationwide floods brought by the 1988 monsoon.” Journal of Clinical Microbiology 29(10): 2273–2279. Akenji, T. N., N. N. Ntonifor, H. K. Kimbi, E. L. Abongwa, J. K. Ching, M. B. Ndukum, D. N. Anong, A. Nkwescheu, M. Songmbe, M. G. Boyo, K. N. Ndamukong and V. P. Titanji (2005).
54
L. Kelly-Hope and M.C. Thomson
“The epidemiology of malaria in Bolifamba, a rural community on the eastern slopes of Mount Cameroon: seasonal variation in the parasitological indices of transmission.” Annals of Tropical Medicine and Parasitology 99(3): 221–227. Akhtar, R. and A. J. McMichael (1996). “Rainfall and malaria outbreaks in western Rajasthan.” The Lancet 348(9039): 1457–1458. Al-Nakib, W., I. L. Chrystie, J. E. Banatvala and F. Al-Sayegh (1980). “Rotavirus and non-bacterial infantile gastroenteritis in Kuwait.” International Journal of Epidemiology 9(4): 355–359. Allan, R., S. Nam and L. Doull (1998). “MERLIN and malaria epidemic in north-east Kenya.” The Lancet 351(9120): 1966–1967. Almendares, J., M. Sierra, P. K. Anderson and P. R. Epstein (1993). “Critical regions, a profile of Honduras.” The Lancet 342(8884): 1400–1402. Amin, M. M. M., A. M. Z. Hussain, M. Murshed, I. A. Chowdhury, S. Mannan, S. A. Chowdhuri and D. Banu (1999). “Sero-diagnosis of dengue infections by haemagglutination inhibition Test (HI) in suspected cases in Chittagong, Bangladesh.” Dengue Bulletin 23: 34–38. Anon (1994). “Rapid assessment of vectorborne diseases during the Midwest flood – United States, 1993.” Morbidiy and Mortality Weekly Report 43(26): 481–483. Anon (1999). “El Nino and associated outbreaks of severe malaria in highland populations in Irian Jaya, Indonesia: a review and epidemiological perspective.” Southeast Asian Journal of Tropical Medicine and Public Health 30(4): 608–619. Anon (2002). “Web-based tool for early warning of malaria epidemics in Africa: monitoring current rainfall anomalies in zones at epidemic risk.” Weekly Epidemiological Record 77(32): 276. Anon (2004). “Informed choices for attaining the Millennium Development Goals: towards an international cooperative agenda for health-systems research.” The Lancet 364(9438): 997–1003. Anyamba, A., K. J. Linthicum and C. J. Tucker (2001). “Climate-disease connections: Rift Valley Fever in Kenya.” Cadernos de Saude Publica 17(Suppl): 133–140. Anyamba, A., K. J. Linthicum, R. Mahoney, C. J. Tucker and P. W. Kelley (2002). “Mapping potential risk of Rift Valley fever outbreaks in African savannas using vegetation index time series data.” Photogrammetric Engineering and Remote Sensing 68(2): 137–145. Armah, G. E., J. A. Mingle, A. K. Dodoo, A. Anyanful, R. Antwi, J. Commey and F. K. Nkrumah (1994). “Seasonality of rotavirus infection in Ghana.” Annals of Tropical Paediatrics 14(3): 223–229. Baird, J. K., S. Owusu Agyei, G. C. Utz, K. Koram, M. J. Barcus, T. R. Jones, D. J. Fryauff, F. N. Binka, S. L. Hoffman and F. N. Nkrumah (2002). “Seasonal malaria attack rates in infants and young children in northern Ghana.” American Journal of Tropical Medicine and Hygiene 66(3): 280–286. Barbazan, P., S. Yoksan and J. P. Gonzalez (2002). “Dengue hemorrhagic fever epidemiology in Thailand: description and forecasting of epidemics.” Microbes Infect 4(7): 699–705. Barker, C. M., W. K. Reisen and V. L. Kramer (2003). “California state Mosquito-Borne Virus Surveillance and Response Plan: a retrospective evaluation using conditional simulations.” American Journal of Tropical Medicine and Hygiene 68(5): 508–518. Barnston, A. G., M. Chelliah and S. B. Goldenberg (1997). “Documentation of a highly ENSOrelated SST region in the equatorial Pacific.” Atmosphere-Ocean 35(3): 367–383. Barrera, R., N. Delgado, M. Jimenez and S. Valero (2002). “Eco-epidemiological factors associated with hyperendemic dengue hemorrhagic fever in Maracay city, Venezuela.” Dengue Bulletin 26: 84–95. Bartley, L. M., C. A. Donnelly and G. P. Garnett (2002). “The seasonal pattern of dengue in endemic areas: mathematical models of mechanisms.” Transactions of the Royal Society of Tropical Medicine and Hygiene 96(4): 387–397. Bavia, M. E., L. F. Hale, J. B. Malone, D. H. Braud and S. M. Shane (1999). “Geographic information systems and the environmental risk of schistosomiasis in Bahia, Brazil.” American Journal of Tropical Medicine and Hygiene 60(4): 566–572.
3 Climate and Infectious Diseases
55
Bavia, M. E., J. B. Malone, L. Hale, A. Dantas, L. Marroni and R. Reis (2001). “Use of thermal and vegetation index data from earth observing satellites to evaluate the risk of schistosomiasis in Bahia, Brazil.” Acta Tropica 79(1): 79–85. Bavia, M. E., D. D. Carneiro, C. Gurgel Hda, C. Madureira Filho and M. G. Barbosa (2005). “Remote Sensing and Geographic Information Systems and risk of American visceral leishmaniasis in Bahia, Brazil.” Parassitologia 47(1): 165–169. Belcher, D. W., F. K. Wurapa, W. B. Ward and I. M. Lourie (1975). “Guinea worm in southern Ghana: its epidemiology and impact on agricultural productivity.” American Journal of Tropical Medicine and Hygiene 24(2): 243–249. Ben Salah, A. B., R. Ben Ismail, F. Amri, S. Chlif, F. Ben Rzig, H. Kharrat, H. Hadhri, M. Hassouna and K. Dellagi (2000). “Investigation of the spread of human visceral leishmaniasis in central Tunisia.” Transactons of the Royal Society of Tropical Medicine and Hygiene 94(4): 382–386. Besancenot, J. P., M. Boko and P. C. Oke (1997). “Weather conditions and cerebrospinal meningitis in Benin (Gulf of Guinea, West Africa).” European Journal of Epidemiology 13(7): 807–815. Bi, P. and K. A. Parton (2003). “El Nino and incidence of hemorrhagic fever with renal syndrome in China.” JAMA 289(2): 176–177. Bi, P., X. Wu, F. Zhang, K. A. Parton and S. Tong (1998). “Seasonal rainfall variability, the incidence of hemorrhagic fever with renal syndrome, and prediction of the disease in low-lying areas of China.” American Journal of Epidemiology 148(3): 276–281. Bi, P., S. Tong, K. Donald, K. Parton and J. Ni (2002). “Climatic, reservoir and occupational variables and the transmission of haemorrhagic fever with renal syndrome in China.” International Journal of Epidemiology 31(1): 189–193. Bi, P., S. Tong, K. Donald, K. A. Parton and J. Ni (2003a). “Climate variability and transmission of Japanese encephalitis in eastern China.” Vector Borne Zoonotic Disease 3(3): 111–115. Bi, P., S. Tong, K. Donald, K. A. Parton and J. Ni (2003b). “Climatic variables and transmission of malaria: a 12-year data analysis in Shuchen County, China.” Public Health Reports 118(1): 65–71. Bi, P., K. A. Parton and S. Tong (2005). “El Nino-Southern Oscillation and vector-borne diseases in Anhui, China.” Vector Borne Zoonotic Disease 5(2): 95–100. Bittencourt, J. A., E. Arbo, A. S. Malysz, R. Oravec and C. Dias (2000). “Seasonal and age distribution of rotavirus infection in Porto Alegre–Brazil.” Brazilian Journal of Infectious Disease 4(6): 279–283. Black, R. E., S. S. Morris and J. Bryce (2003). “Where and why are 10 million children dying every year?” The Lancet 361(9376): 2226–2234. Bodker, R., J. Akida, D. Shayo, W. Kisinza, H. A. Msangeni, E. M. Pedersen and S. W. Lindsay (2003). “Relationship between altitude and intensity of malaria transmission in the Usambara Mountains, Tanzania.” Journal of Medical Entomology 40(5): 706–717. Botto, C., E. Escalona, S. Vivas-Martinez, V. Behm, L. Delgado and P. Coronel (2005). “Geographical patterns of onchocerciasis in southern Venezuela: relationships between environment and infection prevalence.” Parassitologia 47(1): 145–150. Bouma, M. J. (2003). “Methodological problems and amendments to demonstrate effects of temperature on the epidemiology of malaria. A new perspective on the highland epidemics in Madagascar, 1972–89.” Transactions of the Royal Society of Tropical Medicine and Hygiene 97(2): 133–139. Bouma, M. J. and C. Dye (1997). “Cycles of malaria associated with El Nino in Venezuela.” JAMA 278(21): 1772–1774. Bouma, M. J. and H. J. van der Kaay (1994). “Epidemic malaria in India and the El Nino southern oscillation.” The Lancet 344(8937): 1638–1639. Bouma, M. J. and H. J. van der Kaay (1996). “The El Nino Southern Oscillation and the historic malaria epidemics on the Indian subcontinent and Sri Lanka: an early warning system for future epidemics?” Tropical Medicine and International Health 1(1): 86–96.
56
L. Kelly-Hope and M.C. Thomson
Bouma, M. J., C. Dye and H. J. van der Kaay (1996). “Falciparum malaria and climate change in the northwest frontier province of Pakistan.” American Journal of Epidemiology 55(2): 131–137. Bouma, M. J., G. Poveda, W. Rojas, D. Chavasse, M. Quinones, J. Cox and J. Patz (1997). “Predicting high-risk years for malaria in Colombia using parameters of El Nino Southern Oscillation.” Tropical Medicine and International Health 2(12): 1122–1127. Bouvier, P., A. Rougemont, N. Breslow, O. Doumbo, V. Delley, A. Dicko, M. Diakite, A. Mauris and C. F. Robert (1997). “Seasonality and malaria in a west African village: does high parasite density predict fever incidence?” American Journal of Epidemiology 145(9): 850–857. Brengues, J. (1975). La Filariiose de Bancroft en Afrique de L’Ouest. Paris, ORSTOM: 299. Brewster, D. R. and B. M. Greenwood (1993). “Seasonal variation of paediatric diseases in The Gambia, west Africa.” Annals of Tropical Paediatrics 13(2): 133–146. Brooker, S., S. I. Hay, W. Issae, A. Hall, C. M. Kihamia, N. J. Lwambo, W. Wint, D. J. Rogers and D. A. Bundy (2001). “Predicting the distribution of urinary schistosomiasis in Tanzania using satellite sensor data.” Tropical Medicine and International Health 6(12): 998–1007. Broom, A. K., M. D. Lindsay, A. J. Plant, A. E. Wright, R. J. Condon and J. S. Mackenzie (2002). “Epizootic activity of Murray Valley encephalitis virus in an aboriginal community in the southeast Kimberley region of Western Australia: results of cross-sectional and longitudinal serologic studies.” American Journal of Tropical Medicine and Hygiene 67(3): 319–323. Broom, A. K., M. D. Lindsay, A. E. Wright, D. W. Smith and J. S. Mackenzie (2003). “Epizootic activity of Murray Valley encephalitis and Kunjin viruses in an aboriginal community in the southeast Kimberley region of Western Australia: results of mosquito fauna and virus isolation studies.” American Journal of Tropical Medicine and Hygiene 69(3): 277–283. Burke, D., A. Carmichael and D. Focks (2001). Under the weather: climate, ecosystems, and infectious disease. Washington, DC, National Academy Press. Camargo, L. M., G. M. dal Colletto, M. U. Ferreira, M. Gurgel Sde, A. L. Escobar, A. Marques, H. Krieger, E. P. Camargo and L. H. da Silva (1996). “Hypoendemic malaria in Rondonia (Brazil, western Amazon region): seasonal variation and risk groups in an urban locality.” American Journal of Tropical Medicine and Hygiene 55(1): 32–38. Campagne, G., A. Schuchat, S. Djibo, A. Ousseini, L. Cisse and J. P. Chippaux (1999). “Epidemiology of bacterial meningitis in Niamey, Niger, 1981–96.” Bulletin of the World Health Organization 77(6): 499–508. Cazelles, B., M. Chavez, A. J. McMichael and S. Hales (2005). “Nonstationary influence of El Nino on the synchronous dengue epidemics in Thailand.” PLoS Medecine 2(4): e106. Cazelles, B., and S. Hales (2000). Climate cycles and forecasts of cutareous leishnanicesis, nonstationary vector-borne disease. Plos Medicine 3(8): e328. doi: 10/1371/journal. pmed.0030328 Ceccato, P., T. Ghebremeskel, M. Jaiteh, P. M. Graves, M. Levy, S. Ghebreselassie, A. Ogbamariam, A. G. Barnston, M. Bell, J. Del Corral, S. J. Connor, I. Fesseha, E. P. Brantly, M. C. Thomson, Malaria Stratification, climate and epidemic early warning in Eritrea. American Journal of Tropical Medicine (2008). Cedeno, J. E. M. (1986). “Rainfall and flooding in the Guayas river basin and its effects on the incidence of malaria 1982–1985.” Disasters 10: 107–111. Chakravarti, A. and R. Kumaria (2005). “Eco-epidemiological analysis of dengue infection during an outbreak of dengue fever, India.” Virology Journal 2: 32. Chaves, L. F. and M. Pascual (2007). “Climate cycles and forecasts of cutaneous leishmaniasis, a nonstationary vector-borne disease (vol 3, pg 1320, 2006).” Plos Medicine 4(3): 602–602. Chavez, M. R. C., V. P. Sedas, E. O. Borunda and F. L. Reynoso (2005). “Influence of water temperature and salinity on seasonal occurrences of Vibrio cholerae and enteric bacteria in oysterproducing areas of Veracruz, Mexico.” Marine Pollution Bulletin 50(12): 1641–1648. Cheesbrough, J. S., A. P. Morse and S. D. Green (1995). “Meningococcal meningitis and carriage in western Zaire: a hypoendemic zone related to climate?” Epidemiology and Infection 114(1): 75–92. Chen, H. X. and F. X. Qiu (1993). “Epidemiologic surveillance on the hemorrhagic fever with renal syndrome in China.” Chinese Medical Journal (Engl) 106(11): 857–863.
3 Climate and Infectious Diseases
57
Chen, H. X. and F. X. Qiu (1994). “Studies on the environment structure of natural nidi and epidemic areas of hemorrhagic fever with renal syndrome in China.” Chinese Medical Journal (Engl) 107(2): 107–112. Chevalier, V., S. de la Rocque, T. Baldet, L. Vial and F. Roger (2004). “Epidemiological processes involved in the emergence of vector-borne diseases: West Nile fever, Rift Valley fever, Japanese encephalitis and Crimean-Congo haemorrhagic fever.” Revue Scientifique et Technique 23(2): 535–555. Christophers, S. R. (1911). “Malaria in the Punjab.” Scientific Memoirs by the Officers of the Medical and Sanitary Departments of the Government of India 46. Coiro, J. R., M. M. Bendati, A. J. de Almeida Neto, C. F. Heuser and V. L. Vasconcellos (1983). “Rotavirus infection in Brazilian children with acute enteritis: a seasonal variation study.” American Journal of Tropical Medicine and Hygiene 32(5): 1186–1188. Colwell, R. R. (1996). “Global climate and infectious disease: the cholera paradigm.” Science 274(5295): 2025–2031. Connor, S. J., P. Ceccato, T. Dinku, J. Omumbo, E. Grover-Kopec and M. C. Thomson, (2006) Using Climate Information for Improved Health in Africa: Relevance, Constraints and Opportunities. Geospatial Health 1: 17–36. Connor, S. J., J. Da Silva and S. Katikiti (2007). Malaria control in Southern Africa. Climate Risk Management in Africa: Learning from Practice. M. E. Hellmuth, A. Moorhead, M. C. Thomson and J. Williams (eds). New York, International Research Institute for Climate and Society (IRI), Columbia University, p. 116. Constantin, G., M. De, J. F. Guegan, M. Petit and B. Cazelles (2006). “Regional-scale synchrony of cholera epidemics in western Africa with climate variability.” Epidemiology 17(6): S207–S207. Cook, S. M., R. I. Glass, C. W. LeBaron and M. S. Ho (1990). “Global seasonality of rotavirus infections.” Bulletin of the World Health Organization 68(2): 171–177. Cooper, S. N. (1964). “Trachoma – A bio-meteorological approach.” Journal of the All India Ophthalmoological Society 12: 50–58. Cordova, S. P., D. W. Smith, A. K. Broom, M. D. Lindsay, G. K. Dowse and M. Y. Beers (2000). “Murray Valley encephalitis in Western Australia in 2000, with evidence of southerly spread.” Communicable Disease Intelligence 24(12): 368–372. Corwin, A. L., R. P. Larasati, M. J. Bangs, S. Wuryadi, S. Arjoso, N. Sukri, E. Listyaningsih, S. Hartati, R. Namursa, Z. Anwar, S. Chandra, B. Loho, H. Ahmad, J. R. Campbell and K. R. Porter (2001). “Epidemic dengue transmission in southern Sumatra, Indonesia.” Transactions of the Royal Society of Tropical Medicine and Hygiene 95(3): 257–265. Corwin, A. L., D. Subekti, N. C. Sukri, R. J. Willy, J. Master, E. Priyanto and K. Laras (2005). “A large outbreak of probable rotavirus in Nusa Tenggara Timur, Indonesia.” American Journal of Tropical Medicine and Hygiene 72(4): 488–494. Cox, J., and J.A. Abeku (2007). Early warning systems for malaria in Africa: from blueprint to practice. Trends in Parasitology 23: 243–246. Craig, M. H., R. W. Snow and D. le Sueur (1999). “A climate-based distribution model of malaria transmission in sub-Saharan Africa.” Parasitology Today 15(3): 105–111. Craig, M. H., I. Kleinschmidt, J. B. Nawn, D. Le Sueur and B. L. Sharp (2004). “Exploring 30 years of malaria case data in KwaZulu-Natal, South Africa: Part I. The impact of climatic factors.” Tropical Medicine and International Health 9(12): 1247–1257. Cross, E. R., C. Sheffield, R. Perrine and G. Pazzaglia (1984). “Predicting areas endemic for schistosomiasis using weather variables and a Landsat data base.” Military Medicine 149(10): 542–544. Cunin, P., M. C. Fonkoua, B. Kollo, B. A. Bedifeh, P. Bayanak and P. M. Martin (2003). “Serogroup A Neisseria meningitidis outside meningitis belt in southwest Cameroon.” Emerging Infectious Diseases 9(10): 1351–1353. Cunliffe, N. A., P. E. Kilgore, J. S. Bresee, A. D. Steele, N. Luo, C. A. Hart and R. I. Glass (1998). “Epidemiology of rotavirus diarrhoea in Africa: a review to assess the need for rotavirus immunization.” Bulletin of the World Health Organization 76(5): 525–537.
58
L. Kelly-Hope and M.C. Thomson
Cvjetanovic, B., B. Grab and K. Uemura (1978). “Dynamics of acute bacterial diseases. Epidemiological models and their application in public health. Part II. Epidemiological models of acute bacterial diseases.” Bulletin of the World Health Organization 56(Suppl 1): 25–143. da Rosa e Silva, M. L., F. G. Naveca and I. Pires de Carvalho (2001). “Epidemiological aspects of rotavirus infections in Minas Gerais, Brazil.” Brazilian Journal of Infectious Disease 5(4): 215–222. DaSilva, J., B. Garanganga, V. Teveredzi, S. M. Marx, S. J. Mason and S. J. Connor (2004). “Improving epidemic malaria planning, preparedness and response in Southern Africa: Report on the 1st Southern African Regional Epidemic Outlook Forum, Harare, Zimbabwe, 26–29 September, 2004.” Malaria Journal 3(1): 37. Davies, F. G., K. J. Linthicum and A. D. James (1985). “Rainfall and epizootic Rift Valley fever.” Bulletin of the World Health Organization 63(5): 941–943. de Magny, G. C., B. Cazelles and J. F. Guegan (2006). “Cholera threat to humans in Ghana is influenced by both global and regional climatic variability.” Ecohealth 3(4): 223–231. Depradine, C. and E. Lovell (2004). “Climatological variables and the incidence of Dengue fever in Barbados.” International Journal of Environmental Health Research 14(6): 429–441. Dicko, A., C. Mantel, B. Kouriba, I. Sagara, M. A. Thera, S. Doumbia, M. Diallo, B. Poudiougou, M. Diakite and O. K. Doumbo (2005). “Season, fever prevalence and pyrogenic threshold for malaria disease definition in an endemic area of Mali.” Tropical Medicine and International Health 10(6): 550–556. DoD-GEIS (2005). Climate and Disease Connections: Rift Valley Fever Monitor. Department of Defense, Global Emerging Infections System. Done, S. J., N. J. Holbrook and P. J. Beggs (2002). “The Quasi-Biennial Oscillation and Ross River virus incidence in Queensland, Australia.” International Journal of Biometeorology 46(4): 202–227. Elfadil, A. A., K. A. Hasab-Allah and O. M. Dafa-Allah (2006). “Factors associated with Rift Valley fever in south-west Saudi Arabia.” Revue Scientifique Et Technique-Office International Des Epizooties 25(3): 1137–1145. Elnaiem, D. E., J. Schorscher, A. Bendall, V. Obsomer, M. E. Osman, A. M. Mekkawi, S. J. Connor, R. W. Ashford and M. C. Thomson (2003). “Risk mapping of visceral leishmaniasis: the role of local variation in rainfall and altitude on the presence and incidence of kala-azar in eastern Sudan.” American Journal of Tropical Medicine and Hygiene 68(1): 10–17. EM-DAT (2005). The International Disaster Database. Epstein, P. R. (1999). “Climate and health.” Science 285(5426): 347–348. Epstein, P. R. (2001). “Climate change and emerging infectious diseases.” Microbes Infect 3(9): 747–754. Epstein, P. R. (2002). “Climate change and infectious disease: stormy weather ahead?” Epidemiology 13(4): 373–375. Epstein, P. R. (2005). “Climate change and human health.” New England Journal of Medicine 353(14): 1433–1436. Ernould, J. C., K. Ba and B. Sellin (1999). “The impact of the local water-development programme on the abundance of the intermediate hosts of schistosomiasis in three villages of the Senegal River delta.” Annals of Tropical Medicine and Parasitology 93(2): 135–145. Few, R., M. Ahern, F. Matthies and S. Kovats (2004). Floods, health and climate change: a strategic review. Tyndall Centre Working Paper No. 63, Tyndall Centre for Climate Change Research. Franco, A. A., A. D. Fix, A. Prada, E. Paredes, J. C. Palomino, A. C. Wright, J. A. Johnson, R. McCarter, H. Guerra and J. G. Morris, Jr. (1997). “Cholera in Lima, Peru, correlates with prior isolation of Vibrio cholerae from the environment.” American Journal of Epidemiology 146(12): 1067–1075. Franke, C. R., C. Staubach, M. Ziller and H. Schluter (2002a). “Trends in the temporal and spatial distribution of visceral and cutaneous leishmaniasis in the state of Bahia, Brazil, from 1985 to 1999.” Transactions of the Royal Society of Tropical Medicine and Hygiene 96(3): 236–241.
3 Climate and Infectious Diseases
59
Franke, C. R., M. Ziller, C. Staubach and M. Latif (2002b). “Impact of the El Nino/Southern Oscillation on visceral leishmaniasis, Brazil.” Emerging Infectious Diseases 8(9): 914–917. Fun, B. N., L. Unicomb, Z. Rahim, N. N. Banu, G. Podder, J. Clemens, F. P. Van Loon, M. R. Rao, A. Malek and S. Tzipori (1991). “Rotavirus-associated diarrhea in rural Bangladesh: two-year study of incidence and serotype distribution.” Journal of Clinical Microbiology 29(7): 1359–1363. Gagnon, A. S., A. B. G. Bush and K. E. Smoyer-Tomic (2001). “Dengue epidemics and the El Niño Southern Oscillation.” Climate Research 19 35–43. Gagnon, A. S., K. E. Smoyer-Tomic and A. B. Bush (2002). “The El Nino southern oscillation and malaria epidemics in South America.” International Journal of Biometeorology 46(2): 81–89. Gatton, M. L., B. H. Kay and P. A. Ryan (2005). “Environmental predictors of Ross River virus disease outbreaks in Queensland, Australia.” American Journal of Tropical Medicine and Hygiene 72(6): 792–799. Gebre-Michael, T., J. B. Malone and K. McNally (2005). “Use of Geographic Information Systems in the development of prediction models for onchocerciasis control in Ethiopia.” Parassitologia 47(1): 135–144. Giha, H. A., S. Rosthoj, D. Dodoo, L. Hviid, G. M. Satti, T. Scheike, D. E. Arnot and T. G. Theander (2000). “The epidemiology of febrile malaria episodes in an area of unstable and seasonal transmission.” Transactions of the Royal Society of Tropical Medicine and Hygiene 94(6): 645–651. Gill, C. A. (1921). “The role of meteorology in malaria.” Indian Journal of Medical Research 8: 633–693. Gill, C. A. (1923). “The prediction of malaria epidemics.” Indian Journal of Medical Research 10: 1136–1143. Gill, C. A. (1936). “Some points in the epidemiology of malaria arising out of the study of the malaria epidemic in Ceylon in 1934–35.” Transactions of the Royal Society of Tropical Medicine and Hygiene 5: 427–466. Girgis, N. I., J. E. Sippel, M. E. Kilpatrick, W. R. Sanborn, I. A. Mikhail, E. Cross, M. W. Erian, Y. Sultan and Z. Farid (1993). “Meningitis and encephalitis at the Abbassia Fever Hospital, Cairo, Egypt, from 1966 to 1989.” American Journal of Tropical Medicine and Hygiene 48(1): 97–107. Goddard, L., S. J. Mason, S. E. Zebiak, C. F. Ropelewski, R. Basher and M. A. Cane (2001). “Current approaches to seasonal-to-interannual climate predictions.” International Journal of Climatology 21(9): 1111–1152. Gomez, C., A. J. Rodriguez-Morales and C. Franco-Paredes (2006). “Impact of climate variability in the occurrence of leishmaniasis in Bolivia.” American Journal of Tropical Medicine and Hygiene 75(5): 42–42. Gomwalk, N. E., L. T. Gosham and U. J. Umoh (1990). “Rotavirus gastroenteritis in pediatric diarrhoea in Jos, Nigeria.” Journal of Tropical Paediatrics 36(2): 52–55. Gomwalk, N. E., U. J. Umoh, L. T. Gosham and A. A. Ahmad (1993). “Influence of climatic factors on rotavirus infection among children with acute gastroenteritis in Zaria, northern Nigeria.” Journal of Tropical Paediatrics 39(5): 293–297. Greenwood, B. M., A. K. Bradley, P. G. Cleland, M. H. Haggie, M. Hassan-King, L. S. Lewis, J. T. Macfarlane, A. Taqi, H. C. Whittle, A. M. Bradley-Moore and Q. Ansari (1979). “An epidemic of meningococcal infection at Zaria, Northern Nigeria. 1. General epidemiological features.” Transactions of the Royal Society of Tropical Medicine and Hygiene 73(5): 557–562. Greenwood, B. M., I. S. Blakebrough, A. K. Bradley, S. Wali and H. C. Whittle (1984). “Meningococcal disease and season in sub-Saharan Africa.” The Lancet 1(8390): 1339–1342. Greenwood, B. M., A. K. Bradley and R. A. Wall (1985). “Meningococcal disease and season in sub-Saharan Africa.” The Lancet 2(8459): 829–830. Greenwood, B. M., A. K. Bradley, A. M. Greenwood, P. Byass, K. Jammeh, K. Marsh, S. Tulloch, F. S. Oldfield and R. Hayes (1987). “Mortality and morbidity from malaria among children in
60
L. Kelly-Hope and M.C. Thomson
a rural area of The Gambia, West Africa.” Transactions of the Royal Society of Tropical Medicine and Hygiene 81(3): 478–486. Griffith, D. C., L. A. Kelly-Hope and M. A. Miller (2006). “Review of reported cholera outbreaks worldwide, 1995–2005.” American Journal of Tropical Medicine and Hygiene 75(5): 973–977. Grover-Kopec, E., M. Kawano, R. W. Klaver, B. Blumenthal, P. Ceccato and S. J. Connor (2005). “An online operational rainfall-monitoring resource for epidemic malaria early warning systems in Africa.” Malaria Journal 4(1): 6. Gupta, R. (1996). “Correlation of rainfall with upsurge of malaria in Rajasthan.” Journal of the Association of Physicians of India 44(6): 385–389. Gupta, U. C. and V. V. Preobragenski (1964). “Trachoma in India: Endemicity and Epidemiological Study.” Journal of the All-India Ophthalmological Society 12: 39–49. Guthmann, J. P., A. Llanos-Cuentas, A. Palacios and A. J. Hall (2002). “Environmental factors as determinants of malaria risk. A descriptive study on the northern coast of Peru.” Tropical Medicine and International Health 7(6): 518–525. Gyapong, J. O., M. Gyapong and S. Adjei (1996). “The epidemiology of acute adenolymphangitis due to lymphatic filariasis in northern Ghana.” American Journal of Epidemiology 54(6): 591–595. Gyapong, J. O., D. Kyelem, I. Kleinschmidt, K. Agbo, F. Ahouandogbo, J. Gaba, G. OwusuBanahene, S. Sanou, Y. K. Sodahlon, G. Biswas, O. O. Kale, D. H. Molyneux, J. B. Roungou, M. C. Thomson and J. Remme (2002). “The use of spatial analysis in mapping the distribution of bancroftian filariasis in four West African countries.” Annals of Tropical Medicine and Parasitology 96(7): 695–705. Haberberger, R. L. J., E. Fox, P. Asselin, Y. Said-Salah, S. Martinez and E. A. Abbatte (1990). “Is Djibouti too hot and too humid for meningococci?” Transactions of the Royal Society of Tropical Medicine and Hygiene 84(4): 588. Haffejee, I. E. and A. Moosa (1990). “Rotavirus studies in Indian (Asian) South African infants with acute gastro-enteritis: II. Clinical aspects and outcome.” Annals of Tropical Paediatrics 10(3): 245–254. Haines, A. and J. A. Patz (2004). “Health effects of climate change.” JAMA 291(1): 99–103. Hakre, S., P. Masuoka, E. Vanzie and D. R. Roberts (2004). “Spatial correlations of mapped malaria rates with environmental factors in Belize, Central America.” International Journal of Health Geography 3(1): 6. Hales, S., P. Weinstein and A. Woodward (1996). “Dengue fever epidemics in the South Pacific: driven by El Nino Southern Oscillation?” The Lancet 348(9042): 1664–1665. Hales, S., P. Weinstein, Y. Souares and A. Woodward (1999). “El Nino and the dynamics of vectorborne disease transmission.” Environmental Health Perspectives 107(2): 99–102. Hales, S., N. de Wet, J. Maindonald and A. Woodward (2002). “Potential effect of population and climate changes on global distribution of dengue fever: an empirical model.” The Lancet 360(9336): 830–834. Hanlon, P., L. Hanlon, V. Marsh, P. Byass, F. Shenton, R. C. Sanders, M. Hassan-King and B. M. Greenwood (1987). “Epidemiology of rotavirus in a periurban Gambian community.” Annals of Tropical Paediatrics 7(4): 238–243. Hansen, J. W., M. Dilley, L. Goddard, E. Ebrahimian and P. Ericksen (2004). Climate Variability and the Millennium Development Goal Hunger Target. New York, The Earth Institute, Columbia University. Harley, D. O. and P. Weinstein (1996). “The Southern Oscillation Index and Ross river virus outbreaks.” Medical Journal of Australia 165(9): 531–532. Hassan, A. N., L. R. Beck and S. Dister (1998a). “Prediction of villages at risk for filariasis transmission in the Nile Delta using remote sensing and geographic information system technologies.” Journal of the Egyptian Society for Parasitology 28(1): 75–87. Hassan, A. N., S. Dister and L. Beck (1998b). “Spatial analysis of lymphatic filariasis distribution in the Nile Delta in relation to some environmental variables using geographic information system technology.” Journal of the Egyptian Society for Parasitology 28(1): 119–131.
3 Climate and Infectious Diseases
61
Hay, S. I., M. F. Myers, D. S. Burke, D. W. Vaughn, T. Endy, N. Ananda, G. D. Shanks, R. W. Snow and D. J. Rogers (2000). “Etiology of interepidemic periods of mosquito-borne disease.” Proceedings of the National Accademy of Sciences of the USA 97(16): 9335–9339. Hay, S. I., D. J. Rogers, G. D. Shanks, M. F. Myers and R. W. Snow (2001). “Malaria early warning in Kenya.” Trends in Parasitology 17(2): 95–99. Hay, S. I., D. J. Rogers, S. E. Randolph, D. I. Stern, J. Cox, G. D. Shanks and R. W. Snow (2002a). “Hot topic or hot air? Climate change and malaria resurgence in East African highlands.” Trends in Parasitology 18(12): 530–534. Hay, S. I., J. Cox, D. J. Rogers, S. E. Randolph, D. I. Stern, G. D. Shanks, M. F. Myers and R. W. Snow (2002b). “Climate change and the resurgence of malaria in the East African highlands.” Nature 415(6874): 905–909. Hendrickx, G., A. Napala, D. Rogers, P. Bastiaensen and J. Slingenbergh (1999). “Can remotely sensed meteorological data significantly contribute to reduce costs of tsetse surveys?” Memórias do Instituto Oswaldo Cruz 94(2): 273–276. Hendrickx, G., A. Napala, J. H. Slingenbergh, R. De Deken, J. Vercruysse and D. J. Rogers (2000). “The spatial pattern of trypanosomosis prevalence predicted with the aid of satellite imagery.” Parasitology 120(2): 121–134. Hu, W., N. Nicholls, M. Lindsay, P. Dale, A. J. McMichael, J. S. Mackenzie and S. Tong (2004). “Development of a predictive model for ross river virus disease in Brisbane, Australia.” American Journal of Tropical Medicine and Hygiene 71(2): 129–137. Hunter, J. M. (1997). “Geographical patterns of guinea worm infestation in Ghana: an historical contribution.” Social Science and Medicine 44(1): 103–122. Huq, A., R. B. Sack, A. Nizam, I. M. Longini, G. B. Nair, A. Ali, J. G. J. Morris, M. N. Khan, A. K. Siddique, M. Yunus, M. J. Albert, D. A. Sack and R. R. Colwell (2005). “Critical factors influencing the occurrence of Vibrio cholerae in the environment of Bangladesh.” Applied Environmental Microbiology 71(8): 4645–4654. Hurtado-Diaz, M., H. Riojas-Rodriguez, S. Rothenberg, H. Gomez-Dantes and E. Cifuentes (2006). “Impact of climate variability on the incidence of dengue in Mexico.” Epidemiology 17(6): S207–S207. Ijaz, M. K., S. Alharbi, S. A. Uduman, Y. Cheema, M. M. Sheek-Hussen, A. R. Alkhair, A. G. Shalabi, S. S. Ijaz, S. A. Bin-Othman, S. A. Sattar and L. F. Liddle (1994). “Seasonality and prevalence of rotavirus in Al-Ain, United Arab Emirates.” Clinical and Diagnostic Virology 2(6): 323–329. Indeje, M., M. N. Ward, L. J. Ogallo, G. Davies, M. Dilley and A. Anyamba (2006). “Predictabiity of the Normalized Difference Vegetation Index.” Journal of Climate 19(9): 1673–1687. IPCC (Intergovernmental Panel on Climate Change) (2001). IPCC Third Assessment Report – Climate Change 2001: Working Groups II: impacts, adaptation, and vulnerability. J. J. McCarthy, O. F. Canziani, N. A. Leary, D. J. Dokken and K. S. White (eds). Geneva, World Meteorological Office (WMO) and United Nations Environment Programme (UNEP). IRI (2005) Sustainable development in Africa: is the climate right? New York, The International Research Institute for Climate Prediction, Colombia University. IRI (2006) A Gap Analysis for the implementation of the Global Climate Observing System Programme in Africa. New York, The International Research Institute for Climate and Society, Colombia University. Julvez, J., J. Mouchet, A. Michault, A. Fouta and M. Hamidine (1997). “The progress of malaria in sahelian eastern Niger. An ecological disaster zone.” Bulletin de la Societe Pathologique Exotica 90(2): 101–104. Kabatereine, N. B., S. Brooker, E. M. Tukahebwa, F. Kazibwe and A. W. Onapa (2004). “Epidemiology and geography of Schistosoma mansoni in Uganda: implications for planning control.” Tropical Medicine and International Health 9(3): 372–380. Kane, E. M., R. M. Turcios, M. L. Arvay, S. Garcia, J. S. Bresee and R. I. Glass (2004). “The epidemiology of rotavirus diarrhea in Latin America. Anticipating rotavirus vaccines.” Revista Panamericana de Salud Publica 16(6): 371–377.
62
L. Kelly-Hope and M.C. Thomson
Kanojia, P. C., P. S. Shetty and G. Geevarghese (2003). “A long-term study on vector abundance & seasonal prevalence in relation to the occurrence of Japanese encephalitis in Gorakhpur district, Uttar Pradesh.” Indian Journal of Medical Research 117: 104–110. Kelly-Hope, L. A., B. H. Kay, D. M. Purdie and G. M. Williams (2002). “The risk of Ross River and Barmah Forest virus disease in Queensland: implications for New Zealand.” Australia and New Zealand Journal of Public Health 26(1): 69–77. Kelly-Hope, L. A., D. M. Purdie and B. H. Kay (2004a). “Differences in climatic factors between Ross River virus disease outbreak and nonoutbreak years.” Journal of Medical Entomology 41(6): 1116–1122. Kelly-Hope, L. A., D. M. Purdie and B. H. Kay (2004b). “El Nino Southern Oscillation and Ross River virus outbreaks in Australia.” Vector Borne Zoonotic Disease 4(3): 210–213. Kelly-Hope, L. A., D. M. Purdie and B. H. Kay (2004c). “Ross River virus disease in Australia, 1886–1998, with analysis of risk factors associated with outbreaks.” Journal of Medical Entomology 41(2): 133–150. Kelly-Hope, L. A., P. Diggle, B. S. Rowlingson, J. O. Gyapong, D. Kyelem, M. Coleman, M. C. Thomson, V. Obsomer, S. W. Lindsay, J. Hemingway and D. Molyneux (2006). “Negative spatial association between lymphatic filariasis and malaria in West Africa.” Tropical Medicine and International Health 11(2): 1–7. Kilian, A. H., P. Langi, A. Talisuna and G. Kabagambe (1999). “Rainfall pattern, El Nino and malaria in Uganda.” Transactions of the Royal Society of Tropical Medicine and Hygiene 93(1): 22–23. King, R. J., D. H. Campbell-Lendrum and C. R. Davies (2004). “Predicting geographic variation in cutaneous leishmaniasis, Colombia.” Emerging Infectious Diseases 10(4): 598–607. Kleinschmidt, I., J. Omumbo, O. Briet, N. van de Giesen, N. Sogoba, N. K. Mensah, P. Windmeijer, M. Moussa and T. Teuscher (2001). “An empirical malaria distribution map for West Africa.” Tropical Medicine and International Health 6(10): 779–786. Koelle, K., M. Pascual and M. Yunus (2005a). “Pathogen adaptation to seasonal forcing and climate change.” Proceedings of Biological Sciences 272(1566): 971–977. Koelle, K., X. Rodo, M. Pascual, M. Yunus and G. Mostafa (2005b). “Refractory periods and climate forcing in cholera dynamics.” Nature 436(7051): 696–700. Kondo, H., N. Seo, T. Yasuda, M. Hasizume, Y. Koido, N. Ninomiya and Y. Yamamoto (2002). “Postflood–infectious diseases in Mozambique.” Prehospital Disaster Medecine 17(3): 126–133. Konno, T., H. Suzuki, N. Katsushima, A. Imai, F. Tazawa, T. Kutsuzawa, S. Kitaoka, M. Sakamoto, N. Yazaki and N. Ishida (1983). “Influence of temperature and relative humidity on human rotavirus infection in Japan.” Journal of Infectious Disease 147(1): 125–128. Koopman, J. S., D. R. Prevots, M. A. Vaca Marin, H. Gomez Dantes, M. L. Zarate Aquino, I. M. J. Longini and J. Sepulveda Amor (1991). “Determinants and predictors of dengue infection in Mexico.” American Journal of Epidemiology 133(11): 1168–1178. Koram, K. A., S. Owusu-Agyei, D. J. Fryauff, F. Anto, F. Atuguba, A. Hodgson, S. L. Hoffman and F. K. Nkrumah (2003). “Seasonal profiles of malaria infection, anaemia, and bednet use among age groups and communities in northern Ghana.” Tropical Medicine and International Health 8(9): 793–802. Kovats, R. S. (2000). “El Nino and human health.” Bulletin of the World Health Organization 78(9): 1127–1135. Kovats, R. S., M. J. Bouma and A. Haines (1999). El Nino and health. Geneva, World Health Organization: WHO/SDE/PHE/99.4 46. Kovats, R. S., M. J. Bouma, S. Hajat, E. Worrall and A. Haines (2003). “El Nino and health.” The Lancet 362(9394): 1481–1489. Kristensen, T. K., J. B. Malone and J. C. McCarroll (2001). “Use of satellite remote sensing and geographic information systems to model the distribution and abundance of snail intermediate hosts in Africa: a preliminary model for Biomphalaria pfeifferi in Ethiopia.” Acta Tropica 79(1): 73–78. Kuhn, K., D. Campbell-Lendrum, A. Haines and J. Cox (2005). Using climate to predict infectious disease epidemics. Geneva, World Health Organization: 54.
3 Climate and Infectious Diseases
63
Lama, J. R., C. R. Seas, R. Leon-Barua, E. Gotuzzo and R. B. Sack (2004). “Environmental temperature, cholera, and acute diarrhoea in adults in Lima, Peru.” Journal of Health, Population and Nutrition 22(4): 399–403. Lapeyssonnie, L. (1963). “Cerebrospinal meningitis in Africa.” Bulletin of the World Health Organization 28(Suppl): 1–114. Le Sueur, D., F. Binka, C. Lengeler, D. De Savigny, B. Snow, T. Teuscher and Y. Toure (1997). “An atlas of malaria in Africa.” African Health 19(2): 23–24. Leonardo, L. R., P. T. Rivera, B. A. Crisostomo, J. N. Sarol, N. C. Bantayan, W. U. Tiu and N. R. Bergquist (2005). “A study of the environmental determinants of malaria and schistosomiasis in the Philippines using Remote Sensing and Geographic Information Systems.” Parassitologia 47(1): 105–114. Li, C. F., T. W. Lim, L. L. Han and R. Fang (1985). “Rainfall, abundance of Aedes aegypti and dengue infection in Selangor, Malaysia.” Southeast Asian Journal of Tropical Medicine and Public Health 16(4): 560–568. Lindblade, K. A., E. D. Walker, A. W. Onapa, J. Katungu and M. L. Wilson (1999). “Highland malaria in Uganda: prospective analysis of an epidemic associated with El Nino.” Transactions of the Royal Society of Tropical Medicine and Hygiene 93(5): 480–487. Lindblade, K. A., E. D. Walker, A. W. Onapa, J. Katungu and M. L. Wilson (2000). “Land use change alters malaria transmission parameters by modifying temperature in a highland area of Uganda.” Tropical Medicine and International Health 5(4): 263–274. Lindsay, S. W. and C. J. Thomas (2000). “Mapping and estimating the population at risk from lymphatic filariasis in Africa.” Transactions of the Royal Society of Tropical Medicine and Hygiene 94(1): 37–45. Lindsay, S. W., R. Bodker, R. Malima, H. A. Msangeni and W. Kisinza (2000). “Effect of 1997–98 El Nino on highland malaria in Tanzania.” The Lancet 355(9208): 989–990. Linthicum, K. J., F. G. Davies, A. Kairo and C. L. Bailey (1985). “Rift Valley fever virus (family Bunyaviridae, genus Phlebovirus). Isolations from Diptera collected during an inter-epizootic period in Kenya.” Journal of Hygiene (Lond) 95(1): 197–209. Linthicum, K. J., C. L. Bailey, F. G. Davies and C. J. Tucker (1987). “Detection of Rift Valley fever viral activity in Kenya by satellite remote sensing imagery.” Science 235(4796): 1656–1659. Linthicum, K. J., A. Anyamba, C. J. Tucker, P. W. Kelley, M. F. Myers and C. J. Peters (1999). “Climate and satellite indicators to forecast Rift Valley fever epidemics in Kenya.” Science 285(5426): 397–400. Lobitz, B., L. Beck, A. Huq, B. Wood, G. Fuchs, A. S. Faruque,. and R. Colwell (2000). “Climate and infectious disease: use of remote sensing for detection of Vibrio cholerae by indirect measurement.” Proceedings of the National Accademy of Sciences of the USA 97(4): 1438–1443. Loevinsohn, M. E. (1994). “Climatic warming and increased malaria incidence in Rwanda.” The Lancet 343(8899): 714–718. Mabaso, M. L., M. Craig, P. Vounatsou and T. Smith (2005). “Towards empirical description of malaria seasonality in southern Africa: the example of Zimbabwe.” Tropical Medicine and International Health 10(9): 909–918. Mabaso, M. L. H., M. Craig, A. Ross and T. Smith (2007a). “Environmental predictors of the seasonality of malaria transmission in Africa: The challenge.” American Journal of Tropical Medicine and Hygiene 76(1): 33–38. Mabaso, M. L. H., I. Kleinschmidt, B. Sharp and T. Smith (2007b). “El Nino Southern Oscillation (ENSO) and annual malaria incidence in Southern Africa.” Transactions of the Royal Society of Tropical Medicine and Hygiene 101(4): 326–330. Macfarlane, S.B., M. C. Thomson and C. L. AbouZahr (2007). Millennium: invest in country statistical systems Nature 446, 974. Maelzer, D., S. Hales, P. Weinstein, M. Zalucki and A. Woodward (1999). “El Nino and arboviral disease prediction.” Environmental Health Perspectives 107(10): 817–818. Malone, J. B., O. K. Huh, D. P. Fehler, P. A. Wilson, D. E. Wilensky, R. A. Holmes and A. I. Elmagdoub (1994). “Temperature data from satellite imagery and the distribution of schistosomiasis in Egypt.” American Journal of Tropical Medicine and Hygiene 50(6): 714–722.
