BIOENGINEERING, THERMAL PHYSIOLOGY AND COMFORT
Studies in Environmental Science Volume 1 Atmospheric Pollution 1978 Proceedings ofthe 13th International Colloquium, held in Paris, April 25-28,1978 edited by M. M. Benarie Volume 2 Air Pollution Reference Measurement Methods and Systems
Proceedings of the International Workshop, held in Bilthoven, December 12-16, 1977 edited by T. Schneider, H. W. de Koning and L. J. Brasser Volume 3 Biogeochemical Cycling of Mineral-Forming Elements
edited by P. A. Trudinger and D.
1. Swaine
Volume 4 Potential Industrial Carcinogens and Mutagens by L. Fishbein Volume 5 Industrial Waste Water Management by S. E. jnrrgensen Volume 6 Trade and Environment: A Theoretical Enquiry by H. Siebert, 1. Eichberger, R. Gronych and R. Pethig Volume 7 Field Worker Exposure during Pesticide Application
Proceedings of the Fifth International Workshop of the Scientific Committeeon Pesticides of the International Association on Occu ational Health, held in The Hague, October 9-1( 1979 edited by W. F. Tordoir and E. A. H. van Heemstra-Leq uin Volume 8 Atmospheric Pollution 1980 Proceedings of the 14th International Colloquium, held in Paris, May 5-8, 1980 Volume 9 Energetics and Technology of Biological Eli m ination of Wastes Proceedings of the International Colloquium, held i n Rome, October 17-19, 1979 edited by G. Milazzo Volume 10 Bioengineering, Thermal Physiology and Comfort
edited by K. Cena and J. A. Clark
Studies in Environmental Science 10
BIOENGINEERING, THERMAL PHYSIOLOGY AND COMFORT edited by
K. Cena Environmental Physics, Institute of Building Science, Technical University of Wroclaw, Wybrzeze Wyspiahskiego 27, 50-370 Wroclaw, Poland and
J. A. Clark Environmental Physics, Department of Physiology and Environmental Studies, University of Nottingham School of Agriculture, Sutton Bonington, Loughborough, LEI2 5RD, Great Britain
ELSEVIER SCIENTIFIC PUBLISHING COMPANY Amsterdam - Oxford - New Y o r k 1981
Distribution of this book is being handled by the following publishers: For the U.S.A. and Canada ELSEVIER/NORTH-HOLLAND, INC. 52 Vanderbilt Avenue New York, New York 10017 For Albania, Bulgaria, Chinese People’s Republic, Cuba, Czechoslovakia, German Democratic Republic, Hungary, Korean People’s Democratic Republic, Mongolia, Poland, Romania, the U.S.S.R., Vietnam and Yugoslavia ARS POLONA Krakowskie Przedmieicie 7 00-068 Warszawa, Poland For all remaining areas ELSEVIER SCIENTIFIC PUBLISHING COMPANY 355 Jan van Galenstraat P.O. Box 211, 1000 AE Amsterdam, The Netherlands
Library of Congress Cataloging in Publication Data Main entry under title: Bioengineering, thermal physiology and comfort. (Studies in environmental science; 10) Papers presented at a conference held Sept. 4-7, 1978, at Karpacz, Poland, and sponsored by the Technical University of Wroclaw. Bibliography: p. Includes index. 1. Heat - Physiological effect - Congresses. 2. Man - Influence of climate - Congresses. 3. Clothing, protective Congresses. 4. Insulation (heat) Congresses. 5. Bioengineering - Congresses. I. Cena, Krzysztof. 11. Clark, Jeremy Austin, 1938111. Politechnika Wroclawska. IV. Series. [DNLM: 1. Biomedical engineering. 2. Adaptation, physiological. 3. Body temperature regulation. QT34 B6031 QP82.2.514BSG 612’.014462 80-1 6578
-
-
ISBN 0-444-99761-X (voI. 10) ISBN 0-444-41696-X (series)
Copyright @ Wroclaw Technical University Press, 1981
All rights rsservcd. No part of this publication may be reproduced, stored in retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the publisher. Printed in Poland
CONTENTS SYMBOLS AND UNITS
. . . . . . . . . . . . . . . . . . . . . . .
6
I. PHYSICAL PRINCIPLES AND MEASUREMENTS
. . . . . . MONTEITH andA. E. WHELDON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Measurement of thermal balance of man. Y . NISHI 3. Evaluating the effects of clothing on the wearer. R . F. GOLDMAN. . . . . . . . 4. Human skin temperature and convective heat loss. R . P. CLARK . . . . . . . . 1 The physics of the microclimate. J A CLARK. A. J MCARTHUR. J L
13 29 41 57
I1. MODELS AND INDICES OF HEAT EXCHANGE 5 . Rational temperature indices of thermal comfort. A . PHARO GAGGB . . . . . . 19 6. Required sweat rate as an index of thermal strain in industry. J J VOGT. V CANDAS. 99 J P LIBERTand F DAULL . . . . . . . . . . . . . . . . . . . . . . . .
.
. .
..
.
.
7 Modelling of heat transfer in man. Y. HOUDAS. . . . . . . . . . . . . . . . 111
I11. PHYSIOLOGY. WORK AND EXERCISE 8. 9. 10. 11. 12.
.
Exercise physiology and sensory responses. R . R GONZALEZ . . . . . . . . 123 Thermal physiology of man in the aquatic environment. I HOLMERand U BERGH 145 Climatic change and acclimatization. G E . FOLK Jr . . . . . . . . . . . . . . 157 1 69 Man in extreme environments. problems of the newborn and elderly. D ROBERTSHAW Physiological signals for thermal comfort. M CABANAC . . . . . . . . . . . 181
.
.
.
.
.
.
IV COMFORT. ITS SPECIFICATION AND CONSEQUENCES
.
.
13 Design requirements for a comfortable environment. D. A MCINTYRE . . . . . . . . . . . . . . . . . . 14. Prediction of local discomfort for man. P 0 FANGER 15. The dependence of comfortable temperatures upon indoor and outdoor climates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . M A HUMPHREYS 16. The effects of moderate heat stress on mental performance. D . P . WYON I ANDERSEN and G . R . LUNDQVIST. . . . . . . . . . . . . . . . . . . .
. .
. .
.
195 221 229 251
V . SUMMING-UP
.
17 Physics. physiology and psychology. K . CENAand J . A . CLARK . SUBJECT INDEX
. . . . . . . .
..............................
271 285
This Page Intentionally Left Blank
SYMBOLS AND UNITS
The symbols used in the text, the quantities they represent and the appropriate S. I. units are listed below. Some local variants, which appear only briefly, are defined at the appropriate point in the text. Mean values of quantities are denoted by a bar over the symbol, in the conventional way. Most work in the field covered by this book now employs the units of the Systeme Internationale, which has considerably eased the task of the editors. However, there remain a few differences of practice. In particular, while many engineers continue to specify water vapour pressure in Torr (mm of Mercury), the majority of climatologists use the millibar; strictly, neither is an S. I. unit, though the latter sometimes disguise the millibar as the hecto-Pascal (lo2 Pa). On Professor Gagge's suggestion the S. I. unit of pressure, the Pascal, was adopted in this text. The multiple of the kilo-Pascal (kPa) is that used, as it is more convenient numerically. All vapour pressures in the text and f i g m have therefore been expressed in kPa, the form most likely to be met in future. The conversions between kPa and the Torr and millibar are relatively straightforward. When the water vapour pressure appears in the numerator of an equation, or alone, the conversion is 1 kPa
=
10 mbar = 7.5 Torr.
When the vapour pressure is in the denominator the ratios are I kPa-'
= 0.1
mbar-'
= 0.133
Torr-l.
Latin letters Units
surface area surface area estimated from DuBois formula specific heat convective heat exchange body characteristic dimension water vapour pressure water vapour pressure in ambient air saturation vapour pressure at skin temperature evaporative heat transfer insensible perspiration, or diffusive heat loss maximum evaporative heat loss respiratory evaporative heat loss sweating evaporative heat loss acceleration due to gravity = 9.81 (without subscript) body height heat transfer coefficient for convection, in water or air combined sensible heat transfer coefficient (her = h,i-h,) evaporative heat transfer coefficient mass transfer coefficient linear radiation heat transfer coefficient tissue (skin) conductance total heat transfer by sensible heat whole body dry heat transfer total evaporative heat transfer whole body evaporative heat transfer respiratory heat transfer clothing insulation clothing insulation in clo unit
m-2
m2 J kg-'
K-'
w m--2 m kPa kPa kPa
w m-2 w m--2 w m--2 w m-2 W m--2 in s-2 n~ W m--2 K-' W m-2 K-' W m-2 kPa-l m s-l W m-2 K-' W m--2 K-I W m-2 W
w W
W m2 K W-' cia
6
J
k K L M
Mn P
Q r Rn RS RL S sh
t
T
T-a Tb Tc T'l
To Tr
Ts TS" Tw V
v W
rate of heat storage thermal conductivity conductive heat exchange thermal irradiance metabolic rate net metabolic rate barometric pressure rate of heat production per unit volume of tissue resistance to heat or mass transfer net radiant heat transfer net shortwave radiation net thermal radiation rate of sweat production hourly rate of sweat production time temperature air temperature mean body temperature body core temperature dew point temperature operative temperature radiant temperature skin temperature surface temperature wet bulb temperalure air (wind) speed volumetric rates of blood flow, O2 consumption etc, as in text rate of production of external work
w m-2 W m-I K-l W m--2 W m--2 W W m--2 kPa W m--3 s m-1 W m--2
w m--2 W m--2 g m--2 s-* g m-2(hour)-' S
"C or K "C "C "C "C "C "C "C "C "C
m s-l m3 s-l W mW2
Greek letters Y R
e d
P
psychrometer constant latent heat of vaporization of water density Stefan-Boltzmann constant = 56.7 x lo-' ratio of transfer coefficientsfor sensible and latent heat (when not equal toy)
Dimensionless quantities im
F Gr
Le Nu Re RH Sh E
71 w
clothing permeability index for vapour transfer radiation form factor or clothing factor, according to subscript in text Grashof number Lewis number Nussel t number Reynolds number relative humidity Sherwood number emissivity for thermal radiation efficiency, sweating or mechanical work skin wettedness
kPa K-l J kg-l kg m--3 WmV2 K-4 kPa K-'
7
Clothing resistance to water vapour Ioss Water vapour transfer is modified by the presence of clothing, as is sensible heat transfer. The equation for sensible heat transfer has the form of Current
Potential Difference Resistance to Flow
=
'
The equivalent equation for evaporative heat transfer E i s a form of the psychrometer equation
where be is the vapour pressure difference between the skin surface and the environment, in kPa. For Zv to be expressedinunits of m2 K W-l , and therefore be identified with the clothing insulation, I, X must be a constant with dimensions K kPa-l. When sensible heat and water vapour transfer are both at rates well in excess of these possible by molecular diffusion, so that the transfer mechanisms are equivalent, Zv is equal to Z. Then it can be shown that at about 30"C, close to normal skin temperatures, E
=
16.5AejZ.
This gives the fortuitous benefit that when vapour pressures are expressed in kPa and insulation in clo (1') Ae E = 16.5 0.155 I' ' which may be conveniently rounded to give the equation used by GOLDMAN in chapter 3 E
==
100
-.Ae I'
This Page Intentionally Left Blank
PREFACE Man is in many ways a unique animal and one of his most obvious differences from other species is his ability to “bioengineer” his environment, so that he is protected from the thermal extremes of ambient conditions. The physical principles of heat exchange are, however, common to all animals and even poikilotherms of mass greater than a few grams employ behavioural thermoregulation, which is essentially physical. Physiological thermoregulation enables homeothermic animals to tolerate a much wider range of environments than would be possible by behavioural thermoregulstion alone. The third strategy of thermoregulation, “bioengineering” the creation of a tolerable or comfortable environment where none previously existed is largely man’s. In nature the most obvious examples are the nests of small birds and mammals, but man’s control of fire and his building technology, which have become scientifically based in the present century, allow him to live in all terrestrial environments but the highest mountains. For many millenia man’s construction of shelters may have been primarily for protection from predation and from extremes of climate, but his increasing sophistication has long made comfort a prime requirement in buildings. In many parts of the world ‘native’ buildings were evolved, which were satisfactory for their particular conditions. However, the spread of a common technology has usually superseded traditional methods of construction, and only a scientific understanding of building will allow the design of comfortable environmen‘s with new materials and methods in any environment. The design of comfortable environments also presupposes a knowledge of what is comfortable. Hence there have been numerous studies of the thermal physiology of man and its relationships to his sensations of thermal comfort and discomfort, both at rest and during work. Despite man’s ability to engineer comfort over a wide range of ambient conditions, he is still, occasionally, faced with an environment in which his first priority is survival. In these circumstances it is his physiological competence in thermoregulation which is first tested, and secondly his ability to acclimatize to stress. For the very young and very old even a normal indoor environment may present a potentially lethal challenge, but even a healthy adult may experience tliermoregulatory stress during prolonged physical work or athletic endeavour. The thermal challenge presented by the aquatic environment may also provide useful information about man’s thermoregulatory ability. The purpose of this book is to review current knowledge of the physics and thermal physiology of man’s reactions to his thermal environment, outlined above, and how these affect his comfort, well-being and work performance. This is necessary information for the bioengineering of appropriate environmental Conditions for man’s various activities. Though the book was conceived as an independent entity, the Technical University of Wroclaw convened a school using the same title, “Bioengineering, Thermal Physiology and Comfort”, at which the authors came together to present and discuss their papers. This school was held at Karpacz, Poland, o n
10
4-7 September, 1978. Matters raised during the school have been incorporated in the concluding chapter. Neither publication of this book, nor the holding of the school at Karpacz, would have been possible without the generous support of the authorities of the Technical University of Wroclaw. In particular, the editors wish to acknowledge the personal patronage of the school by the President of the University, Professor T. P O ~ B S K I . The editors’ co-operation in preparation of the manuscripts for publication was also greatly facilitated by grants which enabled reciprocal visits to Poland and Great Britain in 1977-1978. Visiting fellowships were received from the Technical University o f Wroclaw (by J.A.C.) and the Science Research Council of Great Britain (by K.C.), and travel grants were awarded by the British Council and the University o f Nottingham (to J.A.C.) and by the Technical University of Wroclaw (to K.C.). On a more personal note the editors wish to thank all the authors, not least for their co-operation in agreeing to the editing changes inevitably required to produce a text of unified format from seventeen individual papers. In addition to some of the authors, J. Narqbski, P. Poczopko and J. A. J. Stolwijk acted as chairmen of sessions at the school. We also wish to acknowledge the help of A. J. McArthur in the initial stages of editing and of E. Sliwiliska, who assisted in the compilation of the index. We are also indebted to the special care and attention of those concerned with preparing the text for printing, in particular M. Gutterwil and E. Sobesto of the Wroclaw Technical University Press.
K. Cena, J. A . Clark
I. PHYSICAL PRINCIPLES EASUREMENTS
This Page Intentionally Left Blank
Chapter 1
THE PHYSICS OF THE MICROCLIMATE J. A. CLARK, A. J. MCARTHLTR, J. L. MONTEITH Environmental Physics, Department of Physiology and Environmental Studies, University of Nottingham School of Agriculture, Sutton Bonington, Loughborough, LEl2 SRD, Great Britain.
A. E. WHELDON Department of Child Health, University of Nottingham Medical School, Queen’s Medical Centre,Nottingham, NG7 2UH. Great Britain.
CONTENTS Introduction Metabolic heat production Conduction Heat storage Radiation Thermal radiation Solar radiation Net radiation Convection Free convection Forced convection Mixed convection Evaporation Evaporation and convection Conclusions
INTRODUCTION
All animals, including man, respond to their surroundings and are particularly sensitive to the state of the thermal microclimate. Many cold-blooded animals attempt to keep body temperature within a preferred range by behavioural methods, moving into sunlight or shade. In contrast, homeotherms can maintain a constant body temperature by physiological responses to changes of microclimate, though these adjustments are often supplemented by ingenious stratagems of behaviour,
14
J. A. CLARKet al.
either by individuals or within groups. Man uses both physiological and behavioural methods of temperature regulation, but he has also developed the skill to control his own microclimate by heating and by air conditioning, or, less extravagantly, by selecting appropriate clothing. In this book, we review the processes by which man achieves thermal equilibrium with his environment. Some of these processes are physiological and involuntary, some are consciously controlled, and some depend on the subtle perception of “comfort”. Whatever process is involved, thermal equilibrium between a man and his environment depends on the physical mechanisms which govern heat transfer from the body core to the skin surface and from the skin through clothing to the environment. A review of the physics of microclimate is therefore an essential introduction to the physiological and psychological aspects of the subject which other authors will describe. The thermal comfort of man in an everyday environment, such as the home, factory or office, depends partly on the rate at which he exchanges heat with the environment and partly on his own level of work. In extreme thermal environments his tolerance of discomfort and, ultimately, his survival are decided by the competence of his thermoregulatory physiology. The heat balance equation, as employed in thermal physiology, therefore has both physical and physiological implications. Equation (1.1) is an application of the First Law of Thermodynamics - the law of energy conservation - and therefore satisfies the requirement that the sum of heat inputs, outputs and storage must be zero (MONTEITH, 1973; CAMPBELL, 1977). For practical purposes the components of the heat balance are usually expressed in power per unit area of the external surface of the body (e.g. W m-”. A convenient algebraic expression is M,,+R+H+G+J== 0. (1.1) When the energy conservation equation is expressed in this form, the balance is between terms which are heat gains by the body and those which involve heat dissipation. The usual sign convention is to ascribe a positive sign to the numerical values of the gains. Therefore metabolism is always positive and the other terms are usually negative. M, is the net flux density of metabolic heat, i.e. the total heat production of the body less the enthalpy transferred to the respiratory air stream per unit time, expressed per unit surface area. For a resting adult M, is about 60W m-’, and may increase by a factor of approximately 20 during exercise. R is the net radiative energy exchange between the surface and its surroundings and H is the total heat transfer by convection with the surrounding medium. Because the environment is usually specified in terms of temperature and humidity, H is usually divided into convection ( C )and evaporation (E),both carried by convective processes, so that
H = C+E. (1 -2) The error incurred by considering sensible and evaporative heat transfer separately is usually negligible. The circumstances in which this is not so are considered later. C is the heat transfer by conduction to solid substrates and J is the rate of change
The physics of rhe microclimate
15
of heat storage in the body. The numerical value of J will therefore be negative when the body temperature is rising, i.e. when storage is contributing to heat dissipation. Subsequent sections of this chapter outline, in turn, the factors which determine each of the components of the heat balance equation. METABOLIC HEAT PRODUCTION
A combination of physical and physiological mechanisms enables the adult human to maintain an almost constant core temperature (T,) in a wide range of environments. Metabolic responses to different thermal environments are best illustrated by considering a simplified energy balance, in which G = 0 and the body is in thermal equilibrium (i.e. J = 0). The remaining terms are M,, E , and the total sensible heat exchange, C + R . If we assume that sensible heat transfer obeys Fourier’s Law of heat exchange (STRUNK,1971) CS-R = h#,-TJ,
(1.3)
where hT is the overall conductance for sensible heat exchange, in W m-2 K-I, and T, is the ‘effective temperature’, i.e. that of the equivalent isothermal environment. Then M,+hT(T,-T,)+E
= 0.
(1.4)
Because man does not pant effectively, respiratory heat exchange does not change independently of metabolism in a given environment. Therefore, if we exclude behavioural thermoregulation, we need to consider only three strategies for the maintenance of thermal balance. These are most easily shown on a diagram showing how M,, C+R and E change with environmental temperature, as in fig. 1.1.
Fig. 1.1. Diagrammatic representation of relationships between heat production, evaporative and non-evaporative heat loss
Envir w e n t o l ternpemture
A zone of hypothermia: B temperature of summit metabolism and incipient hypothermia: C critical temperature; D temperature of marked increase in evaporative loss: E temperature of incipient hypenherma1 rise: F zone of hyperthtrmia; CD zone of least thermoregulatory effort: CE zone of minimal metabolism: BE thcrmoregulatory range. From MOUNT (1974)
The “metabolic diagram’’ (MOUNT, 1974) or “thermoneutral profile” (FOLK, 1974) is typical of a homeotherm with constant external insulation (i.e. a fixed insulation between the skin and the environment), The diagram may be divided into a number of zones, which correspond to the different strategies of physical and
16
J. A. CLARKet al.
physiologicsl control of the heat balance. At a fixed level of food intake and activity, the rate of mctrtbolism is minimal in the “thermoneutral zone”, between the lines C and E in fig. I . I . However, Mount defined a narrower zone, between lines C and D, as the “zone of least thermoregulatory effort”. This zone corresponds to the operation of the first and most economical strategy, that of regulating heat exchange by body conductance - brought about by v‘isoconstriction or vasodilation of capillnry b1oc.d vessels in the periphcral tissue. The more restricted zone of human comfort may be identified with the upper end of this zone. A second strategy is used at lower temperatures. Below the lower critical temperature (C in fig. 1.1.) evaporative heat loss and tissue conductance are minimal and ahnost constant, and h, is a minimum. Hence sensible heat loss from the body increases in proportion to the difference bctween T,and T, and, to maintain a steady core temperature, metabolic heat production must increase to compensate. The slope of the increase is proportional to the total thermal conductance of the body, and therefore depzids on both body insulation and clothing. According to BURTON and EDHOLM (1969), most adult mammals can increase their metabolic heat production by a factor of about three in response to cold stress alone, but much higher rates may be xhieved i n the short term by voluntary exercise. Because of his limited tissue insulation the lowest temperature a t which a naked adult may survive for long periods is about 2 “C, a value which, according to BURTONand EDHOLM (1969), is consistent with historical observations of the habitually naked natives of Tierra del Fuego. Most rc;rders of these p e e s are, however, much less hardy (FOLK, chapter lo). Shivering contributes to the involuntary response to cold. Non-shivering thermogenesis may also be found i n mammals. It is associated with brown fat deposits and has been observed in human babies (HEY,1974) though it is usually absent in adult humans. The zbility to increase metabolic rate in response to cold stress is reduced in the sick, the elderly and babies, in perticular. These groups are therefore more susceptible to hypothermia than the general population (ROBERTSHAW, chapter 11). The third strategy, that of regulating evaporative heat loss, is cmployed in hot conditions when sensible heat transfer is insuscknt to remove metabolic heat, above D in fig. 1.1. This is a strategy for which man is uniquely well equipped. His ability to secrete sweat in large quantities enables the body to control latent heat loss over a wide temperature range, so that the .thermoneutral zone extends well above that for comfort. In most environments the upper limit of this zone, known as the “upper critical temperature”, is determined largely by physical factors, which limit the rate of difliision of water vapour away from tlie skin. However, the ability to sweat may be rostrictcd in both the younz ar?d elderly, and may become “fatigued” even in athletes (CLARK, chapter 4). Above the upper critical temperature ( E i n fig. 1.1) latent heat loss is enhanced by active mechanisms, which themselves increase M and the risk of 3yperthermia. Many maminnls exhibit panting in this region, but, as noted earlier, this is not important for man. In practice, because the highest rates of work Ere rarely sustained for long periods, tlicrnizl storage in the body is often critical for the toleration by men of heat loads exceeding those w!iich may be dissipated by pssive evaporation of sweat.
The pliysics of’die microcliniiite
17
The metabolic diagram is complicated by other factors: real environments are rarely either constant or isothermal. Hence in order to evaluate the thermal environment it is necessary to know the separate transfer coefficients for convection, radiation and evaporation even when the objective is a single index of “environmental temperature” (GAGGE,chapter 5 ) ; the external clothing insulation may be varied at will, while metabolic heat production varies both between people and with time for one person. Generally, the higher is M,,, the wider is the thermoneutral zone and the lower are the lower and upper critical temperatures and the thermal limits of comfort (fig. 1.2). Thus GONZALEZ (chapter 8) observes that exercising subjects are
.-V
0
Fig. 1.2. Diagram of response of net metabolic heat production M , to environmental temperature T,
n
2
\
\
\
\
L-Jb
8
comfortable at lower temperatures than those preferred by resting subjects. Conversely, when the metabolic rate is depressed the thermoneutral zone is narrower and displaced to higher temperatures. Clothing, which is discussed in greater detail by GOLDMAN (chapter 3), controls the rates of transfer of sensible heat and water vapour as well as satisfyings social conventions. Its effect on sensible heat exchange is conveniently considered in terms of insulation. We may replace h, in equation (1.3) by an equivalent thermal resistance IT := lz;’, where IT has units of m2K W-I. If we take the simplest case of an isothermal environment in which the surroundings temperatures for convection and radiation are the same, the total insulation may be divided into three components acting in series: the body insulation, I b ; that of the clothing, I ; and the combined external resistance for convective and radiative transfer to the surrounding environment, I,. Hence
C+R
= (Tc-Te)/Vb+Z+Ie).
(1.5)
In p,ractice the combined external insulation ( l + I e ) is usaally considered as the clothing insulation, since the two are not easily separable and generally I $- I,. 2
- Bioengineering
18
J. A. CLARKet al.
The conversion between m2 K W-' and the clo unit (GAGGE, BURTON and BAZETT, 1941) is 0.155 m2 K W-' = 1 clo. CONDUCIION
Heat transfer by conduction occurs between the human body and solid surfaces such as chairs, beds and floors with which it is in contact. However, it is usually a small component of the total heat balance, for two reasons: first, the contact area is usually small; second, the thermal conductivity of the substrate is usually low. HEY, KATZand O'CONNELL (1970) measured the contact areas of supine babies, and found values of approximately 10% of the total skin surface area on firm substrates. This figure is likely to be representative for any normal posture of children or adults, but the fraction may increase by a few percent on compressible upholstery surfaces. Both wood and padded upholstery - the usual surfaces supporting a sitting or lying human - have low coefficientsof thermal conductivity, so that heat exchange through these is low .However, since conductive heat exchange is directly proportional to the local temperature gradient, contact with cold surfaces of high thermal capacity and conductivity, such as metal, solid plastics and stone, may cause high local rates of heat transfer and consequently local discomfort (FANGER, chapter 14). Most analyses of human heat balance therefore assume that conduction is negligible, though few measurements have been made. HEY, KATZ and O'CONNELL (1970) reported that 20 % of sensible heat loss from naked babies in a perspex metabolism chamber was by conduction through the floor, but this figure was reduced to about 3 % by a 1 cm thick foam mattress. These figures are consistent with measurements of heat loss from farm animals.
HEAT STORAGE
When the mean body temperature is constant the rate of heat storage ( J ) in the body tissues is zero, by definition, and in practice it is negligible over long time periods, in excess of a day for adult humans. However, over short periods J can be an important component of the heat balance, for example following the onset of exercise (GONZALEZ, chapter 8) and in circumstanceswhere J determines the tolerance time for work in severe environments (VOGTet al., chapter 6). The total capacity for heat storage in the body can be estimated as the product of its thermal capacity with the maximum tolerable change in mean body temperature. It therefore depends both on body mass and on the difference between the actual body temperature and the tolerable limits. For a man of mass 75 kg the heat storage changes by about 3Od kJ for each degree Kelvin change in mean body temperature (assuming that the specific heat of the tissues is close to that of water). A body temperature rise of 1 K in an hour therefore represents heat storage at a rate of about 85 watts. Expressed per unit area of surface, J m -45 W m-' in the sign convention used in equation (1.1). Heat storage is therefore likely to be important in man for
The physics of zhe microclimate
I9
periods of up to a few hours. In small animals it is important for periods of just a few minutes and only in the largest mammals is thermal storage a useful mechanism on a diurnal time scale. Children have a lower thermal storage potential than adults, associated with their small body mass and lower mass to surface area ratio. They are therefore at greater risk than adults in “severe” thermal environments, which include water at sea temperatures normal in summer in temperate climates (HOLM~R and BERGH, chapter 9). RADIATION
Both solar radiation and thermal radiation of terrestrial origin are important in the radiative energy exchange between the surface of the human skin or clothing and its environment. Their contribution to the heat balance has been reviewed by CENA(1974). In a terrestrial environment the two bands of electromagnetic radiation are effectively separate, solar or short wave radiation lying largely in a band of wavelengths from 0.3 to 3 pm, while longer wavelength thermal radiation from local sources lies mainly between 3 and 100 pm. THERMAL RADIATION
Thermal radiation transfer is the result of an exchange involving both emission and absorption. All surfaces emit thermal radiation at a rate depending on their surface temperature (T,)and emissivity (E). The emitted flux L(W m-’) is given by the Stefan-Boltzmann equation L =EUC,
(1.6)
where a is the Stefan-Roltzmann constant (a = 56.7 x lo-’ W m-’ K-4). The emissivity of skin and of almost all clothing materials is close to unity, as is their absorptivity for thermal radiation, so that the human surface can be treated as a “black body” for radiation in this band. The net exchange of thermal radiation, RL, is therefore given by the sum of the incident and emitted fluxes RL = L,-aT;,
(1.7)
where L, is the incident thermal radiation. Surfaces in typical human environments have temperatures in the range 270-300 K, so LL is usually between 300 and 450 W m-2. RLis usually negative, and this component of heat loss is similar in magnitude to the convective component from the outside of our skin and clothing. Indoors R, is often the largest pathway of heat dissipation. SOLAR RADIATION
The exchange of shortwave radiation (0.3-3 pm) is both more variable and more complex, since it depends both on surface colour and on the geometry of interception (CLARK and CENA,1976). Outdoors, intercepted solar radiation often exceeds meta-
20
J. A. CLARKe! al.
bolic heat production even in cold climates; indoors other high temperature sources of radiation, such as lamps, furnaces and electric fires, may also provide highly asymmetric sources of radiant energy. The net shortwave radiation flux, R,, is the product of the incident energy flux RSiand absorptivity a (1.8)
The spectral absorptivity and reflectivity of skin and clothing are strongly correlated with their colour. Typical values of a for skin are 0.85 for negroid subjects a n d 0.68 for Caucasians. Extreme values for clothing are M 1 for black and 0.3 for white fabrics for 6000 K (solar) radiation (ROLLERand GOLDMAN, 1967). Representative values of a for 250 K and 1200 K radiation are 0.65 and 0.85 for Caucasian skin, and 0.8 and 0.9 for medium grey clothing, respectively ( FANGER, 1970). NET RADIATION
The net radiant energy exchange obtained by combining equations (1.7) and (1.8) is
R = R,+ R,
=
aR,,-+L,-~T,".
(1.9)
Variations in the net radiation exchange over the body surface can affect comfort. Net radiant fluxes are usually quoted as averages for the whole of the body surface, as this is the quantity required for an evaluation of the overall heat balance, but some measurements of local radiative heat exchange have been made (CLARK, CENAand MONTEITH, 1973). The mean radiant exchange R is obtained by integration
R = - -1 [ R A .
dA
(1.10)
Though mathematically trivial, this equation reminds us that the radiation flux may vary greatly from point to point on the body. Moreover, the area ( A ) available for radiation exchange may be considerably different from the actual skin area, A,, assessed by the DuBois formula. Two factors may work in opposite senses: Because of shape and posture the area of the human body available for radiant exchange is reduced to less than the DuBois area. Considerable areas of skin or clothing are normally in contact with adjacent areas and so not available for radiant exchange. Further, some areas exchange radiation with other areas of the body, a t least in part. Posture enables a wide range of control of the ratio AIA,, from about 0.96 for a spreadeagled man to about 0.5 for the foetal position (GUIBERT and TAYLOR, 1953; MITCHELL,1974b). Contact with substrates such as chairs also reduces the area available for radiation. In contrast, the addition of clothing increases the external surface area. Measured values of the ratjo of the area of the clothed body to A, range from unity for minimal clothing to about 1.5 for the heaviest arctic assemblies. According to FANGER (1970) the ratio is typically about 1.2 for a suited male. Part of the insulative value of clothing is therefore lost by a concomitant increase in the surface area available for radiative and convective heat exchange.
77le physics of the mircoclimate
21
Changing the area for radiant exchange by posture is an effective form of behavioural thermoregulation. Posture also determines convective exchange, but to a lesser extent because convection is eliminated only by physical contact between adjacent body surfaces. In many circumstances, however, social pressures or occupational demands limit postural thermoregulation : it is more practical for an office worker to don an extra layer of clothing to reduce heat loss than to adopt a foetal position. However, the posture of babies nursed naked in incubators may play an important part in their thermoregulation. RUTTER(private communication) observes that full-term babies often adopt a spreadeagled position when the incubator temperature is high. Similar behavioural thermoregulation has been observed in other young mammals (e.g. MOUNT, 1967). It is often convenient to describe the radiation environment 8y its “mean radiant temperature”, TR(K). This is defined by
TR= [ ( L , + C Z R ~ ~ ) / G ] ~ * * ~ .
(1.11)
Provided that TRand the mean skin temperature, T,, do not differ by more than about 20 K, which is normally the case indoors, equation (1.9) may be simplified further, using a linear approximation, by substituting from equation (I .I I), whence R
= h,(TR-Ts),
(1.12)
where h, is a radiation transfer coefficient with a value of about 6 W ni-’ K-’ at temperatures close to 300 K.
CONVECTION
For a naked man, the thermal insulation (I, = l/hJ o f the boundary-layer of air around his body is an important component of his total resistance to sensible heat loss. A decrease in the insulation provided by this layer, due, for instance, to an increase in the rate of air movement, can cause a substantial increase in his sensible heat loss. In contrast, the insulation provided by the boundary-layer around a man immersed in water is small. The skin temperature of a naked man will therefore be almost identical to the temperature of the surrounding water and his sensible heat loss is determined largely by the insulation of his skin and body tissue. This depends on the control of internal convection (HOUDAS ct al., chapter 7). A decrease in the boundary-layer insulation due to an increase in the rate of water flow will in this case have little effect on heat loss. Provision of an insulating layer of clothing (e.g. wet suit) will reduce beat loss considerably and allow the temperature of the skin to rise above. that of the surrounding water. For a fully clothed man in air, a decrease in boundary-layer insulation alone will have little effect on his heat loss in the cold, but in hot conditions it helps to dissipate absorbed solar radiation. In contrast, a decrease in clothing insulation due to wind penetration will markedly increase sensible heat loss (NEWBURGH, 1970). Convective heat transfer from a clothed man is enhanced during exercise, both in air (R. P. CLARK, chapter 4) and in water (HOLMER and BERGH,chapter 9). Ventilation
22
J. A. CLARK et al.
of clothing due to movement also reduces the insulation provided (BELDING et al., 1947; CROCKFORD, 1970). It is difficult to quantify the effects of convection on the insulation of clothing, but there are well accepted theories which can be used to describe the heat exchange across the boundary-layer which separates a man from the bulk of the fluid in which he is immersed. Convection is the interchange of warm and cold fluid. The movements of fluid which carry heat away from the body surface may be driven by two mechanisms: “free c ~ n v e c t i o ndue ~ ~ to density differences in the fluid associated with temperature gradients; or “forced convection”, due to external forces such as wind. In each case the relationship between the rate of heat transfer, C, and the temperature difference between the surface and the air which drives the heat transfer
AT
= (Ts-Ta)
can be written as
C = NUkATfd,
(1.13)
where k (W m-’ K-I) is the thermal conductivity of the fluid, d is the characteristic dimension of the body and Nu is the Nusselt number. Nu is a dimensionless group which expresses the ratio of the actual heat transfer coefficient for convection, h, (W rn-’ K-’ ), to that expected for conductive heat transfer through thickness d of the fluid, i.e.
NU = h,d/k.
(1.14)
The calculation of a rate of beat loss by convection therefore requires the estimation of Nu, which depends on the size and shape of the body, the nature of its surface and the fluid properties (MCADAMS, 1954; MITCHELL, 1974a). Empirical relations are available in the engineering literature for simple shapes such as cylinders and spheres (e.g. KREITH,1958) and Nusselt numbers for human heat loss may be estimated with acceptable accuracy by treating the body as a smooth cylinder of appropriate characteristic dimension (MITCHELL,1974b). FREE CONVECTION
In free convection the fluid adjacent to a warm body becomes less dense than that remote from it and rises to be replaced by colder fluid. The consequent movement of warm air around the human body was first demonstrated by LEWISet al. (1969), using schlieren techniques (see R. P. CLARK, chapter 4). According to MONTEITH (1973), the Nusselt number appropriate for free convection from a human body is given by Nu = 0.63Gr0*25Pro‘Z5,
(1.15)
where Gr, the Grashof number, equals agd3AT/v3and Pr is the Prandtl number for the fluid. The quantities a and v are the coefficients of thermal expansion of the fluid
The physics of rhe riticroclirnafe
23
and its kinematic viscosity respectively, and g is the accerelation due to gravity. In air, for which Pr = 0.71, the above equation simplifies to Nu = 0.58Gr0*25. For values of AT between 1 and 20 K the corresponding values of 11, for a standing human, with d = 1.5 m, lie in the range from about 1.5 to 3 W m-z K - l . For a prone man, the boundary-layer is thinner (d assumed m 0.2 m) and values of h, are larger, in the range 2.5 to about 5.0 W m-’ K- .The respective heat flux densities would be 2.5 and 110 W m-’. In contrast, for a man lying in water the Prandtl number at 10 “C is 9.5, so that Nu = I . I G I - ~ .Calculation ~~. of Nu predicts that in water a AT of 1 K will give a heat flux of about 150 W m-2, a value consistent with the results discussed by HOLMBR and BERGH(chapter 9).
’
FORCED CONVECTION
In forced convection rates of heat transfer are determined by relative motions of the fluid driven by “external” forces, which includes the movement of limbs in swimming, and posture will be less important. The forced convection Nusselt number for a smooth cylinder is given by Nu = 0.26Re0.60Pro-33 for lo3
(1.16)
< Re < 5 x lo4 and Nu = 0.026Re0*81
(1.17)
for4x104 < R e < 4 x 1 0 5 , where the Reynolds number, Re, is the dimensionless group relating the fluid velocity, v, to its kinematic viscosity, Y , and to d (i.e. Re = Y ~ / Y ) .For air, substitution of the numerical value of Pr in equation (1.16) gives Nu = 0.23 For a man with d = 0.2 m this leads to values of h, of about 6 and 23 W m-’ K-’ at air speeds of 0.5 m s-’ and 5 m s-I, respectively. In water at 10 “C,a flow rate of only 0.5 m s-l gives a rate of heat transfer per unit temperature difference of about one-and-a-half kilowatts per m2. MIXED CONVECTION
Between free and forced convection there is a zone of mixed flow, in which both mechanisms contribute significantly to heat transfer. For Gr/Re* < 0.1 forced convection occurs, and free convection dominates when Gr/Re2 > 16 (KREITH, 1958). The upper limit corresponds to an airspeed of about 0.3 ms-’ for adults; so that in typical indoor conditions we lose heat by mixed convection. Unfortunately, estimation of the Nusselt number for this regime is difficult: according to FANGER (1970) the accepted procedure is to calculate Nu for both free and forced convection and use the larger. However, CAMPBELL (1977) suggests that in some circumstances the two processes may be additive. Perhaps the best procedure is to base estimates
24
J. A. CLARK et a).
of Nu for this range on published measurements of heat transfer from cylinders and spheres by mixed convection (e.g. YUGE,1960; OOSTHUIZEN and MADAN,1970). EVAPORATION
Man always loses some heat by evaporation of water through the skin. At low ambient temperatures this loss is almost constant and of the order of I0 yo of M,. In the heat and during exercise, however, the evaporation of sweat removes large quantities of excess heat from the body. Man has a tremendous capacity to produce sweat in response to heat stress: sweat rates can be as high as 0.5 g s-' (about 2 kg h-') for short periods and 0.25 gs-' (0.9 kgh-') over several hours. These correspond to heat losses of 1.2 and 0.6 kW, respectively. For comparison, the metabolic heat production of man increases from about 100W at rest to about 1 kW during strenuous activity. Evaporative heat loss depends, however, not only on the ability of the body to secrete sweat but also on the physical properties of the environment: if the maximum rate of evaporation is less than the rate of sweat secretion, heat dissipation will be impeded. Heat and water vapour transfer through the boundary-layer of the body take place by similar processes. The evaporation rate S', expressed as a mass flux per unit area of skin, can therefore be described by an equation similar to equation (1.13) S' = ShDAX/d, (1.18) where D is the molecular diffusivity for water vapour in air (mZ s-')and Ax the difference in water vapour concentration between the skin surface and the air. The absolute humidity has units of g m-' and is related to the quantity usually measured, the vapour pressure, e, by 3~ = 2170 e/T, (1.19) where c is i n kPa and Tin K . T h e Sherwood number Sh is the mass transfer analogue of the Nussclt number. Whereas the units for the sensible heat transfer coefficient in current literature are almost always those of the S.T. system, W m-' K-' , the presentation of mass transfer is less consistent. This is because three units of vapour pressure are used in conjunction with S.I. units: The Torr or nim of Mercury is used principally by engineers; the millibar (mbar) is employed by meteorologists and climatologists; the Pascal (usually kPa) is the recommended S.I. unit. The conversions are 1 kPa = 10 mbar
= 7.5 Torr.
The kilo-Pascal is employed in the remainder of this text. EVAPORATION AND CONVECTION
Because moist air is less dense than dry air, rates of free convection may be determined partly by gradients of water vapour concentration as well as by those of temperature. When heat and water vapour transfer occur together the temperature
The physics of’the niicrocliinore
25
difference in the Grashof number in equation (1.15) (and in the corresponding mass transfer equation, equation (1.18)) should therefore by replaced by the difference in “Virtual Temperature” T,, given by
(I .20)
T, = Ta(1+0.38e/p),
where p is the atmospheric pressure, in the same units as e. The effects of water vapour gradients on free convection, both in the free atmosphere and within insulating layers, have received little attention. However, MONTEITH (1973) showed that the Grashof number for a naked man could be underestimated by a factor of about three if the effects of vapour pressure gradients were ignored. The corresponding underestimation of heat and mass flux is about 30%. CENAand MONTEITH (1975) have shown that related effects may be significant for free convection within mammalian hair coats, and this factor warrants further examination in clothing. . The Sherwood and Nusselt numbers for free convection are related by Sh = NuL~’.~’,
(1.21)
where Le is the Lewis number. For water vapour in air Le = 0.90. Therefore if either Sh or Nu are known the other can be estimated. However, except in conditions of extremely high heat load, rates of evaporation are usually determined by the rate of sweat secretion rather than by the transfer processes directly, though sweat is secreted so as to maintain energy balance. In the human environment, the appropriate equation for forced convection, corresponding to equation (1.21) is
.
Sh
= Nu
(1.22)
Insertion of the value of Nu appropriate to the characteristic dimensions of an adult human and typical indoor air speeds, corresponding to Re = 7 . 4 lo4, ~ gives Sh = 200. This value is consistent with the measurements of Sh = 160 for man by CLIFFORD et al. (1959). According to RAPr (1970) the agreement between theory and measurements is good. CONCLUSIONS
The thermal environment affects man through the transfer processes between his body and its surroundings. The mechanisms of these processes have been outlined in the present chapter, and their evaluation is reviewed later by NISHI (chapter 2). Complete specification of a thermal environment is obviously a complex problem (MONTEITH, 1974), and in general it is necessary to describe not only the temperature gradient between a subject and the fluid (usually air) in which he is immersed, but also the respective transfer coefficients for evaporation, convected sensible heat and radiation, and the radiant field (MCTNTYRE, chapter 13). Because of the complexity of real environments, within buildings as well as is the open air, effective temperature indices have been developed. The aim of these indices, reviewed by GACCE (chapter 5), is to specify the environment by a single number in units of temperature, which is the most easily understood determinant of heat loss,
26
J.. A. CLARK et al.
Such indices may help to specify comfortable or tolerable environments with greater precision. However, man’s perception of the environment is complicated by his own psychological and physiological reactions (CABANAC, chapter 12): he may become bored with neutrality and be stimulated by a modest thermal stress (WYON et al., chapter 16). Because psychological factors are important, it is difficult to create environments which satisfy everyone (FANGER, 1970), though air conditioning engineers are engrossed in this task. Perhaps they should reassess their aims: HUMPHREYS (chapter 15) has collected results which suggest that habituation plays a targe part in our judgement of comfort, and that the rate of change of temperalure in particular, may be as important for comfort as the precision of control.
REFERENCES BELDING H. S., RUSSELL H. D., DARLINQ R. C., and FOLKG. E. (1947), Analysis offactors concerned in maintaining energy balance for dressed men in extreme cold Efects of activity on the protective value and comfort of arctic clothing, Am. J . Physiol. 149,223-239. BURTON A. C. and EDHOLM 0. G. (1969), Man in a Cold Environment. (Facsimile of 1955 edition), Hafner, New York, London. CAMPBELL G. S. (1977), An Introduction to Environmental Biophysics, Springer-Verlag, New York. CENAK. (1974), Radiative heat loss from animals and, man [In:] Heat Loss from AnimalsandMan, eds: J. L. MONTEITH and L. E. MOUNT.Buttenvorths, London. CENAK. and MONTEITH J. L. (1975), Transfer processes in animal coats. 111. Water vapour dirusion, Proc. R. SOC. Lond. B 188,413423. CLARKJ. A. and CENAK. (1976), Solar and thermal radiative loadr in the energy balance of man, Eng. Med. 5. 75-78. CLARK J. A., CENAK., and M o m m J. L. (1973), Measurements of the local he ntbulance of animal coats and human clothing, J. Appl. Physiol. 35. 751-754. CLIFFORD J., KERSLAKE D. McK, and WADDELL J. L. (1959), The efects of wind speed on maximum evaporative capacity in man, J . Physiol. 147, 253-259. CROCKFORD G. W . (1970), Protective clothing for fishing crews, Proc. R. Soc. Med. 63, 1007-1008. FANGER P. 0. (1970), Thermal Comfort, Danish Technical Press, Copenhagen. FOLK G. E. (1974), Adaptation and heat loss: The past thirty years, [In:] Heat Loss from Animals and Man, 4s.: J . L. MONTEITH and L. E. MOUNT,Buttenvorths, London. A. P., BURTON A. C., and BAZEITH. C., 1941, A practical system of units for the description GAGGE of the heat exchange of m a n with his environments, Science 94.428430. GUIBERT A. and TAYLOR C. L. (1952), Radiation area of the human body, J. Appl. Physiol. 5, 24-37. HEYE. N. (1974). Physiological control of body temperature, [In:] Heat Loss from Animals and Man, eds.: J. L. MONTEITH and L. E. MOUNT,Butterworths, London. HEYE. N., KATZ G., and O’CONNELL B. (1970), The total thermal insulation of the newborn baby, J. Physiol. Lond. 207, 683-698. KREITHF. (1958), Principles of Heat Transfer, Int. Text Book Co., Scranton Pa. USA. LEWISH. E., FOSTER A. R., MULLAN B. J., Cox R. N., and CLARK R. P. (1969), Aerodynamics of the human microenvironment, Lancet, 1273-1277. MCADAMS W . H . (1954), Heut Transmission, 3rd edition, McGraw-Hill, New York. MITCHELL D. (1974a), Convective heat transfer jkom man and other animals, [In:] Heat Loss from Butterworths, London. Auiinals and Man, eds.: J . L. MoNreim and L. E. MOUNT, D. (1974 b), Physical basis of thermoregulation, [In:] Environmental Physiology, ed. : MITCHELL D. ROBERTSHAW, Physiology Series One 7, Buttenvorths, London.
The physics of the microclimate
21
MONTEITH J. L. (1973), Principles of Environmental Physics, Arnold, London. MOUNT L. E. (1967), The CIimatic Physiology of the Pig, Arnold, London. MOUNTL. E. (1974). The concept of thermal neutrality, [In:] Heat Loss from Animals and Man, eds.: J. L. MONTEITH and L. E. MOUNT,Butterworths, London. NEWBURGH L. H. (1968), PhysioIogy of Heat Regulation and the Science of Clothing, (facsimile of 1949 edition), Hafner, New York, London. OOSTHUIZ~N P. H. and MADANS. (1970), Combined convective heat transfer from horizontal cylinders in air, Trans. ASME 92, Series C, 194-196. RAPP G. M . (1970). Convective mass transfer and the coejjfcient of evaporative hear loss from the human skin, [In:] Physiological and Behavioiiral Temperature Regulation, eds. : J . D. HARDY, A. P. GAGOE,and J. A. J. STOLWIJK, Thomas, Illinois. W. L.. and GODMAN R. F. (1967), Estimation of solar radiation environment,Int. J . Biometeor. ROLLER 11, 329-336. STRUNK T. H. (1971), Heat loss from a Newtonian animal, J. Theor. Biol. 33,3541. YVOET., 1960, Experiments on heat transjer from spheres including combined natural and forced convection, J. Heat Transfer, Trans. ASME 82, 214-220.
This Page Intentionally Left Blank
Chapter 2
MEASUREMENT OF THERMAL BALANCE OF MAN Y.
"I1
Hokkaido Institute of Technology, Teine. Sapporo, Hokkaido 061-24, Japan.
CONTENTS The body heat balance Independent variables in the human thermal environment Ambient temperature Ambient vapour pressure Air movement . Mean radiant temperature or effective radiant field Clothing insulation Dependent physiological variables in the body heat balance equation Mean skin temperature Skin wettedness Metabolic energy production Sensible heat exchange by radiation and convection Operative temperature Clothing Radiation exchange Mean radiant temperature and effective radiant field Measurement of radiant exchange Convective heat exchange Evapora! ive heat exchange Conclusions THE BODY HEAT BALANCE
The thermal environment of man begins at the skin surface and extends outward to the surrounding media, which consist of the air we breathe; the clothing we wear; man-made sources of heat and cold necessary for our health and comfort; heat, cold and humidity caused by weather, and exposure to solar radiation. All these factors are characterized by temperature, or they in some way affect the heat transfer from the skin surface by radiation, convection, conduction, or evaporation.
Y. NSHI
30
The thermal exchange of man with his environment may be described by a heat balance equation relating independent environmental variables, dependent physiological factors and some properties of the boundary between the human body and the environment (GAGGE et al., 1971; GAGGE and NISHI, 1977; NISHI and GAGGE, 1971). The heat balance equation describing the thermal exchange between the body and its environment takes the classic form
M+ W*+R*+C*+E*+J*
= 0,
where M is the rate of metabolic energy production, W*the rate of work, J* is the rate of storage of body heat, E* the rate of evaporative heat transfer, R* the rate of radiant heat exchange and C* the rate of convective heat transfer. In equation (2.1) all terms have the units of work, J s-' or watts. For the present analysis the outer skin surface will be considered as the boundary separating the human body and its thermal environment. Alternatively all the terms in equation (2.1) can be expressed in watts per square metre (W m-z) as in equation (2.3) below. The outer skin area may be evaluated by the classic DuBois formula A - 0.202m0.425h0.725 (2.2) D -
9
in which the total skin surface area of the human body (A,) is in square metres, body mass (m) in kilograms, and height (h) in metres. At the skin surface, equation (2.1) takes the form given as equation (1.1) in chapter 1 of these proceedings. M,,+R+C+E+J=
0,
(2.3)
where M , is the net rate of metabolic heat loss from the skin surface, and conductive heat loss is neglected. In applications of equation (2.3) we will assume that the evaporative heat exchange ( E ) always occurs at the skin surface, and that the sensible exchange (i.e., R + C ) from the clothed body surface is the same as the dry heat flow from the skin surface to the clothing surface. The usefulness of the heat balance equation in any application of partitional calorimetry lies in the ability to calculate accurately (withinf5 % of M ) any four of its five terms. The fifth term, containing an unknown factor to be measured, may be found by difference. INDEPENDENT VARIABLES IN THE HUMAN THERMAL ENVIRONMENT
There are seven variables that must be measured to describe any thermal environment experienced by a human subject. AMBIENT TEMPERATURE
The ambient temperature (To,"C) of a gaseous environment surrounding the body, usually a mixture of air and water vapour, is defined as that measured at a point outside the thermal boundary-layer.
Measurement of tliermal balance of tnan
31
AMBIENT VAPOUR PRESSURE
The dew-point temperature (Td)is a fundamental measure of humidity in a moist environment. The ambient vapour pressure (e,) is an alternate fundamental measure of humidity. To a close approximation the water vapour pressure is equal to the at temperature Td,the “dew point”. There are many saturation vapour pressure eScTd, meteorological tables and psychometric charts available in engineering handbooks which relate saturated water vapour pressure (e,) to temperature (e.g. ASHRAE, 1977). A useful empirical relationship between e, and T is e,
= exp [16.6536-4030.183/(T+235)],
(2.4)
where e, is in kiloPascals (kPa). Other measures of humidity, but dependent on the ambient air temperature are relative humidity (RH),which is the dimensionless ratio e,/e, and wet bulb temperature (T,,,). If any two of the five variables, T,, e,, T,, R H , and Td are known, the other three may be found by use of a psychrometric chart or by use of the equation (2.4.1)
and the psychro,meter equation. For sea level e = escrw,-O.O66(T,-T,),
(2.4.2)
where e is in kPa. For e in millibars the psychrometric constant in (2.4.2) becomes 0.66 mb K-’, and for Torr 0.5 Torr K-
’.
AIR MOVEMENT
The movement ( v ) of the ambient air results from a) free buoyant motion caused by warm body in cool air medium; b) forced ventilation of the environment itself; and c) bodily motion caused by activity. Air motion is difficult to measure consistently; anemometers usually measure air movement caused by forced ventilation. Air motion over the body surface is a fundamental consideration, necessary for a complete understanding of both convective and evaporative heat exchange. MEAN RADIANT TEMPERATURE OR EFE‘ECTIVE RADIANT FIELD
The basic environmental variables that govern the exchange of heat by radiation are the mean radiant temperature ( T J , or in the energy mode, effective radiant field, RE which is defined as the radiant energy, in W m-’, exchanged by a human being with the imaginary black enclosure at temperature T,. CLOTHING INSULATION
The practical unit of clothing insulation is d o , which represents the effective insulation provided by a normal business suit worn by a sedentary worker in a comfortable indoor surrounding. The value of one unit d o is arbitrarily set at 0.155 m2 K W-I.
Y.NISHI
32
DEPENDENT PHYSIOLOGICAL VARIABLES IN THE BODY HEAT BALANCE EQUATION MEAN SKIN TEMPERATURE
Skin temperature (T,) may be measured by the use of appropriate sensors. Mean skin temperature, T,, may be defined as the average of a t least eight local values of T,, each weighted by the fraction of the total body surface represented. A useful weighting scale is head (7 oh), chest (17.5 yo),back (17.5 yo), upper arms (7 %), forearms (7 ?"), hands (5 yo),thighs (19 ?&), and legs (20 "/). Accurate measurements of skin temperature make it possible to determine the temperature and vapour pressure gradient that affect both the sensible and insensible heat exchange from the body surface. Skin temperature serves as significant index of the mode of regulation of body temperature. It may also serve as an index of our sensory judgments of heat and cold, including thermal comfort. SKIN WETTEDNESS
Skin wettedness (0)is defined as the ratio of the equivalent skin area covered with water (A,,,) to the total skin surface area ( A D ) ;or in practical terms as the ratio of the actual rate of evaporative heat loss to the maximum possible in the environment. BODY HEAT STORAGE
The rate of storage of body heat ( J , W m-2) is directly related to the rate of change in mean body temperature d TJdr (in K s-') by
in which cb is the specific heat of the body and m is the body mass in kg. Measured values of the specific heat of the body tissues are approximately3.5kJ kg-' K-'. In a state of physiological thermal neutrality, which occurs during rest and when there is no regulation of body temperature by sweating, the preferred mean skin temperature range is 33-34 "Cand rectal temperature is 36.9-37.1 "C.The corresponding range of mean body temperature would be 36.3-36.5 "C. METABOLIC ENERGY P R O D U U I O N
Metabolic energy production, M , in the basic heat balance (equation (2.1)) may be measured by the rate of oxygen consumption using the following equation for the relation between oxygen consumption and heat production, expressed in W m-2
M
= 352(0.23 R+0.77)(
Vo2/AD),
(2.6)
in which R is the respiratory quotient, which varies from 0.83 during rest to 1.0 during moderately heavy exercise. Voz is the rate of oxygen consumption in litres per minute at standard temperature and pressure.
Measurement of' thermal balance of m a n
33
The metabolic energy M may be expended in four ways: as metabolic heat which passes through the skin surface, M,,; as heat of vapourization of respired water vapour, E R ; as heat convected by respiration, C,; and as external work W. Thus the net rhetabolic heat in W m-2, passing through the skin surface is
M, = (M+ER+cR+ V I A , .
(2.7)
Because the ventilation of the respiratory tract is closely linked to the oxygen requirements of the body, both respiratory heat losses are almost proportional to metabolic rate. FANGER (1970) suggested that to a good aproximation ER and C, (for air at sea level) could be described by equations (2.8) and (2.9), respectively.