64
L. Kelly-Hope and M.C. Thomson
MARA/ARMA (1998) Mapping Malaria Risk in Africa/Atlas du Risque de la Malaria en Afrique. Towards an atlas of malaria risk in Africa. First technical report of the MARA/ARMA collaboration. Durban, South Africa: MARA/ARMA, 1998. (http://www.mara.org.za/). Martins, L. M., J. M. Rebelo, M. C. dos Santos, J. M. Costa, A. R. da Silva and L. A. Ferreira (2004). “Eco-epidemiology of cutaneous leishmaniasis in Buriticupu, Amazon region of Maranhao State, Brazil, 1996–1998.” Cadernos de Saude Publica 20(3): 735–743. Mathur, K. K., G. Harpalani, N. L. Kalra, G. G. Murthy and M. V. Narasimham (1992). “Epidemic of malaria in Barmer district (Thar desert) of Rajasthan during 1990.” Indian Journal of Malariology 29(1): 1–10. McLigeyo, S. O. (1998). “Recent infectious disease outbreaks in Kenya: have we been caught unaware?” East African Medical Journal 75(2): 61–62. McMichael, A. J., D. H. Campbell-Lendrum, C. F. Corvalan,., K. L. Ebi, A. Githeko, J. D. Scheraga and A. Woodward (2003). Climate change and human health – risks and responses. Geneva, World Health Organization. Molesworth, A. M., M. H. Djingary and M. C. Thomson (2001). “Seasonality of meningococcal disease in Niger, West Africa: a preliminary investigation.” GEOMED ‘99: 92–97. Molesworth, A. M., L. E. Cuevas, S. J. Connor, A. P. Morse and M. C. Thomson (2003). “Environmental risk and meningitis epidemics in Africa.” Emerging Infectious Diseases 9(10): 1287–1293. Molyneux, D. H. (2004). ““Neglected” diseases but unrecognised successes–challenges and opportunities for infectious disease control.” The Lancet 364(9431): 380–383. Molyneux, D. H., P. J. Hotez and A. Fenwick (2005). “Rapid-impact interventions”: how a policy of integrated control for Africa’s neglected tropical diseases could benefit the poor.” PLoS Med 2(11): e336. Moreira, C. J. E. (1986). “Rainfall and flooding in the Guayas river basin and its effects on the incidence of malaria 1982–1985.” Disasters 10(2): 107–111. Mpabalwani, M., H. Oshitani, F. Kasolo, K. Mizuta, N. Luo, N. Matsubayashi, G. Bhat, H. Suzuki and Y. Numazaki (1995). “Rotavirus gastro-enteritis in hospitalized children with acute diarrhoea in Zambia.” Annals of Tropical Paediatrics 15(1): 39–43. Mulic, R. and D. Ropac (2002). “Epidemiologic characteristics and military implications of hemorrhagic fever with renal syndrome in Croatia.” Croatian Medical Journal 43(5): 581–586. Mulic, R., D. Ropac, Z. Gizdic and N. Sikic (2003). “[What is new in the epidemiologic characteristics of hemorrhagic fever with renal syndrome in Croatia?].” Acta Medica Croatica 57(5): 399–405. Munyekenye, O. G., A. K. Githeko, G. Zhou, E. Mushinzimana, N. Minakawa and G. Yan (2005). “Plasmodium falciparum spatial analysis, western Kenya highlands.” Emerging Infectious Diseases 11(10): 1571–1577. Mutanda, L. N., S. N. Kinoti, W. Gemert and E. O. Lichenga (1984). “Age distribution and seasonal pattern of rotavirus infection in children in Kenya.” Journal of Diarrhoeal Disease Research 2(3): 147–150. Nacher, M., B. Carme, D. Sainte Marie, P. Couppie, E. Clyti, P. Guibert and R. Pradinaud (2001). “Seasonal fluctuations of incubation, healing delays, and clinical presentation of cutaneous leishmaniasis in French Guiana.” Journal of Parasitology 87(6): 1495–1498. Nacher, M., P. Couppie, B. Carme, E. Clyti, D. Sainte Marie, P. Guibert and R. Pradinaud (2002). “Influence of meteorological parameters on the clinical presentation of cutaneous leishmaniasis in French Guiana and on the efficacy of pentamidine treatment of the disease.” Annals of Tropical Medicine and Parasitology 96(8): 773–780. Nacher, M., V. I. Carrara, E. Ashley, R. McGready, R. Hutagalung, J. V. Nguen, K. L. Thwai, S. Looareesuwan and F. Nosten (2004a). “Seasonal variation in hyperparasitaemia and gametocyte carriage in patients with Plasmodium falciparum malaria on the Thai-Burmese border.” Transactions of the Royal Society of Tropical Medicine and Hygiene 98(5): 322–328. Nacher, M., V. I. Carrara, R. McGready, E. Ashley, J. V. Nguen, K. L. Thwai, S. Looareesuwan and F. Nosten (2004b). “Seasonal fluctuations in the carriage of Plasmodium vivax gametocytes in Thailand.” Annals of Tropical Medicine and Parasitology 98(2): 115–120.
3 Climate and Infectious Diseases
65
Nakhapakorn, K. and N. K. Tripathi (2005). “An information value based analysis of physical and climatic factors affecting dengue fever and dengue haemorrhagic fever incidence.” International Journal of Health Geographics 4: 13. Ndiaye, O., J. Y. Hesran, J. F. Etard, A. Diallo, F. Simondon, M. N. Ward and V. Robert (2001). “Climate variability and number of deaths attributable to malaria in the Niakhar area, Senegal, from 1984 to 1996.” Sante 11(1): 25–33. Neronov, V. V. and S. M. Malkhazova (1999). “Relationship between zoonotic cutaneous leishmaniasis morbidity in the Murgab oasis and hydrometeorological factors.” Meditsinskaia parazitologiia i parazitarnye bolezni October–December (4): 22–26. Nihei, N., Y. Hashida, M. Kobayashi and A. Ishii (2002). “Analysis of malaria endemic areas on the Indochina Peninsula using remote sensing.” Japanese Journal of Infectious Disease 55(5): 160–166. Nurgaleeva, R. G., N. M. Mustafin, A. G. Stepanenko and T. I. Savinova (1988). “Effect of weather conditions in an epidemic season on the nature of the morbidity of hemorrhagic fever with renal syndrome.” Zhurnal Mikrobiologii, Epidemiologii, Immunobiologii (3): 40–44. Nwoke, B. E., C. O. Onwuliri and G. O. Ufomadu (1992). “Onchocerciasis in Plateau State; Nigeria: ecological background, local disease perception & treatment; and vector/parasite dynamics.” Journal of Hygiene, Epidemiology, Microbiology and Immunology 36(2): 153–160. Odiit, M., P. R. Bessell, E. M. Fevre, T. Robinson, J. Kinoti, P. G. Coleman, S. C. Welburn, J. McDermott and M. E. Woolhouse (2005). “Using remote sensing and geographic information systems to identify villages at high risk for rhodesiense sleeping sickness in Uganda.” Transactions of the Royal Society of Tropical Medicine and Hygiene 100(4): 354–362. Odongo-Aginya, E., G. Ssegwanyi, P. Kategere and P. C. Vuzi (2005). “Relationship between malaria infection intensity and rainfall pattern in Entebbe peninsula, Uganda.” African Health Sciences 5(3): 238–245. Okuno, T., P. T. Tseng, S. T. Hsu, C. T. Huang and C. C. Kuo (1975). “Japanese encephalitis surveillance in China (Province of Taiwan) during 1968–1971. I. Geographical and seasonal features of case outbreaks.” Japanese Journal of Medical Science and Biology 28(5–6): 235–253. Olaleye, O. D., O. Tomori, M. A. Ladipo and H. Schmitz (1996). “Rift Valley fever in Nigeria: infections in humans.” Revue Scientifique et Technique 15(3): 923–935. Omumbo, J. A., S. I. Hay, C. A. Guerra and R. W. Snow (2004). “The relationship between the Plasmodium falciparum parasite ratio in childhood and climate estimates of malaria transmission in Kenya.” Malaria Journal 3: 17. Omumbo, J. A., C. A. Guerra, S. I. Hay and R. W. Snow (2005a). “The influence of urbanisation on measures of Plasmodium falciparum infection prevalence in East Africa.” Acta Tropica 93(1): 11–21. Omumbo, J. A., S. I. Hay, R. W. Snow, A. J. Tatem and D. J. Rogers (2005b). “Modeling malaria risk in East Africa at high-spatial resolution.” Tropical Medicine and International Health 10(6): 557–566. Organization, P. A. H. (1998). Health Impacts of the Southern Oscillation (El Nino). PAHO, Washington DC. Pascual, M., X. Rodo, S. P. Ellner, R. Colwell and M. J. Bouma (2000). “Cholera dynamics and El Nino-Southern Oscillation.” Science 289(5485): 1766–1769. Pascual, M., J. A. Ahumada, L. F. Chaves, X. Rodó X, and M. Bouma (2006) Malaria resurgence in the East African highlands: Temperature trends revisited. Proceedings of the National Academy of Sciences USA 103: 5829–5834. Patz, J., M. Hulme, C. Rosenzweig, T. Mitchell, R. Goldberg, A. Githeko, S. Lele, A. McMichael and D. Le Sueur (2002). “Climate change: Regional warming and malaria resurgence.” Nature 415(6874): 905–909. Patz, J. A., D. Campbell-Lendrum, T. Holloway and J. A. Foley (2005). “Impact of regional climate change on human health.” Nature 438(7066): 310–317.
66
L. Kelly-Hope and M.C. Thomson
Peterson, A. T., C. Martinez-Campos, Y. Nakazawa and E. Martinez-Meyer (2005). “Time-specific ecological niche modeling predicts spatial dynamics of vector insects and human dengue cases.” Transactions of the Royal Society of Tropical Medicine and Hygiene 99(9): 647–655. Phukan, A. C., D. K. Patgiri,. and J. Mahanta (2003). “Rotavirus associated acute diarrhoea in hospitalized children in Dibrugarh, north-east India.” Indian Journal of Pathology and Microbiology 46(2): 274–278. Phukan, A. C., P. K. Borah and J. Mahanta (2004). “Japanese encephalitis in Assam, northeast India.” Southeast Asian Journal of Tropical Medicine and Public Health 35(3): 618–622. Poveda, G., W. Rojas, M. L. Quinones, I. D. Velez, R. I. Mantilla, D. Ruiz, J. S. Zuluaga and G. L. Rua (2001). “Coupling between annual and ENSO timescales in the malaria-climate association in Colombia.” Environmental Health Perspectives 109(5): 489–493. Ram, S., S. Khurana, S. B. Khurana, S. Sharma, D. V. Vadehra and S. Broor (1990). “Bioecological factors and rotavirus diarrhoea.” Indian Journal of Medical Research 91: 167–170. Ramasamy, R., M. S. Ramasamy, D. A. Wijesundera, A. P. Wijesundera, I. Dewit, C. Ranasinghe, K. A. Srikrishnaraj and C. Wickremaratne (1992). “High seasonal malaria transmission rates in the intermediate rainfall zone of Sri Lanka.” Annals of Tropical Medicine and Parasitology 86(6): 591–600. Raso, G., B. Matthys, E. K. N’Goran, M. Tanner, P. Vounatsou and J. Utzinger (2005). “Spatial risk prediction and mapping of Schistosoma mansoni infections among schoolchildren living in western Cote d’Ivoire.” Parasitology 131(Pt 1): 97–108. Rodo, X., M. Pascual, G. Fuchs and A. S. Faruque (2002). “ENSO and cholera: a nonstationary link related to climate change?” Proceedings of the National Accademy of Sciences of the USA 99(20): 12901–12906. Rogers, D. (1991). “Satellite imagery, tsetse and trypanosomiasis in Africa.” Preventive Veterinary Medicine 11(3–4): 210–220. Rogers, D. J. (2000). “Satellites, space, time and the African trypanosomiases.” Advances in Parasitology 47: 129–171. Rogers, D. J. and B. G. Williams (1993). “Monitoring trypanosomiasis in space and time.” Parasitology 106(Suppl): S77–92. Ropelewski, C. F. and M. S. Halpert (1987). “Global and regional scale precipitation patterns associated with the El Niño/Southern Oscillation.” Monthly Weather Review 115: 1606–1626. Rosenberg, R., R. G. Andre and S. Ketrangsee (1990). “Seasonal fluctuation of Plasmodium falciparum gametocytaemia.” Transactions of the Royal Society of Tropical Medicine and Hygiene 84(1): 29–33. Sachs, J. D. (2004). “Health in the developing world: achieving the Millennium Development Goals.” Bulletin of the World Health Organization 82(12): 947–949; discussion 950–2. Sachs, J. D. and J. W. McArthur (2005). “The Millennium Project: a plan for meeting the Millennium Development Goals.” The Lancet 365(9456): 347–353. Salim, A. R. and H. A. Sheikh (1975). “Trachoma in the Sudan. An epidemiological study.” British Journal of Ophthalmology 59(10): 600–604. Sarkies, J. W. (1967). “Dust and the incidence of severe trachoma.” British Journal of Ophthalmology 51(2): 97–100. Savory, E. C., L. E. Cuevas, M. A. Yassin, C. A. Hart, A. M. Molesworth and M. C. Thomson, M.C. (2006) Evaluation of the meningitis epidemics risk model in Africa. Epidemiology and Infection 14: 1–13. Schemann, J. F., D. Sacko, D. Malvy, G. Momo, L. Traore, O. Bore, S. Coulibaly and A. Banou (2002). “Risk factors for trachoma in Mali.” International Journal of Epidemiology 31(1): 194–201. Schreiber, K. V. (2001). “An investigation of relationships between climate and dengue using a water budgeting technique.” International Journal of Biometeorology 45(2): 81–89. Schwab, L., R. J. Whitfield, D. Ross-Degnan, P. Steinkuller and J. Swartwood (1995). “The epidemiology of trachoma in rural Kenya. Variation in prevalence with lifestyle and environment. Study Survey Group.” Ophthalmology 102(3): 475–482.
3 Climate and Infectious Diseases
67
Shaman, J., J. F. Day and M. Stieglitz (2004). “The spatial-temporal distribution of drought, wetting, and human cases of St. Louis encephalitis in southcentral Florida.” American Journal of Tropical Medicine and Hygiene 71(3): 251–261. Shanks, G. D., S. I. Hay, D. I. Stern, K. Biomndo and R. W. Snow (2002). “Meteorologic influences on Plasmodium falciparum malaria in the Highland Tea Estates of Kericho, Western Kenya.” Emerging Infectious Diseases 8(12): 1404–1408. Singh, N. and V. P. Sharma (2002). “Patterns of rainfall and malaria in Madhya Pradesh, central India.” Annals of Tropical Medicine and Parasitology 96(4): 349–359. Siziya, S., T. E. Watts and P. R. Mason (1997). “Malaria in Zimbabwe: comparisons of IFAT levels, parasite and spleen rates among high, medium and lower altitude areas and between dry and rainy seasons.” Central African Medical Journal 43(9): 251–254. Skalova, R. (1984). “Epidemiology of meningococcal meningitis in Sahel and Mongolia.” Infection 12(Suppl 1): S35–43. Small, J., S. J. Goetz and S. I. Hay (2003). “Climatic suitability for malaria transmission in Africa, 1911–1995.” Proceedings of the National Accademy of Sciences of the USA 100(26): 15341–15345. Snow, R. W., M. H. Craig, U. Deichmann and D. le Sueur (1999). “A preliminary continental risk map for malaria mortality among African children.” Parasitology Today 15(3): 99–104. Speelmon, E. C., W. Checkley, R. H. Gilman, J. Patz, M. Calderon and S. Manga (2000). “Cholera incidence and El Nino-related higher ambient temperature.” JAMA 283(23): 3072–3074. Srivastava, A., B. N. Nagpal, R. Saxena and V. P. Sharma (1999). “Geographic information system as a tool to study malaria receptivity in Nadiad Taluka, Kheda district, Gujarat, India.” Southeast Asian Journal of Tropical Medicine and Public Health 30(4): 650–656. Steele, A. D., J. J. Alexander and I. T. Hay (1986). “Rotavirus-associated gastroenteritis in black infants in South Africa.” Journal of Clinical Microbiology 23(5): 992–994. Steib, K. and P. Mayer (1988). “Epidemiology and vectors of Dracunculus medinensis in northwest Burkina Faso, West Africa.” Annals of Tropical Medicine and Parasitology 82(2): 189–199. Stensgaard, A., A. Jorgensen, N. B. Kabatereine, J. B. Malone and T. K. Kristensen (2005). “Modeling the distribution of Schistosoma mansoni and host snails in Uganda using satellite sensor data and Geographical Information Systems.” Parassitologia 47(1): 115–125. Stockdale, T. N., D. L. T. Anderson, J. O. S. Alves and M. A. Balmaseda (1998). “Global seasonal rainfall forecasts using a coupled ocean-atmosphere model.” Nature 392(6674): 370–373. Strickman, D. and P. Kittayapong (2002). “Dengue and its vectors in Thailand: introduction to the study and seasonal distribution of Aedes larvae.” American Journal of Tropical Medicine and Hygiene 67(3): 247–259. Sultan, B., K. Labadi, J. F. Guegan and S. Janicot (2005). “Climate drives the meningitis epidemics onset in West Africa.” PLoS Medecine 2(1): e6. Suwannee, A., K. Kaew Nualchawee, S. Murai, A. Eiumnoh and K. Honda (1997). Application of remote sensing and geographic information system for vector-borne disease in humans through rice agro-ecosystem. Asian Conference on Remote Sensing (cited In WHO publication: Using climate to predict infectious disease epidemics. 2005), Malaysia. Swaroop, S. (1946). “Forecasting of epidemic malaria in the Punjab, India.” American Journal of Tropical Medicine 29: 1–17. Tedesco, L. R. (1980). “Trachoma and environment in the Northern Territory of Australia.” Social Science and Medicine (Geogr) 14(2): 111–117. Teklehaimanot, H. D., M. Lipsitch, A. Teklehaimanot and J. Schwartz (2004a). “Weather-based prediction of Plasmodium falciparum malaria in epidemic-prone regions of Ethiopia I. Patterns of lagged weather effects reflect biological mechanisms.” Malaria Journal 3(1): 41. Teklehaimanot, H. D., J. Schwartz, A. Teklehaimanot and M. Lipsitch (2004b). “Weather-based prediction of Plasmodium falciparum malaria in epidemic-prone regions of Ethiopia II. Weather-based prediction systems perform comparably to early detection systems in identifying times for interventions.” Malaria Journal 3(1): 44.
68
L. Kelly-Hope and M.C. Thomson
Thammapalo, S., V. Chongsuwiwatwong, D. McNeil and A. Geater (2005). “The climatic factors influencing the occurrence of dengue hemorrhagic fever in Thailand.” Southeast Asian Journal of Tropical Medicine and Public Health 36(1): 191–196. Theander, T. G. (1998). “Unstable malaria in Sudan: the influence of the dry season. Malaria in areas of unstable and seasonal transmission. Lessons from Daraweesh.” Transactions of the Royal Society of Tropical Medicine and Hygiene 92(6): 589–592. Thiongane, Y. and V. Martin (2003). “Systeme sous regional d’alerte et de controle de la Fievre de la Valley du Rift (FVR) en Afrique de l’Oueste.” EMPRES/FAO Bulletin: 7. Thompson, D. F., J. B. Malone, M. Harb, R. Faris, O. K. Huh, A. A. Buck and B. L. Cline (1996). “Bancroftian filariasis distribution and diurnal temperature differences in the southern Nile delta.” Emerging Infectious Diseases 2(3): 234–235. Thompson, R. A., J. Wellington de Oliveira Lima, J. H. Maguire, D. H. Braud and D. T. Scholl (2002). “Climatic and demographic determinants of American visceral leishmaniasis in northeastern Brazil using remote sensing technology for environmental categorization of rain and region influences on leishmaniasis.” American Journal of Tropical Medicine and Hygiene 67(6): 648–655. Thomson, M. C., J. B. Davies, R. J. Post, M. J. Bockarie, P. A. Beech Garwood and J. Kandeh (1996). “The unusual occurrence of savanna members of the Simulium damnosum species complex (Diptera: Simuliidae) in southern Sierra Leone in 1988.” Bulletin of Entomological Research 86(3): 271–280. Thomson, M.C., D. Elnaiem, S. J. Connor and R. W. Ashford (1999). Towards a kala azar risk map of Sudan: mapping the potential distribution of Phlebotomus orientalis using digital data of environmental variables. Tropical Medicine and International Health. 4(2): 105–113. Thomson, M. C., K. Abayonmi, A. Barnston, M. Levy, and M. Dilley (2003) El Niño and drought in southern Africa. The Lancet 361: 437–438. Thomson, M. C., V. Obsomer, J. Kamgno, J. Gardon, S. Wanji, I. Takougang, P. Enyong, J. H. Remme, D. H. Molyneux and M. Boussinesq (2004a). “Mapping the distribution of Loa loa in Cameroon in support of the African Programme for Onchocerciasis Control.” Filaria Journal 3(1): 7. Thomson, M. C., P. J. Erickson, A. Ben Mohamed and S. J. Connor (2004b). Chapter 13: Land use change and infectious disease in West Africa. Ecosystems and Land Use Change R. DeFries, G. Asner and R. Houghton (eds). Washington, DC., American Geophysical Union. Thomson, M. C., S. J. Mason, T. Phindela and S. J. Connor (2005). “Use of rainfall and sea surface temperature monitoring for malaria early warning in Botswana.” American Journal of Tropical Medicine and Hygiene 73(1): 214–221. Thomson, M. C., F. J. Doblas-Reyes, S. J. Mason, R. Hagedorn, S. J. Connor, T. Phindela, A. P. Morse and T. N. Palmer (2006a). “Malaria early warnings based on seasonal climate forecasts from multi-model ensembles.” Nature 439: 576–579. Thomson, M. C., A. M. Molesworth, M. H. Djingarey, K. R. Yameogo, F. Belanger, L. E. Cuevas (2006b) Potential of environmental models to predict meningitis epidemics in Africa. Tropical Medicine & International Health 11(6): 773–780. Thonnon, J., M. Picquet, Y. Thiongane, M. Lo, R. Sylla and J. Vercruysse (1999). “Rift valley fever surveillance in the lower Senegal river basin: update 10 years after the epidemic.” Tropical Medicine and International Health 4(8): 580–585. Tong, S. and W. Hu (2001). “Climate variation and incidence of Ross river virus in Cairns, Australia: a time-series analysis.” Environmental Health Perspectives 109(12): 1271–1273. Tong, S., P. Bi, K. Parton, J. Hobbs and A. J. McMichael (1998). “Climate variability and transmission of epidemic polyarthritis.” The Lancet 351(9109): 1100. Tong, S., P. Bi, K. Donald and A. J. McMichael (2002). “Climate variability and Ross River virus transmission.” Journal of Epidemioly and Community Health 56(8): 617–621. Tong, S., W. Hu and A. J. McMichael (2004). “Climate variability and Ross River virus transmission in Townsville Region, Australia, 1985–1996.” Tropical Medicine and International Health 9(2): 298–304.
3 Climate and Infectious Diseases
69
Tong, S., W. Hu, N. Nicholls, P. Dale, J. S. MacKenzie, J. Patz and A. J. McMichael (2005). “Climatic, high tide and vector variables and the transmission of Ross River virus.” Internal Medecine Journal 35(11): 677–680. Traore-Lamizana, M., D. Fontenille, H. G. Zeller, M. Mondo, M. Diallo, F. Adam, M. Eyraud, A. Maiga and J. P. Digoutte (1996). “Surveillance for yellow fever virus in eastern Senegal during 1993.” Journal of Medical Entomology 33(5): 760–765. Trape, J. E. (1999). “Climate changes and infectious diseases: Malaria and tick-borne relapsing borreliosis.” Medecine et Maladies Infectieuses 29(5): 296–300. Travis, P., S. Bennett, A. Haines, T. Pang, Z. Bhutta, A. A. Hyder, N. R. Pielemeier, A. Mills and T. Evans (2004). “Overcoming health-systems constraints to achieve the Millennium Development Goals.” The Lancet 364(9437): 900–906. UN (United Nations) (2002). The Millennium Project. Investing in Development: A Practical Plan to Achieve the Millennium Development Goals. Uzun, S., C. Uslular, A. Yucel, M. A. Acar, M. Ozpoyraz and H. R. Memisoglu (1999). “Cutaneous leishmaniasis: evaluation of 3,074 cases in the Cukurova region of Turkey.” British Journal of Dermatology 140(2): 347–3450. Valencia Tellería, A. (1986). “Health consequences of floods in Bolivia in 1982.” Disasters 10(2): 88–106. Van Der Hoek, W., F. Konradsen, P. H. Amerasinghe, D. Perera, M. K. Piyaratne and F. P. Amerasinghe (2003). “Towards a risk map of malaria for Sri Lanka: the importance of house location relative to vector breeding sites.” International Journal of Epidemiology 32(2): 280–285. van der Kolk, M., A. E. Tebo, H. Nimpaye, D. N. Ndombol, R. W. Sauerwein and W. M. Eling (2003). “Transmission of Plasmodium falciparum in urban Yaounde, Cameroon, is seasonal and age-dependent.” Transactions of the Royal Society of Tropical Medicine and Hygiene 97(4): 375–379. Vasconcelos, P. F., Z. G. Costa, E. S. Travassos Da Rosa, E. Luna, S. G. Rodrigues, V. L. Barros, J. P. Dias, H. A. Monteiro, O. F. Oliva, H. B. Vasconcelos, R. C. Oliveira, M. R. Sousa, J. Barbosa Da Silva, A. C. Cruz, E. C. Martins and J. F. Travassos Da Rosa (2001). “Epidemic of jungle yellow fever in Brazil, 2000: implications of climatic alterations in disease spread.” Journal of Medical Virology 65(3): 598–604. Vasconcelos, P. F., S. G. Rodrigues, N. Degallier, M. A. Moraes, J. F. da Rosa, E. S. da Rosa, B. Mondet, V. L. Barros and A. P. da Rosa (1997). “An epidemic of sylvatic yellow fever in the southeast region of Maranhao State, Brazil, 1993–1994: epidemiologic and entomologic findings.” American Journal of Tropical Medicine and Hygiene 57(2): 132–137. Vittor, A. Y., R. H. Gilman, J. Tielsch, G. Glass, T. Shields, W. S. Lozano, V. Pinedo-Cancino and J. A. Patz (2006). “The effect of deforestation on the human-biting rate of Anopheles Darlingi, the primary vector of Falciparum Malaria in the Peruvian Amazon.” American Journal of Tropical Medicine and Hygiene 74(1): 3–11. Wanji, S., N. Tendongfor, M. Esum, S. N. Atanga and P. Enyong (2003). “Heterogeneity in the prevalence and intensity of loiasis in five contrasting bioecological zones in Cameroon.” Transactions of the Royal Society of Tropical Medicine and Hygiene 97(2): 183–7. Wanji, S., N. Tendongfor, M. Esum, S. S. Yundze, M. J. Taylor and P. Enyong (2005). “Combined utilisation of Rapid Assessment Procedures for Loiasis (RAPLOA) and Onchocerciasis (REA) in rain forest villages of Cameroon.” Filaria Journal 4(1): 2. Wellde, B. T., D. A. Chumo, M. J. Reardon, D. Waema, D. H. Smith, W. C. Gibson, L. Wanyama and T. A. Siongok (1989). “Epidemiology of Rhodesian sleeping sickness in the Lambwe Valley, Kenya.” Annals of Tropical Medicine and Parasitology 83(Suppl 1): 43–62. Werneck, G. L. and J. H. Maguire (2002). “Spatial modeling using mixed models: an ecologic study of visceral leishmaniasis in Teresina, Piaui State, Brazil.” Cad Saude Publica 18(3): 633–637. Whelan, P. I., S. P. Jacups, L. Melville, A. Broom, B. J. Currie, V. L. Krause, B. Brogan, F. Smith and P. Porigneaux (2003). “Rainfall and vector mosquito numbers as risk indicators for mosquito-borne disease in central Australia.” Communicable Disease Intellegence 27(1): 110–116.
70
L. Kelly-Hope and M.C. Thomson
Woodruff, R. E., C. S. Guest, M. G. Garner, N. Becker, J. Lindesay, T. Carvan and K. Ebi (2002). “Predicting Ross River virus epidemics from regional weather data.” Epidemiology 13(4): 384–393. WHO (World Health Organization) (1999). Infectious Disease Report. Geneva, WHO. WHO (World Health Organization) (2001). Macroeconomics and Health. Geneva, WHO. WHO (World Health Organization) (2004). Using climate to predict infectious disease outbreaks: a review. C. D. S. a. Response. Geneva, WHO. Wu, P. C., H. J. Su, C. Y. Lin, S. C. Lung, H. R. Guo and J. G. Lay (2006). “The effects of temperature and recovery of vector on the spatial distribution of dengue fever occurrences in Taiwan.” Epidemiology 17(6): S206–S206. Wyss, K. (2004). “An approach to classifying human resources constraints to attaining healthrelated Millennium Development Goals.” Human Resources and Health 2(1): 11. Yacob, K. B. M. and S. Swaroop (1945). “Investigation of the long-term periodicity in the incidence of epidemic malaria in the Punjab” Journal of the Malaria Institute India 6: 39–51. Yacob, K. B. M. and S. Swaroop (1946). “Malaria and rainfall in the Punjab” Journal of the Malaria Institute India 6: 273–284. Yang, G., X. Zhou, J. B. Malone, J. C. McCarroll, T. Wang, J. Liu, Q. Gao, X. Zhang, Q. Hong and L. Sun (2002). “GIS prediction model of malaria transmission in Jiangsu province.” Zhonghua Yu Fang Yi Xue Za Zhi 36(2): 103–105. Yang, G. J., P. Vounatsou, X. N. Zhou, M. Tanner and J. Utzinger (2005a). “A Bayesian-based approach for spatio-temporal modeling of county level prevalence of Schistosoma japonicum infection in Jiangsu province, China.” International Journal of Parasitology 35(2): 155–162. Yang, G. J., P. Vounatsou, X. N. Zhou, M. Tanner and J. Utzinger (2005b). “A potential impact of climate change and water resource development on the transmission of Schistosoma japonicum in China.” Parassitologia 47(1): 127–134. Yang, G. J., P. Vounatsou, X. N. Zhou, J. Utzinger and M. Tanner (2005c). “A review of geographic information system and remote sensing with applications to the epidemiology and control of schistosomiasis in China.” Acta Tropica 96(2–3): 117–129. Zhou, X. N., J. B. Malone, T. K. Kristensen and N. R. Bergquist (2001). “Application of geographic information systems and remote sensing to schistosomiasis control in China.” Acta Tropica 79(1): 97–106. Zhou, G., N. Minakawa, A. K. Githeko and G. Yan (2004). “Association between climate variability and malaria epidemics in the East African highlands.” Proceedings of the National Academy of Sciences of the USA 101(8): 2375–2380. Zhou, G., N. Minakawa, A. K. Githeko and G. Yan (2005). “Climate variability and malaria epidemics in the highlands of East Africa.” Trends in Parasitology 21(2): 54–56.
Chapter 4
Integration of Seasonal Forecasts into Early Warning Systems for Climate-Sensitive Diseases such as Malaria and Dengue Stephen J. Connor1, * and Gilma C. Mantilla2
Abstract Effective prevention and control of epidemics has been a key element of global, regional and national disease control policies for many years. Epidemics are by their nature abnormal events and will clearly challenge the normal routine approaches to control and provision of treatment. Epidemics are caused by unusual changes in the existing equilibrium between the human host, the pathogen and its vector. While the level of risk may be exacerbated by social factors, climate variability plays an important role and indeed it is most often abnormality in meteorological and environmental conditions that ‘triggers’ epidemics of the climate sensitive diseases. Malaria and dengue are considered climate-sensitive diseases and in recent years there have been attempts to develop and test integrated early warning systems which seek to provide advance warning of changes in epidemic risk, through incremental indicators that allow control services greater opportunity to plan, choose and implement more timely and focused response in the areas affected. This paper uses the example of malaria early warning system applications in Southern Africa to illustrate the elements of the system, evidence of its potential benefits, including the control options it may provide and some of the current challenges and opportunities for its broader implementation in Africa and elsewhere.
Keywords Malaria, dengue, seasonal forecast, early warning system, climate variability, climate change, global warming
1 International Research Institute for Climate and Society, The Earth Institute, Columbia University, New York, USA 2
National Institute of Health, Ministry of Social Protection, Bogota, Colombia
*Corresponding author: Tel: + 1 845 680 4458; fax: + 1 845 680 4864; e-mail: sjconnor@iri. columbia.edu
M.C. Thomson et al. (eds.), Seasonal Forecasts, Climatic Change and Human Health. © Springer Science + Business Media B.V. 2008
71
72
4.1
S.J. Connor and G.C. Mantilla
Climate and Health
Certain diseases and ill health are associated with particular environmental conditions, season, and climate. This was recognized by the ancient writers of Vedic literature, by Hippocrates, and remains the focus of considerable research today. Climate may impact on health through a number of mechanisms, directly through cold or heat stress or, more commonly, indirectly through: (a) Its role in determining agricultural output and consequently food security which directly affects nutritional status (b) Its role in the economy and income opportunities, which affects the ability to maintain nutritional status, prevent infection and obtain curative health care as necessary (c) Its role in determining seasonal and inter-annual demographic processes (e.g. seasonal labour migration and environmental refugees) (d) Its impact on the spatial and temporal distribution of climate-sensitive infectious diseases – commonly, but not always, transmitted by vectors (Thomson et al. 2004a, b; Kelly-Hope and Thomson, Chapter 3, this volume) The World Health Organization have recently identified a number of climate-sensitive infectious diseases, some of which including: malaria, cholera and dengue; they describe as being promising candidates for the development of climate informed early warning systems (WHO 2005a; Kelly-Hope and Thomson, Chapter 3, this volume).
4.2
Climate and Infectious Disease in Africa
There is currently intense concern over the growing public health problems posed by infectious diseases in the developing world, especially HIV-AIDS, TB and malaria. These three diseases compound each other: HIV increases susceptibility to TB and severe malaria; malaria and TB hasten the progression of HIV-AIDS (WHO 2002b, 2005b) making care of sufferers particularly challenging to resource-poor health services and communities in sub-Saharan Africa. Infectious diseases disproportionately affect the poorer countries and are seen as a significant constraint to economic development and progress towards the Millennium Development Goals, particularly in sub-Saharan Africa. Therefore calls for massive investment in health services and control programmes in the most affected countries have been made (WHO 2001a). The enormous impact and scale of infectious diseases in Africa often leads to their being described as ‘epidemic’. However this term may be inappropriate for a disease such as malaria which has been with us since time immemorial, and where the greatest burden continues to occur in regions where the disease has been continuously present in the community. For instance, malaria transmission throughout much of Africa has long been described as ‘stable’ or endemic transmission, meaning that it is expected to
4 Integration of Seasonal Forecasts into Early Warning Systems
73
affect certain population groups each year with a predictable, usually seasonal, distribution of infections throughout the year. The burden of endemic malaria within the African continent does however appear to have been growing over recent decades (Hay et al. 2004) and it is this worrying trend that often leads to the misuse of the term ‘epidemic’. The increase is largely attributed to demographic changes, increasing drug resistance, declining control infrastructure, the impact of co-infection with other diseases and poverty. Those most vulnerable to endemic malaria are young children who are yet to acquire immunity to the disease, pregnant women whose immunity is compromised during pregnancy, and non-immune migrants/travelers. It must be stressed that malaria is a preventable disease. In the early half of the last century, prior to its eradication, it was endemic throughout much of the United States and Europe (Najera 1999). Today endemic malaria like many infectious diseases persists in the tropics, hence the term ‘tropical diseases’. This is something of a misnomer as malaria transmission has occurred close to the Arctic Circle (Wernsdorfer 1980) and may be seen more as a disease of poverty than geography (Sachs and Malaney 2002). However, where malaria is not adequately controlled the distribution of the disease remains closely linked with the seasonality of the climate and the local environment which is highly favourable for transmission in many parts of tropical Africa. The specialized Anopheles mosquitoes which transmit this disease require specific habitats with surface water to lay their eggs, humidity for adult mosquito survival and the development rates of both the mosquito and parasite populations are influenced by temperature. Endemic areas are therefore regions that have an environment suitable for the interactions of the mosquito vectors, the parasites and the human host. Climate information (long-term mean climatology) has been used to map those regions of the African continent considered climatically suitable for endemic malaria transmission to occur. The most widely known of these maps are those produced by the Africa based MARA Collaboration (MARA 1998; Craig et al. 1999). Along the margins of endemic malaria there are geographic zones where malaria is ‘unstable’ because the full set of conditions required for transmission are only occasionally present. In these unstable, or epidemic, transmission zones the human populations are not frequently exposed to malaria and do not therefore acquire substantial protective immunity. In effect all age groups remain vulnerable to malaria as a life threatening disease, if and when an epidemic does occur. While epidemic malaria occurs along the geographic margins of endemic malaria its impact is far from marginal. It has recently been estimated that more than 124 million Africans live in zones of unstable or epidemic malaria where they suffer 12 million malaria episodes and up to 310,000 deaths annually (Worrall et al. 2004). Clearly control planning, surveillance and intervention needs to be appropriate to these two very different situations (Connor and Thomson 2005). In endemic regions malaria control can be based largely on routine measures, while in epidemic prone regions malaria control needs to be flexible and responsive to rapidly changing circumstances, hence the interest in climate informed malaria early warning systems (WHO 2004).
74
4.3
S.J. Connor and G.C. Mantilla
Identifying Epidemic Malaria
Given that the term epidemic is often used in a confusing manner it is important to try and establish what an appropriate definition is. Classically an epidemic of malaria is a sharp but temporary increase in disease incidence resulting from a disturbance in the pre-existing ecological and epidemiological equilibrium between the human, parasite and mosquito vector populations that inhabit a particular ecoepidemiological setting. However, the many and various determinants of epidemics make their definition complicated and attempts have been made to classify epidemics based on their main causative factor, be it natural or man-made. Epidemics due to natural causes such as abnormal meteorological conditions may be described as ‘classic’ or ‘true’ epidemics. Because classic epidemics occur in regions where the environmental conditions are marginal for vector and parasite development, such as desert-fringes (where it is usually too dry to support transmission) or highlandfringes (where it is usually too cool), they are often periodic, temporary in nature and a return to equilibrium is to be expected. Epidemics may also occur naturally as pronounced peaks in areas with normally low-intensity seasonal transmission. As is the case with classic epidemics these may also occur periodically, as environmental conditions allow, and a return to equilibrium should occur. Some epidemics can be described as man-made. These can be further subcategorized as to whether they occur as a result of a human induced change in the environment (such as an irrigation scheme in a warm-semi-arid environment); the displacement of people between highland and lowland regions; a breakdown in pre-existing levels of malaria control. Epidemics have also occurred, albeit rarely, following the accidental introduction of an exotic vector species into an area. Manmade epidemics are less-likely to be followed by a natural return to the pre-existing equilibrium and a transition from epidemic to endemic transmission patterns may become evident. See Fig. 4.1. Malaria epidemics can be localized events or may affect large populations throughout many countries within a region – often when there are large scale regional climate anomalies. For example: Southern Africa in 1996/1997, East Africa in 1997/1998 and the West African Sahel in 1999/2000. Because they are unusual events they often over-burden existing health services and consequently their non-immune victims suffer high case fatality rates, up to 10 times higher than in non-epidemic years, across all age ranges (Kiszewski and Teklehaimanot 2004). Although epidemics are unusual they are most likely to occur in specific areas that can be identified and mapped. Information on climate variability has been used to help map regions prone to malaria epidemics (Connor 2003; WHO-UNICEF 2003). Analysis of previous epidemics helps understand the nature of their causation and offers opportunities for predicting and monitoring changes in the eco-epidemiology of the disease (Najera et al. 1998; Connor et al. 1999; Cox et al. 1999). These basic activities form the fundamentals of epidemic early warning and, when combined with the essentials of a flexible epidemic control plan and implementation of timely and adequate control activities, provide a basis for epidemic prevention (WHO 2004a).
4 Integration of Seasonal Forecasts into Early Warning Systems
75
Fig. 4.1 A conceptual diagram of three different epidemic scenarios ●
●
●
The first (top) situation represents a ‘true’ epidemic, i.e. an infrequent event occurring in areas where the disease does not normally occur. This type of epidemic is often associated with warm arid and semi-arid regions. This type of epidemic may be cyclical in nature. The second (middle) situation represents an unusually high peak in transmission in areas where malaria normally presents seasonally. This type of epidemic is often associated with hypoendemic or mesoendemic settings such as the highland fringes. These epidemics may also be cyclical in nature. The third (bottom) situation represents a “resurgent out-break” where neglect or breakdown in control allows malaria to attempt to return to its higher ‘pre-control’ level of endemicity. This third type of epidemic may be associated with more complex emergency situations involving political instability and displaced populations (WHO 2002a).
4.4
Malaria Early Warning and Response Systems
Prevention and control of malaria epidemics has been prominent among the key technical elements of the current global malaria control strategy (WHO 1993a, b) and the Roll Back Malaria Partnership (WHO 1998, 2000, WHO-UNICEF 2003). Subsequently national and district malaria control programmes in regions affected have been encouraged to develop better means of predicting, preventing and controlling malaria epidemics. WHO and other partners have supported this initiative through the Roll Back Malaria Technical Support Network on Epidemic Prevention and Control and through the actions of a number of WHO-Inter-Country Programmes in partnership with national Ministries of Health in their member states. A working framework for the development of systems that enable epidemic forecasting, prevention, early detection and control has been developed (WHO 2001b) and its various components and processes tested in a number of countries (WHO 2004). Clearly the primary purpose of early warning is to enable earlier intervention and effective prevention and control of epidemics. Current thinking on the epidemic early warning and response system is based on a framework of five integrated components: (1) vulnerability assessment and
76
S.J. Connor and G.C. Mantilla
Fig. 4.2 A conceptual framework for an integrated epidemic prediction, preparedness, prevention and response system for climate-sensitive diseases
monitoring; (2) seasonal climate forecasting; (3) environmental monitoring; (4) sentinel case surveillance; and (5) planning, preparedness and response. This integrated framework (Connor 2005) is illustrated in Fig. 4.2 and described below.
4.4.1
Vulnerability Assessment and Monitoring
There are many factors that may increase the vulnerability of a population to malaria epidemics and increase the severity of disease outcome should a malaria epidemic occur. Co-infection with other diseases such as HIV-AIDS is a major consideration for countries in sub-Saharan Africa. Resistance to anti-malarial drugs has also become a recent problem throughout much of the region. Drought or floods, food insecurity, civil unrest and associated population movements between areas of differing endemicity combine to make certain populations more vulnerable to epidemics. Information from agencies outside of the health sector may be very useful for assessing and monitoring vulnerability. Intersectoral collaboration, with food security, drought monitoring, livestock and water resource management services, is encouraged. Vulnerability monitoring does not indicate when an epidemic will occur but gives an indication of changes taking place that raise the ‘epidemic potential’ and increase the risk and probability of more severe disease outcome should
4 Integration of Seasonal Forecasts into Early Warning Systems
77
an epidemic occur. This information can be used to stimulate the development of epidemic response plans, assessment of contingent control service capacity and identify roles and responsibilities of multi-sectoral response partners.
4.4.2
Seasonal Climate Forecasting
In recent years there have been significant scientific advances in our ability to predict climate on the seasonal timescale (Goddard and Mason 2002). Where epidemics result from unusual meteorological conditions, then some advance warning of the potential for these conditions could offer several months lead-time for preparation and control planning. The skill associated with the seasonal climate predictions varies from region to region, but is generally higher within the tropics. Climate forecasting inherently involves a level of uncertainty and control services will need guidance on the interpretation of these forecasts and the potential implications for malaria and its control in the coming season. Intersectoral collaboration with regional and national forecasting services is therefore strongly encouraged. The information available from seasonal climate forecasts may allow control services to prepare for preventative interventions, procure effective drug supplies, raise community awareness and education on personal protection prior to the onset of the main rainy season and 3–6 months prior to the peak malaria season.
4.4.3
Environmental/Climate Monitoring
The availability of routine information on environmental/climatic variables pertinent to malaria transmission, such as rainfall, temperatures, humidity, vegetation status (an indication of soil water availability) and flooding can provide control services with useful indicators of changes in epidemic risk. While the lead-times are shorter than those from seasonal forecasts (1–3 months) they are generally more certain as they are based on observations of the variables concerned. The two basic sources of such information are periodic summaries (usually estimated from meteorological satellites) available from specialist centres via the internet; or directly from national meteorological services’ ground-based weather station network. Generally, summary products are available free of charge, whereas the meteorological services may need to charge for station data and this may present difficulties for poorly resourced health services. Control services are encouraged to begin dialogue with their national meteorological services and discuss the specific information requirements and support they may need. The information provided by environmental monitoring allows control services to focus in on changes in conditions at a more localised scale where they may implement preventative control interventions, such as vector control, and encourage use of personal protection, including mosquito nets.
78
4.4.4
S.J. Connor and G.C. Mantilla
Sentinel Case Surveillance and Early Detection
The development of good sentinel case surveillance systems is a pre-requisite for an effective and responsive health service and in the case of epidemic malaria an essential means to detect any unusual increases in case numbers. There are a number of methods used to determine epidemic thresholds and control services are encouraged to explore these to determine which are most likely to be appropriate to their particular setting. While they may be able to detect the early phase of an epidemic, if used in isolation epidemic thresholds offer very short lead-times (1–3 weeks) for preventative control planning and implementation. Once an epidemic threshold has been passed control services need to consider carefully whether they should concentrate their resources primarily on ensuring drug distribution and effective case management, rather than embarking on vector control interventions which they may not be able to mobilise with sufficient time to be effective.
4.4.5
Planning, Preparedness and Response
This component is clearly the most important as it encapsulates the primary purpose of the system, earlier and effective response based on predetermined control planning and contingency measures. Control plans are developed based on vulnerability assessment and monitoring of epidemic prone regions and the communities at risk. Early warning and early detection indicators and their interpretation are agreed upon and the preventative control options they offer are identified and costed. Contingency measures and intersectoral response partners identified and clear directions of roles and responsibilities are agreed. The agencies responsible for control at the local level are clear about when they should call on additional resources from the national level and what level of resources they should expect. The control response options should be implemented on the basis of an agreed set of early warning/early detection indicators such that a sufficient level of confidence in the response activities is achieved for purposes of accountability. Ultimately the integrated epidemic early warning and response system is designed to produce incremental indicators of changes in epidemic potential in epidemic prone regions. Use of these incremental indicators are designed to enable control services to plan, prepare and implement a more effective prevention and control response in regions where malaria control has to be flexible and based on year to year changes in the epidemiology of the disease. The system should be seen as a cyclical process with the knowledge gathered from its development, adoption and use being used to refine the system and make it more intelligent to its setting as time progresses.
4 Integration of Seasonal Forecasts into Early Warning Systems
4.5
79
Best Practice
To date the implementation of an integrated approach to epidemic malaria control has advanced most notably in Southern Africa. The Southern African region has a long and varied history of malaria control with periodic epidemics occurring (Musawenkosi et al. 2004). The countries of the Southern African Development Community are committed to the Abuja Targets for Roll Back Malaria in Africa, and this includes improved detection and response to epidemics (SADC 2003). To meet these targets national malaria control services are expected to detect 60% of malaria epidemics within 2 weeks of onset, and respond to 60% of epidemics within 2 weeks of their detection. They recognize that to achieve these targets they need improved information on what type of epidemics are most likely to occur, where they are likely to occur and ideally some indication of when they are likely to occur. Experience and evidence for use of an integrated MEWS approach within a National Malaria Control Programme has been demonstrated for Botswana over the past few years (Thomson et al. 2005, 2006). Other countries in the Southern Africa Development Community (SADC) region consider this approach to provide a useful framework for planning epidemic preparedness and response strategies and have been supported by the WHO Southern Africa Inter-Country Programme for Malaria Control (SAMC) to explore these tools further (Da Silva et al. 2004). During the last 3 years WHO-SAMC and the National Malaria Control Programmes of the SADC member states have engaged in interdisciplinary preseason conferences, “Malaria Outlook Fora,” to consider the current status of malaria control within their respective countries and the vulnerabilities in the region and consider these in the light of the seasonal climate forecast for the forthcoming malaria season. The seasonal forecasts are produced using ‘state of the art’ multi-model ensembles (Thomson et al. 2006) and with the assistance of regional and nation meteorological services are tailored to the expressed needs of the malaria control community. The NMCP managers then produce epidemic preparedness plans for the coming season. Information on where to obtain routine meteorological information is also discussed and inter-sectoral collaborations are encouraged (Da Silva et al. 2004). However, the SADC countries vary markedly in their levels of endemicity/epidemicity, and current levels of capacity for surveillance and control coverage. Tanzania is for the most part a highly endemic country with an estimated 16–19 million cases per year. Botswana and Swaziland, by contrast, are currently recording cases in the low thousands and hundreds respectively. Zimbabwe’s current political and economic crisis has recently compromised its control programme making its population highly vulnerable should an epidemic occur, and two of the countries, Mozambique and Angola, are in process of reconstructing their control programmes after emerging from major disruption as a result of long-term conflict situations. While an integrated MEWS approach may be relevant to certain districts in all of the countries, emulation of Botswana’s success will be challenging for many.
80
S.J. Connor and G.C. Mantilla
Botswana it is not a poor country, it has enjoyed a stable government since independence, has a good health service infrastructure and has a long history of vector control operations to prevent malaria. A number of its neighbours also have a history of vector control but, with the exception of South Africa and Swaziland, these tend to have declined in their coverage over recent years. The Global Fund for AIDS, TB and Malaria (GFATM) is currently making significant resources available to national malaria control programmes to enable them to set priorities for malaria control and begin a much needed revitalization of their control infrastructure. The 2005–2006 malaria season acted as something of a test case for MEWS in Africa (Connor et al. 2007). The September 2005 SARCOF meeting predicted above-normal rainfall for much of southern Africa ahead of the malaria season. The MALOF meeting in November that year assessed vulnerability to malaria outbreaks based on an updated rainfall forecast. Most countries had had poor rains for the previous three seasons, so the forecast indicated a high probability of malaria epidemics in epidemic-prone areas. The rainfall proved to be both higher than normal and prolonged. The WHO InterCountry Malaria Control Programme’s first bulletin of 2006 reported that ‘malaria epidemics and incidence increase can be expected during the months of April and May. There is an increased relative risk of malaria epidemics in Botswana, Namibia and highlands of Madagascar, Mozambique, Zambia, and Zimbabwe’. Malaria outbreaks were subsequently reported in Botswana, Madagascar, Namibia, South Africa, and Zimbabwe. Zambia also reported a flood-related increase in malaria incidence that triggered an emergency situation. The WHO Inter-Country Programme’s postmortem bulletin reported on the countries’ responses. Below are some excerpts: Namibia: Preparedness for the season was much better this year when compared to the previous season. Most malaria prevention activities were timely and reached satisfactory coverages. However, due to prevailing risk factors and vulnerability, malaria epidemics still occurred in certain areas of the country and they were responded to properly. South Africa: In the current season epidemics were only reported in Limpopo Province in the months of December and January.… Extensive health education was undertaken, targeting communities at risk. The media were also used extensively in getting the message out to communities at risk and surveillance was strengthened with daily updates until the outbreak was brought under control. Zimbabwe: Some districts faced shortage of staff and some logistical problems.… Epidemic investigation and reaction to localized outbreaks was not conducted timely [sic] due to lack of fuel for vehicles. Quality control of spraying activities was also problematic due to transport problems in some parts of the country. Botswana: Malaria epidemics in Okavango region were related to increased rainfall in the Okavango Delta. Indoor residual spraying coverage was 60% and insecticide-treated nets were distributed in response.
As a point of comparison, 1996–1997 was also a wet year that followed several dry years, and there was a high incidence of malaria as a result. The figures obtained so far show that, in Botswana, the number of cases in the 2005–2006 season were about one tenth of those in 1996–1997. Similar results were reported from Swaziland. In Zimbabwe the number was about half, and in South Africa it was
4 Integration of Seasonal Forecasts into Early Warning Systems
81
about one-third. Several factors will have contributed to the reduced number of cases; in particular, control measures such as more effective drug treatments are more widely available than 10 years ago. However, early warning and the lead-time for preparedness that enables epidemic prevention has also contributed.
4.6
Looking Forward: Opportunities for Application Elsewhere
The Botswana/SADC example is considered exemplary and as a result the WHO Inter-Country Programme Teams in the other sub-regional blocks (Great Lakes and Greater Horn, and Sahelian West Africa) are negotiating access to funds from the African Development Bank to develop regional support for malaria control, allowing countries to gain value added benefit from the GFATM’s investment at the national level. This is a promising development as cross-border cooperation in malaria control is vitally important as epidemic prone areas are often based on particular environmental zones rather than administrative boundaries. For example, high rainfall anomalies in Angola may ultimately find their way as increased stream-flow into Botswana and Namibia, and create extensive breeding sites for vectors. Drought, food security, or a range of other factors, may lead to migrations of people across borders from one level of endemicity to another and pose a significant increase in epidemic risk. Development of national epidemic risk maps therefore ought to reflect the situation in neighbouring countries. Seasonal climate forecasts have the potential to contribute as one component in integrated early warning systems for climate-sensitive diseases such as malaria and dengue. The successful implementation of such EWS will however depend on close cooperation among several partners in the health, environment and climate communities. Co-operation is essential at all levels from that of the epidemic prone district to the international community. Opening these channels of communication and strengthening multi-sectoral collaborations will allow public health professionals and climate-environmental scientists and practitioners to incorporate more meaningful variables into the seasonal disease predictions. In addition, it is necessary to exchange information with institutions dealing with vulnerable populations such as food security and refugee agencies, to develop mutually beneficial mechanisms that ensure easy access and utilization of relevant information for planning or decision-making.
4.6.1
Concerns over Global Warming
Recently there have been concerns that the distribution of malaria in Africa may change as a consequence of global warming with increased transmission in areas where temperature is currently a limiting factor. Similar concerns have been
82
S.J. Connor and G.C. Mantilla
expressed in Latin America where malaria, dengue and dengue hemorrhagic fever has been increasing in recent years. In Colombia malaria incidence is known to have increased from fewer than 15 cases per 10,000 inhabitants in 1964 to 58 per 10,000 in 1983 and nearly 156 per 10,000 in 1998. The rate of dengue transmission has also shown an increasing trend with a fourfold increase in cases from 1997 to 2002. Over 80,000 Colombians fell ill to dengue and dengue hemorrhagic fever (DHF) in 2002. On a regional scale, Colombia contributed nearly 35% of reported DHF cases in the Americas from 1996 to 2004, and almost 25% of continental malaria transmission in 2003. Approximately 23 million and 13 million Colombians live in areas of endemic dengue and malaria transmission, respectively (MSPC 1996, 2000, 2003a, b). Careful analyses are required to establish (a) whether or not there are long-term trends in the climate which may be attributable to global climate processes, (b) long-term changes in malaria and dengue transmission and (c) if the evidence for (a) and (b) is positive that there is a causal relationship between the two. Increasing our knowledge of environmentally sensitive diseases, strengthening health systems and making them more intelligent in the face of a changing environment will help us now and in the future. It is recognized that by adopting an integrated approach for malaria and dengue control planning the overall health information and surveillance system will be strengthened as some other diseases have climate and environmental components to their spatial and temporal distribution. This integrated early warning approach will be applied, tested and amended to the needs of both malaria and dengue surveillance and control in Colombia over the next 5 years as a result of a collaborative GEF1 project between the Government of Colombia, the World Bank and the IRI on adaptation to climate change by the health system. The National Institute of Health, under the leadership of the Ministry of Social Protection has developed a comprehensive and integrated approach to fight dengue and malaria in Colombia. Project implementation has been scheduled for a 5-year period (2006–2011). The first year will be devoted to detailed project design (with considerable weight for initial applied research) and strengthening institutional capacities. Years 2–5 will focus on implementation of the early warning system, as well as the implementation of the wider Integrated Dengue and Malaria Surveillance and Control System in the affected areas. It is important to highlight that resources allocated to strengthening institutional capabilities (overcoming weaknesses of the existing health delivery system) represent nearly 43% of the total project costs. In line with the ENSEMBLES project which seeks to inform climate decisions on multiple time scales: inter seasonal, inter annual and inter decadal; this integrated project focuses primarily on developing a better informed infrastructure to control these diseases in the present, recognizing that by increasing our resilience to variability in climate-sensitive diseases today we also help increase our resilience to the changes our future climate may bring. 1 The Global Environmental Facility for Adaptation to Climate Change under the United Nations Framework Convention on Climate Change is administered by the World Bank, UNDP and UNEP.
4 Integration of Seasonal Forecasts into Early Warning Systems
83
References Connor, S.J. (2003). Improved knowledge on the climatic and environmental determinants of malaria distribution in sub-Saharan Africa: implications for improving control planning and reducing vulnerability to malaria (and other climate sensitive diseases). Consultancy Report: DFID Malaria Knowledge Programme, LSTM, 94. Connor, S.J. (2005). Climate and environmental monitoring applications in epidemic early warning. In: Confalonieri, U.E.C & Marinho, D.P., 2005. Remote Sensing and the Control of Infectious Diseases: Proceedings of an Inter-American Workshop. ENSP/FIOCRUZ, Rio de Janeiro, 104 pp. Connor, S.J. and M.C. Thomson (2005). Epidemic malaria: preparing for the unexpected. A Policy Brief. SciDevNet Dossier on Malaria, SciDevNet. Connor, S.J., M.C. Thomson, and D.H. Molyneux (1999). Forecasting and prevention of epidemic malaria: new perspectives on an old problem. Parassitologia 41(1–3):439–448. Connor, S.J., J. DaSilva, and S. Katikiti (2007). Malaria control in Southern Africa. In: Hellmuth, M.E., Moorhead, A., Thomson, M.C., and Williams, J. (eds). Climate Risk Management in Africa: Learning from Practice. New York: International Research Institute for Climate and Society (IRI), Columbia University. Cox, J., M. Craig, D. Le Sueur, and B. Sharp (1999). Mapping malaria risk in the highlands of Africa. Durban: Medical Research Council, 96. Craig, M.H., R.W. Snow, and D. Le Sueur (1999). A climate-based distribution model of malaria transmission in Sub-Saharan Africa. Parasitology Today 15(3):105–111. Da Silva, J., B. Garanganga, V. Teveredzi, S.M. Marx, S.J.Mason, and S.J. Connor (2004). Improving epidemic malaria planning, preparedness and response in Southern Africa. Malaria Journal 3(1):37. Goddard, L. and S.J. Mason (2002). Sensitivity of seasonal climate forecasts to persisted SST anomalies. Climate Dynamics 19:619–632. Hay, S., C.A. Guerra, A.J. Tatem, A.M. Noor and R.W. Snow (2004). The global distribution and population at risk of malaria: past, present and future. The Lancet 4:327–336. Kiszewski, A.E. and A. Teklehaimanot (2004). A review of the clinical and epidemiological burdens of epidemic malaria. American Journal of Tropical Medicine and Hygiene 71(Suppl 2): 128–135. MARA (1998). Towards an Atlas of Malaria Risk in Africa. Durban: MARA/ARMA, 30. MSPC (1996). Guia integral de manejo de las enfermedades transmitidas por vectores. Bogota: Ministerio de Protección Social de Colombia. MSPC (2000). Memorias reunión Impulso a la iniciativa hacer retroceder la malaria. Bogota: Ministerio de Protección Social de Colombia. MSPC (2003a). Malaria: Estrategias para afrontar una prioridad en salud pública. Memorias curso-taller. Bogota: Ministerio de Protección Social de Colombia. MSPC (2003b). Plan nacional de vigilancia y control de la malaria en Colombia 2003–2006. Bogota: Ministerio de Protección Social de Colombia. Musawenkosi, L., H. Mabosa, B. Sharp and C. Lengeler (2004). Historical review of malaria control in Southern Africa with emphasis on the use of indoor residual house spraying. Tropical Medicine and International Health 9(8):846–856. Najera, J.A. (1999). Malaria Control: Achievements, Problems & Atrategies. Geneva: World Health Organization. Najera, J.A., R.L. Kouznetzsov, and C. Delacollette (1998). Malaria Epidemics: Detection and Control, Forecasting and Prevention. Geneva: WHO. Sachs, J.D. and P. Malaney (2002). The economic and social burden of malaria. Nature 415:680–685. SADC (2003). SADC Malaria Report 2003. Gaborone: Southern Africa Development Community, 36.
84
S.J. Connor and G.C. Mantilla
Thomson, M.C., S.J. Connor, N. Ward, and D. Molyneux (2004a). Impact of climate variability on infectious disease in West Africa. Eco-Health 1:138–150. Thomson, M.C., P.J. Erickson, A. Ben Mohamed, and S.J. Connor (2004b). Land use change and infectious disease in West Africa. In: R. De Fries, G. Asner and R. Houghton (eds). Ecosystems and Land Use Change. Washington, DC: American Geophysical Union. Thomson, M.C., F.J. Doblas-Reyes, S.J. Mason, R. Hagedorn, S.J. Connor, T. Phindela, A.P. Morse and T.N. Palmer (2006). Malaria early warnings based on seasonal climate forecasts from multi-model ensembles. Nature 439, 576–579. Thomson, M.C., S.J. Mason, T. Phindela and S.J. Connor (2005). Use of rainfall and sea surface temperature monitoring for malaria early warning in Botswana. American Journal of Tropical Medicine and Hygiene 73(1):214–221. Wernsdorfer, W.H. (1980). The importance of malaria in the world. In: J.P. Kreir (ed). Malaria, vol 1. New York: Academic, 1–93. WHO-UNICEF (2003). Africa Malaria Report 2003. Geneva: World Health Organization, UNICEF. WHO (1993a). Implementation of The Global Malaria Control Strategy: Report of a WHO Study Group on the Implementation of the Global Plan of Action for Malaria Control 1993–2000. Geneva: WHO. WHO (1993b). A global strategy for malaria control. Geneva: WHO. WHO (1998). Roll Back Malaria: a Global Partnership. Geneva: World Health Organisation. WHO (2000). WHO Expert Committee on Malaria, 20th Report. Geneva: WHO, 71. WHO (2001a). Macroeconomics and Health: Investing in Health for Economic Development. Report on the Commission for Macro-Economics and Health. Geneva: World Health Organization, 213. WHO (2001b). Malaria Early Warning Systems: concepts, indicators and partners. A framework for field research in Africa. Geneva: WHO, WHO/CDS/RBM/2001.32. WHO (2002a). Final report on the 3rd meeting of the RBM Technical Resource Network on Epidemic Prevention and Control. Geneva: WHO. WHO (2002b). Strategic framework to decrease the burden of TB/HIV. Geneva: World Health Organization, 49. WHO (2004). Malaria epidemics: forecasting, prevention, early warning and control – From policy to practice. Geneva: World Health Organization, 59. WHO (2005a). Using climate to predict infectious disease epidemics. Geneva: World Health Organization, 54. WHO (2005b). Malaria and HIV interactions and their implications for public health policy. Geneva: World Health Organization, 38. Worrall, E., A. Rietveld and C. Delacollette (2004). The burden of malaria epidemics and costeffectiveness of interventions in epidemic situations in Africa. American Journal of Tropical Medicine and Hygiene 71(Suppl 2):136–140.
Chapter 5
Climate and Environmental Information and the Control of Epidemic Meningitis in Africa Luis E. Cuevas1 ,*, C. Anthony Hart2, and Madeleine C. Thomson3
Abstract Epidemics of bacterial meningitis have plagued the African continent for over a century creating public health emergencies with considerable morbidity and mortality. Although many bacteria can cause meningitis, most epidemics are due to a small number of meningococcal serogroups, especially groups A, C, X, Y and W135 with more than 90% of these epidemics due to group A Neisseria meningitidis. Current group A vaccines have serious shortcomings, although a new conjugate vaccine, likely to be licensed in the near future, is expected to greatly modify the epidemiology of the disease. The distribution of the epidemics has a strong association with the environment and these associations have been used to develop models to predict their location and incidence with varying degrees of success. Here we describe the characteristics of epidemic meningitis in Africa, its association with the environment, the approaches currently available for control and the advantages of the new vaccines. We also discuss the problems that may be faced with the advent of the vaccines and their large-scale implementation and how geographical information systems and climate and environmental information could play a role in informing vaccine strategies monitoring vaccine efficacy, and controlling epidemics in Africa.
Keywords Epidemics, meningococcal meningitis, Africa, geographic distribution, climate, environment, geographical information systems
1
Liverpool School of Tropical Medicine, Pembroke Place, Liverpool, UK
2
Department of Medical Microbiology, University of Liverpool, UK
3
University of Liverpool, and International Research Institute for Climate & Society, The Earth Institute at Columbia University, New York, USA * Corresponding author: Tel: + 44-151-70534219; fax: + 44-151-7053329; e-mail: lcuevas@liv. ac.uk
M.C. Thomson et al. (eds.), Seasonal Forecasts, Climatic Change and Human Health. © Springer Science + Business Media B.V. 2008
85
86
5.1
L.E. Cuevas et al.
Introduction
Meningitis is an infection of the membranes surrounding and protecting the brain (meninges). It is usually caused by an infection with a virus or bacterium. Viral meningitis is relatively mild and often clears up without specific treatment. In contrast bacterial meningitis causes significant mortality, requires rapid treatment with antibiotics and is a common disease in many areas of the world. Here we describes its epidemiology, with an emphasis on the characteristics of the disease in Africa, its association with the environment, the current and future tools available for its control and how geographical information systems and climate and environmental information could enhance surveillance and response activities.
5.2
Epidemiology of Bacterial Meningitis
The signs and symptoms of bacterial meningitis are those of infection and of inflammation of the meninges. The infection starts when the disease causing pathogens are able to colonize the nasopharynx from where they translocate to enter the circulation. The onset of disease is sudden, with fever in most cases, and is often preceded by symptoms of upper respiratory tract infection. Irritation of the meninges is manifest by nausea, vomiting, headache, irritability, confusion, back pain and neck stiffness. In children above 1 month of age and adults, three major pathogens are responsible for over 90% of cases: Streptococcus pneumoniae, Haemophilus influenzae group b and Neisseria meningitidis. Their relative contribution varies considerably for reasons that are still unclear and according to geographical area, climate, age, the presence of immunosuppression and vaccination. Of particular importance, bacterial meningitis has undergone spectacular changes in epidemiology since the introduction of effective vaccines in the early 1990s, such as the H. influenzae b conjugate vaccines (Hib) and meningococcal C, which has resulted in a sharp decline in the incidence of invasive infections by these pathogens in areas where the vaccine is used (Adegbola et al. 2005). Similarly, the availability of conjugate pneumococcal vaccines has led to the near elimination of invasive S. pneumoniae disease, including meningitis in young children (Scheld et al. 2002), in regions that can afford these vaccines. Although bacterial meningitis is relatively unusual in industrialized countries, epidemic meningitis is still common in poor countries, particularly in Africa where explosive epidemics caused by N. meningitidis have continued to affect the ‘Meningitis Belt’ (Fig. 5.1). Epidemics in this region often come in waves and the most recent wide-scale pandemic in the region, which began in 1996, resulted in over 300,000 cases and 30,000 deaths (Stephens et al. 2007). While meningitis may not be the number one killer disease in Africa it is one of the most feared diseases due to its rapid onset (sometimes a few hours) and high mortality (often greater than 10%) and
5 Climate and Environmental Information and the Control of Epidemic
87
Fig. 5.1 The meningitis belt of Africa (From Lapeyssonnie 1963)
its devastating long-term consequences (with some 10–20% of survivors developing permanent sequelae such as epilepsy, hearing loss, or mental retardation) (Stephens et al. 2007). At present, exclusive of epidemics, about 1.2 million cases of bacterial meningitis are estimated to occur annually worldwide with 135,000 deaths (WHO 1998). For N. meningitidis there are at least 15 clinically significant serogroups of the meningococci bacterium, classified according to the antigenic structure of their polysaccharide capsule, with A, B, C, X, Y and W135 being mostly responsible for human disease. Groups A and C meningococci are associated with epidemics and group B with sporadic and endemic disease and hyper-endemicity in some areas such as New Zealand, Norway, Canada and the Pacific Northwest of the USA.
88
L.E. Cuevas et al.
Most epidemics in Africa are due to serogroup A meningococci although group C meningococci occasionally cause epidemics and recently groups W135, X and Y have emerged (Gagneux et al. 2002; Savory et al. 2006). N. meningitidis clones are also able to produce epidemics spreading through the world (Moore 1992; Molesworth et al. 2002). For example, a clone of group A N. meningitidis (III-1) produced an epidemic in China in the 1970s, spread to Nepal causing an epidemic in 1983 and to New Delhi and Pakistan in 1985. In 1987 it was brought to Mecca by pilgrims attending the Haj where the frequent contact between individuals permitted the rapid transmission from asymptomatic carriers. Over 11% of Hadjis leaving Mecca that year were found to be asymptomatic carriers of this highly virulent strain which was subsequently widely disseminated throughout the continent, initiating a wave of epidemics in 1988 (Luo et al. 1998). In Africa, epidemics occur in the dry season, despite the fact that asymptomatic person-to-person spread occurs readily throughout the year. In the region, asymptomatic carriage of the meningococcus is poorly documented and some prospective studies are underway in, for example, Burkina Faso. Carriage is known to vary according to the characteristics of the individual and the environment. For example, carriage is very low in humid areas (Cheesbrough et al. 1995) while between 10% and 35% of the population are asymptomatic carriers in dry areas (Hassan-King et al. 1988). It varies significantly with age, smoking habits (Stuart et al. 1989), the presence of smoke in the household, attending public gatherings (markets, worship centers) and other factors. Carriage of the bacteria often spreads from village to village, ahead of the epidemics waves although the duration of this carriage in an individual is unknown. Asymptomatic carriage among the population of villages affected by epidemics is often similar to carriage rates among residents of villages not experiencing outbreaks (Emele et al. 1999). It is thus thought the seasonality of disease is related to increased invasiveness of the pathogen rather than increased person to person transmission, which might reflect an effect of the dust, extreme dryness and heat on the host’s mucosal defenses (Moore 1992; Hart and Cuevas 1997; Molesworth et al. 2002). There is also an association between the occurrence of meningitis in a person and a history of respiratory tract infection with viruses or mycoplasma (Moore et al. 1990). Among the few other risk factors that are well characterized, household contacts of a case run a high risk of developing infection (De Wals et al. 1981). In most cases colonization of the nasopharynx precedes invasive disease. The time from acquisition of the infection to the development of symptoms can be as short as 2–3 days but secondary cases have been reported as long as 4 months after contact. There is evidence to suggest that the risk of disease is greatest in the period immediately after colonization as 70% of secondary cases occur within the first week of contact and the remaining from the second to tenth weeks (De Wals et al. 1981). Although the mortality from meningococcal meningitis is relatively low in industrialized countries (sometimes less than 1%), its mortality in the ‘Meningitis Belt’ is much higher during inter-epidemic periods and the early stages of an epidemic. In addition, if the infection is complicated by the bacteria entering the bloodstream (septicemia) it can prove rapidly fatal (as fast as 2 h) with profound
5 Climate and Environmental Information and the Control of Epidemic
89
shock and bleeding from capillaries producing petechiae, purpura and ecchymoses on the skin. Fortunately the proportion of cases with a septicaemic component appears to be significantly lower in the Meningitis Belt (Whittle and Greenwood 1976; Salih et al. 1990; Hart and Cuevas 1997). In contrast, in tropical countries outside the African Meningitis Belt, over two thirds of cases have a septicaemic component (Riordan et al. 1995). Whether this difference represents a true difference in susceptibility to septicemia or reflects the difficulties of recognizing a rash on a dark skin or accessibility to hospital services is unclear. However the former seems more likely (Hart and Cuevas 1997).
5.3
Measures for Control
The tools available for large-scale prevention of meningococcal meningitis are not ideal and their limitations need to be considered for a better understanding of current control guidelines. Studies to establish the efficacy of chemoprophylaxis (or the provision of antibiotics to healthy individuals) to eradicate nasopharyngeal carriage of meningococci (or H. influenzae), have demonstrated benefit for the prevention of secondary cases in the families with an initial case (Correia and Hart 2004). Antibiotics that are effective in eradicating susceptible nasopharyngeal meningococci include sulphadiazine, minocycline, rifampicin, ciprofloxacin or ceftriaxone (Cuevas and Hart 1993; Correia and Hart 2004). Although chemoprophylaxis is important to prevent secondary cases in a household, it has limited public health impact during epidemics. For example, although 10% of patients presenting with meningococcal meningitis in Nigeria were secondary cases (Greenwood et al. 1979) mass chemoprophylaxis is not possible for several reasons. First, health services do not have the necessary infrastructure required for contacting families during emergencies. Second, although sulphonamides were used in Africa after the Second World War and were effective in stopping epidemics, this strategy is no longer feasible given the high prevalence of resistance to sulphadiazine and the frequent side effects observed in individuals with HIV. Third, the emergence of resistant meningococci during treatment with rifampicin (Cuevas and Hart 1993) brings with it the possibility of compromising its use as a first-line drug in tuberculosis and lastly, the high cost of alternative antibiotics. Attempts to control epidemics in Africa therefore rely on the rapid deployment of vaccines when an epidemic is detected. Vaccines are based on the capsular polysaccharides of each serogroup, are highly immunogenic and available for all of them. Polysaccharide antigens however are so called T-cell-independent antigens, resulting in poor immunological memory and have particularly poor ability to elicit suitable defenses in infants and young children (Reingold et al. 1985) who require of multiple doses (Robbins et al. 2000). The lack of immunological memory means that re-vaccination is required every few years. Vaccines are thus mostly used to control epidemics once they are already underway and not for routine vaccination.
90
L.E. Cuevas et al.
A meningococcal vaccine incorporating groups A and C capsular polysaccharides (as well as Y and W 135 if required) is available and widely used in Africa despite its limitations. Protective antibodies elicited with a single dose persist for up to 5 years in adults but for only 1–2 years in children under 4 years old. The vaccine does not affect nasopharyngeal carriage (Blakebrough et al. 1983) and does not modify the transmission of the bacteria through asymptomatic carriage. Given that the vaccine is not effective in young children and provides short-term protection, current control activities rely on mass emergency immunization at the start of an epidemic (WHO 2000). This strategy requires good surveillance and sets minimal thresholds for interventions. In populations of over 30,000, an incidence of five cases per 100,000 population per week is an alert threshold when investigation and confirmation of cases is required and surveillance should be enhanced. If there have been no epidemics in the region, and vaccine coverage is less than 80% or it is early in the dry season, then mass immunization is introduced when the incidence reaches 10 cases/100,000 per week; otherwise the vaccination threshold is 15 cases per 100,000 per week. For populations under 30,000 the alert is two cases per week and the vaccination threshold five cases per week or doubling of the number of cases in a 3-week period. Alert thresholds also become epidemic thresholds if there is an epidemic in a neighboring area. Introduction of mass immunization is a major undertaking requiring substantial logistical preparation for vaccine transport, equipment for vaccinations, mobilization of large numbers of health care workers and population access. Despite some success, this strategy has not led to the control of meningococcal meningitis in Africa. Amongst the obstacles encountered, the short shelf life of the vaccine means that health services in the region often need to procure the vaccine at short notice and the lack the human and infrastructural resources to react on time leads to vaccination campaigns often starting when the epidemic has passed its peak. At the end of the 1990s, controversy ensued: since reactive vaccination was fraught with so many problems, some have continued to argue for a strategy of preventive AC vaccination for high risk areas but this view has been hotly contested (Birmingham et al. 2003; Robbins et al. 2003). However this controversy may well die out with the development of new vaccines. Recently, conjugate Group C meningococcal vaccines have been developed and successfully introduced in industrialized countries. In this type of vaccines the polysaccharide of the meningococci is chemically linked to a protein and linkage changes the immunological properties of the polysaccharide, resulting in long-term immunological memory, even in infants (MacLennan et al. 2001). The conjugate Group C vaccine is delivered through the routine immunization services for children in the UK, where it has 92% and 97% protective efficacy in infants and adolescents respectively (Ramsay et al. 2001). This vaccine however is not appropriate for Africa, as the predominant meningococcus in this area is group A. A conjugate A, C, W135 and Y vaccine is already available in the USA, but as for other conjugate polysaccharide vaccines they are much more expensive than the polysaccharide alone and are not marketed for use in Africa. Ideally a conjugate A vaccine that is given to infants and provides life-long immunity (Peltola 2000) at low cost would be ideal for Africa. Candidate vaccines
5 Climate and Environmental Information and the Control of Epidemic
91
are currently being tested by the Meningitis Vaccine Project (http://www. meningvax.org) and new vaccines are expected to be licensed by 2009 at a price of 40 cents per dose. If approved, the vaccines will be delivered as part of routine childhood immunization (Savory et al. 2006) and to the entire population between the ages of 1 and 29. These vaccines are thus likely to dramatically change the epidemiology of epidemic meningitis in the region.
5.4
Geographical Distribution of Epidemics and Their Relationship to the Environment
Bacterial meningitis has a marked association with climate. For example, in humid low lying regions in the Congo basin, S. pneumoniae (Omanga et al. 1980) and H. influenzae b predominate although they do not cause epidemics. In dryer regions, such as in the Meningitis Belt where the climate is semi-arid (e.g. the Sahel in sub-Saharan Africa) the meningococcus causes vast epidemics (Hart and Cuevas 1997). These epidemics, which reoccur within 2–14 years, can reach incidences over 400 cases/100,000 population per year and endemic rates during inter-epidemic periods are often over 50 cases/100,000 per year (Moore 1992). The distinct geographical distribution of the Meningitis Belt suggests that the environment plays a significant role in their location and occurrence. Local populations and health staff are well aware of the link between the start and end of an epidemic and the environment, with most epidemics starting during the dry and dusty season and stopping with the arrival of the rains. A common feature for the epidemics, observed by Lapeyssonnie in his original description of the Meningitis Belt, is their location in the 300–1,100 mm mean annual rainfall isohyets and parts of Africa with high rainfall, such as in the Congo basin, are rarely affected (Fig. 5.2) (Lapeyssonnie 1963). Thus, climatic drivers may govern the distribution of the meningococcal epidemics and have been the subject of studies in the last decade for the development of statistical models for prediction. The geographical distribution and frequency of the epidemics has also changed in recent decades. In the late 1980s and 1990s many countries experienced epidemic waves with outbreaks being reported from areas where epidemics have been historically absent (Moore and Broome 1994; Savory et al. 2006). Although it is possible that new clones were spreading throughout the continent and were affecting new susceptible populations (Moore et al. 1989); it is also possible that environmental changes may have contributed. A systematic review of the epidemics occurring before and after 1975 (a time when a severe drought affected the region) illustrated that the areas affected in the latter period occurred further south than before with a pattern compatible with climate and environmental changes. Furthermore, these changes were not homogeneously distributed, with an increased risk in areas along the southern margins of the Sahel and in East and Southern Africa, while the already high risk in the meningitis belt remained unchanged.
92
L.E. Cuevas et al.
Fig. 5.2 The 300–1,100 mm mean annual rainfall isohyets of Africa (From Molesworth et al. 2002)
These changes have occurred in a pattern that is compatible with climate and environmental changes as shown in Fig. 5.3. The Meningitis Belt has thus expanded since its original description to include Tunisia and Algeria to the north and Somalia, Kenya, Tanzania, Zambia, Uganda and Rwanda to the east and south (Hart and Cuevas 1997; Molesworth et al. 2002). Epidemics have also been reported from Angola, Namibia, Mozambique, the South of the Democratic Republic of Congo (DRC: previously Zaire) and Botswana in the last decade. A map with the cumulative distribution of the epidemics in continental Africa up to the year 2004 is shown in Fig. 5.4. Climate driven statistical models could help to define areas at risk of experiencing epidemics. These would require the quantification of correlations between the observed location of epidemics and selected environmental variables. Their potential was demonstrated by Molesworth et al. (2003), who used ecological approaches to establish statistical relationships between absolute humidity, dust, rainfall, land cover type and population density and the epidemics. Although their ecological design would not allow the exploration of cause effect interactions, the models were
5 Climate and Environmental Information and the Control of Epidemic
93
Fig. 5.3 Changes in the risk of meningitis epidemics after 1975 (From Cuevas et al. 2007)
Fig. 5.4 Geographical distribution of meningitis epidemics in Africa (Modified from Molesworth et al. 2002)
94
L.E. Cuevas et al.
Fig. 5.5 An ecological model to identify areas at risk of epidemics in Africa (From Molesworth et al. 2003) (see Appendix 1)
able to predict the geographical distribution of the epidemics in the continent with 83% sensitivity and 67% specificity (Molesworth et al. 2003) (Fig. 5.5). Although the models do not seem to be used widely, probably due to their lack of an interface to allow their use at the operational level, a prospective evaluation confirmed their sensitivity and specificity (Savory et al. 2006). The association of meningococcal epidemics with the environment is not restricted to Africa (Cuevas et al. 2007). Historically, Group A epidemics occurred in other areas characterized by low humidity, as an extension of the 300–1,100 mm mean rainfall isohyets across the world. These areas include Moscow (Achtman et al. 2001), China (Hu 1990), Nepal (Cochi et al. 1987) and India (Bhatia et al. 1968). Geographical Information Systems could characterize the areas that share the ecological characteristics of the Meningitis Belt to identify populations that would benefit from vaccination programmes in the future and conversely, to further our understanding of why the epidemics are more severe and frequent in the Meningitis Belt than elsewhere in the world. There is ample potential for the development of climate/environmental models to predict the occurrence of epidemics as similar models of other diseases (e.g. malaria) have been successfully developed (Thomson et al. 2006a). Models to predict the incidence of meningitis however are at a much earlier stage of development than those designed to predict their location (Jackou-Boulama et al. 2005; Sultan et al. 2005; Thomson et al. 2006b). For example, Jackou-Boulama et al. reported that
5 Climate and Environmental Information and the Control of Epidemic
95
measuring rainfall has some potential value in predicting meningococcal meningitis incidence in Niger although the data set was limited. Meningitis incidence between 1996 and 2002 increased during the dry season, decreased at the beginning of the rainy season, and had a negative correlation coefficient with rainfall (JackouBoulama et al. 2005). Sultan et al. (2005) reported that the onset of meningococcal meningitis epidemics in Mali correlated with atmospheric indices based on surface wind speed, with the winter maximum (defined by the atmospheric index) occurring the same mean week as the epidemic onset. Thomson et al. (2006b) reported that dust, rainfall, vegetation greenness (NDVI) and cold cloud duration (CCD) anomalies (defined as the mean of each variable minus the observed value) correlated with changes in the incidence of meningitis in Burkina Faso, Niger, Mali and Togo. In this latter retrospective study, annual meningitis incidence anomalies were correlated with monthly anomalies for rainfall and dust before, during and after the epidemic seasons. These correlations varied across the regions and were stronger in savannah areas (Thomson et al. 2006b). The development of more powerful predictive models however is hampered by the lack of long-term longitudinal data covering larger geographical areas. Current models also have predictive values that are not high enough for operational use. The studies described however have highlighted that epidemic and incidence prediction are feasible (Molesworth et al. 2003) and that climate variability may influence epidemic onset. Smaller scale environment/climate-based models to predict the incidence of epidemics have unsuccessfully tried to incorporate additional information (e.g. asymptomatic carriage, vaccine use, epidemics occurrence in previous years or neighboring areas and incidence of malaria and acute respiratory infections). In Ethiopia, studies at Woreda (a sub-district) level that used an ecological approach resembling the studies conducted by Molesworth et al. (2003) had a lower predictive value than pan-continental models. This might be possibly due to the increased variability of elements of the weather on a smaller scale overriding the predictive value of the climate on a large scale. The National Surveillance System for Niger is also aiming to establish associations between the local environment and the weekly incidence of meningitis to test if the preparation of graphic displays of information could facilitate epidemic response. Although this work has an important operational component (Sidikou et al. 2003), there is no information available as to its potential to predict epidemics. Given the complexities of the epidemiology of the epidemics, the limited but improving tools for control and the changing environment, it is legitimate to discuss what are the most appropriate roles for geographical information systems and climate/environmental information for the control of epidemic meningitis. The development of predictive models and their validation requires considerable effort and furthermore significant effort is required to make such models operational. For example, only after 30 years of research for algorithm development relating climatic information with the distribution of malaria has investment in operational product development meant that early warning systems are being adopted by malaria decision-makers in Africa (Thomson et al. 1999). Despite their good predictive value in identifying epidemics in some areas, their slow uptake may
96
L.E. Cuevas et al.
be related to the fact that most malaria related deaths occur in endemic and hyperendemic areas despite the high mortality rate of cases occurring during epidemics. Although not directly comparable to the current situation of meningitis, the changes that are likely to occur in the epidemiology of the latter with the introduction of the new vaccines may create some analogies. Other issues also impede the development and implementation of models, including poor surveillance and a reluctance of health systems to integrate largescale early warning systems which require a flexible response strategy. These barriers however are diminishing with recent economic, technological and political developments. Key issues that are likely to influence control activities however need to be discussed. First, the envisaged annual production of the conjugate vaccines during the first quinquenia after registration will be in the order of tens of million doses, possibly around 40 million. However, it is estimated that about 350 million individuals live in areas affected by the epidemics (Molesworth et al. 2003). Vaccination strategies therefore will need to consider the discrepancy between the number of doses available, the need to immunize young children with several doses, older populations with a single dose and the vaccine stocks required for epidemic control in population groups not covered by the vaccine. Maps of the population at risk that are or are not eligible to receive routine vaccination and where the vaccines have been introduced would facilitate monitoring the effects that the vaccines have in areas frequently affected by epidemics. This information would be useful to qualify the risk of further epidemics and to inform the selection of new priority areas for vaccination (Molesworth et al. 2003). Much of the population data required for these maps is already available electronically (Molesworth et al. 2002) and Molesworth’s or similar maps of the distribution of epidemics may not require long to be updated. The second issue is the need to monitor vaccine efficacy after the introduction of the vaccine. The introduction of the new conjugate vaccines is following the precedent established during the introduction of conjugate vaccines against H. influenzae and group C meningococci using immunological responses as surrogates of protection (Soriano-Gabarro et al. 2004). After their introduction in routine immunization systems, enhanced national surveillance systems documented incidence cases in parallel age cohorts receiving the vaccines at different times. This approach provided strong evidence of efficacy (see http://www.dh.gov.uk/Home/fs/ en for a description of the UK meningitis C vaccine experience). National and regional surveillance systems in Africa however do not have the same sensitivity to monitor age cohorts and may be unable to demonstrate incidence changes in vaccinated and non-vaccinated cohorts. As an alternative, the demonstration that districts, countries or regions introducing the vaccine have a lower incidence and frequency of epidemics in locations with matched ecological risk where the vaccine has not been introduced could provide a powerful tool both to complement routine surveillance systems and advocacy. Early warning systems therefore would need to predict the occurrence of an epidemic at a time when epidemic frequency may start to decrease and (hopefully) disappear. There will however be a lag time between vaccine licensing and
5 Climate and Environmental Information and the Control of Epidemic
97
large-scale vaccine coverage when epidemics would stop. Although these early warning systems may receive lower priority for development, they may still warrant development for the next decade. They need to be developed soon to make an impact.
5.5
Conclusion
Meningococcal meningitis remains a significant public health threat in Africa and current tools for its control have serious shortcomings. A new conjugate vaccine is likely to be licensed in the near future and is likely to have a strong impact on the epidemiology of the disease in the subsequent decades. There is likely to be a time lag between the introduction of the vaccine and achieving full immunization of the population due to the time required for increased production of the vaccines. The region also has relatively weak surveillance systems and will require supplementary monitoring systems to demonstrate population vaccine efficacy after licensure and to further inform vaccination strategies. The distribution of the epidemics in the continent has a strong association with the environment and these associations have been used to develop predictive models of the location and incidence of epidemics. Models for the former could be used over the next decades to inform meningitis control. This would require the development of risk maps identifying areas vaccinated/not vaccinated with the new vaccines and the identification of areas ecologically susceptible to epidemics not covered by the vaccines to facilitate prioritizing of vaccination in the areas with the highest risk. The identification of epidemics in the new environment would require the rapid development of regional ‘early warning systems’ that would be useful for the control of epidemics in populations not covered by the new vaccines. Acknowledgements This paper was funded in part by grants from the Meningitis Research Foundation, UK and the US National Oceanic and Atmospheric Administration Office of Global Programs who facilitated most of the work cited in this paper by Molesworth, Savory, Hart and Cuevas. The views herein contained are those of the authors and do not necessarily reflect the views of MRF or NOAA or any of their sub-agencies.
References Achtman, M., A. van der Ende, P. Zhu, I. S. Koroleva, B. Kusecek, G. Morelli, I. G. Schuurman, N. Brieske, K. Zurth, N. N. Kostyukova and A. E. Platonov (2001). Molecular epidemiology of serogroup a meningitis in Moscow, 1969 to 1997. Emerg Infect Dis 7(3): 420–477. Adegbola, R. A., O. Secka, G. Lahai, N. Lloyd-Evans, A. Njie, S. Usen, C. Oluwalana, S. Obaro, M. Weber, T. Corrah, K. Mulholland, K. McAdam, B. Greenwood and P. J. Milligan (2005). Elimination of Haemophilus influenzae type b (Hib) disease from The Gambia after the introduction of routine immunisation with a Hib conjugate vaccine: a prospective study. Lancet 366(9480): 144–150.
98
L.E. Cuevas et al.
Bhatia, S. L., K. B. Sharma and R. Natarajan (1968). An outbreak of meningococcal meningitis in Delhi. Ind J Med Res 56(3): 259–263. Birmingham, M. E., R. F. Lewis, W. Perea, C.B.Nelson, A. Kabore and D. Tarantola (2003). Routine vaccination with polysaccharide meningococcal vaccines is an ineffective and possibly harmful strategy. Bull World Health Organ 81: 751–755. Blakebrough, I. S., B. M. Greenwood, H. C. Whittle, A. K. Bradley and H. M. Gilles (1983). Failure of meningococcal vaccination to stop the transmission of meningococci in Nigerian schoolboys. Ann Trop Med Parasitol 77(2): 175–178. Cheesbrough, J. S., A. P. Morse and S. D. Green (1995). Meningococcal meningitis and carriage in western Zaire: a hypoendemic zone related to climate? Epidemiol Infect 114(1): 75–92. Cochi, S. L., L. E. Markowitz, D. D. Joshi, R. C. Owens, Jr., D. H. Stenhouse, D. N. Regmi, R. P. Shrestha, I. Lacharya, M. Manandhar, V. L. Gurubacharya, et al. (1987). Control of epidemic group A meningococcal meningitis in Nepal. Int J Epidemiol 16(1): 91–97. Correia, J. B. and C. A. Hart (2004). Meningococcal disease. Clin Evid (12): 1164–1181. Cuevas, L. E. and C. A. Hart (1993). Chemoprophylaxis of bacterial meningitis. J Antimicrob Chemother 31(Suppl B): 79–91. Cuevas, L. E., I. Jeanne, A. Molesworth, M. Bell, E. C. Savory, S. J. Connor and M. C. Thomson (2007). Risk mapping and early warning systems for the control of meningitis in Africa. Vaccine 25S: A12–A17. De Wals, P., L. Hertoghe, I. Borlee-Grimee, S. De Maeyer-Cleempoel, G. Reginster-Haneuse, A. Dachy, A. Bouckaert and M. F. Lechat (1981). Meningococcal disease in Belgium. Secondary attack rate among household, day-care nursery and pre-elementary school contacts. J Infect 3(Suppl 1): 53–61. Emele, F. E., C. N. Ahanotu and C. E. Anyiwo (1999). Nasopharyngeal carriage of meningococcus and meningococcal meningitis in Sokoto, Nigeria. Acta Paediatr 88(3): 265–269. Gagneux, S. P., A. Hodgson, T. A. Smith, T. Wirth, I. Ehrhard, G. Morelli, B. Genton, F. N. Binka, M. Achtman and G. Pluschke (2002). Prospective study of a serogroup X Neisseria meningitidis outbreak in northern Ghana. J Infect Dis 185(5): 618–626. Greenwood, B. M., A. K. Bradley and P. G. Cleland (1979). An epidemic of meningococcal meningitis at Zaria, Northern Nigeria. Trans R Soc Trop Med Hyg 73: 557–573. Hart, C. A. and L. E. Cuevas (1997). Meningococcal disease in Africa. Ann Trop Med Parasitol 91(7): 777–785. Hassan-King, M. K., R. A. Wall and B. M. Greenwood (1988). Meningococcal carriage, meningococcal disease and vaccination. J Infect 16(1): 55–59. Hu, X. (1990). Surveillance of epidemiological features for epidemic cerebrospinal meningitis in China during 1980s. Zhonghua Liu Xing Bing Xue Za Zhi 11(6): 321–324. Jackou-Boulama, M., R. Michel, L. Ollivier, J. B. Meynard, P. Nicolas and J. P. Boutin (2005). Correlation between rainfall and meningococcal meningitis in Niger. Med Trop (Mars) 65(4): 329–333. Lapeyssonnie, L. (1963). Cerebrospinal Meningitis in Africa. Bull World Health Organ 28(Suppl):1–114. Luo, N., C. Perera, J. Holton, H. Ayles and A. Zumla (1998). Spread of Neisseria meningitidis group A clone III-I meningitis epidemic into Zambia. J Infect 36(2): 141–143. MacLennan, J., S. Obaro, J. Deeks, D. Lake, C. Elie, G. Carlone, E. R. Moxon and B. Greenwood (2001). Immunologic memory 5 years after meningococcal A/C conjugate vaccination in infancy. J Infect Dis 183(1): 97–104. Molesworth, A. M., L. E. Cuevas, S. J. Connor, A. P. Morse and M. C. Thomson (2003). Environmental risk and meningitis epidemics in Africa. Emerg Infect Dis 9(10): 1287–1293. Molesworth, A. M., M. C. Thomson, S. J. Connor, M. P. Cresswell, A. P. Morse, P. Shears, C. A. Hart and L. E. Cuevas (2002). Where is the meningitis belt? Defining an area at risk of epidemic meningitis in Africa. Trans R Soc Trop Med Hyg 96(3): 242–249. Moore, P. S. (1992). Meningococcal meningitis in sub-Saharan Africa: a model for the epidemic process. Clin Infect Dis 14(2): 515–525.
5 Climate and Environmental Information and the Control of Epidemic
99
Moore, P. S. and C. V. Broome (1994). Cerebrospinal meningitis epidemics. Sci Am 271(5): 38–45. Moore, P. S., J. Hierholzer, W. De Witt, et al. (1990). Respiratory viruses and mycoplasma as cofactors for epidemic group A meningococcal meningitis. JAMA 264: 1271–1275. Moore, P. S., M. W. Reeves, B. Schwartz, B. G. Gellin and C. V. Broome (1989). Intercontinental spread of an epidemic group A Neisseria meningitidis strain. Lancet 2(8657): 260–263. Omanga U, Nethihinyurwa M, S. D, et al. (1980). Aspectes étiologiques et évolutifs des méningites purulentes de l’enfant à Kinshasa: analyse de 471 cases. Méd d’Afrique Noire 27: 25–34. Peltola, H. (2000). Emergency or routine vaccination against meningococcal disease in Africa? Lancet 355(9197): 3. Ramsay, M. E., N. Andrews, E. B. Kaczmarski and E. Miller (2001). Efficacy of meningococcal serogroup C conjugate vaccine in teenagers and toddlers in England. Lancet 357(9251): 195–196. Reingold, A. L., C. V. Broome, A. W. Hightower, G. W. Ajello, G. A. Bolan, C. Adamsbaum, E. E. Jones, C. Phillips, H. Tiendrebeogo and A. Yada (1985). Age-specific differences in duration of clinical protection after vaccination with meningococcal polysaccharide A vaccine. Lancet 2(8447): 114–118. Riordan, F. A. I., O. Marzouk, A. P. J. Thomson, J. A. Sills and C. A. Hart (1995). Changing presentation of meningococcal disease. Eur J Pediatr 154: 472–474. Robbins, J. B., R. Schneerson and E. C. Gotschlich (2000). A rebuttal: epidemic and endemic meningococcal meningitis in sub-Saharan Africa can be prevented now by routine immunization with group A meningococcal capsular polysaccharide vaccine. Pediatr Infect Dis J 19: 945–953. Robbins, J. B., R. Schneerson, E. C. Gotschlich, I. Mohammed, A. Nasidi, J. P. Chippaux, L. Bernardino and M. A. Maiga (2003). Meningococcal meningitis in sub-Saharan Africa: the case for mass and routine vaccination with available polysaccharide vaccines. Bull World Health Organ 81(10): 745–750; discussion 751–755. Salih, M. A., H. S. Ahmed, Z. A. Karrar, I. Kamil, K. A. Osman, H. Palmgren, Y. Hofvander and P. Olcen (1990). Features of a large epidemic of group A meningococcal meningitis in Khartoum, Sudan in 1988. Scand J Infect Dis 22(2): 161–170. Savory, E. C., L. E. Cuevas, M. A. Yassin, C. A. Hart, A. M. Molesworth and M. C. Thomson (2006). Evaluation of the meningitis epidemics risk model in Africa. Epidemiol Infect 134(5): 1047–1051. Scheld, W. M., U. Koedel, B. Nathan and H. W. Pfister (2002). Pathophysiology of bacterial meningitis: mechanism(s) of neuronal injury. J Infect Dis 186(Suppl 2): S225–S233. Sidikou, F., S. Djibo, M. K. Taha, J. M. Alonso, A. Djibo, K. K. Kairo, S. Chanteau and P. Boisier (2003). Polymerase chain reaction assay and bacterial meningitis surveillance in remote areas, Niger. Emerg Infect Dis 9(11): 1486–1488. Soriano-Gabarro, M., N. Rosenstein and F. M. LaForce (2004). Evaluation of serogroup A meningococcal vaccines in Africa: a demonstration project. J Health Popul Nutr 22(3): 275–285. Stephens, D. S., B. Greenwood and P. Brandtzaeg (2007). Epidemic meningitis, meningococcaemia, and Neisseria meningitidis. Lancet 369(9580): 2196–2210. Stuart, J. M., K. A. Cartwright, P. M. Robinson and N. D. Noah (1989). Effect of smoking on meningococcal carriage. Lancet 2(8665): 723–725. Sultan, B., K. Labadi, J. F. Guegan and S. Janicot (2005). Climate drives the meningitis epidemics onset in west Africa. PLoS Med 2(1): e6. Thomson, M. C., S. J. Connor, U. D’Alessandro, B. Rowlingson, P. Diggle, M. Cresswell and B. Greenwood (1999). Predicting malaria infection in Gambian children from satellite data and bed net use surveys: the importance of spatial correlation in the interpretation of results. Am J Trop Med Hyg 61(1): 2–8. Thomson, M. C., F. J. Doblas-Reyes, S. J. Mason, R. Hagedorn, S. J. Connor, T. Phindela, A. P. Morse and T. N. Palmer (2006a). Malaria early warnings based on seasonal climate forecasts from multi-model ensembles. Nature 439(7076): 576–579.