ER = 1 7 . 3 10-3M(5.87-e) ~
(2.8)
CR = 1 . 4 10-3M(34-TJ, ~
(2.9)
and
where both are expressedin W m-' ,e is in kPa and Tain degrees Celsius. In a typical environment (25 "C, e = 1.6 kPa, 50 % RH)ER accounts for about 1 % of M and is often ignored. Work (W) can be measured accurately on a bicycle ergometer or treadmill. The ratio, W / M , represents the mechanical efficiency (q) of the body doing work. In human beings the maximum mechanical efficiency, measured while pedalling on a bicycle ergometer, is approximately 18-22 % for an average person. Treadmill exercise is about 8-10 % efficient. For level walking and during most stationary activities, the mechanical efficiency is-zero and external work may be ignored. SENSIBLE HEAT EXCHANGE BY RADIATION AND CONVECTION
The exchange of sensible heat from the skin surface is usually accomplished first by conduction through clothing, and next by radiation and convection from the outer clothing or skin surface to the surrounding environment. OPERATIVE TEMPERATURE
The operative temperature (To) of human thermal environment is defined by GAGGE in chapter 5 as the temperature of an isothermal "black" enclosure in which man would exchange the same heat by radiation and convection from his body surface as he would in the actual non-uniform environment. By this definition, the dry heat exchange (HD= RSC) from the body surface, at temperature T,, is given by (2.10)
where hc, is the combined coefficient for heat transfer by radiation and convection in W m-2 K-', or by using ambient air temperature T, and mean radiant temperature, T, HD = h,(T,-Ta)+hr(T,,-Tr), (2.11) 3
- Bioengineering
34
Y . NISHI
in which h, is the convective heat transfer coefficient in W m-' K-' and h, the linear radiation exchange coefficient, also in W m-' K-'. Comparing equations (2.10) and (2.1 l), it follows that operative temperature is defined by
.
To = (hrTr+hcTa)/(hr+hc)
(2.12)
and that
hCr= h,+hc.
(2.13)
Thus by equation (2.12) operative temperature can also be defined as an average of T, and Tr, weighted by the respective heat transfer coefficients. CLOTHING
Sensible heat transfer through the clothing layer may be written as HD
= hd(~'-Tsu)Y
(2.14)
where h,, is the effective clothing conductance, in W m-' K-', and T, is now regarded as the mean temperature of the outer clothing surface. The reciprocal of h, is I , the effective insulation of the clothing worn, in m2 K W-'. The dry heat exchange from the clothing surface at temperature T,, is given by (2.15)
Hence, combining equations (2.14) and (2.1 5) to eliminate
TSu (2.16)
where F, is the dimensionless ratio (2.17) A thermal efficiency factor F,was first proposed by BURTONand EDHOLM(1 969). Thus the effective combined heat transfer coefficient from the skin surface is the product of Fc and her. By eliminating h, in equations (2.15) and (2.16) it may be seen that -
F, = (Eu-To)/(Ts-To).
(2.18)
et al., 1975). Thus, F, may be found by direct measurements of To,T,, and T,(NISHI RADIATION EXCHANGE MEAN RADIANT TEMPERATURE A N D EFFECTIVE RADIANT FIELD
The concept of effective radiant field RE was introduced by GAGGE et a]. (1967) as an aid to better understanding of high-temperature radiant sources. Mean radiant temperature (T,) is included in the definition for the effective radiant field as follows R E = h,(T,-T,),
(2.19)
Measurement of therriial balnnce of man
RE = hcr('o-'a),
35
(2.20)
where RE is in W m-'. The term (RSC)in the basic heat balance equation (2.3), may now be rewritten as (2.21) -
R
= F,[hr(Ts-Ta)-RE].
(2.22)
In establishing the general principles of sensible heat exchange above, the linear radiation exchange coefficient h, has been treated as a constant. For an unclothed = 34 "C at To = 29 "C, the ;,value of h, is subject at thermal neutrality, with 4.5 W m-' K-' for a lightly clothed subject (0.6 clo) with ? , = 31 "Cand To= 24 "C during thermal comfort, h, = 4.7 W m-' K-' ; for a well clothed subject (1.0 clo) = 27 "C and To= 20 "C, h, = 4.8 W rn-'K-'. These coefficients are with expressed per unit skin area, and are therefore lower then the radiative exchange coefficient obtained from the Stefan-Boltzmann relation.
Fs
TSu
MEASUREMENT OF RADIATION EXCHANGE
A simple direct measure of oured Bedford globe
human beings is obtained by using a skin-col15)m, diameter) and by using the formula (2.23)
where v is the ambient air movement in m s-' and Tg is the globe temperature. The first term in the bracket is the value of h, for a sphere at temperature 27 "C. The second term is Bedford's formula for the convective heat transfer coefficient for a 6 inch globe. The factor 0.76 converts the radiation field that the globe "sees" to that appropriate for a sitting human being with 0.6 clo insulation, as, because of its shape much of the body "sees" other parts of itself and in consequence, its radiative area is considerably less than A,. CONVECTIVE HEAT EXCHANGE
Measurements of h, in the laboratory for various standard activities are presented in tables 2.1 and 2.2. The theoretical value proposed by RAPP(1973) in table 2.1 is baeed on a sphere with 75 cm diameter; those by MISSENARD (1971) in table 2.2 are based on a long 17 cm diameter cylinder. Values by the naphthalene method, reported by NISHIand GACCE(1970), represent the first direct measurements of h, during exercise, treadmill walking, and free walking without the use of calorimetry. The values presented have an experimental basis up to speeds of 1.8 m s-'. For greater air velocities they are still useful as first-order estimates. Finally, values of h, vary widely over the surface of the body. Table 2.3 illustrates local h, values observed and GAGGE (1970) during rest and various types of exercises. by NISHI
Y. NISHI
36
T a b l e 2.1 Convective heat transfer coefficient (hJ in normally ventilated environment Condition Seated Seated Standing Seated on bicycle Pedalling bicycle Pedalling bicycle at 60 rpm
h,
Investigator
4.1 2.9&0.9 4.5f0.3 3.4 4.8&0.8 6.0
Remarks
RAPP(1973) WINSLOWet al. (1936) WINSLOWet al. (1936) WINSLOW et al. (1936) WINSLOWet al. (1936) NISHI and GAGGE (1970)
Theoretical Partitional Partitional Partitional Partitional
Calorimetry Calorimetry Calorimetry Calorimetry
Naphthalene method
T a b l e 2.2
Formulae relating the convective heat transfer coefficient (h,) to air velocity Formula
Activity
h, = 11.6 h, = 3.42+5.93~ h, = 0.53 h, = 8.6vfW
Investigator
Seated Seated Treadmill Free walking
Remarks
Y, room air movement (1971) Y, room air movement Nisai and GAGGE(1970) vfw, speed of t~-eadmill NISHI and GAGGE,(1970) Y/W, speed of walking
WINSLOWet al. (1936)
MISSENARD
T a b l e 2.3 Local convective heat transfer coefficient (Ac), in W m-’ K-I, during rest and exercises in normal air movement (0.154.2m s-’)
Body Region Resting
Free walking Bicycle
Sitting
10.9 m s-’ (1.8 ms-’ 60 rpm
Head Chest Back UpwrForearms arms Hands Thighs
3.2 4.2 5.4 7.2 9.5 4.4
2.5 3.6 4.5 4.8 6.7 3.3
4.0 3.2 6.4 4.3 8.3 4.7 6.0 6.7 ‘17.0 3.2 5.3
2.4
3.9 6.6 10.8 11.2 16.3 5.2
4.6 7.2 15.4 11.6 17.2 4.7
2.8 5.0 7.7 8.7 12.5 6.7
Legs
Mean (hc)
3.7 3.1 10.5 5.8 14.4 8.4 11.8 8.4 17.0 12.0 11.1 6.0
The relation of room air movement to the coefficient h, has limited significance. When a subject is active, such as when pedalling a bicycle ergometer, the measured value of v for the ambient air has even less significance. In evaluating the effect of both air movement and activity, it is more practical to consider the resulting value of h, as an index of “relative air movement” rather than use the actual value of the air movement itself.
Measuremelit of thermal balance 01. man
37
EVAPORATIVE HEAT EXCHANGE
The heat loss by the evaporation of sweat is man’s most effective means of survival in the heat. The evaporative heat loss ( E ) itself is the best single physiological index of his environmental stress. Since the beginning of human calorimetry, observed changes in E have been the quantitative basis for measurements of the combined transfer coefficient and the individual transfer coefficients for radiation and convection for the particular experimental arrangement used (WINSLOW et al., 1936; COLIN and HOUDAS,1967). In partitional calorimetryE may be found from the rate of change in body mass A, as measured by a sensitive balance. The most successful balance used for this purpose has been the Potter Bed, described in US Patent 3,224,518 and 3,360,002. This balance is designed without any wearing knife edges and has proven useful for continuous measurement of body weight during both rest and heavy exercise (SALTIN et al., 1970). In such studies E is determined by the following relation
E = 60m A/AD,
(2.24)
where h is the rate change of body mass in grams per minute and 1 is the latent heat of vapourization of water (A = 2450 J g-I). To evaluate E,, the total evaporative heat loss from the body (E)must now be corrected for respired vapour (ER)using equation (2.8). The validity of equation (2.24), as a direct measure of evaporative heat loss, depends on the ability of the sweat produced at a rate to evaporate completely on the skin surface. Without active sweating the average sized person would lose weight by respiratory and skin diffusion at a rate of approximately 0.5 g min-’. When working at 50 % of maximum oxygen capacity in 30 “C To and 50% RH, the value of h would be about 10-12 g min-’. The maximum evaporative heat loss (En,)from a totally wet skin surface is proportional to the water vapour pressure difference from the skin surface to the ambient air. The maximum evaporative capacity per unit area of wet skin may be written as
Em = 16.5hcF,,(es-e),
(2.25)
where e, is the saturation vapour pressure at skin temperature, in kPa. The factor of 16.5 is the reciprocal of the psychrometer constant at sea level, which may be derived from the Lewis relation. The term FeCin equation (2.25), known as Permeation Efficiency Factor (NISHI and GAGGE,1970b), has been added to take account of clothing worn. The factor F,, is analogous to Fc for dry heat exchange. For normal porous clothing, such as worn every day by an average person, it has been shown experimentally (NISHIand GAGGE, 1970b) that :
Fee = (1 +0.92hCI)-’, where I is the clothing insulation for sensible heat transfer.
(2.26)
Y. NISHI
3b
Skin wettedness is an expression of the efficiency of evaporative regulaton. The wetted area of the skin (A,,,)may be defined as that area of the skin which if covered with sweat, would provide the observed rate of skin evaporation under the prevailing condition. Thus by this definition
E
= AwEm/AD = WE,,,,
(2.27)
where w is a dimensionless number between 0 and 1. w -. A J A , = EIE,,,.
(2.28)
When E can be evaluated experimentally, since E,,, is obtained from the basic observations of Ts, To,T,, h, and I , it is possible to evaluate skin wettedness a t any time by equation (2.28). Skin wettedness ( 0 )ranges from a certain minimum value, which occurs when there is no evaporative heat loss by regulatory sweating, to a maximum theoretical value of unity. At the minimum the evaporative heat loss from the skin surface (Ed)is entirely due to the diffusion of water vapour through the outer layers of the skin. When regulatory sweating begins, evaporative heat loss may occur by diffusion as well as by the evaporation of sweat (E,). When the skin surface is completely wet (Le. o = l), E is attributed entirely to regulatory sweating (E,). The ratio EJE,,, describes the skin wettedness due to sweating (ors), the ratio E,/E, is the skin wettedness due to diffusion (cod), and the total wettedness w at any time is given by GAMEet a!. (1971) as 0
= mdf(l-Od)w,.
(2.29)
From data reported by BREBNERet al. (1955), the minimum value of w may be 0.06. The significance of the maximum possible evaporative capacity (EJ for clothed and unclothed humans was first emphasized by BELDINGand HATCH(1956) when they proposed their Heat Stress Index.
CONCLUSIONS
Heat balance equations, describing human heat exchange by radiation, convection, evaporation and conduction from the skin surface with the thermal environment have been outlined in this chapter. The environmental variables in the basic heat balance equation that must be measured are the ambient air temperature, the mean radiant temperature or effective radiant field, the ambient water vapour pressure or humidity, air movement as it affects the convective and evaporativc heat loss, and clothing insulation. The physiological variables in the heat balance equation are skin temperature, skin wettedness, mean body temperature, metabolic energy consumption and the ra?e of external work. A more complete description of the heat balance equation, with more details applicable to special and aquatic environments is given in a previous review by GAGGE and the present author (GACGE and NISHI, 1977).
Measwetirent of thermal bolnnce of
mala
39
REFERENCES American Society of Heating, Refrigerating and Air-conditioning Engineers (ASHRAE) (1977), Handbook of Fundamentals, ed.: Carl W . MACPHEE.,New York. BELDING H. S. and HATCH T. F. (1956), Index for evaluating heatstress in terms of resultingphysiological strain, ASHRAE Trans. 62, 213-236. BREBNER F. F. KERSLAKE D. McK, and WADDEL J. L. (1956). The diffusion of’ water vapour through human skin, J. Physiol. (London) 132, 225-231. BURTON A. C. and EDHOLM 0. G. (1969), Man in a Cold Environment (Facsimile of 1955 edition), Hafner Publishing Co., New York. COLIN J. and HOUDASY. (1967), Experimental determination of coefficient of heat exchange by convection ofthe human body, J . Appl. Physiol. 22, 31-38. FANGER P. O., Thermal Comfort (2nd ed.; facsimile of 1970 edition), McGraw-Hill, New York. GAGGE A. P., RAPPG. M., and HARDY J. D. (1967), The effective radiant field and operative temperature necessary for comfort with radiant heating, ASHRAE Trans.73, 1-9. GAGGE A. P., STOLWIJK J. A. J., and NISHI Y.(1971), An effective temperature s a l e based on a simple model of human physiological regulatory response, ASHRAE Trans. 77, 247-262. GAGGE A. P. and NISHI Y. (1977), Heat exchange between human skin surface and thermalenvironnaent [In:] Handbook of Physiology. Reactions to Environmental Agents, ed.: D. H. K. LEE,Bethesda, Md. Am. Physiol. SOC.,sect. 9, chapt. 5, 69-72. MISSENARDA. (1971), Exchange thermiques du corps humain avec /‘ambiance, Rev. GBn. Therm. 117, 765-770. NISHI Y. and GAGGE A. P. (1970a), Direct evaluation of convective heat tramfer coefficient by naphthalene sublimation, J . Appl. Physiol. 29, 830-838. NISHIY. and GAGGEA. P. (1970b), Moisturepermeation of clothing - a factor governing ihermul equilibrium and comfort, ASHRAE Trans. 76, 137-145. NISHI Y. and GAGGEA. P. (1971), Humid operative temperature. A biophysical index of thermal sensation and discomfort. J . Physiol. (Paris) 63, 365-368. NISHI Y. GONZALEZ R. R., and GAGCEA. P. (1975), Direct measurement of clothing heat transfer properties during sensible and insensible heat exchange with thermal environment, AS H U E Trans. 81, 183-199. RAPPG. M. (1973). Convecfiveherit transferjor nude man, cylinders and spheres at low air velocities, ASHRAE Trans. 79, 75-87. SALTINB., GAGGEA. P., and STOLWIJK J. A. J. (1970). Body temperatures andsweatingduring thermal transients caused by exercise, J . Appl. Physiol. 28, 318-327. WINSLOWC.-E. A., HERRINOKIN L. P., and GAGGEA. P. (1936), Tl7e determination of radiption and convection exchanges by partitional calorimetry, Am. J . Physiol. 116,669-684.
This Page Intentionally Left Blank
Chapter 3
EVALUATING THE EFFECTS OF CLOTHING ON THE WEARER R. F. GOLDMAN U.S.A. Research Institute of Environmental Medicine, Natick, Massachusetts 01760, U.S.A.
CONTENTS Introduction The five levels of analysis Units The permeability index i, Determination of I and im The copper man Physiological chamber trials Conclusions from specific studies Problems in the cold Problems in the heat Impermeable materials Effects of different materials or treatment Effect of drape or venting Summary
INTRODUCTION
Clothing has been designed to a much greater extent by fashion and by iechnological developments in indus’ry than by any scientific analysis of the heat exchange allowed by clothing between the wearer and his environment. However, requirements to maximize survival time, extend performance time and improve the general comfort of soldiers exposed to extremes of Arctic, de:ert or tropic environments have required the development of a multi-disciplinary, multi-level program at this laboratory. These scientific analyses, essential to deal with environmental extremes, can also be applied to suggest clothing design; for less extreme environments and to evaluate the relative contributions of various factors to thermal aspects of clothing comfort.
42
R. F. GOLDMAN THE FIVE LEVELS OF ANALYSIS
A multidisciplinary approach has been evolved at the USA Research Institute of Environmental Medicine in Natick, Massachusetts to assess the thermal interactions between the environment, the uniform worn, the man and his job. Studies are conducted at five different levels of analysis, with each level providing information which can be related to the others, as follows: I ) the physical heat transfer characteristics of the uniform materials are measured using classical heated flat plate theory, and also a unique “sweating” flat plate; 2) complete clothing a: semblies, with and without such additional items as gloves, headgear or back packs etc., are evaluated on a “sweating” copper manikin fo,. the heat transfer characteristics of the clothing assembly; the values obtained are used in calculations in a programmed computer model to predict the wearer’s tolerance limits ; 3) carefully controlled physiological trials are carried out in climatic chambers with volunteer subjects dressed in these clothing systems, to validate or refine the computer predicted limits; 4) controlled small scale studies are conducted in the field or at the work site. Groups of men wear specified clothing systems and carry out specified tasks under conditions of environment, terrain and work rate where physiological problems are anticipated, based upon experience in climatic chamber trials; 5) studies with these clothing systems are carried out, collaboratively, during actual field opera.ions scheduled by Army elements or other groups. Specific details of the methodology for the laboratory studies (i.e. physical plate material studies, biophysical copper man evaluations and tolerance predictions and physiological chamber validation experiments) are presented below. The methods used in field studies are adapted for each problem, are therefore difficult to generalize and will be included when discussing the results. UNITS
Some years ago, physiologists working in the field of clothing and the associated heat transfer from a man developed a technique to determine how much heat would pass through a garment by thermal radiation and convection from the skin (GAGGE et al., 1941). The difference between a man’s skin temperature and the ambient tcrnperature was taken as a gradient across which, to avoid a change in body temperature, he had to eliminate the difference between his metabolic heat production and the heat he could lose by evaporation of sweat from his skin or of water from his lungs. The non-evaporative component was assumed to pass through the clothing by radiation and convection heat transfer. They then defined the insulation I of a clothing system, plus the overlying still air layer. In terms of SI units their “clo” unit is equal to 0.155 m2K W-’. The dry heat transfer (HD)to the surroundings (i.e. convective plus radiative), in units of W m-’, is given by
HD = AT/?
(3.1 a)
Effects of’clorhing
43
or, when insulation is expressed in clo (Z‘)
HD = AT/O.l55Z’,
(3.1 b)
where AT is the temperature difference between the skin and ambient air, in Kelvin. This equation states that heat flow equals the driving force, in this case a temperature difference, divided by a resistance. This basic approach for radiation-convection heat loss yields a quantitative assessment of how good a given uniform is for a resting man in cold weather, since radiation and convection are major avenues of heat loss in the cold. However, for a working man in the cold evaporation of sweat becomes an important avenue of heat loss. Furthermore, radiation and convection heat loss decrease with increasing ambient temperature while evaporative cooling rises. Thus, the insulation value alone is insufficient in the heat. THE PERMEABILITY INDEX, f,
A similar form of equation can be used to predict evaporative heat transfer HE HE = 16.5(eS-eJ1, to allow the use of the same units for I in equation (3.2) as in in equation (3.la) the constant must have the dimensions of K kPa-’, i.e. the inverse of those of the “psychrometer constant”. The derivation of equation (3.2) is given in the introductory chapter. The gradient for evaporative transfer is the difference between the vapour pressure at the skin surface (e,) and the ambient vapour pressure (eJ in kiloPascals (kPa). Using the slope of the wet bulb lines on a psychrometric chart a vapour pressure difference can be converted to an equivalent temperature gradient. One can then determine the evaporative heat loss from a square metre of surface with a given water vapour pressure; e.g. at 35°C (the skin temperature of a sweating man) there would be a vapour pressure of 5.6 kPa at the skin, and the gradient will thus be 5.6 kPa minus of this laboratory the ambient air vapour pressure e,. The late Dr Alan WOODCOCK proposed that the evaporativeheat transfer for a nude man, or for any clothing system, could be expressed as the ratio of the actual evaporative heat loss, as hindered by the clothing, to that of a wet bulb with equivalent insulation (WOODCOCK, 1962). He suggested expanding equation (3.2) to include a dimensionless permeability index (i,) so that: HE = 16.5i,n(es-ea)/I. (3.3a) The index ,i could range from 0, for a system with no evaporative transfer, to 1 for a system which had no more impedance to evaporative heat transfer than the usual wet bulb thermometer. The conventional wet bulb, of course, is ventilated, and the still air layer is greatly reduced. Since a soldier is surrounded by a relatively undisturbed air layer, i, seldom approaches 1.0 for a man, but is limited in still air to about 0.5. When 1 is expressed in clo units (1’)and e in kPa equation (3.3a) may be approximated to HE= 100i,,,(es-ea)/Z‘. (3.3b)
44
R. F. GOLDMAN DETERMINATION OF I AND i,,,
Figure 3.1 a shows the flat plate apparatus used in measurement of the insulation value. The apparatus consists of a test section (A), surrounded on all four sides by a guard section (B) with another guard section ( C ) beneath the entire upper plate. All three sections are instrumented with plate temperature sensors, heating elements and thermostats. The sample to be tested (0) is placed on the surface and the entire assembly is placed in a constant temperature cabinet. In operation, power to the guard sections is controlled so that their surface temperature is identical to that of the test section. Thus there is no gradient for heat loss from the bottom or edges of the test section. After equilibrium is established the power required by the test section equals the heat lost through the insulation, and can be expressed as a flux density per degree of temperature gradient from the plate surface to ambient air. This can be converted to the corresponding insulation for the sample plus the adhering air layer, using equation (3.1). If a thin cotton “skin” is placed on the plate surface and a water level is maintained at the surface of some small holes drilled in it then the “skin” wicks out enough water to maintain a constant, saturated, vapour pressure under the fabric sample. A constant ambient vapour pressure is maintained in the measuring chamber and power requirements measured, just as for the dry plate. The value of i,,, can be determined for a given sample plus its adhering air layer by means of equation (3.3). Figure 3.la shows the “sweating” flat plate and its water supply in the constant temperature and humidity chamber with its control and recorder panel. The flat plate determinations of I and i, are primarily of use in selection of the fabrics to be used in a clothing system. The effects on heat transfer of different weaves, perforations, different finishes or treatments, the effects and best arrangement of multiple layers, etc. can all be established using the sweating flat plate (FONSECA, 1967). Heated, dry and “sweating” cylinders have been developed to mimic the cylindrical shape of the body. These are useful for studying wind penetration through clothing, and effects of spacer materials, but factors of drape, fit, and shape are difficult to simulate even on a cylinder. Also, a complete uniform is made up of a number of different components, protecting various parts of the body, so that evaluation of a complete clothing system requires a more sophisticated model 1965). than a cylinder (FONSECA and BRECKENRIDGE, THE COPPER MAN
The solution has been the development of life sized, heated copper manikins. Figure 3.1 b shows a manikin with his “sweating” cotton skin. The heat provided to the manikin to maintain a constant skin temperature can be measured and the ambient temperature and vapour pressure of the test chamber can be controlled; skin and air temperature and vapour pressure are measured. Thus, the sensible and evaporative heat losses caused by a given gradient of temperature and vapour pressure can be calculated for any clothing system worn. This technique has been in use for the last 15 years. Using the insulation and evaporative transfer indices, with some physiological knowledge, tolerance times can be predicted, for a given
Effects of clothing
45
C
Fig. 3.1. Laboratory Test Methodology a) The "sweating" flat plate apparatus for assessing material characteristics; the diagram inset shows the test section A, upper B and lower C guard sections and the position of the materlal to be evaluated D b) The sweating copper manihin used to assess non-evaporative (I) and evaporative (im) heat transfer characteristics of a complete uniform assembly c) Volunteer subjects, each with a different clothing assembly, seated in the climatic chamber during a rest break. The cables leading overhead from each subject carry rectal and skin tcmpcratuce information
tack, for men in the chambers and in the field. The heat stored by the body must be the difference between the heat produced at work and that lost by evaporation and by radiation and/or convection through the clothing system. Since the average man has 1.8 m2 of surface area, by estimating his skin temperature @"), total dry
R. F. GOLDMAN
46
heat transfer (El,*), in watts, can be calculated for any given ambient dry bulb temperature To as
H i = 1.8HD = 1.8(Fs-Ta)/1
(3.4a)
(3.4b) Similarly one can calculate the maximum evaporative heat transfer from a body, through clothing, for any given ambient vapour pressure
HZ
=
1.8HE= 29.7(5.9-eJ1,
(3.5)
where a 36 "C skin temperature has been assumed for the clothed sweating man, giving e, = 5.9 kPa. If heat production, respiratory heat loss and solar heat load gain are known, one can calculate whether the man can eliminate all the heat he is producing or whether some of it will be stored in his body. The specific heat of human tissue is 3.47 kJ kg-' K-', therefore the body temperature of a 70 kg man will be increased by 1 "Cfor each 240 kJ stored. This allows prediction of tolerance as the time to reach a given body temperature. A computer program has been devised (GIVONIand GOLDMAN,1971, 1972, 1973a, 1973b; PANDOLF et al., 1976) which incorporates many of the significant physiological, physical and environmental factors involved in human heat transfer. If the appropriate values for clothing, environment and metabolic heat production are supplied, the .model will predict the body temperature (rectal and skin) and heart rate response of a n individual under the chosen conditions. However, the predicted responses are frequently checked by actual environmental chamber exposures of men. PHYSIOLOGICAL CHAMBER TRIALS
Standard protocols have been developed for these trials; for example, one requires two fifty-minute walks, separated by a ten-minute break, followed by a one hour rest (see table 3.1). Figure 3.1 c shows subjects in an environmental chamber, during a rest period. The men are seated on benches placed on one of the large four man treadmills. Each subject is wearing a different garment since, as usual, choice of the garments and subjects was randomized on each day. Each subject is wearing a rectal catheter, to measure deep body temperature, and a three point skin temperature harness. Two connecting cables from each man are led outside the chamber to the instrumentation shown in figure 3.1 c. Each subject's rectal temperature is indicated continuously on one of the eight meters at the base of the master timer. Skin temperatures are recorded sequentially on the recorder, along with the rectal and chamber temperatures. Each point printed is simultaneously encoded and punched on the digital punch tape system shown at the right. This tape is used as a permanent record and is also fed into a programmable calculator system (HP 9825) for prompt data reduction and analysis. The results are plotted, frequently superimposed on a preliminary plot o f the predicted responses, so that the agreement can be checked.
T a b l e 3.1 Standard protocols for clothing studies
Cumulative time 0
115 120 170
Schedule: Protocol I Approx. 0800 Report to tropic chamber dressing room. Weigh canteens and urine containers. Subjects strip nude, void, defecate. Weigh nude. Put on catheter, skin harness, uniform, pack. Weigh clothed, with helmet liner and pack. Obtain initial pulse rate, rectal temp., skin temp. Approx. Procedure Cumu- Approx. hour lative hour 0900 Enter tropic wind tunnel maintained at “approtime0 0900 priate”c0ndition. Rest,sittingfor 115 minutes ,without pack. Continuous rectal and skin temp. recording throughout test period. Water ad lib. Pulse rates at 6@ and 115 minutes. 1055 Weigh clothed, with helmet liner and pack; weigh 50 0950 and refill canteens. 1100 Subjects walk for 50 minutes at 5.6 km hr-I 60 lo00 Pulse rates at 25 and 50 minutes. 1150 Rest, sitting, for 60 minutes. Pulse rates at 30 110 1050 and 60 minutes 120 1 loo 180
230
1250
1200
Leave tropic wind tunnel. Weigh clothed, with helmet liner and pack. Strip and dry thoroughly. Weigh nude. Weigh canteens. Tabulate all water intake and output and calculate 2-hour evaporative water loss at rest, total evaporative water loss and total sweat production.
Schedule: Protocol I1 __-
-
Procedure Enter tropic wind tunnel maintained at “appropriate ”condition. Walk for 50 minutes at 5.6 km hr-l. Continuous rectal and skin temp. recording throughout test period. Water ad lib. Pulse rates at 25 and 50 minutes. Rest, sitting for 10 minutes. Pulse rates at 10 minutes. Walk for 50 minutes at 5.6 km hr-l. Pulse rates at 25 and 50 minutes. Weigh clothed, with helmet liner and pack; weigh and refill canteens. Rest, sitting for 60 minutes without pack. Pulse rates at 30 and 60 minutes. Walk for 50 minutes with pack at 5.6 km hr-I. Pulse rates at 25 and 50 minutes.
48
R. F. GOLDMAN CONCLUSIONS FROM SPECIFIC STUDIS
Having established the scientific approaches on which our program is based, we will emphasize some specific studies directed to clothing for extremes of cold and heat. A clothing system to minimize heat loss for an inactive man under Arctic conditions must cause a severe heat stress problem as the man increases his heat production by activity, if he does not remove most of his Arctic protective clothing.
PROBLEMS IN THE COLD
Man’s thermal problems in extreme cold have two possible solutions: increasing heat supply or decreasing heat loss. Inefficient clothing design can increase heat production during activity; for example, “friction” in multiple layer clothing increases heat production by about 16%, and footwear weighing 1 kg increases the energy cost of locomotion as much as 5 kg carried on the torso. However, this approach is not really desirable since such solutions are useless at rest, and during physical work men wearing Arctic clothing usually produce too much heat. Techniques for increasing physiological heat production other than by shivering (e.g. increased dietary protein, non-shivering thermogenesis with cold acclimatization) are generally ineffective, so if extra heat is to be provided it must be by auxiliary heat. For inactive subjects dressed in full Arctic clothing, a i little as 3 watts to each hand and 5 watts to each‘foot has proven adequate for comfort at -57 “C with a 4.5 m s-’ wind for more than six hours (GOLDMAN, 1964). Decreases in heat loss can be accomplished only by reducing convective, radiative or evaporative heat exchange. Sensible heat loss is largely a function of the thickness of insulation. Using the relationship of 0.243 mz K W-’ (1.57 clo) per cm of thickness, much has been accomplished. However, the problems of clothing weight per se, clothing bulk and hindered mobility, and the dramatic increases in the surface area available for heat loss as one increases the radius of insulation around the body’s cylindrical sections- particularly such thin human cylindricalsections as the fingerslimit further improvement. Lightweight, non-wove nbatts of material are increasingly bcing substituted for conventional insulat’ng materials in parka liners, gloves and sleeping bags. lmproved outer wind-proof materials and closures are also desired to reduce the loss in insulation of clothing as wind and/or subject movement rate increases (HOLLIES and GOLDMAN, 1977). The specification of insulation for the combined heat loss by radiant and convective heat transfer, obscures consideration of radiant heat loss per se. Reflective insulation is in disrepute, since a highly publicized reflective lining system for outercoats was a failure in really reducing rdaiant heat loss from the body. However, new techniques deposit reflective materials on tough, light, thin plastic materials which dramatically increase the insulation of non-woven, lightweight batt materials, with little increase in weight or thickness. Unfortunately, most of the measured increase simply represents a return to the 0.24 mzK W-’per cm standard found with conventional materials; the free path length in the non-woven, batt, materia%
Effecls of’clothing
49
reduces the insulation per unit thickness because of increased radiant heat losses through the materials. The use of reflective materials simply reduces these losses. A “sandwich” of several layers of batt insulation between two thin windbreak surface layers, which has an insulating value of 0.54 m2 K W” (3.5 clo), could be improved to 0.84 m2 K W-’ (5.4 clo) merely by adding three reflective layers of about 1 mm thickness each, but 0.74 m2K W-’ (5 clo) would be expected for a conventional material of this same thickness without any reflective materials. At present it is difficult to maintain the critical spacing required to demonstrate these benefits of reflective layers in non-woven materials. During fabrication into clothing the insulating batt is compressed below the critical spacing dimension required; i.e. compression produces reduced thickness and returns free path length toward that of conventional materials. Work is currently underway to develop practical clothing items which achieve the improvements in insulation and reductions in weight that we feel are possible with multilayer batts incorporating such reflective materials. Most of these materials, so far, are impermeable to water vapour. Since for a man at rest about 25% of total heat loss is by evaporation (half from the respiratory tract, half from the skin) vapour bamer materials might be useful; moisture released by the body when absorbed into the clothing substitutes water - a good heat conductor - for trapped air, and thus reduces insulation. Vapour barrier boots have proven highly successful at slowing foot cooling rates; but problems arise when feet are not dried regularly and dry socks donned. The potential for use of reflective materials within such vapour barrier items seems quite promising. Newer reflective materials appear highly vapour permeable, but tend to be so heavy that the benefit of reduction in weight of using the non-woven batts disappears. A number of devices have been proposed over the years for reducing the respiratory heat loss; generally these do not seem to be worth their problems, at least in healthy young individuals. Similarly, face masks to reduce the exposed skin surface have been tested, but again are generally not worth the trouble; a good parka hood with a controllable apkrture adequately shields the face and minimizes exposed surface area. For the few individuals exposed to conditions sufficiently extreme to require such items, modern technology can provide portable life support systems which provide a comfortable microclimate within a completely encapsulating clothing system. In summary, auxiliary heating,’ substitution of lightweight batt insulation, incorporation of reflective insulation where practical and, ultimately, climate controlled clothing systems are the promising new approaches being developed for cold weather clothing. PROBLEMS IN THE HEAT
During heavy exercise in extreme heat, man becomes almost totally dependent on evaporation of sweat for the cooling required to dissipate his heat production (MJ, at rest or at work. M,, ranges from about 58 W m-? (I met) at rest to a peak production of perhaps 580 W m-’ (10 met) for short periods. A value of 175 W m-*(3 met) can be used for an average man’s heat production during light industrial work. We 4
- Bioengineering
50
R. F. GOLDMAN
have shown that our average subject, when allowed to unconsciously adjust his speed while “working hard”, tends to select an average heat production of 290 W m-2 (5 met) (HUGHES and GOLDMAN, 1970). The equivalent evaporative cooling, required when ambient air temperature approaches skin surface temperature, is 520 watts. This in turn requires evaporation of some 750 grams of sweat per hour if the evaporation occurs at the skin surface: the sweat rate required to produce the same cooling can double if evaporation occurs from the clothing surface rather than from fh e skin, as a result of the increase in the insulation between the point where evaporative cooling takes place and the skin surface. In contrast, the maximum sustainable sweat production is 1000 grams per hour for a well heat acclimatized man; the importance of removing heavy clothing during activity is obvious. The problem becomes still more obvious if one considers the difficulty of avoiding dehydration; by getting the litre of water required to replace this body water loss each hour. When the man stops exercising the sweat trapped in his clothing will eventually evaporate and provide undesirable additional cooling. The copper manikin I values reflect both material thickness and the air insulation trapped between the skin and clothing, a function of material stiffness, i.e. drape. Because the skin to air temperature gradient for non-evaporative heat loss is small in the heat, the important value is iJZ; this indicates the fraction of the maximum evaporative cooling possible in a given environment without wind, since these manikin values are “still air” determinationc. Thus, a man wearing just a fatigue uniform (inL N 0.5; I’ rn 1.35 clo) can obtain about 37 % of the maximum cooling possible (iJZ’ = 37 W m-’ kPa-’), while with a heavier clothing system (im- 0.5; I’ w 2.5 clo) only 20 yois available. In chamber studies of the physiology of unclothed 1965) body et al., 1965; IAMPIETROand GOLDMAN, men in the heat (GOLDMAN heat storage of about 330 kJ was enough to make a number of the volunteers stop working; a heat storage of about 660 kJ resulted in heat exhaustion in 50 yo of those who had continued, while almost no one could tolerate 920 kJ. Note the correspondence of these critical levels of heat storage with the heat debt levels of 330 kJ waking a sleeping subject, and the 630 kJ inducing marked shivering. Hard work in a 35 “C, 50 :h relative humidity (RH) environment is so severe a combination that the predicted tolerance time is less than l+ hours; even wearing just the usual fatigue uniform; in reality the fatigue uniform would rapidly become sweat soaked and this, in combination with any natural ambient air motion, would result in a considerably longer tolerance time than this predicted value. Arctic protective uniforms, however, cannot readily wet out with sweat and heat exhaustion can occur as easily in the Arctic, if a high work level is sustained without appropriate reduction in clothing. While the copper man based predictions are useful in ranking clothing systems, they obviously should be supplemented by physiological chamber trials to assess the effects of body motion, wind and the efficiency of sweating, a s shown by HOLLIES and GOLDMAN (1977). Men walking on treadmills under controlled conditions cannot indicate all the problems that will occur in troops attempting to perform their military missions in the field. While thermal effects resulting from the addition of packs, body armour, helmets, weapons, etc. have been measured both on copper
51
Efecis of clothing
manikins and in physiological cha.mber studies (GOLDMAN, 1969; JOY and GOLDMAN, 1968), the problems associated with military tasks other than marching, and with solar heat load, terrain and wind variation are impossible to reproduce indoors. IMPERMEABLE MATERIALS
In the study presented in fig. 3.2, a commercial vinyl raincoat was worn either full length or cut down by 1/4, 1/2 or 3/4 of the total length. It is notable that both the Z and the i,,, /Zvalues determined on the sweating copper man reflect this linearity. During the rest period there is no body heat storage, but by the end of the walk the mean heat storage values rank in the exact order predicted. However, the difference observed between the 1 /4 and 1 /2 length garments is smaller than one would anticipate from the physical values, and reflects a physiological compensation that occurred. The extra sweat a man will produce as his body temperature rises can be evaporated to provide additional cooling if sufficient unimpeded body surface area is available.
-
7
Rest
Rest
.z r
200
Y
y \
aJ
\
-
CJ' 150
\ \
P 0
\
tn
g
\
\
\
4 -.
100
1
x
I\
U
0 m
50
\
0
0 0
X
- 50 I 0
30
90
60
120
150
180
210
Time
(mtn)
Fig. 3.2. Mean body heat storage for eight subjects, at rest and during a 5.6 km hr-' treadmill walk. The sweating copper manikin measurements of I' suggest that the permeability index values are a linear function of the percentage of the surface covered with imMrmeable materials. While the subjects can eliminate all of their resting heat production regardless of raincoat system, the effects of increased impermeable area are clearly evident during the walk. Conditions: air temperature 29.4 "C, wet bulb temperature 22 "C 0 --full length, 1.7 clo (0.26 m 2 K W - I ) ; U-
r: 0
.-.-
3/4 length, 1.6 d o (0.25 m2 K W-'); 1/2 length, 1.5 clo (0.23 rnz K W - ' ) ;
..-..-
114 length, 1.4 clo (0.22 m 2 K W-')
R. F. GOLDMAN
52
The average sweat production did increase as a function of increased impermeable coverage, and in the case of the 1/2 length garment enough of this extra sweat was able to be evaporated to compensate for the increased body area covered by and impermeable layer. The still greater sweat production of men wearing the 3/4 and full !cngth garments could be not evaporated in sufficient quantities to compensate. EFFECT3 OF DIFFERENT MATERIALS OR TREATMENTS
In the study presented in fig. 3.3, fire resistant uniforms made up of either one or two layers of a new “Nomex” synthetic fibre were evaluated in comparsion with a standard cotton “fatigue” and a standard cotton tropical combat uniform. All sets of these four uniform types had been laundered but not starched and, as a substudy, an additional group (8 se’s) of the fatigue uniforms was laundered and then a0 7
Y
al cn
2 0
300
c
vr
c
Q
(Ir
200 1 ,
U 0
0 100
0
I
1
I
1
I
I I ,
I
I
Effects of clothing
53
heavily starched. Results from the copper manikin gave a value of im/Z = 34 W m-* kPa-’ and essentially similar values of insulation for all but the two-layer Nomex uniform. The physiological measurements obtained on subjects during the chamber experiment confirm these physical results. The average heat storage values for subjects wearing the four garments with the same i,/I values ranged between 260 and 290 kJ at the end of the fifty-minute walk, while the corresponding value for the two-layer Nomex garment, which had i,/Z = 31 W rn-’ kPa-’, was 365 kJ. Thus, variation in wicking and wetting characteristics, associated with differences in fibres or finish, which might have produced a difference between results obtained under static laboratory test conditions and those obtained on actual humans in chamber studies, did not; the relative results predicted from the static copper man tests were confirmed by the chamber evaluation.
-
Rest
Walk
200
7
Y
c o,
150
0,
P
2 Ln 4-
100
0
a
r
2 0
50
m
0
* I I
I
-50 30
0
60
90
I 120
I
180
1%
210
Time (min) Fig. 3.4. The physiological effects of shoulder yoke vents: these are more difficult to asscss from the values given in the figure legend, since they are obtained from a stationary, sweating copper manikin. Putting a vent into an impermeable coated fabric raincoat siguificantly increases the wearer’s heat loss while walking; a loose fitting poncho with hood, in theory the worst of the garments because. of the greater area covered, has a very similar physiological effect to a vented impermeable raincoat, because of the flapping and billowing associated with walking in a poncho pumping air. Adding a vent to an already air permeable raincoat (“standard”) produces little or no add itional heat loss. Conditions as for fig. 3.2. coated raincoat, 1.7 clo (0.26 m* K W-’);
0a-.-
A..-. .0 X
poncho, 1.8 clo (0.28 rn2KW-l); coated rainco % with vent, 1.7 clo (0.26
.- . . standard raincoat, 1.7 clo (0.26 I
---
m2
K W-’);
W-’); standard raincoat with vent, 1.7 clo (0.26 m 2 K W-’) 1112 K
54
R. F. GOLDMAN EFFECTS OF DRAPE OR VENTING
The results of another study of raincoats is presented in fig. 3.4; impermeable coated fabric raincoats, both with and without a vent inserted under a flap across the shoulder area, were compared with water repellent but vapour permeable raincoats cut to the same fashion and again both with and without the vent. The standard Army poncho, which is an impermeable loose cape with hood, was also included in the study. The copper man I values were essentially the same for all garments, while the permeabilities showed little or no difference associated with venting, i, was 0.26 for the permeable coats, 0.16 for the impermeable coated fabric coats and 0.11 for the poncho. However, the results of the physiological evaluation showed a significant advantage for the vent in the impermeable coated fabric raincoat, but no advantage for the vent in the permeable coat. In addition, the average heat storage of men wearing the poncho was nearly as low as that of men wearing the vented impermeable garment. Thus, the ability offered by an impermeable garment wi!h a vent, or by a billowing impermeable garment was of significant advantage to subjects walking in such garments; on the other hand, a vent in an already permeable garment seemed to offer little or no additional air exchange. Thus, with respect to factors of fit and drape or cut which provide ventilation features, the static physical test values require careful interpretation. Additional studies, involving flapping the garments during the copper manikin measurements, have been encouraging, but the technique requires additional refinement (HOLLIESand GOLDMAN, 1977). SUMMARY
In summary, the combined measurement of I and i, on heated sweating copper manikins is a valuable tool for predicting a rank order of thermal stress effects for clothing assemblies worn in the cold or heat. Care must be taken if air permeabilities differ widely or if clothing design allows unusual air exchange during subject motion; HOLLIES and GOLDMAN (1977) have recently described modifications to the static manikin measurement to allow for air motion and for subject generated air motion. This allows precise prediction of the rectal temperature (GIVONIand GOLDMAN, 1972) and heart rate (GIVONIand GOLDMAN, 1973a), responses of a man as a function of his clothing, his heat production (GIVONI and GOLDMAN, 1971), his state of acclimatization (GIVONIand GOLDMAN, 1973b) and the ambient environment. Thus assessment of the insulation and evaporative conductance of a clothing system can provide an accurate estimate of the thermal advantages of one garment or fabric relative to another. There are effects of cut, drape and fit which must receive special consideration (HOLLIESand GOLDMAN,1977). The techniques presented are valuabIe tools in clothing design and such evaluations are desirable in studies of the man-clothing-job-environment system for ordinary clothing, as well as for advanced clothing systcins with artificial environmental conditioning. The advantages of carrying out in the laboratory and climatic chamber the detailed evaluations of the thermal protection, and problems, associated with protective clothing systems should be obvious; precision measurement allows assessment of small differences, and prediction is possible for any combination of clothing,
Effects of clothitig
55
work level, wind and temperature. Similar precise techniques are being developed for measurement of the protection afforded to the extremities (hands, feet, face) by gloves, boots, face masks, hoods, etc. Thus, field evaluations of the thermal protective aspects of Arctic gear should never be carried out until the ‘%homework”in the laboratory has been completed, and then only as a validation of the laboratory findings under field conditions. On the other hand, the laboratory cannot provide the wear testing, the blowing snow and the realism of the field environment. Personnel involved in planning, carrying out or evaluating the results of such field evaluations should insist on having the results of the laboratory and/or climatic chamber evaluations of the specific items to be tested, before planning the details of the field evaluation. REFERENCES
FONSECA G. F. (1967), Moisture transjkr through perforated impermeable foam insulations, Textile Res. J. 37, 1072-1078. FONSECA G. F. and BRECKENRIDGE J. R. (1965), Windpenetration through fabric systems, Textile R a . J. 35, 95-103. A. C., and BAZETTH. C . (1941), A practical system of units for the description GAGCEA. P., BURTON of lieat excharge of man with his environment, Science 94,428430. GIVONIB. and GOLDMAN R. F. (1971), Predicting metabolic energy cost, J . Appl. Physiol. 30,429433. GIVONIB. and GOLDMAN R. F. (1972), Predicting rectal temperature response to work, environment and clothing, J. Appl. Physiol. 32, 812-822. GIVONI B. and GOLDMAN R. F. (1973a), Predicting heart rate response to work, environment and clothing, J. Appl. Physiol. 34, 201-204. GIVONI B. and GOLDMAN R. F. (1973 b), Predicting efects of heat acclimatization on heart rate and rectal temperature, J. Appl. Physiol. 35, 875-879. GOLDMAN R. F., GREEN E., and IAMPIETROP. F. (1965), Tolerance of hot wet environments by resting men, J. Appl. Physiol. 20, 271-277. GOLDMAN R.F. (1969), Physiological costs ofbody armor, Mil. Med. 134,204-209. GOLDMAN R. F. (1964), The Arctic soldier: possible research solutions for his protection, Proc. 15th Alaskan Science Conf. 15, 114-135. HOLLIES N. R. S. and GOLDMAN R.F. (1977), Clothing Comfort: Interaction of Thermal, Ventilation, Construction and Assessment Factors, Ann Arbor Science, Michigan. HUGHES A. L. and GOLDMAN R. F. (1970), Energy cost of “hard work”, J . Appl. Physiol. 29,570-572. IAMPIETRO P. F. and GOLDMAN R. F. (1965), Tolerance of men working in hot, hirmid environments, J. Appl. Physiol. 20, 73-76 JOY R. J . T. and GOLDMAN R. F. (1968), A method of relating physiology and military performance, Mil. Med. 33, 458-470. PANDOLF K. B., HAISMANM. F., and GOLDMAN R. F. (1976), Metabolicenergy expenditure and terrain coejficientsfor walking on snow, Ergonomics 19,683-690. WOODCOCK A. H . (1962), Moisture transfer in textile system, Textile Res. J. 32,628-723.
This Page Intentionally Left Blank
Chapter 4
HUMAN SE IN TEMPERATURE AP D CO TVECTIT E HEAT LOSS R. P. CLARK Laboratory for Aerobiology, Medical Research Council, Clinical Research Centre, Hnrrow. Middlescx, HA1 3UJ .Great Britain.
CONTENTS
Introduction Skin temperature measurement Skin temperature distribution at different environmental temperatures The effect of solar radiation on skin temperature Exercise and temperature distribution Temperature distribution over infants nursed in incubators The evaluation of mean skin temperature Convective heat loss from the body Natural convection Heat transfer to the boundary-layer The effect of posture on convective heat loss Forced convection The 'pendulum effect' Heat loss in highly turbulent flows Heat loss in air-conditioned buildings Conclusions
1NTRODUCI'ION
Much has been written about the thermal interaction between man and the environment. Classic works by authors such as WNSLOWand HEBRINGTON (1949) and BURTONand EDHOLM(1955) have long ago described the basic physical and physiological principles which have formed the basis of much further research. Recently new techniques have become available which enable investigations of the human micro-environment to be made in fresh ways. Some of these techniques and their results are discussed in this paper.
58
R. P. CLARK SKIN TEMPERATURE MEASUREMENT
In any study of man's thermal interaction with the surroundings, it is first necessary to measure the physical variables that control heat exchange with the environment. Temperature is one of the most important of these variables, because it is the difference between the skin surface temperature and that of the surroundings that enables the body surface to exchange heat with the environment. In most environments the skin is warmer than the surroundings and thus the heat transfer is away from the body to the ambient air. In some situations, for instance in hot climates, this may be reversed, and the body may gain heat from the environment where the air temperature is greater than the skin temperature. In these circumstances the body will still be able to lose heat by sweating and can achieve thermal balance, but at a higher body temperature (and therefore at a higher skin temperature) than in environments which are cooler than the skin surface. The classical idea of skin surface temperature distribution comes from the concept of a central hot core from which heat emanates to the surrounding tissues. This concept leads to a skin surface distribution where the temperature decreases as the distance from the 'core' increases. As we shall see in the following sections, this gives an accurate description only for a person at rest in moderate and warm environments. In warmer conditions and during exercise, where there are local areas of heat production, or in situations where there is considerable sweating, this classical pattern is greatly modified. Skin temperatures have generally been measured with thermocouple or thermistor surface contact probes. Temperatures read from a number of probes indicate the distribution over an area. However, even if a large number of measurements are taken it is difficult to obtain a detailed pattern of skin temperature distribution. In contrast, the relatively new technique of infra-red thermography enables detailed mapping of temperature distribution. The thermal radiation from the body surface, which is determined by the skin surface temperature, is received by a camera containing an infra-red detector. The signal from this detector is processed electronically to present a television-like picture, where various tones of grey represent particular temperatures. A refinement is to present the thermogram as a coloured display, where specific colours are assigned to particular temperatures. The technique enables temperetures over the whole body to be visualized without contact with the surface. Thus a further advantage is that a subject is not restricted in his activities. Infra-red thermography has recently been used in a number of investigations described in the following sections. SKIN TEMPERATURE DISTRIBUTION AT DIFFERENT ENVIRONMENTAL TEMPERATURES
Figure 4.1 is the thermogram of a young male subject (mean skinfold thickness of 4.2 mm) and shows the steady state skin temperatures at an environmental temperature of 11 "C with minimal air movement. One immediate and obvious feature of this thermogram is the large range of temperature between different regions.
8
Fig. 4.1. The steady state temperature distribution over a subject in a climatic chamber at a dry bulb temperature of 11 "C. Black and white reproduction of infra-red colour thermograms. Adjacent grey arms differ in teinpcratrire by 1 "C
60
R. P. CLARK
The temperature distribution of the skin surface may be related to that of the whole body by regarding the body as having a hot central 'core' away from which temperatures diminish. If the central 'core' is the abdominal cavity, the thorax, the brain, and the spinal cord. The skin over these areas will be warm because of its proximity to the 'core'. For example, the skin of the axilla is normally 1-2 cm from the 'core', whereas that of the buttocks may be as much as 6-7 cm away. At cool and moderate environmental temperatures the infra-red thermogram largely confirms the concept of a warm 'core' and cool periphery. Additional features reflect either the metabolism or the conductivity of the structures underlying the skin surface. For instance, there is a warm area under the axilla and another on the skin over the spine and shoulders. Areas of skin over bony tissue, for instance at the elbows and the knees, and over fat areas, such as the buttocks, are cooler. Although skin temperature is affected by environmental temperature, thermograms taken at 11 "C and 20 "C show many similar features. The main difference is that in the warmer conditions the underlying body structures are less evident, probably due to the greater thermal gradient through the skin in cooler environments and higher rates of local heat loss with less lateral diffusion of heat. At environmental temperatures above 27 "C the features due to the underlying body structures become less distinct. Figure 4.2 shows the skin temperature distribution at an air temperature of 31 "Cout of doors, in shade, with air movement of about 1 m s-I. Large areas of the skin surface are at the same temperature and the features such as the warm spine and the cold buttocks are no longer visible. At these higher air temperatures the gradient between the skin and the surroundings is small and there is consequently little convective cooling. Evaporative heat loss due to sweating is, however, considerableand these changesin the modes of heat loss combine to produce a more uniform temperature over the body. Whereas the skin temperature range at environmental temperatures between 10" and 20 "C is 10-12 "C at 31 "C this range is reduced to about 4-5 "C. THE EFFECT OF SOLAR RADIATION O N SKIN TEMPERATURE
The effect of solar radiation on skin temperatures has been investigated thermographically. The subject was exposed to sunlight at a dry bulb temperature of 31 "C.Areas of the body receiving direct radiation increased in temperature, surfaces such as the hair and swimming trunks reaching 48-50 "C. However skin temperatures increased by 5-6 "C only, to 38-40 "C.These results show differences between 'passive' surfaces such as the hair and clothing and a surface such as skin, which can thermoregulate actively by changes in blood supply and convective and evaporative heat losses. Temperatures over shaded areas actually decreased slightly during exposure to the sun, because of increased sweating. EXERCISE A N D TEMPERATURE DISTRIBUTION
Skin temperature distributions can be affected markedly by other factors. Exercise, in particular, (an greatly modify them. A recent study using infra-red thermography has demonstrated the temperature changes that may occur during running (CLARK,
Skin temperature and convection
61
Fig. 4.2. Steady state temperatures over a subject out of doors in shade at a dry bulb temperature of 31 "C. Air velocity about 1 m s-l. Black and white reproduction of infra-red colour thermograms. Adjacent grey areas differ in temperature by 1 "C
MULLANand PUGH, 1977). The temperature distribution over an Olympic class athlete was observed with the runner stationary, running on a treadmill at 4.5 m s-' in 'still air' and on a treadmill at 4.5 m s-l in the presence of wind at an equal velocity, to simulate outdoor running. These experiments were conducted at an air temperature of 11 "C in a climatic chamber. They were subsequently repeated out of doors on a running track, at 21 "C, when the Olympic class athlete was compared with a good club standard athlete during a one hour run. Measurements showed that the temperature distribution was changed completely during the exercise, reaching steady values after about twenty minutes. The changes were produced by changes in skin blood supply, sweat rate and evaporative cooling. Changes also occur because of convective cooling due to the motion of the limbs through the air, and in the metabolism in underlying structures, e.g. the increased
62
R. P. CLARK
heat output of the active muscles. This resulted in raised temperatures over the muscles, the arms and legs and cooling of the areas not overlying structures of high heat production. Measurements with the runner on a treadmill, both in the presence of wind and in 'still air' enabled an assessment of which temperature changes were caused by changes in blood flow and metabolism, and which were due to environmental cooling. In 'still air' the skin over the abdomen for instance, cooled as the exercise progressed while the temperatures over the active muscles of the thighs increased. Initially the hands and arms were cool; but later in the run there was a large increase in skin blood flow, just before the onset of sweating, which caused the hands and arms to rise by several degrees and become visible on the thermogram, as shown in fig. 4.3. When this experiment was repeated in the presence of wind, the temperature changes were larger and more rapid. The whole body was cooled more but the patterns of increased temperature due to active muscle were still dominant. When the Olympic class athlete was compared with the club standard man out of doors on a running track, the initial temperature distribution and its change during exercise was similar. However the different abilities of the two athletes to thermoregulate were strikingly demonstrated when the club standard man stopped sweating and had to retire with anhydrotic heat exhaustion (PUGH,1972). The thermogram showed that the cessation of sweating was followed by an immediate rise in skin temperature of more than 10 "C. This observation substantiated previous findings that cooling by increased airflow alone is not sufficient to account for the observed fall in skin temperatures during running. There must be sweating as well. Perhaps the most significant finding in this study of athletes was the re-distribution of skin temperature during exercise when the highest temperatures were found to coincide with the surfaces over active muscles. This elevation of skin temperature is evidence of direct heat transfer from the muscles to the skin surface, and is obvious despite other fictors such as local variations in air flow or sweating. Thermography is also sufficiently sensitive to show skin temperature changes associated with transient thermal stimuli. For example fig. 4.4 shows a series of thermograms which display the skin temperature distributions before and after taking a hot drink (CLARKet al., 1977b). TEMPERATURE DISTRIBUTIONS OVER INFANTS NURSED IN INCUBATORS
Measurements of the temperature distributions over infants nursed in a conventional incubator have been made recently (CLARKet al., 1978) at temperatures between 28-33 "C. The results showed a temperature distribution reminiscent of that found over the adult in moderate and warm environments (fig. 4.5); the 'hot core' with temperatures diminishing towards the extremities was clearly demonstrated. Very few features reflected the metabolism or conductance of underlying structures, except at the lower end of the temperature range where some structures were just beginning to appear. Of particular interest is the temperature distribution over the head, which in the infant accounts for some 20% of the total body area. The metabolism in the brain
Fig. 4.3. The temperature distribution over a subject on a treadmill in a climatic chamber at 11 "C.a - before running, b - five minutes after starting to run and c - after twenty minutes running and just before the onset of visible sweating. Black and white reproduction of infra-red colour thermograms. Adjacent grey areas differ in temperature by 1 "C
Skin temperature and convection
65
Fig. 4.5. Steady state temperature distribution over a baby nursed in an incubator. Air temperature 32 "C. Black and white reproduction of infra-red colour thennogram. Adjacent grey areas differ in temperature by 0.6"C
is also a large proportion of the total heat production in the neonate. Temperatures over the head were similar to those found over the warm central part of the body: in this sense the head may be regarded as part of the 'hot core' of the body. THE EVALUATION OF MEAN SKIN TEMPERATURE
A comparison was made of the mean skin temperatures obtained by analysis of infra-red thermograms and those obtained from thermocouple probe readings for the athletes and the neonates. The mean skin temperatures for the whole body agreed to uiithin 1.5 "C However, for particular body regions in the athletes, the mean values obtained by the two methods differed by up to 4 "C.Where large variations of temperature exist over small distances, the positions of probes are critical for their reading to be representative of the area considered. For instance, during exercise the skin temperature of the knee was more than 10 "C lower t?an that over the calf muscle, the distance between these areas was about 10 cm only. For the whole body a number of probe measurements (11-13 here) averaged out such positional inequalities between the two methods. For calculations of heat loss from the body it is essential to have a measure of mean surface temperature for each area. The complex skin temperature distributions shown therefore present a difficulty for models designed to simulate heat transfer from the body surface. Much simpler distributions are found in inanimate objects, which have led to the mathematical forms used to model heat transfer from the body. 5
- Bioengineering
66
R. P. CLARK CONVECTIVE HEAT LOSS FROM THE BODY
The temperature difference between the skin and the surroundings enables the body to exchange heat by conduction, convection, evaporation and radiation, and to achieve thermal equilibrium. Convective heat transfer varies due to the environment and activity and this is described in the following sections. NATURAL CONVECTION
For a subject with a mean skin temperature lower than air temperature, the air adjacent to the skin surface will become heated by conduction and will rise due to buoyancy. This results in an envelope of air moving upwards which surrounds the body (fig. 4.6). This natural convection boundary-layer flow forms an imporiant part of the humm micro-environment. It is the mechanism by which heat is lost from the body by convection, and is also important in the transport of desquamated skin particles which often carry micro-organisms and produce an airborne infection hazard. This air flow has been visualised and filmed using a schlieren optical system (CLARK and Cox, 1974; LEWISet al.. 1969). The system essentially consists of a parallel beam of light focussed to a particular point. If any of the light rays in the parallel beam are bent by passing through air with a different refractive index, the focus of the beam is displaced from its original position. By arrang'ng for the undeviated beam to be focussed through a coloured filter and for the displaced beam to be focussed through a different colour it is possible to visualise heated air streams as one colour against a background of another. Figure 4.7 shows a diagram of the schlieren optical system used to visualise the boundary layer flow over the whole body. These studies revealed a convective upflow of air which is thin and slow moving at the ankles and lower legs, but which
Fig. 4.6. A diagram of the natural convective air streams over a stand-
ing subject
quickly increases in velocity and thickness and is vigorous over higher parts of the body. At face height, for a standing naked subject, the maximum air velocity is between 0.4 and 0.5 m s-' .The corresponding boundary-layer thickness is 0.15-0.20 m. The plume of hot air above the head extends for at least 1.5 m before dispersing into the general room air, The total volume of air entrained is around 600 litres per minute at an air temperature of 25°C.