100
L.E. Cuevas et al.
Thomson, M. C., A. M. Molesworth, M. H. Djingarey, K. R. Yameogo, F. Belanger and L. E. Cuevas (2006b). Potential of environmental models to predict meningitis epidemics in Africa. Trop Med Int Health 11(6): 781–788. Whittle, H. C. and B. M. Greenwood (1976). Meningococcal meningitis in the northern savanna of Africa. Trop Doct 6(3): 99–104. WHO (1998). Epidemic meningococcal disease. Fact sheet 105 (Geneva). WHO (2000). Detecting meningococcal meningitis epidemics in highly-endemic African countries. Wkly Epidemiol Rec 75: 306–309.
Chapter 6
A Multimodel Framework in Support of Malaria Surveillance and Control Daniel Ruiz1,2,*, Stephen J. Connor2, and Madeleine C. Thomson2
Abstract Mathematical models have played a significant role in understanding the complexity of malaria transmission dynamics. This paper describes the first steps in an effort to create an ensemble framework of various mathematical models, an approach that has been widely used to represent uncertainty in seasonal climate forecasting. This tool will be implemented to explore the role that both climatic and non-climatic factors play in fluctuations and trends in malaria incidence, and to offer useful information to effectively guide decision-makers in risk assessment, malaria control investments and choice of interventions. The initiative aims to contribute to a large programme, the Integrated Surveillance and Control System, an ambitious idea that has been proposed as an adaptation strategy to prevent the possible adverse effects of climate change on human health in Colombia. Currently, the ensemble framework is being used to: compare the simulation outputs of the selected models with actual malaria morbidity profiles observed in several endemic/epidemic-prone regions; assess changing climate and future scenarios; simulate the impact of indoor residual spraying campaigns; conduct stability analysis; and stimulate an interactive learning environment.
Keywords Malaria, modelling, dynamical models, ensemble, early warning system
1 Grupo de Profundización en Hidroclimatología, Programa en Ingeniería Ambiental, Grupo de Investigación ‘Gestión del Ambiente para el Bienestar Social – GABiS’, Escuela de Ingeniería de Antioquia, Calle 25Sur No. 42–73, Envigado, Antioquia, Colombia 2 International Research Institute for Climate and Society, The Earth Institute, Columbia University in the City of New York, USA
* Corresponding author: Tel: (57-4) 339-3200; e-mail:
[email protected]
M.C. Thomson et al. (eds.), Seasonal Forecasts, Climatic Change and Human Health. © Springer Science + Business Media B.V. 2008
101
102
6.1
D. Ruiz et al.
Introduction
Climate change is considered to be a major global environmental problem that could have significant effects on human health (Martens 1997; Patz et al. 2005). One of its potential, ecologically mediated, impacts is the change in the incidence and the spatial distribution of climate-sensitive vector-borne diseases. Malaria, dengue, leishmaniasis, and yellow fever are among those diseases considered most likely to increase as rainfall patterns change and global temperatures increase (Epstein 2000). Changes in climatic conditions could influence the behaviour of vectors (proliferation and frequency of blood meal feedings), their geographical distribution (expansion into formerly vector-free territories), and the development rate at which pathogens inside the mosquitoes mature (Epstein 2000). Diseases caused by pathogens that spend part of their life cycle outside humans or other warm-blooded hosts are particularly climate-sensitive (Patz et al. 2000; WHO 2004). Malaria, for instance, is caused by protozoan parasites (Plasmodium) that are transmitted by the bite of infected adult female Anopheles mosquitoes. The disease is widespread in tropical and subtropical regions, infects between 300 and 500 million people every year, and causes an estimated 1–3 million deaths annually, mostly among young children in Sub-Saharan Africa (WHO 1994). The most serious forms of the disease are caused by Plasmodium falciparum and P. vivax parasites. These micro-organisms multiply within the human red blood cells, causing symptoms that include fever, anaemia, chills, flu-like illness, and in severe cases, coma and death. Malaria transmission, distribution, endemicity, and seasonality have been widely linked to prevalent climatic conditions (e.g. Aron and May 1982; Macdonald 1957; Dye 1992), and frequently reported to be strongly affected by extreme weather events such as those associated with El Nino-Southern Oscillation (Poveda and Rojas 1997; Bouma et al. 1997; Poveda et al. 2000, 2001). It is generally accepted that malaria transmission can be reduced by preventing mosquito bites with bed nets and insect repellents, by spraying insecticides inside houses and draining standing water where the mosquitoes lay their eggs (Macdonald 1957), or by treating malaria primary cases. Several articles have suggested that malaria is likely to be the vector-borne disease most sensitive to long-term climate change (Lindsay and Birley 1996; Martens 1997; Martens et al. 1999; Patz et al. 2000, 2002): ‘ongoing warming will have enlarged the zone of potential malaria transmission from an area containing 45% of total world’s population to an area containing about 60%’ by 2100 (Epstein 2000). In 2003 the WHO indicated in the summary document ‘Climate change and human health – risks and responses’ that increases in global temperatures of 2–3°C, already predicted by General Circulation Models, would (a) alter the geographic range (changes in both latitude and altitude) of vector-borne diseases, increasing ‘the number of people who, in climatic terms, are at risk of malaria by around 3–5%’; (b) change the seasonality (timing of peaks in transmission) of the disease in many currently endemic and epidemic areas; and (c) increase the seasonal duration of malaria in epidemic-prone regions. There is now a growing body of literature
6 A Multimodel Framework in Support of Malaria Surveillance
103
arguing that global climate change could be a major driving mechanism behind increasing incidence of this disease (Patz et al. 2000). The WHO (World Health Report 2002) suggested that global warming has been causing deaths and disabilityadjusted life years since the mid-1970s, and estimated that an increase of about 6% of malaria incidence, during 2000, in some middle-income countries could be attributed to climate change. As scientific and technological advances permit better predictions of climatic conditions over the immediate term, there is ‘an evident opportunity to incorporate climate forecasting capabilities into control plans’ aimed at reducing the human health impacts of malaria endemicity and outbreaks (Poveda et al. 2001). Malaria, however, is not only climate-sensitive. It is a highly complex multi-factorial disease. Changing social and economic conditions, including land use changes (which affect the local micro-climate), human migration, access to interventions (including insecticide-treated bed nets, effective medicines, effective indoor-residual spraying, environmental management and effective drugs), nutritional status and co-infection with other parasitic diseases, all have significant effects on the general epidemiology of malaria, the vulnerability of populations and on the timing and severity of disease outbreaks. The effect of global warming on malaria transmission has been generally examined, however, only in terms of increased vectorial capacity, ignoring for instance that the ‘impact of temperature changes could be disproportionately higher in populations with low levels of antimalarial immunity’ (Yang and Ferreira 2000). In many tropical endemic areas the effect of increasing temperatures may be less important than changes in socioeconomic conditions (Yang and Ferreira 2000; Thomson and Connor 2001; Ter Veen 2002; Hay et al. 2002). The WHO (2003) has also suggested that ‘most temperate regions would remain unsuitable for transmission because either they remain climatically unsuitable or socioeconomic conditions are likely to remain unsuitable for reinvasion’. Furthermore, many infectious diseases persist in the Tropics as a function of poverty (Sachs 2001), suggesting that economic development plays an enormous role in controlling infectious diseases. There is currently much momentum behind major disease control initiatives, such as the Global Fund for AIDS, TB and Malaria; the Global Alliance for Vaccines and Immunization and the Global Network for Neglected Tropical Diseases, as well as the wide-ranging anti-poverty agenda exemplified in the Millennium Development Goals, all of which aim to reduce the burden of infectious diseases. While the outcome of climate change on disease is uncertain, we can be sure that the future distribution of climate-sensitive diseases will depend greatly on how successful these and other initiatives are in the next few years, especially if health systems are strengthened as a result. Thus, future scenarios might be uncertain if we do not have a deep understanding of the current epidemiological situation and the complexity of this multi-factorial disease. Such ‘profound’ knowledge of malaria dynamics is possible through a holistic examination of the climatic, biological, socioeconomic and demographic key-factors that are driving fluctuations, changes and trends of malaria incidence. In this context, mathematical models have been used to try and represent the causal relationship between climatic and socio-economic drivers and malaria outcome.
104
D. Ruiz et al.
Models of malaria transmission have been around for more than 90 years and subsequently we have seen the development of increasingly intricate models which try to capture this complex system. Despite inherent limitations in disease transmission modelling activities (due in part to the impossibility of perfectly representing all transmission components, including initial conditions, as they occur in the real world) we may yet find value in such exercises particularly to indicate opportunities for more effective control investment and intervention both in the immediate and medium term. However, instead of using individual models in isolation we may gain more useful insights by developing ‘ensembles’ of different models, where biases in one model may be compensated for by biases in other models and where each model may be rerun using a different set of initial starting conditions (e.g. using different values for the initial infectious population). In this way the combined outputs (represented as probability distributions) may offer useful information to effectively guide decision-makers in risk assessment and choice of control interventions. This approach has been widely used in seasonal climate forecasting where it has been shown to be an effective means of improving model predictions as well as demonstrating the level of uncertainty associated with model outputs (Hagedorn et al. 2005; Doblas-Reyes et al. 2005). Such multimodel ensemble seasonal forecast outputs have already been applied to simple statistical models based on malaria incidence and rainfall (Thomson et al. 2006) for use in malaria control planning (Connor and Mantilla, this volume). Given the complexity of malaria transmission and the need for mathematical models for predicting the impact of interventions we apply the same rationale for using a multimodel ensemble approach in the malaria forecasting which can be used in conjunction with outputs from climate models including those which result from the ENSEMBLES project (See Hewitt, this book). We now plan to apply the process prospectively in a case study based in Colombia which is outlined below.
6.2
Integrated Surveillance and Control Systems: The Colombian Initiative
Even though Colombia is responsible for only 0.25% of the total global emissions of carbon dioxide to the atmosphere (Instituto Nacional de Salud INS de Colombia 2005; Ruiz 2005) it is considered to be potentially one of the most seriously affected developing countries due to its high vulnerability to the negative impacts of climate change (World Bank Group 2006). In its first communication (NC1-2002) to the United Nations Framework Convention on Climate Change (IDEAM 2002), the Colombian government identified the human health as one of the areas of primary concern. The future climate scenario suggests a nationwide increase of 1–2°C in mean annual temperature by 2050, and variations in total annual rainfall records of about ± 15%. As a consequence, Colombia is expecting a dramatic (further) increase in malaria incidence in not only its endemic-prone areas, but also in its
6 A Multimodel Framework in Support of Malaria Surveillance
105
highlands where human populations are characterized by low immunity rates to malaria parasites. Colombia has proposed a very ambitious adaptation strategy to mitigate the possible adverse effects of climate change on human health: the National Integrated Surveillance and Control System (Instituto Nacional de Salud INS de Colombia 2005; Ruiz 2005). This initiative is part of the project ‘Integrated National Adaptation Pilot (INAP): High Mountain Ecosystems, Colombia’s Caribbean Insular Areas, and Human Health’ currently conducted and coordinated by the IDEAM. The public health strategy is based upon five linked components (see Fig. 6.1; Ruiz 2005): an Early Warning System (EWS) Framework; climate forecast, monitoring and analysis of scenarios; epidemiological surveillance and control activities; early diagnosis and treatment of primary cases; and entomological surveillance and control activities. These components are supported by structural axes representing capacity building (primary education and local expertise) and process maintenance and continuity, and are articulated by a ‘conveyor belt’ representing the institutional networks and the financial strategies. The major phases of the implementation of this adaptation strategy include: the design of the EWS; the strengthening of institutional capacity; and the implementation of the knowledge management-, monitoring- and evaluation-system. The national and international communities have been developing climate-based EWS to provide local alerts of different magnitudes at inter-annual, annual, and seasonal timescales (Poveda et al. 2001; WHO 2004). The major goal of these EWS is to enhance malaria surveillance, improve control and response to epidemics, and ‘facilitate early, coupled and environmentally sound public health interventions’ (Poveda et al. 2001). There has been a growing interest in developing frameworks within the EWS that allow a continuous assessment of local risks of malaria transmission in the face of global climate change. Those frameworks, however, have limited their focus to how climatic factors affect the transmission potential of mosquito vectors. Few efforts have explored the role both climatic and non-climatic contributors play in the dynamics of the disease and in the global increase of its incidence. Moreover, ‘most tools were developed as pure research; therefore, neither the extent to which they address specific control decisions nor their utility for planning public health interventions is completely clear’ (Yang 2000). The approach proposed for Colombia will make use of dynamical models that employ the established biological mechanisms of the transmission cycle of malaria parasites, to integrate climatic variables with demographic, epidemiological and entomological data routinely collected in malaria-prone areas. Therefore, the Integrated Surveillance and Control System will include climate forecasting capability, statistical models, eco-epidemiological mathematical models, and other promising tools to support campaigns and mitigation plans aimed at reducing human health impacts of sudden malaria outbreaks (Ruiz 2005). Global forecasting models will be incorporated to estimate future climatic conditions with a lead time of no less than four to six months. In Colombia, the strategy will use such predictions to prevent the concomitant health effects of not only the warm phase of El Niño-Southern Oscillation (ENSO) phenomena, but also the non-ENSO climate
106
D. Ruiz et al.
a
EARLY WARNING SYSTEMS FRAMEWORK
Climate-independent components
EPIDEMIOLOGICAL SURVEILLANCE Institutional networks and financial strategies
Human macro-factors
ENTOMOLOGICAL SURVEILLANCE
Climate change and climate variability
Relevant entomologicalbiological variables
CLIMATIC SCENARIOS MONITORING AND FORECASTING
Education, local expertise, process maintenance and continuity
DIAGNOSIS OF PRIMARY CASES
b EARLY WARNING SYSTEMS FRAMEWORK
Climate-independent components
EPIDEMIOLOGICAL SURVEILLANCE Institutional networks and financial strategies
Human macro-factors
ENTOMOLOGICAL SURVEILLANCE
Climate change and climate variability
Relevant entomologicalbiological variables
CLIMATIC SCENARIOS MONITORING AND FORECASTING
Education, local expertise, process maintenance and continuity
DIAGNOSIS OF PRIMARY CASES
Fig. 6.1 Integrated surveillance and control system under ‘normal’ conditions (a) and during the onset of epidemics (b). Major components are supported by axes and conveyor belts representing capacity building, institutional networks and financial strategies (From Ruiz 2005)
variability. The eco-epidemiological models, on the other hand, will be used to: (a) understand the complexity of the disease, in order to evaluate spatial and temporal risks; (b) estimate the time of occurrence and severity (or possible magnitude) of unexpected malaria outbreaks; (c) analyze the major confounders of the overall
6 A Multimodel Framework in Support of Malaria Surveillance
107
driving factors that seem to be interacting to cause abrupt increases in malaria incidence (thus, interventions can be designed to address key-variables in order to reduce the vulnerability of human populations to epidemics); (d) investigate the current decision making process and provide quantitative goals for effective interventions (timing and magnitude) adapted to the specific ecological circumstances of each endemic area (INS 2005); (e) pose and answer “what if” questions; and (f), perhaps more important, help decision makers learn. The future goal of this ecoepidemiological initiative is to allow regional health authorities to assess the local risk of malaria transmission in the face of local and regional environmental changes, and help them to determine the most appropriate preventative actions that have to be taken in order to prevent epidemics before they begin (INS 2005). It is worth mentioning that these experiences could also be implemented for other vectorborne diseases, in which increases in incidence and prevalence are attributed to climatic forcing.
6.3
Dynamical Models and Multi-Model Ensemble
A collaborative research project is currently being conducted to design and implement a multimodel ensemble approach to malaria transmission modelling. Up to now the mathematical dynamical models (see Table 6.1) discussed by Macdonald (MAC 1957), McKenzie et al. (CDE-I and CDE-II 1998), Ruiz and Yan (CDE-III 2003), Yang (YANG 2000), Martens (MAR 1997), Worrall et al. (WCT 2007), Ruiz et al. (RUIZ 2002a, b, 2003, 2006), and Hoshen and Morse (HM 2004) have been incorporated in the framework. The approach also includes the tool for decision makers proposed by Githeko and Ndegwa (2001). Each of these mathematical tools has its respective schematic diagram (see Fig. 6.2) and a stock-flow model in the multimodel ensemble (see Fig. 6.3). The main blocks in Fig. 6.2 represent the linking components of the dynamical model proposed by Ruiz et al. (2006): mosquito population (vector ecology), human population, and weather patterns. The level variables representing the total number of eggs, larvae and pupae (virtual compartments), susceptible, infected and infectious vectors (top), and susceptible, infected, infectious, and immune host (bottom) are depicted by small rectangles. Mosquito and host populations are linked by the transmission of parasites through mosquito blood meals (vertical). The noncontinuous arrows represent time delays. The ovals on top represent: (left) table functions affecting the main module and denoting the availability of adequate breeding sites, and (right) predator-prey interactions during pre-imago stages. On non-continuous arrows, the circles represent hatching, larva development and emergence success. Variables DT, RH, DP, WP and SM represent mean daily temperatures, mean daily relative humidity values, total daily rainfall records, wind patterns, and soil moistures, respectively. DT and RH affect the vector ecology and the blood meal. DP and SM affect the availability of adequate larval habitats. The ‘virtual’ compartments representing pre-imago stages are only used when vector
108
D. Ruiz et al.
Table 6.1 Exogenous variables of the mathematical models included in the multi-model ensemble Model MAC CDE-I CDE-II CDE-III YANG MAR WCT RUIZ HM Community-based √
Human population at risk Human natural birth Human natural mortality Human induced mortality Spray programme Surveillance
√
√ √ √ √
√ √ √
√
√ √ √
√ √
Parasite Parasite species Sporogony
√
√
√
√
√
Host delay for infectivity √ Host window for immunity √ Immune response Protective immunity Partial immunity Immunologic memory Human resistance Human susceptibility Human recovery
√
√ √
√
√
√
√ √ √ √ √
√ √
√
√ √
√
Human host (individual)
√
√ √
√ √
√
√
√ √ √
√
√
√ √
√ √ √ √ √ √
√ √
√
Mosquito population Vector natality Vector survivorship √ Vector induced mortality Vector feeding √ Vector infectivity √ Vector density √ Environment carrying capacity Vector susceptibility
√ √
√ √
√ √
√
√
√
√
√
√
√ √ √ √ √
√ √ √ √
√ √ √ √ √
√
√ √ √
densities obtained by simulation are required for estimating the Vectorial Capacity (Ruiz et al. 2006). The stock-flow model in the multimodel ensemble (see Fig. 6.3) represents the system of coupled differential equations and the major components, level variables and endogenous variables of each mathematical model. The major exogenous variables are divided into five major groups: community-based, parasite, individual human host, mosquito vector, and environmental variables (see Table 6.1 for the first four groups). Their values can be gathered from published literature or local data, can be directly calculated from field records or mosquito life-table studies, or can be measured through laboratory experiments. Currently, the multimodel ensemble framework is being used to: analyze the general epidemiology of the disease; compare its simulation outputs with actual
6 A Multimodel Framework in Support of Malaria Surveillance
Breeding places availability model
109
Predator-prey-food models
Pre-imago stages
Mosquito population Egg Infectious vector
Larva Infected vector
Pupa Susceptible vector
Climatic patterns BLOOD MEAL Sporozoites inoculation
DT
RH DP
Susceptible
WP SM
Infected
BLOOD MEAL Gametocytes ingestion
Infectious
Immune
Host population Fig. 6.2 Schematic diagram of the dynamical model proposed by Ruiz et al. (2006)
malaria morbidity profiles observed in several endemic/epidemic-prone regions with different eco-epidemiological settings; and simulate several changing epidemiological scenarios, to assess the effects of global warming and changes in local climatic and socioeconomic conditions on malaria transmission. The validation of the multi-model ensemble is based on monitoring malaria prevalence changes to test whether predictions from numerical simulations are consistent with field observations; and examining whether the models applied for one site can correctly predict malaria epidemics in other sites. This document discusses the usefulness of the interactive framework to: (a) understand the role of non-climatic factors; (b) simulate actual scenarios; (c) assess changing climate and future scenarios; (d) simulate interventions; (e) conduct stability analysis; and (f) support learning processes.
6.4
Role of Non-Climatic Factors
In many places, socioeconomic factors could have greater impact on malaria incidence than relevant climatic conditions (i.e. temperature). The mathematical model proposed by Yang (2000) allows us to analyze the effects that both climatic and
110
D. Ruiz et al.
Fig. 6.3 Stock-flow model of the differential-delay-equation ‘compartment’ model discussed by McKenzie et al. (1998). The model was created using the computer software Powersim Constructor Version 2.51 for Microsoft Windows XP-Home Edition Version 2002. Level variables are represented by rectangles: in this case the discussed model has a system of seven ordinary differential equations
non-climatic factors could have on the epidemiology of the disease. Yang’s model assesses the proportionate increase in malaria incidence that might be expected in a hypothetical population at risk under several qualitatively dissimilar climatic and socioeconomic scenarios. The basic reproduction rate (Ro) defined by this model involves several exogenous variables, including the average period of time delayed from the infection (infective bite) until the appearance of malaria gametocytes in the blood of a human host (1/γ1), the natural resistance rate against malaria (θ), the natural and disease-induced mortality rates for human hosts (µ and α), the transmission rate from human hosts to vectors (f), the average period to build up an effective immune response in a human host (1/γ), the duration of sporogony (1/σ2(T)), the natural and induced (for instance, by insecticides) mortality rates of mosquitoes (µ′ and α′), and the inoculation rate from vectors to humans (h). Other exogenous variables included in this tool are the rates π1, π2, and π3 at which protective immunity, partial immunity and immunologic memory decrease in human hosts, the rate of oviposition φ, the rate of eggs becoming non-viable µe(T), and the cycle duration from the egg to the mature adult (1/σ1(T)). Yang and Ferreira (2000) conducted the analyses of the impact of climatic and non-climatic factors on malaria transmission by assuming, initially, three levels of human exposure to the mosquito’s population: RISK = 0, 1 and 2, each of them reflecting representative areas of low, intermediate and high-risk contact, respectively. The level of risk is represented by the parameters h (inoculation rate) and f (transmission rate), which can be dependent on environmental conditions (Yang
6 A Multimodel Framework in Support of Malaria Surveillance
111
and Ferreira 2000). Secondly, the social and economic conditions (SEC) are subdivided into three major categories: SEC = 0, 1, and 2, each of them representing good, intermediate, and deteriorating conditions prevailing in the communities at risk. SEC reflects, for instance, ‘the health care system effectiveness in identifying promptly new malaria cases’, and the sanitation conditions and economic activities of the communities at risk (Yang and Ferreira 2000). As a final point, climatic conditions are represented in Yang’s model by three major categories: low, intermediate and high ambient temperatures. According to Yang and Ferreira, the parameters θ, 1/γ1, 1/γ, π1, π2, π3, µ, and φ are those profoundly influenced by social and economic conditions. The parameters that are strongly affected by temperature are µE(T), 1/ σ1(T) and 1/σ2(T). The authors studied the model by analyzing its steady state equilibrium values and its dynamic state, both highly dependent on the values of the basic reproduction rate. Steady analysis. The system of differential equations of the Yang’s model exhibits an equilibrium point where malaria reaches a specific endemic level. Figure 6.4 depicts the values of the respective basic reproduction rate (Ro) for different levels of risk, socioeconomic (SEC) and ambient conditions. Ro is an estimate of the average number of secondary cases arising in a very large population at risk of completely susceptible humans following the introduction of a single primary case. Ro = 1 is defined as the transmission threshold: for values below, the disease cannot develop itself in the community at risk (Yang and Ferreira 2000) and will recede; for values above, malaria cases will propagate. If Ro > 1, but still at low values, the disease can develop itself in the community but at low endemic levels. Lower values of Ro mean that small efforts can lead to the eradication of the disease (Yang and Ferreira 2000). On the contrary, higher values of Ro mean that only greater efforts can lead to its eradication. As SEC and temperature changes are both subdivided into three classes, there are nine possible combinations of the parameters of Yang’s model for each level of risk (see Fig. 6.4). Communities living in areas with low risk of contact exhibit transitions from disease-free conditions for good and intermediate SEC, to malaria at low endemic levels for deteriorating conditions. Communities living in areas with intermediate risk of contact exhibit transitions from disease-free conditions for good SEC, to malaria at low endemic levels for intermediate and deteriorating conditions. Finally, communities living in areas with high risk of contact always experience disease at endemic levels (see Fig. 6.4); changes from intermediate to deteriorating SEC lead to transitions from disease at intermediate prevalence to malaria at higher endemic levels. It is observed, in general, that changes in social and economic conditions are more harmful than increases in ambient temperatures (see Yang and Ferreira 2000 for further analyses). Dynamic analysis. In the mathematical model proposed by Yang (2000) the fraction of individuals at a given time t who are susceptible is represented by the variable X1(t), the fraction of incubating individuals by X2(t), the fraction of infectious people by X3(t), the fraction of immune individuals by X4(t), the fraction of partially immune hosts by X5(t), the fraction of non-immune hosts but with immunologic memory by X6(t), and the fraction of incubating people after re-infection by X7(t).
112
D. Ruiz et al. GOOD
INTERMEDIATE
LOW
DETERIORATING
SOCIO-ECONOMIC CONDITIONS (SEC)
INTERMEDIATE
HIGH
AMBIENT TEMPERATURES
50 46 HIGH
40
30
BASIC REPRODUCTION RATE (Ro)
30 28 INTERMEDIATE
20
12
18
11
10 LOW-ENDEMIC
8.7
0.7 0.46
5.7 5.3
0.43
5
3.4
0.13 0.082 0.075
1.2 1.1
1.9
2.0
3.3
2.2
0.38 0.62 0.35
2.0 1.9 MALARIA-FREE
0 LOW RISK
INTERMEDIATE RISK
HIGH RISK
( LEVEL OF CONTACT )
Fig. 6.4 Steady analysis of the Yang’s model for different levels of risk, socioeconomic (SEC) and ambient conditions
The fractions of susceptible, incubating and infectious mosquitoes at time t are represented by the variables y1(t), y2(t) and y3(t), respectively. Yang and Ferreira (2000) analyzed the asymptotic equilibrium values reached by the system for three dissimilar situations: (X3,1) a disease-free community but potentially under low risk (RISK = 0, SEC = 0, T = 20°C, Ro = 0.075); (X3,2) disease at low endemic levels and intermediate risk of malaria (RISK = 1, SEC = 1, T = 21.5°C, Ro = 2.172); and (X3,3) disease at high endemic levels and under high-risk of malaria (RISK = 2, SEC = 2, T = 31°C, Ro = 46.100); see Fig. 7.5. The initial values for this analysis are: X1(0) = (0.10; 0.998; 0.998), X2(0) = (0.20; 0.001; 0.001), X3(0) = (0.20; 0.001; 0.001), X4(0) = (0.15; 0; 0), X5(0) = (0.15; 0; 0), X6(0) = (0.10; 0; 0), X7(0) = (0.10; 0; 0), y1(0) = (60%; 100%; 100%), y2(0) = (20%; 0%; 0%), and y3(0) = (20%; 0%; 0%). The Yang’s model is run for a simulation period of 1,825 days (5 years). In the malaria-free community but potentially under low risk, the disease is eradicated during the first year after the onset of an epidemic, despite the initial high prevalence (Yang and Ferreira 2000). In the community where malaria is at low endemic levels but the risk of contact is intermediate, the outbreak is reached approximately a couple of years after the introduction of malaria cases. Thus, malaria can be eliminated from the community during the first days of a controlling effort by driving the Ro to values below unity. In the community where malaria is
6 A Multimodel Framework in Support of Malaria Surveillance
113
1,0
1,0
X3,1 ( LOW RISK AND GOOD SEC )
0,9
0,8
0,8
0,7
0,7
0,6
0,6
0,5
0,5
0,4
0,4
0,3
0,3
X3,3 ( HIGH RISK AND DETERIORATING SEC )
0,2
Fraction of individuals
Fraction of individuals
0,9
0,2
( INTERMEDIATE RISK AND SEC )
X3,2
0,1
0,1
1801
1701
1601
1501
1401
1301
1201
1101
1001
901
801
701
601
501
401
301
201
101
0,0 1
0,0
Time [days]
Fig. 6.5 Time series of Yang’s model results for the epidemiological scenarios low risk, intermediate risk and high risk
at high-endemic levels and the risk of contact is also high, the outbreak is reached more rapidly and it is more accentuated (Yang and Ferreira 2000).
6.5
Simulations of Actual Situations
To analyze the overall transmission dynamics and the skill of the mathematical models to represent malaria incidence, several actual epidemiological scenarios are currently under consideration. In Africa, regions of interest include the District of Kisii on the highlands of Western Kenya and the District of Chobe on the upland plateau of northern Botswana. In South America, endemic areas include the Colombian municipalities of Nuquí on the province of Chocó, rural lowlands of the Pacific Coast, and El Bagre on the province of Antioquia along the floodplains of Cauca and Nechi rivers, on the rural lowlands of the Caribbean Coast. For each scenario several steps are conducted to determine the magnitude of the exogenous variables of the multimodel ensemble. These steps include: (a) the analysis of demographic census, to estimate the total populations at risk in rural and semi-rural areas; (b) the analysis of epidemiological time series (historical positive malaria cases for the prevalent type(s) of infection), to determine current conditions, malaria prevalence, major type of infection, initial conditions, etc.; (c) the analysis of vector population dynamics, to determine the primary vectors incriminated in malaria transmission and their feeding/resting behaviour; and (d) the analysis of climatic time series, to detect the role climatic conditions play in malaria transmission dynamics on each study site. The Colombian epidemiological scenarios are discussed in depth by Ruiz et al. (2006). These experiences are currently being extrapolated to twelve pilot areas of the Integrated National Adaptation Pilot Project, distributed in all the Colombian geographic settings. In this paper preliminary results from the Kenyan highland
114
D. Ruiz et al.
area, the KDH village of Kisii municipality in Kisii District of the province of Nyanza (Zhou et al. 2004; Ruiz and Yan 2004), are used to explain how small changes in climatic conditions could have significant impacts on the entomological variables involved in the epidemiology of the disease. As mentioned before, many tropical and sub-tropical countries are expecting a dramatic (further) increase in malaria incidence in not only their endemic areas, but also in highlands and highland-fringes where human populations are characterized by low immunity rates to malaria parasites. That is the situation we could face in the Colombian Andean region. Even though the highlands tend to exhibit minimum ambient temperatures below 16°C that generally inhibit parasite and vector development, global warming and local climatic and environmental changes seem to be favouring malaria transmission. Preliminary results from Kisii indicate that observed annual maximum and minimum temperatures increased at rates slightly greater than 0.49°C/20 years and 0.56°C/20 years, respectively, over the available historical period (Ruiz and Yan 2004). Not dramatic changes. But such trends have an impact on the histograms representing individual frequencies of monthly maximum, mean and minimum temperatures, as depicted in Fig. 6.6. By separately reviewing the climatic time series for the periods of January 1980–December 1989 and January 1990–December 1999, it is evident that on a monthly timescale all maximum temperatures reached values above the threshold
25
T2
T1
a
Frequency [%]
20
15
10
5
0 14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Temperature [⬚C] 25
T2
T1
b
Frequency [%]
20
15
10
5
0 14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Temperature [⬚C]
Fig. 6.6 Histograms representing individual frequencies of monthly maximum, mean and minimum temperatures recorded in the Kenyan highland area for the periods January 1980–December 1989 (A) and January 1990–December 1999 (B) (From Ruiz and Yan 2004)
6 A Multimodel Framework in Support of Malaria Surveillance
115
of 21.5°C required for sporogony (denoted by T1), but remained below the 40°Cthreshold at which mosquito death occurs. In this range temperatures are favourable for both the successful completion of the sporogonic cycle and mosquito survival, assuming that the duration of sporogony is less than 20 days and (for the sake of simplifying the analysis) the average life span of the female Anopheles mosquito is about 20 days. During 80.8% of the months of 1980–1989 and 74.2% of the months in 1990–1999, the mean monthly temperatures remained below the threshold of 21.5°C and above the minimum temperature of 16°C (denoted by T2), which results in cessation of parasite development (Martens 1997). In this range temperatures are considered ‘favourable’ for parasite development, but the length of the sporogonic cycle is likely to exceed the assumed mosquito lifespan of 20 days. Hence, in theory transmission will not occur within this temperature range. The monthly minimum temperatures fell below this 16°C-temperature threshold during 61.7% of the months in 1980–1989 and 46.7% of the months in 1990–1999. As parasite death and inhibited mosquito development occur below this temperature threshold, in theory transmission will not occur within this temperature range. In summary, between 1980–1989 and 1990–1999 the number of months that experience the favourable mean monthly temperature range of 21.5–40.0°C has increased by 6.6%, while the number of months experiencing minimum monthly temperatures below the 16°C-temperature threshold has decreased by 15%. What is happening? The histograms are moving upward. It is likely that these small changes could be the driving mechanisms behind significant increases in vectorial capacity in these areas: assuming that the feeding interval, the daily survival probability of the mosquito host, and the sporogonic cycle are all function of mean indoor temperatures, vectorial capacity could theoretically have increased by 27% in the given 20-year period, from a rate of 1.5–1.9 potential infective bites per day (Ruiz and Yan 2004). The intra-annual cycles of rainfall, temperature patterns and average monthly malaria incidences observed in Kisii under ‘average’ conditions during the periods 1980–1989 and 1990–1999 seem to show similar changes (see Fig. 6.7). Analysis of the total monthly precipitation records for the periods 1980–1989 and 1990– 1999, indicates that under ‘average’ conditions the rainfall patterns exhibit an intraannual cycle with two peaks. The first peak, a large one representing the long rains, commonly occurs during the trimester April–May–June, and the second one, corresponding to the short rainy season, falls in the trimester September–October– November. Analysis of the mean monthly temperatures for the same periods indicates that under ‘average’ conditions a highly warm season typically occurs in February–March–April, and a ‘cold’ season or period of low temperatures normally occurs in July–August–September. The average monthly incidence of P. falciparum malaria during the period 1980–1999 exhibited two particular patterns. Under ‘average’ conditions, in the period 1980–1989, malaria incidence reached a peak in transmission around the month of June, following the period of long rains and while temperatures were in transition between warm and ‘cold’ seasons. In a number of instances during the period 1990–1999, favourable temperatures during the warm season correspond to a second peak in malaria incidence around the
D. Ruiz et al.
300
a
Precipitation 250
21 150 20 100
Temperature [°C]
200
19
0 02
03
04
05
06
07
08
09
10
11
21
20 100 19
0
12
18 01
02
03
04
05
06
07
08
09
10
11
12
Month
c
Malaria incidence
30
0,040
d
Malaria incidence 0,035
Maximum Monthly Temperature
30
Maximum Monthly Temperature 28
28
0,020 24
0,015
Malaria incidence
26
Temperature [°C]
0,025
0,025
26
0,020 24
0,015
Temperature [°C]
0,030
0,030
Malaria incidence
22
150
Month
0,040 0,035
23
200
50
18 01
Mean Monthly Temperature
250
22
50
b
Precipitation
Precipitation [mm]
Precipitation [mm]
300
23
Mean Monthly Temperature
Temperature [°C]
116
0,010
0,010
22
22 0,005
0,005 0,000
20 01
02
03
04
05
06
07
Month
08
09
10
11
12
0,000
20 01
02
03
04
05
06
07
08
09
10
11
12
Month
Fig. 6.7 Intra-annual cycles of total monthly rainfall and mean monthly temperature patterns observed in the Kenyan highland area under ‘average’ conditions during the periods 1980–1989 (a) and 1990–1999 (b). Intra-annual cycles of average monthly malaria incidence for the same area along with average maximum monthly temperatures for the periods 1980–1989 (c) and 1990– 1999 (d). Error bars indicate the confidence interval for a 0.05 significance level (From Ruiz and Yan 2004)
month of April. The observed new peak in transmission, which was not evident during the preceding decade, could potentially be related to changes in maximum and minimum temperatures. The above-mentioned changes are evident in the main results of two of the models of the multimodel ensemble (see Fig. 6.8): monthly values of Plasmodium falciparum malaria incidence were modelled for the period spanning from October 1, 1989 to September 30, 1998, which corresponds to a simulation exercise of about 3,287 days (approximately 9 years). Based on analyses of demographic data, the initial total human population at risk in the selected region reaches over 13,000 people, the population living in the area at the end of 1989. An average growth rate of about 0.032 individuals per year was used, assuming good conditions prevailing in the community. The initial number of total infected individuals was assumed to be equal to 10 people, equivalent to the total positive cases reported during October, 1989. The initial number of total infectious individuals was assumed to be equal to 30, equivalent to the total positive cases reported during September, 1989. The initial number of total immune individuals was assumed to be 0. The output of the mathematical models WCT and RUIZ suggest an increase in malaria incidence over
6 A Multimodel Framework in Support of Malaria Surveillance
117
Fig. 6.8 Time series of two of the models (INC_W: WCT model; Prev: RUIZ model) of the multi-model ensemble results for the analyzed highland area, for the period October 1, 1989– September 30, 1998. Gray solid bars represent the actual Plasmodium falciparum malaria incidence (IFM) observed during the corresponding period. Top panel: RUIZ model was run using mean monthly indoor temperatures and a constant vector density. Bottom panel: RUIZ model is run using mean monthly indoor temperatures and constant (gray solid line) and simulated (black solid line) vector densities
the 9-year simulation period. Modelling results show that seasonality and observed prevalence of recorded malaria outbreaks are ‘well’ represented. For a constant vector density, however, simulated prevalence exhibits peaks in transmission a few months before recorded outbreaks. The time of occurrence of such events is well simulated when vector density is obtained through simulation of the vector ecology (see Fig. 6.8).
6.6
Changing and Future Scenarios
Slight changes in local environmental conditions have significant impacts on malaria transmission. Land use changes, for instance, may affect malaria indirectly through changing the environmental conditions of mosquito breeding sites
118
D. Ruiz et al.
(Lindblade et al. 2000). Increased standing water, associated with poor drainage in the event of excessive rainfall, create more larval habitats for the development of pre-imago stages of the malaria vector. Land use changes may also alter significantly the microclimates of larval habitats and resting places of vector mosquitoes. When forests are cleared for human activities, they are typically converted into agricultural or grazing lands, which generally exhibit higher ambient temperatures. These more favourable environmental conditions could lead then to an accelerated development from eggs to adult imago, as well as increased mosquito survival and a shortening of the duration of gonotrophic and sporogonic cycles. To analyze the effect of increased standing waters and alterations of microclimates of larval habitats and resting places of vector mosquitoes on the disease dynamics, the following changing scenarios for the same area and for the same period October 1, 1989–September 30, 1998 were simulated using the RUIZ model (see Fig. 6.9): (a) constant vector density and increase of +1.0°C in mean monthly indoor temperatures for the entire simulation period (see VDC + 1 time series on Fig. 6.9); (b) constant vector density and decrease of −1.0°C in mean monthly indoor temperatures for the entire simulation period (see VDC-1 time series on Fig. 6.9); (c) fluctuations in vector density for reduction of 70% in soil infiltration (increase in the availability of breeding sites due to land use changes) and increase of +1.0°C in indoor temperatures (see VDS + 1S_1 on Fig. 6.9). For an increase of +1.0°C in mean monthly indoor temperatures and a constant vector density, RUIZ model results indicate that malaria reaches an average prevalence of about 6%, and maximum peaks in transmission of about 8% of the total population at risk. For an increase of +1.0°C and fluctuations in vector density, model results show seasonal fluctuations of disease incidence, an average malaria prevalence of about 7%, and peaks in transmission during the epochs of expected outbreaks reaching almost 9.5% of the total population at risk.
Fig. 6.9 Time series of RUIZ model results for the simulation period October 1, 1989–September 30, 1998 and for several changing scenarios. Gray bars represent actual Plasmodium falciparum malaria incidence (IFM)
6 A Multimodel Framework in Support of Malaria Surveillance
119
Monthly values of Plasmodium falciparum malaria incidence were also modelled for the period from October 1, 1989 through July 24, 2003, corresponding to a simulation exercise of about 5,045 days (ca. 13 years, see Fig. 6.10). The validation period comprised October 1, 1989–September 30, 1998, while the ‘prediction’ period (malaria data was not available) included October 1, 1998–July 24, 2003. The analysis of this ‘future’ climatic scenario allows us to preliminary estimate the potential impact of both increasing trends in temperatures and alterations of the microhabitats conditions of vector mosquitoes on malaria transmission in the selected region. Annual maximum and minimum temperatures for the period 2000–2002 are estimated using the estimated trends (rates of 0.49°C/20 years and 0.56°C/20 years from the annual maximum and minimum temperatures recorded during the previous period). Monthly maximum, mean and minimum temperatures
Fig. 6.10 Time series of RUIZ model results for the simulation period October 1, 1989–July 24, 2003. (a) The model is run for a constant vector density and three scenarios of monthly indoor temperatures: mean values (near normal-NN), 95% above normal (AN) and 95% below normal (BN); see mean values and error bars on Fig. 6.7b. (b) The model is run for mean monthly indoor temperatures and vector density obtained through simulation. In this case, a reduction of almost 70% in the soil infiltration related to possible land use changes is incorporated in the model to study the dynamics of the vector ecology
120
D. Ruiz et al.
for the period 2000–2002 are estimated using the intra-annual cycles of temperature patterns observed under ‘average’ conditions during the period 1990–1999. Since no increasing or decreasing trends were detected in the annual and monthly total precipitation records, rainfall data for the period 2000–2002 were assumed equal to the intra-annual cycle of rainfall patterns estimated under ‘average’ conditions during the period 1990–1999. Modelling results seem to suggest that, due to increasing trends in temperatures, peaks in prevalence could reach 7–8% of the total inhabitants living in the study site (see top panel on Fig. 6.10), while due to both increasing trends in temperatures and alterations of the microhabitats conditions of vector mosquitoes, peaks in prevalence could reach 8–9% of the total population at risk (see bottom panel on Fig. 6.10a, b).
6.7
Assessment of Interventions and Stability Analysis
The multi-model ensemble could be used to assess the potential reduction in Plasmodium falciparum malaria incidence that could be achieved in a given locality by an indoor residual spraying (IRS) campaign. For instance, the mathematical model proposed by Worrall et al. (2007) was used to simulate malaria transmission in the same area for the same period October 1, 1989–September 30, 1998 and for two different epidemiological scenarios (see Fig. 6.11): (a) un-sprayed population (the probability of a vector surviving each gonotrophic cycle was assumed to be constant and equal to EXP(−1/1.229) according to the value suggested by Worrall et al. (2007); this variable represents the percentage of vectors surviving each feeding cycle in the unsprayed population); and (b) sprayed population (it is assumed a 24% coverage achieved by the spray programme and an effective residual life of the insecticide of about 6 months; in this case, the percentage of vectors surviving each
Fig. 6.11 Time series of the WCT model results for the analyzed highland area, for the period October 1, 1989–September 30, 1998. Solid bars represent the actual Plasmodium falciparum malaria positive cases observed in the area during the corresponding period (IFM time series)
6 A Multimodel Framework in Support of Malaria Surveillance 1,0
Prevalence in humans
1,0
0,8
3 3
3
0,6
0,6
0,4
0,0
3
0,8
3 3
0,2
121
1 2
1 2
1 3
M
2
G
3
MG
1
100
2 2
3
2
MM
GG MMGG
2
0
150
1
1
1 3
0,0 50
1
2
0,4
0,2
2
0
1
1 2
50
100
Time [days]
Time [days]
a
b
150
Fig 6.12 Time course of model results of the classic differential-equation model CDE-I (a) and the differential-delay-equation ‘compartment’ model CDE-II (b). Variables M (or MM) and G (or GG) denote the proportions of infected and infectious humans. Variable MG (or MMGG) represents the overall malaria prevalence
3 Infectious vectors ( FF )
Infectious vectors ( F )
4,0
3,0
2,0
1,0
0,0 0,00
0,10
0,20
0,30
0,40
2
1
0 0,00
0,10
0,20
0,30
0,40
Proportion of infectious hosts (G)
Proportion of infectious hosts (GG)
a
b
0,50
0,60
Fig. 6.13 Phase plane of the classic differential-equation model (a) and the differential-delayequation ‘compartment’ model (b)
feeding cycle is assumed to be equal to EXP(−1/0.4)). WCT model results suggest that malaria positive cases can be reduced by almost 600 individuals when an IRS campaign is carried out in the area (see Fig. 6.11). The multi-model ensemble also helps to analyze the trajectories of the solutions of the mathematical models in their respective phase planes. As an example, the models CDE-I and CDE-II discussed by McKenzie et al. (1998) are used to analyze the dynamics of the disease in a hypothetical population at risk (see Figs. 6.12 and 6.13). In this case, it was assumed a total population at risk of 500 individuals; the total number of susceptible, infected, infectious, and immune individuals was assumed to be equal to 375, 125, 0, and 0 individuals, respectively. Also, a length of the interval between infection and the onset of infectivity in the human host of 20 days, a host delay from infectious to immune in human hosts of 20 days, an average time in the immune state of the human host of 12.55 days, a daily rate of vector natality of 0.50, a length of the interval between infection and the onset of infectivity in the mosquito of 10 days, a daily survival probability of the mosquito
122
D. Ruiz et al.
of 0.95, a daily rate of vector biting of 0.50 day−1, and a vector population size of 5,000 mosquitoes were assumed. In Fig. 6.13, the CDE-II model exhibits an asymptotically stable node or spiral moving towards the equilibrium point as time increases.
6.8
Learning Process and Final Remarks
It is well known that knowledge in the context of malaria incidence relates specifically to information learned about the disease transmission process. The interactive framework could also be used as a powerful tool to support learning processes. It is suggested that, with caution, transmission models driven by climatic variables can help decision-makers understand the relative importance of factors affecting changes in disease over the immediate and medium-term. Even though it was not demonstrated here, it is argued that more could be gained from using a multi-model ensemble framework based on differing starting conditions and that this may more effectively capture some of the uncertainty between the models world and the real conditions. In the short-term future we will express the ensemble outputs as probability distributions and use the model interface as an exploratory learning tool to help disease control managers understand decision risk more fully and support the investment and intervention choices they need to make. Acknowledgements Daniel Ruiz would like to thank Dr. Guiyun Yan from the College of Health Sciences, University of California, Irvine for all his support during the project ‘Biology of African Malaria’. The authors would like to thank Mark Cane from the Lamont-Doherty Earth Observatory of the Columbia University in the City of New York, and Hernan Alonso Moreno, Adriana María Molina and Ana María Zambrano from the Grupo de Profundización en Hidroclimatología, Programa en Ingeniería Ambiental, Escuela de Ingeniería de Antioquia (Colombia) for all their valuable contributions to the work described. The Instituto de Hidrología, Meteorología y Estudios Ambientales (IDEAM), the Instituto Nacional de Salud de Colombia (INS), Paola Mejía from the Mailman School of Public Health-Columbia University in the City of New York, and Hyun M Yang from Departamento de Matemática Aplicada do Instituto de Matemática, Estatística e Ciência da Computação-Universidade Estadual de Campinas, Brasil are thanked for their support throughout the project. Analyses conducted by Daniel Ruiz during the period 2003–2004 were supported by the Department of Biological Sciences-State University of New York at Buffalo. Current modelling activities are being supported by Programa Ingeniería Ambiental-Escuela de Ingeniería de Antioquia, and the International Research Institute for Climate and Society (IRI), Columbia University in the City of New York.
References Aron, J. and R.M. May (1982). The population dynamics of malaria. Anderson, R.M. (ed). The population dynamics of infectious diseases: theory and applications. Chapman & Hall, London, 139–179.
6 A Multimodel Framework in Support of Malaria Surveillance
123
Bouma, M., G. Poveda, W. Rojas, M.L. Quinones, J. Cox and J. Patz (1997). Predicting high-risk years for malaria in Colombia using parameters of El Niño-Southern Oscillation. Trop Med Int Health 2:1122–1127. Doblas-Reyes, F.J., R. Hagedorn and T.N. Palmer (2005). The rationale behind the success of multi-model ensembles in seasonal forecasting. Part II: calibration and combination. Tellus 57A: 234–252. Dye, C. (1992). The analysis of parasite transmission by blood-sucking insects. Annu Rev Entomol 37: 1–19. Epstein, P.R. (2000). Is global warming harmful to health? Scientific American, August 2000. Githeko, A.K. and W. Ndegwa (2001). Predicting malaria epidemics in the Kenyan highlands using climate data: a tool for decision makers. Glob Change Hum Health 2: 54–63. Hagedorn, R., F.J. Doblas-Reyes and T.N. Palmer (2005). The rationale behind the success of multi-model ensembles in seasonal forecasting. Part I: basic concept. Tellus, 57A: 219–233. Hay, S.I., J. Cox, D.J. Rogers, S.E. Randolph, D.I. Stern, G.D. Shanks, M.F. Myers and R.W. Snow (2002). Climate change and the resurgence of malaria in the East African highlands. Nature 415: 905–909. Hoshen, M.B. and A.P. Morse (2004). A weather-driven model of malaria transmission. Malaria J 3:32, doi:10.1186/1475-2875-3-32. Instituto de Hidrología, Meteorología y Estudios Ambientales (IDEAM) (2002). Executive summary of Colombia’s First National Communication to the United Nations Framework Convention on Climate Change, 22 pp. Instituto Nacional de Salud de Colombia (2005). Sistema Integrado de Vigilancia y Control de Malaria y Dengue en Colombia. Lindblade, K.A., E.D. Walker, A.W. Onapa, J. Katungu and M.L. Wilson (2000). Land use change alerts malaria transmission parameters modifying temperature in a highland area of Uganda. Trop Med Int Health 5: 263–274. Lindsay, S.W. and M.H. Birley (1996). Climate change and malaria transmission. Ann Trop Med Parasitol 90: 573–588. Macdonald, G. (1957). The epidemiology and control of malaria. Oxford University Press, London. Martens, W.J.M. (1997). Health impacts of climate change and ozone depletion. An eco-epidemiological modelling approach. Maastricht, The Netherlands. Martens, P., R.S. Kovats, S. Nijhof, P. de Vries, M.T.J. Livermore, D.J. Bradley, J. Cox and A.J. McMichael (1999). Climate change and future populations at risk of malaria. Glob Environ Change 9: 89–107. McKenzie, F.E., R.C. Wong and W.H. Bossert (1998). Discrete-event simulation models of Plasmodium falciparum malaria. Simulation 71(4): 213–217. Patz, J.A., T.K. Graczyk, N. Geller and A.Y. Vittor (2000). Effects of environmental change on emerging parasitic diseases. Int J Parasitol 30(12–13): 1395–405. Patz, J.A., M. Hulme, C. Rosenzweig, T.D. Mitchell, R.A. Goldberg, A.K. Githeko, S. Lele, A.J. McMichael and D. Le Sueur (2002). Regional warming and malaria resurgence. Nature 420: 627–628. Patz, J.A., D. Campbell-Lendrum, T. Holloway and J.A. Foley (2005). Impact of regional climate change on human health. Nature 438: 310–317. Poveda, G. and W. Rojas (1997). Evidencias de la asociación entre brotes epidémicos de malaria en Colombia y el fenómeno El Niño-Oscilación del Sur. Revista de la Academia Colombiana de Ciencia 21:421–429. Poveda, G., N.E. Graham, P.R. Epstein, W. Rojas, M.L.Quinones, I.D. Velez and W.J.M. Martens (2000). Climate and ENSO variability associated with vector-borne diseases in Colombia. In: El Niño and the Southern Oscillation, Multi-scale Variability and Global and Regional Impacts. 1st edn. Diaz H.F. and Markgraf V. (eds). Cambridge University Press, Cambridge/ New York, pp 183–204.
124
D. Ruiz et al.
Poveda, G, W. Rojas, M.L. Quiñones, I.D. Velez, R.I. Mantilla, D. Ruiz, J.S. Zuluaga and G.L. Rua (2001). Coupling between annual and ENSO timescales in the malaria-climate association in Colombia. Environ Health Perspect 109: 489–493. Ruiz, D. (2005). Fase de Preparación (PDF-B) del proyecto Integrated National Adaptation Pilot (INAP): High Mountain Ecosystems, Colombia’s Caribbean Insular Areas, and Human Health – Sub-actividad ‘Modelación Biológica’, Actividad ‘Evaluación del Riesgo’, Área temática ‘Sistema de Alerta Temprana’, Componente ‘Salud Humana’. Grupo de Profundización en Hidroclimatología, Programa en Ingenieria Ambiental, Escuela de Ingenieria de Antioquia. Final report, 210 pp. Ruiz, D. and G. Yan (2003). Biology of African Highland Malaria: report. Department of Biological Sciences, State University of New York at Buffalo, 60 pp. Ruiz, D. and G. Yan (2004). Modeling malaria transmission for Kisii District in the highlands of Western Kenya. Part I: analysis of base scenarios. Department of Biological Sciences, State University of New York at Buffalo, 45 pp. Ruiz, D., G. Poveda, M.L. Quiñones, I.D. Velez, G. Rua, W. Rojas and J.S. Zuluaga (2002a). Modelación sistémica para el diagnóstico de la interacción clima-malaria en Colombia. Aplicación durante El Niño 1997–1998 y La Niña 1998–2000. Meteorol Colomb 5: 41–48. Ruiz, D., G. Poveda, M.L. Quiñones, I.D. Vélez, G. Rúa, W. Rojas, and J.S. Zuluaga (2002b). Modeling entomological-climatic interaction of malaria transmission at Nuquí (Colombian Pacific Coast). Case of study: El Niño 1997–1998 and La Niña 1998–2000. In: Conference Climate Variability and Change and their Health Effects in the Caribbean: Information for Climate Variability and Change Adaptation Planning in the Health Sector. PAHO/WHO. Bridgetown, Barbados. Ruiz, D., G. Poveda, R.I. Mantilla, M.L. Quiñones, I.D. Vélez, G. Rúa, W. Rojas and J.S. Zuluaga (2003). Modelación de la interacción entomológica-climática de la transmisión de la malaria mediante Dinámica de Sistemas1 Rev Colomb Entomol 29: 191–201. Ruiz, D., G. Poveda, I.D. Velez, M.L. Quiñones, G.L. Rua, L.E. Velasquez and J.S. Zuluaga (2006). Modelling entomological-climatic interactions of Plasmodium falciparum malaria transmission in two Colombian endemic-regions: contributions to a National Malaria Early Warning System. Malaria J 5:66, doi:10.1186/1475-2875-5-66 (http://www.malariajournal. com/content/5/1/66). Sachs, J.D. (2001). Macroeconomics and health: investing in health for economic development. Report of the Commission on Macroeconomics and Health – World Health Organization, 213 pp. Ter Veen, J.M.L. (2002). The effect of climate, economics and environment on the transmission of malaria in the USA between 1900 and 1946. London School of Hygiene and Tropical Medicine, 72 pp. Thomson, M.C. and S.J. Connor (2001). The development of Malaria Early Warning Systems for Africa. Trends Parasitol 17(9): 438–445. Thomson, M.C., F.J. Doblas-Reyes, S.J. Mason, R.Hagedorn, S.J. Connor, T. Phindela, A.P. Morse and T.N. Palmer (2006). Malaria early warnings based on seasonal climate forecasts from multi-model ensembles. Nature 439: 576–579. The World Bank Group (2006). Colombia-Integrated National Adaptation Program Project. Documents & Reports. World Health Organization (1994). World malaria situation in Part I. Wkly Epidemiol Rec 72:269–274. World Health Organization (2002). World Health Report 2002: Reducing risks, promoting healthy life. Publications of the World Health Organization, Geneva. World Health Organization (2003). Climate change and human health – Risks and responses: summary. Publications of the World Health Organization, ISBN 92 4 159081 5, Geneva, Switzerland, 38 pp. World Health Organization (2004). Using Climate to Predict Infectious Disease Outbreaks: A Review. Geneva: World Health Organization. (WHO/SDE/OEH/04.01).
6 A Multimodel Framework in Support of Malaria Surveillance
125
Worrall, E., S.J. Connor and M.C. Thomson (2007). A model to simulate the impact of timing, coverage and transmission intensity on the effectiveness of indoor residual spraying (IRS) for malaria control. Trop Med Int Health 12(1): 75–88. Yang, H. (2000). Malaria transmission model for different levels of acquired immunity and temperature-dependent parameters (vector). Rev Saúde Pública 34(3): 223–231. Yang, H. and M. Ferreira (2000). Assessing the effects of global warming and local social and economic conditions on the malaria transmission. Rev Saúde Pública 34(3): 214–222. Zhou, G., N. Minakawa, A. Githeko and G. Yan (2004). Association between climate variability and malaria epidemics in the East African Highlands. PNAS USA 101(8): 2375–2380.
Recent Developments and Next Steps in Seasonal Forecasting and Health Madeleine C. Thomson1, Stephen J. Connor1, and Bettina E. Menne2
Introduction The underlying theme of this session, seasonal forecasting and health, from research to operations, is rapidly moving to centre stage in the global discussions on adaptation to climate variability within a changing climate. Research funded by the European Union and described in the chapter by Hewitt (this book) ‘The impact on human health of climate and climate change: Research in the ENSEMBLES project from seasonal to centennial timescales’ has been the ‘front runner’ in this process. It builds on the work of the previous EU funded DEMETER project led by the European Centre for Medium Range Weather Forecasting (ECMWF) which paved the way by demonstrating that improvements in weather and climate science can be rapidly integrated into improvements in health delivery systems; if one is prepared to invest at the interface between the health–climate, research–operational/ stakeholder, communities.
Partnership Development Since the Wengen 10th Anniversary Workshop in September, 2005, new partnerships between the health and climate communities have been created, and new ideas on how such partnerships can be designed to help deliver better health have been developed. Senior policy makers in developing countries are rapidly becoming more concerned about the implications of climate change on their economies and the welfare
1
The International Research Institute for Climate and Society, Earth Institute at Columbia University, New York 10964, USA 2
The European Centre for Environment and Health, WHO-Europe, via Francesco
M.C. Thomson et al. (eds.), Seasonal Forecasts, Climatic Change and Human Health. © Springer Science + Business Media B.V. 2008
127
128
M.C. Thomson et al.
of their increasingly vulnerable populations. For instance, the Africa Partnership Forum (http://www.africapartnershipforum.org/) is a key forum at a senior political level for discussing and monitoring policy issues, strategies and priorities in support of Africa’s socio-economic development. The APF meets twice each year and is co-chaired on an alternating basis by representatives of the African Union, NEPAD, G8 and non-G8 OECD member countries. In its most recent communiqué following the May 2007 meeting in Berlin, it identified climate change as a major threat to development, including the achievement of the Millennium Development Goals (MDGs) in Africa. Recognising that Africa is a continent that contributes least to the problem, and yet is most vulnerable to the impacts, they stress that more needs to be done both to reduce future emission levels, and to help Africa adapt to the changes – including better management of climate risks including drought, desertification, flooding and associated epidemics – which are already happening.
Climate for Development in Africa (ClimDev-Africa) This concern has been carefully and strategically considered in Africa: within the AU-NEPAD Environmental Action Plan (July 2003); the AU-NEPAD Disaster Risk Reduction Strategy (July 2004); and recently through the organization of the Climate for Development Needs Stakeholders Workshop: an action plan for Africa (ClimDev-Africa), initiated in 2006 by the Global Climate Observing System (GCOS) in response to the G8–2005 Summit. This latter initiative is led by the Joint Secretariat of the African Union (AUC), the African Development Bank and the United Nations Economic Commission for Africa. The ClimDev-Africa Action Plan sets out to address priority actions geared towards raising awareness, fostering appropriate policy and practice, and undertaking climate risk management in climate sensitive sectors (including health), providing climate services, and upgrading climate observations, infrastructure and data management where shown to be necessary. Documents developed to support this process include: ●
Sustainable Development in Africa: Is the Climate Right?1
This document, prepared for the G8 Gleneagles Summit, sets out the rationale for managing climate risk towards the achievements of the MDGs in Africa. ●
A Gap Analysis for the Implementation of Global Climate Observing System in Africa2
In particular, this document attributed the weaknesses of climate science and services, and in the use of information to manage climate risk across sub-Sahara Africa
1 2
http://iri.columbia.edu/outreach/publication/report/05-01/report05-01.pdf http://iri.columbia.edu/outreach/publication/report/06-01/report06-01.pdf
Recent Developments and Next Steps in Seasonal Forecasting
129
to “market atrophy;” arguing that lack of effective demand has both aggravated and resulted from poor access to relevant climate information and services. Identifying elements of best practice in climate risk management was an explicit request of the GCOS stakeholder meeting held in Addis Ababa in April 2006 which initiated the ClimDev programme. In response to this request the IRI dedicated the inaugural issue of a new report series ‘Climate and Society’ to capturing the results of a number of case studies which show elements of best practice in climate risk management including ‘demand-led’ innovations in water, health, food security and rural agricultural development: Called ‘Climate Risk Management in Africa: Learning from practice’, http://iri. columbia. edu/climateandsociety the report describes current efforts that are helping societies better adapt to climate risk, and shows that when climate information successfully reaches vulnerable populations, and related service agencies, it can be used to improve livelihoods and economies, and even save lives. Detailing five examples of how climate risks are being effectively managed in Africa, the book was launched on January 30 during a special session of the 8th African Union Heads-of-State Summit held in Addis Ababa, Ethiopia. Interest in improving the outcome of climate sensitive health events has also resulted in new advocacy organizations within the development community. For instance, the Health and Climate Foundation (http://www.hc-foundation.org/) a Swiss-based non-governmental organization, was established in 2007 with the following objective: ‘to increase the well-being of people in developing countries through better understanding of the connections between health, climate and other environmental factors; and the use of this knowledge at international, national and local levels to combat disease and poverty’.
Part II
Climate Change and Health Martin Beniston1 and Ricardo García-Herrera2
Introduction The impacts of climatic change on human health are complex to ascertain, because populations have different vulnerabilities to climate and susceptibility to disease. However, there are likely to be a number of distinct types of impacts, for example, those related to (a) the physiological effects of heat and cold on the human body; (b) the impact of heat, cold, CO2 on allergenic plants, pollens and pollution; (c) the spatial and temporal distribution of vector-borne disease; and (d) vulnerability of the population to poor health outcomes as a result of climate impacts on livelihoods.
Pollen, Allergies, and Climate This short session addressed issues of shifting plant distributions in a warming climate and the potential impacts of in- and out-migration of allergenic species on human health response, e.g., pollinosis and asthma. It was shown on the basis of French and Swiss studies that there has been a systematic shift towards earlier blooming of plant and tree species and changes in the timing of release of pollen into the atmosphere. In addition, invasive species such as the very aggressive Ambrosia (ragweed), whose spread into hitherto Ambrosia-free zones is related not only to climatic change but also to land-use changes, has substantially increased the number of pollen-related allergies. This is particularly the case in Switzerland, where Ambrosia has spread into the western part of the country from France and into the southern part from Italy, with signs of a rise in plant numbers at increasingly higher altitudes. As a means of understanding and ultimately forecasting the mechanisms by which 1 Climate Change and Climate Impacts, University of Geneva, Battelle, Building D 7 route de Drize, CH 1227 Carouge, Switzerland 2
Departamento Física de la Tierra II, Facultad de Físicas, Universidad Complutense, Ciudad Universitaria, 28040, Madrid, Spain
M.C. Thomson et al. (eds.), Seasonal Forecasts, Climatic Change and Human Health. © Springer Science + Business Media B.V. 2008
131
132
M. Beniston and R. García-Herrera
pollen is transported and dispersed in the atmosphere, preliminary results from numerical simulations of the diurnal cycle of ragweed release, transport/dispersion, and deposition were presented. This innovative approach to pollen studies, using a regional climate model within which a pollen dispersion module is embedded, can help in developing early warning systems for public health authorities, and for quantifying the potential changes in the seasonality and amount of pollen in the atmosphere under various scenarios of climatic change.
Extreme Events and Health The session on extreme climatic events and health examined the relationship between human responses to the occurrence of extreme climatic episodes, such as the 2003 summer heat wave in Europe that resulted in 30,000 excess deaths in Europe, and both the human and institutional responses to such relatively uncommon events. For example, it appears that humans respond different absolute climatic thresholds depending on their geographical situation, but that appropriate responses by emergency management entities can significantly modify the impacts of such extremes. A broader context for the meeting was to explore the connections between climate science and health early warning systems on the one hand, and emergency preparedness entities and the general public on the other. It was noted that crises arising from major disasters are much more effective at enabling effective mitigation actions, by providing “learnful moments” that focus the attention of public institutions. Two case studies related to the 2003 summer heat wave in Europe provided a useful context to illustrate the above statement about mitigation and response to climatic extremes. The 2003 heat wave was blamed for the occurrence of over 20,000 excess deaths in France alone. It was noted that the same mistake was not likely to be repeated, when the next major heat wave affects the region. On the other hand, in Spain, the response, which at first appeared to be one of denial of the fact that spikes in mortality were being record in various parts of the country, may only have changed to the extent of reporting the impacts of future climatic extremes as they unfold, but not necessarily including plans to undertake preventative actions proactively in order to reduce such impacts. Health risks associated with increased surface ozone and their climatic associations through wind regimes and location of pollution sources were also discussed.
Water- and Vector-Borne Diseases Despite significant uncertainties related to climate change, projections suggest important consequences for the evolution of infectious diseases that already have considerable impacts on morbidity, mortality, and general health status. When
Climate Change and Health
133
potential changes in climate are combined with evolution of other human activities such as migration, changing production strategies, infrastructure developments, there is even higher probability that the geographic distribution of disease and the number of people at risk for infection could increase. A presentation on the “Potential Impact of Climate Change and Water Resource Development on the Transmission of Schistosoma japonicum in China” and the discussion of “Land Use/Cover Changes and Rising Malaria Cases: The Case of South-western Amazonia” were stark illustrations of such possibilities. In China, models show that temperatures are rising and that the critical threshold for development of host and parasites for schistosomiasis may also be increasing, thus exposing millions of new people to the disease. In addition, the second largest water infrastructure project (after the Three Gorges Dam), intended to deliver water to northern provinces of China, will provide new pathways for the spread of the disease. Similar mechanisms appear to be affecting the spread of malaria in the southwestern Amazon, where, in addition to climate change, road construction is leading to important ecological changes in land cover, favouring the development of malaria transmitting mosquitoes, which can effect a human population that is growing as it moves into the region, following the path of the roads.
Policy and Adaptation Response to these kinds of processes requires both information and public policy response. One paper addressed the first issues with his discussion of “A French Initiative on Remote Sensing and Public Health”. The endeavour will use satellites to provide real-time data on epidemiological, social, biophysical, hydrological, and climate data. These are intended to become the foundations for health early warning and information systems. Information gathering and response strategies incur substantial costs, however. The general problem of cost and willingness to pay was analyzed in Canada for the specific problem of “Water Quality and Climate Change: What are Consumers Willing to Pay? The paper presented methodologies aimed at evaluating the additional costs individuals are willing to assume in order to reduce collective risks of microbial infections or cancer from their water supplies. An Internet survey was conducted in Canada to provide some concrete figures on the values people attached to improvements in the safety of their drinking water. This single example raised the important question of cost and benefits that must be posed about the development of warning systems and health infrastructure developments but that cannot be divorced from other social factors such as changing livelihoods, economic status, population movements and concentrations. The session highlighted the tight connections between health, climate, and society and underscored the need for interdisciplinary and integrative approaches.
Chapter 7
The RedGems* Information System: Climate, Environment and Public Health Yves M. Tourre1, Jean P. Lacaux1, Antonio Güell2, M. LaFaye2, Philippe Sabatier3, Eliane Coëffier4, and Laurent Braak5
Abstract Since 1999, a French multidisciplinary consortium on Spatial Surveillance of Epidemics (or S2E), has been developed, lead by the national space agency (CNES) and its subsidiaries (MEDES, MEDIAS-France, and CLS), along with french organizations involved with public health: i.e., Institut Pasteur, National Institute for Agronomic Research (INRA in french), Veterinary School of Lyon (ENVL in french). The consortium is to facilitate innovative environmental monitoring and surveillance schemes from space (Guell 2005) as well developing telemedicine (MEDES) and applications for public health improvement. An offspring of the consortium, following the French contribution during the Johannesburg Summit 2002, is the brand-new interactive health information system or RedGems (Re-Emergent Diseases Global Environment Monitoring from Space). The backbone of the site is tele-epidemiology which implies real-time monitoring of environment and intensive use of processed high resolution SPOT-5 images. The site is to contribute to the development of early warning systems (EWS), health information systems (HIS) and risks’ mapping using new tools and processed data and products obtained from space. It is believed that this new initiative will help tackling the upcoming public health challenge associated with climate variability/change, and contribute to capacity building and multidisciplinary networking worldwide. Keywords EWS, RVF, High-Res.Images, RedGems
* Re-Emergent Diseases Global Environment Monitoring from Space
1
MEDIAS-France
2
CNES, Toulouse, France
3
INRA/Veterinary School of Lyon, France
4
Institut Pasteur, Paris, France
5
MEDES, Toulouse, France
M.C. Thomson et al. (eds.), Seasonal Forecasts, Climatic Change and Human Health. © Springer Science + Business Media B.V. 2008
135
136
7.1
Y.M. Tourre et al.
Introduction
Since 1999, and in the context of climate variability and climate change, a French multidisciplinary consortium on Spatial Surveillance of Epidemics (or S2E), has been developed and includes innovative environmental monitoring and surveillance schemes from space. The core organizations and agencies of the S2E Consortium are the national space agency or CNES and its subsidiary (i.e., MEDES, MEDIASFrance, CLS), INRA (Agronomic Research Institute), the Veterinary School of Lyon (ENVL), and Institut Pasteur. The consortium collaborates with scientific laboratories as deemed necessary, such as CIRAD, CEA, IRD, Universities, and/or industrial partners such as EADS-MS2I, SOGREAH, SPOT-Image, among others. During the last 6 years, the consortium has been involved in several projects such as: S2E Dengue (Dengue fever in Guyana, see Sabatier et al. 2004), EMERCASE (Rift Valley Fever in Senegal and Southern Mauritania), S2E Argos (tropical diseases in Niger and Burkina Faso), S2E Migrating (Avian Flu with INSERM, see also http://www.shanghaipasteur.ac.cn/en/eindex.asp). Bi-lateral projects have also been implemented such MATE (Dengue fever in Argentina) between CNES and the Argentinean Spatial Agency (CONAE), or are in the implementation phase such as BIBO (Avian Flu Manuguerra 2005) with the Chinese Space Agency (CRESDA). A ‘Vibrio Project’ including cholera in the Mediterranean is being developed directly by CNES. Following the French contribution and presentation during the Johannesburg Summit 2002, among others and under the aegis of the S2E Consortium, a new interactive health information system or RedGems (Re-Emergent Diseases Global Environment Monitoring from Space), is born. Its two pillars: i.e., tele-epidemiology, and telemedicine (MEDES, http://www.medes.fr/TPS.html) are to facilitate realtime monitoring of human and animal health (i.e., epidemiology, clinical data, entomological data, etc.). The primary mission of RedGems is to analyze the health-climate/environment relationships in order to contribute to the development of early warning systems (EWS). This is to predict and mitigate public health impacts from epidemics. The integrated and multidisciplinary approach includes three main objectives: – Deployment of in situ health information systems for data information, exchange and management in real-time – Use of bio-mathematical models for epidemic dynamics (see Porphyre et al. 2005) – Remote sensing monitoring of climate and environment linking epidemics with ‘confounding factors’ such as vegetation, hydrology, forest and water extent, and population dynamics (see Lacaux and Tourre 2005; Epstein et al. 1998) The goals are to: 1. Make extensive usage of multidisciplinary information obtained from space and in situ: monitored and processed data/products and metadata 2. Disseminate state-of-the-art and integrated knowledge
7 The RedGems Information System
137
Fig. 7.1 Home page of RedGems, where the facet called ‘Diseases’ has been highlighted with a list of diseases under consideration displayed to the right. Other facets can be highlighted by moving the mouse pointer
3. Assemble and distribute latest results: dynamic, mechanisms, paradigms, and applications (Sabatier et al. 2001) 4. Promote newest, innovative and added-values products, from funded on-going projects on climate/environment and public health (Bicout and Sabatier 2004; Lacaux et al. 2006) The site is online since January 2006, the homepage is displayed in Fig. 7.1, and it is available at: http://medias.dsi.cnrs.fr/redgems.
7.2
RedGems Organization and Content
The site is organized in such ways that any user can navigate and gather the most relevant information associated with a disease. It is constructed so that diseases which have been investigated during S2E projects (see the end of Section 7.1) are being presented first. As such the Rift Valley Fever (RVF) is being considered as a case study (see EMERCASE project at http://medias.obs-mip.fr/emercase2/emercase/index; Ndione et al. 2003), and the RVF ‘template’ will be used for other diseases such as Dengue fever, Meningitis, and Malaria. Linkages with other organizations and national/international sites
138
Y.M. Tourre et al.
Fig. 7.2 The RVF page with all latest information available including bibliography
(hyperlinked) dealing with the same diseases (see the scrolling menu to the left of Fig. 7.2) are established. At the end of each page a button ‘Ask the Expert’ is to obtain additional information or answer to a specific question online. One of the principal results from the multi-disciplinary approach of RedGems and partners is that in some cases including the RVF, the knowledge of ponds’dynamics on small scales and over the Sahel, is a pre-requisite in order to apprehend mosquitoes breeding and hatching habits (Mondet et al. 2005). This lead the consortium to initiate the analysis of high resolution SPOT-5 10 m images. The processed data using new theoretical and computational approaches allowed for new products to be displayed through RedGems. RedGems is thus a ‘window’ to present examples of processed SPOT-5 10-m colored images for a particular geographical site. In Fig. 7.3, for example, ponds over northeast Senegal where the RVF is prevalent, are being detected at a given time. This is a new S2E product or NDPI (Normalized Deviation Pond Index, Lacaux et al. 2006). Ponds’ dynamic through the NDPI can be also obtained through a series of processed images during a full rainy season. This type of information can be used later-on by entomologist working on Aedes and Culex mosquitoes, carriers of RVF.Another type of information, displayed on the RedGems site is the evaluation of ponds’ turbidity and ponds with/ without vegetation (see Fig. 7.4). This is another S2E new product or NDTI (Normalize Deviation Turbidity Index).
7 The RedGems Information System
139
Fig. 7.3 False color compositing of the NDPI during August 26, 2003 (Copyright CNES 2003, distribution SPOT-Image S.A., all rights reserved). Scale: 2 cm = 750 m (= 75 pixels) (see Appendix 2)
Fig. 7.4 Ponds classification in terms of turbidity or NDTI, also for August 26, 2003 (Copyright CNES 2003, distribution SPOT-Image S.A, all rights reserved). Same scale as in Fig. 10.3. Coding: green = vegetation covered; blue = non-turbid water; purple = turbid water (see Appendix 2)
Experimental and other innovant products can be displayed as available such as the Zone Potentially Occupied by Mosquitoes or ZPOM (see Fig. 7.5). From Ba et al. (2005) the presence of mosquitoes can be as far as 500 m from the ponds. The ZPOM can then be identified after vectorizing ponds. It should be noted that the density of mosquitoes is decreasing as a function of the distance from the center of the pond, hence the coloring in Fig. 7.5 for approximately every 70 m (dark blue to light brown). MEDIAS-France expertise is also to develop generic software for Metadata Catalog. RedGems is using this opportunity to display sub-catalogs for each disease under consideration. The Metadata forms can thus be filled-in by multi-disciplinary partners on a continuous basis, using login and password, while users can have access to full catalogs and addresses of databases’ owners. This tool is to facilitate multidisciplinary research at the highest levels.
140
Y.M. Tourre et al.
Fig. 7.5 ZPOM for the same date as Figs. 7.3 and 7.4 (Copyright CNES 2003, distribution SPOT-Image S.A, all rights reserved). Same scale as in Fig. 7.3 (see Appendix 2)
7.3
Conclusion
In this paper the model for the RVF presentation through RedGEMS is shown. Readers are strongly encouraged to visit the site to appreciate fully its potentiality and content. Through the AMMA-Europe projects funded through FP6, a specific workpackage on Climate-Environment and Health is in its implementation phase. The RVF monitoring from space and in situ, as well as modelling activities is one contribution. It is foreseen to apply similar approach, techniques, and Metadata to Malaria and Meningococcal Meningitis over Niger and Burkina Faso. The latter will then be naturally implemented naturally through RedGems. Finally, RedGems is to contribute to the development of Early Warning Systems (EWS), and Health Information Systems (HIS), for effective diseases’ strategy control. It is believed that this new initiative will help tackling the new public health challenge associated with climate variability/change, to capacity building and multidisciplinary networking worldwide. In the near future the site will include additional biosphere data and products (land cover, water extent, etc.), hydrosphere data and products (rainfall, stagnant waters, wind stress, sea surface temperature, altimetry, algae blooms, etc.), atmospheric data and products (winds, convection, aerosols, etc.), and socio-economical data and products (habitat, migration, turmoil, etc.). It is argued that RedGems can become a key information system for the European Center for Disease Control and Prevention and in the challenging context of climate variability and change. Acknowledgements The authors would like to thank the constant support from CNES even during difficult implementing periods. We would like to thank Helene Gauthier-Bourret for managing the RedGems site and Cecile Vignolles, in-house teledetector who processed SPOT-5 images, and Hélène Gauthier-Bourret keeping an eye on the RedGems site, both from MEDIASFrance. Support also provided by AMMA-Europe (FP6).
7 The RedGems* Information System
141
References Ba, Y., D. Diallo, C. M. Fadel Kebe, I. Dia, and M. Diallo, 2005: Aspects of bio-ecological of two Rift Valley Fever virus vectors in Senegal (West Africa): Aedes vexans and Culex poicilipes (Diptera: Culicidae). J. Med. Entomo., 42, 5, 740–750. Bicout, D. J. and P. Sabatier, 2004: Mapping Rift Valley Fever vectors and prevalence pattern using rainfall variations. Vector Borne Zoon Dis., 4, 1, 33–42. Epstein, P. R., H. F. Diaz, S. Elias, G. Grabherr, N. E. Graham, W. J. M. Martens, E. MosleyThompson, and J. Susskind, 1998: Biological and physical signs of climate change: focus on mosquito-borne diseases. Bull. Am. Meteor. Soc., 79, 3, 409–417. Guell, A., 2005: Satellite and Epidemics. First AREVA-PASTEUR Forum 2005: Impacts of Climate Change on Emerging Diseases. Shangaï, China. http://www.shanghaipasteur.ac.cn/en/ eindex.asp Lacaux, J. P. and Y. M. Tourre, 2005: Une approche déterministe, expérimentale et par modèlisation, des relations entre le changement climatique et santé. Plan National Santé Environnement et Plan Santé Travail: Nouvelles Perspectives de Recherches. Changements Globaux et Impacts sur la Santé. Ministère Délégué de la Recherche. 31 Mars–1 Avril 2005. Lacaux, J. P., Y. M. Tourre, C. Vignolles J.A. Ndione and M. Lafaye, 2007: Classification of ponds from high-spatial resolution on remote sersing: application to Rift Valley Fever epidemics in Seregal. Rem, Sers, Env., 106: 66–74, Elsevier Publishers. (sub-judice). Manuguerra, J. C, 2005: Avian Flu. First AREVA-PASTEUR Forum 2005: Impacts of Climate Change on Emerging Diseases. Shangaï, China. http://www.shanghaipasteur.ac.cn/en/ eindex.asp Mondet, B. A. Diaïté, J. A. Ndione, A. G. Fall, V. Chevalier, R. Lancelot, M. Ndiaye, and N. Ponçon, 2005: Rainfall patterns and population dynamics of Aedes (Aedimoerphus) vexans arabiensis, Patton 1905 (Diptera: Culicidae), a potential vector of Rift Valley Fever virus in Senegal. J. Vector Ecol., 30, 1, 102–106. Ndione, J.A., J. P. Besancenot, J. P. Lacaux, and P. Sabatier, 2003: Environnement et épidémiologie de la fièvre de la vallée du Rift (FVR) dans le basin inférieur du fleuve Sénegal. Env. Risques Santé, 2, 3, 1–7. Porphyre, T., D. J. Bicout, and P. Sabatier, 2005: Modelling the abundance of mosquito vectors versus flooding dynamics. Ecol. Model., 183, 173–181. Sabatier, P., M. A. Dubois, R. Lancelot, P. Hendrickx, and J.P. Lacaux, 2001: Space surveillance and mathematical modelling of the Rift Valley Fever. “Observing our environment from space. New solutions for a new millenium”. G. Begni and A.A. Balkema, Lisse, The Netherlands, pp. 317–323. Sabatier, P., J. M. Babouchkine, J. Morvan, L. Polidori, J. P. Lacaux, L. Braak, M. A. Dubois, and H. Chaudet, 2004: S2E.Dengue. Surveillance spatiale de la Dengue. Conception d’outils d’acquisition, de cheminement, d’analyse et de prévision épidémiologique. In ITBM/RBM 25, 292–296.
Chapter 8
Impacts of Heat and Ozone on Mortality Risk in the New York City Metropolitan Region Under a Changing Climate Kim Knowlton1, Christian Hogrefe2, Barry Lynn3, Cynthia Rosenzweig3, Joyce Rosenthal4, and Patrick L. Kinney1
Abstract Climate change may lead to both increased heat and ozone (O3) levels in urban areas over the coming century. To assess potential human health impacts of these changes, models are needed for projecting regional-scale temperature and O3 changes under climate change, and for characterizing the independent and joint health effects of heat and O3. To meet these needs, mortality transfer functions for summer heat and O3 were developed and applied in a regional health risk assessment for the New York City metropolitan region. The objective was to analyze and project the relative impacts of climate-related changes in mean daily temperature and 1-hour maximum O3 concentrations on acute non-accidental mortality from all internal causes of death. Exposure-response relationships were developed using a 10-year record of daily summer observations for the region (1990-1999). This was done using a time series Poisson regression model that jointly estimated O3 and temperature effects on mortality, controlling for time trends and day of week effects. To project impacts into future decades, we developed a integrated modeling system that took global scale climate projections for the 2020s, 2050s, and 2080s, using the Intergovernmental Panel on Climate Change (IPCC) A2 and B2 emission scenario assumptions, and down-scaled these to a 36 km grid using regional models for climate and air quality. Regional downscaling was carried out using the GISS-MM5 linked global-regional model system for climate and the Community Multiscale Air Quality (CMAQ) model for air quality. Mortality risks were projected using the transfer functions estimated from the 1990s data. Results showed that both O3 and heat stress had measurable impacts on mortality risk, but that the relative impacts changed over time. This modeling strategy could be applied in other metropolitan areas and for other health outcomes to assess health impacts of heat and O3 under a changing climate.
1 Department of Environmental Health Sciences, Mailman School of Public Health, Columbia University, New York, USA 2
State University of New York, Albany, NY, USA
3
Goddard Institute for Space Studies, New York, USA
4
Graduate School of Architecture, Planning & Preservation, Columbia University, New York, USA
M.C. Thomson et al. (eds.), Seasonal Forecasts, Climatic Change and Human Health. © Springer Science + Business Media B.V. 2008
143
144
K. Knowlton et al.
Keywords Global warning, Climate change, heat, mortality, ozone, modeling, downscaling, health impacts, risk assessment
8.1
Introduction
Numerous previous studies have documented short-term increases in human mortality in response to acute heat stress and high ambient O3 concentrations, from a variety of causes but mainly from cardiovascular and respiratory illnesses. Extreme heat events have been linked extensively to short-term increases in death counts (Basu and Samet 2002; Semenza et al. 1996). Though uncertainties remain, sameday mean temperature is a useful predictor of mortality risk (Basu et al. 2002, 2005; Hajat et al. 2002; O’Neill et al. 2003). The shape of the temperature-mortality transfer function differs by location and possibly over time (Curriero et al. 2002, 2003), yet estimation of heat transfer functions in time series models has seldom taken account of air pollution effects (Katsouyanni et al. 1993). Mortality effects of O3 have been demonstrated in time series studies, controlling for temperature and other pollutants (Bell et al. 2004, 2005; Ito et al. 2005; Kinney and Ozkaynak 1991; Levy et al. 2005; Thurston and Ito 2001). A survey of 95 large US communities found that a 0.52% increase in daily mortality was associated with a 10 ppb increase in the previous week’s O3 (Bell et al. 2004). In urban areas, O3 formation stems from the release of primary precursor compounds – nitrogen oxides and volatile organic compounds – from stationary and mobile combustion sources and other natural processes. In the presence of sunlight these compounds react to form O3. While the relative contributions of different precursor emission sources vary locally, regional O3 concentrations are affected by precursors from upwind sources, as well as by global background O3 levels. Ozone formation is locally sensitive to temperature (Kleinman and Lipfert 1996), and some studies suggest that changes in regional climate may have a greater effect on ambient O3 concentrations than the effects of increased anthropogenic emissions or boundary conditions (Hogrefe et al. 2004a). The potential effects of climate change on O3 and associated mortality risk have not been examined extensively. From the wide range of potential human health impacts, why focus on heat and O3? Because heat has had among the most direct, discernable impacts and receives considerable public attention, especially after the Chicago 1995 and Europe 2003 summer heat waves. Frequently during such summer heat waves in the US peak electrical demand surges, owing to extensive use of air conditioning (A/C), causing power blackouts just when A/C is most needed to ameliorate heat stress. Over the last 20 years, the New York City metropolitan region remains an O3 non-compliance area, and high summer O3 concentrations extend into non-urban areas. At present, there are no reliable estimates of regional precursor emissions extending into the middle 21st century. Furthermore, the complex formation chemistry of O3 means that ambient concentrations do not necessarily change in direct relation to changes in precursors (Seinfeld and Pandis 1998). Because the future evolution of local O3 precursors and resultant O3 concentrations is highly uncertain, one approach for projecting future conditions is to hold O3 precursor emissions constant at current levels, in an effort to isolate the effects of climate changes upon O3 concentrations (Kinney et al. 2005).
8 Impacts of Heat and Ozone on Mortality Risk in the New York City
145
Among a wide range of adverse heat- and O3-related health outcomes that includes increases in respiratory and cardiovascular hospitalizations, mortality was the initial focus since it carries the highest “cost” in economic terms (US EPA 1997). Furthermore, looking at mortality allowed us to consider the whole region as our study area and move beyond the city limits in assessing impacts, because mortality data were available for the broad New York City metropolitan region. The central question we sought to examine was, can we assess potential future health impacts of heat and air quality at regional to local scales resulting from global climate change? Few previous impacts assessments have applied dynamically downscaled GCM-to-RCM model systems with human health impacts projections as the goal driving the study. This research provides one of the first examples of applying a risk assessment to temperature and O3 concentrations simulated from one dynamically downscaled GCM-to-RCM model system. By simultaneously projecting meteorological and air quality conditions, changes in the health impacts of these two factors can be compared.
8.2
Methods
The New York Climate and Health Project (or NYCHP) was designed to address this need and project the relative health impacts of local climate-related changes in temperatures and ground-level O3 concentrations. We compared acute summertime nonaccidental mortality at present (using data from the 1990s) to several future decades (2020s–2050s–2080s). We used a four-part methodology to assess region-specific mortality impacts. First, we sought to develop mortality transfer functions for temperature and O3 effects on summer mortality, using historical (1990–1999) death, weather, and air quality data for the study area. Next, we developed an integrated modeling system that included modules for global climate, regional climate, and regional air quality. Third, the retrospective epidemiological analysis was combined with the projective integrated climate–air quality model system through application of a health risk assessment, and current vs. future mortality was compared to assess potential mortality risks in the New York City metro area in the 21st century. Lastly, a sensitivity analysis examined alternative greenhouse gas (GHG) growth scenarios in order to assess how reduced GHG emissions might reduce potential adverse health impacts of climate change.
8.2.1
Retrospective Impacts Assessment: Epidemiological Analysis
In order to estimate the region’s population-specific mortality transfer functions for summer heat and O3, we developed a Generalized Additive Model (GAM) which included both variables for the 31-county study area (see Fig. 8.1, right panel).
146
K. Knowlton et al.
Health Impacts Domain 36 km modeling domain
NYCHP 36 km grid cells
Fig. 8.1 (top) The NYCHP modeling domain (top from Hogrefe et al. 2004b); (bottom) NYCHP 31-county study area, with 36-km modeling grid superimposed
To limit potential confounding by seasonal trends in mortality, temperature and O3, the analysis considered only three summer months (June 1–August 31) each year. Summer is the season with highest ambient O3 and heat conditions and was the model simulation period for the NYCHP. Retrospective epidemiological analysis to develop transfer functions for summer months thus facilitated their application within the prospective risk assessment framework. 8.2.1.1
Data Sources
Mortality data were obtained from the US National Center for Health Statistics (NCHS) for 1990–1999. Daily deaths counts within each of the 31 counties for all internal causes (ICD-9 codes 0–799.9 for 1990–1998 and ICD-10 codes A00-R99
8 Impacts of Heat and Ozone on Mortality Risk in the New York City
147
for 1999) were pooled, excluding accidental causes and those among nonresidents, to obtain a set of daily summer regional death count totals. Air quality data were obtained from the US EPA’s Aerometric Information Retrieval System (AIRS) for O3 monitoring stations within the study area. Of 39 reporting stations in the study area with O3 data on any of the 920 summer days from 1990–1999 (10 summer * 92 days/summer), those with fewer than 80% non-missing days were removed from further analyses. For the 16 remaining stations, there were 13,743 monitor-days with data (93.4%) and 977 monitor-days (6.6%) for which data was interpolated. None of the 920 study days had region-wide average O3 concentrations based wholly on imputed data. An exploratory analysis compared O3 β coefficient values without PM10 included in the modeling to β estimates for the O3 effect with daily PM10 included in the GAM modeling. To test region-specific effects of PM10 on O3 (because PM10 data are only collected every 6 days), the model estimated effects for the subset of summer days 1990–99 (n = 153) on which PM10 data was available from stations in the study area. The possible confounding effect of PM10 on the relationship between O3 and mortality was evaluated by comparing the following: (1) the O3 β coefficient unadjusted for PM10 on those 153 days for which PM10 data were available; (2) the O3 coefficient adjusted for same-day PM10, i.e. including it as a term in the Poisson regression. Thus the relationship between O3 and mortality was compared with and without PM adjustment, using the identical dataset. There was no significant change in the O3 β coefficient in either of the cases tested, and the widely overlapping estimates for models with and without PM10 do not provide evidence of confounding. This work along with recent studies that did not find PM10 to be an important confounder of O3-related mortality (Bell et al. 2005; Ito et al. 2005; Levy et al. 2005) supported the omission of PM10 for the central analyses. Daily mean temperature (Tave) and dewpoint temperature (both in °F) data were obtained from the National Oceanic and Atmospheric Administration (NOAA) National Climatic Data Center (NCDC) data inventory. Stations within the study area with at least 80% non-missing Tave data included 16 meteorological stations. Only six airport stations had daily dewpoint data for the years in question, and humidity was not included in the final model. 8.2.1.2
Statistical Modeling
A statistical model was developed using Poisson GAM regression with log daily death counts as the outcome variable. From β and Standard Error (or SE) estimates the incremental changes in the relative risk of mortality were calculated for Tave and maxO3 (h)ave(48). Here the choice of lag structure was made a priori from results of previous studies that found significant effects on acute mortality associated with same-day mean temperatures (Basu et al. 2005; Basu and Samet 2002; Hajat et al. 2002; O’Neill et al. 2003). Similarly, a number of studies have found O3-mortality effects at relatively short-term lags of 0–1 days for O3 (Bell et al. 2005, 2004; Ito et al. 2005; Levy et al. 2005; Thurston and Ito 2001), thus the mean of lag days 0 and 1-h max O3 concentrations (maxO3 (h)ave(48)) was applied.
148
K. Knowlton et al.
The Poisson GAM model tested under a range of different degrees of freedom for the smoothing spline of time, was: log (daily mortality) = DOW + s(time,df) + b(Tave)t + b(maxO3(h)ave(48))
(8.1)
where DOW = Day of week (indicator variable) s is the spline of time used to smooth time-varying trends in the mortality data df is the number of degrees of freedom (here 30 df) time is a variable indicating the actual day within the study b’s are the parameter estimates for the matrix of predictors (Tave)t is the daily average temperature on day t (maxO3 (h)ave(48)) is the mean 1-h max O3 averaged from same and previous day Several different combinations of temperature terms were evaluated, i.e. linear, quadratic, cubic, etc. based on previous studies that suggested a non-linear temperature–mortality relationship (Basu and Samet 2002). To evaluate possible interaction, i.e. joint effects between O3 and heat, a stratified analysis was used which divided the 1990s summer days into above- vs. below-median O3 and temperature strata. The full model was re-run within each stratum, looking for differences in the β coefficient estimates for heat and O3 across the strata.
8.2.2
Integrated Modeling System for Projecting Future Temperature and O3
The overall NYCHP project model linkages and downscaling process illustrated in Fig. 8.2 are described in detail in related papers (Kinney et al. 2006; Lynn et al. 2004, 2006; Hogrefe et al. 2004b). The GISS coupled global ocean/atmosphere model (Hansen et al. 2002; Russell et al. 1995) was driven by two different IPCC greenhouse gas scenarios. A2 projects relatively high population growth until the 2050s, increasing per capita consumption and a more efficient and balanced global use of energy sources. Since A2 per capita emissions may be similar to current levels, as GHG reductions from technological efficiencies are balanced by increasing energy demand and increasing population, it can be considered one probable evolution of emissions.B2 instead represents “smarter” growth, i.e. GHG emissions level off from 2050 onwards with a shift toward other energy sources (IPCC 2000). The GISS GCM was linked via initial and boundary conditions to the Penn State-NCAR Mesoscale Model 5 (MM5) regional climate model (Grell et al. 1994; Dudhia 1993). MM5 was run on two nested domains of 108 and 36 km over the USA. To simulate O3 air quality, the Community Mesoscale Air Quality (CMAQ) model (Byun and Ching 1999) was run at 36 km and took its initial conditions from the GISS-MM5 simulations (Hogrefe et al. 2004b). For CMAQ the 1996 US EPA National Emission Trends (NET) database was processed by the Sparse Matrix Operator Kernel Emissions Modeling System (SMOKE) (Houynoux et al. 2000) and
8 Impacts of Heat and Ozone on Mortality Risk in the New York City
149
Fig. 8.2 The New York Climate and Health Project (NYCHP) model system, showing linkages among model components
for some simulations scaled by IPCC scenarios. The simulation periods were June–August 1993–1997; June–August 2023–2027; June–August 2053–2057; and June–August 2083–2087. The MM5 model simulated temperatures and CMAQ simulated O3 concentrations across the model domain in summers for these four future decades. Gridded temperatures and O3 concentrations were interpolated to county centroid latitude/longitude coordinates using inverse distance weighting (IDW) from the three nearest station data to individual county centroids, for use in the county-level mortality risk assessments. All counties were classified into urban or non-urban, on the basis of Census 2000 population density (urban being those with 10,000 or more residents per square mile of land area; non-urban with less than 10,000 residents per square mile). ANOVA for the 1993–1997 summers showed that the urban stations’ mean temperature was significantly greater than non-urban counties. Thus in order to maintain this observed differentiation in temperatures when interpolating to geographic centroids, an additional criteria was used in interpolation. If any of the three nearest stations were located in a county with opposite urban/non-urban character, that station’s daily Tave was not included in that county’s calculation. No interpolation distance was equal to or greater than 100 km (60 miles).