Skin temperature arid convection 'B'
'A
L i ght
67
mrabollc mirror
source
2nd Mirror
. Y
, 1 '
outer strips
Fig. 4.7. Diagram of the schlieren optical system
The natural convective boundary-layer flow may be characterised as being either laminar or turbulent. The parameter that describes the state of the flow is the Grashof number (Gr). This non-dimensional group is the ratio of viscous forces to buoyancy forces in the flow and is given by
where g is the acceleration due to gravity, v is the kinematic viscosity of the air, T, and T,are the mean air and skin temperatures, respectively, in degrees Kelvin, and h is the vertical height on the body. Figure 4.8 shows the relationship between the Grashof number and the vertical height on the human body, indicating regions of laminar, transitional and turbulent flow. When the Grashof number is less than lo9 the flow is laminar; when it exceeds 1Olothe flow is fully turbulent. For a standing naked subject with an 8-10 "Ctemperature difference between his skin and the air, the flow is laminar up to a height of 1 m, and becomes fully turbulent at 1.5 m; the region between the navel and the head is subject to transitional flow. Clothing modifies this pattern by reducing the external temperature gradient, but this is not as important in defining the flow as the vertical height, which appears as h3 in the equation. However, for a clothed man, when the temperature difference between the clothing and the air is small, it is possible to have a laminar boundary layer over the whole body. The velocity and temperature profiles have been evaluated for convective boundary-layers, around cylinders and flat plates. Measurements have been made in the flows around the human body, using thermocouple probes and constant temperature hot wire anemometers (CLARK,1973), but the complex shape of the body makes analysis difficult. Typical temperature and velocity profiles found in the convective
R. P. CLARK
65
r
Turbulent f l o w
Ir;
I
j;
Grashof number
Fig. 4.8. The variation of Grashof number with vertical height over the body surface, for a skin temperature of 33 "C and an ambient temperature of 25 O C
\U15 11
0
01
03
05
07
13
15
09
Dlstance f r o m heated surface ( c m )
Fig. 4.9. Typical velocity and temperature profiles in the natural convective boundary-layer flow
over a standing man
flow are seen in fig. 4.9. The velocity must be zero at the skin surface, increasing to reach a maximum value at about one third of the boundary-layer thickness. Subsequently, the velocity falls to ambient air velocity at the outside of the layer. The temperature profile is a smooth curve from the skin temperature at the body surface to the temperature of the surrounding air at the outer margin of the boundary layer. HEAT TRANSFER TO THE BOUNDARY-LAYER
Heat transfer from the body surface to the surroundings equals k dT/dy, (where k is the thermal conductivity of the air, and dT/dy is the slope of the temperature profile at the skin surface). The convective heat loss is therefore directly proportional
Skin temperature and convection
69
to the slope of the temperature gradient at the body surface. Thus, the thickness of the boundary-layer flow at any point will determine the local heat transfer from the body surface. Where the boundary-layer is thick, e.g. over the face and the upper parts of the body, heat transfer rates will be low. In contrast, over the ankle and lower legs the boundary-layer is thin and the rate of heat transfer will be high. The thick flow over the upper parts of the body such as the face and head afford a degree of thermal protection to the body by reducing the heat loss. THE EFFECT OF POSTURE ON CONVECTIVE HEAT LOSS
From the above description of the relationship between convective heat loss and boundary-layer thickness, it can easily be appreciated that posture will have an effect on convective heat loss because of the dependence of the convective flow pattern on vertical height. For example, the boundary-layer flow over the head and face of a standing man has developed over the full height of the body; the flow is thick the temperature gradient is shallow and consequently local convective heat loss is low. In contrast, when the subject is lying down the flow over the head and face is quite different, as illustrated in fig. 4.10. In this case the flow over the face
Fig. 4.10. The natural convective flow patterns over the head, for two postures
5tcnt!nq
and head has developed over the head only. The maximum air velocity over the face in a supine position is about 0.05 m s-' only, compared to 0.5 m s-' when standing. The insulating properties of the boundary-layer flow are effectively lost lying down. Recent experiments to measure convective heat output in these two postures have shown that the heat loss from the head in the lying position is about 30% higher than in the standing posture (CLARKand TOY, 1975a). These experiments used small surface calorimeters which measured the local heat loss from several areas of the face directly (CLARKet al., 1972). The relationship between convective boundary-layer flow and heat loss was clear; when the boundary-layer was thin higher heat transfer rates were recorded than from the areas where the boundary layer was thicker.
70
R. P. CLARK
When sitting, the boundary-layer flow is more complicated than in either the lying or standing positions. The flow over the horizontal knees and thighs interacts with the flow developed over the upper part-of the body so that the boundary-layer flow and heat loss over the head and face is little different from that in the standing posture. FORCED CONVECTION
When the body is exposed to a wind or is moving through the air, the natural convective boundary-layer flow is displaced and the body loses heat by forced convection. The variables that influence forced convection are the mean air velocity and the nature of the flow (i.e. whether it is laminar or turbulent), and the flow direction. The degree of turbulence and its scale can have a profound effect upon the heat loss. Many studies of heat loss from man in forced convection have been carried out, both indoors in climatic chambers and out of doors. The results of such studies are generally expressed in terms of the relationship between the heat transfer coefficient, h, (W m” K-’), and the mean air velocity to which the subject is exposed. However, most measurements have been made with unidirectional airflows. KERSLAKE (1972) has reviewed the results of such experiments, and gives the expression h, = = 8 . 3 ) 4 , where v is the air velocity in m s-I. Measurements of the heat loss distribution round the human head have demonstrated the dependence of forced convection on air flow direction in a unidirectional air stream (CLARKand TOY, 1975b). In practice the body is rarely subjected to truly linear forced convective flows; out of doors the wind is invariably turbulent. During movement the trunk and head only, of a person walking or running, perform a motion approaching straight line translation through the air. The moving limbs perform swinging motions ; the thighs and upper arms behave as pendulums and the lower legs and forearms have whiplash movements. These swinging movements modify the boundary-layer flow. THE ‘PENDULUM EFFECT
Studies using the schlieren optical system have shown the air-flow patterns around moving limbs (CLARKet al., 1974); complementary measurements of the local convective heat loss were also made. Visualization of the air flows around the legs of a runner have shown that the ‘pendulum effect’ produces completely different flow patterns to those found in linear flows. The flow around a swinging thigh forms a bow wave and a trailing wake and these are alternately established and reversed by each change in direction of the swinging leg. The flows around the lower legs and forearms, which perform a whiplash movement, are similar in nature, although more complex. Classical fluid dynamics and heat transfer theories are inappropriate for these conditions, as the movements of the body during walking and running are far more complicated than those associated with man-made structures on which the theory is based. Translation of the body through the air produces additional complications; a unidirectional airflow is superimposed on the alternating flows produced by the
Skin temperature and convection
71
'pendulum effect'. Schlieren visualisation shows similar flows around swinging and translating heated cylinders, used to simulate the action of the limbs during movement. Measurements of local convective heat loss around the thigh of a runner on a treadmill were made in a climatic chamber. The results show that, both in still air and in the presence of wind, the distribution of convective heat loss around the circumference of the thigh is different from that in a unidirectional airflow. Graphical integration was used to obtain a value for the overall heat transfer coefficient around the thigh. The coefficient was about twice as high as expected in a unidirectional wind equal to the mean velocity of the oscillating leg. A linear wind, representing the effect of translation of the body, further increased the convective heat loss. On the basis of these results it is estimated, that the heat loss from the whole body (assuming the trunk, head and face to have uniform translation) will be about twice that expected by considering the whole body to be in a uniform wind and applying the formula h, = 8.3 h. HEAT LOSS IN HIGHLY TURBULENT FLOWS
Heat losses may also be amplified when a subject is stationary in the presence of a turbulent or 'buffeting' wind. In this case, if the scale of turbulence is large in relation to the size of the man, the air will be continually changing direction. Such conditions have recently been studied in the harsh environment found beneath a hovering helicopter(CLAR(Ket al., 1976a; CLARKet al., 1976b; CLARKet al., 1977a). This work formed part of the evaluation of clothing for personnel who work on the landing decks of ships a t sea in the arctic. Measurements made beneath hovering helicopters confirm that the highly turbulent nature of the flow produced, for a given mean wind speed, much higher cooling coefficients than expected for the mean velocity. For instance, coefficients of up to 80 W m-* K-l were found at mean wind speeds of 20 m s-' ;the convective cooling predicted by the 'traditional' formula at this windspeed is about 30 W m-' K-' . One consequence of such high rates of heat loss is that people exposed to air movements of this kind at low temperatures require clothing with better thermal insulation than has hitherto been thought necessary. A particularly vulnerable area is the face, which is often left unprotected. In such environments subjects wearing highly insulating clothing may lose up to 40 % of their total body heat loss from the exposed face. The effects described in the preceding section have been discussed by CLARK(1976), and are summarised in fig. 4.11. HEAT LOSS IN AIR-CONDITIONED BUILDMGS
Thermal comfort in the environment where people live and work has received 1977). One problem much attention in recent years (HUMPHREYS, 1975; FANGER, has been to investigate comfort during exposure to air streams from different directions and at different levels of turbulence (OSTERGAARD et al., 1974). Such air streams will modify the convective boundary-layer flow around the body, and these modi-
R. P. CLARK
72
fications may be sensed by the bodyss the effeccomfort or discomfort. Recently studies have been carried out to asesses the effects of unidirectional downflowing air streams on patients and staff in a hospital ward (CLARKand MULLAN,1977). Fwc&
ccmectiw situatlons
Subject in a
Hwt loss in the highly turbu1eP.t flow)
inifa-m
b?fwoth a MicC$W h :%D L?/m-2K-i ct .%.in3s:Wds
SubJect runrmg wind h,- 303 \Nm-2K-’ h , ~ 8 3 ~ ‘ T W r n - ~ K - ’ at L.5 m s - ’
cf
and 9.w
ms-l
Fig. 4.1 1. Forced convective cooling for three kinds of air movement, h, denotes the convective heat loss coefficients
This work was carried out at Mount Vernon Hospital, Northwood, England during a three year clinical trial to assess the use of special air systems in the treatment of major burn injuries. This trial involved the construction of a new burns unit equipped with sophisticated air conditioning systems. The unit was designed to provide an acceptable thermal environment both for the badly injured patients and the staff who attended them. At the same time the air conditioning was required to produce bacteriologically clean conditions to reduce the chance of airborne cross infection, which is a major problem in treating large burns. The unit had four special care rooms where patients were nursed during the initial period of their treatment. These rooms had canopies suspended from the ceiling which produced a linear downflow of ultra-clean air beneath which the patients were nursed. It was possible to study convective cooling at two different velocities of the downflowing air (either 0.65 m s-’ or 0.30 m s-I). The results were compared to those obtained in the same room when the air conditioning was turned off. In the experiments a heated model was used to determine overall convective cooling coefficients. This model was an elliptical cylinder with a total surface area of 1.2 mz and thus represented the average bodysurface area. It was fitted with electrical heaters, the power was supplied from a control console where the voltage and current were measured. Thermocouple thermometers were used to measure the temperatures of the model surface and of the surrounding air. Experiments were carried out to determine convective cooling coefficients with the model freely suspended in air both vertically and horizontally. In addition, the tests were repeated with the cylinder resting horizontally on a hospital bed. The coefficients were determined with the cylinder completely uncovered and also when it was covered with a sheet; this represented the degree of bedding that was used within the burns unit. The relationship between posture and the convective boundary-layer flow has been discussed carlier, where it was shown that the convective flow over a standing
Skin temperature and convection
73
subject was quite different from that found oter a prone subject. The modification to the boundary-layer flow by the linear downflowing air streams is illustrated diagrammatically in fig. 4.12. Table 4.1 shows the relationship between cooling coefficients, posture and the speed of the linear downflowing air. For the vertical cylinder the downflowing air streams were too slow to produce forced convection.
n n n
Fig. 4.12. Diagram of the modification of the convective upflow by downflow in different postures
T a b l e 4.1 Convective cooling coeficients for a cylindrical model horizontal and vertical in 'still air'and when exposed to linear downflowing airstreams at two different velocities Convective cooling coefficient h, (W m-z K-') Downflow Downflow 'still air' 0.30 m s - I 0.65 m s-' Vertical cylinder Horizontal cylinder
6.7 6.6
5.9 5.7
4.5 10.8
However, the convective cooling coefficient decreased as the velocity of the downflowing air increased; the downflow competed with the vigorous convective upflow generated over the whole height of the cylinder. The interaction of the two air streams slowed down and thickened the boundary-layer flow, decreasing the temperature gradients; smaller gradients lowered convective heat loss. The effects for the horizontal cylinder were different. When the downflow w a s 0.30 m s-l the cooling coefficient was lower than in 'still' air. Conditions of forced convection were produced when the downflowing air speed was increased to 0.65 m s-l, and this was reflected in a large increase in the convective cooling coefficient.
P. R. CLARK
74
The convective heat loss coefficients for the heated cylinder, freely suspended in air, on a mattress uncovered, and covered by a sheet are shown in fig. 4.13. The arrows in the top part of the figure indicate the direction and thickness of the convective air-streams generated around the cylinder. The lower part of the diagram shows the convective heat loss 'envelope' where the radial distance from the cylinder circumference to the edge of the dotted area is proportional to ,the local convective heat loss from the cylinder surface.
Cylinder treely suspended
Cylinder on a mattress
Cylinder on a mottress and covered with a sheet
Mat tress h,= 1 2 8 Wrn-'K-'
h, = 9 L W m-'z K -
h,= 6 1 W rn'2K-'
Fig. 4.13. Diagrams of air flows and convective cooling coefficient 'envelopes' for a horizontal cylinder in three configurations, h, denotes the overall heat loss coefficient
The effect of the mattress is to eliminate part of the convective envelope and the sheet further modifies part of the envelope, indicated by the larger dots. The overall cooling coefficients determined are shown beneath eacb diagram, but, because of the difficulty in separating heat loss due to convection, radiation and conduction, the coefficients are presented here as total heat loss coefficients (he, Wm-2 K-'). Table 4.2 summarizes the coefficients measured in these experiments. The results showed that if the air speed was too high, the patient, who was horizontal, was at a thermal disadvantage compared with attendant nursing staff who were upright; this was an important factor in determining the correct air temperature for the treatment of the patients. It was shown that comfort for patients and staff could be prejudiced by high ventilation rates, even though they were considered desirable from a microbiological point of view. Nevertheless, when the air flows in the burns unit were permanently reduced to 0.30 m s-' , it was found that over a period of many months, the bacterial cleanliness was substantially the same as previously at the higher air flow rate of 0.65 m s-'. Experience in this unit and in other air conditioned hospital areas suggests that air velocities greater than 0.30 m s-l are not necessary from either a thermal comfort or microbiological standpoint;
Skin temperature and convection
75 Table 4.2
Overall cooling coefficients for a horizontal cylindrical model freely suspended and when placed on a bed in ’still air’, exposed to linear downflows at two different velocities Overall cooling coefficient h, (W rn-I K-I) Downflow Downflow ‘still air’ 0.30 m s-l 0.65 m s-I ~~
Horizontal cylinder freely suspended in air Horizontal cylinder on a bed, uncovered Horizontal cylinder on a bed, covered with a sheet
~~
12.8
11.0
17.4
9.4
13.3
12.5
6.1
9.1
5.8
higher air speeds produce nursing management problems and patients may be exposed to excessive heat loss, especially in areas where they may be nursed for long periods. CONCLUSIONS
Thermography produces a very detailed pattern of skin temperature, which is seen to be much more complex than has been revealed by thermocouple and other techniques. However, it is perhaps comforting to those who have used thermocouple techniques that the mean skin temperatures measured by the two methods agree reasonably well. Nevertheless, regional skin temperatures measured by thermocouples and thermography may be very different. The pattern of temperature change in different activities or in different environments, as shown by thermography, would also be very difficult to predict or demonstrate without this technique. This is particularly true in studies of the changing temperature distribution during exercise, where rapid changes occur over the whole body due to the effects of environmental cooling, extra heat production in the muscles and changed skin blood flow. The results obtained by thermographic and schlieren methods have modified a number of ideas on temperature distribution and convective heat loss. The complex nature of skin temperature distribution, and its variation with different environments and with exercise, can be demonstrated over the whole body with a detail which would be difficult to achieve with probe techniques. The description of the convective boundary-layer flow given by visualization of the moving air streams presents information which is largely different to that previously found in physiological textbooks. In the past great emphasis has been placed on the study of man in extremes of heat or cold. In contrast many of the current problems in environmental physiology are concerned with more normal environments, where the body’s response is often more difficult to measure. .The techniques described here are sensitive enough to provide visual demonstration of many of the thermal effects to which the body responds, and which otherwise are difficult to describe adequately.
76
R. P. CLARK REFERENCES
BURTON A. C. and EDHOLM 0. G. (1955), Man in a Cold Environment, Edward Arnold, London. CLARK R. P. (1976), Convective heat loss from the human body, Eng. Med. 5, 67-68. CLARK R. P. (1977). Environmental physiology and thermography, Proceedings of the Society of Photo-Optical Instrumentation Engineers, Vol. 110. CLARKR. P. (1973), The roIe of the human micro-environment in heat transfer andparticle transporl, PhD thesis, The City University, London. CLARKR. P. and Cox R. N. (1974). An applicationof aeronautical techniques to physiology, 1 - me human micro-environment and convective heat transfer, Med. Biol. Eng. 12, 270-274. CLARKR. P., Cox R. N., and TOYN. (1972). A surface plate calorimeter for measuring local heat transfer in fiee and forced convection within the human micro-environment,J. Physiol. 223,10-12. CLARK R. P., CROSSK. W., GOFFM. R., MULLAN B. J. STOTHERS J. K. and WARNER R. M. (1978), Neonatal whole body thermography, J . Physiol. 280, 2P-3P. B. J. (1977a). Heat loss studies beneath hovering helicopters, CLARKR. P., GOFFM. R., and MULLAN J. Physiol. 267, 6P-8P. B. J. (1977b), Skin temperaturesduringsunbathing and some CLARK R. P., GOFFM. R., and MULLAN observations on the effect of hot arid cold drinks on these temperatures, J. Physiol. 267, 8P-9P. R. P. and MULLAN B. J., (1977), Convectivebody cooling in air conditioned buildings, J. Physiol. CLARK 267, 9P-llP. B. J., and GOFFM. R., (1976a), A study using infra-red thermography of CLARKR. P., MULLAN cIothing assemblies for use by personnel working beneath operating helicopters, Royal Naval Personnel Committee Report ES 4/76. CLARK R. P., MULLAN B. J., and GOFFM. R. (1976b). Air velocity and convective cooling coeficient measurements beneath a Sea King helicopter, Royal Naval Personnel Research Committee Report ES 1/76. CLARK R. P., MULLAN B. J., and PuGH L. G. C. E. (1977), Skin temperatures during running - a study using infra-red colour thermgraphy, J. Physiol. 267, 53-62. B. J., PUGHL. G. C. E., and TOYN. (1974), Heat Iosses from the moving limbs CLARKR. P., MULLAN in running: the ‘pendulum’ effect, J. Physiol. 240, 8P-9P. CLARK R. P. and TOYN. (1975b), Forced convectionaround the human head, J . Physiol. 244,295-302. CLARKR. P. and TOYN. (1975a), Natural convectionaroundthe human head, J . Physiol. 244,283-293. FANGER P. 0. (1977), Thermal comfort in indoor environments, F:]Thermal Analysis - Human Comjort- ZndoorEnvironments,SymposiumProceedings, Nat. Bur. Stand. Specialpublicat ion 491, and J. E. HILL. Washington, eds.: B. W. MANGUM M. A. (1975), Field studies of thermal comfort compared and applied, Proceedings of the HUMPHREYS Symposium on Physiological Requirements of the Microclimate in Industry, Building Research Establishment England, Current papers, CB 76/75. D. McK. (1972), The Stress of Hot Environments, Cambridge University Press. KERSLAKE LEWISH. E., MULLAN B. J., FOSTER A. R., Cox R. N., andCLARK R. P. (1969), Aerodynamics of the human micro-environment, Lancet (i), 1273-1277. G. W. and ROSENBAUM J. C. (1963), Surface temperature measurement with thermocouples, MOLNAR [In :] Temperature, Its Measurement and Control in Science and Industry, Vol. 3, Part 3, Biology Chapman and Hall, London. and Medicine, ed.: J. D. HARDY, J., FANGER P. O., OLESENS..and MADSENTh. L. (1974). The efecr on man’s comforr OSTERGAARD of a uniform airjlow from diferent directions, ASHRAE Trans. 80, 142-157. PUGHG. (1972), The gooseflesh syndrome - acute anhydrotic heat exhaustion in long distance runners, Br. J. Physical Educ. 3, 50-56. L. P. (1949), Temperature and Human Lve, Princeton University WINSLOW C. E. A. and HERRINGTON Press.
.
11. MODELS AND INDICES OF HEAT EXCHANGE
This Page Intentionally Left Blank
Chapter 5
RATIONAL TEMPERATURE INDICES OF THERMAL COMFORT A. PHAROGAGGE John B. Pierce Foundation Laboratory end Department of Epidemiology and:Public Health, Yale University. New Haven Connecticut 06519, U.S.A.
CONTENTS Introduction Thermal comfort Thermal indices Operative Temperature The role of clothing in sensible heat exchange Temperature indices of insensible heat exchange Humid Operative Temperature The New Effective Temperature (ETC) Standard indices Standard Operative Temperature Standard Operative Vapour Pressure Development of a psychrometric chart for estimating Standard Effective Temperature (SET*) Sweating SET*, clothing and environmental acceptability The relation between ET* and SET* and warm discomfort during exercise Indices of cold discomfort Discussion Conclusions INTRODUCTION
The search for a simple physical index-of human response to the thermal environment has been vigorously pursued for the past half century. Engineers, psychologists and physicians, have each had an objective unique to their own professional interests. Ventilating engineers, who started in the 1920'~~ sought methods to predict the response of occupants to temperature and humidity in homes and buildings with
80
A. PHARo GAGGE
the newly developed central heating and air conditioning. Psychologists were interestcd in how human sensations of warmth and cold relate to the temperature of the environment. Physiologists were interested in how the thermal living environment affects the regulation of the human body temperature and how the effector processes necessary for this regulation (i.e. sweating, vascular changes, shivering and behavioural changes) affect human judgments of heat and cold (HALDANE,1905). Finally, the physician was interested in the health of persons exposed to extremes of cold, heat and humidity. All of these professional groups have at one time or another generated physical and physiological indices, which have served as predictive indices of thermal comfort, of thermal sensation, of heat and cold tolerance and of performance. The present chapter will describe physical or physiological indices of the thermal environment which have already been proven useful or which may prove valuable in future. There will be no attempt to present a complete review and evaluation of past work. Basic principles will be established by which indices, for thermal comfort in particular, may be developed and their usefulness judged.
THERMAL COMFORT
A definition of thermal comfort is necessary, if it must be related to any “index”. In the past such a definition has never proven to be simple. Cold and warmth were not recognized as separate human sensations until the latter part of the 19th century. Thermal comfort, as a measurable entity, was first recognized by the heating and ventilating engineers (HOUGHTEN and YAGLOU, 1923). They defined a “comfortable” environment as one sensed by the occupant as neither warm nor cold. Until the 196O’s, engineers have continued to measure comfort quantitatively by the use of category scales involving the following word sequence: cold -cool - slightly cool -comfortable - slightly warm -warm hot For purposes of statistical analyses each category is assigned a number either in a sequence from 1 through to 7 or on a positive, and negative scale (+3 to -3). to distinguish warmth or cold, with zero for neutral or comfortable, The above pragmatic approach to thermal comfort proved acceptable at first to engineers and perhaps to the early psychologists. However, it has never proven fully acceptable to physiologists, who recognized early that a state of thermal discomfort was not associated always with positive thermal sensation but with the nature of the tbermoregulatory response (by sweating, vascular changes, and shivering) to the physical environment itself. WINSLOW, HERRINGTON and GAGGE(1937) thus introduced the sensory categories scale of L6plea~nt’7-‘gunplea~ant.)-L’VeTYUnPleasant7’, which sensations were considered distinct and different from those with a temperature basis. In a later study GAGGE,STOLWIJK and HARDY(1967), used categories of “comfortable”-“slightly uncomforta ble’7-“~eryuncomfortable”:it was demonstrated that sensations of “slightly cool” or “slightly warm” could also be associated with
-
Indices of thermal confort
81
pleasantness or comfort. The studies by CHATONNET and CABANAC (1965), and later by CABANAC (1971) and STEVENS, MARKSand SIMONSON (1974) have revolutionized our understanding of the latter sensations. These show that local temperature sensations on the skin surfaces, whether warm or cold, may be either pleasant or unpleasant, depending on the mean body temperature and/or skin temperature. Currently, the clearest definition of “comfort” is the one used by the American Society of Heating, Refrigerating and Air-conditioning Engineers in their Standard for Thermal Environmental Conditions for Thermal Occupancy. This defines comfort as “that state of mind which expresses satisfaction with the thermal environment”. By such a definition comfort is given both a physiological and sensory basis. Thermal discomfort in either warmth or cold can be treated as a single variable which has been shown to follow a power law (STEVENS, MARKSand GAGGE, 1969). Alternatively, where we are concerned with a population rather than individuals, a state of “thermal comfort” is reflected by the percentage of those persons exposed to a given environment who indicate that it is “acceptable”. There never will be a perfect environment; there will always be a small percentage who will be dissatisfied. On statistical grounds, FANGER (1970) has set the ideal figure at 95 yo satisfied. For engineering applications the ASHRAE Standard has set the percentage satisfied at 80 for a comfortable environment. THERMAL INDICES
Because of the complexity of man’s thermal environment, even in indoor conditions, there have been many attempts to produce a satisfactory index to specify his environment by a single number. The usual dimension of such indices is temperature, which must be based on the heat balance equation, discussed by J. A. CLARKet al. and NISHIin chapters 1 and 2 of this book. The following sections of this review outline the derivation of some of the rational indices and discuss their application for the prediction of comfort and safety in work. OPERATIVE TEMPERATURE
In addition to the basic temperatures, ambient (To),and mean radiant (F,),the simplest derived environmental temperature index for sensible heat exchange is Operative Temperature (To); which is defined as the temperature of a uniform black (4n) enclosure in which a human occupant would exchange the same amount of heat by radiation (R) and convection (C) as i n the actual non-uniform environment. By this definition, RSC = h c r ( ~ s , , - ~ o ) , (5.1)
R+ C = h, (Tsu -Tr)+h, (Fsu-T>.
(54
where C and R are in units of W m-2 and T,,, is the mean temperature, in “C,of the outer surface of the occupant, whether it be skin or clothing, h, is the linear radiation 6
- Bioengineering
82
A. PHARO GAGGE
exchange coefficient given by
h, = (0.72)(5.67X 10-8)4[0.5(T,+Ta)+27313, when T, > T,,
> To,
or by
h, = (0.72)(5.67 x lo-') 4 [0.25 (Tr+T,+2 TSJ+273l3, when T,,> (T,+TJ/2, where in each case the factor 0.72 is the fraction of the total human body surface contributing to radiation exchange and the term in the second bracket is the Stefan-Boltzmann constant, in W m-2K-4. h, is the convective heat transfer coefficient, a function of air velocity Y , and pressure. From equations (5.1) and (5.2)
T, = (h,T,+~,TJ/h,Y (5.3) where the combined coefficient h,, = h,+h,. Operative Temperature is thus an average of T, and To,weighted by the respective linear radiative and convective heat transfer coefficients observed at the body exterior surface. Equation (5.3) may be rewritten as or
where 2ZB, Effective Radiant Field, represents the radiant heat received by the human occupant only from those radiative surfaces in the surrounding enclosure whose temperatures direr from To. When the various temperatures contributing to T, differ widely (e.g. solar heat against cold walls of a room) the relative absorptance of the various body surfaces must be considered. Skin and clothing surfaces may be considered black for thermal radiation. Operative Temperature, as derived and defined here, has appeared in the literature in many forms. Perhaps the earliest was the Tempdrature Resultante of Missenard ; other descriptions are Equivalent Temperature and Corrected or Adjusted Dry Bulb Temperature, used by the English and American engineers. Operative Temperature is a direct measure of the environmental heat stress on a human subject due to sensible heat exchangealone, and is not to be confused with the Effective Temperature and YAGLOU (I923), which is an empirical index of stress (ET), used by HOUGHTEN caused by both sensible and insensible heat exchange. THE ROLE OF CLOTHING
IN SENSIBLE HEAT EXCHANGE
Clothing adds a resistance to the flow of heat from the body surface to the outside. We may assume that the sensible heat transfer from the outer clothing surface (at mean temperature, T,J to the environment at Tois the same as the heat transferred
Indices of thermal comfort
83
through clothing, it can be shown that
R S C = hcr(T,--T0)7
(5.6)
R S C = h,F,(~,-T,) or hb(T,-To),
(5.6')
where the thermal efficiency factor, F,, equals l/(l+hJ) and I is the clothing insulation in units of m2 K W-l . Any form of sensible heat transfer which approximately obeys Newton's Law of Cooling may be described by a product hLr(T,--T2, where h:, is the effective sensible heat transfer coefficient involved (in W m-z K-l)' in our case equal to (hcrFc).For any environment, hir is primarily a function of Y, I and atmospheric pressure (in gases), as well as of the density and conductivity of the medium, which may be water, another fluid or various gas mixtures. A rigorous treatment of all these variables is being presented in other chapters in the present book, and has been reported elsewhere (GAGGE and NISHI,1977) for normal living environments. TEMPERATURE INDICES OF INSENSIBLE! HEAT EXCHANGE
Insensible heat loss by the evaporation of water and sweat from the skin surface must involve the gradient of vapour pressure from the skin surface to the ambient air. The evaporative heat transfer from an area of completely wet skin surface in air, as for sensible heat exchange, is a function of air movement, clothing insulation worn ( I ) and barometric pressure ( p ) in air. The permeability of clothing to water vapour is also a significant factor. A well established theorem of mass and heat transfer is that both convective and evaporative heat transfer from a wet surface are affected in the same way by air movement and air density (LEWIS,1922). This general analogy between convection and evaporation has been extended to normal everyday clothing. Thus the maximum evaporative heat loss (Em, W m-2) from a wet skin surface may be described by an equation analogous to equation (5.6) for sensible heat, as follows: E,,, = 16.5h,Fe,(e,-e,), (5.7) where 16.5 is the reciprocal of the psychrometric constant (K kPa-') at sea level and varies inversely with p . This may be derived from the Lewis relation. h,, the convective transfer coefficieEt, varies as the product ( V ~ ) O . FeC ~ ~ .is the permeation efficiency of water vapour through clothing, which for normal clothing is equal to the ratio l/(l+0.92hCZ).e, and e, are the saturation vapourpressures at Tsand Td,respectively, where Tdis the dew point temperature. The effectiveinsensible heat transfer coefficient hi (later in equation (5.9)) is described by the term hi = 16.5h,Fec. For a normally clothed subject ( I = 0.09 mz K W-' or 0.6 clo) in still air, F, would be 0.8 and the effective insensible heat transfer coefficient 43 W m-2 kPa-' . An accurate algorithm (Antoine equation) relating temperature and saturation vapour pressure is es = [16.6536-4030.183/(T+235)], where T is in "C and e, in kPa.
A. PHAROGAGGE
84
Em,as described in equation (5.7), represents the maximum heat loss by evaporation from a completely wet surface. Under these conditions the skin has a “wettedness” ( 0 )of unity. Generally, only a fraction of the skin surface is wet with moisture and it is never completely dry. When there is no sweating (E, = 0 ) , evaporation on th: skin surface is caused by diffusion of vapour through the outer layers of the skin (Ed). For normal comfortable but slightly cool conditions, this is equivalent to a wettedness ( w ) of approximately 0.06 (BREBNERet al., 1956). Skin wettedness therefore ranges from a minimum of 0.06 to unity during the course of body temperature regulation by sweating. The total evaporative loss ( E ) can be described by the equations: E = EdI-E,, (5.8) 0.94E,+0.06Em,
(5.8‘)
WE,.
(5.8“)
= =
Thus whenever E, = 0, E = E d ; when E, = Em,Ed is always zero. The total heat loss from the skin surface (H,) is thus the sum of the sensible and insensible terms given, in W m-2, by H, = hr,(T,-T,)+wh:(e,-e,).
(5.9)
When rearranged this becomes (eu-4
=
(-do)[ ~ o - ( ~ s - ~ s / ~ ; ) l ,
(5.9’)
where y = hf,/h: is the ratio of the transfer coefficients for sensible and insensible heat, and has units of kPaK-’. The significance of equation (5.9) may be realized by representing it on a psychrometric chart with Toon the abscissa and e on ordinate (see fig. 5.1). Equation (5.9‘)
Fig. 5.1. A psychrometric chart showing the effects of vapour pressure (el, temperature and skin wettedness (w) on the heat balance. Other symbols as defined in text
Indices of thermal comfort
85
describes a straight line between two points; point OP, whose coordinates are the Operative Temperature and vapour pressure actually observed at the location of the subject, and an imaginary point CP, whose coordinates are TS--Hs/h:,, and e,. The slope of the locus CP-OP is -y/w. The ratio y = hb/h: is a unique transfer characteristic of the total environmental heat exchange for a human subject. It is a function of air movement, clothing worn and barometric pressure. The CP-OP locus represents the combinations of Toand e that would result in the same total heat loss (H,) from the skin surface. When o = 1, the' resulting locus would represent the upper theoretical limit of the zone of evaporative cooling. When o w 0.06, there is no regulatory sweating, and the resulting locus will represent the lower threshold for sweating. y has a value of 0.11 kPa K-' for a normally clothed sedentary subject in still air and decreases with clothing insulation, as well as with increasing air movement and barometric pressure. Since thermal comfort is associated with T, M 33" to 34 "C and with o = 0.06, clothing insulation has a major effect on the T, necessary for comfort but has a less significant effect on the upper limit of evaporation. Increasing air movement raises both the T, for comfort and the upper limit of evaporative regulation. Increasing barometric pressure, over the range 1-4 atmospheres, has little effect on the T, for comfort, but significantly narrows the zone of evaporative regulation. y is inversely related to the permeability index, ,i described in GOLDMAN'S review (chapter 3). According to his definition i,,, = (l6.5y)-I, and i, is dimensionless. HUMID OPERATIVE TEMPERATURE
The Humid Operative Temperature (To&is defined as the uniform temperature of an imaginary black enclosure at 100% relative humidity (RH) in which a moist object or human subject would have the same total heat exchange by radiation, convection and evaporation as he (or it) would in the actual environment. By this definition equation (5.9) becomes (5.10)
By equating (5.9) and (5.10), Humid Operative Temperature is obtained by solution of the equation: (-.-eS(Toh,)
= -(Y/d@o-To/J,
(5.1 1)
which may be solved by iteration. Graphically, Tohis the abscissa at the intersection of the CP-OP locus in fig. 5.1 with the 100% RH (saturation) curve. As the skin surface dries (i.e. o +O), Tohapproaches To in value. For the case of a normally clothed subject (y = 0.11) the value for To,is 23" when o = 0.06, and 33", when o = 1. The values so derived are numerically similar to those given by the old 6Lclassical"EffectiveTemperature (ET) scale of HOUGHTEN and YAGLOU (1923).
86
A. PHmo GAWE THE NEW EFFECTIVE TEMPERATURE (@T*)
The new Effective Temperature (ET+)index, as now used by the American Society of Heating, Refrigerating and Air-conditioning Engineers (ASHRAE) is defined as the temperature of an imaginary uniform black enclosure at 50% RH in which a human subject would have the same total heat exchange by radiation, convection and evaporation as he would in the actual environment. By this definition, and by analogy with equations (5.10) and (5.11) above, ET* is obtained by solution of the following equation (5.12) = - ( Y / 4 Vo-TET’) eo-0.5 %*),,( or (5.12‘) 0.5%(,,*)+(Y/4 TET’ = e,+(dw) To * Thus in fig. 5.1, ET* is the abscissa at the intersection of the CO-OP locus with the 50% RH curve. Again, for our clothed sedentary subject the range of ET+ for o = 0.06 to w = 1 is from 24” to 42” C. The virtue of an ET+ type scale is that the values 24” for ‘%omfort”and 42 “C for “hot-very uncomfortable” match the “temperature” experience of an average North American or North European far better than the corresponding index range of 23.5” to 33 “C for To,,; saturated environments are relatively unknown to an average person. When rewritten as equation (5.12‘), ET+ can be solved by iteration with a pocket calculator (NISHI, 1977). For both To, and ET+ the same environmental characteristic, ly, applies to both the imaginary enclosure and test environment. The indices, calculated from (5.11) and (5.12), may be used to represent any values of Toand e in terms of the standard environment with a fixed relative humidity, such as 100% or 50%. STANDARD INDICES
The value of a “standard” index is its ability to describe the total heat exchange in any environment in terms of the heat exchange that would occur in a known familiar environment. Alternatively, if the human response to a well tested environmental condition is known, it should be possible to predict human response to any test condition in terms of that to a known condition. Our ability to predict the equivalence of different environments by the use of rationally derived indices will save the thermal physiologist or heating and ventilating engineer a great deal of experimental effort in the revalidation of known laboratory results. To apply successfully any such equivalencytest, it is necessary to know the effectivesensibleheat transfer coefficients, including the clothing insulation concerned, in both the test and standard environments. This is now a distinct reality. There are basically two types of Standard Indices - Standard Operative Temperature, for sensible heat transfer and Standard Operative Vapour Pressure for the insensible heat transfer. STANDARD OPERATIVE TEMPERATURE
Standard Operative Temperature (T’Jis defined as the temperature of a uniform enclosure (i.e. with T,= To)in which an occupant in still air at sea level, while wearing 0.6 clo as standard, would lose the same Eensible heat as he would in the test environment.
Indices of thermal comfort
87
By this definition for the test environment R+C
= h:,(Ts-To)
(5.13)
and for the standard environment
R+C
=h ~ s ( ~ s - ~ s o ) ,
(5.13')
where his is the effective sensible heat transfer coefficient for the standard eiiviron nient. Thus Tso = ( h ~ ~ / ~ ~ ~ ~ ) T , + ( l - Tj* h~,/h;j) (5.14) Equation (5.14) provides a method for predicting combinations of clothing insulation, air movement and barometric pressure that would have the Same sensible heat exchange as in the standard environment. STANDARD OPERATIVE VAPOUR PRESSURE
Standard Operative Vapour Pressure (e,,) is defined as the vapour pressure oa uniform enclosure in which an occupant, while wearing 0.09 mzK w-' (0.6 clo) standard, would lose the same insensible heat from his skin surface as he would in the test environment. If the insensible heat transfer coefficients are hi for the test and hi, for the standard environments, respectively, by definition E = wh:(e,-e,) = wh:(~&-eso)
(5.15) (5.159
thus
eso = ( ~ i / h ~ s ) e ~ - ( l - ~ ~ / h ~ ~ ) e s .
(5.16)
By use of equations (5.14) and (5.16), it is now possible to plot any beat transfer equation, such as equation (5.97, in terms of T, and e,, on a psychrometric chart analogous to fig. 5.1. The slopes of the various loci fot constant wettedness would now be h;,/(oh:,) or - y / w . The definitions of (a) Standard Humid Operative Temperature (TJoh)and (b) Standard Effective Temperature (SET") now follow logically: (a) Tsohis defined as the temperature of an imaginary enclosure at 100% RH in which a sedentary human occupatlt, dressed in standard clothing (I = 0.09 mz K W-' or 0.6 clo) in stillair, would lose the same totalheat by sensibleand insensible heat transfer as he would in the actual environment. TJohis given by the solution of an equation (similar to equation (5.11) : eSO-eS(Tsoh)
= -(vs/W)(Tso-Tsoh)
(5.17)
Graphically Tsohis the abscissa at the intersection of a locus with slope (-vJ/o) through the point (Tso,eso)on a psychrometric chart, with the 100% RH saturation curve. (b) SET* has a similar definition to Tsohexcept that the enclosure would have 50% RH. It is obtained by solution of the following equation, similar to equation (5.121, es0-0.5 es(sET*)= -(~Jm)(Tso-T&p*) (5.18)
88
A. PHARO GAGGE
or 0.5 es(SET*)+(Y’s/‘”))
TSET* = eso+(ys/o)
(5.18‘)
Tso9
or the intersection of the line with the 50 % R H curve. How equations (5.17) and (5.18) can be used to convert a psychroinetric chart for the test condition to the equivalent under the standard condition will be illustrated later, in fig. 5.6. DEVELOPMENT OF A PSYCHROMETRIC CHART FOR ESTIMATING STANDARD EFFECTIVE TEMPERATURE (SET*)
The standard environment chosen for a SET*-comfort chart is one describing typical heat transfer factors in a North American or European home or office building for a sedentary subject. The probable metabolic rate is between 55 and 65 W m-2 and the average total heat loss from the skin surface between 50 and 58 W m-” during thermal equilibrium. The applicable heat transfer coefficients, to be chosen as standard, arc listed in table 5.1. The values for the combined coefficients h,, It,, h:, and y Table 5.1 Standard heat transfer coefficients in air Heat transfer coefficients Activity
Sedentary M,, = 55 to 65 W rn-z (still air)
sensible (W rn-I K-’ ) hC = 3.0 h, = 4.4 h,, = 7.4 hLr = 4.4 tp
Moderate activity M,, = 165 to 180W rn-2 v = 0.8 ms-‘
=
=: =
h, = 50
1 = 0.09 m 2K W-I
11: = 40
(0.6 do) F = 0.6 I;; = 0.8
he
12.1 7.8
tp
I and F
0.11 kPaK-I
h, = 7.6 A, = 4.4 h,,
insensible (W rn+ kPa-’)
= 0.083
=
125
11; = 95
1 = 0.05 rn2K W-l (0.3 do) F = 0.6
r;, = 0.75
kPa K-I
Symbols as in text.
will increase slightly with Operative Temperature, due to the 4th power radiation law. Those here apply during “comfort”, when T, = 33 to 34 “C and Fsuis about 30 “C and To = 25 “C. The standard environment, defined in table 5.1, is very familiar to our daily experience and one for which there exists much experimental evidence (obtained at the Pierce Laboratory, Kansas State University, the Technical University of Denmark, and elsewhere) especially over the Operative Temperature range 15 to 40 “Cand relative humidity range from 20 % to 70 %. The chart in fig. 5.2 represents a series of loci for constant SET, any point of which will satisfy equation (5.9), the basic surface heat balance equation, as well a s
Indices of tltertnal cornfort
15
20
x 30 35 Operahe ternperdure ( T o r)r Ts0)
LO
L5
OC
Fig. 5.2. Psychrometric chart for evaluating Effective Temperature (ET*) from Operative Temperature (To) and humidity (eJ. The relative humidity lines apply when To = T, = T,. The category scales indicated are those observed in many studies in the field of heating and ventilating engineering. Since the standard coefficients for sedentary subjects (Table 5.1) are used here, the present chart can be used for SET* determinations in terms of Tsoand e,
equation (5.12) which defines ET* or SET,since the standard coefficients from table 5.1 are used in the calculations. In practice, during the regulation of body temperature the values of Fs, and E, (due to regulatory sweating) both increase with increasing environmental heat stress. In the cold E, is zero and 2", drops, in part due to vasoconstriction. In deriving fig. 5.2, a two-node model of body temperature regulation (GAGGE et al., 1971;NISHI and GACCE, 1977) has been used to develop the necessary realistic values for both Tsand o. The SET values represent those expected after one hours exposure to the conditions concerned. A category scale for thermal sensation, comfort and acceptability has also been marked in fig. 5.2, which summarizes our experience for the standard environment. Studies of the physiological background of fig. 5.2 show that, below 22" SET is essentially a linear function of Ts, temperature sensation and cold discomfort (see fig. 5.7); both psand internal body temperature are functions of the time of exposure. Below 15 "C,the body compensates by shivering. In heat stress, once skin wettedness has reached a maximum (w l), body heating sets in. Loci of constant SET above 41.5" all have the same slope, v j s , and each degree above 41.5" would represent a rise, after one hours exposure, of approximately 0.25"C in mean body temperature for an average size person (70 kg). For SET values above 41.5", the principal problem for the engineer and physiologist is tolerance time, discussed by VOGTet al. in chapter 6. If a total rise in body temperature of 1 "C is set as the limit of tolerance for health and performance, then on our chart this tolerance limit would be reached in the 45-46" SET range after 1 hour of exposure.
<
A. PHAROGAGGE
90
Figure 5.2 may be used in two ways. First, it may be used as a comfort chart for sedentary subjects by using observations of operative temperature and humidity in the occupied space and assuming the standard clothing insulation of 0.09 m 2K W-' (0.6 clo). Alternatively, the observed values for air movement, humidity, clothing worn and barometric pressure can be used to evaluate the heat transfer characteristics of the test environment. By adjusting the observed Toand e, to T, and e,,, using equations (5.14) and (5.16), the value of SET and the sensory response may be thus predicted. SWEATING
The Standard Operative Temperature (T'J may also be used as an index of sweating. In fig. 5.3 the sweat rates observed by the various authors indicated (GAGGE and HARDY,1967), have been plotted versus the T, observed for each experimental condition. The sweat rate ( E ) in each case was evaluated from the reported rate of
Relative humidity 30- LO%
0
A
A o
A
8
o
7
3b
2;
,o!
T,
Bed-ERF Chair -ERF Chair - T O M Clo -ERF Supine -calorimeter Supine -wind tunnel 2 node model
5; Lo L; 5h clothed, sedentary ("C)
A
;5
I
25
30
35
Equiv ,T
LO
15
unclothed. sedentary
50
("C)
Fig. 5.3. Chart showing the unique relationship between the observed evaporation (Es)and Standard Operative Temperature. Two abscissa scales are shown for T,, for a clothed and an unclothed subject. For references to observed data see fig. 8 in GAGOE and HARDY (1967)
weight loss of the subject, measured by a precision balance and normalized by the DuBois area formula. Values of T,,, have been found by using values of Fsand the transfer coefficients reported by the author, The experimental conditions represented cover a wide range of air movements, T,, Trand clothing, and may be related to fig. 5.3. When these data are reduced to a common T,, scale the common nature of temperature regulation by sweating during a resting state is evident. The slope of E, against T,,, is 4.2 W m-' K-' for a clothed subject, for the unclothed subject the slope would have been 6.2 W m-' K-'. Both slopes are slightly lower than those expected, 4.7 and 7.2 W m-' K-' respectively, if p , had remained constant through-
Indices of thermal comfort
91
out. This chart demonstrates that sweating during a given activity is proportional to Standard Operative Temperature. Therefore if either E, or T,, are known they can be used as a predictive index for the other, as long as skin wettedness is less than unity. SET*, CLOTHING AND ENVIRONMENTAL ACCEPTABILITY
Considerable energy savings can be achieved by lowering the dry bulb temperatures in work spaces during the winter and by reducing the need for air conditioning in the summer. Extension of the range of acceptable environments can be accomplished by proper choice of clothing. The use of a Standard Effective Temperature Index to estimate the required clothing insulation for comfort and acceptability is illustrated in fig. 5.4. The SET* temperatures on the ordinate may be interpreted in terms of those illustrated in fig. 5.2. The SET range 22.2 to 25.6 "C, has been observed to be acceptable to 80 yo of office workers, during a recent New York survey (GAGGE et al., 1976). This range was chosen earlier, as the optimum range of acceptability in the ASHRAE Comfort Standard (1966). In fig. 5.4 horizontal lines at the two temperatures mark the 80 % acceptable range for "slightly warm-slightly COO^'^. Fig. 5.4 can also be used to show for each ambierlt temperature the clothing insula-
Fig. 5.4. Humidity and air movement chart showing those combinations of ambient air temperature and clothing insulation that result in the same Standard Effective Temperature (on the ordinate). The paired horizontal lines indicate the ranges for 80% and 90% acceptability, based on a recent survey in a large office building (from GAGGE et al., 1976)
18
20
22
2L
28
26
Air temperature
30
32
"TI
tion necessary for comfort. The 80 % acceptability range (22.2-25.6 "C), as observed, applies to an average clothing insulation of 0.09 m2 K W-' (0.6 clo). For winter clothing of 0.155 m2 K W-' (1.0 clo), the air temperature range 19-22.5 "C should prove 80 % acceptable. For a summer clothing assembly of 0.05 m2 K W-' (0.3 clo) the ambient air temperature range of 24-27 "C should prove 80 % acceptable. Figure 5.4 can be used to indicate what clothing levels would be desirable for any chosen winter or summer conservation guidelines, samples of which are marked also on the figure. Corresponding charts for other levels of humidity and air movement can be developed using the same analytical equations (GAGGE et al., 19761.
92
A. PHAROGAGGE
THE RELATION BETWEEN ET* AND SET* AND WARM DISCOMFORT DURING EXERCISE
The data from the exercise study of GONZALEZ et al. (1977) illustrated in figs. 5.5 and 5.6 present observations of human responses during 45 minute exposures to the thermal environment while working unclothed on a bicycle ergometer ( M w 175 W m-2) in a test chamber. Air movement was about 0.75 m s- . The physiological observations associated with these tests are reported by GONZALEZ in chapter 8.
’
$ .oO
&
o\o
O ” \
Observed conditlons
Stondord mronment
Fig. 5.5. Two psychrometric charts in which the same test measurements are plotted (A) in terms of ambient air temperature (T,) and vapour pressure (ea),and (B) in terms of Standard Operative Temperature (T,) and Standard Operative Vapour Pressure (eso). Note the compression of the T‘, and e,, scales
Figure 5.5A is a psychrometric chart showing the observed e, for 111 different test conditions, with humidities ranging between 20 and 90% R H and T,(= T,) between 25 and 50 “C. Test values for h,, and he were respectively 12.6 W m-’ K-’ and 126 W m-2 kPa-‘. For each point each subject made a magnitude estimate of warm discomfort and the observed values of w , T, and e, were used to derive ET* using equation (5.12’). Trend lines for ET* have been drawn in fig. 5.5 A. However, in fig. 5.6 the comfort observations have been plotted versus the derived ET*. The estimates of discomfort are significantly correlated with ET* (r2 = 0.85). In terms of verbal categories an ET* of 27.5 “C was associated with “comfort”, while 43 “C was associated with “extremely uncomfortable” and “very hot”. Under the latter conditions T,approached 37 “C and oesophageal temperature approximated 39 “C, while skin wettedness was unity with sweat dripping off the skin surface. ET* also had a significant linear correlation with mean body temperature, T b ; as well as with heart rate. One difficulty of using any ET index scale is that it always applies, by definition, to a single activity. At any given activity, it is possible to “normalize” any operative temperature Toand ambient vapour pressure, e,, to give the T, and e,, of a reference environment. For the “sedentary” subject the 0.6 clo - still air situation was chosen
Indices of thermal comfort 5
25%
93
to2max
M = 175 Wrn2
t oi
G
c
0 L a,
a,
T l 3
-
I
oi
z
U h
1
rn
I
a
/
3
3 I
I
25
20
30
3 I
35
A -
15
LO
ET’ ( T I
is
M
25
30
SET
35
LO
‘<
!TI
Fig. 5.6. For each of the test conditions indicated in fig. 5.5 magnitude estimates of comfort and discomfort compared with both the Effective Temperature (ET*) and Standard Effective Temperature (SET*)
-
as standard. Activity at 175 W m-* (3 met) is li common work level, which can be maintained indefinitely. In practice a subject would wear light protective clothing, I = 0.05 m2 K W-’ (0.3 clo), during such an activity. By using this combination as a new standard (see lower part of table 5.1)and the observed Fsand w it would be possible to calculate SET from equation (5.18’) in terms of T,, and e8,. The resulting T,,, e,, and SET scales have been shown on both fig. 5.5B and 5.6. In terms of the “3-met” SET scale the temperature range 23-45 “C is again associated with the same thermal comfort-discomfort scale as was used for sedentary subjects. For the 3 met level of activity, “comfort” is now associated with an observed wettedness et al. (1937), FANGER (1970), GAGGE of 0.15-0.25 rather than with 0.06-0.1. WINSLOW et al. (1969) have observed that the level of both skin sweating and wettedness during comfort increase linearly with the level of activity. Because it assumes that human subjects will wear lower clothing insulation with increasing activity the SET comfort chart can well apply to any type of activity generally found in everyday living. The combinations of clothing and activity for equivalence of SET for discomfort were developed by a 2-node model of human and GAGGE,1977) and have been validated experitemperature regulation (NISHI mentally. These are: 60 W m-* (1 met) with 0.09 m2 K W-’ (0.6 clo); 120 W m-2
94
A.
hAR0
GAG~E
(2 met) with 0.07 m2 K W-* (0.45 clo); 175 W m-* (3 met) with 0.05 mz K W-' (0.3 clo) and 350 W m-' (6 met) for unclothed subjects.
%-
0
e
006 006 A09
E
I
e
-
a -;o;
01
/2
01
05 05
>A ' --CmI
4
:
@@-Coid
:
A
I
2s -
/ I n
,J'
sedentary
- slightly
COG;
Fig. 5.7. Chart showing the relation between mean skin temperature and Standard Operative Temperature in the cold. The small sized symbols represent model data under the conditions indicated. The large size circles and squares represent observed values for clothed and unclothed subjects
and for clothed and unclothed subjects. The sensation categories indicated are essentially those reported in the study on humans. The inflection at about SET* = 17 "C shows the presence of shivering. The observed data showed a 10 yorise in metabolism in the cold.