8.2.3
Health Impacts Risk Assessment
The health risk assessment (RA) model framework is illustrated in Fig. 8.3. In order to isolate the impacts of climate changes on future regional mortality, we held population constant at the Census 2000 county totals. We also held anthropogenic O3 precursor emissions constant at the NET1996 inventory levels and assumed
150
K. Knowlton et al.
Study Population
Year 2000 county population
X
Base Rate
X
Baseline county mortality rate
Change in Environmental Conditions
Projected from Integrated Model: ∆ temp, DO3
Mortality Transfer Function
X
=
Number of Additional Deaths
% increase in mortality per unit D
Fig. 8.3 Health risk assessment model applied in the NYCHP
mortality rates would remain constant at county-specific mean 1990s reference rates. The daily MM5 and CMAQ model simulations at 36 km were used as the “exposure assessment” component of the RA. The mortality transfer functions estimated for the 1990s were applied in the risk assessment for both current and future decades.
8.2.3.1
Heat-Mortality Risk Assessment
Following standard practice in the climate impacts field, changes in temperatures over time were calculated by taking monthly differences between future simulated temperatures in each decade vs. the 1990s simulated temperatures for individual counties. For example, the mean simulated temperature for the five Julys of the 1990s for Bergen County (New Jersey) was computed. This value was then subtracted from the mean simulated temperature for the five Julys of the 2020s for Bergen County. This was done for each month of the decade for each county, producing decade-specific monthly differences in temperatures resulting from climate change. Note that in the climate impacts field these differences are referred to as “anomalies.” Here the more intuitive term climate-induced differences or CIDs is coined and applied: CIDMM 5 = TMM 5 _ fut − TMM 5 _ 90 s
(8.2)
where CIDMM5 is the month-, decade- and county-specific climate induced difference TMM5_fut is the mean MM5-simulated temperature for that county, decade, and month TMM5_90s is the mean MM5-simulated temperature for the same county and month, in the 1990s
8 Impacts of Heat and Ozone on Mortality Risk in the New York City
151
To characterize temperature exposures in future decades, the month- and decadespecific CIDs were added to the county-specific 1990s observations to create MM5adjusted temperature exposures: TMM5adj = Tobs + CID MM5
(8.3)
where TMM5adj is the projected daily temperature in a particular decade, month, and county Tobs is the daily observed 1990s temperature for that county, interpolated to its centroid CIDMM5 is the month-, decade- and county-specific climate-induced difference. MM5-adjusted daily county temperatures were applied in the heat-mortality risk assessment.
8.2.3.2
Ozone-Mortality Risk Assessment
To project changes in summer O3-related mortality relative to the 1990s, the risk assessment was applied to daily 1-h maximum O3 concentrations in five 1990s summers from station observations versus from five mid-decade summers from CMAQ simulations (i.e. 1993–1997 vs. 2023–27, etc.) at 36 km horizontal resolution. The mean concentrations from lag days 0 and 1, i.e. the same and previous days were calculated so that the corresponding transfer function estimates from the Poisson GAM model could be applied in the O3-mortality RA. The RA model was run for each decade, using SAS version 9.0 (SAS Institute 2002) to apply the linear-quadraticcubic heat and the linear O3 effects.
8.2.4
Sensitivity Analyses
The projected mortality changes developed above were compared to projections obtained using previously published transfer function values. A heat-mortality value developed for New York City from data for 1973–1994 suggested a 13.05% increase in daily mortality per 10°F increase in daily Tave above 73.54°F (RR = 1.1305, 95% CI 1.1186–1.1424), and found little or no increase in mortality associated on days with Tave 52.13–73.54°F (Curriero et al. 2002, 2003). For the O3-mortality transfer function, a previously published value from a meta-analysis of several studies in which non-linear temperature terms were used suggested a 5.6% increase in daily mortality per 100 ppb increase in daily 1-h maximum O3 (RR = 1.056, 95% CI 1.032–1.081), this being a non-threshold mortality effect (Thurston and Ito 2001). A second sensitivity analysis compared the main mortality risk assessment results for the 2050s A2 to 2050s projections using B2 scenario results, as a way of evaluating possible health benefits of GHG reduction efforts.
152
8.3 8.3.1
K. Knowlton et al.
Results Retrospective Impacts Assessment: Epidemiological Analysis
Preliminary analysis of the mortality and temperature data from our study area similarly suggested that days with mean temperatures below 63.6°F be eliminated from the heat-mortality analysis, since they were outliers in the temperaturemortality scatterplot, which also suggested a higher-order polynomial be used for the temperature-mortality relationship. Figure 8.4 shows a scatterplot of daily mortality vs. daily Tave with the model fit drawn through the points. The resulting Poisson GAM model estimated a linear/non-threshold transfer function for O3 and a threshold temperature of 63.6°F above which linear, quadratic and cubic temperature terms were included. On the 859 (95%) of 1990s summer days with Tave > 63.6°F, the mean daily death count was 362.0 (SD 29.17), the mean Tave was 72.72°F (SD 4.565), and the mean concentration for maxO3 (h)ave(48) was 68.94 ppb (SD 19.090). The final model was: log (daily deaths) = DOW + spline (time) + β1-3 (mean T ave )1-3 + b4 (maxO3 (h)ave (48) ) The heat-mortality effect increases with increasing temperatures owing to the higher-order temperature terms in the model. For example, for a 5°F increase in
Fig. 8.4 Scatterplot of daily mortality vs. daily Tave, with model fit indicated
8 Impacts of Heat and Ozone on Mortality Risk in the New York City
153
daily Tave from 70°F to 75°F is associated with a 1.1% increase in daily mortality across the study area, but increasing Tave by 5°F from 80°F to 85°F is associated with an 8.0% increase. The O3 effect on the other hand is linear, thus percent mortality increases remain constant across a range of incremental changes in concentration. Heat and O3 each had effects upon daily internal-cause mortality in the study area. Stratified analysis suggests that some interactive mortality effects may exist between O3 and heat. On low-temperature days, there appeared to be a significantly greater O3-mortality effect. On low-temperature days, the relative risk of O3-related mortality was 1.092 per 100 ppb increase in daily 1-h maximum O3 (95% CI 1.048, 1.138, p < 0.0001). This may suggest that at low temperatures, the O3-mortality effect is less prone to confounding by increased heat-mortality, whereas at higher temperatures the O3-mortality effect is not discernable from the large (non-linear) heat-mortality effect. However, for the preliminary analysis presented here, the risk assessment evaluated heat- and O3-related mortality effects based on the full, nonstratified analysis (see Table 8.1).
8.3.2
Projecting Potential Future Health Impacts
Temperature and O3 simulation results from the linked GISS-MM5 model system and CMAQ have been discussed at length in papers by Lynn et al. (2004, 2006) and Hogrefe et al. (2004b). Here we focus on the risk assessment results. Our risk assessment evaluated the daily summer heat-related mortality increase by application of β coefficient estimates from the epidemiological analysis in the formula: Additional heat-related mortality = (population/100,000) * (county daily mortality rate) *[exp((temp * 0.29193) + ((temp ^ 2) * (–.00434)) + ((temp ^ 3) * (.00002152)))-1]
(8.4)
For the O3 effect the daily O3-related mortality increase was evaluated as:
Table 8.1 Absolute and relative changes in climate-related mortality in future decades under the A2 scenario Decade
Regional summer heat-related mortality
Regional summer O3-related mortality
1990s 2020s A2
1,116 1,542 38% increase vs. 1990s 2,347 110% increase vs. 1990s 5,533 396% increase vs. 1990s
1,059 1,174 11% increase vs. 1990s 1,108 5% increase vs. 1990s 1,266 20% increase vs. 1990s
2050s A2 2080s A2
154
K. Knowlton et al.
Additional O3-related mortality = (population/100,000) * (daily mortality rate) * [exp((maxO3(h)ave(48)) * (0.00045738))–1]
(8.5)
Mortality for a typical summer in each decade was evaluated and compared to that in a typical 1990s summer. For example, Fig. 8.5 shows the regional distribution of percentage changes in heat-related mortality by the 2050s under the A2 assumptions, and Fig. 8.6 shows the O3 effects. The absolute and relative (percentage) changes in climate-related mortality in future decades under the A2 scenario assumptions are shown in Table 8.1. These findings suggest that O3-related mortality could decrease between the 2020s and 2050s in the New York metro region, holding O3 precursor emissions and population constant. It should be noted that O3 concentrations and related health impacts within the 31-county study area may not be representative of O3 changes across the entire eastern US modeling domain (see Fig. 8.1, top left). The downscaled GISS-MM5 model system at 36 km resolution allows us to see how population density could interact with environmental conditions. Higher population density counties tended to show greater absolute number of heat effects since more people are exposed there to heat stress, augmented by the urban heat island effect. Yet the greatest percentage heat-mortality increases by the 2050s A2 relative to the 1990s (see Fig. 8.5) occurred in non-urban counties on the perimeter of the study area, in part because of low 1990s heat-related mortality estimates there
164 180
172
163
122
158
163
133 152
158 159
103
133 102
151
114
111
100 106 110 101 100 86
134
119
% increase in summer heat-mortality vs. 1990s
125 137
151
123
102
129
81 - 100% increase 101 - 120% 121 - 140% 141 - 160% 161 - 180%
Fig. 8.5 Map of percentage changes in heat mortality impacts by 2050s A2 vs. 1990s (see Appendix 2)
8 Impacts of Heat and Ozone on Mortality Risk in the New York City
155
0.7 1.5
0.9
0.5
4.0
1.4
0.3
3.2 3.0
1.0
2.7
4.2
2.3 4.3
3.4
4.7 4.3
4.8
4.6
4.2 6.1
5.5
5.2 9.6
5.1
6.1
5.2
% increase, 2050sA2 - 90s O3 mortality
6.3 6.8
6.1
0 - 1.5% increase 1.5 - 3.0% 3.0 - 4.5% 4.5 - 6.0% > 6.0% increase
5.3
Fig. 8.6 Map of percentage changes in O3 mortality impacts by 2050s A2 vs. 1990s (see Appendix 2)
Number of Related Regional Deaths, Typical Decadal Summer
Climate-Related Mortality, Current vs. A2 Future Model Simulations 6'000
5'000
4'000
summer heat-related mortality summer O3-related mortality
3'000
2'000
1'000
0 1990s
2020s
2050s
2080s
Decade
Fig. 8.7 Evolution of projected trade-offs between heat- and O3-related climate impacts through the 2080s in the NYCHP study area
156
K. Knowlton et al.
that tended to inflate the effect of even modest increases in the absolute numbers of deaths. For O3, while “enhanced air pollution” occurred within the urban region as a whole owing to upwind precursor emissions as well as local vehicle and utility emissions, higher concentrations by the 2050s spread beyond the urban core into non-urban counties along the SW–NE prevailing wind directional axis (see Fig. 8.6). Figure 8.7 shows the projected evolution over time of comparative heat vs. O3 impacts. As the decades progress, temperature becomes far more important as a mortality effect than O3, in particular by the 2080s. In the 1990s, summer O3related mortality (estimated at 1,059) was on par with heat-related (1,116) in a typical summer, but by mid-century this could change with heat-mortality (2,347) approximately doubling as compared to the 1990s, while O3-mortality (1,108)
Table 8.2 Sensitivity analyses (a) comparing A2 vs. B2 greenhouse gas emission scenarios and their relative effects on summer mortality impacts by the 2050s; and (b) comparisons of mortality risk assessment projections made using newly developed transfer functions from the current epidemiological analysis vs. applying transfer functions from previous mortality health effects literature (Curriero et al. 2002, 2003 for heat; Thurston et al. 2001 for O3 as previously applied in Knowlton et al. 2004 for O3-mortality) (a) Comparing A2 vs. B2 heat- and O3-related mortality projections
Projected summer heat-related mortality
1990s
2050s B2 (lower CO2 emissions)
2050s A2 (higher CO2 emissions)
1,116
2,013
2,347
80% increase relative to 1990s 1,139
110% increase relative to 1990s 1,108
7.6% increase relative to 1990s
4.6% increase relative to 1990s
Projected summer O3- 1,059 related mortality
(b) Synthesizing combined A2 vs. B2 heat- and O3-related mortality projections, applying different transfer functions
Decade 1990s 2050a A2
Heat O3 Heat
O3 a
Mortality RA calculation applying new transfer function values
Mortality RA calculation applying previous transfer function valuesa
1,116 1,059 2,347 110% increase vs. 1990s (more than double) 1,108 4.6% increase vs. 1990s
1,418 1,275 2,764 95% increase vs. 1990s (nearly double) 1,335 4.7% increase vs. 1990s
Curriero et al. (2002, 2003) for heat and Thurston et al. (2001) for O3.
8 Impacts of Heat and Ozone on Mortality Risk in the New York City
157
increased by just 5%; and by the 2080s, heat-related mortality (5,533) could be over four times that from O3 (1,266). Sensitivity analyses compared B2 2050s mortality projections to A2 (Table 8.2a) and found approximately 27% fewer heat-related B2 deaths. These may represent potential savings in health impacts under GHG regulatory schemes. While larger O3-related mortality was projected for the New York metro region under the B2 scenario assumptions, different patterns across the eastern USA exist; domain-wide, O3 is projected to increase more under the 2050s A2 scenario than under B2. Compared to risk assessment results using previously published heat transfer functions (Curriero et al. 2002, 2003), our projections are broadly consistent and yield only slightly lower heat-mortality (Table 8.2b). Compared to Thurston and Ito (2001), our O3-mortality projections show no appreciable difference.
8.4 8.4.1
Discussion Policy Implications
This work represents an effort to more closely tailor projections of some possible health impacts of climate change for a specific regional population, in this case for the New York City metropolitan region. In the USA, health policy decisions (emergency planning, hospital surveillance, etc.) are often made by county health departments, so climate impact projections are likely to be most meaningful if framed at the county level. Furthermore, greenhouse gas emissions (GHGs) are not currently regulated at the federal level in the USA but regional initiatives are emerging among states, notably the Regional Greenhouse Gas Initiative (or RGGI) which would limit GHGs among a group of northeastern states (including these three) under a cap-and-trade scheme. If in the future the potential health impacts of climate change are monetized and become part of cost-benefit regulatory schemes, then risk assessments such as this could provide information useful not only to public health care infrastructure planning but also to regulators and legislative policymakers. Utility to policymakers is one of the explicit goals of integrated impacts assessments, and this project has met with local success via a request to submit testimony regarding a proposed New York City Council local law to regulate municipal GHG emissions. Acclimatization over the course of one season or over multiple years is possible as warming trends continue, yet this has not been accounted for in this study. Recent work suggests that not all communities have equal access to A/C based on race (O’Neill et al. 2005) and socioeconomic status (O’Neill et al. 2003), thus one cannot assure that A/C as an adaptation will negate summer heat stress equitably across the population. This type of risk assessment may help target those geographic communities most in need of alternative adaptive measures for heat stress and O3.
158
8.4.2
K. Knowlton et al.
Future Research Areas
For this initial analysis we chose to use the unstratified model results for the risk assessment. We acknowledge that ignoring this interaction may have underestimated O3 effects on mortality, especially at low temperatures. However, to integrate the interactive effects into the risk assessment, a more extensive analysis of interaction terms in modeling the observed 1990s data would have been necessary. That will be the focus of future work, which will test formal interaction models and, if significant, incorporate these findings into the mortality risk assessment. In the future, the inclusion of other alternative global and regional climate models, as in the “ensembles” concept, would help articulate a fuller range of potential impacts, given a wider set of input assumptions. Furthermore, it would be useful to develop PM2.5 estimates using the integrated modeling system, since fine particulates are the most health-damaging component of particulate matter (PM). We would like to include Apparent Temperature (AT) calculated from dewpoint and temperature data into the statistical modeling in order to include humidity effects on mortality. Other health outcomes in addition to mortality, in particular respiratory and cardiovascular hospitalizations, would be valuable additions to the risk assessment projection efforts. Inclusion of an adaptation module or some alternative analysis like the “analog summer” to re-evaluate potential population responses in the hottest summers (Hayhoe et al. 2004) would allow calculation of possible alternative transfer function values that could then be re-applied in the risk assessment. Lastly, in order for these suggested results to find their way to a broader stakeholder audience it will become increasingly important to continue to get more people involved in the training and funding of climate impacts professionals and the communication of their research results.
8.5
Summary
This work describes how location-specific projections of heat and O3-related transfer functions and deaths associated with changing climate have been developed for the New York City metro area. Both temperature and O3 were associated with daily deaths when included simultaneously in a Poisson GAM regression model. A dynamically downscaled global-to-regional climate/air quality modeling system was developed to estimate temperature and O3 in future decades at 36 km model resolution. Risk assessment then revealed the geographic distribution of environmental impacts differed for temperature and O3. For this study area, the relative mortality impact of temperature was projected to increase over the course of the 21st century, relative to the O3-mortality effect. The downscaled GISS-MM5 regional model system can distinguish regional variations at the county level and how those patterns interact with population density variations (i.e. size of exposed population) across the metro region.
8 Impacts of Heat and Ozone on Mortality Risk in the New York City
159
Acknowledgements This work was supported by the US Environmental Protection Agency under Science To Achieve Results (STAR) grant R828733. Additional support was provided by National Institute of Environmental Health Sciences Center grant ES09089 and from the National Aeronautics and Space Administration/Goddard Institute for Space Studies Climate Impacts Group. Although the research described in this article has been funded wholly or in part by the US EPA, it has not been subjected to the Agency’s required peer and policy review and therefore does not necessarily reflect the views of the Agency and no official endorsement should be inferred.
References Basu R, Samet JM. 2002. Relation between elevated ambient temperature and mortality: a review of the epidemiologic evidence. Epidemiol Rev 24:190–202. Basu R, Dominici F, Samet JM. 2005. Temperature and mortality among the elderly in the United States: a comparison of epidemiologic methods. Epidemiology 16:58–66. Bell ML, McDermott A, Zeger SL, Samet JM, Dominici F. 2004. Ozone and short-term mortality in 95 US urban communities, 1987–2000. JAMA 292:2372–2378. Bell ML, Dominici F, Samet JM. 2005. A meta-analysis of time-series studies of ozone and mortality with comparison to the National Morbidity, Mortality, and Air Pollution Study. Epidemiology 16:436–445. Byun DW, Ching JKS (eds.). 1999. Science algorithms of the EPA Models-3 Community Multiscale Air Quality (CMAQ) modeling system. EPA/600/R-99/030, US Environmental Protection Agency, Office of Research and Development, Washington, DC 20460. Curriero FC, Heiner KS, Samet JM, Zeger SL, Strug L, Patz JA. 2002. Temperature and mortality in 11 cities of the eastern United States. Am J Epidemiol 155: 80–87. Curriero FC, Samet JM, Zeger SL. 2003. Re: “On the use of generalized additive models in timeseries studies of air pollution and health” and “Temperature and mortality in 11 cities of the eastern United States” (Letter to the Editor). Am J Epidemiol 158:93–94. Dudhia J. 1993. A nonhydrostatic version of the Penn State/NCAR mesoscale model: Validation tests and simulation of an Atlantic cyclone and cold front. Mon Wea Rev 121:1493–1513. Grell GA, Dudhia J, Stauffer D. 1994. A description of the fifth-generation Penn State/NCAR Mesoscale Model (MM5). NCAR Technical Note, 138 pp., TN-398 + STR, National Center for Atmospheric Research, Boulder, CO. Hansen J, Sato M, Nazarenko L, Ruedy R, Lacis A, Hall T, Shindell D, Santer B, Stone P, Novakov T, Thomason L, Wang R, Wang Y, Jacob D, Hollandsworth S, Bishop L,, Logan J, Thompson A, Stolarski R, Lean J, Willson R, Levitus S, Antonov J, Rayner N, Parker D, Christy J. 2002. Climate forcing in Goddard Institute for Space Studies SI2000 simulations. J Geophys Res-Atmos 107:4347, doi:10.1029/2001JD001143]. Hajat S, Kovats RS, Atkinson RW, Haines A. 2002. Impact of hot temperatures on death in London: a time series approach. J Epidemiol Community Health 56:367–372. Hayhoe K, Cayan D, Field CB, Frumhoff PC, Maurer EP, Miller NL, Moser SC, Schneider SH, Cahill KN, Cleland EE, Dale L, Drapek R, Hanemann RM, Kalkstein LS, Lenihan J, Lunch CK, Neilson RP, Sheridan SC, Verville JH. 2004. Emissions pathways, climate change, and impacts on California. Proc Natl Acad Sci USA 101(34):12422–12427 (August 24, 2004), www.pnas.org/cgi/doi/10.1073/pnas.0404500101 Hogrefe C, Lynn B, Civerolo K, Ku J-Y, Rosenthal J, Rosenzweig C, Goldberg R, Gaffin S, Knowlton K, Kinney PL. 2004a. Simulating changes in regional air pollution over the eastern United States due to changes in global and regional climate and emissions. J Geophys Res 109, D22301, doi:10.1029/2004JD004690. Hogrefe C, Biswas J, Lynn B, Civerolo K, Ku J-Y, Rosenthal J, Rosenzweig C, Goldberg R, Kinney PL. 2004b. Simulating regional-scale ozone climatology over the eastern United States: model evaluation results. Atmos Env 38:2627–2638.
160
K. Knowlton et al.
Houynoux MR, Vukovich JM, Coats Jr., CJ, Wheeler NJM, Kasibhatta P. 2000. Emission inventory development and processing for the seasonal model for regional air quality. J Geophys Res-Atmos 105:9079–9090. Intergovernmental Panel on Climate Change (IPCC) 2000. Special Report on Emissions Scenarios (SRES). Nakicenovic N, Swart R (eds.). Cambridge University Press, Cambridge, 612 pp. Ito K, DeLeon SF, Lippmann M. 2005. Associations between ozone and daily mortality: analysis and meta-analysis. Epidemiology 16:446–457. Katsouyanni K, Pantazopoulou A, Touloumi G, Tselepidaki I, Moustris K, Asimakopoulous D, Poulopoulou G, Trichopoulous D. 1993. Evidence for interaction between air pollution and high temperature in the causation of excess mortality. Arch Environ Health 48:235–242. Kinney PL, Rosenthal JE, Rosenzweig C, Hogrefe C, Solecki W, Knowlton K, Small C, Lynn B, Civerolo K, Ku J-Y, Goldberg R, Oliveri C. 2006. Assessing potential public health impacts of changing climate and land uses: the New York Climate and Health Project. In: Ruth M, Donaghy K, Kirshen P (eds.), Regional Climate Change and Variability: Impacts and Responses. Cheltenham/Northampton, MA: Edward Elgar, pp. 161–189. Kinney PL, Ozkaynak H. 1991. Associations of daily mortality and air pollution in Los Angeles County. Environ Res 54:99–120. Kinney PL, Knowlton K, Hogrefe C. 2005. Ozone: Kinney et al. Respond to Schwarty J, Michacls P, Davis RE, Ozone: Unrealistic scenarios. Environ Health Perspect 113:A86–87. Kleinman LI, Lipfert FW. 1996. Metropolitan New York in the Greenhouse: Air Quality and Health. In: Hill D (ed.), The Baked Apple? Metropolitan New York in the Greenhouse Ann NY Acad Sci 790:91–110. Knowlton K, Rosenthal JE, Hogrefe C, Lynn B, Gaffin S, Goldberg R, Rosenzweig C, Civerolo K, Ku J-Y, Kinney PL. 2004. Assessing ozone-related health impacts under a changing climate. Environ Health Perspect 112:1557–1563. Levy JI, Chemerynski SM, Sarnat JA. 2005. Ozone exposure and mortality: an empiric Bayes metaregression analysis. Epidemiology 16:458–468. Lynn BH, Druyan L, Hogrefe C, Dudhia J, Rosenzweig C, Goldberg R, Rind D, Healy R, Rosenthal J, Kinney PL. 2004. On the sensitivity of present and future surface temperatures to precipitation characteristics. Climate Res 28:53–65. Lynn BH, Rosenzweig C, Goldberg R, Hogrefe C, Rind D, Healy R, Dudhia J, Biswas J, Druyan L, Rosenthal J, Kinney PL. 2006. The GISS-MM5 regional climate modeling system. Part I: Sensitivity of simulated current and future climate to model configuration. J Appl Meteorol (in revision). O’Neill MS, Zanobetti A, Schwartz J. 2003. Modifiers of the temperature and mortality association in seven US cities. Am J Epidemiol 157:1074–1082. O’Neill MS, Zanobetti A, Schwartz J. 2005. Disparities by race in heat-related mortality in four US cities: the role of air conditioning prevalence. J Urban Health 82:191–197. Russell GL, Miller JR, Rind D. 1995. A coupled atmosphere-ocean model for transient climate change studies. Atmos-Ocean 33:683–730. SAS Institute. 2002. SAS System 9 for Windows. Cary, NC: SAS Institute. Seinfeld JH, Pandis SN. 1998. Atmospheric Chemistry and Physics: From Air Pollution To Climate Change. New York: Wiley. Semenza JC, Rubin CH, Falter KH, Selaniko JD, Flanders WD, Howe HL, Wilhelm JL. 1996. Heat-related death during the July 1995 heat wave in Chicago. New Engl J Med 335: 84–90. Thurston GD, Ito K. 2001. Epidemiological studies of acute ozone exposures and mortality. J Exp Anal Environ Epidemiol 11:286–294. US Environmental Protection Agency (US EPA). 1997. The benefits and costs of the Clean Air Act: 1970–1990 (EPA 410-R-97-002) [http://www.epa.gov/air/sect812/copy.html] (Appendix I).
Chapter 9
Physiological Equivalent Temperature as Indicator for Impacts of Climate Change on Thermal Comfort of Humans Andreas Matzarakis1 and Bas Amelung2
Abstract Using the measure of Physiologically Equivalent Temperature (PET) it is analysed how changes in the thermal environment can affect human well-being. Historical data used in this study have been assembled for the normal climate period 1961–1990 (CNTRL). Future conditions are calculated based on the period 2071–2100, for which simulated datasets are available, based on GCMs integrated with scenarios. The scenarios used here are the Intergovermental Panel on Climate Change (IPCC) second report on emission scenarios (SRES) A1F and B1A, which represent a worst and a moderate climate case. The results are shown for December, January and February (DJF, winter months in the northern hemisphere), and for June, July and August (JJA, summer months in the northern hemisphere). Areas with extreme and uncomfortable thermal conditions and heat stress affections can be identified. In many regions of the world, e.g. the Mediterranean and North America, changes in thermal perception by humans are shown to outpace changes in air temperature. This has major implications for the assessment of the health effects of climate change. It is highly likely that the effects of climate change on human health and well-being have been underestimated in past studies, because these were based on air temperature changes rather than changes in PET, which describes the effects of meteorological and thermo-physiological parameters.
Keywords Thermal comfort, physiologically Equivalent Temperature, Heat Stress
9.1
Introduction
Throughout the 21st century, air temperature will continue to rise, according to computer simulations performed with global circulation models (GCMs). In addition to air temperature, the output of these GCMs includes a range of climate variables,
1
Meteorological institute, University of Freiburg, Freiburg, Germany
2
International Centre for Integrative Studies, University of Maastricht
M.C. Thomson et al. (eds.), Seasonal Forecasts, Climatic Change and Human Health. © Springer Science + Business Media B.V. 2008
161
162
A. Matzarakis and B. Amelung
such as air humidity, wind speed and cloud cover. Based on these variables, further analysis of thermal comfort and the impacts of extreme heat conditions on humans can be undertaken. Humans have always been aware that weather and climate affect their health and well being. Two thousand five hundred years ago, Hippocrates described regional differences of climate and their relationship to states of health. Fevers vary seasonally and so do people’s moods and various psychological disturbances. Aches and pains in joints flare up in winter, while in summer heat waves debilitate and kill (World Meteorological Organisation 1999). Locations with extreme heat conditions may also result in health problems (e.g. caused by heat stress, UV-radiation, air pollution and heat strokes). Cause–effect relations between the atmospheric environment on the one hand, and human health and comfort on the other can be analysed with a human-biometeorological classification that takes into consideration: – The thermal complex (comprises the meteorological elements which have a thermophysiological effect on humans) – The air pollution complex (comprises solid, liquid and gaseous natural and anthropogenic air pollutants which have an effect on human health) – The actinic complex (comprises the visible and ultraviolet spectrum of solar radiation which has a direct biological effect) In this analysis, only the thermal complex is considered. It includes the meteorological factors air temperature, air humidity, and wind velocity, as well as short and long wave radiation, which affect humans thermo-physiologically in indoor and outdoor climates. This thermal complex is relevant to human health because of a close relationship between the thermoregulatory mechanisms and the circulatory system. Effects of the thermal environment of humans are best determined with the aid of thermal indices based on the energy balance of the human body (Verein Deutscher Ingenieure 1998). Common applications are PMV (Predicted Mean Vote) (Fanger 1972), PET (Physiologically Equivalent Temperature) (VDI 1998; Höppe 1999; Matzarakis et al. 1999), SET* (Standard Effective Temperature) (Gagge et al. 1986) or Outdoor Standard Effective Temperature (Out_SET*) (Spagnolo and de Dear 2003) and Perceived Temperature (Tinz and Jendritzky 2003). These well-documented thermal indices have varying foci, but are essentially different combinations of the same set of important meteorological and thermophysiological parameters (Matzarakis 2001). Unfortunately, data on several of these parameters, such as short and long wave radiation, are generally not available in climate records. As a result, climate assessments and thermal comfort studies have often resorted to the use of climate indices that do not include these key factors. For example, the Intergovermental Panel on Climate Change report (IPCC 2001) describes the effect of weather and climate on humans with a simple index based on a combination of air temperature and relative humidity. The exclusion of important meteorological (wind speed and radiation fluxes) and thermo-physiological (activity of humans and clothing) variables seriously diminishes the significance of the results. From synoptic, climatological and astronomical data, estimates for short and long wave radiation fluxes can be obtained (Verein Deutscher Ingenieure 1998; Matzarakis et al. 2000). These estimates
9 Physiological Equivalent Temperature as Indicator for Impacts of Climate
163
are used in this paper to explore the effects of climate change for the thermal environment of humans around the world. The objective of this article is twofold: (1) to give a brief overview of the assessment methods for human bioclimate and (2) to discuss some exemplary results indicating the current quality of human bioclimate at the end of the 21st century.
9.2
Data
Scenarios were used to determine the meteorological parameters needed to predict PET values. Future climatic conditions cannot be predicted with any degree of certainty, as several unpredictable factors are involved. Future socio-economical and technological developments will determine to a large extent the amount of human-induced emissions of greenhouse gases. To get a feeling for the range of possible climate conditions that may be common by the end of the century, a range of scenarios developed by the Intergovernmental Panel on Climate Change (IPCC) were used. IPCC undertook an exploration of the possible changes in socio-economical conditions and population (IPCC 2000, 2001), which resulted in a range of plausible scenarios (known as the SRES scenarios). From these, Greenhouse gases (GHG) emissions and atmospheric concentrations of greenhouse gases could be estimated, which in turn have been used to explore the response of the climate system. Among the four main SRES scenarios, the A1F and A2A represent cases of rapid climate change, while the B1A and B2A scenarios represent more moderate levels of change (Table 9.1). The historical data that were used have been retrieved from the CRU CL 1.0 historical dataset, containing 0.5° 1961–1990 mean monthly gridded data, assembled by the Climatic Research Unit, University of East Anglia, Norwich, UK (New et al. 1999, 2000). The CRU CL 1.0 grid-based dataset is based on a dataset of 1961–1990 climatological normals, which was produced by numerous weather stations around the world. The station data were interpolated to obtain a 0.5° latitude × 0.5° long grid-based dataset, covering the entire landmass of the earth except Antarctica; ocean space is not included. From the historical data sets the mean monthly data of air temperature, relative humidity and wind speed of each grid of the globe have been used. For the calculation of the global radiation the mean monthly sunshine fraction has been manipulated to cloud cover. The way of production
Table 9.1 Description of used emission scenarios for PET-calculations Scenario Description A1F A2A B1A B2A
A world of rapid economic growth and rapid introductions of new and more efficient technologies A very heterogenous world with an emphasis on family values and local traditions A world of “dematerialization”and introduction of clean technologies A world with an emphasis on local solutions to economic and environmental sustainability
164
A. Matzarakis and B. Amelung
of the grid data and their uncertainties of the climate variables are described in New et al. (1999). The mean radiant temperature of each grid of the globe has been calculated based on the possible global radiation and the mean monthly cloud cover by the RayMan model. The mean radiant temperature and the PET can be calculated in one run with the RayMan model based on input parameters (air temperature, relative humidity, wind speed and cloud cover). The dataset of future climatic conditions was based on an integration of the Hadley Centre’s HadCM3 model forced with the SRES emissions scenarios (Johns et al. 2003). The HadCM3 model produces gridded data with a spatial resolution of 2.5° latitude × 3.75° longitude, which is significantly coarser than that of the CRU 1.0 dataset. The used HadCM3 dataset consists of monthly averages for four time slices: 1961–1990, 2010–2039, 2040–2069, and 2070–2099. The uncertainties and difficulties of the climate projections data are described in Amelung (2006) and Hulme et al. (2002). All variables that were needed for the analysis of PET were available from the CRU 1.0 and HadCM3 datasets (air temperature, relative humidity and wind speed) or could be calculated from them (mean radiant temperature). The procedure of calculation of PET for the scenarios is the same as for the historical data sets.
9.3
Methods
Since the 1960s, heat balance models of the human body have become more and more accepted in the assessment of thermal comfort. The basis for these models is the human energy balance equation. One of the first and still very popular heat balance models is the comfort equation defined by Fanger (1972). Fanger introduced the thermal indices “Predicted Mean Vote” (PMV) and “Predicted Percentage Dissatisfied” (PPD) to help air-conditioning engineers create thermally comfortable indoor climates. Two decades later, Jendritzky et al. (1990) managed to make Fanger’s approach applicable to outdoor conditions by assigning appropriate parameters to adjust the model the much more complex outdoor radiation conditions. This approach, which is also known as the “Klima Michel Model”, is now increasingly being applied. Since this model was designed only to estimate an integral index for the thermal component of climate and not to represent a realistic description of thermal body conditions, it is able to work without the consideration of fundamental thermo-physiological regulatory processes. For example, in Fanger’s approach the mean skin temperature and sweat rate are quantified as “comfort values”, being only dependent on activity and not on climatic conditions (Höppe 1999). More universally applicable models take into account all basic thermoregulatory processes, like the constriction or dilation of peripheral blood vessels and the physiological sweat rate (Höppe 1993, 1999). They enable the user to predict “real values” of thermal quantities of the body, i.e. skin temperature, core temperature, sweat rate or skin wetness. The Munich energy balance model for individuals” (MEMI) (Höppe 1993) is such a thermo-physiological heat balance
9 Physiological Equivalent Temperature as Indicator for Impacts of Climate
165
model. It is the basis for the calculation of the physiologically equivalent temperature (PET). In detail the MEMI model is based on the energy balance equation (9.1) for the human body: M + W + R + C + ED + ERe + ESw + S = 0
(9.1)
Where, M the metabolic rate (internal energy production), W the physical work output, R the net radiation of the body, C the convective heat flow, ED the latent heat flow to evaporate water diffusing through the skin (imperceptible perspiration), ERe the sum of heat flows for heating and humidifying the inspired air, ESw the heat flow due to evaporation of sweat, and S the storage heat flow for heating or cooling the body mass. The individual terms in this equation have positive signs if they result in an energy gain for the body and negative signs in the case of an energy loss (M is always positive; W, ED and Esw are always negative). The unit of all heat flows is in Watt (Höppe 1999). The individual heat flows in Eq. 9.1, are controlled by the following meteorological parameters (Verein Deutscher Ingenieure 1998; Höppe 1999): – – – –
Air temperature: C, ERe Air humidity: ED, ERe, ESw Wind velocity: C, ESw Mean radiant temperature: R
Thermo-physiological parameters are required in addition: – Heat resistance of clothing (clo units) – Activity of humans (in Watt) The human body does not have any selective sensors for the perception of individual climatic parameters, but can only register (by thermoreceptors) and make a thermoregulatory response to the temperature (and any changes) of the skin and blood flow passing the hypothalamus (Höppe 1993, 1999). These temperatures, however, are influenced by the integrated effect of all climatic parameters, which are in some kind of interrelation, i.e. affect each other. In weather situations with less wind speed, for instance, the mean radiant temperature has roughly the same importance for the heat balance of the human body as the air temperature. At days with higher wind speeds, air temperature is more important than the mean radiant temperature because it dominates now the increased enhanced convective heat exchange. These interactions are only quantifiable in a realistic way by means of heat balance models (Verein Deutscher Ingenieure 1998; Höppe 1999). PET is defined to be equivalent to the air temperature that is required to reproduce in a standardised indoor setting and for a standardised person the core and skin temperatures that are observed under the conditions being assessed (Verein Deutscher Ingenieure 1998; Höppe 1999). The standardised person is characterised by a work metabolism of 80 W of light activity, in addition to basic metabolism; and by 0.9 clo of heat resistance as a result of clothing.
166
A. Matzarakis and B. Amelung
The following assumptions are made for the indoor reference climate: – Mean radiant temperature equals air temperature (Tmrt = Ta). – Air velocity (wind speed) is fixed at v = 0.1 m/s. – Water vapour pressure is set to 12 hPa (approximately equivalent to a relative humidity of 50% at Ta = 20°C). The calculation of PET includes the following steps: – Calculation of the thermal conditions of the body with MEMI for a given combination of meteorological parameters. – Insertion of the calculated values for mean skin temperature and core temperature into the model MEMI and solving the energy balance equation system for the air temperature Ta (with v = 0.1 m/s, VP = 12 hPa and Tmrt = Ta). Finally the resulting air temperature is equivalent to PET. PET allows the evaluation of thermal conditions in a physiologically significant manner, too. With respect to this, Matzarakis and Mayer (1996) transferred ranges of PMV for thermal perception and grade of physiological stress on human beings (Fanger 1972) into corresponding PET ranges (Table 9.2). They are valid only for the assumed values of internal heat production and thermal resistance of the clothing. It is worth mentioning that the VDI-guideline 3787 part 2 “methods for the human-biometeorological evaluation of climate and air quality for urban and regional planning, part I: climate“(Verein Deutscher Ingenieure 1998) recommends the application of PET for the evaluation of the thermal component of different climates to emphasize the significance of PET more further. This guideline is edited by the German Association of Engineers (‘Verein Deutscher Ingenieure’ VDI).
Table 9.2 Ranges of the physiological equivalent temperature (PET) for different grades of thermal perception by human beings and physiological stress on human beings; internal heat production: 80 W, heat transfer resistance of the clothing: 0.9 clo (According to Matzarakis and Mayer 1996) PET 4°C 8°C 13°C 18°C 23°C 29°C 35°C 41°C
Thermal perception
Grade of physiological stress
Very cold
Extreme cold stress
Cold
Strong cold stress
Cool
Moderate cold stress
Slightly cool
Slight cold stress
Comfortable
No thermal stress
Slightly warm
Slight heat stress
Warm
Moderate heat stress
Hot
Strong heat stress
Very hot
Extreme heat stress
9 Physiological Equivalent Temperature as Indicator for Impacts of Climate
167
PET can be calculated with the radiation and bioclimate model RayMan, which is suitable for the calculation of the radiation fluxes and thermal indices a.e. PET in easy and complex environments (Matzarakis et al. 2000). RayMan includes the MEMI model and the calculation procedure for PET and is free available software.
9.4
Results
In order to have reference values for present climate conditions PET has been calculated from historical data sets. For climate projections have been used the scenarios results of HadCM3 model for the PET calculations. The analyses have been carried out for two seasons and two time slices (i.e. intervals). The time segments represent seasons consisting of the combined months of December, January, and February, and the combined months of June, July, and August, coinciding with the winter and summer seasons in the northern hemisphere and the other way round in the southern hemisphere. Analysis has been carried out for the historical period 1961–1990 (CNTRL) and the future period 2071–2100. The PET values have been calculated with the RayMan model (Matzarakis et al. 2000). Figure 9.1 shows the PET conditions for CNTRL (a, top) in the JJA season, and the expected changes according to the A1F scenario (b, middle panel) and the B1A scenario (c, bottom panel). The top panel can be taken as a proxy for actual bioclimatic conditions. A comparison of current and future conditions, projected by the scenarios, shows remarkable changes. The A1F projections show a shift towards warmer conditions in all regions of the world. Many parts of the world, including the Mediterranean and areas in North America show changes in PET values in excess of 10°C, which are much higher than the expected changes in air temperature for these regions. Especially in the Mediterranean and areas in North America the PET can increase more than 15°C, which corresponds to three levels of increased physiological strain for humans according to Table 9.2 classification. While PET values will increase in most areas, slight cooling will occur in a relatively small area around Gabon in Africa, and in a somewhat larger area around Burma and Thailand in Asia. As expected, the B1A projections show more moderate results, with PET conditions ranging between A1F and CNTRL. In some parts of the world (particularly in the southern hemisphere) the conditions will not change significantly and some small areas, i.e. Gabon will be have lower PET as the CNTRL conditions. Figure 9.2a (top panel) shows the PET conditions for the DJF season. In the A1F projections, the DJF PET values will be higher than they are in the current CNTRL situation (see Fig. 9.2b, middle panel). This holds for all the world’s regions, with the greatest changes occurring in the northern latitudes (sometimes in excess of 10°C) and the smallest changes taking place in the middle and lower latitudes. Analogous to the results from the JJA season, B1A projections indicate smaller changes in the DJF season than the A1F projections (Fig. 9.2c, bottom panel). Apart from the area close to the North Pole in which the level of change is
168
A. Matzarakis and B. Amelung
Fig. 9.1 (continued)
considerable, changes are moderate to small. A small area of West Africa will even experience slight cooling. In comparison to the maximum changes of PET the physiological strain value will be lower (more than one class) than the A1F projections, according to the classification of Table 9.2.
9 Physiological Equivalent Temperature as Indicator for Impacts of Climate
169
Fig. 9.1 (continued) Basic PET conditions (a) and differences between time slice (2070–2100) minus (1961–1990) for A1F (b) and B1A (c) for JJA (see Appendix 2)
Fig. 9.2 (continued)
170
A. Matzarakis and B. Amelung
Fig. 9.2 (continued) Basic PET conditions (a) and differences between time slice (2070–2100) minus (1961–1990) for A1F (b) and B1A (c) for DJF (see Appendix 2)
9 Physiological Equivalent Temperature as Indicator for Impacts of Climate
9.5
171
Conclusions
The main conclusion from this paper is that the thermal consequences on humans of climate change should have been underestimated. Changes in the overall bioclimatic conditions for humans are expected to be considerably greater than changes in air temperature alone. Changes in non-temperature factors such as short and long wave radiation appear to reinforce the first-order effects of temperature change. In most regions of the world, the projected climate change will produce bioclimatic conditions that are more stressful (a PET of more than 35 means extreme hot conditions for Europeans) to people and affect their health and well being. Regions with PET >35°C will be increase compared to present bioclimatic conditions and the possibility of heat waves will also increase. In addition, the changed thermal conditions will lead to higher energy consumption (and higher emissions of greenhouse gases) as a result of the increased need for cooling. The results presented in this paper have to be considered as a first approach. The analyses could be further elaborated by undertaking more detailed studies comparing different GCMs, including regional climate effects, extreme events (Heat waves) and expected future land use changes. It has to be mentioned that the uncertainties in the input data, which are included in the data derived from the climate scenarios have an influence in the thermal bioclimate conditions. These uncertainties a.e. an increase of air temperature of 2°C in thermal neutral conditions (air temperature 20°C) is associated with an increase of PET of 2.4°C. Nevertheless, the information and results in their current form are already very likely to be assisting in decision making on various levels, including health, tourism and regional planning. Acknowledgements Thanks to Nikola Sander for proofreading and editing the manuscript. Thanks to Markus Zygmuntowski for the global maps produced by IDL software.
References Amelung, B. (2006) Global (environmental) change and tourism. Issues of scale and distribution. Univeritaire Pers Maastricht. Fanger, P. O. (1972) Thermal comfort. McGraw-Hill, New York. Gagge, A. P., Fobelets, A. P., Berglund, L. G. (1986) A standard predictive index of human response to the thermal environment. ASHRAE Trans 92, 709–731. Höppe, P. (1993) Heat balance modelling. Experientia 49, 741–746. Höppe, P. (1999) The physiological equivalent temperature – a universal index for the biometeorological assessment of the thermal environment. Int J Biometeorol 43, 71–75. Hulme, M., Jenkins, G. J., Lu, X., Turnpenny, J. R., Mitchell, T. D., Jones, R. G., Lowe, J., Murphy, J. M., Hassell, D., Boorman, D., McDonald, R., Hill S. (2002) Climate change scenarios for the UK: UKCIP02 scientific report. Tyndall Centre, 112 pp. IPCC (2000) Emission scenarios: A special report of working group III of the Intergovernmental Panel on climate change. Cambridge University Press, Cambridge.
172
A. Matzarakis and B. Amelung
IPCC Climate Change (2001) The Scientific Basis: Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change. In: Houghton, J. T. et al. (eds). Cambridge University Press, Cambridge. Jendritzky, G., Menz, G., Schmidt-Kessen, W., Schirmer, H. (1990) Methodik zur räumlichen Bewertung der thermischen Komponente im Bioklima des Menschen, Akademie für Raumforschung und Landesplanung, Hannover. Johns, T. C., Gregory, J. M., Ingram, W. J., Johnson, C. E., Jones, A., Lowe, J. A., Mitchell, J. F. B., Roberts, D. L., Sexton, D. M. H., Stevenson, D. S., Tett, S. F. B., Woodage, M. J. (2003) Anthropogenic climate change for 1860 to 2100 simulated with the HadCM3 model under updated emission scenarios. Clim Dynam 20, 583–612. Matzarakis, A. (2001) Die thermische Komponente des Stadtklimas, Wiss. Ber. Meteorologisches Institut der Universität Freiburg No. 6. Matzarakis, A., Mayer, H., (1996) Another kind of environmental stress: Thermal stress. WHO collaborating centre for Air Quality Management and Air pollution Control. NEWSLETTERS 18, 7–10. Matzarakis, A., Mayer, H., Iziomon, M. (1999) Heat stress in Greece. Applications of a universal thermal index: physiological equivalent temperature. Int J Biometeorol 43, 76–84. Matzarakis, A., Rutz, F., Mayer, H. (2000) Estimation and calculation of the mean radiant temperature within urban structures. In: Biometeorology and Urban Climatology at the Turn of the Millenium. In: de Dear, R. J., Kalma, J. D., Oke T. R. Auliciems A. (eds). Selected Papers from the Conference ICB-ICUC’99, Sydney, WCASP-50, WMO/TD No. 1026, 273–278. New, M., Hulme, M., Jones, P. (1999) Representing twentieth centrury space-time climate variability. Part 1: development of a 1961–1990 mean monthly terrestrial climatology. J Climate 12, 829–856. New, M., Hulme, M., Jones, P. D. (2000) Representing twentieth century space-time climate variability. Part 2: development of 1901–96 monthly grids of terrestrial surface climate. J Climate 13, 2217–2238. Spagnolo, J., de Dear, R. (2003) A field study of thermal comfort in outdoor and semi-outdoor environments in subtropical Sydney Australia. Build Environ 38, 721–738. Tinz, B., Jendritzky, G. (2003) Europa- und Weltkarten der gefühlten Temperatur. In: Chmielewski, F.-M., Foken, Th. (eds.) Beiträge zur Klima- und Meeresforschung, Berlin und Bayreuth. pp. 111–123. Verein Deutscher Ingenieure (1998) VDI 3787, Part I: environmental meteorology, methods for the human-biometeorological evaluation of climate and air quality for the urban and regional planning at regional level. Part I: climate. VDI/DIN-Handbuch Reinhaltung der Luft, Band 1b, Düsseldorf, 29 pp. World Meteorological Organisation (1999) Climate and human health. World Climate News, 14, 3–5.