Indices of thermal comfort
95
DISCUSSION
The present rational indices are based on the physical properties of the environment plus knowledge of the physiology of the skin surface, which are involved in both heat and-mass transfer. Our understanding of the various avenues of heat exchange at the skin surfacenow make possible measurement of nearly every transfer coefficient involved in the human’s heat exchange with his thermal environment. The significanceof the various factors is now better understood. For example, measurement of clothing surface temperature is less important, as long as skin temperature can be accurately evaluated. Estimates of skin temperature based on thermal sensation alone are beginning to follow consistent patterns (CABANAC, chapter 12). Clothing insulation measurements for use in the zone of evaporative regulation need be accurate only within A35 % (NISHIet al., 1976) but in the cold these measurements should be within &lo %, especially for application of the present standard indices. Measurement of ambient air movement is an insignificant measurement compared to some direct measurement of the convective heat transfer coefficient itself. Measurement of air movement becomes even less significant when clothing is worn. Some of the present indices have their analogues in those of the past. Humid Operative Temperature (Tofi)is numerically and quantitatively close to the older ET of HOUGHTEN and YAGLOU.Tsofiis closely related to the adjusted Effective Temperature of the English school (BEDFORD, 1948), when both wind movement and clothing are considered. Loci of constant old ET, when plotted on a psychrometric chart, prove to be parallel lines and do not have the convergence typical of the new ET* loci. This convergence is caused by the increased wettedness of the skin surface in the zone of evaporative regulation, associated with higher values of ET*. As a consequence, the old ET tended to exaggerate the importance of humidity in the cold and underestimated it in the heat. At the middle of the zone of evaporative and ET fall in the 26-3 1 “Crange, there is good numerical regulation, where both Tsoh agreement between the two indices. Many combinations of instruments have been used in the past for measurements used to predict “Effective Temperature” and comfort. The simplest has been the black 15 cm diameter globe thermometer used by BEDFORD for measurement of T,. The uncorrected temperature of the globe (T’), especially with paint (pink) simulating the skin absorptance, can be used as a direct measure of Operative Temperature (To) over a wide range of air movement up to 2 m s-’. YACLOU(1949) successfully developed the WBGT index based on a weighted average of T, (lo%), Tg(20%), and unventilated wet bulb temperature (70%) to simulate his ET. Other indices for effective temperature use various weighted averages of the dry and wet bulb (e.g. the Oxford Index; LEITHEAD and LIND,1964). During the summer the U.S.A. Weather Bureau uses a Temperature-Humidity-Index (THI) (THOM,1971) in which T, and the wet bulb temperature, T,, are averaged. In 1905, HALDANE suggested the ventilated wet bulb temperature alone as a good index of heat stress. Loci of constant temperature for all the above physical indices prove to be parallel lines when plotted on a psychrometric chart - but with varying negative slopes. The steepest is the THI-Index which simulates comfort (see fig. 5.2). The slope of the WBGT or old E T
96
A. PHAROGAGCE
indices are parallel to those for SET values of 25-30 “C. The slope for unventilated wet bulb loci and those of the Oxford Index would be parallel to SET between 36 and 37 “C. The ventilated T,”loci would be the least steep and come close to those for our y , = 0.8. In short, there is no one all purpose physical instrument to estimate comfort and the stress of the thermal environment. Only rationally derived indices can be used for this purpose. Other analytical methods have been used for prediction of thermal sensation, including discomfort. The most well known is FANGER’S “Comfort Equation”, which is based on two experimental observations: (1) that the F, for comfort decrease linearly with increasing metabolic rate or activity; and (2) that the sweating required for comfort also increases linearly with metabolic rate. By eliminating both T, and E in the heat balance equation, FANGER (1970) can describe any combination of M , T,, T r yair movement and clothing insulation, which would result in a state of thermal comfort. Deviations from this base line are used as an index of his “Predicted Mean associates temperature Vote” (PMV) based on a -3, 0, +3 scale. In the cold FANGER sensation with body cooling (i.e. negative storage). In the heat the difference between the evaporative loss for thermal equilibrium, i.e. [M,-(R+C)],and the comfort threshold value of E is associated with warmth sensation. He has introduced a third factor, which reduces the sensitivity of his PMV to these stresses as level of activity is increased. His equation does not include the effect of humidity and Em on maximum skin wettedness but assumes that all evaporation occurs at the skin surface. His PMV scale is closely parallel to the temperature sensation scale, that would follow from the use of ET* and SET* at humidities below RH of 60%. Successful computer models simulating temperature regulation of the human body are being developed continuously (e.g. STOLWIJK and HARDY,1966; GAGGE and NISHI,1977; AZERand Hsu, 1977; and chapter 7). The purpose of all is to predict, for the activity concerned and for all basic environmental variables, the resulting skin temperature, skin blood flow, regulatory sweating and internal body temperature as well as to interpret how these physiological values will affect temperature sensation and comfort. CONCLUSIONS
With the balance equation for human heat exchange with the thermal environment as a starting point, the following rational temperature indices have been developed: ( I ) Operative Temperature (70), which is a linear average of the ambient air temperature (T,) and mean radiant temperature (Tr),weighted by their respective linear transfer coefficients;(2) Humid Operative Temperature (Toh)and the New Eflective Temperature (ET*). Both Tohand ET* are indices of the total heat exchange from the skin surface, while Tois a measure of the sensible heat exchange only; Tohis based on an equivalent saturated environment, while ET* on one with 50% RH. The effective transfer coefficient governing the evaporative loss from the skin surface is always inseparable from the associated skin wettedness factor. Transfer coefficients for sensible and insensible heat from the skin surface are modified by clothing insulation.
Indices of thermal comfort
97
When a reference environment is used with standardized clothing and air movement it is possible to describe any test environment in terms of (3) a Standard Operative Temperature (Tso);(4) a Standard Ambient Vapour Pressure (e,,) and (5) the Standard Efective Temperature combining both sensible and insensible hea.t transfer. As was shown in fig. 5.3, any environment can be described in terms of a standard environment. Thus human response, whether physiological or sensory, to any complex thermal environment, can be compared to its known response in a standardized and familiar environment. Finally, by use of a standard temperature index scale, a prescriptive temperature range, based on controlled laboratory experiments, can be established for comfort and various levels of discomfort as well as tolerance to heat and cold. REFERENCES American Society of Heating, Refrigerating and Air-Conditioning Engineers, New York 1966, Standard 55-66, “Thermal Comfort Conditions”. AzeR N. Z. and Hsu S. (1977), The prediction of thermal sensation from a simple model of human physiological regulatory response, ASHRAE Trans. 83. BEDFORD T. (1948), Basic Principles of Ventilation and Heating, H. K. Lewis, London. D. F., KEWLAKE D. McK., and WADDEL J. L. (1956), The d i r i f o n of water vapour through BREBNER human skin, J . Physiol. London 132, 225-231. CABANAC M. (1971), Physiologicd role of pleasure, Science 173, 1103-1 107. CHATONNET J., CABANAC M. (1965), Theperception of thermal comfort, Int. J. Biometeor. 9, 183-193. FANGER P. 0. (1970), Thermal Comfort. Analysis and Application in Environmental Engineering, Danish Technical Press, Copenhagen. J. D. (1967), Thermal radiation exchange of human bypartitional calorimetry, GAGGE A. P. and HARDY J . Appl. Physiol. 23, 248-258. GAGGE A. P. and NISHI Y. (1977), Chapter 5. Heat exchange between human skin surface and thermal environments. [tn :] Handbook of Physiology, Section 9: Reactiom to Environmental Agents, ed.: D. H. K. LEE, American Physiological Society, Bethesda, Maryland, pp. 69-92. GAGGEA. P., NISHIY., and NEVINSR. G. (1976), The role of clothing in meeting energy conservotion guidelines, ASHRAE Trans. 82, 234-247. GAGGE A.P., STOLWUK J. A. J., and HARDY J. D. (1967). Comfort andthermal sensations and associated physiological responses at various ambient temperatures, Environ. Res. 1, 1-20. GAGGE A. P., STOLWJK J. A. J., and NISHIY. (1969), The prediction of thermal comfort when thermal equilibrium is maintained by sweating, A S H R A E Trans. 75, 108-125. GAGGEA. P., STOLWIJK J. A. J., and NISHIY. (1971), An effective temperature scale based on a simple model of human physiological temperature response, ASHRAE Trans. 77, 247-262. GAGGE A. P., WINSLOW C. E. A., and HEFWJNGTON L. P. (1938), The influence of clothing on physiological reactions of human body to varying environmental conditions, Am. J. Physiol. 124, 30-50. GONZALEZ R. R., NISWY., and GAGGE A. P. (1977), Mean body temperature andefective temperature as indices of human thermoregulatoryresponse to warm environment, [In:] New Trends in Thermal Physiology, eds.: Y.HOUDAS and J. D. GuIeu,Masson, Paris. HALDANE J. S. (1905), The influence of high temperature, J. Hygiene 5, 494513. HOUGHTEN F. C. and YAGWUC. P. (1923), Determining lines of equal comfort, ASHVE Trans. 28, 163-176 and 361-384. L e m m C. S. and LINDA. P. (1964), Heat Stress and Heat Disorders, Churchill, London. LEWISW. K. (1922), The evaporation of a liquid into a gas, ASME Trans. 44, 325-335. NISHIY. (1977), Field assessment of thermal characteristics of man and his environment by using a programmable calculator, ASHRAE Trans. 83, 103-124. 7
- Bioengineering
98
A. PHAROGAGGE
NISH!Y . and GAGGEA. P. (1977), Effective temperature scale useful for Iiypobaric and hyperbaric environments, Av. Space Env. Med. 48, 97-107. NISHLY., GONZALEZ R. R., NEVINS R. G., and GAGGEA.P. (1976). Field measurement of clotliing thermal insirlation, ASHRAE Trans. 82, 248-259. STEVENSJ. C., MARKSL. E., and GAGGEA. P. (1969), The quantitative assessment of thermal discomfort, Environ. Res. 2, 149-165. STEVENS J. C., MARKS L. E., and SIMONSON D. (1974), Regional sensitivity and spacial summation in the warntth sense, Physical Behavior 13, 825-836. STOLWIJK J. A. J. and HARDYJ. D. (1966), Temperature regulation in man - a tlieoretical study, Pflugers Arch. 291, 129-162. THOM H . C. S. (1971), ASHRAE Bull. WA-71-5. WINSLOW C-E.A, HERRINGTON L. P., and GAGGEA. P. (1937), Relations between atmospheric conditions, pliysidogical reactions and sensations of pleasantness, Amer. J. Hyg. 26, 103-1 15. YAGLOUC. P. and MINARDD. (1957), Control of heat casualties at military training centers, Am. Med. Ass. Archs and Hlth, 16, 302-316. C. P. (1968), Indices oJcomfort. [In :] Physiologyof Heat Regularion, ed.: L. H . NEWBURGH, YAGLOU Hafner, New York and London.
Chapter 6
REQUIRED SWEAT RATE AS AN INDEX OF THERMAL STRAIN IN INDUSTRY J. J. VOGT, V. CANDAS, J. P. LIBERTand F. DAULL C . N. R . S . Bioclimatic Rcsearch Centre, 21 Ruc Bccquerel, 67087 Strasbourg-CCdex, France.
CONTENTS Introduction Calculation of the Required Sweat Rate Validation of the Required Sweat Rate Discussion Required Sweat Rate as an index of heat strain in industry Conclusions
INTRODUCTION
In industrial work situations, such as in the steel, glass, ceramic and textile industries or in mines, man is often exposed to heat stress. If the resulting physiological strain is too great acute or chronic damage may develop. Therefore estimation of the stress is required, because it is the only objective means to compare different work situations and to put upper limits for industrial exposures. This is especially important in view of the present emphasis on working conditions and their role in improving the quality of life. Estimation of the total heat load is difficult because of the number of variables to be taken into account: metabolic heat production, clothing insulation, air temperature, mean radiant temperature, air humidity and air velocity. The fluctuations of these variables during exposure to hot conditions require frequent measurements
100
J. J. VOGT et al.
during the work shift. The difficulties of such an estimation must be overcome in view of its many practical consequences. Man’s well being, health, safety and efficiency may depend upon its accuracy. The ergonomist may employ estimates of heat loads; first, to define the existence and the nature of risks due to the heat stress; second, where a risk does exist, to predict a maximum exposure time to the environment; third, where no risk exists, to compare the respective heat stresses in work places and to propose appropriate improvement. Many indices have been proposed for industrial purposes since YAGLOU (1927) gave us his “Effective Temperature” scale (ET). This was based on the subjective judgement of equal thermal sensation by subjects moving from one set of conditions to another. Nomograms were proposed for men wearing shorts or overalls and various corrections included in order to take into account radiant temperature. A number of critical studies of ET, performed between 1950 and 1960, were analysed by KJZRSLAKB (1972). Despite its limitations this scale was recommended by a World Health Organization Committee (1969), as a general index of thermal stress in industry. Safe limits of heat exposure were suggested for acclimatized and nonacclimatized men, at three levels of metabolic heat production. The “Predicted Four Hour Sweat Rate” index (P4SR) was developed from experiments on young acclimatized naval ratings (MCARDLEet al., 1947). The index combines the effects of air temperature, radiant temperature, humidity, air movement, metabolic heat production and two levels of clothing (shorts and overalls). The empirical nomograms derived from these experiments allow the prediction of the basic Four Hour Sweat Rate. Calculations allow corrections for clothing and metabolism. The procedure is rather complicated, but a number of investigations have confirmed that its predictions are quite accurate (-LAKE, 1972; LHTHEAD and LIND, 1964). The safe limit of P4SR for healthy and acclimatized young men seems to be about 4.5 litres. Lower values have to be considered for non-acclimatized men. The “Heat Stress Index” (HSI) of BELDING and HATCH(1955) is based on the physical analysis of heat exchanges. Two quantities are estimated: the rate of evaporative heat loss required for maintaining the heat balance E,. and the maximum evaporative capacity with fully wetted skin Em. The ratio EJE, gives the index value corresponding to the skin wettedness (GAGGE,1937). Several simplifications of the calculation of this index have been proposed. Corrections for clothed subjects were introduced by HERTIGand BELDING (1963). Safe limits of this index are taken as a combination of an index value lower than 100 and required evaporation lower than 1 litre per hour. The “Wet Bulb Globe Temperature” index (WBGT) was developed in 1957 by YAGLOU and MINARD.The WBGT uses the “natural” wet bulb temperature, TL, of an unventilated wet bulb thermometer exposed to the ambient radiation, and the globe temperature T,. For indoors the formula is (WBGT) = 0.7 Tk+0,3 T,and for outdoors the formula includes the air temperature T,, becoming (WBGT) = 0.7 T;+O.2 T,+O.IT,. Due to its simplicity (no measurementsof air velocityare required) and success in recent years, the National Institute of Occupational Safety
Sweat rate and tliermai strain
101
and Health in 1972proposed its use as a general indexof thermal stress in the U.S.A. Safe limits of heat exposure are predicted by using different WBGT values for three levels of metabolic heat production and two levels of air velocity. The “Index of Thermal Stress” (ITS) was described by GIVONIin 1963. This index allows the prediction of the sweat rate required for thermal equilibrium in subjects wearing summer clothing. As in the case of HSI, E, and Em are first estimated the next step is to calculate their ratio and to find from this the efficiency of sweating. The index value is then calculated as a predicted required sweat rate. Despite the complex calculations involved, even for nude subjects, we consider this index the most comprehensive to date for general evaluation of any industrial heat stress; other indices previously described allow only semi-quantitative comparisons. No other index allows a logical evaluation of safe exposure times. For this evaluation it is necessary that the index takes into account the heat balance. The only other index which is based on heat exchanges (HSI) does not allow this, because it takes into account only evaporation and not the sweat rate, which is the limiting physiological variable. The sweat rate may exceed evaporation because, as shown by GIVONI (1963) and LIBERTet al. (1975), some sweat drips from the body when skin wettedness is between 0.7 and 1.0. Area I of fig. 6.1 corresponds to such an underestimation for the Heat Stress Index scale. The Predicted Four Hour Sweat Rate does take into account sweat loss. But this index, which is founded on empirical correlations between observed sweat rates and a set of influencing factors, gives no precise estimate of the proportion of sweat secretion which can be evaporated. Consequently P4SR cannot predict safe exposure times. Area 2 of fig. 6.1 represents underestimation by P4SR.
t
Fig. 6.1. Theoretical required sweat rate of 709 g h- plotted on the psychrometric chart, simplified as full straight lines. The dashed straight lines represent skin wettedness w = 0.7 and o = 1.0. Tbe line ---.- corresponds to required evaporation of 700 g h-I (as in Heat Stress Index). Area I correspondsto underestimation of sweat rate by HSI; Area 2 corresponds to the underestimation of sweat rate (as in P4SR). Conditions: metabolism 200 W, sweat rate 700 g h-l, clothing insulation 0.05 m2 K W-’ (0.3 do), wind speed 0.5 m s - l
J. J. V o ~ et r al.
102
Further discussion of the different heat stress indices may be found in the books and LIND(1964) and of KERSLAKE (1972) and in the paper by GAGGE of LEITHEAD and NISHI(1976). Our purpose was to work out and validate an index of thermal stress. This is based analytically on the heat balance equation and estimates the sweat rate required to maintain thermal equilibrium, following the principle of GIvoNI’s Index of Thermal Stress. The calculation procedure for required sweat rate uses laboratory determination of the heat exchange coefficients, whereas GIVONI’Sequation bears little relation to the physical processes ; his coefficients were determined empirically. Validation of our procedure was carried out in industrial situations in order to compare Required Sweat Rate to observed sweat rates during work. CALCULATION OF THE REQUIRED SWEAT RATE
During steady state in hot conditions, the heat balance equation is
M,+ W+Ee +Ej+E,+RSC
=0
(6.1)
where all terms are expressed in units of watts per square metre of the external surface of body (W m-’) and M , is the net metabolic energy expenditure estimated from the oxygen consumption; W i s the external mechanical work; E, is the respiratory evaporative heat loss; Ei is the insensible skin evaporative heat loss; E, is the sweating evaporative heat loss; R is the radiant heat flow; C is the convective heat flow. Here E, represents the evaporative heat loss required for heat balance. When the actual heat loss by sweating (E,) is less than E,; the thermal state of the body is no longer steady, and a heat storage term must be included in the balance equation. For a naked subject the maximal evaporative heat loss of the fully wetted skin is given by E m = he (es-ea),
(6.2)
where Em is the maximal evaporative heat loss, he is the evaporative heat exchange coefficient (W m-2 kPa-’), and e, and e, are the saturated water vapour pressure at the skin temperature and the water vapour pressure of the ambient air respectively. Evaporative heat transfer coefficients are different at each site of the skin surface. When skin wettedness increases it becomes inevitable that some skin areas exhibit excess sweating (KERSLAKE, 1963: GIVONI,1963; LIBERTet al., 1975; CANDAS et al., 1978). When dripping sweat appears the sweat efficiency, q , defined by the ratio of EJI to the rate of sweat secretion S (in g m-’ s-’) decreases. Their relationship is of the form q
= E,/AS = ae-bw,
(6.3)
where a and b are numerical coefficients with values of 6.3 and 2.3, respectively, determined by measurement; 1 is the latent heat of vapourization of water and o is the skin wettedness, which corresponds to the ratio o = E,/E,.
(6.4)
Sweat rate and thermal sttain
103
The ratio EJE,, represents the skin wettedness required in order to realize the required evaporation, taking into account the maximal evaporative capacity of the environment. When w > 1 prolonged exposure is unsafe because E, cannot be realized. Safe exposure times must then be computed from the difference @,-Em), which is a heat storage rate, by fixing the maximal heat quantity whose storage will not increase the body temperature by more than an accepted amount. If w < 1 the Required Sweat Rate, S,, can be computed by dividing E, by the product of the efficiency of sweating and A . S, = E,/?lA
(6.5)
S, (g m-’s-’) represents the sweat rate needed to achieve thermal balance. It allows both for the evaporation and dripping of sweat, and may be compared to S,, the possible sweat rate that an individual can produce according to his acclimatization status. If S, < Sp no special problem arises. If S, > S,, it is necessaiy to compute safe exposure times from their difference, (S,-S,,). Clothing reduces radiative and convective heat transfer as well as the removal of heat by evaporation. It must therefore be taken into account in the calculation of Required Sweat Rate. The thermal efficiency factor and permeation efficiency factor given by NISHI (1965) were used. The required sweat secretion rate (Shr) in grams per hour for a subject of area A = 1.8 m2 is S,,, = 3600 S,A, where S, comes from equation (6.5). Therefore
For industrial application, we neglected E, and Ei in equation (6.1), since they are usually small compared to the other terms of the heat balance equation. Heat transfer coefficients and clothing factors used in our calculations were based on the work of MISSENARD (1973) and NISHIand GAGGE (1969), respectively. VALIDATION OF THE REQUIRED SWEAT RATE
In order to validate the use of the Required Sweat Rate, for assessing physiological strain in actual work and heat stress situations, the Required Sweat Rate was compared to sweat rate measurements on miners working in hot dry conditions. Calculation of the Required Sweat Rate necessitated measurement of six variables in each work place. Oxygen consumption iiieasurements were made with a Douglas bag method. Each measurement took at least five minutes. All typical work operations were sampled. From these results mean values of energy expenditure for each work period and for each work shift were computed. For the different work phases measured energy expenditures were between 105 W and 490 W. The mean values over whole work shifts ranged between 140 W and 325 W. Clothing insulation was estimated from tables (FANGER, 1970). Values were between 0.0155 and 0.062 m2 K W-’ (0.1 and 0.4 clo, respectively). Air temperatures
J. J. VOGTet al.
104
were measured by mercury thermometers, the range was 24 to 46 "C.Mean radiant temperatures, calculated from globe temperature measurements, were between 26 and 48 "C.Wet bulb temperature, read on an aspirated psychometer, was between 19 and 31 "C.Air velocity was measured near the miners body with a vane anemometer, measured values were between 0.1 and 4.2 m s-I . Sweat rates were computed from time changes of body weight. At the beginning of the work shift the miner was weighed and his clothing was noted. During the work shift an observer noted every ingestion of water or food and the volume of urine excreted. Body weight was measured as often as possible in order to determine the time course of sweat loss. Figures 6.2 and 6.3 show the relationships between the observed sweating rate and the Required Sweat Rate, computed for the whole work shift either by taking the mean values of the Required Sweat Rates computed for each work period (fig. 6.2) or taking the sweat rate calculated from mean values of energy expenditures, clothing and environment for the whole duration of the working shift (fig. 6.3). Correlations are highly significant (p < 0.001). The standard error of estimates is less than 100g h- and corresponds to the lack of precision of the estimated coefficientsin the heat exchange equation as well as to the possible inaccuracy of the different measurements. The two regression equations do not differ significantly. The rates of sweat loss observed ranged from 100 g h-' to 1400 g h-', and the total sweat loss by individuak was between 900 g and 3400 g for a complete shift. /
lOWr
q I
L
y=11079x
1133) 1002 r=O88L
+
n.58 I 0
I
I
m
I
I
LOO
I
I
I
Eo3
I
800
I
1
1wo
Required Sweat Rate lgh-')
Fig. 6.2. Observed sweat rate plotted against Required Sweat Rate. Required Sweat Rates are mean values computed for each work period
Sweut rate and thermal sweat
y : l12L2x
n.59 1
0
I
?a,
I
,
601 ? 9 3 2 r.0892
,
I
MD
LCC
Required Sweat
+
,
105
1
BW
I
I
1OW
Rate (ch-’)
Fig. 6.3. Observed sweat rate plotted against Required Sweat Rate. Required Sweat Rates computed from mean values of metabolic heat production, clothing and environmental conditions for the whole duration of the working shift
DISCUSSION
This study shows that the two procedures give almost identical values of Required Sweat Rates (see figs. 6.2 and 6.3). However, the Required Sweat Rate assumes steady sweating only. Our study shows that, despite this simplification, the actual sweat rate can be predicted with sufficient accuracy, at least in the observed working conditions. S,,, underestimated the sweat rate actually observed by about 100 g h-‘ (figs. 6.2 and 6.3). For comparison the P4SR index is plotted in fig. 6.4. For all conditions P4SR underestimates observed sweat rate by approximately 100 g h-’ . The two indices (Required Sweat Rate and P4SR) are significantly correlated. When both are expressed in g h-’, for nude men, (P4SR) = (1.013 S,,-40) with an error of f 6 8 g h-’ (r = 0.95 and n = 58). For clothed men (P4SR) = (1.120 Sh,-50) with an error of &SO g h-’ (r = 0.94 and n = 58). The increased barometric pressure in mines explains this underestimation. Figure 6.5 shows the relationship between the observed sweat rate and the index currently used in French mines, the “Temperature RCsultante Minitre”. In this case correlation is poor. As the observed sweat rate is representative of physiological strain imposed by working conditions, we can conclude that Required Sweat Rate is a good index.
J. J. VOGTet al.
106
f
.
y = (0905 x
n.58
OL
, 0
3
I
I
I
I
LW
200
predicted Four Hour
151 89) t 8L.S
+
r.0921
,
.
I
I
1
1
I
1WG
Bm
6M)
Sweat Rate i g h - ’ )
Fig. 6.4. Observed sweat rate plotted against Predicted Four Hour Sweat Rate (PIISR) values
g
c
@lJ-
B ml0
200 -
,il 20
y = i38L6 x
-
n.70
,
56LL)
1553
I
1
15
?
r -0689
30
Temphture R k l t a n t e Minibre
35 (OC!
Fig. 6.5. Observed sweat rate plotted against “Tempkrature R6sultante M mi6xe”
”
Sweat rate and therniol strain
107
REQUIRED SWEAT RATE AS AN INDEX OF HEAT STRAlN IN INDUSTRY
As shown above, the Required Sweat Rate correlates very well with observed sweat rates. Can it be usedforassessing heat strain in industry? In order to answer to this question the different conditions which such an index has to satisfy must be examined. These are: that the index shall take into account the six influencing factors : metabolic heat expenditure, clothing, air and mean radiant temperature, humidity and air velocity; that the index shall allow an evaluation of the degree of discomfort as well as of safe exposure times; and that the measurements and calculations required shall be simple and rapid. How far does Required Sweat Rate satisfy these conditions? Being derived from the heat balance equation, S, is directly related to the physiological strain; the larger the Required Sweat Rate, the larger the physiological strain. The Required Sweat Rate corresponds to the sweat rate which must be realized in order to match the heat stress. Its value depends very little on heat acclimatization, which increases only the maximal sweating capacity making tolerable an otherwise intolerable heat stress. The ratio of Required Sweat Rate to maximal possible sweat rate indicates the strain in a given situation. Estimations of discomfort and of tolerance times require different limits of sweat rate. FANGER, (1970) proposed a comfort value of sweat rate depending on metabolic energy expenditure according to the empirical equation So = (0.6 Mn-35), where So is the optimal sweat rate (g m-' h '). If Required Sweat Rate lies between zero and So, we can consider that man is in his comfort range. Upper tolerable values can also be proposed for two cases: for maximal sweat rate and for maximal sweat loss. The maximal sweat rate depends on the acclimatization of the individual. Non-acclimatized people cannot sweat more than 600 to 700 g h-l, while acclimatized people may sweat much more. Values near 2500 g h-' have been observed, but for industrial applications we consider that 1000 to 1300 g h-' is the likely maximum, Conditions in which Required Sweat Rate is lower than this threshold are tolerable. Figure 6.6 shows on the psychrometric chart maximal conditions for Required Sweat Rate in a range between 1000 and 1300 g h-'. These values can be compared with limits proposed by MULLERand WENZEL (1961) for young acclimatized men. For Required Sweat Rates from 400 to 700 g h-' comparison can be made with limits proposed in the U S A . by National Institute of Occupational Safety and Health (KAMON,1975). The upper tolerable limits of total sweat loss must also be set. Dehydration will occur when the conditions are such that drinking during the exposure cannot compensate for the amount of water lost (LEITHEAD and LIND, 1964). We consider that a sweat loss of 4 to 5 litres is the tolerable maximum for a workshift. In 1966 the World Health Organization proposed a total sweat loss of 4 litres per shift as an upper limit of sweat loss, even if the sweat rate itself is tolerable. From this limit we can calculate a Recommended Working Time (RWT), in hours, as (RWT)
= 4000/S,,,.
(6.7)
J. J. Vocr et al.
108
Ambient wpour pressure [(POI
-.-.
----.........
(OCI
(ms-’)
or ~ c l o )
(wI
37
09 01-09
0 0
3L
05
-
<1 6
005iCl3) 0051031
232 232 232 232
-
l@X
Fig. 6.6. Required Sweat Rate, sj,,, plotted on psychrometricaxes. These lines can be compared to the upper tolerable values proposed by MULLERand WENZEL(1961) for young, healthy, acclimatized men; and to the Wet Bulb Globe Temperature index (WBGT) for general industrial use (National Institute for Occupational Safety and Health; KAMON,1975)
RWT is applicable to just tolerable conditions only, as defined by LEITHEAD and LIND. Intolerable conditions introduce the problem of setting safe exposure times. Conditions are intolerable if the required evaporation (E,) exceeds maximal evaporative capacity (Em).In these conditions heat balance cannot be achieved, and the rate of heat storage is equal to the difference (Er-Em). The safe exposure time corresponds to the time taken to accumulate the maximum acceptable heat storage. PANDOLF (1977) suggested a maximum heat storage of 0.34 MJ. This is the lowest value proposed in the literature, but for safety in industrial applications we must use the lowest criterion. CONCLUSIONS
In conclusion, analytical indices such as the Required Sweat Rate should be preferred to the others available. Here the heat stress evaluation has been improved by applying knowledge gained in the laboratory. ‘The index may be used to predict discomfort as well as safe exposure times, which allow complete evaluation of the thermal conditions during work. The calculation of Required Sweat Rate is complex. but either diagrams and nomograms (fig. 6.7) or a programmable calculator may be used for its rapid estimation.
109
Sweat rate and therrnal strain Ambient
vapour
pressure
(kPa1
v = 2.0 m s-1
v = 0.5ms-' 60
50
3 0
9
LO
c
1
30
20
a a 60
E
L
50
5 0
LO c
I: 30
L
0 I
5
20
.1
A
n E
60
Q
2m
50
N I!
LO
30
20
---
-.- 700 gh-' Fig. 6.7. Required Sweat Rates (400 - - -, 700 -. and lo00 Required Sweat
Rate :
- LOO gh-'
lOOOgh-'.
-* g h-l) plotted on psychrometric axes for three levels of metabolic heat production (150.200 and 250 watts) and for two levels of air velocity (v = 5 and 2.0ms-l). w is the skin wettedness. Conditions: clothing insulation 0.05 m2 K W-' (0.3 clo), mean skin temperature 36 "C ~
REFERENCES
BELDMO H. S. and HATCH T. F. (1955), Index for evaluating heat stress in terms of resulting physiological strains, Heat Piping Air Cond. 27, 129-135. CANDAS V., LIBERTJ. P., HOEFT A., and VOGTJ. J. (1978), The required wettedness and the sweat rate, [In:] New Trendrin ThermalPhysiology, eds.: Y . HOUDASand J. D. GUIEU, Masson, Paris. FANOER P. O., (1970), Thermal Comfort, Danish Technical Press, Copenhagen, 224 pp. GACGE A. P. (1937), A new physiological variable associated with sensible and insensibleperspiration, Am. J. Physiol. 120, 277-287. GACGE A. P. and NISHI Y . (1976), Physicalindices of the thermalenvironment, ASHRAE J. 18,47-51. GIVONIB. (1963), Estimation of the effect of climate on man. Development of a new thermal index, Building Research Station Report, Technion, Haifa. Israel, 145 pp.
110
J. J. VOGTet al.
HERTIGB. A. and BELDINGH. S. (1963), Evaluation of health hazards, [In:] Temperature, Its Measurement and Control in Science and Industry, Vol. 3, ed.: J. D. HARDY,Reinhold, New York, pp. 347-355. KAMON E. (1975), Ergonomics of heat and cold, Texas Reports on Biology and Medicine 33, 146-182. KERSLAKE D. McK. (1963), Errors arising from the use of mean heat exclrange coejficients in the calculation of the heat exchange oj'a cylindrical body in a transverse wind, [In:] Temperature, Its Measurement and Control in Science and Iiidicstry, Vol. 3, ed.: J. D. HARDY,Reinhold, New York, pp. 183-190. URSLAKE D. McK. (1972), The Stress of Hot Environments, Cambridge University Press, 315 pp. LEITHEAD C . S. and LINDA. R. (1964), Heat Stress and Heat Disorders, Cassel, London, 303 pp. LIBERTJ. P., V ~ G J. T J., CANDAS V., and HOEFT A. (1975), Influence des conditions hygrothermiques anibiantes sur le rendement e'vuporatolre de la sudation thermique, J. Physiol. (Paris) 70, 717-735. MCARDLEB., DUNHAM W., HOLLONG H. E., LADELLW. S. S., Sco-n J. W., THOMSON M. L., and WEINERJ. S. (1947), The prediction of the physiological effects of warm und hot environments, Medical Research Council, Report No. 47/391, H. M. Stationnary Office, London. MISSENARD F. A. (1973), Coefficients d'ichange de chaleur du corps humainpar convection, en fonction de la position, de I'activite' du sujet et de I'environment,Arch. Sci. Physiol. 27, A45-A50. MOLLERE. A. and WENZELH. G. (1961), Die Beurteiliing des Arbeitsklimas, [In:] Handbuch der gesamten Arbeitsmedizin, ed.: G. LEHMANN, Band l, Urban und Schwarzenberg, Berlin. National Institute for Safety and Health (I 972), Occupational exposure to hot environments.Washington, U.S. Dept. of Health, Education and Welfare, 101 pp. NISHI Y. and GAGGEA. P. (1969), Moisture permeation of clothing - a jactor governing thermal eqiiilibrium and comfort, ASHRAE Trans. 76, 137-145. PANDOLF K. B. (1977), The influence ofhot environments on human performances of muscular exercise, [In:] Proc. o f Int. Union of Physiol. Sciences, p. 743; Pitit-Salp&itre, Paris, Vol. XII, 813 pp. World Health Organization (1966), Problemes de santC associh au travail dans des conditions de contrainte thermique. Rapport technique no 412, Geneve, World Health Organization, 54 pp. YAGLOU C. P. (1927), Temperature, humidity and air movement in industries: The effective temperature index., J . ind. Hyg. 9, 297-309. YAGLOU C. P. and MINARDD. (1957), Control of heat casualties at military training centers, Am. Med. Ass. Archs. ind. Hlth. 16, 302-316.
Chapter 7
MODELLING OF HEAT TRANSFER IN MAN Y. HOUDAS Department of Thermoregulation, Faculty of Medicine, University of Lille. Place de Vcrdun, 59045 Lille, France.
CONTENTS Introduction Preliminary observations Fundamental equations Conduction Convection External exchanges Models Partial or sub-system models Whole body models One-cylinder models Multi-layer models Multielement models Conclusions
INTRODUCTION
In a recent International Congress of Physiological Sciences, HARDY(1970) reviewed “Models of Temperature Regulation”. He pointed out the considerable increase in work on this subject and emphasiz edtheinter est and usefulness of modelling techniques for a better understanding of the biological regulatory processes. The temperature regulation field is perhaps one of the best applications of these modelling techniques because the thermal function involves physical phenomena, and, moreover, is strictly related to the physical parameters of the environment. The word “model” is used today in various senses. For instance, an anatomical scheme is sometimes presented as a model. In other circumstances, if an experiment is conducted on a dog, the dog is also considered as a model. It is thus first necessary to define accurately the sense in which the word “model” is used in this paper. A model is an analogous representation of biological phenomena and of their relationships. The most powerful mode of expression of a model is the mathematical one. However, mathematics can be used as a unique means of expression. It can be also used t o
112
Y. HOUDAS
describe the relationships between physical elements. Thus, the border between purely mathematical and physically realistic mathematical models is often uncertain. A complete model of thermal regulation would involve the representation of two kinds of processes: a) the heat tranefers within the body, and between the body and its environment (the so-called “passive or controlled system”) ; b) the mechanisms by which the body can control its exchanges to maintain its temperature within close limits (the ‘‘active or controlling system”). However, the models cannot generally be complete. They concern one of these, aspects rather than the two problems together, because the physical and the mathematical procedures, and perhaps also the aims of work on these two kinds of problems, are different. Such a difference has been emphasized by GRODINS (1970). He distinguished “descriptive models” and “explanatory models”. Schematically, the first correspond to the models of thermal exchanges (or “thermokinetic” models). They can be very complex because of the great number of variables and parameters involved. However, if they are as complete as possible, they can be used in simulation and design for physiologists and practical engineers (called “engineering” models by HWANGand KONZ,1977). Conversely, the aim of the true regulatory models is primarily to explain the “physical” structure of the thermal reguIator. They do not necessarily use the same mathematical language. For instance, a special mathematical procedure, involving the so-called ”transformation“ is generally used for describing these models (see HOUDAS et al., 1978). As they are used especially to support a general concept of the thermal regulation, they can be simpler than the descriptive models, but they cannot be used in thermo-physiologicalengineering. From this point of view, there is a certain confusion in the models presented in thermal physiology. Some of those described as models of thermal “regulation” are thermokinetic models and not truly regulatory models. Of course, a thermokinetic model must be built on a certain assumption concerning the regulatory mechanisms. However, this assumption can be very simple. It is an error to use this kind of model for supporting a hypothesis on the nature of the thermal regulation. For instance, it is possible to describe the sweating rate as related to the difference in core temperature between its initial value T and its value T+ AT at a later time. This relationship results only from the choice of the mathematical representation. It does not allow us to assume that the regulation of the sweating rate implies the existence of a set-point, which is T, and that the sweating rate is determined by the difference AT. It is physically obvious that it is the difference AT which is determined by the sweating rate. The regulatory models, because they have little practical interest, are small in number. They have been reviewed recently (HOUDAS,1977; HOUDM et al., 1978). For this reason, this paper is specifically devoted to the thermokinetic models. PRELIMINARY OBSERVATIONS
Temperature is a very special variable. For instance, the mass of a given body is a physical property of the object and is invariable. Numerous physical meamres are, as is the mass, specificto the body. On the contrary, temperature is not a specific
Modelling of'heat tramfir
113
property of the body. It results only from the thermal equilibrium between a body and its environment. This fundamental idea that temperature is not a specific property of the body must lead the reasoning of the physiologists. From this point of view, the value of the internal temperature of homeothermic animals must not be considered as a dogma. It results from the heat produced and exchanged by the body. A homeothermic organism exhibits a very important thermal inertia. For this reason, when a thermal disturbance occurs, a new steady state is attained only with great delay. Consequently, during transients, the relationships between the different variables involved in the thermal exchanges vary with the time. Almost all models describing the thermal exchanges are designed for 'steady state. It is important to keep in mind this fundamental aspect of these models. As the heat exchanges with the environment occur principally at the skin surface in man, it is obvious that the thermal steady state of the body implies that the mean skin temperature must be lower than the central temperature. If not, the heat produced by the metabolism cannot be transferred to the periphery. FUNDAMENTAL EQUATIONS
All thermokinetic models use the same general equations to describe the heat transfer. Heat is tiansferred from core to shell and hence to the skin surface by conduction and the blood convection. CONDUCTION
The fundamental law of heat conduction was developed by FOURIER. Let us consider a small cube inside a given body. This cube produces heat at a rate Q (W me3), and exchanges heat with its surroundings. The general equation below expresses the fact that, at any time, the sum of the heat transferred through the three sides of the cube and the heat produced by it, is equal to the rate of temperature variation 6T/Bt, at any point;
d2T
d2T
B2T
Q
cp 6T
+--6y2 + -+-=---6x2 6z2 k k
__
Bt
(7.1)
where k is the thermal conductivity, c is the specific heat of the body and e is the density of the body in consistent units. CONVECTION
The rate of heat transferred by the blood convection of Qb is generally represented as follows:
where vb is rate of blood flow per unit volume of tissue, c, is specific heat of the blood and eb is density of the blood, Tveand T,, are the temperatures of the venous blood and the arterial blood, respectively. 8
- Bioengineering
114
Y. HOUDAS EXTERNAL EXCHANGES
Heat is transferred from the skin to the environment by conduction K , convection C, radiation R,and evaporation E. The equations describing these transfers are well known and have been often presented (MITCHELL et al., 1969; RAPP, 1973; HOUDAS and GUIEU,1977). The thermal balance of the body is then represented by the equation M+K+C+R+E+J=
0
(7.3)
in which M is the rate of metabolicaheatproduction and J the rate of heat storage (J = 0 in steady state), all terms being usually expressed in units of watts per square metre of body surface. Equation (7.3) when written by physiologists, often includes the mechanical work W, while M represents the total energy expenditure. This procedure depends on the methods used for determining these variables. In a thermal study jt seems better to introduce only the thermal factors. A thermokinetic model can be partial or complete according to whether it describes the thermal exchanges occurring in one part of the body or in the entire body. MODELS PARTIAL OR SUB-SYSTEM MODELS
These models describe the heat exchanges in a part of the body, for instance between the different tissue layers of a limb. However they are generally added together and included in a complete representation of the heat exchanges of the whole body. For this reason, they will be studied below. There is, however, one exception: the counter-current exchanges occurring in the limbs. When a vein, leading the blood back from the extremity of the limb, is in close contact with the corresponding artery, which leads the blood toward the extremity, heat exchange occurs by conduction between these two vessels. The warm arterial blood gives heat to the cool venous blood, so that in the artery the blood is progressively cooled, while that in the vein is progressively warmed. There is then a partial short-circuit for heat flow in the extremity. This mechanism occurs in the arms and legs of the human body but, except in cold conditions, this thermal exchange appears to be of little importance. A model and WEVER of this counter-current heat exchange has been developed by ASCHOFF (1958). Although this model was simple and probably not quantitative enough, it demonstrated the possible importance of this exchange. It is also interesting because it was one of the first attempts to apply modelling in the thermo-physiological field. WHOLE-BODY MODELS
The first attempt to represent the body from a thermokinetic point of view was made in 1911 by LEFEVRE in his book Chaleur animale et Bioinergktique. He considered the body as a sphere with a core producing heat and a shell losing heat to the environment. This simple model has survived as the first step for a general view of the thermo-physiological problems.
ModeIIing of heat transfer
115
ONE-CYLINDER MODELS
The first of these models was proposed by BURTON(1934). He assumed the body to be a cylinder with uniform properties throughout. He studied the temperature distribution within the model assuming the body was first immersed in water at 32.5 "C and then the water was rapidly heated to 36 "C. To simulate cutaneous vasodilat.ltion, BURTONassumed that the apparent thermal conductivity of the tissues was increased four times by heating. Although very simple this model showed interesting results, especially that some hours were required before complete thermal equilibrium. The importance of the thermal inertia of the body was then demonstrated. A new one-cylinder model was presented about 30-40 years later by GAGGB et al. (1971). The human body is considered to be a single cylinder with a central core and an outer layer which is the skin. Heat M is generated inside the core and transferred to the skin both by the blood (convection), and tissue conduction. From the skin, heat is transferred to the surroundings by the classical ways (see above). The mathematical development is based on equation (7.3). The main steps in this development are : 1 . The determination of the respective roles of core and shell in the total heat storage, assuming that the initial temperatures are 34.1 "C for the skin and 36.6 "C for the deep part of the body. 2. The computation of the relationships between the skin and deep body temperatures and, on the one hand, the skin blood flow and, on the other hand, the rate of sweat production. This model was able to predict the probable values of the main variables related to the judgement of comfort, and allowed establishment of an "Effective Temperature Scale" to be used by the engineers. MULTI-LAYER MODELS
In their model (fig. 7.1), CROSBIE et al. (1963) assumed the heat flow from the core to the skin to be unidirectional. This one-dimensional model is divided into three layers: the core (viscera, skeleton, etc.) which is the source of basal metabolism; the muscle layer which is the source of increased metabolism caused by exercising
Fig. 7.1. Three-layer model of CROSBIE,HARDY and FESSENDEN (1963) T - skin temperature, Tm - musclo temperature, T i - deep body temperature. Layer thickness is given in crn
layer -source ot increased rnetaboli srn caused by shivering or exercising
Y. HOUDAS
116
or shivering; and the skin. Assuming that the regulated temperature is the average body temperature and that the temperature regulation involves three basic types of contra1 mode (proportional control, rzte control, on-off control), the simulator is able to predict steady state values of skin and deep body temperature, metabolic rate, and evaporative heat loss. If the time constants for the various thermal changes are introduced, the simulator can also predict the dynamic responses to sudden shifts in environmental temperature. In the ATKINS-WYNDHAM model (ATKINSand WYNDHAM, 1969), the body is represented by a cylinder containing four concentric layers (fig. 7.2): the core. the muscles, the deep part of the skin, and the superficinl, bloodless layer of the skin,
,-Outer
/
skln
/-Skin and tissue
Heart,
lungs
Fig. 7.2. Four-layer model of ATKINSand WYNDHAM(1969)
- deed
T - air temperature, Ts- skin temperature, Tb body temperature
As it is assumed there is no axial heat flux in the cylinder, this model is essentially one-dimensional and can be compared with the slab model of CROSBIE et al., (1963). The basic equation of uni-dimensional heat flow, which derives by simplification from equation (7.1), is solved by the finite difference technique. The cylinder is divided into a number of finite steps, a differential equation being allocated to each step. The heat transferred by the blood is determined from equation (7.3), while the heat exchanged with the environment is computed by equation (7.2). Finally, the control of sweat rate is assumed to be caused by signals issuing from the skin and the hypothalamus. However, the equation proposed by the authors uses the rectal and skin temperatures of 36.5 and 33 "C,respectively, as base values. In the model developed by BROWN(1963), the body is represented by four concentric layers : central core, somatic muscles, subcutaneous tissues, and epidermis. This model was used in an analogue computer. The particular interest of this model is the inclusion of clothing and its use also for simulation of the responses to water immersion.
Modelling of heat transfer
117
The linear model recently developed by CENAand CLARK (1978) also involves the clothing. It appears to be useful in application to natural microclimates of strong solar irradiance, and to some industrial environments. MULTI-ELEMENT MODELS
The aim of the model developed by WISSLER(1963) (fig. 7.3) was to interpret temperature profiles in the nude man. He introduced the following factors: local heat production ; heat conduction between tissue layers, heat transfer by circulating blood thermal insulation due to fat layers; counter-current heat exchange between adjacent arteries and veins; the heat loss by the respiratory tract ; sweating; heat generated by shivering, heat storage ; the environmental factors involved in the thermal exchanges; and tbe geometry of the body. The body is represented by six elements: the trunk, the head, the two arms and the two legs. Each element is assumed to be a homogeneous cylinder of bone and tissue, covered by a layer of fat and skin. The six elements are linked together by a central blood compartment and longitudinal heat transfer is neglected.
Fig. 7.3. Sixelement model of WISSLER(1963). Each element is assumed to be a homogeneous cylinder. All elements are linked together by a single central countercurrent heat exchanger, marked HE, in which arteries and veins meet together
Applying the classical heat transfer equations to each element this model allowed the steady state temperature profiles in each segment for a resting man to be determined. The same basic model has been used by STOLWIJK (1970). Each of the six elements is divided in four concentric layers as in the ATKINS-WYNDHAM model. Thus, with the central blood compartment, there are a total of 25 compartments. The input for the controller is represented by the signals coming from these 25 compartments. For each compartment, the distribution of different tissue types, their volume, heat capacitance, basal metabolic rate, and thermal conductance are determined. The heat exchanged by each element with the environment is calculated from the respective combined heat transfer coefficients. The controlling system is divided in three hypothetical parts: the thermoreceptors, the integrative system, and the effectors. The general scheme represents the thermoreceptor output for each of the 25 compartments by a general equation (7.4)
118
Y. HOUDAS
which states that the thermoreceptor error output E, from compartment n is equal to the difference between the instantaneous temperature T, and the threshold temperature Tsn to which is added the product r, of the dynamic sensitivity factor of the receptors in compartment n , and the rate of temperature change (dT/dt,J.Then, the value of E, is tested for each compartment and is classified warm if positive, cold if negative, etc. The total receptor output from the skin is obtained by summing values from the six skin compartments. The effector mechanisms are sweating, vasodilatation, vasoconstriction and shivering. Each of these mechanisms can be represented by an equation, for instance for sweating. This states that the sweating rate depends on the sum of: the warm output of the head multiplied by the coefficient for sweating command from head core; the warm output of the skin multiplied by the coefficient for sweating command from skin; and a combined core and skin output multiplied by the appropriate coefficient for sweating command. It is not possible in this paper to describe the entire program which has been developed by STOLWIJK (1970) and used, for instance, by the U.S. Space Agency in simulation of space environments. GORDON et al., (1977) have developed a model especially to simulate and predict the physiological responses of the human to cold exposure. It uses much new physiological data. The body is idealized as 14 cylindrical and spherical angular segments, as shown in fig. 7.4. Each segment consists of several concentric tissue layers, for
SECTION A - A
Fig. 7.4. Model of GORDON, ROE= and HORVATK (1977) consisting of 10 cylinders (c), 2 cylindrical segments (cs) and 2 spherical segments (ss)
instance 5 for the abdomen (core, bone, muscle, fat, skin), but only 4 for the arms and feet (bone, muscle, fat, skin). Each tissue layer is subdivided into annular shells and a finite-difference formulation is written for the central modal point of each such shell.
Modelling of heat transfer
119
The control system of this model is characterized by the input signals. These signals are based on the variation in the head core temperature, the skin temperature and also of the heat flux through the skin over the whole body. The role of this last factor has sometimes been challenged. However, it agrees with some assumptions on the thermal regulation which suppose that the constailcy of temperature is the result of a heat rate regulation rather than a true regulation (HOUDAS et al., 1978; WEBBet al., 1978). The initial values are those of thermal neutrality. However, some signals are not continuously used. For instance, the skin temperature error is not used until the mean skin temperature is below 28.5 "C.This model has been used to simulate the results of the experiments of RAVEN and HORVATH (1970) who exposed subjects to a step change in environmental temperature from neutrality to 5 "C. The general agreement between the model and the experimental values is quite good.
CONCLUSIONS
The models presented here are only a small part of the great number which have been proposed. However, they constitute the elementary basis of the models which are now used by physiologists and engineers to predict the physiological behaviour of a subject in a given environment. The advantages and disadvantages of each model depend on the experimental conditions which have led to their formulation. This perhaps determines their limits. Because the thermal factors involved in the heat exchange of a man in his own work environment are not all determinable, these models are valuable only when the thermal conditions of the man and his environment are both known or may be controlled,
REFERENCES
ASCHOFF J. and WEVER R. (1958), Ken1 und Schale im Wirmehaushalt des Menschen, Natunvimnschaften AS, 477485. A. R., WYNDHAM C. H. (1969), A study 01temperature regulation in the human body with ATKINS the aid of an analogue coniputer, Pflugers Arch. 37, 104-119. BROWN A. C. (1963), Analog computer simulation of temperature regulation in man, Technical Report AMRL-TDR-63-116, Wright-Patterson A. F.B. A. C. (1935), The average temperature of the tissues of the body, J. Nutr. 9,264-280. BURTON CENAK. and C L A RJ,~A. (1978), A linear model of heat transfer through human clothing, [In:]New Trends in Thermal Physiology, eds.: Y . HOUDAS and J. D. GU~EU, Masson, Paris. CROSBIE R. J., HARDYJ. D. and FESSENDEN E. (1963), Electrical analog sirnulalion of temperature regirlation in man, [In :] Temperature : Its Measurement and Control in Science and Industry, ed. : J. D. HARDY,Part 111, Rheinhold, New York, pp. 627-635. J. A. J. and NISHIY. (1971), An efective iemperature scale based on a siwle GAGGE A. P., STOLWIJK model of himan physiological regulatory response, ASHRAE Trans. 77, 247-262. R. B. and HORVATH S. M. (1977), A mathematical model of' the human GORDON R. G., ROEMER temperature regulatory system, Biomed. Eng. 23, 434-444. F. S. (1970), Theories ond models in regulatory biology, [In :I Physiological and Beliavioural GRODINS
120
Y. HOUDAS
Temperature Regulation, eds.: J. D. HARDY,A. P. GAGGE and J. A. J. STOLWIJK, C. C. Thomas, Springfield, pp. 722-726. HARDY J. S. (1972), Models oJ- temperature regulation, [In :] Essays on Temperature Regulation, eds. : J. BLIGHand R. MOORE,North-Holland, Amsterdam, pp. 163-186. HOUDASY. (1977), Developnient of a new concept on temperature regulation in man, Acta physiol. Polon. 28. suppl. 1, 19 p. HOUDAS Y.and GLJIEU J. D. (1977), La fonction thermique, SIMEP-Editions, Lyon. HOUDAS Y., LECROART J. L., LEDRUC.. CAREITE G., and GUIEUJ. D. (1978), The thermoregulatory mechanisms considered as a follow-up system, [In:] New Trencls in Thermal Physiology, eds.: Y. HOUDAS and J. D. GUIEU,Masson, Paris. HWANGC. L. and KONZS. A. (1977), Engineering models of the human thermoregulatory system. A review, Biomed. Eng. 24, 309-325. LEFEVRE J. (1911), Energitique et chaleur animale, Masson, Paris. MITCHELL D., WYNDHAM C. H., VERMEULENA. J. HOGDSONT., ATKINSA. R., and HOPMEYR H. S. (1969), Radiant and convective heat transfer of nude men in dry air, J. Appl. Physiol. 26, 111-1 18. RAPPG. M. (1973), Convective heat transfer and convective coefjcients of nude man, cylinders and spheres at low air velocities, ASHRAE Trans. 79, 75-87. RAVENP. R. and HORVATH S. M. (1970), Variability of physiological parameters of unacclimatized males during a two-hour cold stress at 5 "C,Int. J. Biorneteorol. 14, 309-320. STOLWIJK J. A. J . (1 970), Mathematical model of tliermoregulation[In :I Physiologicaland Behavioural Temperature Regulation, eds.: J . D. HARDY,A. P. GAGGE, and J. A. J. STOLWIJK, C. C. Thomas, Springfield, pp. 703-721. WEBBP., ANNISJ. F. and TROUTMAN S. J. (1978), Heat Jlow regulation, [In:] New Trends in Thermal and J. D. GUIEU,Masson, Paris. Physiology, cds.: Y.HOLJDAS WISSLERE. H. (1963), An analysis of factors affecting temperature levels in the nude human, [In:] Temperature, Irs Measurement and Control in Science and Industry, cd.: J. D . HARDY,Part 111, Rheinhold, New York, pp. 603-612.
Iii. PHYSIOLOGY, WORK AND EXERCISE
This Page Intentionally Left Blank
Chapter 8
EXERCISE PHYSIOLOGY AND SENSORY RESPONSES R. R. GONZALEZ John B. Pierce Foundation and Department of Epidemiology and Public Health, Yale University, New Haven, Connecticut 06591. U S A .
CONTENTS Introduction Comparison of physiological systems Circulatory and respiratory response Thennoregulatory changes Thermal sensation in exercise Thermal discomfort in exercise Thermal transients Heat acclimation Effective temperature Sense of effort in exercise Conclusions
INTRODUCTION
Altering a prior resting state by physical exertion generally involves the coordination of multiple systems in the body. These comprise adjustments of muscular, respiratory, circulatory and thermoregulatory responses, which are more pronounced the more intense the muscular effort. More subtle responses comprise the chemical and autonomic changes involved in exercise, which encompass the anaerobic and oxidative processes which supply energy to the exercising muscle. Less clearly understood are the responses to exercise of the central nervous and endocrine systems mediating behavioural processes. These are a direct consequence of exercise and play a part in its control, through the thermal sensations of comfort and the perception of effort. In this review we shall direct attention to some recent work on the physiological systems and their interactions with sensory responses during positive continuous exercise. Sensation will be defined as a central nervous response of the body as a direct result of stimulation of a sense organ. Perception involves the combination (either cortical or subcortical) of different sensations and past experience.
124
R. R. GONZALEZ COMPARISON OF PHYSIOLOGICAL SYSTEMS
During the first few minutes of exercise energy is produced by anaerobic biochemical processes which are minimal sources in steady light to moderate exercise, but become the major supply in very heavy exercise. Oxygen debt often carries over after the activity (WASSERMAN et a]., 1967). HARRISet al. (1977) have shown a general pattern, regardless of type, duration and intensity of exercise. Their data indicate that high energy phosphate (ATP/ADP) decreases and the increase of lactate acid are coupled. They postulate that the hydrogen ion concentration plays an important part in the regulation of anaerobic energy utilization. A successful attempt to model another controlling variable has been made by MIfcHELL et al. (1972). Changes in oxygen tension in muscle from a set value of 3.3 kPa serve as an error signal. As this increases there are reductions in the resistance of vascular beds and increases in muscle blood flow. CIRCULATORY AND RESPlRATORY RESPONSE
After the initial transient, which may last no more than 90 s (MITCHELL et al., 1972), a steady-state ensues during continuous exercise. Oxygen delivery to the tissues from the atmosphere becomes the primary source of oxidation for energy production and is linked with circulatory and ventilatory responses. It is clear that whatever the metabolic demand of an active muscle, the circulation must meet it by an appropriate cardiac output (EKELUND and HOLMGREN, 1967). Oxygen extraction from the blood, estimated from arteriovenous oxygen concentration differences, increases rapidly from rest (- 4 to 5 cm3 per 100 cm3 blood, or vol. %) to become 8-10 vol. % during submaximal exercise and as high as 16 vol. yoin maximal exercise. Increases in oxygen delivery (above resting) occur as a result of changes in heart rate, stroke volume, and oxygen extraction. Stroke volume is usually almost constant (RUSHMER, 1965; ROWELL,1974), so increases in cardiac output result mainly from increases in heart rate. Table 8.1 shows the partition of the cardiac output to different body areas during rest and at a typical level of continuous exercise (about 2.2 1 min-', i.e. 60% of maximum aerobic capacity). Also shown are the possible neural controls, via the autonomic nervous system and primary local chemical controls. Extensive research by ROWELL (1974) and MASON(1968) indicates that despite increased baroreceptor traffic (which, at rest decreases sympathetic activity and increases vagal discharge to slow the heart by reflex) during continuous exercise, vasoconstrictor nerves in muscle vessels become active in proportion to the intensity and duration of exercise. In the active muscles, locally produced metabolites cause vasodilation (MITCHELL et al., 1972; HARRIS, 1977). In muscles not involved in the exercise increased vascular tone diverts a proportion of the cardiac output to working muscles (70-90%) or to skin circulation. When exercise is severe, O2 is also completely extracted in veins draining nonworking muscles; vasoconstriction at these sites slows blood flow to minimal rates and 0, extraction may be as complete as in active sites (RUSHMER, 1965). Skin blood flow is high during exercise as it facilitates heat dissipation. All
Exercise physiology
125
these responses have also been demonstrated during heat stress in moderate or maximum upright exercise. ROWELL (1974) found that when core and skin temperatures were raised both cardiac output and stroke volume decreased. In other studies NADEL et al. (1977) have recently shown that with moderate exercise (- 60 % ko2 max) increases in core temperature are accompanied by a decrease in plasma volume, though skin temperature, cardiac output, and stroke volume are constant. It is difficult to distinguish the mutual interactions of particular systems. During exercise, cardiac output depends on blood volume, the vascular system, the sympathetic nervous activity and posture. However, the product of cardiac output and total peripheral flow resistance decides the mean arterial pressure. Table 8.1 also shows the effect of sympathetic nervous activity on total peripheral resistance. Other important variables to consider during exercise are the ventilatory equiv. . . . alents for O2 (VE/Vo2)and for COz (VE/Vco2).Ventilatory output, cardiac output and metabolism must safeguard against acidosis; but alkalosis must also be prevented, so that ventilation must not be increased out of proportion to metabolic demand (WASSERMAN et al., 1967; HARRIS et al., 1977). For measured body core temperatures (TJ between 36 and 37.8 "C, the ventilatory equivalents for O2 and C 0 2 remain constant (PETERSEN and VEJBY-CHRISTIANSEN, 1973). However, they found that for T, > 38 "C, hE/Vo2 and VE/kco2may increase without acidosis. They suggest either that increases in body temperature above this threshold serve as an additional drive to increase ventilation or that there are other, perhaps nonthermal stimuli. THERMOREGULATORY CHANGES
Since only 20 % of the energy is utilized as work, 80 % of the energy must be dissipated as heat. The internal thermal load, occurring as a result of physical exertion, therefore constitutes an important physiological challenge to the thermoregulatory system. Rates of heat production can be as high as 20 times the basal level (1 160 W rnvZ or 20 met). For continuous exercise internal body temperature rises by about 0.14 "C for each multiple above the resting state. In order to maintain thermal equilibrium, autonomic responses activating heat production and heat dissipation are called forth when internal and skin temperatures are changed from one steady-state level to another. In exercise the net metabolic heat load (heat production, M , less external work done, W )must be dissipated to the environment or stored by the body. Heat storage most often occurs when the thermoregulatory response mechanisms are inefficient, as in untrained or unacclimated individuals. For an average male, about 335 kJ of heat storage (equivalent to about a 1.4 "C rise in T,) is a typical tolerance limit (GAGGE et al., 1971). Heat storage also occurs during the initial transient of exercise. During the first 20 minutes the rate of metabolism may greatly exceed the heat dissipation required to maintain homeostasis. Heat balance is re-established mainly by matching the' dissipation of heat by evaporation ( E ) with the heat load. Figure 8.1 shows the sweating response to moderate exercise in a wide environmental range, from To= 26 "C to 50 "C and 29% to 90% relative humidity (GONZALEZ et al., 1978).
I26
R. R. GONZALEZ
Typical function and control of regions of the body
Region
Function
Percent of cardiac Oxygen uptake of Primary output (%) to region region (ml min -I) control of function Rest Exercise 60% yoz Rest ma. (max) ~
Skin
heat exchange
central
metabolic
Visceral organs Kidney
9
2
5
190
local, some neurogenic
20
86
97
1470
absorption
central
25
1
55
134
central, local
20
1
14
67
Brain
volume, electrolyte balance metabolic
local
12
3
48
134
Heart
metabolic
local
4
7
26
134
gas
local,
exchange
mechanical -
67 2176 16.4
Skeletal muscle
Pulmonary other
k02 Cardiac output
.
-
(mi min-I) (1 min-')
exchanges all received 10
0
17 262
5.8
25
5.8
Compiled from ROWELL, 1974; MASON, 1968; MITCHELL et al., 1973; WADEand BISHOP,1962; HARRISet al., 1977.
There is a great deal of evidence that the stimulus for heat dissipation responses is a function of temperature signals from both internal and peripheral thermoreceptors (CABANAC et al., 1971;HARDY et al., 1971). These signals are integrated primarily in the anterior/preoptic hypothalamus. The integrated thermal signal, which represents a weighted mean body temperature (SNELLEN, 1966; GONZALEZ et al., 1978), is compared to a reference temperature giving a combined signal which serves as an efferent neural thermal drive. Thermoregulatory responses are proportional to the deviation from the reference. A number of thermoregulatory responses occur in parallel during exercise. Figure 8.2 shows typical physiological responses to steady exercise at metabolic rates up to 70 yoof the maximum. During thermal equilibrium a person's internal body temperature is independent of air temperature. However, body temperature may change accordingto the intensity of exercise or energy metabolism (M. NIELSEN,1938). The equilibrium body temperature is related to the work intensity, as a percentage and HERMANSEN, 1966). of the maximal aerobic capacity of an individual (SALTIN Figure 8.2 also shows that during exercise in ambient temperatures below 30 "C,the mean skin temperature is primarily a function of the ambient temperature of the
Exercise physiology
127
T a b l e 8.1 during rest and exercise in a thermoneutral zone
-__
Neurogenic control Primary control
Responses constrictor (= C )
sy rn pat het ic adrenergic (a)
adrenergic (a) and cholinergic
C D
sympathetic adrenergic (a)
C
(same as visceral)
-
Secondary control
Responses
sympathetic cholinergic to sweat glands a, high stimulation /3, normal
dilation (= D)
-
C D
any of these: low Po2,high Pco,, ' K hyperosmolarity, pH, local metabolites
p, high stimulation
D D
-
-
-
-
-
y , dopamine -
sympathetic adrenergic
C (not significant)
sympathetic adrenergic
C
Chemical control RePrimary chemical sponse
-
D
parasympathetic cholinergic
D
high PCO,
D
n, high stimulation
C D
adenosine, metabolites, lowPoz
D
low Poz (hypoxia)
C
p, normal stimul-
ation minimal autonomic innervation, vessels almost fully dilated
-
n
-E
E
3
c
I
W
E
0
E, l W m F 2 1
Fig. 8.1. Relationship between observed evaporative heat loss ( E ) and that required for heat balance Er (from GONZAL~Z et al., 1978). Er is taken as the difference between metabolic heat production and total sensible heat loss Conditions: 28 %
co
2
max.,
Y
-
0.75 rn .'-s
1
- subjects unacclimatcd. Symbols represent n d o w rubjects
R. R. GONZALEZ
128
37.5
2 37.0
e-b-b
-------Rest o---'_o--o-o
Fig. 8.2. Relation of evaporative heat loss E at the skin, rectal temperature Tc and mean skin temperature 7, to metabolism and ambient temperature T,, at rest and at three levels of bicycle ergometer exercise (from GAWE et al., 1969) Ambient temperature:
m 200 Metobdism
300
Loo
500
10
20
a
@
- 10T; to2max.:
- 30 'C,A - 20 ' C , - 70%, A - SO%,
- 30% environment and is independent of the level of work. (B. NIELSEN, 1969; SNELLEN, L W rn-2I
T,
loci
0
1966; STOLWIJK et al., 1968). Only when the combined metabolic and external heat load exceed the body's ability to dissipate heat by evaporation, caused either by an insufficiency of sweating or by a decrease in evaporative capacity of the environment et al., 1978). do skin and internal body temperatures rise (GONZALEZ I n steady conditions for a constant work intensity (yofO2 max), skin evaporation increases with To(= T,) and T'. Also, for a constant environment regulatory sweating (E,) is closely associated with heat production and mean body temperature (SNELLEN, 1966) (fig. 8.3 a). Additionally, E, is closely related to skin conductance (h,) (fig. 8.3 b). The value of h, is low when mean body temperature is below 35.5 "C and E, M 0, but increases almost linearly with E,, up to 300 W m-2. Two processes contribute to M,, :convective heat transfer by the blood circulation and heat transfer by conduction through the outer layers of skin and tissue. Skin conductance h,, as used in fig. 8.3b, is obtained by dividing the net metabolic heat dissipated at the skin (M,) by the difference between core and skin temperatures, (TC--T'), i.e. h s = Mnl(Tc-TJ* Problems in the evaluation of h, occur when the effective core to skin temperature gradient is small but the net heat flux (M,,) is high, as in exercise in hot, humid situations (GONZALEZ et al., 1978).
Exercise phjsioloEy
a
129
b
Rest r\A--c--,
/o
__
35
36
37
38
Mean body temperoture(OC1
20
I
I
LO
60
hS(W m-2K-'I
Fig. 8.3. For the same data used in fig. 8.2 the relation of cutaneous evaporative heat loss (a) to mean body temperature and (b) to skin conductance, h, Ambient temperature: 0
- 30°C. A - 20°C.
0
- 10°C
The neural control of sweating rate and thermoregulatory skin blood flow have recently been investigated by GONZALEZ et al. (1974), W E N G E Ral. ~ ~(1975) and BRENGELMANN et al. (1977). Exercise transients were used to cause an internal heat stress and regional skin conductance measured by water perfused heat-flow capsules or venous occlusion plethysmography. Local sweating may also be measured by resistance hygrometry (NADELet al., 1971). It has been possible to confirm that thermoregulatory control of skin blood flow in the arm and sweating are clearly linked together and to internal temperature (see table 8.2). According to JOHNSON et al. (1974) skin beds respond by vasoconstriction when skin and core temperatures are excessive. Chest skin conductance and arm blood flow are also reduced during upright exercise when core and skin temperatures become higher than 38 "C (fig. 8.4; GONZALEZ et al., 1974). BRENGELMANN et al. (1977) attribute the decreased skin blood flow response to blood pressure regulation taking precedence over vasodilatory control, at core temperature above a threshold temperature (ca.38 "C). Some animal studies also show reduced heat loss at high core tenlperatures (BOULANT and GONZALEZ, 1977). In rabbits, warm peripheral temperatures markedly decreased the thermosensitivity of heat loss by panting to preoptic temperature. Similar neuronal responses occur in hyperthermic conditions, which suggests competition between peripheral excitatory afferents and local excitatory (preoptic) influences in the control of thermoregulatoryheat loss. Alternatively, the vasoconstriction responses shown in fig. 8.4 may be a consequence of norepinephrine build-up during hyperthermia. The discussion so far shows that connections between various physiological responses during exercise are complex. Behavioural and sensory responses during exercise are also complicated by interactions with other systems. However, some progress has been made in ascertaining the thermoregulatory components. It has 0
- Bioengineering
130
R. R. GONZALEZ T a b l e 8.2 Responses of thermoregulatory skin blood flow and sweating to the factors listed
arm and finger skin blood flow
Primary controls
Secondary controls
sympathetic vasoconstrictor drives and central (T,) and peripheral temperatures
local skin temperature, rate of change of T, on vasoconstriction
Modifying effects
Independent
multiplicative work factors with additive components
V5.)
active vasodilation, metabolic prodcoordinated with strongly T, dependent ucts associated sweating at T8 (the relative importwith sweating (bra- > 34 "C,but not a t ance of core and skin dykinin) all skin sites temperatures ranges from 7.5:l to 129) sweating : whole body and local
additive central inte- local skin temstrongly multipli- work gration of core and perature; rate of cative factors peripheral temperaskin cooling (more influential than t ures (the relative importskin warming); ance of core and skin skin hydration temperatures ranges (wettedness) from 5:l to 1O:l) Compiled from WENGERet al., 1975; JOHNSON et al., 1974; NADELet al. 1971; ROWELL,1974 WYss et al., 1974; GONZALUet al., 1978.
36
37
3.9
39
LO
Oesophageal terrperoture K )
10
35
38
37
N
Tb loci
Fig. 8.4. Relationship of arm blood flow to oesophageal temperature, during upright exercise, (redrawn from BRENGELMANN et al., 1977) and chest skin conductance to mean body temperature (Tb) (from GONZALEZ et al., 1974); work rate: 123 W
also been possible to relate sensory to effector responses, these are discussed in the following sections. THERMAL SENSATION IN EXERCISE
The complex thermoregulatory behaviour occurring in humans is similar to that in animals, where it has been shown that small changes in skin or hypothalamic temperature evoke appropriate responses to keep an animal in a thermoneutral
Exercise physiofogy
131
zone (STITTet al., 1970). CABANAC et al. (1964) found that dogs will voluntarily seek a warmer environment when large quantities of cold water are ingested; a colder environment was sought after ingestion of warm water. Thermeregulatory behaviour in animals is determined by a combination of skin and hypothalamic temperatures. CORBIT (1969) suggested that skin temperature was the dominant stimulus. In another study (ADAIR,1977) found that extrahypothalamic temperatures alter the thermosensitivity of a behavioural response to hypothalamic preoptic temperature. STITTet al. (1970) showed that both physiological and behavioural thermoregulation, in monkeys, are closely linked to the same thermocontroller drive, as earlier suggested by GACCEet al. (1967). In humans, thermoregulatory behaviour is also influenced by thermal sensations, which may be perceived as pleasant, neutral or unpleasant. Whether air movement, ambient temperature etc., are perceived as pleasant or unpleasant depends on internal temperature (CHATONNET and CABANAC, 1965) and will therefore be affected by the level of activity. In the resting state during thermoneutrality, when internal body temperature is constant around 37 "C, the affective components of thermal sensation are modified by skin temperature, which depends in turn on air temperature. MARKSand GONZALEZ (1974) obtained magnitude estimations of the pleasantness or unpleasantness of thermal sensations aroused by radiant heat applied to the forehead. Environmental temperatures varied from 5 to 30 "C.When no radiant stimulus was presented, local sensation on the forehead was absent. As the irradiance was increased, the local sensation first became pleasant, increased to a maximum pleasantness, then declined through neutrality to greater and greater unpleasantness. The response curve depends strongly on air temperature. In cold air, the pleasantness reached relatively high values, and the radiation was unpleasant only when very strong. In warm air, on the other hand, even weak radiation was unpleasant. A distance runner or daily jogger knows, by common experience, how exercise level tempers his affective thermal response to an environment. After running for 3-4 minutes the jogger may feel an air temperature of 11-15 "Cpleasant, with a little breeze and in minimal clothing. He would find this unpleasantly cold while walking. The distance runner feels too warm after the second hour at this air temperature and his perception of discomfort is enhanced for a multitude of thermal and nonthermal sensations: stones in his path, weight of his shoes, how wet he is. When they stop running on a cold day both subjects experience an immediate sensation of cold, but discomfort appears after a time. The responses may be related to phylogenetic behavioural responses in lower animals, where skin temperature alone regulates thermal sensation and is affected by the high dynamic sensitivity of thermal receptors (HARDY et al., 1971). MARKSand STEVENS (1972) have recently clarified some aspects of thermal sensation in humans. They found that when small areas were cooled the sensation of cold correlated with the change of local skin temperature. When large areas of skin are exposed, the more rapidly the skin is cooled the more intense the cold sensation. Warmth sensation, however, did not correlate so well with change in skin temperature. The results of these studies are consistent with earlier observations of whole body discomfort due to cold or warmth (STEVENS et al.,
R. R. GONZALEZ
132
1969). Here, the rate of growth of discomfort due to skin cooling was more than twice that of discomfort due to warmth. GAGCEet al. (1969) showed that the sensation of warmth or cold was also strongly affected by skin temperature, which is in turn related to ambient temperature for wide range of continuous exercise levels (30 to 70% ko, max). Changes in thermal sensation also accompany changes of skin temperature during heat acclimation as shown in fig. 8.5. During exercise in a 50 “C environment, the ievel of warmth judgments were 70 to 80 % lower after heat acclimation. Yet the skin temperatures, albeit lower than prior to acclimatization due to a greater sweat output, would probably be judged “warm” by a resting individual. A recent study by PANDOLP et al. (1977), in trained individuals who underwent a nine d2y heat acclimation also showed a lower mean warmth sensation by 16 to 20 %. /-4
39
0
13
20
30
10
Exwcise duration (minutes)
50
mark represents 1 unit on a linear voting scale). From GONZALEZ et al., 1978
Thermal sensation is subjective, though it is a product of information from more than one kind of receptor, and how much and what kind of a sensation there is will depend on the general activity of the central nervous system at the time and on previous experience (GACCEet al., 1967). There is, however, a clear distinction between thermal sensation and thermal discomfort, which becomes greater during exercise. HARDY et al. (1971) suggest that the perception of thermal discomfort is a product of an integrated state of thermoregulation.
THERMAL DISCOMFORT IN EXERCISE
, THERMAL TRANSIENTS
Warm discomfort occurs whenever physiological mechanisms such as sweating and increase in skin blood flow are activated to bring heat loss into balance with metabolic heat production. Cold discomfort arises predominantly from vasoconstriction
Exercise physiology
133
and a subsequent decrease in skin temperature, necessary for heat conservation. In humans, cold discomfort imposes a greater threat behaviourally than body heating, because vasodilation and sweating are ready fuses for maintaining thermal balance. Although either cold discomfort or cold sensation can adequately describe sensory responses to cold environments, a quantification of warm discomfort or warmth sensation is possible only by assessing multiple physiological responses. Warmth sensation essentially follows from skin heating; but, warm discomfort is composed of an iqteraction between physiological and physical factors including sweating, the fraction of the skin area wet with sweat, peripheralblood flow and central and peripheral temperatures. What has been unclear, until recently, is whether thermal discomfort during exercise is directly affected by body temperature or is associated more with the activation of physiological responses. YAGLOU (1949) recognized that during exercise a skin temperature comfortable for the resting state (33-34 "C)may be judged uncomfortable. FANGER (1970) postulated that as activity increases the evaporation rate required for comfort increases, so that low levels of regulatory sweating become somewhat pleasant. The studies of GAGGEet al. (1969) of moderate exercise (up to 290 W m-2), first provided information that the comfortable level for skin wettedness may also rise, from 0-10 %, for rest, to 20-25 % during exercise in thermal equilibrium. These responses are consistent with the rise in internal temperature during exercise. During exercise, as pointed out in previous sections, regulatory sweating lowers T,. However, the preferred temperature also drops in response to the elevation in core temperature. The degree of discomfort is therefore displaced in exercise. The study of GACGEet al. (1969) also showed that feelings of discomfort, in steady state exercise at constant air temperature, were reduced in fit individuals: their most fit subject experienced minimal discomfort over a wide range of exercise intensity. Otherwise, thermal discomfort was closely associated with the afferent responses of sweating and vasomotor activity (skin conductance). In a recent study GONZALEZ et al. (1977) assessed the influence of central and skin temperature respcnses on estimates of discomfort by naked subjects during transient air temperature changes. Ambient temperature (at a constant 4050 % RH) was increased or dccreased at rates 0.3, 0.6 or 0.9 "Cmin-' until 10 "C above or below an expected comfortable environment. Measurements on each of 4 subjects were repeated 2 to 4 times at each activity levels: rest, 26 yo and 40 % ?o, max. The experiments were conducted in random order. The method of magnitude estimation described by MARKSand STEVENS (1972) and MARKSand GONZALEZ (1974) was used to assess thermal discomfort. In this procedure, a subject makes repeated assessment of the strength of his thermal discomfort in response to a particular physical stimulus (operative temperature), which occurs more than once during the experiment. Figure 8.6 shows the averaged responses to one thermal transient. After an initial time lag at the start of exercise (at zero time) the rise of sweat secretion and its evaporation for the 26 % and 40 % max. levels is closely related to the rise of internal temperature. Skin temperature is little affected by the exercise period, but changes with the ambient temperature. After about 30 minutes internal tempera-
vo2
R. R. GONZALEZ
134
-31 0
I
1
I
I
I
I
10
20
30
LO
50
60
J
Time (minutes) Fig. 8.6. Averagechangesoccurring for rest (a), 26 % ( 0 )and 40% (0) V o , max. during thermal et al. (1977). A D is an arbitrary measureof the transients (-0.3, +0.3 "C min-'). From GONZALEZ change in comfort from the initial state v
- 0.1
ms--l
ture and sweat loss are steady. The changes in discomfort also followed the initial rise in internal temperature and were dependent on the % of Vo2 max. During the reduction in T,, discomfort estimates closely followed the decrease in T,, but changed least for the highest exercise level. At rest cold discomfort was maximal. As T, returns to its original value we see an over-reaction in discomfort judgment,which is greatest at 40 yoko2max. Figure 8.7 shows the relationship between magnitude estimates (taken as geometric means) and operative temperature. This figure indicates that the preferred
135
Exercise physiologv
operative temperature drops with increasing activity. The optima occur at 26.lf 1.5"C for sedentary subjects, 21.8rt1.0 "C for 25% Vo, max'and 20.7f0.9 for 40% ko, max. What is clearly evident from fig. 8.7 is that, whereas warm discomfort increases for each higher activity level up to a To of 30"C, the increase
1
18
20
22
26
2L
Operative temperature, To
I
28
30
(OC]
Fig. 8.7. Relationship of discomfort estimates to operative temperature during thermal transients. Activities: 0. - rest: A. A - 2 6 % ; 0, - 40% V b max. Opensymbols - cold,closed symbols - warm 1
conditions
in cold discomfort is much sharper at equivalent activity levels. Similarly, the sensitivity of cold discomfort to decreases in To is more pronounced at higher levels of steady exercise than when resting. The skin wettedness w (estimated from the ratio of EJE,,,), perceived as neither warm nor cold uncomfortable by these unclothed subjects, was 0.06-0.08 for sedentary, 0.18 for 26 % ko, max. and 0.28 for 40 % max. Warm discomfort was more sensitive to skin wettedness at the higher level of exercise. At a skin wettedness above 0.35 (for 40-50% RH), warm discomfort increased progressively, regardless of the level of activity. Table 8.3 summarizes the results. SCARPERI et al. (1976) have recently done studies in which subjects exercised in a water bath set at 29, 32 and 35 "C. Subjects chose the most preferred back skin temperature, by controlling the temperature of a waterperfused pillow. The basis
eo,
Table 8.3 Optimal comfort conditions in exercise
7 25 40
26.111.5 21.8&1.0 20.7h0.9
33.7 32.7 32.2
36.8 36.20 0.08 37.3 36.40 0.18 37.5 36.41 0.28
T , = (Ts+4Tc)/5. To is the operative temperature.
R. R. GONZALEZ
136
of the study was the observation of CABANAC et al. (1971) that a difference between preferred temperature (T,) and 33 "C serves as a measure of behavioural change. Figure 8.8a is a; example from this study. When the water temperature is comfortable i.e. equal to the mean skin temperature the subject's Tpis equal to the water temperature and the behavioural output is zero. Figure 8.8a indicates that at higher b
a
9'.
-201
m
Fig. S.8a. AT(= T'-T,)
;.I 15
,;$
c
-01
37
34
35
36
37
38
as a function of oesophageal temperature(T,) (from SCARPERI et al., 1976)
TH,o is water temperature damping the total skin temperature
Fig. 8.8b. Plot of preferred temperature (Tp)against weighted mean body temperature for the subjects in fig. 8.8a, where T = (TS+47',)/5
levels of steady work, in which T, is higher, the difference, AT, between Tp and T, is negative if water temperature is too hot or positive (i.e. calling for more heat) if the T, is low and water temperature is low. Figure 8.8 b shows the corresponding responses of preferred temperature to mean body temperature, Tb = (Ts+4T,)/5. It is clear that, within the bounds of normal physiological response, the higher Tb due to exercise level, the lower is the preferred temperature. It is also clearly shown that exercise level does not change dramatically the slope of the response of T, to Tb,indicating that only the integrated body temperature is important in the control. 4 but this For example, a water temperature of 28 "C corresponds to T b ~ 3 6 . "C, combination only exists for higher work levels. HEAT ACCLIMATION
One other component of thermal discomfort that has been only touched on, up to now, is the change occurring during heat acclimation. If discomfort reflects adequately an integrative response to environmental conditions, one can predict that a change in discomfort occurs in similar fashion to changes in other thermoregulatory responses; e.g., increasing in response to an internal temperature signal or other physiological change, with or without a change in gain from a lowered internal body temperature reference point.
137
Exercise physiology
Recent studies (GONZALEZ et al., 1974; NADEL et al., 1974) have shown that during heat acclimation the sweating response is mediated by a lowered internal temperature reference point, with similar thermal sensitivity within the ranges of tolerable internal temperatures. However, in combination with a n exercise training or hot-dry acclimatization regimen, increases in responsiveness of the sweat gland occur for every increase in internal temperature. Figure 8.9 shows results from the study by GONZALEZ and GAGGE(1976). Mean magnitude estimates of discomfort are plotted at 10 minute intervals for exercise, at similar work rates, for days I and 6 of acclimatization in humid and dry conditions.
Fig. 8.9. Magnitude estimates of warm discomfort as a function of exercise duration for days 1 and 6 during exposure to 40 "C with high humidity (circles) and low humidity (traingles). 0,A unacclimated subjects; 0 , A heat acclimated subjects. (From GONZALUand GAGGE,1976)
Q
M
a
t (mnutesl
The most significant change during the daily heat stress was that discomfort during exercise had increased with each successive day of humid acclimation but decreased and levelled off in the hot-dry experiments. In hot-dry environments thermal balance was attained earlier during exercise as acclimation progressed ar.d average T, and heart rate never increased more than 0.7 "C and 43 beats min-' , respectively, above the resting level. Total sweating increased with each day of an experiment, and it is possible that the pleasant sensation of this evaporation of sweat contributed to the decreases in discomfort dur$g acclimation (see also fig. 8.5). For humid heat acclimatization, the discomfort increases were determined to a large extent by signals from internal temperature and skin temperature increases. In fig. 8.10a we see that the discomfort level for both days 1 and 6 is related in the same way to local sweating rate, but that a lower Tb is required for comfort on day 6. The important fact is that discomfort and sweat gland output respond to a similar centraI thermal drive. Thus we see that a heat accIimated individual couples discomfort judgments to a changed thermoregulatory response. In recent work (GONZALEZ, WENGERand ROBERTS,unpublished) we have shown that estimates of discomfort are closely associsted with forearm skin blood flow, measured by electrocapacitance plethysmography during an exercise training and heat acclimation study. In this study, four untrained subjects went through an exercise training and heat stimulation regime (ROBERTSet al., 1977). Averaged results (as geometric means) are shown in fig. 8.1 1. In brief, subjects initially exercised at 60 yo Ifo2 max, in a neutral environment (T,= 25 "C and low humidity). They then underwent 10 days of exercise training (new change was w 14 % in Vo2max.) followed by another
R. R. GONZALEZ
138 a
b
Fig. 8.10. Relation of warm discomfort to (a) forearm sweating rate and (b) to mean body temperature (7'') for heat acclimation experiments during exposure to 40 "C with high humidity (30 minutes of exercise for day 1;25 minutes of exercise for day 6). Conditions: T, = 40 "C, e, = 1.4 to 6.5 kPa 0-
day I ,
A arm blood flov. Iml /inin 100mlj
-
day 6 ; large symbols - r a t
A chest sweoting [gm-2min-')
Fig. 8.11. Changes in mean magnitude estimates of warm discomfort AD plotted against changes in (a) arm blood flow, (b) oesophageal temperature and (c) chest sweating during upright exercise 0 (at 60% Vo2 max.) before ( 0 )and after training ( 0 )and heat acclimation routines (0).AD is an arbitrary measure of the change in comfort from the initial state, as in fig. 8.6. (GONZALEZ, WENGERand ROBERTS, unpublished)
vo2
run at 25 "C. The heat stimulation consisted of 6 5 0 yo max. exercise at T, = 35 "Cand 34 "Cwet bulb temperature. Magnitude estimates, taken each minute, on the final day of each treatment (at 25 "C, 1.3 Ha) show a common relationships between discomfort and skin blood flow. The relationships between discomfort estimates and AT, or chest sweating changes were unaffected by exercise training, but altered during heat acclimation. After heat acclimation, sweating rates were higher at lower AT,: but, for the same changes in discomfort, a shift occurred in the response at
Exercise physiology
139
lower body temperatures (GONZALEZ and GAGGE,1976). The discomfort sweat rate relationship was displaced to lower temperatures in the heat acclimatized state 'but had the same slope. Since the exercise experiments were in a neutral environment, where no sensation of warmth is expected, the responses indicate predominately a change in central mechanisms. EFFECTIVE TEMPERATURE
The final question of discomfort response during exercise is whether some predictive integrated index of thermal strain is possible. GAGGE,in chapter 5, has reviewed this topic more fully. It is clear that physiological responses correlate best with a n Effective Temperature that can be used in the heat balance equation. This Effective Temperature (ET) may be determined for steady moderate exercise over a wide range of environments and is defined as the dry bulb temperature, at 50 % RH, at which total skin heat loss would be the same as in an actual environment. Figure 8.12 indicates a graphical way of determining ET based on heat balance. Where RH = SO%, ET = T,;, where RH > SO%, ET is higher than To; where RH is < 50 %, ET is less than T,. ET may also be determined analytically as discussed by GAGGE(chapter 5).
Fig. 8.12. Graph which can be used for estimation of an Effective Temperature (ET, "C)for exercise The derivation of ET and other indices is discussed in chapter 5
Figure 8.13 shows results from a recent study by GONZALEZ et al. (1978) of discomfort at moderate exercise intensities over a wide span of T, and humidity. Warm .discomfort, as a behavioural response, is strongly related to integrated body temperature (?=') and is also closely associated with effective temperature. In fig. 8.13 there is some suggestion that the sensitivity of discomfort to Tb becomes reduced above Tb = 38 "C, associated with a rapid increase when ET > 42 "C. The temperature
R. R. GONZALEZ
140
,.p'
5L -
V
= 0.5
r
= 0.92
0.
5
3-
.W
2'-
B E
g.
1-
0-
I
3.6
L
20
25
30
35
LO
I I
L5
37
-
I
1
38
39
Tb (OC)
Fig. 8.13. Warm discomfort as a function of Effective Temperature (ET index) and mean body temperature Tb.From GONZALEZ et al. (1977). Here Tb is taken as (T,+9TC)/1O,where T, is measured as the oesophageal temperature. This relationship applies in hot conditions
of 38 "C has been noted before as the threshold where the sensitivity of skin blood . . et al., 1977) and increases in VE/Vo, flow to T, becomes reduced (BRENGELMANN and eE/VCo2 occur (PETERSEN and VEJBY-CHRISTIANSEN, 1973). denotes a respiratory rate. The importance of having a rational index, developed in the above manner, is that prediction of severe heat strain for exercise is possible. For moderate exercise an ET of 25 "Ccould be adduced as a comfortable level, 35 "C as excessive, and above 42 "C signalling circulatory and ventilatory problems.
eE
SENSE OF EFFORT IN EXERCISE
In our example of the jogger and long distance runner, it was pointed out that sensory responses of thermal sensation and discomfort are closely tied with associated physiological responses. These, in turn, are controlled as an outcome of centrally and peripherally integrated thermal signals. If we were to assign these runners to a task such as a mile run, the effort of this run, as well as the aerobic and anaerobic energy yields, would be dissimilar in the two subjects. The trained runner may be able to run a 4 minute mile, a t about 70 yoaerobic and 30 % anaerobic yields. The jogger, for an 8 minute mile, may require 85 % aerobic and only 15 % anaerobic yields. The sensation of effort would be the intensity of the exercise related to the capacity of each runner. This sensation may be independent of environmental (thermal) stimuli but the sensation appears to have central and peripheral components. BORG(1962) found a relationship between estimates of exertion (perceived exertion) and heart rate during continuous exercise. This parameter, as well as respiratory rate or minute ventilation ( i T E ) , have been suggested as part of the central component of the perceived effort. KAMONet al. (1974) found reasonable close correlations of heart rate, ventilation and oxygen uptake with ratings of perceived effort. However, for equivalent
141
Exercise physioiogy
strain due to exercise and heat stress, PANDOLF et al. (1972) noted differences i n effort which were not related solely to responses of heart rate and ventilation. As such, peripheral components (muscle fibre or metabolites) have been attributed as the primary forcing input in central nervous integration of the sensation of effort (PANDOLP et al., 1973). CAFARELLI (1977) has recently clarified some aspects of the derived sensation during continuous exercise. By varying pedalling rates on a cycle ergometer at equivalent work outputs, he showed that the sensation of effort is greater at slow rates (e.g. 30 revs min-l) than at high rates (e.g. 60 revs min-'). Figure 8.14 shows some of these relationships. This figure also shows that the sense
L,l L
' 8' ' 12'
I
I
I
16 Integrated electromyogmm ( ~ 1 0 2units. min-1)
Work output I Kgm min-1)
'
20
Fig. 8.14. Magnitude estimation of effort as a function of work output and integrated electromyogram for two cycling rates (from CAPARELLI, 1977) A - 60 revs min-',
A -
30 revsmin-'
of effort, as a function of integrated electromyogram, is greater at the slow pedalling rates. CAFARELLI attributes the greater sensation of effort to a direct consequence of muscle fibre activity at any given contraction. He suggests that sensation of effort responds primarily to a peripheral signal related to both force of muscle contraction and rate of muscle contraction and consequent metabolic demand. Of course, the peripheral signal may also be effected by hydrogen ion and lactic acid production (HARRISet al., 1977) and inadequate O2 delivery to muscles ('MITCHELL et al., 1972). CONCLUSIONS
This chapter has pointed out some patterns which occur within the various body systems during positive continuous exercise. Answers are needed to important questions concerning how these systems interact during intermittent and negative work, which were not covered here. The thermoregulatory system treats the thermal load derived from negative work much the same as in positive work, except that the work output for a given intensity is added to the net heat flux. Although skin temperature increases, the internal body temperature rise is less per given work intensity than in positive work. However, no studies have focused directly on warm discomfort
142
R. R. GONZALEZ
during negative work. The indication is that for a given work intensity, skin blood flow would be greater than in positive work. In respect to the neural thermal drive, an increase in the sensitivity ofthe warm discomfort to body core temperature relationship would occur, primarily because skin wettedness and skin blood flow are higher than in positive work. This hypothesis is suggested by the differences found in core to skin temperature gradients obtained in positive compared to negative exercise (B. NIELSEN, 1969; STOLWIJK and NADEL, 1973). Other studies (BONDE-PETERSEN et al., 1973) indicate that during extensive training with negative work, sensations of perceived exertion become reduced for a given heart rate increase. It is unclear whether the sensitivity of the effort sense to heart rate or body core temperature would be reduced with extensive negative work training. Other basic questions, which are beyond the scope of this review, concern how the regulation of systematic, cellular and sensory responses change during heavy exercise training. Some studies of humans have shown an increased response to oxidation of free fatty acids by active muscle (HOLLOSZY and BOOTH,1976) thus sparing glycogen stores; attenuation of plasma epinephrine and norepinephrine increases during exercise (50 to 60 % lower) (WINDER et al., 1978); and a lower ACTH release by the pituitary gland (FRENKL et al., 1975). These studies suggest adaptive mechanisms which are centrally mediated. Such responses would also affect (and reduce) behavioural processes of endurance trained individuals. As more data are acquired the logical sequence is to model many of the above variables into a form Euch as that developed for the thermoregulatory system (GAGGE et al., 1971). One such attempt has been made, using a systems analysis approach which includes strength, endurance, skill and some psychological factors during exercise (CALVERT, 1976). REFERENCES ADAIRE. (1977), Skin preoptic and core temperatures influence behavioral thermoregrtlation,J. Appl. Physiol. 42, 559-564. EONDE-PETERSON F., HENRIKSSON J., and KNUTTGENH. G . (1973), Effect of training with eccentric muscle contractions on skelefaf muscle metabolites, Acta Physiol. Sand. 88, 564-570. BORGG. A. V. (1962), Physical performance andperceived exertion, Studia Psychologica et Paedagogica, 11, 1-64. BOULANT J. A. and GONZALEZ R. R. (1977), The effect ojskin temperature on the hypothalamic control of heat loss and heat production, Brain Res. 120, 367-372. BRENCELMANN G . L., JOHNSON J. M., HERMANSEN L., and ROWELL L.B. (1977), Altered control of skin blood flow during exercise at high internal temperature, J. Appl. Physiol. 43, 790-794. CABANAC M. and CHATONNET J. (1964), Influence de la temperature interne sur le catactere affectif d’une sensation rhermique cutanee, J. Physiol. (Paris) 56, 540-541. , CABANAC M., CUNNINGHAM D. J., and STOLWJK J. A. J. (1971), Thermoregulatory set point during exercise: a behavioural approach, J. a m p . Physiol. Psych. 76, 94-102. CAFARELLI E. (1977), Peripheral and central inputs to the effort sense during cycling exercise, Europ. J. Appl. Physiol. 37, 181-189. CALVERT T. W., BANISTER E. W., SAVAGE M. V., and BACHT. (1976), A systems model of the effects of training on physical perjormance, IEEE Trans. 6, 94-102. CHATONNET J. and CABANAC M. (1965), The perception of thermal comfort, Int. J. Biometeor. 9, 183-193.
Exercise physiology
143
CORBIT J. D. (1969), Behavioural regulation of hypothalamic temperature, Science 166, 256-258. EKELUND I,, G. and HOLMGREN A. (1967), Central hernodynamics during exercise, Circl. Res. SUPP. I. 20, 1-33-1-43. FANGER P. 0. (1970), Thermal Contfbrt: Analysis and Applications in Environmental Engineering, Copenhagen: Danish Technical Press. L., and C~AKUARY G. (1975), Further experimental results concerning the relationFRENKL R., CSALAY ship of muscular exercise and adrenal function, Endokrinologie 66, 285-291. GAGGE A. P., STOLWIJK J. A. J., and HARDY J. D. (1967), Comfort and thermal sensationsand associated physiological responses at various ambient temperatures, Env. Res. 1, 1-20. GAGGE A. P., STOLWIJK J. A. J., and SALTIN B. (1969), Comfort, thermal sensation and associated physiological responses during exercise at various ambient temperature. Env. Res. 2, 209-229. GAGGE A. P., STOLWIJK J. A. J., and NISHIY.(1971), An eflective temperature scale based on a simple model of human physiological regulatory response, ASHRAE Trans. 77, 247-262. GONZALEZ R. R., and GAGCEA. P. (1976), Warm discomjort and associated thermoregulatory changes during dry- and humid-heat acclimatization, Israel J. Med. Sci. 12, 804-807. GONZALEZ R. R., PANWLFK. B., and GAGGE A. P. (1974), Heat acclimation and decline in sweating during humidity transients, J. Appl. Physiol. 36, 41 9-425. GONZALEZ R. R., NSHIY., and GAGGE A. P. (1 979, Magnitude estimation ofthermal comfort during alternations in activity level and ambient temperature, [In :I Physiological Requirements on the Microclimate in Industry and Problems of Their Technical Realization, ed.: M . JOKL, Inst. of Hygiene and Epidemiology, Prague, Czechoslovakia, pp. 62-73. GONZALEZ R. R., NISHI Y., and GAGGE A. P. (1977). Mean body temperature andeflective temperature as indices of human thermoregulatory response to warm environments, [In -1 New Trends in Thermal Physiology, eds.: Y . HOUDASand J. D. GUIEU,Masson, Pans. GONZALEZ R. R., BERGLUND L. G.. and GAGGE A. P. (1978), Indices of thermoregulatory strairz for moderate exercise in the heat, J. Appl. Physiol. 44. 889-899. HARDYJ. D., STOLWKJ. A. J., and GAGGE A. P. (1971), [In:] Comparative Physiology of Thermoregulation, ed.: G. C. WHITTOW, Academic Press, New York-London, Vol. 11, p. 327. HARRISR. C., SAHLIN K., and HULTMAN E. (1977), Phosphagenandlactate contents of m. quadriceps femoris of man, J. Appl. Physiol. 43, 852-857. HOLLOSZY J. O., and BOOTHF. W. (1976), Biochemicaladaptations to endurance exercise in muscle, Ann. Rev. Physiol. 32, 273-291. JOHNSON J. M., ROWELL L. B., and BRENGELMANN G. L. (1974). Mod$cation ofskin blood flow-body temperature relationship by exercise, J. Appl. Physiol. 37, 880-886. KAMONE., PANWLP K. B., and CAFARELLI E. (1974), The relationship between perceptual information andphysiological responses to exercise in the heat, J. Human Ergol. 3,45-54. MARKS L. E., and STEVENS J. C. (1972), Perceived coldand skin temperature asfunctions of stimulation level and duration, Am. J. Psych. 85, 407419. MARKSL. E., and GONZALEZ R. R. (1974), Skin temperature modifies the pleasantness of thermal stimuli, Nature 247, 473-475. MASOND. T. (1968), The autonomic nervous system and regulation of cardiovascular performance, Anesth. 29, 670-680. MITCHELL J. W., STOLWIJK J. A. J., and NADEL E. R. (1972), Model simulation of blood f b w and oxygen uptake duriiig exercise, Biophysical J. 12, 1452-1466. NADEL E. R., BULLARD R. W., and STOLWUK J. A. J. (1971), Importance of skin temperature in regulation of sweating, J. Appl. Physiol. 31, 80-87. NADEL E. R., PANDOLF K. B., ROBERTS M. F., and S T o L W l K J. A. J. (1974), Mechanisms of thermal acclimation to exercise and heat, J. Appl. Physiol. 37, 515-520. NADELE. R., WENGER C. B., ROBERTS M. F., STOLWIJK J. A. J., and CAFARELLI E. (1977), Physiological defenses against hyperthermia of exercise, Ann. N. Y. Acad. Sci. 301, 98-109. NIELSEN B. (1969), Thermoregulationin rest andexercise, Acta Physiol. Scand. Suppl. 323, 1-74. NIELSEN M. (1938), Die Regulation der KcYrpertemperature bei Muskelarbeit, Scand. Arch. Physiol, 79, 193-230.
R. R. G O N ~ A L E Z
144
K. B., CAPARELLI E., NOBLEB. J., and METZK. F. (1922), Perceptual responses during PANDOLP prolonged work, Percept. Mot. Skills 35, 975-985. PANDOLF K. B. and NOBLE B. J. (1973), The effect ofpedalling speed and resistance changes onperceived exertion for equivalent power outputs on the bicycle ergometer, Med. Sci. Sports 5, 132-136. PANDOLF K. B., BURSER. L., and GOLDMAN R. F. (1977), Role ofphysicalfitness in heat acclimatization, decay and reinduction, Ergonomics 20, 399408. PETERSEN E. S. and VEJBY-CHRISTIANSEX H. (1973), Efect ofbody temperature on steady-state ventitation and metabolism in exercise, Acta Physiol. Sand. 89, 342-351. ROBERTS M. F., WENGER C. B., STOLWJK J. A. J., a n d N A ~ mE. R. (1977), Skin blood flow and sweating changes following exercise training and heat acclimation,J. Appl. Physiol. 43, 133-137. ROWELLL. B. (1974), Human cardiovascular adjustments to exercise and thermal stress, Physiol. Rev. 54, 75-159. RILEY R. (1965), Gus exchange and transportation, Chapt. 40, [In:] Physiology and Biophysics, eds.: T . C. RUCHand H. D. PATTON,Philadelphia, Saunders, pp. 761-787. RUSHMER R. F. (1965), Control of cardiac output, Chapt. 33, [In:] Physiology and Biophysics, eds.: T. C. RUCHand H. D. PATTON, Philadelphia, Saunders, pp. 644-659. SALTINB. and HERMANSEN L. (1966), Esophageat, rectal and muscle temperature during exercise, J . Appl. Physiol. 21, 1757-1762. SCARPERI M., BEHLING K., and BLEICHERT A. (1976), Investigation of the possible role of a work facfor in human behavioral rhermoregulation, Israel J. Med. Sci. 12(9). 1065-1067. SNELLEN J. W. (1966), Mean body temperature and the control of thermal sweating, Acta Physiol. Pharmacol. Med. 19, 99-174. STEVENS J. C., MARKS L. E.. and GAGGE A. P. (1969), The quantitative assessment of thermal comfort, Env. Res. 2, 149-165. STJTT J. T., ADAIRE. R., NADELE. R., ~ ~ ~ S T O L W J. A. IJK J. (1970). The relation between behaviour and physiology in the thermoregulatory response of the squirrel monkey, J . Physiol. (Paris) 63, 425427.
STOLWIJK J. A. J., SALTINB., and GAGGE A. P. (1968), Physiotogicalfactors associated with sweating during exercise, Aerospace Med. 39, 1101-1 105. STOLWIIK J. A. J. and NADEL E. R. (1973), Thermoregulationduringpositiveandnegative work exercise, Fed. Proc. 32, 1607-1613. WADE0. L. and BISHOPJ. M. (1962), Cardiac Output and Regional Blood Flow, Oxford, Blackwell. WENGER C. B., ROBERTSM. F., SKJLWIJK J. A. J., and NADELE. R. (1975), Forearm blood flow during body temperature transients produced by leg exercise, J. Appl. Physiol. 38, 58-63. WASSERMAN K., VANKESSELA. L., and BURTONG. (1967), Interaction ofphysiological mechanisms during exercise, J. Appl. Physiol. 22, 71-85. WINDERW. W., HICKSON R. C., HACBERG J. M.,EHSANIA. A.. and MCLANEJ. A. (1978), Time course of training-induced lowering of plasma catecholamine and glucagon response to exercise, Fed. Proc. 37(3), 664 (abst.). WYSSC. R., BRENGELMANN G. L., JOHNSON J. M., ROWELL L. B.. a n d N i ~ n ~ R s a ~M. o ~(1974), R Control of skin blood flow, sweating, and heart rate; role of skin vs. core temperature, J. Appl. Physiol. 36, 726-733. YAGLQU C. P. (1949). Indices of comfort, Chapt. 9, [In:] Physiologyof Heat Regulation, ed.: L.H. NEWBURGH, W.B. Saunders, Philadelphia.
Chapter 9
THERMAL PHYSIOLOGY OF MAN IN THE AQUATIC ENVIRONMENT I. H O L M ~ R Work Physiology Division, National Board of Occupational Safety and Health, Fack. 5-17 184. Solna, Sweden.
U. BERGH Department of Physiology 111, Karolinska Institutet, Lidingovsgen 1, S-I14 33 Stockholm. Sweden.
CONTENTS
Introduction Body temperature Metabolic response Heat transfer Performance Swimming Diving Survival INTRODUCTION
Although as BURTONonce remarked, “man is not constructed to spend much time in water”, recreational, occupational, military, and other purposes involve more and more people in water activities. The physical properties of water are very different from those of air. Accordingly, the environmental stress easily becomes severe and may even threaten life. On immersion heat exchange between the body and its environment is enhanced. Water has a volumetric specific heat 4000 times greater than that of air, and a conductivity 25 times that of air. Water therefore serves as a gigantic heat sink around the body. As a consequence a nude man will have difficulty reducing heat loss enough to prevent body cooling, even in water at a moderate temperature (20-25 “C). Several reviews have discussed the thermal requirements of man in water from different viewpoints (KEATINGE, 1969; WEBB, 1975; NADEL, 1977). It is the purpose of this article to review the thermal responses and energy exchanges of swimming man with special reference to his physical work capacity and various kinds of water activity. 10
- Bioengineering
I. HOLM~R and U. BERGH
146
BODY TEMPERATURES
In contrast to the situation in air, water exposure of a nude person will bring the temperature of the skin close to that of water within minutes (fig. 9.1). The skinto-water temperature gradient is affected by water temperature, water movement and swimming movements. Still water and/or low water temperature will induce a higher
Fig. 9.1. Temperature gradient (T,--T,) between skin and ambient air or water in resting and exercising subjects in relation to ambient temperature (T,), Subjects were unclothed. Air data from NIELSEN (1969) and SALTINet al., (1972). Water data from NADEL et al. (1974)
0
i
-5
I
1
-
rest, air;
2 - rest. still water; 3 water; 4 swimming
-
-
rest,
flowing
gradient for a given flow of heat. The core temperature on the other hand, is affected little during the first 5 minutes of immersion, irrespective of water temperature and metabolic rate (NADEL et al., 1974; HOLMBRand BERGH,1971 ; HAYWARD et al., 1975a). A longer exposure will, except in a rather narrow 'thennoneutral' range of water temperatures, induce a change in core temperature. The rate of this change is dependent on factors including water temperature, and the subjects metabolic rate and skinfold thickness. Thus, the temperature zone in which man can stay in water without excessive heating or cooling is very narrow compared to air. This is illustrated in fig. 9.2, which shows the oesophageal temperature of a 26-year-old male subject during 30 min of exposure (swimming or resting) at 3 different water temperatures. Swimming in 36 "C water increased core temperature by 1.5 "C while swimming in water at 18 "Cdecreased core temperature by 3.6 "C. Under similar conditions in air, the core temperature would have the same steady state level regardless of air temperature (NIELSEN,1938). Body heating is rarely a problem in water exposure; the problem is usually to prevent body cooling. In a nude person, the minimal rate of heat flow from core to water is determined by subcutaneous insulation : the simplest measure related to et al., 1958; NADEL et at., 1974). Fat people will this is skinfold thickness (CARLSSON have a slower cooling rate than lem ones (KEATINGE, 1960; NADEL et al., 1974; HOLMJ~R and BERGH,1974), a difference most pronounced in very cold water (fig. 9.3). However, in warm water changes in core temperature are much less dependent on the skinfold thickness.
Physiology in aquatic environment
147
I0
L
8
12
16
20
28
2:.
32
36
LO
Tine ( r n l n l
Fig. 9.2. Oesophageal temperature (T,) in one subject during 30 min. exposure to water at thret different temperatures and subsequent recovery in air. Metabolic rate during swimming was higher the lower the water temperature and varied between 7OO-1ooOW duringswimming, and between 14U-620W when resting. From HoLMER and BERGH(unpublished results) -swimming 0.6 m s-' ,----- resting in 0.6 m s flow
-'
1 -
.
Fig. 9.3. Change in oesophageal temperature (T,) in relation to subject skinfold thickness. Data from five subjects after 20 min of submaximal swimming in water and BERGH at three different temperatures. From HOLM~R
c
-2[Sub)
t ttt
t
IH KEKBJF
PK
(1 974)
METABOLIC RESPONSE
To what extent is it possible to balance the large heat loss in cold water by an increased heat production induced by shivering or voluntary muscle exercise? This question is discussed below in the sections dealing with heat balance and survival. Body cooling will activate cold receptors in the skin and in the central nervous system, causing the body to respond in an attempt to retain a normal temperature.
I. HOLMBR and U. BERGH
148
Thus, heat loss is decreased by vasoconstriction and heat production is increased by shivering. The magnitude of the shivering response is a function of both skin and core temperatures, the former being the more important stimulus (NIELSEN1976). If both skin and core temperatures are low the shivering response is greater than if 1976; BERGHand EKBLOM,1978). These one of these stimuli is present (NIELSEN, effects occur at rest as well as during exercise. Cold stress may increase the oxygen uptake by more than 1 1 min-’. One effect of shivering is that the energy cost of a given rate of work, e.g. by swimming, will increase (fig. 9.4). 30
t
; 2 5 1 E
l
vv’/\\ A *
\
0--0 I
36
37
T,
3.9
[“Ci
Fig. 9.4. Oxygen uptake ( V O , ) in relation to oesophageal temperature (T,) in five subjects during submaximal swimming. Subject’s skinfold thicknesses are listed below (in mm). From HOL- and BERGH,1974 x - KB 8.5. 0-KB 7.5, - JF 10.0, A - 1H 5.3, VL- A K 14.0
Chronic cold water exposure induces only mild adaptive responses. HONG (1965, 1973), RENNIE (1965), and RENNIE et al. (1962) reported a lower tissue conductance and higher resting metabolic rate in Korean diving women, at a given water temperature and skinfold thickness, compared with controls, probably as a result of a morekffective restriction of skin circulation. HEAT TRANSFER
Heat produced in the body core and the limbs is transported by blood convection and tissue conduction to the skin surface for dissipation. The heat flux per unit surface area C(W m-’), is determined by the temperature gradient (T,c-~s) between the body core and the skin and the tissue and skin conductance, h, (W m-’ C-I).
C = hs(Tc-!Fs).
(9.1)
Heat arriving at the skin surface will be dissipated to the colder water by convection at a rate determined by the skin to water temperature (T,,,)gradient (Ts--T,,,) and the heat transfer coefficient for external convection, hc(W m-’C-’).
This is the only important heat transfer at the skin because in water evaporation of sweat cannot be used as a heat dissipating mechanism and heat loss due to radiation is negligible. Values for the convective heat transfer coefficient have been derived analytically (RAPP, 1971), measured on a heated copper manikin placed in water (WITHERSPOON et al., 1971) or determined from experimental data on humans (NADEL et al., 1971; BOUTELIER et al., 1977). RAPP (1971) presented values for h,
Physiology in aquatic environmetrt
149
of 105 W m-' C-I in still water, increasing nearly linearly to 411 W m-' C-' at 0.5 ms-'. WITHERSFQON et al. (1971), obtained values very similar to those of RAPP. The values published by BOUTELIER et al. (1977) are much lower: in still water at 15 "C they gave h, 60 W m-* C-' and in 0.25 m s-', h, -- 200 W m-' C-'. NADEL et al. (1974) monitored heat loss from five skin surface sites using heat flow discs. hfeasurements were made on three subjects either resting or swimming in a flume at various temperatures. Weighted mean heat flows were used to determine =I
.:;'" ,d2
-
-
&
A1 02
03
04
200
0
S
Fig. 9.5. Mean weighted heat flow from three subjects as a function of the skin to water temperature gradient. Average data are presented for swimming (top panel) and rest in water (bottom panel). The slope of the relation is the convective heat transfer coefficient, h,, as defined by eq. (9.2). Tw is water temperature. Data from NADELet al. (1974)
-
Water velocity (ms-'): 1 0, 2 - 0.5. 3 - 0.75. 4 - 0.95 Water temperature ('C): closed symbols - 18, half-closcd - 26, o w n - 33
LOO
$
O
w
/ ,
,
,
,
,
230
c5
10
Ts- T,
15
10
("C)
h, from eqation (9.2) using skin temperature values obtained at the same loci and time. In fig. 9.5 the slope of the line gives the appropriate h, values for the prevailing conditions. h, was found to be 230 W m-2 C-' at rest in still water, 460 W m-2C-' at rest in moving water and 580 W m-2 C-' while swimming, regardless of swimming speed. The explanation for the independence of speed could be that during swimming (breaststroke), the swimmer creates such a degree of water turbulence that the effective water velocity around the swimmer is not substantially influenced by swimming speed. During the transient stage of cold water immersion there is a redistribution of heat from the subcutaneous zone into the body core (CRAIGand DVORAK, 1975). A significant amount of heat is lost from the body's peripheral shell, thus diminishing the total body heat storage. This heat loss is normally not accounted for with indirect calorimetric measurements, but can be accurately monitored by use of the heat flow discs or direct calorimetry. Due to the very high coefficient of heat transfer from the skin, man is largely dependent on internal mechanisms to prevent excessive heat loss in water. This is done by minimizing internal heat conductance to the skin. In the theoretical model and RAPP(1970), the flow of heat from core to skin is affected proposed by BULLARD by two parallel resistors, one a fixed resistance corresponding to the layer of subcutaneous fat, and the other a variable resistance associated with the peripheral circulation. The minimal value of skin conductance in a given individual, Iz,, is determined largely by the thickness of the subcutaneous fat layer (CARLSON et al.,
150
I. HOLM~R and U. BERGH
1958, NADEL et al., 1974). Variations in body heat conductance are obtained by controlling peripheral (especially skin) blood flow at a rate determined by the thermoregulatory demands. This mechanism can be partially degraded by other factors; e.g. shivering and exercise involving a change in peripheral blood flow. The increased heat production due to shivering and exercise are accompanied by an increased muscle blood flow, and hence a higher heat conductance to the skin. In cold water the consequent increase in heat conductance may compensate for the extra heat production, leading to a minimal gain in heat storage or even losses. Several authors have reported greater falls in body temperature during exercise in cold water than when resting (KEATINGE, 1961; NADEL et al., 1974; HAYWARD et al., 1975a). A swimmer in thermal balance must produce as much heat per unit of time as is lost to the water. The rate of heat loss is very high from lean subjects during swimming in cold water. The subject's ability to sustain work at high metabolic rates is therefore important for heat balance. NADEL (1977) recently presented a theoretical energy balance for lean and fat individuals at rest and swimming in 18 "C water. This predicts a considerable decrease in internal temperature for the lean individual while the fat individual is able to resist a decrease. For balance the lean individual needs a greatly increased metabolism. PERFORMANCE
It is well known that exposure to cold water will reduce physical performance et al., 1974; HOLMER and BERGH,1974; DAVIES et al., 1975). in most subjects (NADEL Hot water also depresses work capacity. It is not, however, the water temperature per se that is most important, but its effect on body heat balance and temperatures. Reduced physical performance in hypothermia is associated with a reduced peak oxygen consumption (poJ and an increased energy cost of submaximal exercise. These effects are approximately linearly related to core and to muscle temperature (HOLMBR and BERGI-I, 1974; BERGH and EKBLOM, 1978). In extreme cases the depressed peak oxygen uptake will approach the combined requirements for a submaximal exercise and shivering (fig. 9.6). This will, of course, seriously impair physical performance, both in short and in long term exercise. It is still unknown why hypothermia reduces peak aerobic power. It is striking, however, that this reduction is accompanied by a decreased heart rate (NADEL et al., R BERGH,1974). Since there is no reason to believe that the stroke 1974; H O L ~and volume of the heart should increase during hypothermia, cardiac output is most likely decreased. Furthermore, the enzyme activities in the muscles are probably lowered, since many of the enzymes are temperature dependent. Whether this effect will be of importance we do not know. In warm water, the problem is to dissipate heat at a sufficient rate. To avoid excessive heating high rates of heat production must be avoided, thus physical work capacity is limited. Divers are expected to work with their hands in a variety of aquatic environments. Hands are extremely difficult to keep warm in cold water. Accordingly, manual performance is worsened by the effects of cold. Tactile sensitivity, manual dexterity,
151
Physiology in aquatic environment
and muscular strength decrease in cold hands (BOWEN,1968; STANGand WEINER, 1970; VANGGAARD et al., 1975). BADDELEYet al. (1975) studied cognitive efficiency during 1 hour dives in water temperatures of 4 "C and 26 "C.They concluded that Fig. 9.6. Oxygen uptake (as percent of maximal oxygen uptake) for one subject (skinfold thickness 5 mm) during swimming in water at three different temperatures. With decreased deep body temperature maximal oxygen uptake is reduced in cold water. Oxygen uptake is increased due to shivering during swimming at a given submaximal speed. Theoretically, for each submaximal swimming speed there is a lowest water temperature at which the energy cost would equal the peak aerobic power for that particular individual. In the diagram this occurs at 12 "C and 16 "C for swimming speeds of 0.5 m s-l and 0.75 m s-l, respectively
100 -
a
Y
B
80-
a
60-
Q
0
0
-
LO -
E
r"
20-
s
t
, 10
, 11.