Chapter 10
Weather, Ambient Air Pollution and Bronchial Asthma in Athens, Greece Panagiotis Th. Nastos
Abstract The associations between various meteorological parameters, concentrations of PM10, SO2, and O3 pollutants and bronchial asthma of residents of the wider region of Athens are examined in this study. For this purpose, 1,288 patients’ admissions (412 males and 876 females) recorded in 13 pneumonological clinics of Sotiria Hospital, which is the major Hospital for respiratory diseases in Athens, were analyzed for the period 1/1/2001–31/12/2002. The meteorological data were available by the National Observatory of Athens and the concentrations of PM10, SO2, and O3 pollutants were obtained by the air pollution network of 10 stations of the Ministry of Environment (Directorate of Air and Noise pollution Control). The evaluation of the possible relationship between the bronchial asthma admissions (BAA) and the meteorological variables was achieved by the application of Generalized Linear Models with Poisson distribution, because the medical dataset presents large divergence from a Gaussian distribution. The results showed that a statistically significant (p < 0.01) negative correlation between all examined air temperature variables, water vapor pressure, evaporation, sunshine, total solar irradiance, and BAA exists. Moreover, the findings of the analysis showed that a statistically significant relationship between the examined pollutants and BAA on the same day does not exist, with the exception of O3, which is correlated negatively (p < 0.01) with BAA. Nevertheless, there is a statistically significant lag effect (7–8 days) between the increase in BAA and the peak in the concentrations of PM10, SO2, and O3 pollutants. During the cold period of the year (October–April), BAA is significantly associated with O3 lag 2 day. The interannual variation of bronchial asthma admissions (BAA) reveals peaks within the transitional seasons of the year (spring and autumn), while the main minimum is apparent during summer period and especially in August.
Keywords Weather variability, ambient airpollution, bronchial asthma, Generalized Linear Models, Athens, Greece
Laboratory of Climatology and Atmospheric Environment, University of Athens, Greece
M.C. Thomson et al. (eds.), Seasonal Forecasts, Climatic Change and Human Health. © Springer Science + Business Media B.V. 2008
173
174
10.1
P. Th. Nastos
Introduction
The influence of weather conditions and air pollution on health is a major focus of research. IPCC (Technical Summary, 2001) highlights that global climate change will have various impacts on human health; some of which are positive, but mostly negative. Changes in the frequencies of extreme heat and cold, the frequencies of floods and droughts, and the profile of local air pollution and aeroallergens would directly affect population health. Increased bronchial asthma admissions (BAA) is associated with low air temperature (Greenburg et al. 1966; Yuksel et al. 1996; Grech et al. 2002; Bartzokas et al. 2003). Altitude and the annual variation of temperature and relative humidity outdoors were negatively associated with asthma symptoms (Weiland et al. 2004). Admissions to emergency room for asthma count were negatively correlated with ambient temperature and strong wind existence on previous days. They were also positively correlated with ambient relative humidity (Berktas and Bircan 2003). Hashimoto et al. (2004) suggest that childhood asthma increases when climate conditions show a rapid decrease from higher barometric pressure, from higher air temperature and from higher humidity, as well as lower wind speed while the presence of mist and fog causes the exacerbation of asthma in children (Kashiwabara et al. 2002). Furthermore, a study by Goldstein (1980) showed that almost every asthma epidemic in both New Orleans and New York City was preceded by the passage of a cold front (by 1–3 days) followed by a high pressure system. Several recent studies have reported associations between ambient pollutants and BAA (Bascom 1996; Delfino et al. 1997; Jaffe et al. 2003; Trasande and Thurston 2005). More specifically, Dockery and Pope (1994) have found 3.4% increase in emergency department visits for asthmatics per 10 µg/m3 PM10 and 1.9% increase in hospital admissions for asthmatic attacks per 10 µg/m3 PM10, while the daily counts of emergency room visits were significantly associated with PM10 exposure on the previous day (Schwartz et al. 1993) or on the previous 3 days (Galán et al. 2003). Besides, Walters et al. (1994) showed that during winter a rise of 100 µg/m3 in SO2 might result in four more asthma admissions each day, for residents of Birmingham. With respect to ozone, positive and statistically significant associations were found between hospital respiratory admissions and both ozone recorded on the day of admission and up to 3 days prior to the date of admission (Burnett et al. 1994; Galán et al. 2003). On the contrary, many researchers did not find any significant associations between ambient ozone concentrations and doctor asthma visits (Jalaludin et al. 2004; Braun-Fahrlander et al. 1992; Hoek and Brunekreef 1995; Ostro et al. 1999). The goal of this study is to examine whether an association between BAA, meteorological variables and the concentrations of SO2, O3 and PM10 pollutants could be observed in Athens.
10 Weather, Ambient Air Pollution and Bronchial Asthma in Athens, Greece
10.2
175
Data and Analysis
The medical dataset analyzed consists of 1,288 patients’ admissions recorded in 13 pneumonological clinics of Sotiria Hospital, which is the major hospital for respiratory diseases in Athens, were analyzed for the period 1/1/2001–31/12/2002. The age distribution of the patients is depicted in Fig. 10.1. The children and the adolescents comprise the minority of the recorded patients while the elderly people (age > 70 years) dominate in the upper level of the BAA. Therefore, there is a remarkable plateau in the age distribution regarding the registered patients with age from 20 to 70 years. The number of females are twofold the number of males within each age class after 40 years. There has not been found any reason for this evidence, possibly it is due to females’ sensitivity and modern way of living. The meteorological data were available by the National Observatory of Athens and concern daily values of: – – – – – – – – – –
Mean Air Temperature (Tmean, °C) Day-to-Day Change of Mean Air Temperature (∆Tmean, °C) Maximum Air Temperature (Tmax, °C) Day-to-Day Change of Maximum Air Temperature (∆Tmax, °C) Minimum Air Temperature (Tmin, °C) Day-to-Day Change of Minimum Air Temperature (∆Tmin, °C) Diurnal Air temperature Range (Trange, °C) Day-to-Day Change of Diurnal Air temperature Range (∆Trange, °C) Mean Water Vapor Pressure (e, mm Hg) Day-to-Day Change of Mean Water Vapor Pressure (∆e, mm Hg) 280 260
Bronchial asthma admissions
240 220
Females Males
200 180 160 140 120 100 80 60 40 20 0 <10
[10, 20) [20, 30) [30, 40) [40, 50) [50, 60) [60, 70)
Years
Fig. 10.1 Age distribution of the patients with bronchial asthma in Athens
>70
176
– – – – – – – –
P. Th. Nastos
Mean Atmospheric Pressure (P, hPa) Day-to-Day Change of Mean Atmospheric Pressure (∆P, hPa) Total Evaporation (E, mm) Day-to-Day Change of Total Evaporation (∆E, mm) Total Sunshine (S, h) Day-to-Day Change of Total Sunshine (∆S, h) Mean Total Solar Irradiance (I, W m−2) Day-to-Day Change of Mean Total Solar Irradiance (∆I, W m−2)
The air pollution data measured by the air pollution network of 10 stations of the Ministry of Environment (Directorate of Air and Noise pollution Control) include daily values of: – Mean Concentration of Particulate Matter diameter <10 µm (PM10 mean, µg m−3) – Day-to-Day Change of Mean Concentration of Particulate Matter diameter <10 µm (∆PM10 mean, µg m−3) – Maximum Concentration of Particulate Matter diameter < 10 µm (PM10 max, µg m−3) – Day-to-Day Change of Maximum Concentration of Particulate Matter diameter <10 µm (∆PM10 max, µg m−3) – Mean Concentration of Sulphur Dioxide (SO2 mean, µg m−3) – Day-to-Day Change of Mean Concentration of Sulphur Dioxide (∆SO2 mean, µg m−3) – Maximum Concentration of Sulphur Dioxide (SO2 max, µg m−3) – Day-to-Day Change of Maximum Concentration of Sulphur Dioxide (SO2 max, µg m−3) – Mean Concentration of Ozone (O3 mean, µg m−3) – Day-to-Day Change of Mean Concentration of Ozone (∆O3 mean, µg m−3) – Maximum Concentration of Ozone (O3 max, µg m−3) – Day-to-Day Change of Maximum Concentration of Ozone (∆O3 max, µg m−3) In order to determine the relationship between the admissions of BAA and the aforementioned meteorological and ambient air pollution variables, Generalized Linear Models (GLM), described by McGullagh and Nelder (1997), were applied. The class of models known as Generalized Linear Models, or GLMs, was formally introduced by Nedler and Wedderburn (1972). The components of a GLM are as follows: – The aim is to model the distribution of a stochastic response variable, y, in terms of stimulus variables x1, x2, …, xp, or known mathematical functions of them. – The distribution of y depends on the stimulus variables through a single linear predictor:
∑
p j =1
x j b j where, in general, the xj,s are known mathematical function
functions of the stimulus variables, not necessarily simply the variables themselves. – The mean of y is related to n by a known function called the link function: E[y] = m = l−1 (n), n = l(m).
10 Weather, Ambient Air Pollution and Bronchial Asthma in Athens, Greece
177
Note that the link function transforms the mean into the linear predictor and not the other way round. Hence it acts in the same direction as a transformation of the response itself, from which the idea arose. – The variance of y is a function of the mean: Var[y] = ju (m)/A where j is a possibly unknown, positive scale parameter, A is a known prior weight, and u (m) is a known function of m called the variance function. – The distribution of y has a density of known form, namely ⎡A ⎛ f ⎞⎤ fY ( y; m , j ) = exp ⎢ { yq ( m ) − g (q ( m ))} + t ⎜ y, ⎟ ⎥ ⎝ A⎠ ⎦ f ⎣ This distributional form can be shown to include the normal, gamma, Poisson and binomial distributions, as well as several others such as beta, inverse Gaussian and negative binomial. Note that the relationship between the canonical parameter q, and the mean, m, will depend on the particular distribution, and the relationship between m and n is defined by the link function. The link function establishes the connection between the linear predictor, n, and the mean of the distribution m. There is a ‘natural link’ for each distribution. It is important to note that although the link function is in some way similar to a transformation function, it only establishes a mathematical connection between the mean and the response variables. A transformation function when applied to observations may be intended to simplify the connection between the mean and the response variables. It may also achieve other goals such as to stabilize the variance. The natural link for the Poisson distribution is the log link: n = log(m), m = en, the variance function is u (m) = m and as in the case of the binomial distribution, the scale parameter is 1. Poisson models with log links are often called log-linear models and are used for frequency data. The evaluation of the possible relationships between meteorological parameters, concentrations of PM10, SO2, O3 pollutants and BAA was achieved by the application of Generalized Linear Models with Poisson distribution, because the medical dataset presents large divergence from a Gaussian distribution. In the models fitting procedure the dependent variable was the daily totals of the BAA, filed by the Hospital and the independent covariates were the aforementioned meteorological parameters and ambient air pollutants. Models’ goodness-of-fit was evaluated through deviance residuals (McGullagh and Nelder 1997).
10.3
Results and Discussion
Table 10.1 depicts the mean and maximum daily values of the pollutants’ concentrations and the meteorological parameters examined. With respect to the Air Quality Guidelines of the World Health Organization (WHO2000) [Ozone: 120 µg/m3 per 8 h, SO2: 125 µg/m3 per 24 h and 50 µg/m3 per year, PM10: no guideline values were set for particulate matter because there is no evident threshold for effects on
178
P. Th. Nastos
morbidity and mortality], there were some days with exceeding values for ozone, mainly during the summer time. This is mainly due to the geographical location and to the climatological conditions (high sunshine and air temperature favor the O3 formation). Nevertheless, the national threshold for undertaking extraordinary measures (360 µg/m3 in 1 h) was not exceeded in any station of the network. According to the WHO and national thresholds for the pollutants, there are no exceedances for the SO2 concentrations during the examined period, in any station of the network. There is an important decrease trend of the SO2 concentrations, which is associated with the reduction of the sulfur content in diesel petroleum since 2000. The particulate matter PM10 is one of the new pollutants, which began to be measured recently in the European Union, and constitute an important problem for the majority of the countries, especially the southern ones of the European Union. This pollutant is considered being at high levels. The national threshold is 70 µg/m3 averaged over 24 h and for 35 days within the year. The mean daily PM10 concentrations of all the stations of the air pollution network exceeded the national threshold for 186 days during the 2 examined years. Furthermore, the Environmental Protection Agency (EPA) uses its Air Quality Index (EPA 2006) to provide general information to the public about air quality and associated health effects. An Air Quality Index (AQI) of 100 for PM10 corresponds to a PM10 level of 150 µg/m3 averaged over 24 h, and people with respiratory disease, such as asthma, should limit outdoor exertion. The annual evolution of the pollutants concentrations is illustrated in Fig. 10.2. The primary pollutant SO2 show high values during the cold period of the year, and this is mainly due to central heating of the buildings, while O3 presents high concentrations during summer time, because of the increased sunshine and high air 80
PM10 SO2 O3
Concentration of pollutants (µg/m3)
70 60 50 40 30 20 10 0 1
2
3
4
5
6
7
8
9
10
11
12
Months
Fig. 10.2 Mean annual evolution of the pollutants concentrations, for the period 2001–2002, in Athens
10 Weather, Ambient Air Pollution and Bronchial Asthma in Athens, Greece
179
temperature, which play a determinative role in the photochemical procedures. Particulate matter includes both solid particles and liquid droplets found in air and many man-made and natural sources produce PM10 directly or emit other pollutants that react in the atmosphere to form PM10. There is no remarkable seasonal variation because the sources emitted PM10 exist all over the year. The Generalized Linear Models were applied individually to every considered variable, in such a way that the dependent variable, within the constructed univariate model, was the daily totals of the BAA and the independent variable was each meteorological or pollutant variable. The results (Table 10.2) showed that a statistically significant (p < 0.01) negative correlation exists between all examined air temperature parameters, water vapor pressure, evaporation, sunshine, total solar irradiance, and BAA (shaded parameters in Table 10.2). No statistically significant correlations exist between the BAA and the day-to-day changes of the meteorological parameters. More specifically, on the same day, a decrease of 10 units in daily mean, maximum and minimum air temperature, diurnal air temperature range, mean water vapor pressure, total evaporation, total sunshine and total solar irradiance links to an increase 14%, 12%, 16%, 19%, 26%, 38%, 18% and 0.9% of the probability having BAA, respectively. These results are in agreement with the findings of other researchers (Yuksel et al. 1996; Grech et al. 2002; Berktas and Bircan 2003; Hashimoto et al. 2004; Hajat et al. 2004). A brief description of the prevailing synoptic conditions in Greece is considered useful, in order to understand the covariability of the meteorological parameters examined and their effect to BAA. During the cold period of the year (October–April) Table 10.1 Mean and maximum daily values for PM10, SO2, O3, and meteorological variables in Athens for the period 1/1/2001–31/12/2002 Meteorological Mean daily parameters values Tmean (°C) ∆Tmean (°C) Tmax (°C) ∆Tmax (°C) Tmin (°C) ∆Tmin (°C) Trange (°C) ∆Trange (°C) e (mm Hg) ∆e (mm Hg) P (hPa) ∆P (hPa) E (mm) ∆E (mm) S (h) ∆S (h) I (W m−2) ∆I (W m−2)
19.2 ± 7.4 0.0 ± 1.6 23.9 ± 8.4 0.0 ± 2.3 15.6 ± 6.8 0.0 ± 1.7 8.3 ± 2.6 0.0 ± 2.3 10.2 ± 3.4 0.0 ± 1.5 1,003.1 ± 5.6 0.0 ± 3.5 3.2 ± 2.1 0.0 ± 1.0 8.0 ± 4.3 0.0 ± 3.5 188.2 ± 94.2 0.1 ± 51.4
Maximum daily values Air pollutants 34.1 5.6 40.2 7.7 29.3 7.4 17.0 7.9 18.6 5.6 1,021.3 12.1 10.0 3.4 14.1 10.7 352.0 193.0
PM10 mean (µg/m3) ∆PM10 mean (µg/m3) PM10 max (µg/m3) ∆PM10 max (µg/m3) SO2 mean (µg/m3) ∆SO2 mean (µg/m3) SO2 max (µg/m3) ∆SO2 max (µg/m3) O3 mean (µg/m3) ∆O3 mean (µg/m3) O3 max (µg/m3) ∆O3 max (µg/m3) – – – – – –
Mean daily values
Maximum daily values
59.75 ± 23.32 0.01 ± 20.88 128.14 ± 61.08 −0.04 ± 61.93 14.34 ± 9.56 0.03 ± 6.96 35.81 ± 26.16 0.03 ± 24.22 44.00 ± 21.05 −0.02 ± 11.77 83.69 ± 32.01 −0.01 ± 19.58 – – – – – –
171.96 95.45 521.75 458.50 64.76 29.45 151.29 98.87 97.60 53.78 210.00 81.17 – – – – – –
180
P. Th. Nastos
northern winds prevail in Greece because of the existence of anticyclones over Europe and Siberia and low barometric pressure over Mediterranean Sea. The blow of these winds is interrupted by the blow of southern winds due to passage of troughs in Mediterranean Sea and Europe. Polar continental air masses cover frequently the whole the country, producing a cold regime, which combines low air
Table 10.2 Results of the application of Generalized Linear Models (GLM) with Poisson distribution (dependent variable is the daily number of outpatients with bronchial asthma, while independent covariate is each one of the aforementioned meteorological parameters and the ambient air pollutants). The shaded parameters indicate statistically significant relationships with BAA Meteorological parameters
b coefficient ± Significance Ambient air standard error level p pollutants
b coefficient ± standard error
Significance level p
Tmean (°C)
−0.0153 ± 0.0039 −0.0064 ± 0.0167 −0.0131 ± 0.0034 −0.0042 ± 0.0118 −0.0169 ± 0.0043 −0.0076 ± 0.0157 −0.0213 ± 0.0108 0.0002 ± 0.0122 −0.0302 ± 0.0083 0.0162 ± 0.0186 0.0036 ± 0.0050 −0.0038 ± 0.0076 −0.0479 ± 0.0138 −0.0253 ± 0.0277 −0.0194 ± 0.0062 −0.0055 ± 0.0076 −0.0009 ± 0.0003 −0.0006 ± 0.0005
0.0005 ± 0.0012
0.697836
0.0001 ± 0.0013
0.947794
0.0001 ± 0.0005
0.892464
∆Tmean (°C) Tmax (°C) ∆Tmax (°C) Tmin (°C) ∆Tmin (°C) Trange (°C) ∆Trange (°C) e (mm Hg) ∆e (mm Hg) P (hPa) ∆P (hPa) E (mm) ∆E (mm) S (h) ∆S (h) I (W m−2) ∆I (W m−2)
0.000092 0.701445 0.000138 0.724132 0.000080 0.626525 0.049104 0.989784 0.000263 0.382841 0.467488 0.620220 0.000525 0.361491 0.001839 0.471294 0.003225 0.239556
PM10 mean (µg m−3) ∆PM10 mean (µg m−3) PM10 max (µg m−3) ∆PM10 max (µg m−3) SO2 mean (µg m−3) ∆SO2 mean (µg m−3) SO2 max (µg m−3) ∆SO2 max (µg m−3) O3 mean (µg m−3) ∆O3 mean (µg m−3) O3 max (µg m−3) ∆O3 max (µg m−3)
−0.0001 ± 0.0004 0.885435 0.0040 ± 0.0029
0.165814
−0.0015 ± 0.0042 0.720442 0.0012 ± 0.0010
0.261381
−0.0012 ± 0.0012 0.295670 −0.0036 ± 0.0014 0.008275 −0.0019 ± 0.0024 0.427757 −0.0033 ± 0.0009 0.000354 −0.0021 ± 0.0015 0.164866
10 Weather, Ambient Air Pollution and Bronchial Asthma in Athens, Greece
181
temperature, low vapor pressure and evaporation and short duration of sunshine. These weather conditions seem to be responsible for worsening the BAA. The prevailing weather type during warm period of the year (May–September) is that of Etesians winds (periodical winds of the north section). This type is established in Greece, when a North Atlantic anticyclone extends over Europe covering the Balkans simultaneously with the Indian low over Asia Minor and the Eastern Mediterranean Sea. The blow of Etesians winds transfers polar continental air masses to Northern Greece and the result is the appearance of precipitation and frontal thunderstorms. Another effect of the Etesians regime is the summer drought and the uniform weather conditions in Greece. More specifically, the weather is identified by high air temperature, high absolute humidity and evaporation, high total solar radiation and sunshine (Karalis 1969; Kotinis-Zambakas 1983). Regarding the effect of the aforementioned pollutants on BAA on the same day, the findings of the analysis showed that a statistically significant relationship between these pollutants and BAA does not exist, with the exception of O3, which is correlated negatively (p < 0.01) with asthma. One reason for the lack of effects between air pollution and BAA may be the low levels of ambient air pollution measured within the studied period, which ranged between low to medium levels. Another possible explanation is the public response to early warnings by the local authorities once the concentrations of the pollutants cross a particular threshold. People respond to this alert by decreasing their exposure to outdoor air pollution (or increasing their avoidance behavior), and this leads to fewer hospitalizations than would have otherwise occurred. Jalaludin et al. (2004) came to the same conclusions, studying acute effects of urban ambient air pollution on BAA in a cohort of Australian children. They found no association between ambient ozone concentrations and BAA, but there was, however, an association between PM10 concentrations and BAA. Also, in another study conducted in Finland where ambient air pollution levels were similarly low, Timonen and Pekkanen (1997) did not remark any consistent associations between PM10, NO2 and SO2 concentrations and respiratory symptoms in a group of children with asthma or cough. Furthermore, even in conditions with high levels of PM10 and O3, Hoek and Brunekreef (1995) did not find a positive association between these pollutants and respiratory symptoms. None the less, epidemiological and clinical studies have shown that O3 exposure is associated with worsening of athletic performance, reductions in lung function, shortness of breath, chest pain with deep inhalation, wheezing and coughing, and asthma exacerbations among those with asthma (Bates and Caton 2002). Taking into account the clear seasonality of the O3 concentrations within the year, with peak during the warm period (Fig. 10.2), an investigation of a possible relationship between O3 concentrations and BAA, only during the cold periods of the years examined, is performed. As it is mentioned before the prevailing synoptic conditions do not favour the incidence of ozone episodes, but there are some spells with high sunshine and high air temperature. The results of the GLM application pointed out a statistically significant positive relationship between BAA and O3 lag 2 day. More concretely, an increase of 10 µg/m3 in daily mean concentrations is associated with an increase 5% of the probability of having BAA (p = 0.04).
182
P. Th. Nastos
This evidence shows that even low to medium concentrations of O3 are responsible for worsening BAA. No relationships were found between the other examined pollutants (PM10, SO2) and BAA, during the cold period of the year. The investigation for possible relationships within the warm period of the year results in no evidence of significant correlations between BAA and all the pollutants. The age distribution of the patients manifests that the most vulnerable are the elderly people (age > 70 years), so the application of Logistic analysis, taking into account the WHO thresholds of the pollutants, is considered the most appropriate to reveal relative risks. Regarding SO2 threshold (125 µgr/m3), no relative risk was found, but PM10 exceedances of the threshold of 70 µgr/m3 seems to double the risk (odds ratio = 1.70, 95% CI 1.20–2.40, significant level = 0.003) of observing the daily number of outpatients at age 70 years and older compared to all others. Relevant results were extracted for the O3 concentrations above 120 µgr/m3 (odds ratio = 1.70, 95% CI 1.01–2.89, significant level = 0.048). In the process, the delays between pollutants maxima and BAA were examined with the following results (statistically significant at p = 0.05): we found that BAA were lagged on PM10, SO2 and O3 by 7–8 days. An increase of 10 µgr/m3 on lagged daily mean PM10 concentrations is associated with an increase of 3% for BAA, while an increase of 10 µg/m3 on lagged daily mean SO2 concentrations links to an increase of 9% for BAA. As far as the O3 concentrations are concerned, an increase of 10 µg/m3 on lagged daily mean O3 concentrations is associated with a decrease of 4% for BAA and vice versa. As it is mentioned before, ambient ozone is not a primary pollutant and is positively correlated to sunshine and air temperature, climatic factors that play a determinative role in the photochemical procedures. Besides, decreases of BAA are significantly associated to increases in air temperature. Consequently, ambient ozone is negatively associated to BAA. Gratziou et al. (2001), investigating the relationship between ambient air pollution and respiratory health in children, found that 7-day mean of PM10 was associated with upper respiratory symptoms, while Neuberger et al. (2004), observed an increase of respiratory diseases after a lag of 10–11 days of PM2.5, in Vienna, Austria. Besides, a lag of up to 5 days for asthma visits is associated with an increased risk from exposure to air pollution (Lipfert 1993; US EPA 1996). A more descriptive analysis is shown in Figs. 10.3–10.5 where the relative frequency (%) of the BAA per 10-Days Intervals as a function of the minimum air temperature and the water vapor pressure percentiles (Fig. 10.3), the atmospheric pressure and sunshine percentiles (Fig. 10.4) and the PM10, SO2 and O3 percentiles (Fig. 10.5), along with the variation of the total number of admissions per 10-Days Intervals (yellow line) and the polynomial fitting (brown line) are depicted. The stacked bars shown in each interval represent the percentages of BAA associated with the particular parameters’ percentiles. This section is considered useful to demonstrate in a more tangible way the impact of the percentiles distribution of the meteorological and air pollution variables to the total counts of BAA for the whole study period, within 10-Days Intervals. For instance, it is clear from Fig. 10.5 that even in cold periods of the examined years the so rarely appeared maximum values of O3 seem to exacerbate BAA. The construction of the stacked bar for the 3rd
10 Weather, Ambient Air Pollution and Bronchial Asthma in Athens, Greece 0-25%
25-50%
50-75%
75-100%
100%
60
90%
55
80%
50
70%
45
60%
40
50%
35
40%
30
30%
25
20%
20
10%
15
0%
Total Number of BAA
Frequency (%) of BAA (2001-2002)
Minimum Air Temperature Percentiles
183
10
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 35 36
10-Days Intervals Vapor Pressure Percentiles
0-25%
25-50%
50-75%
75-100% 60
55
80%
50
45
60%
40
35
40%
30
25
20%
Total Number of BAA
Frequency (%) of BAA (2001-2002)
100%
20
15
0%
10
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 35 36
10-Days Intervals
Fig. 10.3 Frequency (%) of the BAA per 10-Days Intervals (2001–2002) as a function of the minimum air temperature and water vapor pressure percentiles, along with the variation of the total number of BAA per 10-Days Interval (yellow line) and the polynomial fitting (brown line). Three reference lines (white lines) concerning the mean, the mean + SD and the mean – SD are also depicted (see Appendix 2)
interval concerning the impact of O3 (Fig. 10.5) is briefly described. During the whole study period, 53 BAA are aggregated for the two 3rd 10-Day Intervals (20–31 January) from which 37 counts are associated with the first percentile of O3, 7 counts with the second percentile, 3 counts with the third percentile and 6 counts with the forth percentile. In the process the relative frequencies are evaluated and these are the components of the stacked bar in the graph.
184
P. Th. Nastos Atmospheric Pressure Percentiles
0-25%
25-50%
50-75%
75-100%
100%
60
80%
50
45
60%
40
35
40%
30
Total Number of BAA
Frequency (%) of BAA (2001-2002)
55
25
20%
20
15
0%
10
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 35 36
10-Days Intervals
0-25%
25-50%
50-75%
75-100% 60
90%
55
80%
50
70%
45
60%
40
50%
35
40%
30
30%
25
20%
20
10%
15
0%
Total Number of BAA
Frequency (%) of BAA (2001-2002)
Sunshine Percentiles 100%
10
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 35 36
10-Days Intervals
Fig. 10.4 The same as in Fig. 10.3 but for atmospheric pressure and sunshine percentiles (see Appendix 2)
The total numbers of BAA are higher in winter months, April and October, while a trough is appeared in August. This pattern is similar to the seasonality of asthma admissions in Malta (Grech et al. 2002) and in Ankara, Turkey (Berktas and Bircan 2003). The air temperature, the water vapor pressure and the sunshine within the highest percentile is well associated with the trough of BAA during summer time, while the highest percentile of the atmospheric pressure coincides with the winter peak of BAA. As far as ambient air pollutants are concerned, there is no clear association between BAA and PM10 concentrations, on 10-Days analysis. On the other hand, the highest percentile of SO2 is in agreement with the peak of BAA during
10 Weather, Ambient Air Pollution and Bronchial Asthma in Athens, Greece PM10 Percentiles
0-25%
25-50%
50-75%
185
75-100% 60
100%
55 50 45 60%
40 35
40%
30
Total Number of BAA
Frequency (%) of BAA (2001-2002)
80%
25 20%
20 15
0%
10 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 35 36
10-Days Intervals
SO2 Percentiles
0-25%
25-50%
50-75%
75-100%
100%
60
50 45
60%
40 35
40%
30
Total Number of BAA
Frequency (%) of BAA (2001-2002)
55 80%
25 20%
20 15
0%
10 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 35 36
10-Days Interval
O3 Percentiles
0-25%
25-50%
50-75%
75-100% 60
100%
50
80%
45 40
60%
35 30
40%
Total Number of BAA
Frequency (%) of BAA (2001-2002)
55
25 20
20%
15 0%
10 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 35 36
10-Days Interval
Fig. 10.5 The same as in Fig. 10.3 but for PM10, SO2 and O3 percentiles (see Appendix 2)
186
P. Th. Nastos
January and February and the minimum of BAA occurs within the highest percentile of O3. The application of Generalized Linear Models on the basis of 10-Days Intervals revealed that an increase of 10 units in 10-Days mean concentration of SO2 links to an increase 16% of the respective totals of BAA. The findings of the above analysis show evidence that weather and air pollution influence BAA, but a further investigation will be carried out since only 2 years were available for the medical datasets.
10.4
Conclusions
The bronchial asthma admissions in the major hospital of Athens for respiratory diseases, during the period 1/1/2001–31/12/2002, show a clear pattern with peaks in winter months, April and October and minimum in August. The application of GLM analysis to the meteorological and medical datasets revealed that there is a statistically significant negative relationship among daily mean, maximum and minimum air temperature, diurnal air temperature range, mean water vapor pressure, total evaporation, total sunshine, mean total irradiance and BAA. On the other hand, there is no statistically significant association between PM10, SO2 concentrations and BAA on the same day while a weak negative relationship between O3 concentrations and BAA is extracted. Nevertheless, a statistically significant positive relationship between BAA and O3 is revealed when the cold periods of the examined years were considered. Besides, increases in PM10 and SO2 concentrations were associated with increases in BAA, while increases in O3 concentrations were associated with decreases in BAA, which become evident after lag 7–8 days. However, there is a need for further research to explain the mechanisms for the observed associations between BAA, weather and ambient air pollution. Acknowledgements Thanks to Helena Nikolopoulos for proofreading and editing the manuscript and to the unknown reviewers for the constructive comments.
References Bartzokas, A., Kassomenos, P., Petrakis, M., Celesides, C. (2003) The effect of meteorological and pollution parameters on the frequency of hospital admissions for cardiovascular and respiratory problems in Athens. Indoor Built Environ., 13, 271–275. Bascom, R. (1996) Health effects of outdoor air pollution. Am. J. Respir. Crit. Care Med., 153, 3–50. Bates, D.V., Caton, R.B. (2002) A citizen’s guide to air pollution. David Suzuki Foundation. Pittsburgh: Dorrance Publishing. Berktas, B.M., Bircan, A. (2003) Effects of atmospheric sulphur dioxide and particulate matter concentrations on emergency room admissions due to asthma in Ankara. Tuberk Toraks, 51(3), 231–238.
10 Weather, Ambient Air Pollution and Bronchial Asthma in Athens, Greece
187
Braun-Fahrlander, C., Ackermann-Liebrich, U., Schwartz, J., Gnehm, H.P., Rutishauser, M., Wanner, H.U. (1992) Air pollution and respiratory symptoms in preschool children. Am. Rev. Respir. Dis., 145, 42–47. Burnett, R.T., Dales, R.E., Raizenne, M.E., Krewski, D., Summers, P.W., Roberts, G.R., RaadYoung, M., Dann, T., Brooke, T. (1994) Effects of low ambient levels of ozone and sulfates on the frequency of respiratory admissions to Ontario hospitals. Environ. Res., 65, 172–194. Delfino, R.J., Murphy-Moulton, A.M., Burnett, R.T., Brook, J.R., Becklake, M.R. (1997) Effects of air pollution on emergency room visits for respiratory illness in Montreal, Quebec. (1997) Am. J. Respir. Crit. Care Med., 155, 568–576. Dockery D.W., Pope III C.A. (1994) Acute respiratory effects of particulate air pollution Annu. Rev. Publ Health, 15, 107–132. EPA (2006) Guideline for Reporting of Daily Air Quality – Air Quality Index (AQI). US Environmental Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park, North Carolina 27711, EPA-454/B-06-001. Galán, I., Tobías, A., Banegas, J.R. (2003) Short-term effects of air pollution on daily asthma emergency room admissions. Eur. Respir. J., 22, 802–808. Goldstein, I.F. (1980) Weather patterns and asthma epidemics in New York City and New Orleans, USA. Int. J. Biometeorol., 24(4), 329–339. Gratziou, C., Katsouyanni, K., Gratziou, M., Manalis, N., Kalandidi, A., Roussos, C. et al. (2001) Air pollution short term effects on respiratory health of children living in Athens. Proceedings of the 7th International Conference on Environmental Science and Technology, Ermoupolis, Syros island, Greece, pp. 291–294. Grech, V., Balzan, M., Asciak, R.P., Buhagiar, A. (2002) Seasonal variations in hospital admissions for asthma in Malta. J. Asthma, 39(3), 263–268. Greenburg, L., Field, F., Reed, J.I., Erhardt, C.L. (1966) Asthma and temperature range.II. 1964 and 1965 epidemiodological studies of emergency clinic visits for asthma in three large New York City hospitals. Arch. Environ. Health, 12, 561–563. Hajat, S., Bird, W., Haines, A. (2004) Cold weather and GP consultations for respiratory conditions by elderly people in 16 locations in the UK. Eur. J. Epidemiol., 19(10), 959–968. Hashimoto, M., Fukuda, T., Shimizu, T., Watanabe, S., Watanuki, S., Eto, Y., Urashima, M. (2004) Influence of climate factors on emergency visits for childhood asthma attack. Pediatr. Int., 46(1), 48–52. Hoek, G., Brunekreef, B. (1995) Effect of photochemical air pollution on acute respiratory symptoms in children. Am. J. Respir. Crit. Care Med., 151, 27–32. IPCC (2001) Climate Change 2001: Impacts, Adaptation and Vulnerability. Technical Summary. Jaffe, D.H., Singer, M.E., Rimm, A.A. (2003) Air pollution and emergency department visits for asthma among Ohio Medicaid recipients, 1991–1996. Environ. Res., 91, 21–28. Jalaludin, B.B., O’Toole, B.I., Leedor, S.R. (2004) Acute effects of urban ambient air pollution on respiratory symptoms, asthma medication use, and doctor visits for asthma in a cohort of Australian children. Environ. Res., 95(1), 32–42. Karalis, J.D. (1969) Types of weather in Greece, Ph.D. thesis, University of Athens, Athens, pp. 32–54, 76. Kashiwabara, K., Kohrogi, H., Ota, K., Moroi, T. (2002) High frequency of emergency room visits of asthmatic children on misty or foggy nights. J Asthma 39, 711–717. Kotinis-Zambakas, S. (1983) Contribution to the study of the climate of Greece. Normal weather per month. Research Center for Atmospheric Physics and Climatology, Academy of Athens, Publication 8, pp. 99–127. Lipfert, F.W. (1993) A critical review of studies of the association between demands for hospital services and air pollution. Environ. Health Perspect., 101, 229–268. McGullagh, P., Nelder, J.A. (1997) Generalized Linear Models, 2nd Edition, London: Chapman & Hall, p. 450. Nedler, J.A, Wedderburn, R.M. (1972) Generalized linear models, J. Roy. Stat. Soc. Ser A 135, 370–384.
188
P. Th. Nastos
Neuberger, M., Schimek, M., Horak, F., Moshammer, H., Kundi, M., Frischer, T., Gomiscek, B., Puxbaum, H., Hauck, H., AUPHEP-Team (2004) Acute effects of particulate matter on respiratory diseases, symptoms and functions: epidemiological results of the Austrian Project on Health Effects of Particulate Matter (AUPHEP). Atmos. Environ., 38, 3971–3981. Ostro, B.D., Eskeland, G.S., Sanchez, J.M., Feyzioglu, T. (1999) Air pollution and health effects: a study of medical visits among children in Santiago, Chile. Environ. Health Perspect., 107, 69–73. Schwartz, J., Slater, D., Larson, T.V., Pierson, W.E., Koenig, J.Q. (1993) Particulate air pollution and hospital emergency room visits for asthma in Seattle. Am. Rev. Respir. Dis., 147, 826–831. Timonen, K.L., Pekkanen, J. (1997) Air pollution and respiratory health among children with asthmatic or cough symptoms. Am. J. Respir. Crit. Care Med., 156, 546–552. Trasande, L., Thurston, G.D. (2005) The role of air pollution in asthma and other pediatric morbidities. J. Allergy Clin. Immunol., 115, 689–699. US Environmental Protection Agency (US EPA) (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants. EPA/600/ P-93/004cF. US Environmental Protection Agency, Research Triangle Park, NC. Environ. Res., 65, 56–68. Walters, S., Griffiths, R.K., Ayers, J.G. (1994) Temporal association between hospital admissions for asthma in Birmingham and ambient levels of sulphur dioxide and smoke. Thorax, 49, 133–140. Weiland, S.K., Husing, A., Strachan, D.P., Rzehak, P., Pearce, N. (2004) Climate and the prevalence of symptoms of asthma, allergic rhinitis, and atopic eczema in children. J. Occup. Environ. Med., 61(7), 609–615. WHO (2000) Air Quality Guidelines for Europe, 2nd edn. WHO Regionalm Publications, European series, No 91. Yuksel, H., Tanac, R., Tez, E., Demir, E., Coker, M. (1996) Childhood asthma and atmospheric conditions. Acta Paediatr. Jpn. (Overseas Edition), 38(6), 606–610.
Chapter 11
Outreach from the IAI Collaborative Research Network on Climate Variability and Human Health Impacts in the Tropical Americas Joan L. Aron
Abstract The Inter-American Institute for Global Change Research (IAI) conducted a 2-week training institute on Climate and Health in the Americas on November 7–18, 2005 at the University of the West Indies in Kingston, Jamaica (see http:// www.institutes.iai.int). This activity was jointly coordinated as an outreach project of the Collaborative Research Network supported by the IAI to address Climate Variability and Human Health Impacts in the Tropical Americas with a focus on malaria and dengue (see http://chiex.net). The central objective of the Training Institute was to help strengthen local and regional capacity to address the impacts of climate variability and climate change on human health in the populations of the Americas, particularly Latin America and the Caribbean. The Training Institute had three core components. (1) The Science component provided an overview of climate and health impacts in the Americas. (2) The Applications component provided a perspective based on operational experience in disaster response and communication with decision-makers. (3) The Proposal Development component provided guidance on interdisciplinary proposal development as well as time for self-organized participant teams to develop and receive feedback on pre-proposals (in English or Spanish) for Training Institute Seed Grants to support follow-on projects. One set of English–Spanish simultaneous translators was available for the entire Training Institute. The approach was fundamentally cross-sectoral and cross-disciplinary in that the health sector must engage professionals in climate, weather, ecosystems, water resources, agriculture, and various social sciences as well as communicate across its own specialties. The approach took into account recent advances in understanding knowledge–action systems, which are the institutional structures needed to create effective systems for the transfer of knowledge to action that are applicable to climate-related issues. Participants had opportunities for more active learning in two field trips and one laboratory experience: inspecting breeding sites for the mosquito that transmits dengue; observing landforms prone to landslides; manipulating images and data used in geographic information systems and remote sensing. Recommendations
Science Communication Studies, Columbia, Maryland, USA
M.C. Thomson et al. (eds.), Seasonal Forecasts, Climatic Change and Human Health. © Springer Science + Business Media B.V. 2008
189
190
J.L. Aron
for future Training Institutes included incorporating new technologies and communicating with policy-makers to develop more proactive societal strategies to manage risks. Keywords Public health, Climate variability, Climate change, Capacity-building, Latin America and the Caribbean
11.1
Introduction
The Inter-American Institute for Global Change Research (IAI) conducted a 2-week training institute on Climate and Health in the Americas on November 7–18, 2005 at the University of the West Indies (UWI) in Kingston, Jamaica (see http:// www.institutes.iai.int). This activity was jointly coordinated as an outreach project of the Collaborative Research Network supported by the IAI to address Climate Variability and Human Health Impacts in the Tropical Americas with a focus on malaria and dengue (see http://chiex.net). The organization of the Training Institute was motivated by the growing recognition that public health is vulnerable to climate change and variability. Concerns about the adverse impacts of climate change and variability cover a variety of health-related issues: infectious diseases, natural disasters, food and nutrition security, water resources, heat stress, air pollution and asthma as well as interactions with stratospheric ozone depletion. There is a great need to strengthen the capacity to understand and assess these adverse health impacts with the aim of providing better tools for decision-makers to develop adaptation strategies to reduce or prevent such impacts. This paper provides an overview of the Training Institute in terms of the objectives, the organizers and participants as well as the pedagogical approaches. Distinctive features are emphasized along with recommendations for future activities that emerged from the Training Institute.
11.2
Objectives
The central objective of the Training Institute on Climate and Health in the Americas was to help strengthen local and regional capacity to address the impacts of climate variability and climate change on human health in the populations of the Americas, particularly Latin America and the Caribbean. It built upon the concept of knowledge– action systems for linking science and action [1]. Effective knowledge–action systems have six components. Those components especially relevant for setting the context for the Training Institute are the ones pertinent to boundary organizations, design for learning and long-term investments in people who can work across the disciplines. 1. Problem definition that is collaborative but user-driven 2. Complete inclusion on the continuum of decision-maker to knowledge-producer
11 Outreach from the IAI Collaborative Research Network
191
3. Boundary organizations that act as intermediaries between nodes in the system – most notably between scientists and decision-makers 4. Design for learning rather than knowing (flexible processes and institutions) 5. Funding strategies tailored to the dual private/public character of such systems and with sufficient continuity to foster long-term relationships between producers and users 6. Long-term investments in people who can work across disciplines, issue areas, and the knowledge–action interface An important antecedent is the workshop on climate and health effects in the Caribbean that was organized in 2002 by the Pan American Health Organization (PAHO) and the World Health Organization (WHO) [2]. Its objectives, which were developed by an international partnership headed by WHO/PAHO, the World Meteorological Organization (WMO) and the United Nations Environment Program (UNEP), define the steps required to provide better tools and approaches for adaptation strategies in public health. 1. Generate awareness of the impact of climate variability and change on health. 2. Understand how climate data are and could/should be used in health planning. 3. Identify the elements of a framework for proactive health/climate actions to assess vulnerabilities and implement mitigation and adaptation strategies in relation to adverse health impacts of climate variability and change. 4. Discuss and define the roles of health and climate professionals in the implementation of the framework for proactive health/climate actions. 5. Identify key partners and assess institutional/organizational arrangements that must be strengthened and what new entities must be put in place at the national and regional levels to assess vulnerabilities and implement mitigation and adaptation strategies in relation to adverse health impacts of climate variability and change. 6. Identify follow-on capacity-building activities to address climate variability and change and health nationally and regionally. Within this broader context, the Training Institute had three specific programmatic objectives. 1. Provide participants with a theoretical and methodological framework to better understand the interactions of global climate change and health and in particular their potential impacts on health sectors of IAI member countries, communities and nations. 2. Present new scientific information on the evolution of the main theme and subthemes of the Training Institute and on the relevance, effectiveness and efficiency of existing research and policy initiatives. 3. Build relations and improve communication between health professionals, natural and social scientists, resource managers, decision-makers and other professionals, and maximize the use of existing capability mechanisms. The IAI also articulated its own institutional objectives for the activity.
192
J.L. Aron
1. Develop a cooperative network of professionals in the Americas that can exchange knowledge and experience in future activities, share data and promote exchange and training of local professionals, as well as the development of partnerships among governments, industries and communities. 2. Connect local and regional professionals and institutions in the region with initiatives and networks of other organizations dealing with the same subject (e.g. International Research Institute for Climate and Society (IRI), PAHO, WHO). 3. Inform local and regional professionals about the funding opportunities available to support projects dealing with climate variability and change, health, and related fields.
11.3
Organizers and Participants
The principal organizers were Joan L. Aron (the author) as the Science Program Coordinator and Simon R. Young as the Applications Program Coordinator. Dr. Aron’s area of expertise is ecosystem change and public health with a background in mathematical models applied to epidemiology. She is a member of the IAI Collaborative Research Network on climate and health. Dr. Young’s area of expertise is in natural hazards assessment, working with PAHO and other organizations. He has a background in geology, with a specialty in volcanology. Anthony Chen and Michael Taylor, faculty members in UWI’s Climate Studies Group, made scientific contributions as well as local organizational arrangements. As the IAI Communication, Training, and Outreach Officer, Marcella Ohira provided overall management and administrative support with assistance from Celine Leite. The US National Science Foundation (NSF) provided the core funding through a grant to IAI. UWI was the local host institution. Other agencies and organizations contributed funds to support the Training Institute: International Human Dimensions Program (IHDP), WMO, International Development Research Centre of Canada (IDRC), System for Analysis, Research, and Training/Assessments of Impacts and Adaptations to Climate Change in Multiple Regions and Sectors (START/AIACC). PAHO contributed speakers and a planning session. The intent that the participants continue with activities after the Training Institute guided the selection of participants who had professional experience and could provide leadership in building networks and projects on the theme of climate and health. The 20 participants work in 14 countries in Latin America and the Caribbean in diverse institutions (see Table 11.1). This list was distributed to the participants and guest lecturers in advance so that they would have a better understanding of the group. During the Training Institute, participants also had an opportunity to introduce themselves and their work to the group.
Name
Surname
Albert Asinto Eleutherio Jairo Alberto
Martis Garcia Giraldo
Country of citi- Country of resizenship dency Institution
Background
Interests
Gender
The Netherlands The Netherlands Meteorological Service Antilles Antilles Colombia Colombia Hydrology, Meteorology & Environment Institute Brazil Brazil National Institute of Meteorology Cuba Cuba Meteorology Institute
Applied Physics, Meteorology Geography, Meteorology
Climate
M
Environment and health
M
Meteorology
M
Bioclimatology
Climate and weather forecasts Air pollution and climate change Extreme weather and health
Diniz
Ana Cristina
Pompa Toledano
Cuba
Cuba
Mario Alberto
Lamfri
Argentina
Argentina
Alex Sinclair
Ifill
Barbados
Barbados
Marilyn Silvia Aparicio Efen
Bolivia
Bolivia
Ricardo Jose
Brazil
Brazil
David
de Paula Souza e Guimaraes Joa Espinal
Dominican Republic
Dominican Republic
Sherine
Huntley
Jamaica
Jamaica
Wallo Vasquez
Ministry of Science, Technology and Environment National Space Commission Univ. West Indies and Barbados Water Authority Ministry of Sustainable Development and Planning National Institute of Space Research National Center for Tropical Disease Control Ministry of Health
Meteorology
Physics
M F
Remote sensing, landscape epidemiology, vectorborne disease Environmental Biology Climate change, water resources, early warning for dengue fever Development Policies, Climate change and human Medicine health; malaria and leishmaniasis Remote Sensing, Remote sensing; Biology, Pathology environment and health; schistosomiasis Medicine, Population Climate change and health; and Development malaria control
M
Medical Entomology
F
Vector control, dengue
M
F
M
M
(continued)
193
Francisco de Assis Antonio
11 Outreach from the IAI Collaborative Research Network
Table 11.1 Participants
194
Table 11.1 (continued) Name
Surname
Country of citizenship
Country of residency
Jose H.
Azoh Barry
Cote d’Ivoire
Mexico
Institution
Background
Interests
Gender
Social Council of the Northern Border University of Guadalajara National Environment Authority
Anthropology
Human health education
F
Chemical Engineer
Hydrochemistry, sedimentology Climate change monitoring and impacts
M
Climate change and variability, health impacts Climate change and variability, vector-borne disease Climate and malaria
F
Hermes Ulises Ramirez Sanchez Mexico
Mexico
Gladys Haydee Villareal Madrid
Panama
Panama
Palmira R.
Ventosilla Lopez
Peru
Peru
Institute of Tropical Medicine
Karen Alicia
Polson
Jamaica
Trinidad and Tobago
PAHO/Caribbean Entomology, Epidemiology Centre Parasitology
Jorge E.