,
,
,
18
22
26
Water temperature
,
,
M
31.
(OCI
the competence of a motivated diver may be unimpaired, despite a marked drop in deep body temperature in cold water. On the other hand, continued exposure to cold can produce gradual decay in the performance of cognitive tasks (BOWBN, 1968; STANGand WEINER,1970). SWIMMING
Recreational swimming is enjoyed by a great number of people in most countries. Voluntary swimming usually presents few thermal problems, since the subjective feeling of discomfort from cold normally brings an end to the exposure. However, competitive swimming recruits most of its participants among boys and girls from 8 to 20 years of age. The swimmer often spends up to two hours in the water during one training session. During the spring and summer the best swimmers repeat this programme two or even three times a day for five to six days per week. The exposure is combined with a very high sustained metabolic rate (500-800 W m-2). Since the mechanical efficiency of swimming is low ( H o L ~ ,1974; DIPRAMPEROet al., 1974), around 90 % of the energy yield is converted to heat. The water temperature of the swimming pools is seldom below 24°C. Accordingly, the heat production should be able to balance or in warmer water even exceed the heat losses. However, there are few measurements of heat balance during swimming training. Swimmers are known to weigh more than other athletes of the same height and age due to a higher percentage of fatty tissue. This may develop as an adaptation to the water, resulting in better buoyancy and thermal insulation. Children, especially those with little body fat, easily become hypothermic in cold water due to their high surface area to mass ratio (SLOANand KJZATINGE,1973). To ensure comfortable
152
I. HOLMLR and U. BERCH
conditions and good learning opportunities the pool temperature should therefore be 28-30°C in swimming schools for children. In long distance swimming water temperature is of great importance. Studies (1955) and BERGHet al. (1978) have indicated that success by PUGHand EDHOLM in long distance swimming competitions (lasting several hours) may be more dependent on the athlete’s capacity to withstand cold than on actual swimming capacity. Long distance swimming races have become increasingly popular in many countries. More emphasis should be given to the risks associated with this activity. BERCH et al. (1978) measured the rectal temperature response of 49 subjects to a 3.2 km race in water at 19 “C. After the race rectal temperatures were below 36.0 “C in 18 of the subjects and below 35.0 “C in 10. Five subjects were helped out of the water displaying signs of severe hypothermia. Lean subjects with poor swimming capacity were most affected by the cold water. DIVING
Diving is primarily a “cold” activity, but modern saturation diving has emphasized diverse problems, not only of cold stress (RAWLINS,1972; BRAUER, 1973; WEBB,1975) but also of heat stress (LEMAIRE and MURPHY,1977). The thermal problems connected with diving in cold water have recently been reviewed by WEBB(1975). Shallow diving and swimming present similar problems of cold stress. Except for dives in tropical water, divers use some kind of thermal protection to reduce body heat loss. A number of difficulties must be overcome to create satisfactory working conditions.. Insulating garments that work well in shallow diving become compressed at depth and may lose two thirds or more of their insulation (GOLDMAN et al., 1966). Hands and feet are difficult to keep warm. Respiratory heat loss, negligible at sea level, becomes a major problem at depth and during saturation diving. There are two types of diver’s dress, “wet” or “dry”. The wet-suit, made of closed-cell foamed neoprene, allows water to seep between the suit and the skin. The water warms up and forms a temperate microclimate. The water is slowly replaced by surrounding water, as a result of movement. In the dry suit, an undergarment made of material with high insulative properties is worn under an outer suit impermeable to water. The wet suit is mainly suitable for shallaw diving (DAVIS et al., 1975). Like the dry suit, most of the insulation is lost due to compression at increasing depth. To counteract compression the suit can be inflated with a gas, e.g. expired air. Under extremely cold conditions it is necessary to supply heat within the diving suit, e.g. by free-flooding water systems or electrical wire systems (RAWLINS, 1972). Solution of the problems under all diving conditions, however, awaits further technical improvements. Loss of heat from the respiratory tract is a major route in diving. It threatens the control of body temperature because none of the thermoregulatory responses affects it, and because the loss of heat is directly from the core, i.e. from the central blood stream (WEBB, 1975). The problem is especially important in saturation diving when using helium/oxygen breathing gas mixtures (up to 98 % helium and 2 %
Pliysiology in nqrratic erivironriieni
153
oxygen at 300 m). The physical properties of helium enhance heat transfer to the extent that hyperbaric helium/oxygen atmospheres can transfer heat almost a: well as water (RAYMOND,1975). Guidelines for the design of pressurized chambers therefore become very similar to those determining thermal comfort for a nude man in a water bath, i.e. ambient temperature very close to comfortable skin temperature (33-35 "C). Respiratory heat loss with a breathing gas at 30 atmospheres and 4 "C would be about 100 W, alone equal to the heat generation of a resting man. Under these circumstances exercise does not restore the heat balance, since ventilation (and respiratory heat loss) increase in direct proportion to increasing metabolism (WEBB, 1975). During saturation dives at greater depths it is necessary to provide heating both of the breathing gas and the suit microclimate to ensure heat balance. It is beyond the scope of this review to consider further the many fascinating problems of thermal physiology involved in saturation diving. The reader is referred to the recent publications by WEBB (1975), and RAYMOND (1975). SURVIVAL
The principal cause of death in accidental cold water immersion is most likely hypothermia, not drowning (KEATINGE, 1969). This implies that the chances of survival are very dependent on the potential for prevention of body cooling. As mentioned earlier fat people stand a much greater chance than lean ones. Other 1961). types of protection, such as clothing, will further reduce heat loss (KEATINGE, The rate of heat loss varies between different parts of the body. The face, the groin, and the sides of the thorax are the sites of the largest heat losses (HAYWARD et al., 1973). Measures to minimize heat loss should include special protection of these parts, e.g. by pulling the knees close to the chest, clinging close to companions, or by wearing specially constructed life jackets. The effect of huddling behaviour is to reduce the effective body surface area and hence diminish heat loss (HAYWARD et a]., 1975b). The optimal method of rewarming from hypothermia depends on several factors such as the type of hypothermia (acute or prolonged), the stage of hypothermia reached, the available facilities etc. This subject is beyond the scope of this article and therefore the interested reader is referred to more extensive reviews of the topic (KEATINGE, 1969; WEBB,1975; MACLEANand EMSLIE-SMITH, 1977; COLLISet al., 1977). An important question is whether or not one should swim in order to increase the heat production. The answer to this question is very complex. It depends on several factors, such as type of clothing, skinfold thickness, water temperature and water velocity. In very cold water, lean people, dressed in clothes that are easily penetrated by water, should try to stay as still as possible. On the other hand, fat people immersed in cool water are often better off if they are swimming. However, in most cases it is advasible to avoid body movements in cold water, since the net result of swimming is most often an increased cooling rate SEATINGE 1961; NADEL
I. HOLMER and U. BERCH
154
eta]., 1974; HAYWARD et al., 1975b). The choice whether to swim or not is, of course, also dependent on the distance to land and the chances of rescue. Survival time prediction for men in cold water is very difficult, since cooling rate depends on many factors. Furthermore, the lethal core temperature is not well defined. MOLNAR(1946) collected data on exposure time and water temperature for rescues of subjects accidentally immersed in the ocean. His prediction of survival time, in hours, in relation to water temperature is given in fig. 9.7 (left dotted line). He concluded that all combinations of immersion time and water temperatures to the left of this curve in the diagram would be lethal. Modifications of Molnar’s and BECKMAN, 1962; WHIITINGHAM, 1965) resulted in a similar data (DEFOREST curve (right dotted line) with combinations to the right of the line considered to be safe. Recently, HAYWARD et al. (1975a) presented survival time predictions based on measurements of cooling rates obtained from men and women deliberately exposed to cold ocean water. The solid line in fig. 9.7 depicts the conditions under
y/
( r1 7
/
1 5 -
3 -
/
Fig. 9.7. Prediction of survival times for men in cold water, based on the studies by MOLNAR(1946) / / et al. (1975) (solid (broken line), and HAYWARD // -’ safe _ - - ~ ’line). A prediction line for safe exposure has been , , derived from MOLNAR’Soriginal data. The line of 0 5 10 15 20 HAYWARD et al. predicts 50 % survival time. Adapted Watrr ternperotcrre i°Ci from HAYWARD et al. (1975) /
/
/
/
2
/
/
1
/
_-_---
which an average person will reach a deep body temperature of 30 “C, considered to be lethal in 50 % of the cases. The major difference between the predictions occurs at low water temperatures, where Molnar’s data give a shorter survival time in comparison with the new data. The data of HAYWARDet al. are the best available at present, but more research is needed to provide a fully satisfactory basis for survival time prediction. REFERENCES BADDELEY A. D., CUCCARO W. J., EGSTROM G. G., WELTMAN G., and WILLISM. A., (1975). Cognitive efficiency of divers working in cold water, Hum. Factors 17, 44-54. BERGHU. and EKBLOM B. (1978), Aerobic power durhg exercise at varying body temperatures, International Series on Sport Sciences, Vol. 6. Swimming Medicine IV,University Park Press Baltimore, pp. 323-326. U., EKBLOM B., HOLM&I., and GULLSTRAND L. (1978),&dy temperature response to a long BERGH distance swimming race, International Series on Sport Sciences, Vol. 6. Swimming Medicine IV, University Park Press, Baltimore, pp. 342-344.
Physiology in aquatic environment
155
BOUTELIER C., B o u ~ u sL., and TIhlRAL J. (1977), Experimental study of convective heat transfer coefficient for the Iiuman body in water, J. Appl. Physiol: Respirat. Environ. Exercise Physiol. 42, 93-100.
BOWEN H. M., (1968), Diver performance and the effects of cold, Hum. Factors 10, 445-463. BRAUER R. W. (1973), Temperature regulation in high pressure environments, [In :] The Pharmacology of Thermoregulations, eds. : A. SCHONBAUM and P. LOMAX, Karger, Basel, pp. 99-1 11, BULLARD R. W. and RAPPG. M. (1970), Problems of body heat loss in water immersion, Aerospace Med. 41, 1269-1277. CARLSON L. D., HSEHA. C., FULLINOTON F., and ELSNER R. W. (1958), Immersion in cold water and body tissue insulation, J . Aviat. Med. 29, 145-152. COLLIS M. L., STEINMAN A. M., and CHANEY R. D.'(1977), Accidentalhypothermia: An experimental study of practical rewarming methods, Aviat. Space Environ. Med. 48, 625-632. CRAIGA. B. and DVORAK M. (1975), Heat exchanges between man and the water environments, [In:] Underwater Physiology, Vth Symposium of Underwater Physiology, ed.: C. J. LAMBERSTEN, Federation of the American Societies of Experimental Biology. DAVIESM., EKBLOM B., BERGHU., and K A N ~ U P - J E N SLL. E N (1975), The effects of hypothermia on subniaximal and maximal work performance, Acta Physiol. Sand. 95, 201-202. DAVISF. M., BADDELEY A. D., and HANCOCK T. R. (1975), Diver performance: the efect of cold, Undersea Biomedical Research 2, 195-21 3. DEFOREST R. E. and BECKMAN E. L. (1962), Some contraindications to use of lfejacket for survival, Arch. Environ. Health 4, 56-64. DIPRAMPERO P. E., PENDERGAST D. R.,WILSOND. W., and RENNIED. W. (1974), Energetics of swimming in man, J. Appl. Physiol. 37, 1-5. GOLDMAN R. F., BRECKENRIDGE J. R., REEV~S E., and BECKMAN E. L.(1966),' Wet' versus 'dry' suit approaches to water immersionprotective clothing, Aerospace Med. 3 1 , 4 8 5 4 7 . HAYWARDJ. S., COLLLS M., and ECKERSON J. D. (1973). Thermographic evaluation of relative hear loss areas of man during cold water immersion, Aerospace Med. 44, 708-711. HAYWARD J. S., F~KERSON J. D., and COLLISM. L. (1975a), i"7rermalbalanc-e andsurvivaltimeprediction of man in cold warer, Can. J. Physiol. Pharmacol. 53.21-32. HAYWARLI J. S., ECKERSON J. D., and Corns M. L. (1975 b), Effect of behavioral variables on cooling rate of man in cold water, J. Appl. Physiol. 38, 1073-1077. HOLM~R I. (1974), Propulsive efficiency of breast stroke and freestyle swimming, Europ. J . Appl. Physiol. 33, 95-103. HOLMBR I. and BERGHU. (1974), Metabolic and thermal response to swimming in water at various temperatures, J . Appl. Physiol. 37, 702-705. HONGS. K. (1965), Heat exchange and basal metabolism of Arna, [In:] Physiology of Breath-Hold Diving and the Ama of Japan, eds.: H . RAHNand T. YOKOYAMA, (Publ. 1341), Washington, D.C.: National Academy of Science, pp. 303-314. HONGS. K. (1973), Pattern of cold adaptation in women divers of Korea (Ama), Fed. Proc. 32, 16141622.
KEATINGB W . R. (1960). The eflects of subcutaneousfat and of previous exposure to cold on the body temperature, peripheral blood flow and metabolic rate of men in cold water, J. Physiol. 153, 166178. -TINGE
W. R. (1961), The efect of work and clothing on the maintenance of the body temperature
in water, Quart. J. Exp. Physiol. 46,69-82.
'
KEATINGE W. R. (1969), Survival in cold water, Oxford and Edinburgh, Blackwell. LEMAIRE C. and MURPHY E. L. (1977). Heurt rate and core temperature as indicators ojheat stress during deep underwter activity, Aviat. Space Environ. Med. 48, 146-148. MACLEAN D. and EMSLIE-SMITH D. (1977), Accidentul Hypothermia, London. Oxford, Edinburgh and Victoria, Blackwell. MOLNAR G. W. (1946), Survival of hypothermia by men immersed in the ocean, J. Am. Med. Assoc. 131. 1046-1050.
156
I. HOLMER and U. BERGH
NAoeL E. R., H O L ~I., RBERGHU., ASTILAND P.-O., and STOLWIJK J. A. J. (1974), Energy exchanges of swimming man, J. Appl. Physiol. 36, 465-471. NADEL E. R. (1977), Thermal and energetic exchanges during swimming, [In:] Problem with Temperatlire Regulation during Exercise, ed.: E. R. NADEL, Academic Press Inc., New York, San Francisco, London. NIELSENB. (1969), Thermoregulation in rest and exercise, Acta Physiol. Scand. Suppl. 323. NIELSEN B. (1976), Metabolic reactions to changes in core and skin temperature in man, Acta Physiol. Scand. 97, 125-138. NreLsEN M. (1938), Die Regulation der Korpertemperatur bei Muskelarbeit, a n d . Arch. Physiol. 79, 193-230. PUGHL. G. C. and EDHOLM 0. G. (1955), Thephysiology of Channel swimmers, Lancet, 761-768. RAPP G. M. (1971), Convection coefficients of man in a forensic area of thermal physiology: heat transfer in underwater exercise, J . Physiol. (Paris) 63, 392-396. RAWLINS J. S. P. (1972). Thermal balance in divers, J . Royal Nav. Med. Serv. 58, 182-188. RAYMONDL. W. (1 975). Temperature regulation in helircm-oxygen atmospheres, Lancet, 807. RENNIED. W., C ~ V I N B.OG., HOWELL B. J., SONGS. H., KANGB. S.. and HONGS. K. (1962), Physical insulation of Korean diving women, J . Appl. Physiol. 17,961-966. RENNIED. W. (1965), Thermal insulation of Korean diving women and mn-divers in water, [In:] Physiology ofBreath-Hold Diving and the Ama of Japan, eds.: H . RAHNand T. YOKOYAMA, (Pub]. 1341), D. C. National Acad. Sci., Washington, pp. 315-324. SALTINB., GAGGEA. P., BERGHU.,and STOLWJK J. A. J. (1972), Body temperatures and sweating during exhaustive exercise, J . Appl. Physiol. 32,635-643. SWANR. E. G. and KEATINGE W. R. (1973), Cooling rates of young people swimming in cold water, J. Appl. Physiol. 35, 371-375. STANGP. R. and WEINER E. L. (1970), Diver performance in cold water, Hum. Factors 12,391-399. VANGAARD L.(1975), Physiological reactions to wet-cold, Aviat. Space Environ. Med. 46,33-36. WEBBP. (1975), Cold exposure, [In:] The Physiology and Medicine ofDiving and Compressed Air Work, eds.: P. B. BENNETT and D. U. ELLIOIT, Bailliere-Tindall, London, pp. 285-306. WHI'ITZNGHAM P. D. G. V. (1965), Factors affecting the survival of man in hostile environments, [In :] A Textbook of Aviation Physiology, ed.: J. A. GILLIFS,Pergamon Press Inc. Elmsford, New York, pp. 479438. WITHERSPOON J. M., GOLDMAN R. F.,and BRECKENNDGE J. R. (1971). Heat transfer coef3cients of humans in cold water, J. Physiol. (Paris) 63 459-462.
Chapter 10
CLIMATIC CHANGE AND ACCLIMATIZATION G . E. FOLKJr. Department of Physiology and Biophysics, University of Iowa, Iowa City. Iowa 52242, U.S.A.
CONTENTS
Introduction What is acclimatization The thermoneutral zone Acclimatization to heat The costs of acclimatization to heat Acclimatization to cold The costs of acclimatization to cold Climatic changes of the past The effects of climatic change Conclusions
INTRODUCTION
Acclimatization must be taken into account when we consider the thermal comfort of human populations. We will consider here the definition of acclimatization; the physiological events associated with it; the occurrence of acclimatization in different seasons; but, most important, the acclimatization which must take place when climate itself changes. To enlarge on the last point, we might consider just that acclimatization which takes place in man and his domestic animals as the seasons change, but it seems more important in the present circumstances of civilization to look at the limits of the physiological capacity of acclimatization. To begin with a specific example, a mechanism for man’s response to cold is vasoconstriction in the skin. As the person acclim?tizes, over days or weeks, the degree of vasoconstriction is altered. This, and other changes in physiological status, result in greater comfort for the individual exposed to cold. It is my intention to inquire whether climatic changes 50, or perhaps 1,000 years from now, must result in a different acclimatization for many populations on this globe and possibly new
158
G. E. FOLK
types of acclimatization. Of course, we must firstjustify this by estimating the probability of changes in climate. We will also consider the importance of acclimatization in a society dominated by the “energy crisis”. WHAT IS ACCLIMATIZATION
We will use as a working definition the statement: “acclimatization is the functional compensation or physiological adjustment, over a period of days or weeks, in response to changes of environmental factors”. We may examine a person or domestic animal to see if tolerance or survival is extended, comfort improved or energyexpenditure decreased. If the phenomenon meets one of these criteria, then “acclimatization” has occurred. The improvementshould be clearly definable and detectable by measurable physiological changes as exposure progresses. There are two other terms, c‘acclimation’’and “habituation”, which represent phenomena closely related to acclimatization. These have been discussed in detail in two other volumes (FOLK, 1974a; MONTHTH and MOUNT,1974) and will not be used in this chapter. Some authors substitute the term “adaptation” for ”acclimatization”. Because adaptation has a triple meaning, many environmental physiologists prefer to avoid the term altogether. PRossm (1973) states that “maybe the word ‘adaptation’ is no longer precise enough for serious usage”. For example, the term is used frequently by physiologists to describe the rapid decay of an excitatory process, as in “sensory adaptation”. I choose in this chapter to use a single term only, that of “acclimatization” although when human subjects and animals are exposed in environmental chambers, strictly speaking a slightly different definition should be used. THE THERMONEUTRAL ZONE
One physiological principle, that of thermoneutrality, must be explained, because it is related to the process of acclimatization. When the metabolic rate of a human subject or a domestic animal is plotted against temperature on a graph, for a series of cold and warm environments, a U-shaped curve is obtained (see chapter 1). The region of minimum energy expenditure is referred to as the thermoneutral zone and corresponds to basal or resting conditions. I prefer to refer to the entire curve as the thermoneutrsl profile (FOLK,1974b). Some authors refer to the point on this curve below which energy expenditure must rise in response to cold stress as “the lower critical temperature”. How does acclimatization relate to the thermoneutral profile? In some cases acclimatization changes the basal metabolic rate and there are two completely different curves, one for the acclimatized animal and the other for the non-acclimatized animal. In other cases, where the basal metabolic rate does not change in acclimatization, the mean metabolism of the animal is higher than the mean metabolism of the unacclimatized animal. Cold acclimatization, with its higher energy consumption, is of obvious importance in the nutritional consequences of changes in climate.
Climatic change and acclimatization
159
Because of the important relationship between energy metabolism and climate, it is necessary to discuss other aspects of the thermoneutral zone, especially those which apply to domestic animals. MOUNT(1974) points out that “it is useful to define a number of environmental zones which are ‘neutral’ in different respects”. He suggests the following zones: 1) Minimal metabolism, bounded on each side by rising metabolic rate; 2) Least thermoregulatory effort, bounded at the colder limit by rising metabolic rate and at the warmer limit by increased evaporative loss; 3) Zones defined for particular purposes, e.g. the comfort zone. Zones which are optimal in any given respect, such as animal productivity, growth rate, or the development of thermoregulation in the young animal do not necessarily coincide with either minimal metabolism or least thermoregulatory effort. The usefulness of the thermoneutral profile has been improved by using the equivalent black-body temperature instead of air temperature only. For example, MAHONJXand KING(1977) found that their experimentalanimals, when given achoice, of several air temperatures and intensities of simulated solar radiation in a laboratory preferendum test, spent most of their time in environments in which the equivalent black-body temperature is near the lower critical temperature. ACCLIMATIZATION TO HEAT
The average European is fortunately not challenged by environmental heat exposure for more than a few days at a time. However, other populations are: Arizona’s major cities, beginning in May, experience over 100 days of shade temperature over 38 “C each year. Approximately one sixth of India has this same experience, and almost all of the rest has between 60 and 100 days of shade temperature over 38 “C.As the populations in these areas adjust to the first days of heat waves, a number of physiological events occur (FOLK,1974b); there are changes in skin physiology which include : 1) increased peripheral conductance, 2) increased sweating capacity, 3) a fall in the threshold skin temperature for the onset of sweating, 4) a better distribution of sweat over the skin. Successive exposures to extreme heat cause progressive falls in rectal temperatures and pulse rate. A final characteristic of improvement in the response to heat is subjective; the disagreeable sensations associated with heat exposure are progressively reduced. Many of these measures still appear to be dropping after 24 days of exposure. It is important to distinguish between natural acclimatization and artificial exposure. When an artificially acclimated group was compared with a group acclimatized outdoors, there was a rapid decline in the performance of the artificially acclimatized group (EDEIOLM, 1969). Another contribution from the same laboratory demonstrated that exercise is not necessary for the development of many of the attributes of acclimatization to heat (Fox, 1974). There are racial differences in sweating: Japanese show a higher rate of sweating than Caucasians. Most interesting is the sweating of a chronically cold-exposed race, the Eskimo. These show 50 % less sweating than
160
G . E. FOLK
Caucasians under the same conditions of exposure to heat. As a rule the homeostatic response is to substantially lower the rectal temperature as acclimatization progresses. However, the lowering of core temperature is not essential, in spite of acclimatization. MARONand HORVATH (1978) recorded numerous cases of athletic success with high temperatures. Maron himself finished the last 314 hour of a marathon, run at 75 yo of his maximum metabolism, with a rectal temperature of 41.9 "C. He placed tenth among 105 runners. Case histories of domestic animals improving in physiological condition under continuous exposure to heat are surprisingly difficult to find. The beagle dog does show some of the changes found in human acclimatization, except that sweating is replaced by panting (FOLKand WHIE, 1970). A few cases of acclimatization by cattle are listed in MONTEITH and MOUNT(1974). For smaller animals the ability to conserve plasma volume in the heat, in spite of the decrease in total body water, apparently explains the animals ability to withstand heat stress and dehydration. The question is whether acclimatization to heat potentiates this ability to maintain plasma volume. HOROWITZ et al. (1978) recently studied a desert rodent and found that on exposure to heat, the control animals showed a failure in retention of plasma volume and an outflux of albumen. This outtlux was prevented in acclimatization. They concluded that the improvement in the retention of plasma volume was accomplished by a reduction in the permeability of the capillary bed with acclimatization. Finally, it should be pointed out that acclimatization may sometimes fail, or it may not be achieved. For example, in the International High-altitude Expedition (DILL,1938) approximately one third of the party were ill continuoudy at some of the high altitude camps, especially at 5500 m, while other members of the party were gradually able to achieve a comfortable and normal existence.
THE COSTS OF ACCLIMATIZATION TO HEAT
Large areas of the world have a relatively mild climate and are heavily populated. One possibility of climatic change is that these areas will experience more years of heat and drought. Individuals unable to acclimatize will suffer, and early mortality may result. One of the more predictable problems would be of water supply. In many cases there are already water shortages despite or because of modern technology. Changes in the extent of hot climates could not be solved by migration of human populations. Associated with acclimatization to heat is an increased use of water for drinking and for comfort. At present in countries like India, houses are built where possible in the cooling microclimate of trees. However, one of the resource problems of India and Africa is deforestation and desertification. Thus, numerous factors will conspire against the comfort and indeed the survival of some populations, if climatic changes substantially increase the hot regions of the globe. The need for acclimatization to heat has already been encountered in the summer in the home and work areas in the United States; brownouts and blackouts suddenly remove air conditioning and expose the population to heat stress.
Clinuiric change and acclitnatizarioii
161
ACCLIMATIZATION TO COLD
Physiological.studies on human subjects exposed to the cold present very different results from those on small animals. There have been fewer measurements of the acclimatization to cold of large domestic animals. Mechanisms to combat cold exposure, in contrast to acclimatization to cold, have been described in aborigine subjects in Australia and in the Ainu, the Japanese women divers (FOLK,1974b). I n both of these groups, their remarkable ability to tolerate cold did improve as exposure continued. It has been difficult to show similar improvements on Caucasian subjects. LEBLANC(1975) found a change in the condition of eight soldiers living outdoors all winter at Fort Churchill. At first their oxygen consumption was high; this declined throughout the winter and there was a concurrent fall in internal and JOHNSTON body temperature. Similar results were also demonstrated by DAVIS (1961) who reported a decrease in shivering in subjects exposed to cold. The subjective changes, as demonstrated in groups of men living in the Arctic and the Antarctic for long periods, are probably equally important. It is a common experience that, as such individuals continue their exposure under cold conditions, they wear less clothing when outdoors. In small animals there are conspicuous changes in thermogenesis on cold exposure. These occur less frequently in larger domestic animals and in human subjects. When rodents such as rats and rabbits are exposed to cold, considerable shivering occurs at first, and this is gradually replaced by non-shivering thermogenesis. The presence of non-shivering thermogenesis is easily demonstrated by the norepinephrine test for cold acclimatization. If norepinephrine is injected into the control animal, the metabolic rate will usually rise by approximately 25 %. If at room temperature the same injection is given to the cold-acclimatized animal, the increase in metabolic rate may be as much as 300%. In the case of bats HAYWARD (1968) demonstrated that the metabolism increased by a factor of ten. These generalizations apply to mammals; however domestic fowl have not yet been shown to have the mechanism and PONIEWIERSKI, 1977). of non-shivering thermogenesis (POCZOPKO There are other aspects of cold acclimatization in small mammals: LYNCH, (1972) detected two functional syndromes as physiological adjustment developed. The earliest is associated with the basal metabolic and liver activity, and the second is associated with non-shivering thermogenesis, food consumption and brown adipose tissue. Changes of all these phenomena took aproximately three and onehalf weeks in mice. Other indices of cold acclimatization not discussed here are listed in table 10.1. Numerous monographs now exist on non-shivering thermogenesis (JANSKP, 1975). These describe the particular organs contributing the extra tissue heat production required to replace that produced by shivering (GRUBB and FOLK,1976). A particularly interesting findings is that increased sodium pump activity plays a major part in enhanced tissue thermogenesis in skeletal muscle, in liver, and in the kidney of cold acclimatized rodents. Its role in thermoregulatory heat production during cold acclimatization is probably similar in other mammals (GUERNSEY and STEVENS, 1977). The evidence for cold acclimatization in man, the hamster, and the rat is summarized in table 10.2. 11
- Bioengineerlng
G. E. FOLK
1 62
T a b l e 10.1 Summary of indices of cold acclimatization based on all animals which have been studied Index
Response to prolonged cold exposure (1) Decreases, then stabilizes or returns to
Body weight
Time of acclimation 1 week to 3 months
control Organ weight (kidney, heart, liver, adrenal) Food consumption Body temperature
Urine catecholamine excretion Epinephrine Norepinephrine Adrenal cortex secretion Plasma corticosteroids Urine volume Urine excretion Body fluids
(2) Decreased rate of growth (1) Increases, then stabilizes
1 week to 3 months
(1) Increases, then stabilizes
1 week to 3 months
(1) No change (2) Decreases, then stabilizes or returns to control
2 weeks
(1) N o change (2) Increase, return to control (1) Increases, then stabilizes (1) Increases, then returns to control or lower (1) Increases, then returns to control or lower (1) Increases, then stabilizes (1) Increases, then stabilizes (1) No change (2) Decrease, then stabilize
1 week 2 days to 1 week 1 day to 1 week
2 weeks 1 week 1 week
.
1 week to 3 months
T a b l e 10.2 The changes which take place in man, rat, and hamster, during the process of cold acclimatization Index
Man
Rat
Hamster
d ? i d, nc
d i i d, nc
d i d, nc
d
d
d
nc i i
i i i i i
nc i i i i i d, nc
-
Body weight Organ weights (some) Food consumption Body temperature Water consumption per food consumption Urinary catecholamines Epinephrine Norepinephrine Adrenal steroids Kidney hypertrophy Urine volume Urine excretion Body fluids
? i i d, nc
1. increases; d, decreases; nc, no a p p a m t change; n, data minimal.
i d, nc
i
Climatic change and acclimatization
163
THE COSTS OF ACCLIMATIZATION TO COLD
An alternative possibility of climatic change is a considerably colder winter. Even escape to the Mediterranean or to Florida in the wintertime may be less satisfactory and less comfortable. This change would have to be accompanied by changes in the habits and apparel of affected populations. Retired people have already been urged to remain in the North in the winter, where the thermostat can provide them with a mediterranean environment, instead of migrating South. Unfortunately it is already evident that the concept of comfortable temperate zone homes no longer applies in winter. The problem of home heating will rapidly become worse, although this is not yet accepted by the population at large. For example, a recent poll showed that only 48% of United States citizens know that we must import oil to meet the needs of heating homes and supplying industry (ABELSON, 1977a). Although many United States citizens are not lowering their thermostats, intelligent individuals are, for the first time, keeping the winter temperature in their homes at 18 "C during the day and much lower at night. If the weather grows increasingly cold over the next 50 years, cold in the home may be combined with more severe outdoor exposure e.g. in long waits at bus stops. We must also interpret potential changes in climate in terms of caloric consumption. Shivering before acclimatization to cold, and non-shivering thermogenesis after acclimatization, must require additional energy consumption in both man and domestic animals. Surprisingly little attention has been given to this. In winter the environments of European homes are generally much colder than in America (especially in the British Isles!) Theoretically, in the 1950s and 1960s, British subjects should have eaten more than Americans in the winter if they wore the same clothing. I have not found figures in the literature for this, but CONSOLAZIO et a]. (1963) do provide some data. Their studies of the energy cost of a fixed activity in various temperatures showed: 1) that for a given task, the energy expenditure is related to the body weight of clothing; 2) that cold exposure will raise the metabolic rate; and 3) that in extreme heat, a given task causes a higher metabolic rate than the same task in a moderate environment. Measurements during cold acclimatization of a group of men showed that during the first week 232 kJ were consumed per kilogram of body weight per day, and 271 kJ day-' during the third week of cold exposure. The cost of acclimatization to cold for resting human subjects is an increase in energy consumption of approximately 17 % (CONSOLAZIO et al., 1963). In principle there are similar costs for small domestic animals, such as rabbits and the domestic fowl. Presumably if the climate becomes colder, food consumption must increase and be provided for. Our own results for four species of rodents give similar figures: the food consumption on the tenth day of cold acclimatization was increased by 84 % for brown lemmings (Lemmus),42 % for white lemmings (Dicrostonyx), 112% for the golden hamster (Mesocricetus), and 52 % for the red-backed vole (Microrus). Large domestic animals are, however, much less affected by cold. The comfort aspects of the adjustment to cold of human subjects in North America and also in Europe are complicated by economical and social considerations.
.
164
G . E. FOLK
In the United States oil production has dropped 12% since 1973 and now meets only 50 % of our needs (ABELSON, 1977a). Because of the shortage and the increased cost of oil, many houses are now maintained at 18 "C. Because of the wind chill factor it is not simply a matter of adding more clothing in order to tolerate this lower temperature. Many American homes have heating systems designed for rapid air movement, which also provide summer air conditioning. People may be uncomfortable due to the draught, a circumstancewhich did not occur at 22 "C. As the energy crisis increases further, so that houses must be maintained at 15 "C,people will probably wear more clothing in the house, especially in old age. These conditions have prevailed in England for many years and are just now appearing in the United States. The difference is that old people in England leave an uncomfortably cool house for a cool outside environment. In northern regions of the United States the uncomfortably cool house has an outside environment that may be as cold as -40 "C. If the above generalizations prove correct, we may see an increase in the incidence of cold related illness in the United States. CLIMATIC CHANGES OF THE PAST
What is the probability of climatic changes in the next 50 years? This is best assessed by looking at the past. Most authorities in the field agree that there could be conspicuous changes, although it is possible that a warming trend, due to a factor such as atmospheric COz, may cancel out a cooling trend, due to a factor like dust, leaving our climate unchanged. This is clearly a controversial field but, since the majority of climatologists predict a change, it is reasonable for us to consider the possibilities of both cooling or warming. We are currently more than midway through an inter-glacial period which began 10,OOO years ago, referred to as the Holocene. At the peak of the glaciation the site of NCWYork lay under a sheet of ice half-a-mile thick and glaciers were formed in the mountains of Hawaii. It is important to note, however, that at that time some climates of the world were the same or even slightly warmer than at present. Thus parts of the globe may change their climate dramatically in one direction while others change in the opposite direction. Some processes may affect the entire atmosphere of the planet. For example when Krakatoa erupted in 1883 the global climate cooled by at least 0.5 "C. There have been other recent and similar climatic changes (BROECKNER, 1975). A specific example from several thousand years ago, will also illustrate recent climatic changes. A series of animal skulls found in Peccary Cave in the southern USA included both animals which now live in Canada and Aldska only, or in Central or South America only. The paradoxical pairs are: jaguar and wolverine; arctic lemming and giant armadillo; and musk ox and tapir (QUINN,1972; SEMKEN, 1974). How did this combination of animals, presently separated by thousands of miles coexist? SEMKEN explains that animal life is not limited by the mean annual temperature. It is the hottest day, week, or month in the summer and the coldest day, week, or month in the winter that limits them. Thus, the temperature extremes
Climatic change and acclimatifation
165
at the time of this cave fauna cannot have been as great as at present. We learn from this that it is inadequate to relate problems of acclimatization of human populations merely to the average temperature. The information provided from Peccary Cave tells us that only a few thousand years ago the climate of the southern United States was totally different from today. It must have been more like that of present-day Seattle where even today the tropical jaguar and the arctic wolverine would be able to live and reproduce side by side. The complex and very unpredictable picture of weather and climate depends upon many variables, listed in table 10.3. These factors will tip the balance sheet in favour of cooling or warming. A few authors believe the trend will be slow and T a b l e 10.3 Processes and systems which could contribute to climatic change Natural changes
Man-made changes
1. 2. 3. 4. 5. 6. 7.
Volcanoes Jet stream repositioning Inherent cycles Shifting of Gulf Stream Circumpolar vortex Forest fires Solar flares and Sunspots 8. Alignment of the Earth, Planets, Moon and Sun 9. Axial tilt
Particles from supersonic transports Aerosols from smokes Dust from agriculture Atmospheric fluorocarbon emissions 5. Carbon dioxide from fossil fuels 6. Vegetation changes
1. 2. 3. 4.
predictable. For example, BRYSON(1974) states that the northern hemisphere has been cooling irregularly since 1945. In the north Atlantic, the changes have been about four times as great as the average of the hemisphere. In Minnesota, his studies by pollen analysis suggest that many climatological variables may change significantly in the time scale we are considering. Table 10.4 shows estimated values for five T a b l e 10.4 Maximum rate of climatic change for Minnesota (since last Ice Age)
Time period in years
250 500
lo00 to 2000
Average temperature, hours, July July
Growing season
hrs
%
OC
A°C days
243 256 243 260 228 260
5
21.1 22.2 21.1 22.3 22.6 21.1
1.1
7 14
1.2 1.5
112 134 112 134 142 110
% 20 28 29
Winter snowfall
Precipitation in growing SeaSon
cm
%
cm
%
130 99 127 165 165 94
31
41 46 41 46 41 46
12
30
76
13
13
Table provided by courtesy of R.A. BRYSON.The changes above were not all simultaneous.
166
G. E. FOLK
important measures of the climate; for example the mean winter snow depth probably changed by about 30 yo in a 250 year period. However, BRYSONconsiders the possibility of rapid changes in the future, and states “that the climate nearly always shifts from one condition to another very rapidly” and that “we do not know how close we arc:to the limit where another Ice Age could be triggered”. If anumber of the factors in table 10.3 act together to produce cooling and if another volcano such as Krakatoa were to erupt, then the balance sheet could be changed suddenly in favour of extreme cooling. Of course, this illustration is greatly oversimplified; ABELSON (1977b) recently cited 27 variables which he claimed must be monitored to obtain data needed for studies of climatic dynamics. An evaluation of these variables has been made by a committee of the National Acadeniy of Sciences. They state that it is indeed possible to have a change of the mean annual temperature of the globe of 2 “C within 100 or 200 years; this couldcertainlyaffect theacclimatization and thermal 1977). comfort of human populations (REVELLE, THE EFFECTS OF CLIMATIC CHANGE
It would be difficult to leave this topic without imagining the effects of a potential new Ice Age. HAYSet al. (1976) calculate that the present interglacial period may end within 1,000 to 3,000 years. It is reasonable to picture New York and Warsaw under half a mile of ice. The change could be gradual, so that natural selection would produce changes in temperature regulation. The acclimatization factors which we have been discussing, if they are advantageous, can be fixed by natural selection. The genotype sets the limits for such environmentally-induced variations. This time span of one to three thousand years does not seem long, however, in the perspective of the Vikings in Greenland (1,OOO years ago) and the citizens of Babylon (4,000 years ago). The temperature regulation of these people must have functioned like our own today. Would physiological changes allow Homo sapiens to adjust to an ice age in a few thousand years and would it matter that the portion of the Earth utilized for agriculture would be greatly restricted? Climatologists agree that governments and civil populations are complacent about the possible effects, in the next half century, of changes of climate upon human populations. SCHNEIDER (1976) even states that, in the long run, climatic change is far more important to the world than nuclear power and technology. Shorter term variations can also be important; the possibility of famine due to bad weather troubles many climatologists. The Indian monsoon has failed twice in the 1970s in six years, compared to once in the 1960s. According to BRYSON (1974) there was a remarkably uniform and favourable climate from I955 to 1970. The weather of the monsoon lands is very variable. It affects perhaps 300 to 400 million people, and unfavourable variations restrict the breeding of domestic animals and affect the thermal comfort of human populations. CONCLUSIONS
This paper has described some of the physiological adjustments during acclimatization. The climatological literature suggeststhat, within 50years,heavily inhabited regions may experience significant changes of climate. The proof lies in the story
Climatic change and acclimatization
167
of the Norsemen in Greenland. If similar changes to those they experienced occur again, and are prolonged, large human populations may need to develop more profuse sweating or greater capacity for non-shivering thermogenesis. Domestic animals may present special problems if breeding is affected. An extreme example of environmental limitation is found, not in a domestic animal, but in the American oyster (Ostreu). This can grow throughout an ambient temperature range from 8 "C to 30 "C, but can only reproduce from 16 O to 18 "C. In summary, we must not examine the temperature regulation mechanisms of human subjects only when they are first exposed to an extreme environment, but also in their best physiological condition, i.e., when they are acclimatized. To prepare for the future we must extrapolate to expected environmental demands. When they occur large populations may experience acclimatization for the first time, or perhaps others must experience new versions of acclimatization carried to its limits. The greatest effect of changing climate on acclimatization may be in the area of nutrition. There are few papers, if any, which report on the acclimatization to heat or cold of the malnourished individual, yet, according to one calculation there were 460 million malnourished people in the world in 1976. If malnourished people cannot acclimatize, or they acclimatize to a lesser extent than the well fed subjects, then they are at greater risk from climatic change. Certainly investigators studying thermal physiology should give more attention to the malnourished human. REFERENCES ABELSON P. H. (1977a), Public opinion and energy use, Science 197, 4311. ABELSON P. H. (1977b), Energy and climate, Science 197, 4307. BROECKER W. S. (1975), Climatic change: Are we on the brink of a pronounced global warming Science 189, 460-463. BRYSONR. A. (1974), A perspective on climatic change, Science 184, 753-760. CONSOLAZIO C. F., JOHNSON R. E., and PECORA L. J. (1963), Physiological Measurements of Metabolic Functions in Man, McGraw-Hill, New York. DAVIS T. R. A. and JOHNSTON S. R. (1961), Chamber cold acclimatization in man, J. Appl. Physiol. 16, 221-225. DILLD. B. (1 938), Life, Heat, and Altitude, Harvard Univ. Press, Cambridge, Mass. EDHOLM 0. G. (1969), Acclimatization to heat, J.Roy. Coll. Physns. Lond. 4,27-36. FOLKG. E., Jr. (1974a), Textbook of Environmental Physiology (2nd ed.), Lea and Febiger, Philadelphia. FOLKG. E., Jr. (1974b), Adaptation and heat loss: Thepast thirty years, [In:] Heat Lossfrom Animals and Man, eds.: J. L. MONTEITH and L. E. MOUNT,Butterworths, London, pp. 119-146. FOLKG . E., Jr. and WHITEJ. G . (1970), Acclimation fo heat of the beagle dog, Int. J. Biometeor. 14, 95-101. Fox R. H. (1974), Heat acclimatizatwn and the sweating response, [In:] Heat Loss from Animals and Man, eds.: J. I,. MONTEITH and L. E. MOUNT,Butterworths, London, pp. 277-303. GRUBB B. and FOLKG. E., Jr. (1976), Effect of cold acclimation on norepinephrine stimulated oxygen consumption in muscle, J. Comp. Physiol. 110, 217-226. GUERNSEY D. L. and STEVENSE. D. (1977). The cell membrane sodium pump as a mechanism for increasing therrnogenesis during cold acclimation in rats, Science 196, 908-910. HAYSJ. D., IMBRIE J., and SHACKLETONN . J. (1976), Variations in the Earfh's orbit: Pacemaker of the Ice Ages, Science 194, 1121-1126.
168
G . E. FOLK
HAYWARD J . S. (1968), The magnitude of NOR-induced thermogeriesis in the bat, Can. J. Physiol. and P h m . 46, 713-718. HOROWITZ M., SAMUELOFP S., and ADLERJ. H. (1978), Acute dehydration: body water distribution in acclimated and nonaccliniated Psammomys obesus, J. Appl. Physiol. 44, 585-588. JANSK~,L. (ed.) (1975), Depressed Metabolism and Cold Thermogenesis, Charles University Press, Prague. LEBLANCJ. (1975), Man in the Cold, Amer. Lecture Series 986, Charles C. Thomas, Springfield, Illinois. LYNCHG. R. (1972), Effects of temperature and photoperiod on thernioregulation in the mouse, Ph. D. Thesis, Univ. Iowa (May), 100 pp. MAHONEY S . A. and KINGJ. R. (1977), The use of the equivalent black-body temperature in the thermal energetics of small birds, J . Thermal Biol. 2, 115-120. MARONM. and HORVATH S. (1978), Thermoregulatoryresponsesduring competitive marathon running, J. Appl. Physiol. (in press). MONTEITH J . L. and MOUNTL. E. (eds.) (1974), Heat Loss from Animals and Man, Butterworths, London. MOUNTL. E. (1974), Tfie concept of thermal neutrality, [In :] Heat Loss from Animals and Man, eds. : J. L. MONTEITH and L. E. MOUNT,Butterworths, London, pp. 425-439. POCZOPKO P. and PONIEWLERSKI W. (1977), The effect of gallamine on the metabolism rate and body temperature of’birds,Bull. de I’academie Polonaise des sciences 25,483436. PROSSER C. L. (1973), Comparative AnimaZPhysiology,Vol. 1 (3rd ed.), W. B. Saunders, Philadelphia. QUINNJ. A. (1972), Extinct mammals in Arkansas and related CI4dates c. 3000 years ago, Proc. 24th Internat. Geol. Congr. 12, 89-96. REVELLE R. R. (1977), The panel on energy and climate, Physics Today 30,17-18. SCHNEIDER S. H. (1976), The Genesis Strategy, Plenum, New York. SEMKEN H. (1974), Micromammalian distribution and migration during the Holocene, Proc. Amer. Quaternary Assoc., Third Biannual Meeting, Univ. of Wisconsin.
Chapter 11
MAN IN EXTREME ENVIRONMENTS, PROBLEMS OF THE NEWBORN AND ELDERLY D. ROBERTSHAW School of Medicine, Indiana University, Bloomington, Indiana 47401, U.S.A. Current address; College of Veterinary Medicine and Biomedical Sciences, Colorado State University. Fort Collins Colorado 80523, U.S.A.
CONTENTS Introduction Newborn Cold exposure Heat exposure Elderly Cold exposure Heat exposure Conclusions
JNTRODUCTlON
Man is more ubiquitous in his distribution than any other mammal and is able to survive and thrive in most areas of the world; only high altitudes prevent his permanent residence. With the evolution of an effective physiological thermoregulatory system, combined with his technological advancement, man is able to withstand the extremes of temperature experienced in the world. Thus it is only under unusual circumstances that ambient conditions threaten his well-being and these are usually self-imposed. For example the hypothermia of mountaineers or swimmers who fail to take adequate precautions against excessive heat loss; or the hyperthermia of that particular group of fanatics who run the marathon under hot humid conditions. There are two occasions in the life of an individual when the thermoregulatory system may be incapable of meeting what for the normal adult would be considered rclatively minor changes in ambient conditions. These are at birth, and in old age. At these times control of ambient conditions becomes necessary because of partial or complete failure of the thermoregulatory system. The purpose of this chapter is to
170
D. ROBERTSHAW
examine the physiological basis for the incomplete development, in the case of the neonate, or the deterioration, in the case of the aged, of effective homeothermic mechanisms. NEWBORN
The newborn of many species are equipped with an almost complete thermoregulatory control system and are capable of existence away from the mother without any hazard to body temperature control. The newborn pig, lamb, and rabbit are such species (MOUNT,1959; ALEXANDER, 1962; HULL, 1965; HULLand SEGALL, 1965). Other species, such as the dog and rat, are born in a stage of relatively incomplete development and without a fully functional thermoregulatory system. Such species require either the close proximity of the mother or the building of a nest. The newborn child would be closer to the latter category, although its anatomical development e.g. palpebral fissure separation, is more complete. The stage of physiological development of the newborn child is probably determined by the fact that the “natural” system of management requires that the newborn be carried close to its mother for a large part of the first few months of life, and its ability to control body temperature therefore is related to the close contact it has with its mother. The most common system. of carrying a child in African communitiesis on the back of the mother leaving the back of the child as the area where most heat exchange with the environment takes place. It may be significant therefore that the heat-generating tissue, brown fat, is found subcutaneously between the scapulae (AHERNE and HULL,1966). The practice of carrying, cuddling or slee*ping close to newborn children has been largely abolished in our civilization and so the provision of artificial environments has brought with it the recognition that the newborn has an imperfect thermoregulatory system and that hypothermia and hyperthermia are potential hazards to its well-being. COLD EXPOSURE
The zone of thermoneutrality and the zone of thermal comfort are approximately the same, although newborn pigs and adult man tend to prefer the lower end of the thermoneutral zone (MOUNT, 1963; GAGGEet al., 1965). The lower critical temperature, defined as “the ambient temperature below which the rate o i metabolic heat production . , . increases . . . to maintain thermal balance” (BLIGHand JOHNSON, 1973) is a parameter of obvious practical importance. Its value will be affected by the metabolic rate and by the level at which body temperature is maintained, being elevated by a decrease in basal metabolism and lowered by a fall in the controlled body temperature. The newborn baby shows both these phenomena i.e. a relatively low metabolism and poor temperature control. The “premature” infant is usually defined as being less than 2.5 kg at birth, although gestational age should also be considered on physiological grounds, since a baby of low birth weight for gestational age reflects some degree of retarded uterine growth and is potentially a greater health risk than a baby of low birth weight which 1968). Irrespective of body weight is normal for gestational age (HILLand ROBINSON,
Man in extreme environments, newborn and elderly
171
a t birth the oxygen consumption of all babies is the same within the first 12 hours of birth when expressed on weight specific basis (HILLand ROBINSON, 1968; HEY, 1969) and is approximately 5 ml O2 kg-' min-' (fig. 11.1). By the second day of life oxygen consumption increases to 7 ml O2 kg-' min-' and at 2 months of age reaches 9 ml O2 kg-I min-' (HILLand RAHIMTULLA, 1965; HILLand ROBINSON, 1968; HEY,
5
I
2
:bl I'L1
I
/'
5 t b
3t
F r 6 '
Fig. 11.1. The relationship between oxygen consumption and birth weight. Measurements of oxygen consumption were made within 12 hr of birth (from HILLand ROBINSON, 1968)
[,/',
0
I
1
Body bmght at
I
,
L
birth
[kg)
I
1969). The low metabolic rate at birth is close to that of the foetus near term and that of the maternal organism (ACHESONet al., 1957). Babies which are small for their gestational age show a particularly marked postnatal rise in metabolic rate, but the onset of the rise occurs later than that of babies of normal birth weight (HILLand ROBINSON, 1968; HEY, 1969). The rise in metabolic rate is associated with a decrease in lower critical temperature; thus the baby is able to withstand a lower environmental temperature without having to increase its heat production to maintain body temperature. The reason for the increase in metabolism with age is not known, although SINCLAIR (1972) has speculated that a reduction in the extracellular fluid volume within the first few days of life increases the mass of metabolically active tissue in relation to body weight. SULYOKet al. (1976), from measurements using a gradient-layer calorimeter, concluded that the relative high metabolism of infants which are small for their gestational age is due to an unfavourable surface area to mass ratio, as well as tp low tissue insulation. All babies show a prompt rise in metabolic rate when exposed to temperatures below their lower critical temperature, even when born as much as 11 weeks pre-term (HEY,1969). However, within the first 12 hours of life the magnitude of the response is insufficient to maintain body temperature and a slow, but apparently controlled, fall in rectal temperature occurs (HILLand RAHIMTULLA, 1965; HEY, 1969). This phenomenon has also been observed in the rabbit (HULL,1965) and may be a mechanism that conserves some of the limited energy reserves of the newborn in the early days of life. Four days after birth the metabolic response per degree fall in ambient temperature has approximately doubled and a greater degree of thermal stability results. However, infants weighing less than 2 kg at birth show a slower development of true homeothermy with age (HEY, 1969).
172
D. ROBERTSHAW
The maximal thermogenic response, usuzlly referred to as Summit Metabolism, is about 2.5 times the basal metabolic rate in children over 2 days old, but lower in infants of low birth weight who may take up to three weeks to achieve an equivalent thermogenic capacity (Hm, 1969). Below the lower critical temperature the slope of the line relating metabolic rate to air temperature gives a measure of the total insulation; the greater the slope the lower the insulation. Tissue insulation is a function both of vasomotor control and the presence of tissues with a poor thermal conductance such as fat. Although (1965) the slope of the line increases with increasing age, HILL and RAHIMTULLA maintain that this is due to the thermolability of the young infant rather than to a genuine decrease in insulation. Direct measurements of tissue insulation by HEY and KATZ(1970) and HEYet al. (1970) confirm that there is no evidence for a lack of vasomotor tone in either babies of low birth weight or those of normal weight. However, the tissue insulation of small babies is lower than for babies weighing over 2 kg a t birth; the difference has been attributed to a lack of subcutaneous fat in the smaller infants (HEY and KATZ, 1970). The relatively low tissue insulation of babies compared to that of adults is probably due as much to the small depth of body tissue as it is to any lack of subcutaneous fat, which is usually present in abundance in the full-term baby. The surface to air insulation is also lower for small babies because of the influence of body shape (HEYet al., 1970); convective heat loss per unit surface area is increased from surfaces with a small radius of curvature. The newborn human increases its rate of heat production by neural stimulation of brown fat tissue rather than by shivering. The sympathetic nerves to the brown fat release norepinephrine which stimulates lipolysis. As a result glycerol and free fatty acids arc released. Brown fat metabolizes the free fatty acids for the generation of heat and only glycerol is released into the circulation. Lipolysis of white adipose tissue, on the other hand, yields increments in the plasma levels of both glycerol and free fatty acids (DAWKINS and SCOPES,1965). Thus a rise in plasma glycerol levels unaccompanied by a change in free fatty acids has been used as indirect evidence for the stimulation of brown fat in newborn children. Brown fat is located between the scapulae and around the internal organs of the thorax and abdomen (AHERNE and HULL,1966). If gestation is shortened, brown fat does not continue to grow as it would if intrauterine growth had not been interrupted (ALEXANDER et al., 1973), and this may explain the poor thermogenic capacity of pre-term infants. Although brown fat is known to be the main source of heat (HULLand SEGALL, 1965) other tissues, particularly skeletal muscle, are thought to provide some of the heat increment of coid exposure, being stimulated by some unknown substance which is released from brown fat (HIMMS-HAGEN, 1975). Some infants show no response to cold exposure and rapidly become hypothermic. The reason for this is not known (HULL,1976). Under these conditions thermoregulatory failure could be due to defective thermogenesis consequent upon depletion of brown fat stores. Jn summary, the newborn is able to show the normal responses to cold exposure of vasoconstriction and thermogenesis, but has a relatively high critical temperature
MUIIin extreme environments, newborn and elderly
173
on account of a relatively poor total insulation due in turn to a high surface area to mass ratio and a small depth of tissue between core and skin. This is exacerbated in the pre-term infant, who additionally has a limited therniogenic capacity and a low basal metabolism.
HEAT EXPOSURE
The need to provide adequate insulation to the newborn child is obvious from the above discussion. Similarly the control of the temperature in incubators at the upper end of thermoneutrality is particularly difficult, since undetected changes in the radiant environment and air flow rate may be potentially lethal. Tissue insulation of babies generally decreases on heat exposure, suggesting that even babies of low birth weight possess adequate control of non-evaporative heat loss (HEYand KATZ,1970). It is well established that the vasodilation associated with heat exposure in the adult is only, in part, due to 1 0 s of vasoconstrictor tone; a part of the vasodilation may be due to bradykinin release from the simultaneously stimulated sweat glands (Fox and EDHOLM, 1963). It is not known if such a mechanism exists in the newborn. All infants born within 3 weeks of term studied by HEYand KATZ(1969) more than doubled their total evaporative water loss on the day of birth. Since there were no changes in either respiratory frequency or respiratory water loss it can be concluded that the newborn human depends on sweating as its main mode of evaporative heat loss. Studies by FOSTER et al. (1969) using a ventilated capsule, revealed that full-term babiesare capable ofsweating on the day ofbirth. However, sweating tends to be localized to the skin of the head and more general sweating appears after a few days on the trunk and limbs. The threshold rectal temperature for the onset of sweating also decreases gradually with time. Babies with a poytconceptual age less than 250 days i.e. 3 weeks before term, show a poor sweat response which, if present, is limited to the head region. However, they develop a sweat gland response at an earlier postconceptual age than infants born close to term (FOSTER et al., 1969). The postnatal increase in responsiveness of the glands to thermal stimuli is paralleled by an increase in sensitivity to direct stimulation using acetylcholine (FOSTER et al., 1969) or pilocarpine (UCHINO, 1939). This presumably reflects the development of cholinergic receptors at the secretory cells of the sweat glands, the development of which may be suppressed in utero but stimulated postnatalIy. Since the total number of sweat glands is determined before birth and no new glands develop after birth, the density (i.e. the number per unit surface area) declines with growth. However, the number of functional glands and the secretory capacity of individual glands increases approximately threefold with increasing age. In the human the net result is that maximal sweat rate per unit surface area increases with age (FOSTER net al., 1969). This is in contrast to studies on the calf, where the decline in sweat gland density with age is not matched by a proportionate increase in secretory ability of individual glands, and total sweat rate therefore declines with age
D. ROBEREHAW
114
(HALESet al., 1968). However the secretory ability of individual sweat glands of the newborn child is approximately 30 times that of the newborn calf. In summary, the ability of the newborn to lose heat by evaporation is related to its conceptual age; infants born more than three weeks before term have a defective sweat gland system and require strict control of their environmental conditions.