Moreno
Venezuela
Venezuela
Ruben Anibal
Bejaran
Argentina
Argentina
Guillermo Leon Rua Uribe
Colombia
Colombia
Norma
Peru
Peru
Institute of Higher Biology, Studies Dr. Gabaldon Entomology University of Buenos Atmospheric Sciences Bioclimatology, climate variability, health Aires Malaria early warning University of Antioquia Medical system Entomology University of Piura Industrial Engineering, Climate research, El Nino, Management risk management
Ordinola Ipanaque
Meteorology, Environment and Natural Resources Molecular Biology, Biochemistry
F
F
M M M F
J.L. Aron
11 Outreach from the IAI Collaborative Research Network
11.4
195
Pedagogical Approaches
The Training Institute had three core components. (1) The Science component provided an overview of climate and health impacts in the Americas. (2) The Applications component provided a perspective based on operational experience in disaster response and communication with decision-makers. (3) The Proposal Development component provided guidance on interdisciplinary proposal development as well as time for self-organized participant teams to develop and receive feedback on pre-proposals (in English or Spanish) for Training Institute Seed Grants to support follow-on projects. The participant teams submitted full proposals 2 months after the Training Institute. The list of lecture presentations for all components is shown in the appendix. The combination of these three components was a distinctive feature of the Training Institute. Another distinctive feature was the use of English-Spanish translators for the entire Training Institute, which is essential for bringing together Latin America and the English-speaking Caribbean. It was not feasible to accommodate Portuguese for the Brazilians. The topics spanned a broad range of disciplines. Within infectious disease, a variety of mosquito-borne illnesses were discussed (malaria, dengue, different kinds of viral encephalitis) as well as water-borne (diarrheal) disease and algal blooms in oceans and estuaries. Other topics were wind storms and floods, drought, food security and nutrition along with air quality and disease. The exacerbating effect of global warming on stratospheric ozone depletion and associated ultraviolet exposure was also presented. The multi-hazard perspective provided greater context on adaptation responses and suggested an examination of geological interactions and climate, notably volcanic eruptions and rates of sea level rise. The topics crossed over different specialties of public health, corresponding to three different areas of PAHO: sustainable development and environmental health; communicable disease control; emergency preparedness and disaster relief. The lecture format was supplemented by discussion and breakout groups to allow the participants to delve more deeply into the themes. Greater focus on selected topics served to highlight teams, especially the IAI collaborative network on climate and health, its START/AIACC partner network on climate and dengue in the Caribbean, and UWI expertise in natural hazards and mapping. Speakers also came from organizations that might become part of future networks, such as the IDRC Ecohealth project, IRI, Environment Canada, PAHO and the Organization of American States (OAS). Strategies of active learning through field trips and laboratories enabled the Training Institute to take advantage of local resources at UWI and the Kingston area in general. Active learning produces a greater depth of understanding and a foundation on which the learner can build. An entomological field trip to the village of Hellshire was led by Dave Chadee of UWI’s campus in St. Augustine, Trinidad working with the Jamaican Ministry of Health. Participants had an opportunity to walk in small groups with health inspectors entering yards to inspect breeding sites for larvae and pupae of the mosquito that transmits the dengue virus. A geological
196
J.L. Aron
field trip on natural hazards in the Kingston area was led by Rafi Ahmad of UWI in Kingston and Simon Young, the Applications Coordinator. Participants could observe areas prone to landslide and see the evidence of land movement. A computer laboratory session on geographic information systems and remote-sensing tools was led by Parris Lyew-Ayee of UWI in Kingston and Simon Young. The topics were topography and slope in Jamaica, 3-h satellite rainfall data, daily atmospheric aerosol data, and imagery for Hurricane Wilma. IAI conducted a written and oral evaluation of the participants independently of the Science and Applications Coordinators. The participants uniformly agreed that the Training Institute was very productive for them. Of the 19 participants who completed the written evaluation form, all agreed that the Training Institute was successful in building the IAI’s “human network” and in providing an understanding of how other disciplines (social sciences, natural sciences, policy and decision-making process, etc.) can help address problems. Eighteen agreed that the Training Institute was successful in bringing together professionals from several backgrounds (natural/social scientists, decision-makers, stakeholders, etc.). The main limitation noted was the paucity of social scientists (only one participant was a social scientist).
11.5
Recommendations for Future Activities
Recommendations to emerge from the Training Institute lie principally in two areas. First, the participants are very interested in the use of new technologies and want to learn more about remote sensing tools especially. Second, the participants want to communicate more with policy-makers to develop more proactive societal strategies to manage risk. Supporting both of these interests will require some creative logistical planning. Access to new technologies is limited by cost and by the complexity of use. Access to policy-makers is limited by their time availability and unpredictable schedules as urgent matters take priority over scheduled presentations. Nevertheless it should be possible to adapt formats for interaction that aid the expansion of networks of professionals in the Americas who can help societies address the public health impacts of climate variability and change. The next immediate step was the awarding of four small peer-reviewed seed grants to Training Institute participants (see Table 11.2). To further strengthen the networking, two scientists per research team will be invited to an August 2006 meeting of IAI Principal Investigators organized in collaboration with Brazil’s Center for Climate and Weather Forecasting (CPTEC) to include a training event on climate, land use and modeling. Acknowledgements I greatly appreciate the collaboration of the Training Institute organizers, especially Simon Young, Tony Chen, Michael Taylor, Marcella Ohira and Celine Leite, who made the Training Institute happen in Jamaica. The enthusiasm and interest of the Guest Lecturers and the Training Institute participants made it a great experience. I am grateful for travel support to Wengen, Switzerland from the Wengen workshop organizers and the Collaborative Research Network on Climate Variability and Human Health Impacts in the Tropical Americas.
Ref.
Project title
TISG-J-1 “Evaluación de Eventos Climáticos Extremos y su Impacto en la Salud en América Latina: casos contaminantes atmosféricos y su influencia en enfermedades respiratorias agudas en la zona metropolitana de Guadalajara; y olas de calor y su incidencia en enfermedades cardiovasculares en Buenos Aires”
Other participating countries
Mexico
Cubaa, Panama Brazil Bolivia
PI’s institution
Hermes Ramírez
Universidad de Guadalajara, Instituto de Astronomía y Meteorología
Rubén Bejaran
Universidad de Buenos Aires, Argentina Departamento de Ciencias de la Atmósfera y los Océanos Universidad de Antioquia, Colombia Programa de Estudio y Control de Enfermedades Tropicales
TISG-J-2 “Evaluación Eco-Regional del Efecto de la Variación Climática Guillermo Rúa sobre la Transmisión de Dengue”
TISG-J-3 “Validación Local de Modelos de Predicción de Enfermedades transmitidas por Vectores y asociadas al Clima”
PI’s country
Name of PI(s)
Palmira Ventosilla Universidad Peruana Cayetano Heredia, Instituto de Medicina Tropical Alexander von Humboldt Jairo García Instituto de Hidrología, Meteorología y Estudios Ambientales, IDEAM
Bolivia Collaboration of: Mario Lamfri, Argentina and Jose Azoh, Mexico
Peru
Colombia
11 Outreach from the IAI Collaborative Research Network
Table 11.2 Training institute seed grants
Argentina Bolivia, Brazil Dominican Republic (continued)
197
Table 11.2 (continued) Project title
Name of PI(s)
TISG-J-4 “A cross-cultural Study on Vulnerability to Dengue Jose Azoh Transmission in Latin America and the Caribbean. Insights on two sets of environmental risk factors and elements for human behavior change in three impoverished urban communities” Guillermo Rúa
a
PI’s institution
PI’s country
Consejo Social de la Frontera Norte (COSOFO)
Mexico
Universidad de Antioquia, Programa de Estudio y Control de Enfermedades Tropicales
Colombia
Other participating countries Peru Venezuela The Netherlands Antilles The Netherlands Antilles Trinidad & Tobago Peru Argentina Jamaica Voluntary Advisor: Dr. Dave Chadee, Trinidad & Tobago
198
Ref.
All Cuban activities will be funded by the IAI Director’s Special Fund (non-US resources).
J.L. Aron
11 Outreach from the IAI Collaborative Research Network
199
Appendix: Lecture Presentations Keynote Lecture Climate Change, Climate Variability and Weather Professor Anthony Chen, Department of Physics, University of the West Indies, Mona, Jamaica
Introduction to Training Institute Climate and Health in Relation to Global Change: Overview Joan L. Aron, Training Institute Science Program Coordinator Natural Hazards, Social Vulnerability, Adaptive Capacity, Early Warning, and Communicating with Decision-Makers: Overview Simon R. Young, Training Institute Applications Program Coordinator
Session: Natural Hazards, Early Warning and Social Dimensions Climate Change and the Frequency/Severity of Geological and Hydro-Meteorological Hazards and Risks Simon R. Young, Training Institute Applications Program Coordinator Climate Change Hazards and Emergency Preparedness: A Risk Management Approach Don MacIver, Environment Canada USAID/OFDA Delivering Disaster Assistance Alan Ross, Consultant to USAID (Caribbean Region)
Session: Diverse Aspects of Climate and Health Monitoring and Forecasting Mosquito Abundance and Encephalitis Virus Transmission Risk in California Forrest Melton, California State University, Monterey Bay & NASA, USA Chris Barker, University of California, Davis, USA Eutrophication, Harmful Algal Blooms and Health in Coastal and Estuarine Systems Under Increasing Climate and Human-Driven Environmental Changes Gustavo J. Nagy, Universidad de la Republica, Uruguay Saharan Dust and Human Health Michele Monteil, University of the West Indies, St. Augustine, Trinidad Stratospheric Ozone Depletion and Human Health Janice Longstreth, Institute for Global Risk Research, USA
200
J.L. Aron
Session: Climate and Mosquito-Borne Diseases: Malaria and Dengue Climate Variability and Mosquito-Borne Diseases: The Case of Dengue in Mexico Horacio Riojas-Rodriguez, Instituto Nacional de Salud Pública, Mexico The Effect of Climate Variability on Malaria and Dengue in Venezuela Yasmin Rubio-Palis, Ministerio de Salud y Desarrollo Social (MSDS) & BIOMED, Universidad de Carabobo, Venezuela Development of an Early Warning System to Prevent Epidemics of Malaria in Colombia Guillermo L. Rua, Universidad de Antioquia, Colombia Moving Average Temperature and Early Warning for Dengue: Temperature Indices and Onset of Dengue Dharmaratne Amarakoon, University of the West Indies, Mona, Jamaica Anthony Chen, University of the West Indies, Mona, Jamaica Early Warning Systems for Climate Sensitive Diseases Stephen J. Connor, International Research Institute for Climate and Society, USA Malaria: From Basics to GIS Based Epidemiology into Ecohealth Approach to Human Health Juan Eugenio Hernandez, Instituto Nacional de Salud Pública, Mexico Ecohealth and Vector-Borne Diseases Roberto Bazzani, International Development Research Centre (Canada), Uruguay Field Approaches for Determining Dengue Risk Factors Dave Chadee, University of the West Indies, St. Augustine, Trinidad
Session: Project Development and Training Institute Seed Grants Trans-disciplinary Collaborations and Partnerships: Making the Whole Greater than the Sum of the Parts Leigh Abts, Johns Hopkins University, USA Climate Change and Health in the Americas: A Review of State of Knowledge and Project Experience Marco Alcaraz, Organization of American States
Session: Natural Disasters and Climate Change: Case Studies Living with Natural Hazards in Jamaica Rafi Ahmad, University of the West Indies, Mona, Jamaica Early Warning Systems and Vulnerability Allan Lavell, Facultad Latinoamericana de Ciencias Sociales y La Red de Estudios Sociales en Prevención de Desastres en América Latina-LA RED, Costa Rica From El Niño to Seasonal Forecasts and the Monitoring of Climate Change in the Caribbean Michael Taylor, University of the West Indies, Mona, Jamaica
11 Outreach from the IAI Collaborative Research Network
201
Session: Water Security and Quality Climate Change and Freshwater Resource Management in the Caribbean: Case of Jamaica Balfour Spence, University of the West Indies, Mona, Jamaica Climate in Relation to Food and Nutrition Security in Central America Ana Victoria Roman, Institute of Nutrition of Central America and Panama, PAHO Joan L. Aron, Science Program Coordinator Achieving Millennium Development Goal (MDG) #7: A Critical Step Towards the Achievement of Health-Related MDGs (Water and Health in the Americas) Ana Treasure, PAHO, Jamaica Elizabeth Ward, Ministry of Health, Jamaica
Session: Project Development and Preparing Training Institute Seed Grants Proposal Evaluation Paul Filmer, US National Science Foundation
References 1. National Research Council Roundtable on Science and Technology for Sustainability. 2005. Knowledge-Action Systems for Seasonal to Interannual Climate Forecasting: Summary of a Workshop. National Academies Press, Washington, DC. 2. Aron, Joan L, Corvalán, Carlos F and Philippeaux, Harry, eds. 2003. Climate Variability and Change and their Health Effects in the Caribbean: Information for Adaptation Planning in the Health Sector. Conference May 21–22, 2002, Workshop May 23–25, 2002, St. Philip, Barbados, West Indies. World Health Organization, Geneva. ISBN 92 4 159071 8 (English).
Recent Developments and Next Steps in Climate Change and Health Patrick L. Kinney1, Ricardo Garcia-Herrera2, and Stephen J. Connor3
Abstract Results presented at the Wengen conference and reported here make clear that the science and art of climate change health impact assessment is advancing rapidly on many fronts, including the empirical understanding of mechanisms, methodologies for modeling future impacts, access to relevant data resources, and education and training. We can expect further advances on all of these fronts as the field of climate change health impact assessment continues its rapid development in coming years. While short-term and seasonal forecasts will continue to be of practical importance in short-term planning, medium to long-range climate/health projections also have an important role to play in longer-range planning, especially as it relates to large-scale infrastructure investments that have long lifetimes.
Keywords Climate: Impacts, Climate: Adoption, Climate: Training, Climate: Technical Assistance
Examples Presented For example, Knowlton et al.’s work illustrates a systematic approach to develop place-based empirical knowledge regarding climate and air quality impacts on human mortality and then to apply this knowledge to model potential future health impacts related to climate change. Integrated assessments of this kind will be of value to policy makers charged with adaptive planning to avoid future environmental health risks. For example, in the USA, long-range plans are developed to bring
1 Department of Environmental Health Sciences, Mailman School of Public Health, 60 Haven Ave, B-1, New York, NY 10032 2 Departamento Física de la Tierra II, Facultad de Físicas, Universidad Complutense, Ciudad Universitaria, 28040, Madrid, Spain 3 International Research Institute for Climate and Society, The Earth Institute at Columbia University, Lamont Campus, Palisades, NY 10964, USA
M.C. Thomson et al. (eds.), Seasonal Forecasts, Climatic Change and Human Health. © Springer Science + Business Media B.V. 2008
203
204
P. Kinney et al.
regions into compliance with standards for outdoor air quality and to assess the potential health benefits of regulations to reduce pollution emissions. This often involves running air quality models for future emissions scenarios. Knowlton et al.’s demonstration of potentially increased future risks of ozone- and heat-related mortality due to climate change in the eastern USA makes it clear that climate can no longer thought of as a constant in such analyses. New aspects of the work included the dynamically downscaled climate forecasts and the joint modeling of air quality and heat effects. Modeling the impacts of heat waves on human health has been the subject of much empirical research, with much of the literature focused on temperature alone. It is increasingly clear however that thermal discomfort and resulting ill-health is determined not just by ambient temperatures, but also by humidity, wind, solar intensity and potentially other factors. Matzarakis shows future climate change scenarios that examine changes in physiologically effective temperature (PET) which takes into account a broader set of met variables. He argues that the increases in PET are more significant and larger from a health perspective than are expected rises in temperature alone. Even more complicated are the relationships between ambient conditions and those experienced by vulnerable people in their daily lives, which would take into account factors such as housing quality, access to air conditioning, and baseline health status. Future projections of health impacts must be firmly based on empirical knowledge. Nastos presented time series analysis of asthma admissions to Athens hospitals over 2 years in relation to meteorological and pollution variables. The results illustrate well the complexities of analyzing asthma in relation to weather and pollution in a time series context, where seasonal cycles can obscure or confound underlying associations. Another critical aspect of climate/health is the need for ongoing training programs for public health practitioners. Practitioners need a basic understanding of the emerging science along with operational experience in using available data and models to understand, forecast, plan for and adapt to climate-induced health stresses. There is also a need for ongoing access to latest tools from climate science. Aron reported on an innovative training and grant development institute held at University of West Indies on climate impacts and health. Participants called for better access to new technologies like remote sensing and more opportunities for interactions with policy makers. Programs of this kind will be in great demand as adaptation to climate change rises to the top of many country’s environmental planning agendas. As noted, an important aspect of the knowledge needed is better access to data and models. The work of Tourre and colleagues illustrates an innovative program to package data products in forms that are useful to climate–health practitioners. Tourre and his French team presented a new web portal to a system with acronym RedGems that pulled together remote sensing and surveillance data to track public health problems, especially related to vector-borne diseases in developing countries. From the examples presented in this book it is clear that adverse health outcomes related to climate variability and change are not confined to developing countries. European initiatives such as Euro-Heat (http://www.euro.who.int/globalchange)
Recent Developments and Next Steps in Climate Change and Health
205
and health focused case studies presented in the recent international WMO Conference in Madrid on Social and Economic Benefits of Weather, Climate and Water Services (http://www.wmo.int/pages/madrid07) suggest this is an area of rapidly growing development with respect to learning to be adaptive to a changing climate both within the developed and developing world.
Color Plates Appendix 1
Fig. 1.1 An illustrative example of the predicted change, due to a doubling of CO2, in summeraverage precipitation (mm/day) over Europe from a small ensemble of 12 slab model simulations
207
208
Appendix 1
Fig. 1.2 Common domain for regional climate model simulations. The colours show the topographic heights
Appendix 1
209
Fig. 2.3 Correlations between January–March 1961–2000 mean sea-level pressure and the Southern Oscillation Index for the same period. Correlations stronger than ± 0.30 (covering about 40% of the globe) are statistically significant at a 1% level of significance. The mean sea-level pressure data are from the European Centre for Medium-Range Weather Forecasts (ECMWF) Reanalysis (Uppala et al. 2005), which is a model’s best estimate of the state of the global atmosphere given scattered observations
210
Appendix 1
Source: IRI (2005). Sustainable development in Africa: is the climate right? Technical Report NO. IRI-TR/05/1, International Research Institute for Climate and Society, New York. USA
Appendix 1
211
Fig. 5.5 An ecological model to identify areas at risk of epidemics in Africa (From Molesworth et al. 2003)
212
Appendix 1
Source: Control panels of the mosquito vector (adult and pre-imago sages) and parasite exogenous variales. Image Source: ars technica the art of technology www.media.arstechnica.com and Institute for Science and Technology in Medicine www.keele.ac.uk/research
Color Plates Appendix 2
Fig. 7.3 False color compositing of the NDPI during August 26, 2003 (Copyright CNES 2003, distribution SPOT-Image S.A., all rights reserved). Scale: 2 cm = 750 m (= 75 pixels)
Fig. 7.4 Ponds classification in terms of turbidity or NDTI, also for August 26, 2003 (Copyright CNES 2003, distribution SPOT-Image S.A, all rights reserved).
213
214
Appendix 2
Fig. 7.5 ZPOM for the same date as Figs. 7.3 and 7.4 (Copyright CNES 2003, distribution SPOTImage S.A, all rights reserved).
164 180
172
163
122
158
163
133 152
158 159
103
133 102
151
125 102
129
114
111
100 106 110 101 100 86
134
137
151
123
119
% increase in summer heat-mortality vs. 1990s 81 - 100% increase 101 - 120% 121 - 140% 141 - 160% 161 - 180%
Fig. 8.5 Map of percentage changes in heat mortality impacts by 2050s A2 vs. 1990s
Appendix 2
215
0.7 1.5
0.9
0.5
4.0
1.4
0.3
3.2 3.0
1.0
2.7
4.3
3.4
4.7 4.3
4.8
4.6 5.1
4.2 6.1
5.5
5.2 9.6
4.2
2.3
6.1
5.2 6.3 6.1
6.8
5.3
% increase, 2050sA2-90s O3 mortality 0 - 1.5% increase 1.5 - 3.0% 3.0 - 4.5% 4.5 - 6.0% > 6.0% increase
Fig. 8.6 Map of percentage changes in O3 mortality impacts by 2050s A2 vs. 1990s
Fig. 9.1 (continued)
216
Appendix 2
Fig. 9.1 (continued) Basic PET conditions (a) and differences between time slice (2070–2100) minus (1961–1990) for A1F (b) and B1A (c) for JJA
Appendix 2
Fig. 9.2 (continued)
217
218
Appendix 2
Fig. 9.2 Basic PET conditions (a) and differences between time slice (2070–2100) minus (1961– 1990) for A1F (b) and B1A (c) for DJF
Appendix 2
219 0-25%
25-50%
50-75%
75-100% 60
90%
55
80%
50
70%
45
60%
40
50%
35
40%
30
30%
25
20%
20
10%
15
0%
Total Number of BAA
Frequency (%) of BAA (2001-2002)
Minimum Air Temperature Percentiles 100%
10
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 35 36
10-Days Intervals Vapor Pressure Percentiles
0-25%
25-50%
50-75%
75-100% 60
55
80%
50
45
60%
40
35
40%
30
25
20%
Total Number of BAA
Frequency (%) of BAA (2001-2002)
100%
20
15
0%
10
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 35 36
10-Days Intervals
Fig. 10.3 Frequency (%) of the BAA per 10-Days Intervals (2001–2002) as a function of the minimum air temperature and water vapor pressure percentiles, along with the variation of the total number of BAA per 10-Days Interval (yellow line) and the polynomial fitting (brown line). Three reference lines (white lines) concerning the mean, the mean + SD and the mean – SD are also depicted
220
Appendix 2 Atmospheric Pressure Percentiles
0-25%
25-50%
50-75%
75-100%
100%
60
80%
50
45
60%
40
35
40%
30
Total Number of BAA
Frequency (%) of BAA (2001-2002)
55
25
20%
20
15
0%
10
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 35 36
10-Days Intervals
0-25%
25-50%
50-75%
75-100% 60
90%
55
80%
50
70%
45
60%
40
50%
35
40%
30
30%
25
20%
20
10%
15
0%
10
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 35 36
10-Days Intervals
Fig. 10.4 The same as in Fig. 10.3 but for atmospheric pressure and sunshine percentiles
Total Number of BAA
Frequency (%) of BAA (2001-2002)
Sunshine Percentiles 100%
Appendix 2
221 PM10 Percentiles
0-25%
25-50%
50-75%
75-100% 60
100%
55 50 45 60%
40 35
40%
30
Total Number of BAA
Frequency (%) of BAA (2001-2002)
80%
25 20%
20 15
0%
10 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 35 36
10-Days Intervals
SO2 Percentiles
0-25%
25-50%
50-75%
75-100%
100%
60
50 45
60%
40 35
40%
30
Total Number of BAA
Frequency (%) of BAA (2001-2002)
55 80%
25 20%
20 15
0%
10 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 35 36
10-Days Interval
O3 Percentiles
0-25%
25-50%
50-75%
75-100% 60
100%
50
80%
45 40
60%
35 30
40%
25 20%
20 15
0%
10 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 35 36
10-Days Interval
Fig. 10.5 The same as in Fig. 10.3 but for PM10, SO2 and O3 percentiles
Total Number of BAA
Frequency (%) of BAA (2001-2002)
55
Author Index
A Abayonmi, K., 46 Abeku, J. A., 49 Abeku, T. A., 34, 35, 49 Achtman, M., 94 Adegbola, R. A., 86 Afari, E. A., 34 Ahern, M., 45, 46 Ahmed, M. U., 39 Akenji, T. N., 34 Akhtar, R., 34 Allan, R., 45, 47 Allan, R. J., 16 Almendares, J., 44 Al-Nakib, W., 39 Amelung, B., 161–171 Amin, M. M. M., 39 Anderson, D. L. T., 19 Anon, 32, 35, 37, 48 Anyamba, A., 37 Armah, G. E., 39 Aron, J., 102 Aron, J. L., 189–201
B Baird, J. K., 34 Barbazan, P., 44 Barker, C. M., 37 Barnston, A. G., 2, 19, 22, 47 Barrera, R., 39 Barsugli, J., 21 Bartley, L. M., 49 Bartzokas, A., 174 Bascom, R., 174 Basu, R., 144, 147, 148 Bates, D. V., 181 Bavia, M. E., 36, 43 Ba, Y., 139
Belcher, D. W., 37, 43 Bell, M. L., 144, 147 Bengtsson, L., 22 Beniston, M., 131–133 Ben Salah, A. B., 36 Berktas, B. M., 174, 179, 184 Besancenot, J. P., 40, 41 Bhatia, S. L., 94 Bicout, D. J., 137 Bi, P., 34, 35, 38, 45, 48 Bircan, A., 174, 179, 184 Birley, M. H., 102 Birmingham, M. E., 90 Bittencourt, J. A., 39 Bjerknes, J., 18 Black, R. E., 32, 52 Blakebrough, I. S., 90 Bodker, R., 34 Botto, C., 33, 36 Bouma, M., 102 Bouma, M. J., 34, 35, 46, 50 Bouvier, P., 34 Braak, L., 135–140 Braun-Fahrlander, C., 174 Brengues, J., 43 Brewster, D. R., 35, 40, 50 Brooker, S., 33, 36 Broom, A., 38 Broom, A. K., 38 Broome, C. V., 91 Brunekreef, B., 174, 181 Burke, D., 32 Burnett, R. T., 174 Byun, D. W., 148
C Camargo, E. P., 34 Camargo, L. M., 34
223
224 Campagne, G., 40 Caton, R. B., 181 Cazelles, B., 39, 47, 51 Ceccato, P., 49 Cedeno, J. E. M., 34 Chakravarti, A., 39, 44, 45 Chaves, L. F., 36 Chavez, M., 41 Chavez, M. R. C., 41 Cheesbrough, J. S., 40, 41, 88 Chen, H. X., 38 Chevalier, V., 51 Ching, J. K. S., 148 Christophers, S. R., 34, 42, 43 Cochi, S. L., 94 Coeffier, E., 135–140 Coiro, J. R., 39 Collins, M., 7 Colwell, R. R., 47 Connor, S. J., 1–4, 32, 33, 35, 48, 71–82, 101–122, 127–129, 203–205 Constantin G., 41, 52 Cook, S. M., 39 Cooper, S. N., 40, 41, 43 Cordova, S. P., 38 Correia, J. B., 89 Corwin, A. L., 39 Cox, J., 49, 74 Craig, M., 73 Craig, M. H., 34, 43 Cross, E. R., 36, 43 Cubasch, U., 6 Cuevas, L. E., 85–97 Cunin, P., 40 Cunliffe, N. A., 39 Curriero, F. C., 144, 151, 156, 157 Cvjetanovic, B., 40
D Da Rosa e Silva, M. L., 39 DaSilva, J., 2, 35 Da Silva, J., 79 Davies, F. G., 37 De Dear, R., 162 Delfino, R. J., 174 De Magny, G. C., 41, 52 Depradine, C., 39, 45 De Wals, P., 88 Dicko, A., 34 Doblas-Reyes, F. J., 25, 104 Dockery, D. W., 174 Done, S. J., 38, 48 Dudhia, J., 148 Dye, C., 35, 46, 102
Author Index E Elfadil, A. A., 37 Elnaiem, D. E., 36 Emele, F. E., 88 Epstein, P. R., 45, 50, 102, 136 Ernould, J. C., 51
F Fanger, P. O., 162, 164, 166 Ferreira, M., 103, 110–113 Few, R., 45 Folland, C. K., 6, 21 Franco, A. A., 40 Franke, C. R., 36, 47 Fun, B. N., 39
G Gagge, A. P., 162 Gagneux, S. P., 88 Gagnon, A. S., 35, 39, 47 Galán, I., 174 García-Herrera, R., 131–133, 203–205 Gatton, M. L., 38, 43, 49 Gebre-Michael, T., 36, 43 Giha, H. A., 34, 44 Gill, C. A., 34, 35, 42–44, 46 Girgis, N. I., 40 Githeko, A. K., 107 Glantz, M. H., 16 Goddard, L., 20, 22, 46, 48, 77 Goldstein, I. F., 174 Gomez, C., 36 Gomwalk, N. E., 39 Graham, R. J., 22 Gratziou, C., 182 Grech, V., 174, 179, 184 Greenburg, L., 174 Greenwood, B. M., 35, 40, 41, 50, 89 Grell, G. A., 148 Griffith, D. C., 52 Griggs, D. J., 7 Grover-Kopec, E., 35, 48 Guell, A., 135–140 Guérémy, J.-F., 22 Gupta, R., 34 Gupta, U. C., 40 Guthmann, J. P., 34 Gyapong, J. O., 36, 42
H Haberberger, R. L. J., 40, 41, 44 Haffejee, I. E., 39
Author Index Hagedorn, R., 25, 104 Haines, A., 32 Hajat, S., 144, 147, 179 Hakre, S., 34, 51 Hales, S., 39, 47, 51 Halpert, M. S., 46 Hanlon, P., 40 Hansen, J., 148 Hansen, J. W., 32 Harley, D. O., 38 Harrison, M. S. J., 16 Hart, C. A., 85–97 Hashimoto, M., 174, 179 Hassan, A. N., 36 Hassan-King, M. K., 88 Hayhoe, K., 158 Hay, S., 73 Hay, S. I., 34, 35, 39, 50, 103 Hendrickx, G., 35 Hewitt, C. D., 5–10, 104, 127 Hoek, G., 174, 181 Hoerling, M. P., 24 Hogrefe, C., 143–158 Höppe, P., 162, 164, 165 Hoshen, M. B., 107 Houynoux, M. R., 148 Hulme, M., 164 Hunter, J. M., 37 Huq, A., 40, 44 Hurtado-Diaz, M., 39 Hu, W., 38 Hu, X., 94
I Ijaz, M. K., 40 Indeje, M., 49 Inwards, R., 15 Ito, K., 144, 147, 151, 157
J Jackou-Boulama, M., 94, 95 Jaffe, D. H., 174 Jalaludin, B. B., 174, 181 Jendritzky, G., 162, 164 Johns, T. C., 164 Julvez, J., 35
K Kabatereine, N. B., 33, 36, 42 Kane, E. M., 39 Kanojia, P. C., 38, 44 Karalis, J. D., 181
225 Kashiwabara, K., 174 Katsouyanni, K., 144 Kelly-Hope, L. A., 31–53, 72 Kilian, A. H., 35 King, R. J., 36 Kinney, P. L., 143–158, 203–205 Kiszewski, A. E., 74 Kittayapong, P., 45 Kleinman, L. I., 144 Kleinschmidt, I., 34, 43 Knowlton, K., 143–158 Koelle, K., 40, 41, 49 Kondo, H., 45 Konno, T., 39 Koopman, J. S., 39 Koppe, C., 6 Koram, K. A., 34 Kotinis-Zambakas, S., 181 Kovats, R. S., 32, 39, 46 Kristensen, T. K., 36 Kuhn, K., 32, 33, 41 Kumar, A., 24, 25 Kumaria, R., 39, 44, 45
L Lacaux, J. P., 135–140 Lafaye, M., 135–140 Lama, J. R., 40, 41, 47 Lapeyssonnie, L., 40, 41, 43, 44, 87, 91 Leonardo, L. R., 34, 36, 42, 43 Le Sueur, D., 34, 43 Levy, J. I., 144, 147 Li, C. F., 39, 44, 45 Lindblade, K. A., 35, 51, 118 Lindesay, J. A., 21 Lindsay, S. W., 35, 36, 102 Linthicum, K. J., 37, 43, 47 Lipfert, F. W., 144, 182 Lobitz, B., 41, 47 Loevinsohn, M. E., 34, 35, 50 Lovell, E., 39, 45 Luo, N., 88 Lynn, B., 143–158 Lynn, B. H., 148, 153
M Mabaso, M. L., 34, 44 Mabaso, M. L. H., 44, 47 Macdonald, G., 102, 107 Macfarlane, S. B., 53 MacLennan, J., 90 Maelzer, D., 38 Maguire, J. H., 36
226 Malaney, P., 73 Malkhazova, S. M., 36 Malone, J. B., 36 Mantilla, G. C., 48, 71–82, 104 Manuguerra, J. C., 136 Marriott, P. J., 15 Martens, P., 102 Martens, W. J. M., 102, 107, 115 Martin, P. M., 37 Martins, L. M., 36 Martin, V., 2 Mason, S. J., 13–26, 77 Mathur, K. K., 34 Matzarakis, A., 161–171 Mayer, H., 166 Mayer, P., 37, 43, 44 May, R. M., 102 McArthur, J. W., 32 McGullagh, P., 176, 177 McKenzie, F. E., 107, 110, 121 McLigeyo, S. O., 47 McMichael, A. J., 6, 32, 34, 45, 50 Menne, B. E., 1–4, 127–129 Mimmack, G. M., 23 Mitchell, J. F. B., 6 Molesworth, A. M., 40, 41, 49, 88, 92–96, 211 Molyneux, D. H., 42 Mondet, B. A., 138 Moore, P. S., 88, 91 Moosa, A., 39 Moreira, C. J. E., 34 Morse, A. P., 107 Mpabalwani, M., 39 Mulic, R., 38 Munyekenye, O. G., 34 Musawenkosi, L., 79 Mutanda, L. N., 39
N Nacher, M., 34, 36 Najera, J. A., 73, 74 Nakhapakorn, K., 39, 43 Namias, J., 21 Nastos, P. Th., 173–186 Ndegwa, W., 107 Ndiaye, O., 35 Ndione, J. A., 137 Nedler, J. A., 176 Nelder, J. A., 176, 177 Neronov, V. V., 36 Neuberger, M., 182 New, M., 163, 164 Nihei, N., 34
Author Index Nurgaleeva, R. G., 38 Nwoke, B. E., 37
O Odiit, M., 35, 43 Odongo-Aginya, E., 35 Okuno, T., 38, 43 Olaleye, O. D., 37 Omanga, U., 91 Omumbo, J. A., 34, 43 O’Neill, M. S., 144, 147, 157 Orlove, B. S., 15 Ostro, B. D., 174 Ozkaynak, H., 144
P Palmer, T. N., 6, 19, 22 Pandis, S. N., 144 Parton, K. A., 38 Pascual, M., 36, 41, 47, 50 Patz, J., 50 Patz, J. A., 32, 50, 102, 103 Pekkanen, J., 181 Peltola, H., 90 Peterson, A. T., 39 Phukan, A. C., 38, 39, 44 Pope, III C. A., 174 Porphyre, T., 136 Poveda, G., 35, 102, 103, 105 Preobragenski, V. V., 40
Q Qiu, F. X., 38
R Rajagopalan, B., 26 Ramasamy, R., 34 Ram, S., 39, 44 Ramsay, M. E., 90 Raso, G., 36, 43 Reingold, A. L., 89 Riordan, F. A. I., 89 Robbins, J. B., 89, 90 Robertson, A. W., 26 Rodo, X., 41, 47 Roeckner, E., 24 Rogers, D., 35 Rogers, D. J., 35 Rojas, W., 102 Ropac, D., 38
Author Index Ropelewski, C. F., 46 Rosenberg, R., 34 Rosenthal, J., 143–158 Rosenzweig, C., 143–158 Ruiz, D., 101–122 Russell, G. L., 148
S Sabatier, P., 135–140 Sachs, J. D., 32, 73, 103 Saha, S., 22 Salih, M. A., 89 Salim, A. R., 40, 41, 43 Samet, J. M., 144, 147, 148 Sarkies, J. W., 41 Savory, E. C., 49, 88, 91, 94 Scheld, W. M., 86 Schemann, J. F., 40 Schreiber, K. V., 39 Schwab, L., 40, 41 Schwartz, J., 174 Seinfeld, J. H., 144 Semenza, J. C., 144 Shaman, J., 37 Shanks, G. D., 34 Sharma, V. P., 34 Sheikh, H. A., 40, 41, 43 Shukla, J., 19 Sidikou, F., 95 Singh, N., 34 Siziya, S., 34 Skalova, R., 40 Small, J., 34, 35 Snow, R. W., 34, 43 Soriano-Gabarro, M., 96 Spagnolo, J., 162 Speelmon, E. C., 41, 47 Srivastava, A., 51 Steele, A. D., 39 Steib, K., 37, 43, 44 Stensgaard, A., 36, 43 Stephens, D. S., 86, 87 Stockdale, T. N., 22, 48 Strickman, D., 45 Stuart, J. M., 88 Sultan, B., 40, 41, 49, 94, 95 Suwannee, A., 38 Swaroop, S., 34, 35, 42, 43
T Tebaldi, C., 7 Tedesco, L. R., 40
227 Teklehaimanot, A., 74 Teklehaimanot, H. D., 34, 35, 49 Ter Veen, J. M. L., 103 Thammapalo S., 39, 45 Theander, T. G., 44 Thiongane, Y., 2, 37 Thomas, C. J., 36 Thompson, D. F., 36 Thompson, R. A., 36 Thomson, M. C., 1–4, 6, 23, 31–53, 72, 73, 79, 85–97, 101–122, 127–129 Thonnon, J., 37 Thurston, G. D., 144, 147, 151, 156, 157, 174 Timonen, K. L., 181 Tinz, B., 162 Tong, S., 38 Tourre, Y. M., 135–140 Traore-Lamizana, M., 37 Trape, J. E., 35 Trasande, L., 174 Travis, P., 32 Tripathi, N. K., 39, 43
U Uppala, S. M., 19, 209 Uzun, S., 36
V Valencia Tellería, A., 35 Van Den Dool, H. M., 16 Van Der Hoek, W., 34 Van Der Kaay, H. J., 34, 35, 46 Van Der Kolk, M., 34 Vasconcelos, H. B., 37 Vasconcelos, P. F., 37 Vitart, F., 3 Vittor, A. Y., 41
W Walker, N. D., 21 Walters, S., 174 Wanji, S., 33, 37 Ward, M. N., 21 Wedderburn, R. M., 176 Weiland, S. K., 174 Weinstein, P., 38 Wellde, B. T., 35 Werneck, G. L., 36 Wernsdorfer, W. H., 73 Whelan, P. I., 38 Whittle, H. C., 89
228 Williams, B. G., 35 Woodruff, R. E., 38, 47 Worrall, E., 73, 107, 120 Wu, P. C., 39 Wyss, K., 32
Y Yacob, K. B. M., 34, 42, 43 Yan, G., 107, 114–116
Author Index Yang, G., 34 Yang, G. J., 36, 43, 51 Yang, H., 103, 105, 107, 109–113 Yuksel, H., 174, 179
Z Zebiak, S. E., 17 Zhou, G., 35, 114 Zhou, X. N., 36
Subject Index
A Adaptation, 3, 82, 101, 105, 127, 133, 157, 158, 190–192, 195, 204 Africa, 2, 9, 10, 19, 20, 34–37, 39–44, 46–50 African Eye Worm, 37 African Monsoon Multidisciplinary Analysis (AMMA), 140 African trypanosomiasis/sleeping sickness, Ngana, 35, 41 Air quality, 7, 143, 145, 147, 148, 158, 166, 177, 178, 195, 203, 204 Ambient air pollution, 176, 181, 182, 186 Analogue, 16, 21 Anomaly/anomalies, 1, 2, 6, 15, 16, 20, 21, 24, 46–49, 74, 81, 95, 150 Atlantic ocean, 19 Atmosphere, 2, 13–24, 26, 46, 48, 104, 131, 132, 148, 179, 209 Avian flu, 136
B Bacterial, 31, 33, 40, 41, 85–87, 91 Best practice, 79–81, 129 Bioclimate, 163, 167, 171 Bronchial asthma admissions, 173–177, 179–186, 219–221
C Capacity-building, 105, 106, 135, 140, 191 Case surveillance, 48, 76, 78 Changing climate, 6, 51, 101, 109, 127, 143, 158, 205 Cholera, 2, 40, 44, 45, 47, 49, 52, 72, 136 Climate models, 5–7, 9, 10, 49, 104, 132, 148, 158, 167, 208 variability, 5, 7, 8, 10, 17, 21, 32, 47, 48, 52, 71, 74
Climate change, 3, 5–10, 32, 45, 50, 51, 82, 101–106 impacts, 6–7 projections, 7, 10 Climatic shifts, 31 Colombia, 3, 35, 36, 39, 46, 71, 82, 101, 104, 105, 193, 194, 197, 198, 200 Communication with decision-makers, 189, 195 Community Multiscale Air Quality(CMAQ) Model, 143, 148–151, 153 Cross-disciplinary, 189 Cross-sectoral, 189
D Dengue, 3, 31, 39, 41, 43–45, 47, 49–52, 71, 72, 81, 82 Dengue hemorrhagic fever, 39, 41, 43, 82 Disaster response, 189, 195 Dynamic models, 21–25, 105, 107, 109
E Early warning systems, 1–4, 6, 31, 33, 41, 45, 48–50, 71–73, 81, 82 Early detection, 49, 75, 78 Ecological models, 94, 211 El Niño-Southern Oscillation (ENSO), 2, 3, 18, 19, 46–48, 105 Emission scenarios, 143, 156, 161, 163 English-Spanish simultaneous translation, 189 Ensemble forecasting, 48 ENSEMBLES project, 5, 7–10, 82, 104, 127 Environmental/climate monitoring, 2, 3, 76, 77, 135, 136 Epidemics, 2, 6, 31–33, 41, 44, 46, 49, 60, 62, 71, 74–77, 79 Error, 15, 23, 25, 26, 116, 119, 147, 180 Exposure-response relationship, 143 Extreme events, 2, 3, 5–8, 10, 45–46, 132, 171 229
230 F Feedback, 9, 21, 22, 189, 195 Forecasts, 1–3, 13–15, 19–24, 26, 52, 77, 79, 81, 193, 200, 203, 204, 209
G General Circulation Models, 10, 24, 102, 161, 171 Generalized Linear Models, 173, 176, 177, 179–181, 186 Geographic distribution, 31, 133, 158 Geographic information systems, 43, 44, 47, 189, 196, 200 Global climate models, 49 Guinea worm, 37, 43, 44
H Health information systems, 135, 136, 140 Health surveillance, 82 Heat stress, 7, 42, 143, 144, 154, 157, 161, 162, 166, 190 Hemorrhagic fever with renal syndrome, 38, 45, 48 Human energy balance, 164 Humidity and meningitis, 40, 94
I Infectious disease, 1, 6, 31–33, 43, 51, 52, 72, 73, 103, 132, 190, 195 Initial conditions, 2, 7, 23–26, 104, 113, 148 Integrated modeling system, 143, 145, 148, 158 Inter-American Institute for Global Change Research (IAI), 189–192, 195, 196, 198 Inter-annual variability, 31, 45–48 Interdisciplinary proposal development, 189, 195 Integrated systems, 3 Interventions, 3, 31–33, 42, 43, 49, 52, 53, 73, 75, 77, 78, 90, 101, 103–105, 107, 109, 120, 122
J Japanese encephalitis, 38, 41, 43, 45, 51, 52
K Knowledge-action system, 189, 190
Subject Index L La Niña, 6, 13, 19, 46, 48 Lag effect, 173 Latin America and the Caribbean, 189, 190, 192, 198 Lead-time, 3, 21, 77, 78, 81, 105 Learning process, 109, 122 Leishmaniasis, 31, 36, 41–43, 47, 50, 52, 102, 193 Liverpool School of tropical Medicine, 31, 85 Lymphatic filariasis, 33, 36, 42, 43, 53
M Malaria, 2, 3, 6, 31, 34, 35, 42–53, 71–82 control, 73–75, 78–81, 101, 104, 193 epidemics, 46, 47, 74–76, 79, 80, 109 surveillance, 2, 3, 48, 49, 73, 76, 78–80, 82, 86, 90, 95 Malaria Early Warning and Response Systems, 75–78 Mapping malaria risk, 43 Mathematical models, 101, 103–105, 108, 109, 111, 113, 116, 120, 121, 192 Meningitis belt, 41, 86–89, 91, 92, 94 Meningitis control, 97 epidemiology, 86–89, 91, 95 epidemics, 93, 95 and environment, 85–97 Meningococcal carriage, 88, 89 meningitis, 2, 31, 40, 41, 43, 44, 49, 88–90, 95, 97, 140 vaccines, 90 Metadata catalog, 139 Model evaluation, 94 dynamical, 21–25, 105, 107, 109 statistical, 21, 23, 26, 91, 92, 104, 105, 147 Modelling, 21, 104, 107, 120, 140 Mortality, 3, 31, 33, 44, 52, 85, 86, 88, 96, 108, 110, 132, 143–146 Mosquito-borne illnesses, 195 Multidisciplinary networking, 135, 140 Multi-model ensembles, 6, 26, 48, 79, 107–109, 117, 120–122 Murray Valley encephalitis /Australian encephalitis, 38, 41
N NDVI and meningitis, 95 Network of professionals, 192
Subject Index New York, 31, 71, 85, 101, 127, 143–145, 149, 151, 154, 157, 158, 174, 203, 210 Normalized Deviation Pond Index, 138, 139, 213
O Ocean, 2, 3, 13, 16–19, 21, 22, 24, 26, 46–49, 148, 163, 195 Onchocerciasis/River Blindness, 33, 36, 42, 43, 53 Operational experience, 189, 195, 204 Outreach project, 189, 190 Ozone, 132, 143, 144, 151, 174, 176–178, 181, 182, 190, 195, 199, 204
P Pacific Ocean, 13, 16–18, 46, 49 Parasitic, 31, 33, 34, 103 Physiologically Equivalent Temperature (PET), 161–171, 204, 216, 218 Poisson distribution, 173, 177, 180 Poisson regression modeling, 143 Prediction, 2, 5–7, 9, 10, 13–15, 20–26, 45, 48, 76, 77, 81, 91, 95, 103–105, 109, 119 Predictive models, 31, 33, 95, 97 Prevention, 1, 71, 74–76, 78, 80, 81, 89, 140 and response, 76 Probabilistic information, 3 Probability/probabilistic, 2, 3, 5, 7, 8, 10, 22, 26, 76, 80, 104, 115, 120–122, 133, 179, 181
R Rainfall and meningitis, 41, 91, 94, 95 RayMan, 164, 167 Re-Emergent Diseases Global Environment Monitoring from Space (REDGEMS), 135–140, 204 Regional climate models, 9, 10, 132, 148, 158, 208 Remote sensing, 43, 44, 47, 133, 136, 189, 193, 196, 204 Response, 2, 4, 6, 33, 71, 75–80, 86, 95, 96, 102, 105, 108 Rift Valley Fever, 2, 3, 37, 41–43, 47, 49, 51, 52, 136–138, 140 Risk analyses, 31, 33 areas, 49, 90 mapping, 31, 43, 49, 81, 97 Ross River virus/Epidemic polyarthritis, 38, 41, 43, 47–49 Rotavirus, 31, 39, 43, 44, 52
231 S Sahel, 41, 42, 74, 91, 138 Schistosomiasis / Bilharzias, 31, 33, 36, 42, 43, 50–53, 133, 193 Sea surface temperatures (SSTs), 2, 3, 6, 17, 18, 20, 21, 24, 41, 46–49, 140 Seasonal climate forecasting, 3, 6, 13–16, 20–22, 26, 76, 77, 101, 104 Seasonality, 31, 33, 43–44, 52, 73, 88, 102, 117, 132, 181, 184 Sea-surface temperatures / sea temperatures seasonal climate forecasting Southern Oscillation teleconnections tercile trade winds uncertainty weather Simulation, 5, 8–10, 24, 25, 101, 108, 109, 112, 113, 116 Spatial Surveillance of Epidemics, 135–138 SPOT-Image, 136, 139, 140, 213, 214 St. Louis encephalitis, 37, 41 Sub-Saharan Africa, 9, 72, 76, 91, 102 Surveillance, 2, 3, 48, 49, 73, 76, 78–80, 82, 86, 90, 95, 96
T Teleconnections, 16, 19 Tele-epidemiology, 135, 136 Tercile, 2, 23 Thermal comfort, 161, 162, 164 perception, 161, 166 Tourism, 171 Trachoma, 40–43, 52, 53 Trade winds, 17–19 Training institute, 189–192, 195–197, 199–201 Transmission dynamics, 6, 42, 50, 101, 113 Telemedicine, 135, 136
U Uncertainties, 2, 5, 7, 8, 13, 21–26, 77, 101, 104, 122, 132, 144, 163, 164, 171 University of the West Indies (UWI), 189, 190, 199–201 Urban climate, 143, 166
V Viral, 31, 33, 37, 86, 195 Vulnerability assessment and monitoring, 76–78
232 W Water-borne disease, 195 Weather, 1, 5, 6, 13–17, 19–23, 25, 26, 45, 46, 48, 49 WHO Southern Africa Inter-Country Programme for Malaria Control (SAMC), 79
Subject Index Y Yellow Fever, 37, 41, 46, 52, 102