ELDERLY
Mortality in the elderly population increases dramatically during extreme seasonal temperatures. Although the increase in deaths of people over 50 during heat waves is rarely attributed directly to hyperthermia, an impaired thermoreylatory system may enhance vulnerability to cardiovascular disease and to a lesser extent cerebrovascular disease. From 100 cases of heat stroke AUSTINand BERRY(1956) noted that 70 were older than 60 years and 40 of these were between 71 and 90 years. Accidental hypothermia of the elderly was not recognized as a reiatively common disorder until two decade6 ago (EMSLIE-SMITH, 1958). It probably was not diagnosed as such since the physician did not use a low-reading thermometer and attributed a failure of the mercury to move to a faulty instrument! COLD EXPOSURE
WAGNER et al., (1974) examined the relationship between age and the physiological response to cold *exposure. They observed that the resting body temperature was inversely related to age and that on cold exposure the rectal temperature of boys rose whereas that of men aged 46-67 fell (fig. 11.2). The fall in temperature was Iinb
36 -
c
P
32-
a
t
c
30-
Y
0
30
Tim? 1 miri
60
Fig. 11.2. Group means of rectal and skin temperatures from resting male subjectsof various ages during 60-min exposures to thermoneutral air temperature (30 "C) folIowed by 30-min exposures to room temperatures falling to 17 "C (from WAGNERet al., 1974)
attributed to two phenomena; a higher peripheral blood flow leading to a higher tissue conductance, and a smaller delayed metabolic response. It is apparent from their studies that their elderly subjects had not reached steady-state conditions, so
Moil in extreme environments, newborn and elderly
175
it is not known if they would have achieved thermal equilibrium had the experiments continued. The delay in response of the older subjects suggests that some degree of diminished thermosensitivity accompanies ageing. In this context it is interesting to note that there is a marked reduction in temperature perception with age. Thus COLLINS et al. (1977) reported that, in the cold, young subjects could, on average, perceive mean temperature differences of 0.8f0.2 "C while elderly subjects could discriminate a mean temperature difference of 2.3f0.5 "C. Conscious temperature perception triggers behavioural responses and, although there is some evidence for the separation of behavioural and autonomic control of temperature regulation (CABANAC, 1974), the poor temperature sense of the elderly may be reflected in a similar deficiency in autonomic thermosensitivity. Fox et al., (1973) noted that 24% of the subjects they investigated lived in dwellings considered to be cold and had mouth temperatures less than 35.5 "C.Nevertheless, many of them neither felt cold nor shivered, and were thus unable to comprehend that the heating was inadequate. Such people are dangerously close to the temperature considered to be clinical hypothermia (35 "C)and any event which immobilizes them, or further cold exposure, could precipitate hypothermia. Any loss of thermosensitivity, whether it be to cold or to heat will lead to a greater degree of thermolability. Thus the greater fall in rectal temperature upon cold exposure observed in old people, when compared with young adults (WAGNER et al., 1974;HORVATH et al., 1955), may be a manifestation of thermolability as much as of a defect in affector mechanisms. WAGNER et al. (1974)noted a higher conductance in their older subjects, probably due to a reduced ability to vasoconstrict. COLLINS et al. (1977)in a study essentially within the thermoneutral zone, also noted that some of their subjects were unable to vasoconstrict in the cold. They correlated this defect with the presence of some degree of orthostatic hypotension, which is another important manifestation of dysfunction in the autonomic nervous system, and suggest that the presence of orthostatic hypotension is likely to be a factor in the loss of thermoregulatory efficiency in the cold. The summit metabolism of old people has not been determined, and one can only speculate on whether or not it is likely to be reduced. Both WAGNER et al. (1974) and HORVATH et al. (1955) observed that the resting metabolic rate of the elderly was lower than that of their younger subjects. Since summit metabolism is the sum of basal metabolism and that due to shivering, a lower basal metabolism will also reduce summit metabolism. On the other hand, the maximal-increase in metabolism that can be achieved by shivering may also be reduced. It is well known that the maximal oxygen consumption achieved by exercise declines with age ; even champion runners who maintain a high degree of physical activity are unable to prevent the fall (ROBINSON et al., 1976). The effect of age is largelydependent on a decline in the capacity of the circulatory system to deliver oxygen from the lungs to the working muscles. By the same reasoning, it might be expected that age would also reduce the thermogenic capacity resulting from cold exposure. Thus it would appear that ageing is associated with a degree of thermolability that may be due to a decline irr thermosensitivity. Accompanying the loss of precise
D. ROBERTSHAW
176
temperature control is a higher tissue conductance and a possible lower thermogenic response to cold exposure. It is difficult to determine to what extent the fall in body temperature results from defective thermoregulatory responses and to what extent it represents some loss of homeothermy, similar to that seen when low caloric intake accompanies cold exposure (HAMMEL, 1964). It may be a controlled fall in body temperature, the fall being brought about by a reduction in tissue insulation and shivering thermogenesis rather than a partial failure of heat conservation mechanisms. The hypothermia observed when energy intake is reduced probably favours energy conservation, at the expense of some loss of functional efficiency and manual dexterity due to central and local cooling. However, under conditions of low energy intake hypothermia is achieved by a delayed thermogenic response and tissue insulation remains high (HAMMEL, 1964). In contrast, in the aged tissue insulation is relatively low. Thus the loss of homeothermy with age is unlikely to be analogous to the physiological energy conservation mechanism described by HAMMEL. HEAT EXPOSURE
Not only is ageing associated with a reduction in cold thermosensitivity but it is well established that there is also an elevated threshold for the onset of sweating (FENNELL and MOORE,1973; FOSTER et al., 1976; COLLINS et al., 1977). This applies to sweating induced by both heat exposure and exercise. Thus young men store less heat and have higher rates of cutaneous moisture loss than older men. Furthermore the increase in sweat rate per degree rise in rectal temperature is also greater in younger men, indicating a greater sensitivity of the sweating mechanism to thermal stimuli (WAGNER et al., 1972). As in the cold, conscious temperature sensitivity to heat is reduced in older subjects when compared to young adults, the values being 2.5&0.8 "C and 0.9&2 "C, respectively (COLLINS et al., 1977). This may contribute to the reduced heat loss of the elderly. The reduction in sweat gland function with age has been intensively studied by FOSTER et al. (1976). They compared sweating at different sites on the body surface and noted that the decrease was considerably more marked on the extremities than the chest although at all sites the decrement was significant. For example, fig. 11.3 shows the decline of the maximal sweat rate from the forearm with age. Since there
L
100
L
-3-
e
? !
g
0 0
"80
0
060
L
v)
0
OL-
008
02-
oL
. .
I
,
I
,
0
,
0 .
, t
Fig. 11.3. Forearm sweat rates in men of various ages. Sweating was elicited by immersion of feet and legs in hot water and measuring sweat rate at a core temperature of 38 "C. Chemical stimulation of sweating was by means of intradermal acetylcholine injections. Thermal response (open symbols), chemi-
Man in extreme envirortrnenls, newborn arid elderly
177
is no significant difference in the number of active sweat glands with age the quantity of sweat produced per gland must be reduced. Their studies revealed an interesting sex difference; the threshold temperatures for the onset of sweating of young female subjects were higher and more variable than in young males, particularly on the extremities. Since they did not relate their findings to any particular phase of the menstrual cycle it is possible that the variation they noted was due to the higher threshold that is linked to the higher body temperatures of the luted phase of the cycle (ROBERTSHAW and BEIER,1977). Maximum sweat rates per gland of young men are only slightly higher than those cf young women. FOSTER et al. (1976) found responses of women older than 70 years either non-existent or very poor for all parts of the body. The old women would all be postmenopausal. It is known that oestrogens and progesterone can modify the sweating threshold, and their absence in the postmenopausal state may in some way reduce the secretory ability of the glands. These findings suggest that older women are at greater risk during heat waves than older men. This is not necesarily so; the increase in mortality during heat waves is predominantly due to ischemic heart disease and cerebrovascular disease (ELLISet al., 1975) and men are more susceptible to these diseases than women. The failure of the sweating system seems mainly in the secretory cells of the sweat gland, rather than in the innervation to the glands, since there was a parallel decline in response of the glands to stimulation by both heat and acetylcholine (fig. 11.3). In addition to a reduced sweating ability in the elderly there is also evidence that cutaneous blood flow may be lower resulting in a diminished thermal conductance (WAGNERet al., 1972). Thus precise control of cardiovascular tone of the cutaneous blood vessels is lost in old age, and.effective thermoregulation is prevented, not only in the heat but also in the cold. CONCLUSIONS
When thermoregulation of the young adult is compared to that of newborn children and t$e elderly, there is evidence for the incomplete development of hcmeothermy in the neonate and a gradual functional deterioration in the aged. Thus, whereas the young adult is able to tolerate extremes of temperature with very little threat to his life, the thermal envrionment of the newborn and old require special attention. In the case of the newborn this is particularly important within the first few days of life; in England half the babies who die within the first year of life do so within the first 72 hours (CROSS,1975). Although these deaths have not necessarily been attributed to hypo- or hyperthermia it may have been a complicating factor. One of the main problems of temperature regulation in the elderly is related to their inability to recognize an extreme environment. In addition they have a poorly controlled thermoregulatory system. Accidential hypothermia or hyperthermia thus represents either a direct or indirect threat to the life of older people. 12
- Bioengineering
178
D. ROBERTSHAW REFERENCES
ACHESON G. M., DAWESG. S., and Morr J. C. (1957), Oxygen consumptionand the arterial oxygen saturation in foetal and newborn lambs, J. Physiol. 135, 623-642. AHERNE W, and HULLD. (1966), Brown adipose tissue and heat production in the newborn infant, J. Pathol. Bacteriol. 91, 223-234. ALSXANDERG., (1962), Temperature regulation in the newborn lamb V. Summit metabolism, Aust. J . Agric. Res. 13, 100-121. ALEXANDER G., NICOLD., and THORBURN G. (1973), Thermogenesisinprematurely delivered lambs, [In:] Foetal and Neonatal Physiology, eds. : R. S. COMLINE, K. W. CROSS, G. S. DAWES and P. W. NATHANSIELZ, Cambridge University Press, London, pp. 410-417. AUSTINM. G. and BERRYJ. W. (1956), Observations on 100 cases ofheat stroke, J. Am. Med. Assoc. 161, 1525-1529. BLIGH J. and JOHNSON K. G. (1973), Glossary of terms for thermalphysiology, J. Appl. Physiol. 35, 941-961. CABANAC M. (1974), Therrnoregulatory behavior, F:] MTP Review of Science, Vol. 7, Environmental Buttenvorths, London, pp. 231-269. Physiology, ed. : D. ROBERTSHAW, COLLINS K. J., DORBC., EXTON-SMITH A. M., Fox R. H., MACDONALD I. E., and WOODWARD P. M. (1977), Accidental hypothermia and impaired temperature homeostasis in the elderly, Br. Med. J. 1, 353-356, CROSS K. W. (1975), Prenatal Research, Br. Med. Bull. 31,2. DAWKINS M. J. R. and SCOPES J. W. (1965), Non-shivering thermogenesis and brown adipose tissue in the human newborn infant, Nature, 206,201-202. ELLISF. P., NELSONF., and PINCUSL.(1975), Mortality during heat waves in New York City, July I972 and August-September, 1973, Environ. Res. 10, 1-13. EMSLIE-SMITH D. (1958), Accidential hypothermia, a common condition with a pathognomonic electrocardiogram, Lancet 2,493-495. FENNELL W. H. and MOORER. E. (1973), Responses of aged men to passive heating, J. Physiol. 231, 118-119. FOSTER K. G., ELLISE. P., Do& C., EXTON-SMITH A. N., and WEINERJ. S. (1970, Sweat responses in rhe aged, Age Ageing 5, 41-101. FOSTER K. G., HEYE. N., and KATZG. (1969), The response of the sweat glands of the newborn baby to thermal stimuli and to intraderma1 acetylcholine, J. Physiol. 203, 1330. FOX R. H. and EDHOLM 0. G. (1963), Nervous control of the cutaneous circulation, Br. Med. Bull. 19, 110-114. Fox R. H., MCGIBBON R., DAVIES L., and WOODWARD P. M.(1973), Problem of the old and rhe cold, Br. Med. J . 1, 21-24. GAGGB A. P., STOLWIJK J. A. J., and HARDY J. D. (1969, A novel measurement o*an's heat exchange with a complex radiant environment, Aerosp. Med. 36,431-435. HALES J. R. S.,FINDLAY J. D., and ROBERTSHAW D. (1968), Evaporative heat loss mechanisms of the newborn cag Br. Vet. J. 124, 83-87. . WMEL T. (1964), Terrestrial animals in cold: recent studies on primitive man, [In:] Handbook of Physiology, Adaptation to the Environment, ed.: D. B. DILL, Am. Physiol. Soc., Washington, D. C., pp. 413-434. HEYE. N. (1969), The relation between environmental temperature and oxygen consumption in the newborn baby, J. Physiol. 200, 589-603. HEYE. N. and KATZ G. (1970), f i e range ojthermal insulation in the tissues of the newborn baby, J. Physiol. 207, 667-681. HBY E. N. and KATZG. 1969, Evaporative water loss in the newborn baby, J. Physiol. 200,605-619. HEYE. N., KATZ G., and O'CONNELLB. (1970), The total thermal insulation of the newborn baby, J. Physiol. 207, 683-698. HILLJ. R. and RAHIMTULLA K. A. (1965), Heat balance and the metabolic rare of'newborn babies
Man in extreme environments, newborn and elderly
179
in relation to environmental temperature, and the effect of age and of weight on basal nretabolism, J. Physiol. 180, 239-265. HILLJ. R. and ROBINSON D. C. (1968), Oxygen consumption in Flormally grown, small-for-dates and large-for-dates newborn infants, J. Physiol. 199, 685-703. ~ S - H A G EJ. N (1975), Role ofadrenal medulla in adaptation to cold, F:]Handbook of Physiology, , . SAYERSand A. D. SMITH,Am. Physiol. Sot., Endocrinology, VI, eds.: H . B W C ~ O G Washington D. C., pp. 637-665. HORVATH S. M., RADCLIFFE C. E., H u n B. K., and SPURRG. B. (1955), Metabolic responses of older peopfe to a cold environment. J. Appl. Physiol. 8, 145-148. HULLD. (1965), Oxygen consumption and body temperature of newborn rabbits and kittens exposed to cold, J. Physiol. 177, 192-202. HULLD. (1976), Temperature regulation and disturbance in the newborn infant, Clinics in Endocrinology and Metabolism 5, 39-54. HULLD. and SEOALL M. M. (1965). The contribution of brown adipose tissue to heat production in the newborn rabbit, J . Physiol. 181, 449-457. MOUNTL. E. (1959) The metabolic rate of the newborn pig in relation to environmental temperature and to age, J. Physiol. 147, 333-345. MOUNT L. E. (1963), Environmental temperature preferred by the young pig, Nature 199, 1212-1213. ROBERTSHAW D. and BEIERC. B. (1977), The relationship bezweenplasma levels of sodium and ionized calcium and thermoregulatory thresholds at differentphases of the human menstrual cycle, Physiologist 20, 80. ROBINSON S., DILL D. B., ROBINSON R. D., TZANKOFF S. P., and WAGNER J. A. (1976), Physiologiuzi ageing of champion runners, J. Appl. Physiol. 41, 46-51. SINCLAIRJ. C. (1972), Thermal control in premature injants. Ann. Rev. Med. 23, 129-148. SLILYOK E., JEQUIERE., and J?ROD'HOM L. S. (1976), Relationship between body size, thermal balance and thermal insulation of term infants under various ambient conditions, Biol. Neonate 28,42-56. UCHINO S. (1939), Sweatingfunction of newborn babies (in Japanese), Sanka Fujinka Kiyo 22,239-267. WAGNER J. A., ROBINSON S., and MARINOR. P. (1972), Heat tolerance and acclimatisation to work in the heat in relation to age, J . Appl. Physiol. 33, 616-622. WAGNER J. A,, ROBINSON S., and MARINO R. P. (1974), Age and temperature regulation of humans in neutral and cold environments, J . Appl. Physiol. 37, 562-565.
This Page Intentionally Left Blank
Chapter 12
PHYSIOLOGICAL SIGNALS FOR THERMAL COMFORT M. CABANAC Laboratory of Physiology, Faculty of Medicine Lyon-Sud, Claude Bernard University, P.O.Box 12,69600 Oullins, France.
CONTENTS Motivation and comfort Role of internal temperature Alliesthesia Quantification of alliesthetic changes Spatial summation Thermoregulatory autonomic responses Perception of hypothalamic temperature The set point for temperature regulation Fever Circadian and oestral cycles Pleasure and comfort
system, receiving inputs (8related to environmental variables and producing comfort/discomfort as its
Environment
Comfort
w
MOTIVATION AND COMFORT
There is a close correlation between thermoregulatory behaviour and the motivation for this behaviour, if the subject has no other priorities. While it is unlikely that behaviour occurs without motivation, motivation can occur in tbe absence of behav-
M. CABANAC
182
iour; this divorce may occur even in well-planned experiments. Such a dissociation between motivation and its behavioural expression has been emphasized in other types of behaviour, such as food intake (MILLERet al., 1950). It is therefore interesting to explore motivation as such. It is difficult to speak of the mental state or of the perception of an animal, particuIarly when the animal is low on the phylogenetic scale, But humans can communicate; thus, by using human subjects the experimenter can explore this new field. It is assumed that the affective part of conscious perception of thermal states is the motivation for thermoregulatory behaviour. Thermal comfort, like other conscious perceptions, is a mental phenomenon, it is therefore a function of the nervous system in the human, as in other species. This system, which permits the perception of thermal comfort, is shown in greatly simp& fied form in fig. 12.2. Consciousness
-E
TS
Fig. 12.2. Simplifieddiagram of temperature regulation and possible shivering and pathways to consciousness. Skin temperature (Ts), sweating (E) are perceived signals. Internal temperature (T') is possibly perceived directly (interrupted arrow). I t modifies the pleasure/displeasure of peripheral sensation. Alliesthesia - oblique arrow (from CHATONNET and CABANAC, 1965)
The central nervous system is the circle. This receives signals from the skin (T,) and from the central core (T'), including the hypothalamus itself. The thermoregulatory nervous system produces autonomic responses : shivering, sweating, piloerection, cutaneous vasomotor control. These are represented by E. On this simple diagram, one may add consciousness, as thermal comfort is a conscious experience. What, in the lower part of the diagram, can pass into the region of consciousness? Skin thermal sensitivity arouses a sensation : this obvious point calls for n o demonstration. Let us consider the other arrows. ROLE OF INTERNAL TEMPERATURE
Fig. 12.3 represents a case in which a patient received a gastric probe in which a refrigerated liquid circulated at -4 "C (CHATONNET et al., 1966). Mean skin temperature fell but stayed above 35 "C, in the warm range of sensation. Even epigastric skin, only a few centimetres from the probe, dropped no further than 34 "C. This temperature is perceived as tepid. Therefore, no peripheral cold message reached the central nervous system. Nevertheless the subject feit uncomfortably cold. The cold signal originated from the thermal core. The strong influence of internal temperature on the percepticn of thermal comfort is shown by responses and ratings given by subjects immersed in a water bath (fig.
Physiological signals for thermal comfort
* f
I
+ -,,
V
183
b
+-+
k
37
i E
@
36
'\+-+,+-+-+'
% -A!
'
35
3L
10
20
30
LO
50
60
70
80
90
100
110
120
Time (mtn) Fig. 12.3. Time course of body temperatures in a patient. a) - the gastric probe is frozen down to -4 "C;b) - the patient complains of cold feeling. Air temperature was 25 "C (from CHATONNET et al., 1966)
+
- rectal temperature, 0 - mean skin temperature. A - epigastric skin temperature
Fig. 12.4. Votes on thermal comfort by a group of subjects immersed in a warm bath. .At uniform and constant skin temperature in a waterbath of 38.5"C, various stages of internal warmth-sensation are experienced as internal cranial temperature increases. The stages range from pleasant indifference (stage 1) to almost unbearable heat (stage 4) (from BENZINGER, 1963)
366 368
370 372
37L
37.6 378 380
Tyrnponic temperature
5.2
(OC)
12.4). In a well-stirred bath, skin temperature is coupled to the bath temperature. If comfort changes in such a bath, the origin of the change cannot be attributed to signals from the skin. In a bath, thermal discomfort seems to depend on internal temperature. Subjects feel uncomfortably warm when hyperthermic (CHATONNET and CABANAC, 1965; BENZINGER, 1963) and uncomfortably cold when hypothermic (CHATONNET and CABANAC, 1965). In an air environment, where mean skin temperature cannot be artificially controlled but can still be monitored, this temperature is not a dependable criterion of thermal comfort (GAGGE et al., 1967; PIRLET, 1962; TOPLIFF and LIVINGSTONE, 1970). These observations agree with observations on
184
M. CABANAC
animals which show that thermoregulatory behaviour depends on internal temperature. But it is possible to go further in the analysis of the conscious phenomenon. Internal temperature may influence thermoregulation in a variety of ways : 1. by a direct perception of internal temperature; it is possible that a sensation may be initiated in the hypothalamus or in other internal detectors; 2. by thermoregulatory reactions, such as sweatingand shivering, elicited by internal temperature displacements and perceived by the consciousness. This perception may be at the origin of thermal discomfort; 3. through alliesthesia. ALLIESTHESIA
One of the primary components of thermal comfort is the pleasure aroused by a peripheral stimulus; this, in turn, depends on internal temperature. To a hypothermic subject, a cool (20 "C)stimulus will feel unpleasant, but to a hyperthermic subject, the same stimulus will feel very pleasant. Therefore, the pleasure or displeasure given by the stimulus will depend on the subject's internal state. This phenomenon is called alliesthesia. The internal signal which gives rise to an alliesthesic change is the difference between the set temperature (or level of regulation) and the actual internal temperature (CABANAC,1969). QUANTIFICATION OF ALLIESTHESIC CHANGES
The preferred skin temperature Tp selected by human subjects is a function o their internal temperature T, and of their mean skin temperature Ts.Two models have been proposed: T, = a(Tc-T)Fs+b, (12.1) and T, = a+bT,+cT,, (12.2) where T is a reference temperature, and a , b , c are dimensionless constants. These models are similar. Whether the sensation and control are a product or a sum of signals, any change in deep body temperature or in mean skin temperature will be followed by a change in preference (CABANAC et al., 1972; BLEICHERT et al., 1973; MARKS and GONZALEZ, 1974). In addition, a change in set point will have the same result (CABANAC and MASSONNET, 1974; CABANAC et al., 1976; CUNNINGHAM and CABANAC, 1971). A warm hand temperature is chosen by hypothermic subjects, and a cold hand temperature, by hyperthermic subjects. Thus, those skin stimuli which are pleasurable will have the effect of correcting internal temperature in proportion to the magnitude of its deviation from the set point. SPATIAL SUMMATION
Although alliesthesic changes, by themselves, seem capable of motivating thermoregulatory behaviour, there is only indirect evidence for their participation in the global perception of thermal comfort. It'is probable that thermal comfort is made up
Pliysiologicnl signals for tlieuriznl comjort
18.5
of a sum of elementary effects identical to the hand alliesthesia described above. When one examines only the discriminative part of sensation, warm sensation shows spatial summation (STEVENS and MARKS,1971). It is likely that alliesthesia affects this global peripheral sensory input (BLEICHERT et al., 1973). If this is true, thermal comfort is the sum of all peripheral sensations, plus an affective factor determined by internal temperature. This way of looking at comfort is tempting, because it explains comfort and discomfort in various circumstances. For example, comfort persists when one side of the body is cooled and the other side warmed (BBJEet al., 1948; HALLand KLEMM,1969) or when the body receives asymmetric thermal radiation producing a difference of up to 13 "C between front and back temperatures. (OLESEN et al., 1972). THERMOREGULATORY AUTONOMIC RESPONSES
Any displacement of internal temperature in humans is followed by' the corrective autonomic reactions of shivering or sweating. In the warm, the onset of sweating coincides with the onset of discomfort (HINDMARSH and MACPHERSON, 1962; WINSLOW et al., 1937), and there is a direct proportionality between warm discom-. fort and skin wetness (GAGGEand GONZALEZ, 1973). In the cold, cessation of shiv-. ering coincides with an improvement in comfort (FAYand SMITH,1941). The question arises whether these reactions, which are conscious but not voluntary, are at the origin of the perception of thermal discomfort. In some languages to shiver is. to feel cold and to sweat is to feel warm. However, so far there is no experimental evidence for the participation of autonomic reactions in thermal discomfort. Conversely, acclimatized humans sweat more during heat stress, and this added sweating is perceived as comfortable, because of the pleasant cooling of the skin (GONZALEZ et al., 1973). In a cold environment, non-shivering curarized rats can be trained to switch on infra-red heat (fig. 12.5), which they obtain by operant conditioning with a visceral response (CABANAC and SERRES,1976). It should be noted, in this context, that ectotherms have no autonomic defence but still display a behaviour very well adapted to temperature regulation. Conversely, shivering elicited by cooling the spinal cord (CORMARECHE-LEYDIER and CABANAC,1973) is not accompanied by a thermoregulatory operant behaviour. From these dissociations between autonomic and behavioural responses, it may be suspected that autonomic reactions are not the major component of thermal discomfort. They simply have a common determinant. PERCEPTION OF HYPOTHALAMIC TEMPERATURE
It is possible that internal temperature is consciously perceived as a sensation.. As manipulations of internal temperature have many effects which can be perceived,. it is difficult to confirm or rule out this possibility. Although the only way to solve this problem would be to thermally stimulate the hypothalamus in humans, preferably in the experimenter himself, animal experimentation has yielded valuable information on this question. Cooiing and warming the brain is followed, in rats, by a corrective-
186
M. CABANAC
behaviour only if the skin is warm or cool; as if there is no discomfort with a neutral and HARDY,1970). skin, whatever the hypothalamic temperature (MURGATROYD However, CORBIT (1969) has provided direct evidence that rats actively change their hypothalamic tehperature. He gave rats the opportunity to thermally stimulate their own brain. When heat-stressed, these animals cool their hypothalamus by pressing a lever. The rate of bar-pressing for brain cooling is quite similar to the
Fig. 12.5. A curarized rat cannot shiver; it is maintained alive by artificial respiration. In the absence .of shivering, i t shows its motivation for external heat while hypothermic by turning the infra-red lamp on from changes in heart frequency (from CABANAC and SERRES,1976)
rate measured when rats press to cool their skin by the same amount, and the same model describes both types of behaviour. Nevertheless, this experiment does not answer the question of whether hypothalamic temperature is directly perceived. It is possible that this intracranial self-stimulation is perceived via alliesthesic changes of peripheral sensations or via autonomic reactions. However, in a second experiment (CORBIT and ERNITS,1974), the rats had the choice between two possible types of behaviour: pulling a chain to cool their skin or pressing a lever to cool their hypothalamus. To an increase in hypothalamic temperature alone, they responded by pressing the lever. To an increase in ambient temperature alone, they responded by pulling the chain. To an increase in both hypothalamic and ambient temperature, they responded with both types of behaviour. This result does not prove that there is a hypothalamic “sensation”,but it shows that signals coming from the skin or from the hypothalamus produce different conscious states, because the rats were able to discriminate between them and to make appropriate responses. It is quite remarkable that these
Physiological signals for thermal comfort
187
successful experiments showed self-cooling of the brain : in the opposite situation pigs and rats given the opportunity to warm their brain did not use the operant behaviour to do so (CARLISLE, unpublished; BALDWIN and INGRAM, 1967). THE SET POINT FOR TEMPERATURE REGULATION
The signals recognized above are those of temperature regulation. There is an additional implicit signal: the set point. Internal sensors provide a message analogous to the actual value of the variable it detects. This message is compared to an implicit value “wanted” by the organism. This implicit value will be called the “set point”. The set point is, therefore, the value to which the corrective responses attempt to return the system. The set point may be established in the central nervous system by neurons insensitive to the variations of the particular variable, as hypothesized by HAMMEL (1965) in the case of temperature regulation. It could equally derive from a non-linear response in the sensors; the threshold in a sensor can act as a set point. The activating signal for the regulatory responses is the difference between the actual value and the set point. When a differenceis detected the organism produces the available corrective responses. We have already seen, in the case of temperature sensation, that the pleasure of thermal stimuli depends upon hyper- or hypothermia. Hyper- and hypothermia are defined as deviations of deep body temperature from the thermostat set temperature, the latter being approximately 37 “C in healthy subjects. The determinant of pleasure, which triggers behaviour, and of autonomic responses, such as shivering and sweating, is, thus, apparently the same. Three cases where the set point in deep body temperature is shifted confirm this hypothesis. The first case is fever; the second and third cases concern cyclic resetting of the set point. FEVER
In fever the thermostat is set at a higher value, which all thermoregulatory responses will defend. Though at a higher temperature than during health, the organism will react as if hypothermic when its temperature drops below the fever set point. The pleasure and displeasure of thermal sensation are identical to that observed in the same subject during health, except that the inner temperature criterion is the fever set temperature. This is seen both when subjects rate imposed stimuli (fig. 12.6a; CABANAC, 1969) and when they choose for themselves the most pleasurable skin temperature (fig. 12.6b; CABANAC and MASSONNET 1974; WILMORE et al., 1975). In both cases, the behaviour is identical to control behaviour, but for a higher set point. CIRCADIAN A N D OESTRAL CYCLES
These provide examples where the set point in body temperature is, as shown in studies of autonomic responses, raised and lowered cyclically during periods of, respectively, 24 hours (ASCHOFF,1970) and 28 days (BITTELand HENANE,1975). During the daily cycle (CABANAC et al., 1976) as well as during the oestral cycle (CUNNINGHAM and CABANAC, 1971), the pleasure of thermal sensation is precisely
M. CABANAC
188
adapted to the defence of the cyclic set point of body temperature. One might wonder why, when the set point is shifted, there is no change in the preferred ambient temperature for humans at thermal equilibrium (FANGER et al., 1974; FANGER, 1973). This is because the actuating signal is the difference between set and actual body
i :.
a very pleasant
g
.-
pleasant
c
f f
.0000 O O
0
0
*
indifferent
o
a
& ? .
I
0
E
-0
unpleasant very unpleasant
0 0
L
0 0
0
0 0 50
00
I
lo
I
I
23
30
I
1
40
50
Stimuli temperature
("C)
0
-
k'
u
?1
c !
a
Warm Cold
f l5
36 5
37
375
38
Oesophageal temperature
('C)
Fig. 12.6. Response to thermal stimuli (a), and thermal preference (b) in feverish subjects a) 0 - hyperthemia, 0 - fever. Bath temperature was 33 "C. In both cases deep body temperature was the same, 38.2 < Tb < 38.5 "C (from CABANAC, 1969) b) thermal choice (ordinate) plotted against oesophageal temperature. Ordinate 33-45 "Crefers to the left branch of each Y shaped curve. Ordinate 33-15 "C refers to the right branch of each Y shaped curve (from CABANAC and MASSONNET, 1976)
temperatures. When the set point is shifted, changes in behaviour (or comfort) occur only during the transient period, until the actual temperature equals the see temperature once again. At the end of this transient period, the subject reaches a new thermal equilibrium. In this situation a temperature difference of plus or minus one degree between core and environmental temperatures is probably not large enough to be significant in terms of heat transfer in an air environment.
Physiofogical signals for thermal con&rt
189
PLEASURE AND COMFORT
Perception of thermal stimuli is present at birth even in premature infants (BRUCK et a]., 1962; BRUCK,1968) it is therefore certainly innate and not acquired. It depends on the internal state of the organism, and not only on the nature of the stimulus. Useful stimuli are pleasant, and the usefulness as regards the body is determined by comparison with the set point of the function involved. The result is a modulation of sensation from pleasant to unpleasant (fig. 12.6 and fig. 12.7). The peripheral stimulus combines with the internal information according to the pattern shown for temperature in fig. 12.7. Skin temperature, whether local or mean, is pleasurable very pleasant
pleasant indifferent
unpleasant very unpleasant
21
2L 27 30 33
36
3Y L2 LS
Strmulus ternperature I°C) STIMULUS
Fig. 12.7. Thermal vote of subjects in response to peripheral thermal stimuli in hypo-, normo- and 1977). Below these, results are expressed as a two hyperthermia. Experimental data (from MOWER, dimensional matrix. Framed in heavy black, is the heterogenous population of boxes which correspond to ASHRAE definition of comfort P - pleasant, Z - indifferent, U - unpleasant
in 2 cases out of 9. Pleasure is observable here in transients only, when the stimulus aids the subject’s return to neutrality. Then all stimuli lose their strong pleasure component and tend to be indifferent or unpleasant (middle row). This lack of pleasure may be more apparent than real, because perfect temperature regulation
M. CABANAC
190
is never achieved. The measurement of autonomic responses, metabolic heat production, evaporative heat loss, and vasodilatation shows no interval between warm and MASSONNET, defense and cold defense (JESSEN and CLOUGH,1973; CABANAC 1977). The result is that the affective response is almost an all-or-nothing one between 1969; MOWER$ hypo- and hyperthermia, and a rare state without pleasure (CABANAC, 1976). This is particularly obvious during exercise when humans switch from rest to activity, their behavioural choice is to lower their mean skin temperature rapidly by more than 2 "C (OLESEN et al., 1972; BLEICHERT et al., 1973). When they can affect a limited skin area only (the hand) they may choose to lower its temperature by more than 20 "C (CABANAC et al., 1971). What, then, is thermal comfort? According to the ASHRAE definition, thermal comfort is the absence of unpleasant thermal sensations. In fig. 12.7 comfort includes, therefore, the cases where there is no discomfort (middle column) and the cases where the subject perceives pleasant stimuli (P boxes). Tbis is a non-homogenous group and one should differentiate between dynamic states, generating pleasure from correcting stimuli, and thermal neutrality where all stimuli are indifferent. Pleasure is the motivation for a behaviour which tends to restore neutrality and indifference. As a result, mean skin temperature is kept so constant by behavioural means in humans and rats that it has been considered as the basis for comfort. In a neutral thermal environment and a subject at rest, 33 "Cconstant skin temperature is the result of a thermal equilibrium. It is thus quite remarkable that this skin temperature is not only a source of information, but also an interface between the environment and the thermal core resulting in thermal neutrality.
REFERENCES ASCHOFFJ. (1970), Circadian rhythm of activity and of body temperature, [In:] Physiological and Behavioral Temperature Regulation, eds.: J. D. HARDY, A. P. GAGGEand J. A. J. STOLSpringfield, Ill., Thomas,chapt. 60,pp. 905-919. BALDWINB. A. and INGRAM D. L. (1967), Tle effect of heating and cooling the hypothalamus on behavioural thermoregulation in the pig, J. Physiol. 191, 375-392. BENZINGER T.H. (1963),Peripheral cold and central warm reception, main origins of human discomfort, Proc. Nat. Acad. Sci. 49, 832-839. BITEL J. and HENANE R. (1975), Comparison of thermal exchanges in men and women under neufral and hot conditions, J. Physiol. 250, 475490. BLEICHERT A., BEHLING K., SCARPERI M., and SCARPEXI S. (1973), Thermoregulatorybehavior of man during rest and exercise, mug. Arch. 338,303-312. BBJE0..NIELSEN M., and OLESENJ. (1948), Unders0gelser over betydningen of ensidig straalingsafieling, Boligopvarmningsudvalgets Med. 9, 31 pp. BXUCKK.(1968), Which environmental temperature does the premature infaa prefer? Pediatrics 41,
27-30. BRUCKK., PARMELW A., and BXUCKM. (1962), Neutral temperature range and range of thermal comfort in premature infants, Biol. Neonat. Suisse 4, 32-51. CABANAC M. (1969),PIaisir ou dkplaisir de la sensation thermique et homeothermie, Physiol. Behav. 4,
359-364. CABANAC M., CUNNINGHAM D. J., and STOLWIJK J. A. J. (1971), Thermoregulatory set point during exercise: a behavioral approach, J. comp. Physiol. Psychol. 76, 94-102.
Physiological signals for thermal comfort
191
CABANAC M., HILDEBRANDT G., MASSONNET B., and STREMPEL H. (1976), A study of the nycthemeraF cycle of behavioural temperature regulation in man, J. Physiol. 257, 275-291. CABANAC M. and MASSONNET B. (1974), Temperature regulation during fever: change of set point or change of gain. A tentative answer from a behavioural study in man,J. Physiol. 238, 561-568. CABANAC M. and MASSONNET B. (1977), Thermoregulatory responses as a function of’core temperature in humans, J. Physiol. 265, 587-596. CABANAC M., MASSONNET B., and BEWCHER. (1972), Preferred hand temperature as a fwction of internal and mean skin temperatures, J. Appl. Physiol. 33,699-703. CABANAC M. and SERRES P. (1976), Peripheral heat as a rewardjor heart rate response in the curarized rat, J. comp. Physiol. Psychol. 90, 435-441. CHATONNET J. and CABANAC M. (1965), The perception of thermal comfort, Int. J. Biometeor, 9, 183-193. CHATONNET J., THIERS H., CABANAC M., and PASQUER J. (1966), Sur I’origine de I’impression consciente de confort thermique, Lyon med. 50, 1387-1392. CORBIT J. D. (1969), Behavioral regulation of hypothalamic !emperatwe, Science 166,256-257. CORBIT J. D. and ERNITST. (1974), Specific preference for hypothalamic cooling, J. comp. Physiol. Psychol. 86, 24-27. CORMARECHE-LEYDIER G. and CABANAC M. (1973), Influence de stimulations thermiques de la moelle &pinidresur le comportement thermoregulateur du chien, Pfliig. Arch. 341, 313-324. CUNNINGHAM D. R. and CABANAC M.(1971), Evidence from behavioral rhermoregulatoryresponses of a shift in set point temperature related to the menstrual cycle, J. Physiol. (Paris) 63, 236-238. FANGER P. 0. (1973), Assessment of man’s thermal comfort in practice, Brit. J. Industr. Med. 30, 313-324. FANGER P. O., H~JBJERRE J., and THOMSEN J. 0. B. (1974). Thermal comfort conditions in the morning and in the evening, Int. J. Biometeor. 18, 16-22. FAYT. and SMITH.G. W. (1941), Observations on reflex responses during prolonged period of human refrigeration, Arch. Neurol. Psychiat. 45, 215-222. GAGGE A. P. and GONZALEZ R. R. (1973), Physiological bases of warm discomfort for sedentary man. Arch. Sci. Physiol. 27, A 409-424. GAGGE A. P., STOLWIIK J. A. J., and HARDY J. D. (1967), Comfort and thermal sensations and associatedphysiological responses at various ambient temperatures,J. Envir. Res. 1, 1-20. GONZALEZ R. R., PANDOLP K. B., and GAGGE A. P., (1973), Physiologicalresponsesandwarm discomfort during heat strain, Arch. Sci. Physiol. 27, -4563-571. HALLJ. F. and KLEMM F. K. (1969), Thermal comfort in disparate thermal environments, J. Appl. Physiol. 27, 601-606. HAMMELH. T. (1965), Neurons and temperature regulation, [In:] Physiological Controls and Regulaand J. R. BROBECK, Philadelphia-London, W. W. Saunders Co., tions, eds.: W. S. YAMAMOTO chapt. 5, pp. 71-97. HINDMARSH M. E. and MACPHERSON R. K. (1962), Thermal comfort in Australia, Aust. J. Sci. 24, 335-339. JESSENC. and CLOUGH D. P. (1973), Evaluation of hypothalamic thermosensitivity by feedback signals, PBiig. Arch. 345, 43-59. MARKS L. E. and GONZALEZ R. R. (1974), Skin temperature modifies rhe pleasantness of thermal stimuli, Nature 247, 473-475. MILLERN. E., BAILEY C. J., and STEVENSON J. A. F. (1950), Decreased “hunger” but increased food intake resulting fioh hypothalamic lesions, Science 112, 256-259. MOWERG. D. (1976), Perceived intensity of peripheral thermal stimuli is independent of intetnat body temperature, J. comp. Physiol. Psychol. 90, 1152-1 155. MURGATROYD D. and HARDYJ. D. (1970), Central andperipheral temperature in behavioral thermoregulation of the rat, [In:] Physiological andBehaviora1TemperatureRegulation, eds. :J. D. HARDY A. P. GAGGE, and J. A. J. STOLWJK, Springtleld ,111. Thomas, chapt. 58, pp. 874-891. OLESENS., BASSING J. J.. and FANGERP. 0.(1972), Physiological comfort conditions at sixteen conibinations of activity, clothing, air velocity and ambient temperature, ASHRAE Trans. 78, 199-206.
192
M. CABANAC
OLESEN S., FANGER P. O., JENSEN P. B., and NIELSEN 0. J. (1972), Comfort limits for man exposed to a asymmetric thermal radiation, Proc. Symp. on Thermal Comfort and Moderate Heat Stress, London, CIB Commission W45. PIRLETR. (1962), Die Yerstellung des Kerntemperatur - Sollwertes bei Kaltelbelastung, Pfliig. Arch. 275, 71-94, STEVENS J. C. and MARKSL. E. (1971), Spatial summation and the dynamics of warmth sensation, Perception Psychophysics 9, 391-398. TOPLIFF E. D. L. and LIVINGSTONE S. D. (1970), Thermal comfort in relation to mean skin temperature, Can. J. Physiol. Pharmacol. 48, 98-102. WILMORE D. W., ORCUTT T. W., MASONA. D., ~ ~ ~ P R UB.I A. TT (1975), Afterations in hypothalamic function following thermal injury, J. Trauma 15, 697-703. WiNSLOW C. E. A,, HERRINGTON L.P., and GAGGB A. P. (1937), Physiological reactiom of the human body to various atmospheric humidities, Am. J. Physiol. 120.288-299.
IV. COMFORT, ITS SPECIFICATION AND CONSEQUENCES
This Page Intentionally Left Blank
Chapter 13
DESIGN REQUIREMENTS FOR A COMFORTABLE ENVIRONMENT D. A. MCINTYRE Electricity Council Research Centre, Capenhurst, Chester CH1 6ES, Great Britain.
CONTENTS Warmth sensation Subjective temperature Comfort range Field data Non-thermal factors Age and sex Climate, season and time of day Surroundings Causes of discomfort Radiation draughts Assymmetric radiation Temperature variations Humidity Floor temperature Air movement Non-specific symptoms Discussion APPENDIX
Specification of environmental variables Metabolic rate Clothing insulation Air temperature and speed Mean radiant temperature Plane radiant temperature Vector radiant temperature
Decisions about the thermal environment in a building must be made at the design stage. In recent years, there has been a growirg awareness of how the building structure and climate interact to determine the thermal behaviour of the building interior, and of the important part that environmental considerations have played
196
D. A. MCINTYRE
in determining the very shape of buildings (BANHAM,1969). No longer it is acceptable for an architect to design his building, and then call in the heating engineer only at the last moment; in the past this has led to too many failures to control the environment inside totally imsuitable structures. The design requirements for a comfortable environment must be specified to the design team in physical terms. It would not be reasonable simply to specify that the building should be “comfortable”; the requirements must be set out in physical units of temperature, air speed and so on. Realistically, the requirements must include both an optimum value for the physical variable in question, and a permitted range. The activities of research workers over the past fifty years or so have produced a great deal of information, and this paper summarises what is known about comfort requirements. Response to a potentially uncomfortable stimulus varies greatly between different people, and from time to time for the same person. Further, the strength of objection to an uncomfortable stimulus is very greatly affected by the circumstances, and it is the unfortunate experience of many office managers that the more money that is spent on a building, the greater seems to be the sensitivity of the occupants to any minor discomfort. In any modern office there is always at least one person who is more sensitive to air movement than any anemometer yet devised by man. Experience shows that by far the most important factor determining comfort is the general feeling of warmth. Many other things influence comfort, and they will be dealt with later in this paper, but the first rule remains: get the warmth right and only then worry about any remaining causes of discomfort. They may well have disappeared. WARMTH SENSATION
By comfort, we mean a state of satisfaction, i.e. a person is comfortable if he says he is comfortable. Comfort cannot be predicted from first principles, nor solely from a knowledge of physiology and the physics of heat loss. The prime data on comfort conditions is obtained by exposing subjects to different environments and asking them how they feel. ROHLES and NEVINS (1971) performed what has become the classic experiment on thermal sensation in the environment chamber at Kansas State University; 1600 subjects were exposed to 160 different combinations of temperature and humidity, and they rated their sensation of warmth on the seven point scale shown in fig. 13.1. Part of the data is displayed in fig. 13.2. It can be seen why so many subjects were needed; the variation between people is so great that it is necessary to work with large numbers of subjects to obtain meaningful averages. By performing a regression analysis on the data of fig. 13.2, we can say that the comfort temperature for sedentary people, wearing light clothing of insulation 0.093 m2 K W-’ (0.6 clo), in an air speed of 0.15 m s-’ at 50 % relative humidity (RH), is 25.6 “C. Clearly it is out of the question to repeat the experiment for all combinations of clothing, activity, thermal radiation, humidity and air speed of interest, and some method of predicting comfort temperatures is necessary.
Design requirements for a comfortable environment
Too warm
197
_---___ Comfort range
-------Much t o o cool
Cold
1 A
Bedford Scale
ASHRAE scale
Numerical category
Underlying continuum
Fig. 13.1. The two seven point scales of warmth in current use. The central three categories are conventionally regarded as the comfort range
A more direct method of determining comfort temperature is that developed at the Laboratory of Heating and Air Conditioning at the Technical University of Denmark. In this method a subject sits alone in an environment chamber, and maintains the chamber at his optimum temperature by requesting any necessary changes in temperature. The comfort temperature found by this technique is termed the preferred temperature ; the comfort temperature found by regression analysis of warmth votes, as in fig. 13.2, is termed the neutral temperature. In some circumstances the two temperatures may not be the same (MCINTYRE, 1978~). The variables which affect a person’s thermal sensation may conveniently be divided into two groups, termed the personal and the physical. The personal variables are the activity level and the clothing insulation worn. A more active person requires a lower air temperature for comfort; increased clothing also allows a lower air temperature. There are four physical variables affecting overall thermal sensation: air temperature, mean radiant temperature, air speed and atmospheric humidity. These are described in more detail in the Appendix to this chapter. The engineer concerned with the indoor environment commonly wishes to describe the warmth of a space in terms of the physical variables under his control. Over the. years many indices have been proposed ; each index has normally reflected the interests of its originator, and so tends to have a limited range of application. It must be realised that all those which combine the physical variables into a single index are theoretically inadequate as a predictor of warmth. In particular, the effects of air temperature T, and airspeed interact with clothing temperature T,, e.g. an increase of air speed v increases the feeling of warmth if T, > T,,, and vice versa for T, < Tcl.Since T,, is a function of metabolic rate and clothing insulation, the effects of T, and v cannot be considered in isolation from the personal variables.
198
D. A. MCINTYRE
The only entirely satisfactory way of tackling this problem is to deal with all variables at the same time, as has been done in FANGER’S (1972) comfort equation or GAGGE’S derivation of Standard Effective Temperature (SET) (GAGGE,NISHIand GONZALEZ, 1973). However, these relations are rather complex for normal use, and also make it impossible to describe the warmth of an environment without knowing about the people in it. SUBJECTlVE TEMPbRATURE
To simplify the presentation of the comfort requirement for the indoor environment, we shall use a simple index that gets round some of these difficulties. The index is termed subjective temperature Tsub,and is defined as the temperature of a uniform enclosure (with T, = T,, the mean radiant temperature of the enclosure; v = 0.1 m s- and RH = 50 yo)which would produce the same feeling of warmth as the actual environment. This definition was proposed by PARCEWSI~I and BEVANS(1965). It is implied that the personal variables are the same in the actual and in the standard environment. This definition is similar to that of the effectivetemperature described by GAGGE et al. (1971) but does not imply the use of any particular model of heat loss. Having got a convenient index, the requirements for comfort may be stated as two questions : firstly, what subjective temperature is needed for comfort, as a function of metabolic rate M and clothing insulation I ; secondly, what combinations of thc physical variables will produce that Tsl,b. The subjective temperature required for comfort is a function of metabolic rate and clothing insulation. Where I is expressed in clo (I!), as in most work in this field:
cub = 33.5-31’-(0.08+0.051‘)M.
(13.1)
This equation is plotted in fig. 13.3 and is derived empirically from FANGER’S equation. Over the range of conditions M < 150 W m-2 and I < 0.233 m2 K W-‘ (1.5 clo) the error introduced by the simplification is less than 0.5 K (MCINTYRE, 1976b). Although equation (13.1) has no theoretical basis it is a successful empirical approximation of FANGEK’S equation. FANGER’S equation, in turn, has been shown to fit experimental data well. Where a person is performing external work (W), not all of his free energy production appears as heat. In this case M in equation (13.1) should be replaced by the internal heat production H (W m-2), where H = M- FV = ( l - q ) M , and 7 is the mechanical efficiency of the task performance. The efficiency of a task is normally low, but may rise to 0.1 for tasks such as sawing, or lifting. The maximum found is about 0.2 for uphill walking or cycling, at a high metabolic rate. For many indoor environments, air and mean radiant temperature are approximately equal and the air speed is low; equation (13.1) then simply gives the air temperature required for comfort. For non-standard conditions, we have to provide an expression for Tsubin terms of the physical variables. The following expressions were derived by MCINTYRE (1976b), for v 0.1 m s-l
<
TsLlb= 0.56T,+0.44Tr,
(13.2)
Design requirements for a comfortable environment
18
LO
25
Temperature
30
199
35
("C)
Fig. 13.2.The warmth vote of 1296 subjects as a function of ambient temperature. The area of each circle is proportional to the number of respondents. Data from Nevms et al. (1966) and FANGER (1972). The best fitting regression line is shown
and for Y > 0.1 m s--'
Tsub = 0.44Tr+0.56 [5- (10 17)"~ (5-T')]/[0.44+0.56(10~)~*~].
(13.3)
The equations represent a reasonable compromise between simplicity and accuracy. Equation (13.2) is supported by evidence from several experiments, summarised in MCINTYRE and GRIFFITHS (1975b). The more general form in equation (13.3) is derived by assuming that at speeds above 0.1 m s-l the convective loss varies as the square root of the air speed, and that the clothing surface temperature is 5 K above the subjective temperature. The derivation is similar to that given by MACKEY (1944). Equation (13.3) has been checked against FANGER'S comfort equation over the range of subjective temperatures from 18 to 28 "C,within the range 0 < v < 1 m s-' and (T,--T'I < 6 K; the discrepancy was always less than 0.6 K, and generally better than this. The relationship between Tsrrb and the physical variables is shown in fig. 13.4. Figures 13.3 and 13.4 allow us to predict the comfort temperature for a person with known clothing and activity, and to ensure that the actual set of conditions combine to produce this comfort temperature. The figures have no theoretical basis; they are an empirical simplification of FANGER'S comfort equation, which they fit with excellent accuracy for indoor conditions. Neither the equations, nor FANGER'S comfort equation itself, allow the prediction of thermal sensation at temperatures
200
\ 5 clo
Fig. 13.4. The relation between subjective temperature Tsub, and the physical variabIes of air temperature T,, mean radiant temperature T,, and air speed Y. The differential (T,-T,) is determined by the heating system and building structure
Design requirements for a comfortable environment
201
which are not comfortable. This paper is concerned only with conditions near comfort, and the problem of discomfort is not considered here. Several models have been proposed for the prediction of sensation; FANGER’S “Predicted Mean Vote” (1972) is based on an extension of the comfort equation, and AZERand Hsu (1977) developed a comprehensive thermoregulatory model for predicting thermal sensation. In warm humid conditions thermal sensation and discomfort behave differently; discomfort is well predicted by skin wettedness, which may be estimated by the Standard Effective Temperature. COMFORT RANGE
Thermal comfort is not an exact concept, nor does it occur at an exact temperature. A person may be comfortable over a range of temperatures and if the temperature is changed so that it moves outside this iange the onset of discomfort is not sudden. There is nothing in the physiological system which behaves like a snap action thermostat. A person’s reaction to a temperature which is less than perfect will depend very much on his expectations, his personality and on what else he is doing at the time. All this makes it difficult to talk precisely of a comfort range. The generally accepted convention is to treat thermal discomfort in terms of the scale of warmth sensation, and the comfort range is taken to be the three central categories of the seven point scales shown in fig. 13.1. On the BEDFORD scale, sensations of “comfortably cool”, “comfortable” and “comfortably warm” are taken to imply an acceptable thermal condition; the equivalent central three categories of the ASHRAE scale are also taken as a comfortable range. An individual can only describe his sensation using one of the discrete categories of the scale. However, it is legitimate when performing analysis to treat the scale as continuous, when deriving such statistics as means and standard deviations. On a continuous scale, the comfort range runs from 2.5 to 5.5, i.e. from the transition between categories 2 and 3, to the transition between categories 5 and 6. A great amount of data has been collected giving the variation of warmth vote with temperature. Typically, a subject sits in an environmental chamber for a time, and then records his warmth sensation on a seven point scale. In most work that has been done, each subject attends only once. The data obtained may then be analysed to produce the best regression line. Figure 13.2 shows the regression line for seated, but not inactive, lightly clothed people. The regression line predicts the mean thermal vote of a group of people, as a function of the ambient temperature. In fig. 13.2 the change of temperature corresponding to a change in mean vote from 2.5 to 5.5 is from 18.5 to 27.5 “C,so at first sight the comfort band is a surprising 9 K. However, this ignores the fact that there is a considerable variation between people. At the neutral temperature of the population, i.e. where the mean vote is 4, there are some people who are too hot, and some who are too cold. If the temperature moves above the optimum, the number of people feeling too hot increases rapidly. Knowing the standard deviation of the votes, which is 0.8 of a scale interval, it is possible to construct a curve showing the proportion of people uncomfortable as a function of the mean vote. This is the well known curve of PPD (Predicted
D. A. MC~NTYRE
202
Percentage Dissatisfied) produced by FANGER (1972), shown in fig. 13.5: It shows the PPD as a function of the mean votes of a population of people, all of whom have similar clothing and activity.
R 1
I
I
I
I
I I I
Mean Vote
~
I
- 6 - 5 - L -3 - 2 -1
I
I
C
1
I
I
I
I
I
I
2
3
L
5
6
Temperature deviatior; from optimum ( K
Fig. 13.5. The Predicted Percentage Dissatisfied (PPD) of a group of people as a function of the mean warmth vote of the group. The PPD is also shown as a function of the deviation of the temperature from the optimum value. A temperature range of rt1.5 K about the optimum ensures a PPD of less than 10%. After FANGER (1972)
Dissatisfaction is defined as a vote outside the central three categories. It may be noted that the curve must pass through 50% PPD at a mean vote of 2.5 and 5.5. If the mean vote is 5.5, then 50 yoof the population must be voting more than 5.5, and so are too hot. The PPD curve shows that good temperature control is important when dealing with a large group of people. Even at the optimum temperature, 5 yo are dissatisfied. As the temperature moves away from optimum, the proportion uncomfortable rises rapidly. If we accept 10 % dissatisfied as a working maximum, the temperature must be controlled within a band of 51.5 K about the optimum. The variation in warmth vote of a group of people at a constant temperature has two components. The first component is the “between subjects” variance. This is because people differ from each other; one person may consistently (on average) require a warmer temperature than another. The second component is the “within subjects” variance. A person is not perfectly consistent and will not necessarily feel the same in the same conditions on different occasions. Data from experiments in which each subject attends on several occasions enable us to estimate the two variances separately (MCINTYRE,1978b) ; the results are surprising. The “within subjects” variance is no smaller than the “between subjects” variance, and both have a standard deviation of about 0.8 scale units. There is a real difference between people in their neutral temperature: some people are consistently warmer or cooler. However, this variation is smaller than the scatter produced by unexplained variation within subjects. The causes of this variation have not been investigated in detail. Although we can control the activity of a subject, it does not follow that the metabolic rate is controlled; previous activity, recent food intake and emotional factors may all contribute to a variation in heat production or storage. Thermal sensation will also be influenced by vasomotor tone, which may vary under the influence of internal, as well as external, factors.
Design requirements for a comfortable environment
203
The fact that the “within subjects” variance is similar to the “between subjects” variance means that we can use FANGER’S PPD curve unaltered to apply to a single individual. The ordinate, PPD, now becomes the percentage of occasions that one person will feel too hot or too cold as a function of temperature. At a temperature which is 4.5 K below the optimum, the person’s mean vote, averaged over several occasions when he experiences that temperature, will be 2.5, i.e. on the borderline between comfort and discomfort. We would therefore expect the person to feel too cold on 50% of the occasions that he experiences this temperature. The comfort range of an individual is thus the same as the comfort range of a group of people. Taking 10% dissatisfied as the criterion, the range is f1.5 K on either side of the optimum. FIELD DATA
A group of people sitting in the same temperature, who all have similar clothing, exhibit a spread of warmth sensations. This is demonstrated in fig. 13.5, which shows that at best one can make 95 yo of the population comfortable at one time, and that the proportion uncomfortable increases rapidly as the temperature moves away from the optimum comfort temperature. The implication is that a room having a large number of people must be carefully controlled within a small temperature band around the optimum temperature. In the real world, people do not wear standard clothing, and are free to modify the insulation value. The range of insulation of clothing acceptable in Europe is roughly 0.05 to 0.19 m2 K W-’ (0.3 to 1.2 clo). Inspection of equation (13.1) shows that this implies that an individual may adjust his comfort temperature over a range of 6 K, and it should therefore be possible for most people to achieve comfort indoors by suitable modification of clothing. To find out if this happens, we must turn to the results of field surveys. If people are making behavioural adjustment to compensate for a change in temperature, they will act so as to reduce their thermal discomfort and the slope of the regression line of warmth vote against temperature will be reduced below the value found in laboratory studies. Many studies carried out in environmentchambers have consistently found that the rate of change of warmth sensation on a seven point scale with respect to temperature is 0.33 scale units per Kelvin. Table 13.1 T a b l e 13.1 Regression coefficients of warmth vote on temperature Sample Climate chamber studies (standard clothing) Field studies, extending ovzr weeks Field study, over year Between studies regression (after HUMPHREYS, 1976)
Regresion coefficient (scale units K - l ) 0.33 0.23
0.16 0.05
204
D. A. MCINTYRE
gives the regression coefficientslisted by HUMPHREYS (1976) in his important study of the results of thirty-three separate field surveys. It appears that people in real life compensate slightly for short term changes in temperature, presumably by changing clothing, and so reduce the regression coefficient to 0.23 scale units per Kelvin. This figure is the average of the coefficients obtained for the different studies, and thus represents the rate of change of sensation one would expect to find for a group of people in an office as the temperature changed from day to day. The position changes remarkably when long term effects are considered, and a regression of all the votes from all the surveys on temperature shows them hardly to be affected by temperature at all. This phenomenon is shown in fig. 13.6 where the neutral tem-
Mean air or globe temperature l°Cl
Fig. 13.6. The neutral temperature of a group of people as a function of their thermal experience. Each point represents the result of an entire field survey. The abscissa is the average temperature (1976) recorded during the survey. After HUMPHREYS
perature for each survey is plotted against the mean temperature of the whole survey; each point thus represents the findings of a whole survey. It can be seen that the people have been remarkably successful in adjusting to their surroundings temperature. However, this adjustment seems to take some time to accomplish, since in any particular building a deviation from the neutral temperature occurring from day to day will produce feelings of warmth, cold and perhaps discomfort. The more extreme adjustment at the ends of the range may well take generations to accomplish, since they involve whole patterns of life style and building, as well as of dress. It should not, therefore, be deduced from fig. 13.6 that any temperature will do. Rather, the temperature provided in a building should be consistent with the local climate and culture, and should also be steady from day to day. A temperature that changes unpredictably from day to day does not allow people to wear appropriate clothing. It was noted
Design requirements jbr a comfortable environment
205
by HUMPHREYS (1976) that air conditioned offices, with a very steady temperature, were successful in achieving high levels of satisfaction with the temperature. The very wide range of neutral temperatures shown in fig. 13.6 cannot be completely explained by clothing adjustment alone. Two possibilities suggest themselves; it is possible that people adapt to an extreme temperature in such a way that comfort corresponds to a different physiological state, so that the people who are comfortable at 34 “C have a different bodily state from those who are comfortable at 18 “C. The other possibility is that the meaning of the words of the seven point warmth scale change their meaning with a change in climate, so that people in a warm climate may wish to be “comfortably cool”, while those in a cold climate may wish to be “comfortably warm”; this implies that the neutral temperature of fig. 13.6 may not in fact represent the ideal temperature. There is some evidence to support this point of view (MCINTYRE, 1978c), but the position cannot yet be said to be resolved. NON-THERMAL FACTORS
The first section of this chapter showed how the mean comfort temperature of a group of people can be predicted from a knowledge of their activity and clothing, and how this temperature is modified by thermal radiation and air speed. Several other factors might be expected to affect comfort temperature, and these are dealt with in the following sections. AGE AND SEX
There is little evidence to suggest that healthy old people require a different temperature from young people. Basal metabolic rate decreases with age, but this is fortuitously compensated by a decrease in insensible evaporation (FANGER, 1972). Experimental studies, including people over seventy years of age, have not found any important difference between the comfort temperatures of the old and young people (FANGW, 1973; GRWFITHS and MCINTYRE, 1973; ROHLES and JOHNSON,1972). Surveys of old people admitted to hospital have shown an appreciable number with low body temperatures, and hypothermia in the elderly is recognized as a potential risk to life. It would appear that there is not a systematic shift of comfort or thermoneutral temperature with age. Rather, it seems that the ageing process results in a deterioration of both behavioural and physiological thermoregulation, which, particularly if coupled with poverty or social isolation, may lead to the old person living in too low a temperature, with a consequent fall in deep body temperature. Children have a higher basal metabolic rate per unit surface area than adults, (1977) found and this does seem to lower their comfort temperature. HUMPHREYS that primary school children, aged between 7 and 9 years, generally felt rather warm in school, but there seemed little relationship between mean warmth and mean classroom temperature; however, over a period of some days warmth sensation recommended a temdid follow variations in classroom temperature. HUMPHREYS perature in the range 19-21 “C for general classwork, which encompasses a range of sedentary and standing activities.
206
D. A. MC~NTYRE
There is no a priori reason why men and women should have the same comfort temperature; the metabolic rate per unit surface area tends to be lower for women than for men, and there are other physiological differences in temperature regulation. Nevertheless laboratory experiments in which men and women wear standard clothing have not shown any important difference in comfort temperature between the sexes (ROHLESand NEVINS,1971; FANGER and LANGKILDE, 1975). It has, however, been found that the slope of the regression of warmth vote on temperature is consistently greater for women than for men, i.e. women are more sensitive to temperature changes away from the comfort temperature (MCINTYRE, 1978b). Several surveys have found that in offices women tend to wear lighter clothing than men, particularly in summer (GAGGEand NEVINS,1977), and this may be reflected in a higher comfort temperature. In practice, there is no need to differentiate between the sexes for comfort temperature. CLIMATE, SEASON AND TIME OF DAY
There are definite physiological adaptions to heat and cold in man. People moving to hot climates show changes in sweating response, and there are parallel, though less marked, changes in shivering and vasoconstrictive responses in people exposed regularly to cold. However, there does not appear to be any change in preferred temperature produced by long term exposure to heat or cold. Several studies in Denmark have failed to find any variation of the mean preferred temperature of groups of cold store workers, winter swimmers or dwellers in the tropics (FANGER et al., 1977), though HUMPHREYS (1976) survey of field studies found a de facto alteration of neutral temperature with prevailing climate. The possible implications of this were discussed above. The same findings apply to the effects of season, and laboratory studies have generally failed to find any difference in warmth sensation between times of year (MCNALLet al., 1968). Bodily functions show a definite 24 hour rhythm. Deep body temperature swings by about 0.7 K between night and day, and one would expect this to be reflected in a change in preferred temperature. There is no suggestion that this happens, and determination of preferred temperature throughout a 24 hour period found no significant variation (FANGER et al., 1974). Workers at Kansas State University also found no difference between afternoon and evening (NEVINS et al., 1966). Shortening the time scale still further; it seems that the temperature which the subjects experience before an experiment has no effect on their sensation during the actual session. MCINTYRE (1975b) showed that half an hour's exposure to either 28 or 19 "C had no influence on a person's preferred temperature measured over the next 2 hours. Similarly ROHLES and WELLS(1977) found that a 1 hour exposure to 16 or 32 "C produced no effect on warmth vote when the subjects returned to a 23 "C standard condition; the warmth sensation returned to normal after only 5 minutes. A rather more extreme experiment by GAGGE et al. (1967) exposed nude subjects to a comfortable (29 "C) room for 1 hour, and then transferred them to a cold (17.5 "C)room for 2 hours. After this period they returned to the 29 "C room.
Design requirements Jor a comfortable
eirviroiiiiieiit
207
Their thermal sensation returned to neutral very quickly, long before skin and rectal temperatures had returned to normal. SURROUNDINGS
Many claims have been put forward for the effects of colour on various aspects of man’s feelings. In particular, it has often been suggested that the use of “warm” colours in a room would allow a lower air temperature to be used for comfort. This has never stood up to experimental test. While a person might well prefer “warm” colours in winter, colour does not affect thermal sensation (BENNETT, 1972; FANGER et al., 1977). These results agree with the general finding that man’s thermal sensation as a function of ambient temperature is little affected by extraneous variables. For instance, GRIFFITHS and MCINTYRE (1975) found that mental concentration while performing a difficult reasoning task had no effect on either neutral or preferred temperature; in the experiment questions were asked about both sensation and preference, to deal with the possibility that the subjects might feel the same sensation while concentrating, but might prefer a change in sensation. In another experiment at the Electricity Council Research Centre (MCINTYRE and GRIFFITHS, 1974) it was found impossible to bias the warmth vote of subjects by giving them false information about temperature changes. Recently, however, ROHLESand WELLS(1977) noticed a difference in warmth vote when subjects experienced the same temperature in different chambers. They followed up this observation with a further experiment and found that a chamber furnished with carpet, panelled walls and soft lighting, produced a higher warmth vote from the subjects than the same chamber with bare walls and fluorescent lighting. The difference in neutral temperature, estimated from the results on a nine point warmth scale, was 1.3 K. CAUSES OF DISCOMFORT
So far we have dealt with those factors which influence a person’s feeling of warmth. It is emphasised again that warmth and cold are the dominant sensation influencing thermal comfort. There is, for instance, little profit in worrying about the humidity level if the temperature is wrong. However, once the temperature is right, a person may still be uncomfortable, and we now consider what other factors might cause discomfort. RADIATION DRAUGHTS
It may be that a room which is comfortable over most of its area contains low temperature surfaces which cause discomfort to people nearby, by producing an excessive radiation loss from one side of the body. The most common instance is, of course, the “radiation draught” felt when sitting near a cold glass surface. Two experiments have dealt with this problem. ANQUEZand CROISET(1969) treated the problem directly, with subjects sitting in front of a cold window. OLESEN et al. (1973) worked with unclothed subjects, but extended their results to deal with clothed subjects. It was shown by MCINTYRE (1975a) that if the findings of the two
D. A. MCINTYRE
208
experiments are expressed in terms of the thermal radiation field, they show very good agreement, and the findings can be expressed as follows: Discomfort will be experienced at positions where the plane radiant temperature (for definition see Appendix, p. 215) facing the cold surface is more than 8 K below the mean radiant temperatures in the main part of the room remote from the window. This applies to normal indoor conditions with low air speed and normally clothed people, and it is assumed that the rest of the room is comfortable. This takes into account both general cooling, due to the reduction of mean radiant temperature, T,, near the window, and the additional discomfort from local cooling on one side of the body. This may be written quantitatively for recommended values as F'(Tr--T,) +Fw (Tr- TJ = 8,
(13.4)
where
F, - form factor, plane element to window, Fw - form factor, plane element to wall, Tg - temperature of window, T,,, - temperature of wall. Other surface temperatures are assumed to be equal to T,. For design British conditions, when the outside temperature is -1 "C we may take T, to be 7.5 "C for single glazing; the inside wall temperature is 18 "C,and T, = 22 "C. Figure 13.7 shows the distance from the centre of a rectangular window which just satisfies equation (13.4). 5
I
I
0
Height of window
'
10
Distance from wlndow (ml
Fig. 13.7. The minimumcomfortabledistance from the centre of a rectangLilar window, calculated for single glazing with an outside temperature of - 1 "C
ASYMMETRIC RADIATION
Most radiant heating systems produce a radiant environment which is asymmetric, to a greater or lesser degree. High levels of asymmetry may be disliked by the occupants, and it is necessary to set some sort of standard. The first step, of course, is to ensure that the heating system produces a thermally neutral environment; any problems from asymmetry will be magnified if the temperature is incorrect.
Design requirements for a comjortable environment
209
The degree of asymmetry may be quantified by the vector radiant temperature T,. This quantity may be visualised as the difference between the average surface temperatures of opposite halves of the room. It is defined in the Appendix. Several workers have tackled the problem of relating discomfort to degree of asymmetry 1970; OLESENet al., 1973). It appears that a vector (e.g. MCNALLand BIDDISON, radiant temperature from ceiling heating of 20 K does not worsen the comfort vote of a group of people when they are asked to rate their discomfort in separate experiments. However, the subjects can detect the asymmetry, and it was found by MCINTYRB (1977a) that they tended to blame the heated ceiling for causing discomfort, even though their average discomfort vote did not worsen, when compared with a uniform environment. To be on the safe side, it is recommended that the T, should not exceed 10 K. When designing panel heating, it is sufficient to assume that all unheated room surfaces are at the same temperature. We can then WE& dwmrelations for the T, and Tv at a point in the room: I
where Fci - form factor, plane element to ceiling, F, - form factor, small sphere to ceiling, T, - ceiling temperature, T, - temperature of unheated room surfaces. The form factors are geometrical functions of the positions of the test point and the hot panel. The equations have been solved for a point under the centre of an overhead panel, and the results are shown in fig. 13.8.
50
40
*. 1c
30
I
CL
Cz
Fig. 13.8. The permitted elevation of the temperature Tci (“C) of an overhead square panel above the mean radiant temperature T, (“0in a room, as a function of the size of the panel. The heated panel has dimensions a2(mZ),and is at a height h above head level, in metres 14
- Bioengineering
20
lo
” 05
hic
10
21 0
D. A. MCINTYRE
Heated ceilings are not the only sources of overhead radiation; lighting installations may contribute thermal radiation. Measurements have shown that the illuminance and thermal irradiance are directly related (MCTNTYRE, 1976a), and the illuminances which produce a T, of 10 K are 850 lux for tungsten filament lamps, 4000 lux for warm white fluorescent, 8000 lux for white fluorescent lamps. Radiation from fluorescent lamps is therefore unlikely to cause trouble, but tungsten filament lamps produce a considerable amount of thermal radiation. TEMPERATURE VARlATIOhUS
Experiments which have exposed people to changing temperatures have generally found that the mean warmth vote of a group of subjects corresponds to the instantaneous temperature, and isnot-affected systematically by the direction or rate of and MCINTYRE, 1974; BERGLUND and GONZALEZ, change of temperafure (GRIFFITHS 1977). NEVINS et a’~.(1975;) used a cyclic change of temperature, with amplitude I0 K and a rate of change as high as 18 K h-l. The subjects rated their sensation on a seven point warmth scale, and the rate of change of sensation with respect to temperature was 0.3 scale units per Kelvin, which is only slightly less than the figure for steady temperatures. In general, then, warmth sensation is determined by the instantaneous temperature, with little effect of rate of change of temperature. It is therefore possible to use the PPD of fig. 13.5 when dealing with changing temperatures. In real life, people generally have the opportunity to alter their clothing insulation to compensate for changes in temperature, but there is not much evidence that they do this in the short term (HUMPHREYS, 1976). There is no doubt that occasional changes in sensory stimulation are pleasurable, and this has been demonstrated experimentally by HABER(1958), but for the rather restricted case of finger temperature only. Temperature variation, as a pleasurable stimulus and as a relief from “temperature boredom”, has its advocates (GERLACH, 1974), but there is, as yet, little convincing evidence of its desirability. WYONet al. (1973) measured performance of mental tasks during temperature swings. Small amplitude swings of 2 K, at head level, were found to reduce arousal and depress performance. Larger oscillatiots increased arousal and restored performance to normal. However these were found to be uncomfortable, and WYON concluded that there was little argument for any beneficial effects. HUMIDITY
The effect of humidity on warmth is strongest at high air temperatures, where a person is sweating and consequently above his comfort temperature. The effect of humidity on comfort temperature is small : FAKGER’S comfort equation predicts that a change in relative humidity from 20 yo to 75 % would reduce preferred temperature by 1 K. Such a small effect would be difficult to detect, and indeed experimental investigations of warmth have not shown any significant effects of humidity on warmth at air temperatures below about 23 “C (ROHLESand NEVINS, 1971; ANDERSEN et a]., 1973; MCINTYRE and GRIFFITHS,1975a; MCINTYRE,1978d).
Design requirements for a comjortable environment
21 1
It is quite clear that at high air temperatures high humidities increase both warmth and discomfort. At comfortable air temperatures any effect of humidity on warmth is negligible; this leaves the question of whether humidity may affect comfort in other ways. Low humidities, which may occur in winter in heated buildings, are commonly held to dry the nose and skin, producing complaints of sore throats and headaches. The experimental evidence on the effects of low humidities has not yet formed a consistent body of findings. Inhibition of mucus flow in the nasal passages at low humidities was demonstrated by EWERT(1965), yet careful experiments by ANDERSEN (1974) failed to confirm it. Several experiments have found humidities below 30% to be less comfortable than medium levels (ANDERSEN et al., 1974; CARLETON and WELCH,1971; MCINTYRE,1978d) but other measurements have shown no correlation. There is suggestive evidence that it is a combination of low humidity and atmospheric pollutants (e.g. tobacco smoke) which produces the ill effects, and so well controlled experiments eliminate the discomfort when eliminating pollutants. The question of whether there is an upper limit for humidity at comfortable air temperatures has received less attention. There is no support for the belief in the ill effects of “damp cold”; at low air temperatures, high humidities have no detectable effect on heat loss or warmth. If a person’s activities are variable enough to produce occasional sweating, then the effect of high humidities in inhibiting sweat evaporation may be noticeable. Clothing kept in high humidities will absorb moisture, and the latent heat required to evaporate this after putting on the garment may produce a sense of chill. The arguments against high humidity in buildings lie elsewhere. Prolonged humidities above 70 % encourage the growth of moulds, which, once established, are very difficult to eradicate. There is epidemiblogical evidence to link the incidence of colds with low humidities. Much of the work (GREEN,1974) has been in countries where the outside air temperature, and consequently indoor relative humidity, falls to values far below those found in Britain. The evidence shows an increase in incidence of infection as the RH falls from 40% to 20%. At low humidities the act of walking on a carpet is able to charge a person to a potential of several kV, and this can produce an unpleasant electrostatic,shock when an earthed object is touched. A review by BRUNDRETT (1977) showed that for most carpet materials the problem occurs only at relative humidities below 40 yo. A range of humidities from 40 % to 70 yo will prove acceptable at comfortable air temperatures, both from the point of view of comfort and other considerations. Higher humidities bring the risk of condensation and mould growth. Lower humidities may be felt as uncomfortable, and increase the generation of static electricity. Occasional excursions down to 30 yo should not be troublesome. FLOOR TEMPERATURE
Where people walk barefoot, floor temperature has an important effect on local comfort. OLESEN(1977) investigated people’s perception of floor temperature, the subjects were lightly clothed, and in general thermal comfort. As expected, the
D. A. MCINTYRE
21 2
thermal conductivity of the floor had an important effect on comfort for bare feet; the lower conductivity of the floor material, the wider the range of temperatures that was tolerated. Table 13.2 shows his recommended ranges of floor temperature, based on experiments with 10 minute exposures, and with the limit taken to be 15 yo dissatisfied. Table 13.2 Comfortablefloor temperatures Floor material
_-___ Carpet Wood Concrete
Acceptable temperature range ("C) Bare feet Shod feet 20-28 23-28 26-29
20-28 20-28 20-28
Figures are from OLESEN(1977), and are based on 10 min. exposure. For longer exposures the upper temperatures should be reduced.
When people are normally shod, the flooring material turns out to have little importance. OLESENfound an optimum floor temperature of 23 "C for standing people, with a permitted range from 20-28 "C, corresponding to 8 % uncomfortable; this is a more stringent criterion of discomfort than adopted for bare feet, and suggests that floor temperature is of little importance. The results from NEVINS at al. (1964) were included in the analysis. The experiment by CHRENKO (1957) and general British experience with underfloor heating suggest that OLESEN'S limits err on the high side, and for prolonged occupation the recommended maximum floor temperature is 25 "C, with excursions up to 27 "C permitted. While this is apparently recommendation, it has considerable enginonly a small difference from OLESEN'S eering importance, since it is difficult to provide sufficient heat transfer to a heated room at low floor temperatures. Time of exposure is also important. A floor temperature that is initially pleasantly warm may, over a few hours, produce unpleasant local vasodilation. BURTON(196 3) warned strongly against the danger of overheating the feet. Cold floors are usually associated with vertical gradients of air temperature. Experience indicates that where the general room air temperature is a comfortable 22 "C, a reduction in air temperature at foot level by 3 K or more will produce discomfort. Current research at the Gas Corporation laboratories in England and at the Danish Technical University should provide further data. AIR MOVEMENT
The effect of general air movement on feelings of warmth has been discussed above. In warm conditions, air movement may be employed to reduce discomfort. The disturbance of an air movement above 1 m s-l in itself can cause annoyance, so that people given a free choice of air speed tend to under-compensate for a raised
Design requirements for a comfortable envfronment
213
air temperature, and act to minimise the combined discomfort of warmth and air movement (MCINTYRE,1978e). The more general problem is that of localised air movement iu an otherwise comfortable environment i.e. the problem of draughts. Some sort of air movement is inevitable in an air conditioned or mechanically ventilated building. Strong draughts, above about 0.4 ms-’, are usually very localzied and caused by faulty design or commissioning of the system. Surveys of buildings have demonstrated a number of typical draught producing situations (DICKSON,1977). At the other end of the speed range, it is common to find complaints of draughts where is very little air movement; usually the effect of a low floor temperature has been misinterpreted as a draught. The specification of comfort limits for draughts has proved difficult, and is not yet completely resolved. By virtue of the natural convection loss from the body, a person produces a rising layer of warm air flowing over the head. The speed of the flow depends on the differencebetween body surface temperature and air temperature, but is typically about 0.2 m s-l; the boundary-layer may be several cm in thickness (LEWISet al., 1969). Since draughts have about the same velocity as this protective layer, there is a rather complex interaction between the two. In laboratory experiments on the sensitivity to air movement it is possible either to expose the subject to a free air jet or else to duct the moving air to within a few cm of the skin surface, and consequently lead it through the boundary-layer (e.g. PEDERSEN, 1977). Experiments with a ducted draught have generally found lower discomfort thresholds than those using a free jet. MCINTYRE (1978a) found a 30 min exposure to a draught of 0.25 m s-’ on the face to be rated significantlyworse than one of 0.15 m s-l; ambient and draught temperatures were both 21 “C. However, a speed of 0.2 m s-l was no worse than 0.15 m s-l, and seemed quite acceptable. This agrees with the findings of HOUGHTEN et al. (1938); but according to PEDERSEN (1977) over 20% of his subjects would find this speed uncomfortable. Our general experience of air movement in offices indicates that a speed of 0.2 m s-l at head level is acceptable. In practice, air movement varies with both time and position, and it is not possible to describe a draught by speed and temperature alone. PEDERSEN (1977) found a greatly increased sensitivity on the part of his subjects when the air speed varied regularly; a frequency of 0.3 Hz gave the greatest effect. At this frequency, a draught of temperature 22 “Cand mean speed 0.05 m s-’ was found uncomfortable by 20 % of subjects, but it seems unlikely that people would respond unfavourably to a free jet of this magnitude. NON-SPECIFIC SYMPTOMS
In each of the above sections, the effect on comfort of a single physical factor has been discussed. The effects have been studied in the laboratory by varying the magnitude of the physical factor, and asking about symptoms in the subjects. In the field, the position is reversed. The symptoms are there, and the researcher has to try to identify the cause. We hwe often heard complaints of “stuffiness” or “lack of oxygen” from the occupants of air conditioned offices, in situations where measure-
214
D. A. MCINTYRE
ments of air quality showed nothing wrong. Other symptoms, such as headaches, tiredness or drowsiness, may be blamed on the environmental system. The problem is widespread, and various cures have been proposed, such as negative ionisation or electric fields. These systems have not been generally accepted. Research is needed to investigate these non-specific discomforts, and find if they can be related to environmental factors. DISCUSSION
Most of the recommendations given in this paper imply that if the strength of a potentially uncomfortable stimulus is increased, there comes a point when the subject finds it no longer acceptable. In reality, however, a person’s decision as to whether the stimulus is unacceptable depends on more factors than “strength”. Clothing and physical activity have an important effect : in general they act to reduce a person’s sensitivity to a stimulus, but study of this topic is only beginning (MCINTYFW and GONZALEZ, 1976). There are other less quantifiable factors. Mental activity will determine how much attention will be paid to a thermal stimulus; a minor discomfort will go unnoticed if there are more important matters to be attended to. Judgements of acceptability are greatly affected by context. A gentle stroll across the office produces a self-inflicted draught well in excess of any comfort recommendation, and a coal fire flouts every guideline on discomfort from thermal radiation. POULTON (1977) warns strongly of the dangers of quantitative subjective assessments, pointing out inter a h that subjective ratings of a stimulus depend on the range of stimuli experienced. He gives examples of judgements of “noisiness”, and concludes that whatever the range of noise a person is exposed to, he will rate the loudest noises as “too loud”. This may provide some crumb of comfort to air conditioning engineers who find that standards of expectation seem to rise just as fast as the standard of air conditioning. Thermal sensation judgements behave more reliably than evaluativejudgements, and it is possible to display information on the spread of sensation of a group as in fig. 13.5. The PPD curve makes the assumption that anyone voting outside the central three categories is dissatisfied. This may not be so, and where it has been tested (BERGLUND and GONZALEZ, 1978) the ratings of “unacceptability” did not coincide with the PPD criterion. When asking subjects to give ratings of discomfort or unacceptability, the precise wording of the question and the manner of asking it may have an important effect on the answer. There is the danger that techniques producing consistent results may unwittingly introduce bias. In an experiment on ceiling 1977a) the subjects produced quite different answers to the heating (MCINTYRE, two questions “How uncomfortable are you?” and “Is the heat from the ceiling causing you any discomfort?”. The subjects were ready to ascribe discomfort to the heated ceiling whether or not its temperature was raised. The comfort criteria in this paper have been collected from a wide range of published papers, and so inevitably include different assumptions of what constitutes “unacceptable” discomfort. In the future, attention must be paid to this problem, so that criteria of unacceptability may be set in the context of the experience of the group to which they apply.
Design requirements for a comfortable environment
21 5
APPENDIX SPECIFICATION OF ENVIRONMENTAL VARIABLES The main body of this chapter has described comfort criteria in terms of the various measures of the environment. Further notes on their definition, and tables of typical values are given in this Appendix. METABOLIC RATE
The activity level of a person is described quantitatively as his metabolic free energy production M, expressed per unit body surface area. Some of this energy may be used to perform external work, but most is lost as heat from the body. The metabolic rate of a person may be estimated by measuring his rate of oxygen consumption. A great deal of information is available in the literature on the metabolic rate associated with different activities. Metabolic rates for a list of activities are shown in table 13.3. T a b l e 13.3 Metabolic rates for different activity levels Metabolic rate (W m-2)
Activity Basal metabolic rate Seated at rest Standing at rest General office work Light industrial work Heavy manual work Walking on level ground, at 4 km h-' Walking on level ground at 6 km h-' Walking up 5% slope at 4 km h-' Walking up 15% slope at 4 kin h-'
45 60 65 75 150 (q = 0.1) 250 (7 = 0.1) 140 200 200 (7 = 0.1) 340 (q = 0.2)
Approximate mechanical efficiency (q) shown in brackets. CLOTHING INSULATION
The thermal insulation of clothing acts as a barrier between skin and air, and so allows man to be comfortable in cool air temperatures. The insulation is often expressed in clo units, where 1 clo corresponds to a thermal resistance of 0.155 m2 K W-' . The resistance is measured from the inner to the outer surface of the clothing, and does not include the thermal resistance of the external air film. The clo unit was introduced towards the beginning of the last war (GAGGEet al., 1941) to provide an easily understood descriptive unit: 1 clo unit was the insulation value of the contemporary business suit. Clothing has become lighter since, and table 13.4 gives a list of the clo values of some typical assemblies. The values are obtained from measurements made on a heated mannikin (SEPANNENet al., 1972), or on a volunteer (NISHI et al., 1975). SPRAGUEand MUNSON (1974) give details of a method by which the insulation value of an assembly may be estimated from a description of the individual garments. AIR TEMPERATURE AND SPEED
The description and measurement of air temperature presents few problems, though it is necessary to use an aspirated thermometer to avoid radiation errors. The description of air speed is more difficult, since air movement is notoriously variable in both space and time. An anemometer
D. A. MCINTYRE
21 6
T a b l e 13.4 Insulation values of some clothing outfits Insulation
Clothing ~
_
_
_
(CW
_
Nude Light sleeveless dress, cotton underwear Light trousers, short sleeve shirt Warm, long sleeve dress, full length slip Light trousers, vest, long sleeve shirt Light trousers, vest, long sleeve shirt, jacket Heavy three piece suit, long underwear (I clo
=
0
0.2 0.5 0.7 0.7
0.9 1.5
0.155 m2 K W-I)
a
of the heated sphere type is the most useful, since it gives a non-directional measure of the general cooling effect of the air motion. MEAN RADIANT TEMPERATURE
The mean radiant temperature T, is defined as the temperature of a uniform enclosure with which a small sphere at the test point would have the fame radiation exchange as it would have in the real environment. In a non-uniform enclosure, the Tr varies with position, and the term test point is used to describe the position of interest. When the surroundings consist of i surfaces, each of temperature 6,
T,=
2 FiC
so long as the do not differ by more than about 20 K from ambient; if they do, the weighted average of the fourth powers of the absolute temperatures should be used. Fi is the form factor from the small sphere to the i-th surface, i.e. the fraction of radiation leaving the sphere which reaches the surface. The effect of small high temperature sources on the T, is more easily dealt with by a different method, since a source such as a bar fire or infrared lamp presents neither a simple shape nor uniform temperature, making both F and T difficult to evaluate. If the source contributes an irradiance R(W mP2) at the test point this increases the mean radiant temperature a t that point by AT, = R/(160T,3), where IJ is the Stefan - Boltzmann constant. At normal temperatures it is sufficient to use the approximation
AT,
= 0.043R.
This is accurate for small A Tr ; the error increases to about 5 + ' ,! for A Tr = 20 K. The value of the irradiance may be obtained from the manufacturer's data for heating panels, or measured with a radiation thermopile. PLANE RADIANT TEMPERATURE
The plane radiant temperature Tpris the surface temperature of the inside of a uniform hemisphere which produces the same irradiance on a small plane element at the test point as exists in the real environment. The element lies in the basal plane of the hemisphere. Plane radiant temperature is a function of direction as well as position; the direction is specified by the outward normal
Design requirements for a confortable environment
217
to the test element. The irradiance R(W is the total radiant energy falling on a surface per unit area; it is not the radiant exchange, and is independent of the temperature of the surface. When the surroundings consist of i surfaces of temperature T i ,
As before, the fourth power law must be used if the temperatures differ by more than about 20 K from ambient. Fpi is the form factor from a small plane element to the i-th surface. The plane radiant temperature is simply related to the irradiance R by
R
= GT;.
The irradiance and hence the plane radiant temperature, is readily measured using a net radio. meter fitted with a reference cavity (see fig. 17.2). VECTOR RADIANT TEMPERATURE
At other times it is necessary to describe the asymmetry of the radiation environment. Consider a small plane element at the test point. The plane radiant temperature seen by the front surface is denoted by Tdr, and that by the rear surface by Tj;.If the element is now pointed in different directions it is found that (TLr - T;;) has a maximum value in a unique direction. The normal to the element is now pointing along the direction of flow of radiant energy, and the value of (Td, - TL:)is the vector radiant temperature T,. This quantity has both magnitude and direction, and obeys the laws of vector addition. The vector radiant temperature is the difference between plane radiant temperatures for opposite directions and is therefore easily visualised as the average surface temperature of one half of a room minus the average of the other half. The equivalent of vector radiant temperature in units of power is the radiation vector, alternatively, known as the differential radiant flux or net radiation (e.g. CENA,1974). The radiation vector R,(Wm-2) is the difference in irradiance on opposite sides of a plane element. T , = RJ(4 d T:)
.
At room temperature, T, = 0.17 R,. The radiation vector is easily measured with a net radiometer. The above concepts are discussed in detail by MCINTYRE(1974) and also in chapter 17.
REFERENCES I., LUNDQVIST G., R., and PROCTOR D. F. (1973), Human perception of liumidity under ANDERSEN four controlled conditions, Arch. Environ. Health 26, 22-27. ANDERSEN I., LUNDQVIST G . R., JENSEN P. L., and PROCTOR D. F. (1974), Human response to 7 8 hr exposure to dry air, Arch. Environ. Health 29, 319-324. ANQUEZ I. and CROISET M. (1969), Thermal comfort requirements adjacent to cold walls - application to glazed opening, NBS-TN-710-4, May 1972, U.S. Department of Commerce, Washington D.C. AZERN. Z . and Hsu S. (1977), The prediction of tl~ermalsensation from a simple model of human physiological regulatory response, ASHRAE Trans. 83, 88-102. R. (1969), Architecture of the Well-Tempered Environment, Architectural Press, London. BANHAM BENNETTC. A. (1972), What's so hot about red? Human Factors 14, 149-154. BERGLUND L. G . and GONZALEZ R. R. (1978), Application of acceptable temperature drifts to built environments as a mode of energy conservation, ASHRAE Trans. 84, 110-121. BILLINGTON N. S. (1948), The warmth of floors - a physical study, J. Hyg. (Camb.) 46, 445-450. BRUNDRETTG. W. (1976), A review ofthe factors influencing electrostatic shocks in offices, J. Electrostatics 2, 295-315.
218
D. A. MCINTYRE
BURTON A. C. (1963), The pattern of response to cold in animals and the evolution of homeothermy, [In:] Temperature: Its measurement and control in science and industry, Vol. 3, Part 3, Biology and Medicine, ed.: J. D. HARDY, Reinhold, New York. CARLETON W. M. and WELCHB. E. (1971), Fluid balance in artificial environments, USAF School of Aerospace Medicine, NASA Report CR-114977. CENAK. (1974), Radiative heat loss from animals and man, [In:] Heat Loss from Animals and Man, eds.: J. L. MONTEITH and L. E. MOUNT,Butterworths, London. CHRENKO F. A. (1957), The effects of the temperature of thefloor surface and of the air on the thermal sensations and the skin temperature of the feet, Br. J. Industr. Med. 14,13-20. DICKSON D. J. (1977), Air movement in of$ces, ECRC/N1012, Electricity Council Research Centre. EWERTG. (1965), On the mucusflow rate in the human nose, Acta Otolaryng. (Stockholm), 200, supp. 21. FANGER P. 0. (1972), Thermal Comfort, McGraw-Hill, New York (Polish translation: Komfort cieplny, Arkady, Warszawa 1974). FANGER P. 0. (1973), The inzuence of age, sex, adaptation, season and circadian rhythm on thermal comfort criteria for man, Annexe 1973-2 au Bulletin de 1'Institut International du Froid, pp. 91-97, Proceedings of Meeting of Institut International du Froid (Commission El), Vienna. N. O., and JERKING E. (1977), Can colour and noise influence man's thermal FANGER P. O., BREUM co&fort ? Ergonomics 20, 11-18. FANGER P. 0. H~JBJERRE J., and THOMSON J. 0. B. (1977), Can winter swimming cause people to prefer lower room temperatures? Int. J. Biometeor. 21, 44-50. FANGER P. 0. and LANGKILDE G. (1975), Inter-individual direrences in ambient temperature preferred by seated persons, ASHRAE Trans. 81, 140-147. FANGER P. O . , OSTBERG 0. McK., NICHOLL A. G., BREUMN. O., and JERKING E. (1974), Thermal comfort conditions during day and night. Bur. J. Appl. Physiol. 33,255-263. GAGGE A. P., BURTON A. C., and BAZETT H. C. (1941), A practical system of unitsfor the description of the heat exchange of man with his thermal environment, Science 94,428-430. GAGGE A. P. and NEVINS R. G. (1977), EfSect of energy conservationguidelineson comfort acceptability and health, Symposium: Thermal Analysis - Human Comfort - Indoor Environments, National Bureau of Standards Special Publication 491, US Department of Commerce, Washington. GAGGE A. P., NISHIY., and GONZALEZ R. R. (1973), Standard Effective Temperature, Conceil International du Batiment, Commission W4.5. Symposium: Thermal Comfort and Moderate Heat Stress, Watford 1972, HMSO, London. J. A. J., and HARDY J. D. (1967), Comfort and thermal sensations and assoGAGGE A. P., STOLWIJK ciated physiological responses at various ambient temperatures, Environmental Research 1, 1-20. 6 A G G E A. P., STOLWIJK J. A. J., and NISHIY.(1971), An effective temperature scale based on a simple model of human physiological regulatory response, ASHRAE Trans. 77, 247-262. GERLACH K. A. (1974), Environmental design to counter thermal boredom, J. Arch. Res. 3, 15-19. GKEENG. H. (1974), The effect of indoor relative humidity on absenteeism and colds in schools, ASHRAE Trans. 80, 131-141. GRIFFITHS I. D. and MCINTYRE D. A. (1973), The balance of radiant and air temperaturesfor warmth in older women. Environmental Research 6 , 382-388. GRIFFITHS I. D. and MCINTYRE D. A. (1974), Sensitivity to temporal variations in thermal conditions, 4.1 Ergonomics 17, 499-507. GRIFFITHS I. D. and MCINTYRE D. A. (1975), The effect of menfal efort on subjective assessments of warmth, Ergonomics 18, 29-33. HABERR. N. (1958), Discrepaizcyfrom adaptation level as a source of afect, J. Exp. Psychol. 56,370315. HOUGHTEN F. C., GUTBERLETC., and WITKOWSKI E. (1938), Draft temperatures and velocities in relation to skin temperature and feeling of warmth, Trans. ASHVE 44,289-308.
Design requirements for a comfortable environment
219
HUMPHREYS M. A. (1976), Field studies of thermal comfort compared and applied, Building Services Engineer 44, 5-23 and 27. HUMPHREYS M. A. (1977), A study of the thermal comfort of primary school children in summer, Building and Environment 12, 231-239. LEWISH. E., FOSTER A. R., MULLEN B. J., Cox R. N., and CLARKR. P. (1969), Aerodynamics of the human micro-environment, Lancet, 1273-1277. MACKEY C. 0. (1944), Radiant heating and cooling, Part 11, Cornell University Engineering Expt. Station Bulletin No. 33. D. A. (1974), The thermal radiation field, Building Science 9, 247-262. MCINTYRE D. A. (1975a), Radiation draughts, Building Services Engineer 43, 136-139. MCINTYRE D. A. (1975 b), Determination of individual preferred temperatures, ASHRAE Trans. 81, MCINTYRE 131-139. MCINTYRE D. A. (1976a), Radiant heatfiom lights, Lighting Research and Technology 8, 121-128. D. A. (1976b), Subjective temperature. A simple index of warmth, ECRC/M916, Electricity MCINTYRE Council Research Centre. D. A. (1977a), Sensitivity and discomfort associated with overhead thermal radiation, MCINTYRE Ergonomics 20, 287-296. MCINTYRE D. A. (1977b), Overhend radiation and comfort, Building Services Engineer 44, 226-234. D. A. (1978a), Draughts on the face, ECRC/M1119, Electricity Council Research Centre. MCINTYRE D. A. (1978b), Seven point scales of warmth, Building Services Engineer 45, 215-226. MCINTYRE D. A. (1978c), Three approaches to thermal comfort, ASHRAE Trans. 84, 101-109. MCINTYRE D. A. (1978d), Response to atmospheric humidity at comfortable air temperatures: a comMCINTYRE parison of three experiments, Ann. Occ. Hyg. 21, 177-190. MCINTYRE D. A. (1978e), Preferred air speeds for comfort in warm conditions, ASHRAE Trans. 84, 246-277. MCINTYRE D. A. and GONZALEZ R. R. (1976), Man’s thermal sensitivity during temperature changes at two levels of clothing insulation and activity, ASHRAE Trans. 82, 219-233. D. A., and GRIFFITHS I. D. (1974), The effect of dummy and real controls on thermal MCINTYRE comfort and perceived warmth, ECRC/N761, Electricity Council Research Centre. D. A. and GRIFFITHS I. D. (1975a), Subjective responses to atmospheric humidity, EnMCINTYRE vironmental Research 9, 66-75. D. A. and GRIFFITHSI. D. (1975b), The effect of uniform and asymmetric radiation on MCINTYRE comfort, Clima 2000. 111-03, pp. 1-22, 6th International Congress of Climatistics, Milan. R. E. (1970), Thermal and comfort sensations of sedenlary persons MCNALLP. E. and BIDDISON exposed to asymmetric radiant fields, ASHRAE Trans.‘76, 123-176. MCNALLP. E., RYANP. W.,and JAAXJ. (1968), Seasonal variation in comfort conditions for college-age persons in the Middle West, ASHRAE Trans. 74, 1-9. A. M. (1967), Effects offloor surface temperature on comfort. Part IV NEVINS R. G. and FEYERHEM Coldfloors, ASHRAE Trans. 73, 1-8. Y.,and GAGGEA. P. (1975), Efects of changes in ambient NEVINSR. G., GONZALEZ R. R., NISHI temperatureand [everof humidity on comfort and thermalsensations,ASHRAE Trans. 81,169-182. NEVINS R. G., MICHAELS K. B., and FEYERHEM A. M. (19641, The effect of$oor surface temperature on comfort, Parts I and 11, ASHRAE Trans. 70,2943. F. H., SPRINGERW., and FEYERHEM A. M. (1966), A temperature-humidity NEVINSR. G., ROHLES chart jbr thermal comfort ojseatedpersons, ASHRAE Trans. 72,283-291. R. R., and GAGGEA. P. (1975), Direct measurement oj-clothing heat transfer NISHIY.,GONZALEZ properties during sensible and insensible heat exchange with the thermal environment, ASHRAE Trans. 81, 183-199. OLESEN B. W. (1977), Thermal comfort reqmirementsforfloors, Proceedings of Conference of Institut International du Froid, Belgrade, pp. 337-343. S., FANGER P. O., JENSEN P. B., and NIELSEN0. J. (1973), Comjbrt limits for men exposed OLESEN to asymmetric thermal radiation, Conceil International du Batiment, Commision W45, Symposium on Thermal Comfort and Moderate Heat Stress, HMSO, London.
220
D. A. MCINTYRE
PARCEWSKI K. I. and BEVANS R. S., A new method of rating the quality of the environment in heated spaces, ASHRAE J. 7, 80-86. PEDERSEN C. J. K. (1977), Komfortkrav ti1 lufbevaegelser i rum, Thesis, Danish TechnicalUniversity. POULTONE. C. (1977), Quantitative subjective assessments are almost always biased, sometimes completely misleading, Br. J. Psychol. 68, 409-425. M. A. (1972), TIermal comfort in the elderly, ASHTrans. 78, 131ROHLESF. H. and JOHNSON 137. R. G. (1971), The nature of thermal comfort for sedentary man, ASHRAE ROHLESF. H. and NEVINS Trans. 77, 239-246. F. H. and WELLSW. V. (1977), The role of environmental antecedents on subsequent thermal ROHLES comfort, ASHRAE Trans. 83, 21-29. O.,MCNALLP. E., MUNSON D. M., and SPRAGUE C. H. (1972), Thermal insulating values SEPANNEN for typical indoor clothing ensembles, ASHRAE Trans. 78, 120-123. C.H. and MUNSOND. M. (1974), A composite ensemble method for estimating thermal SPRAGUE insulating values of clothing, ASHRAE Trans. 80, 120-129. T. L., KJERULF-JENSEN P., and FANGER P. 0. (1973), The efects of WYOND. P., ASGEIRSDOTTER ambient temperature swings on comfort, performance and behaviour, Arch. Sci. Physiol. 27, A441-A458.
Chapter 14
PREDICTION OF LOCAL DISCOMFORT FOR MAN P. 0.FANGER Laboratory of Heating and Air Conditioning, Technical University of Denmark, 2800 Lyngby, Denmark.
CONTENTS
Introduction Asymmetric radiation Draughts Floor temperature Vertical air temperature gradients Local discomfort and energy requirements in buildings
INTRODUCTION
The main purpose of most heating and air conditioning systems is to provide thermal comfort for human beings. For the design and operation of such systems and for the thermal design of buildings it is essential to establish quantitative comfort requirements. Thermal comfort is defined here as “that condition of mind which expresses satisfaction with the thermal environment” (ASHRAE, 1974). A first requirement for comfort is that a person feels thermally neutral for the body as a whole, i.e. that he does not know whether he would prefer a higher or lower ambient temperature level. Man’s thermal neutrality depends on his clothing and activity and on the following environmental variables: the mean air temperature, mean radiant temperature, mean air velocity, and water vapour pressure in the ambient air. The thermal effects of clothing, activity and environmental variables on man have been studied quite intensively during the last decade (FANGER, 1973; NEVINS et al., 1973; GAGGEet al., 1971) and et al., 1966; MCNALLet al., 1967; ROHLES combinations which will provide thermal neutrality may be predicted from the comfort equation and the corresponding comfort diagrams (FANGER, 1973).
222
P. 0.FANGER
But thermal neutrality as predicted by the comfort equation is not the only condition for thermal comfort. A person may feel thermally neutral for the body as a whole, but he might not be comfortable if one part of the body were warm and another cold. It is therefore a further requirement for thermal comfort that no local area of discomfort exists at any part of the human body. Such local discomfort may be caused by an asymmetric radiant field, by a local convective cooling (draught), by contact with a warm or cool floor, or by a vertical air temperature gradient. Each of these cases will be dealt with in the following sections, and limits for avoiding local discomfort will be discussed. ASYMMETRIC RADIATION
Asymmetric radiant fields in spaces may be caused by cold windows, uninsulated walls, or by warm or cool panels in the wall or in the ceiling. Limits for asymmetric radiation from vertical surfaces were studied by OLESFN et al. (1972). Proposing an acceptable limit of 5 % of sedentary thermaIIy neutral subjects feeling uncomfortable due to such asymmetry, he recommended the following formula for estimating the corresponding temperature difference between the temperature of a vertical surface and the mean radiant temperature: -2.4 -11.61<
A T F < 3.9+11.61,
(14.1)
where I is the insulation of clothing in m2 K W-’ or -2.4 --1.81’< A T F < 3.9+1.81’,
(14.2)
where I’ is the insulation of clothing in clo units, F is the angle factor between a sedentary person and the vertical radiant area, which can be found from angle 1973); and A T is the temperature difference between the factor diagrams (FANGER, radiation source and the mean radiant temperature. This formula is in reasonable agreement with the results obtained in similar experiments by MCNALL and BIDDISON (1970) and by MCINTYRE and GRIFFITHS (1975). The formula indicates that a higher degree of asymmetry is permissible when wearing heavy clothing than when light clothing is worn. It allows larger temperature deviations for vertical surfaces than those which occur normally in residential or commercialbuildings, but it should be noted that in OLESEN et al.’s experimentsas well as in those by MCNALLand BIDDISON and by MCINTYRE and GRIFFITHS, the change in temperature of the vertical surface was balanced by an opposite change of the other surfaces to maintain a constant mean radiant temperature. This is somewhat unrealistic; in practice the air temperature is usually changed to balance a low or high surface temperature. Further studies with normally clothed sedentary subjects are recommended to establish limits for acceptable asymmetric radiation, when the air temperature is modified to balance the effect of warm or cool vertical surfaces. Recent (as yet unpublished) studies at the Technical University of Denmark indicate that man is much more sensitive to thermal asymmetry when exposed to
Tliermal discomfort
223
overhead warm radiation (e,g. from heated ceilings). Accepting 5 yo feeling uncomfortable (due to warm head or cool feet), the followiiig formula was found for sedentary, normally clothed subjects : IATFJ< 2. (14.3) t.
This result agrees quite well with the limit recommended by CHRENKO (1953), although he predicts a higher percentage of people feeling uncomfortable.
DRAUGHTS
A draught is defined as an unwanted local convective cooling of the body. It is perhaps the most common reason for complaints in ventilated spaces. As mentioned earlier, the mean air velocity around the body influences the ambient temperature necessary for thermal neutrality for the body as a whole (BURTON et al., 1975; OLFSEN et al., 1972; OSTERGAARD et al., 1974; ROHLESet al., 1974). However, in spite of thermal neutrality, local velocities can provide an unwanted cooling (i.e. draught) of some parts of the body; the neck and the ankles seem to be the most sensitive parts of normally clothed persons. Widely varying limits for velocities are given in standards in different countries, but very few experimental draught studies have been reported in the literature. The most comprehensive earlier study was performed by HOUGHTEN (1938) who studied ten male subjects exposed to constant local velocities at the back of the neck and at the ankles. In practice, however, the air velocity in spaces is never constant but fluctuates, in a random way around a mean value. Extensive draught studies have recently been performed at the Technical University of Denmark (FANGERand PEDERSEN, 1977). More than 100 college-age students have been involved in experiments in which subjects were exposed at the neck and ankles to periodically fluctuating and uniform air flows with different mean velocities, with different amplitudes and frequencies of the fluctuating velocity, and with different air temperatures. Based on the subjective reactions of the subjects, it has been possible to establish a mathematical model which predicts the percentage of unccmfortable persons due to the draught, as a function of the above factors. It was found that fluctuation frequencies around 0.3-0.5 Hz were most uncomfortable but thslt the most typical frequencies in practice seem to be lower than 0.1 Hz. The maximum velocity is also important; people accept a lower mean velocity at higher maximum velocities. Figure 14.1 shows an example of the results. The diagram applies for a frequency of the fluctuating velocity equal to 0.1 Hz, and a maximum veldcity twice as high as the mean velocity. These values are believed to be typical, but further field studies are recommended to study the characteristics of fluctuating velocities in different types of spaces. Further laboratory studies, where subjects are exposed to the more random velocity fluctuations which exist in spaces in practice, are also recommended. Figure 14.1 shows the percentage of people feeling uncomfortable as a function of the mean velocity, and the difference between the local air temperature and the air temperature which is felt as neutral for the body as a whole. This neutral temperature
224
P. 0.FANGER
depends on the clothing, and two scales of the abscissa have been calculated for clothing typical of summer and winter. If the temperature during winter is kept lower than neutral, to save energy, it is especially important to keep the mean velocity at a low level by proper design of the air distribution in the space. If low velocities are not maintained in the occupied zone it is likely that a higber ambient temperature will be required (involving higher energy consumption) to decrease complaints of draughts. x -30%
'OYo fortable due to draught as a function of the local mean
' ;/Ie-----
--
_/I*
c : s
,
I
l
I
,
I
fluctuating velocity
-
0.1 Hz, m a . velocity = twice the mean velocity)
The allowable mean velocities, according to fig. 14.1, are lower than HOUGHTEN'S (1938) results and lower than limits recommended in existing comfort standards (ASHRAE, 1974; Deutsche Normen, 1960). But it should be .kept in mind that HOUGHTEN exposed his subjects to a constant velocity. It should also be noted that fig. 14.1 is based on experiments in which the subjects were exposed to an air flow from a horizontal duct situated very close to the back of the neck. In practice the natural convection current rising along the human body will interfere with the air flow in the surrounding space. The parts of the curves below 0.1 m s-' are therefore probably unrealistic, and should be used with caution. Further studies on the effect of these very low fluctuating velocities are recommended. The subjects were found to be significantly less draugbt-sensitive on their (unclothed) ankles than on the neck. No difference between the draught sensitivity of men and women was found. FLOOR TEMPERATURE
Due to the direct contact between the feet and the floor, local discomfort of the feet can often be caused by too high or too low a floor temperature. Studies on comfort limits for floor temperatures have recently been performed at the Technical University of Denmark by OLESEN(1975), who found the following main results. For floors occupied by people with bare feet (in swimming halls, gymnasia, dressing rooms, bathrooms, bedrooms, etc.) the flooring material is important. From the results of experiments on 16 subjects, and heat transfer theory, OLESEN
Thermal discomfort
225
estimated optimal temperatures. He recommended temperature intervals (given in table 14.1) for a number of typical hooring materials. For 10 minutes' occupancy about 10 % of persons can be expected to experience discomfort at the optimal floor temperature, while for longer occupation less than 15% are expected to be uncomfortable within the recommended temperature interval. To save energy, and to eliminate a desire for higher ambient temperatures caused by cold feet, flooring materials with a low contact coefficient (cork, wood, carpets) or heated floors should be chosen (OLESEN, 1977). Table 14.1 Comfortable temperatures ("C) of floors occupied by people with bare feet -.
Flooring material
Pinewood floor Oakwood Aoor PVC-sheet with felt underlay on concrete Hard linoleum on wood 5 mm tesselated floor on gas concrete Concrete floor Marble
Optimal floor temperature for 1 min 10 min occupancy occupancy
Recommended
floor temperature interval
25 26
26 26
22.5-28 24.5-28
28 28
27 26
25.5-28 24 -28
29 28.5 30
27 27 29
26 -28.5 26 -28.5 28 -29.5
For floors occupied by people with normal indoor footwear the flooring material is without significance. Based on his own experiments and a re-analysis of the results of NEVINS (1964a, 1964b, 1967). OLESENfound an optimal temperature of 25 "C for sedentary and 23 "C for standing or walking persons. At the optimal temperature 6% of the occupants felt warm or cold discomfort at the feet. If one accepts up to 8 % uncomfortable, the floor temperature should be within the interval 22-30 "C for sedentary and 20-28°C for standing or walking persons. At floor temperatures below 20-22 "C the percentage of people experiencing cold feet increases rapidly (OLESEN,1975). Although heavier clothing can provide thermal neutrality at a lower ambient temperature, there is a risk of cold discomfort at the feet if they are not protected correspondingly by well-insulated footwear. If normal light footwear is worn, higher ambient temperatures will be required to counteract the coldness at the feet. VERTICAL AIR TEMPERATURE GRADIENTS
In most spaces in buildings the air temperature is not constant from the floor to the ceiling; it normally increases with the height above the floor. If the gradient is sufficientlylarge local warm discomfort can occur at the head and/or cold discom15
- Bioengineering
226
P. 0. FANCER
fort can occur at the feet, although the body as a whole is thermally neutral. Little information on this subject has been published, but preliminary results from studies at the Technical University of Denmark by SCHBLER (1976), and results by MCNAIR (1974) and ERIKSSON (1975) indicate that the risk of local discomfort is negligible provided that the air temperature difference between head and feet level is less than 2-3 K. LOCAL DISCOMFORT AND ENERGY REQUIREMENTS IN BUILDINGS
To achieve true comfort the temperature should be maintained at a level which enables thermal neutrality for the body in general, and no local thermal discomfort should occur, e.g. caused by radiant asymmetry, local air velocities, warm or cold floors, or air temperature gradients. If local cold discomfort in a space occurs during winter, it is likely that the occupants will require a higher ambient temperature level. This will decrease the local discomfort but will at the same time cause a slightly warm thermal sensation for the body as a whole. The higher the temperature the higher the energy consumption (about 5-15% increase per “C). For both ccmfort and energy conservation it is thus essential that the building, as well as the hating and air conditioning system, be desjgncd s o that occurrence of local cold discomfort on the body is unlikely. For summer conditioiis the same applies for local warm discomfort. On the otlier hand, local heating or cooling of the body may provide a means of energy conservation during the winter or summer respectively. This may be considered primarily under conditions where deviations from optimal comfort would be acceptable. REFERENCES American Society of Heating, Refrigerating and Air-conditioning Engineers (1974) Thermal comfort conditions, ASHRAE Standard 55-74, New York. BURTON D. R., ROBEBON K. A., and NEVINSR. G. (1975), Tlze effect ojtemperature on preferred air velocity jbr sedentary subjects dressed in shorts, ASHRAE Trans. 81, 157-168. CHRENKO F. A, (1953), Heaiedceilingsandcomjbrt, J. of the Inst. of Heating and Ventilating Engineers 20, 375-396 and 21, 145-154. Deutsche Normen DIN 1946, Luftungstechnische Anlagen, April 1960. ERIKSSON H.-A (1 973, Virme och ventilation i traktorhyfter, S-25 Ultuna, Jordbruks-tekniska Institut. P. 0. (1973), Thermalcomfort, New York, McGraw-Hill Book Co., 244p (Polish translation: FANGER Komfort cieplny, Arkady, Warszawa 1974, 231 p.). FANGER P. 0. and PEDEXSEN C. J. K. (1979, Discomfort due to air velocities in spaces, Proc. of the meeting of Commission El (Air Conditioning) of the International Institute of Refrigeration, Belgrade. A. P., STOLWIJK J. A. J., and NISHI Y . (1971), An effective temperature scale based on a simple GAGGE model of human physiological regulatory response, ASHRAE Trans. 11,241-262. HOUGHTEN F. C . (1938), Draft temperatures and velocities in relation to skin temperature and feeling of’warmth, ASHVE Trans. 44, 289-308. MCINTYRE D. A. and GRIFFITHS I. D. (1975), The effect of unijorm and asymmetric radiafion OIZ comfbrf, Proc. of the 6th International Congress of Climatistics “Clima 20W, Milan.
Thermal discomfort
227
MCNAIRH. P. (1974), A preliminaryjfurther study of the subjective effectsof vertical air temperature gradients, London, British Gas Corporation, Project 552. J., ROHLES F. H., NEVINS R. G., and SPRINGER W.(1967). Thermal comfort MCNALLP. E., Jr., JAAX (thermally neutral) conditionsfor three levels of activity, ASHRAE Trans. 73, 1-14. MCNALLP. E., Jr. and BIDDISON R. R. (1970), Thermal and comfort sensations of sedentary persons exposed to asymmetric radiant fields, ASHRAE Trans. 76, 123-136. NEVINS R. G., MICHAELS K. B., a n d F E m m m A. M. (1964a), The effect offloor surface temperature on comfort, Part I : CoZlege age males, ASHRAE Trans. 70, 29-36. R. G., MICHAELS K. B., and FEYERHERM A. M. (1964b), The effect offloor surface temperature NEVINS on comfort, Part 11: College age females, ASHRAE Trans. 70, 3743. NEVINSR. G., ROHLESF. H., SPRINGER W., and FEYERHERM A. M. (1966), Temperature-humidity chart for thermal comfort of seatedpersoizs, ASHRAE Trans. 72,283-291. A. M. (1967), Effect offloor surface temperature on comfort, Part IV: NEVINS R. G. and FEYERHERM Coldfloors, ASHRAE Trans. 73, 1-8. OLESENB. W. (1975), Termiske komfort krav ti1 gulve, (Thermal comfort requirements for floors). Ph. D.-Thesis, Laboratory of Heating and Air Conditioning, Technical University of Denmark. OLFSENB. W. (1977), Thermal comfort requirements for floors occupied by people with bare feet, ASHRAE Trans. 83, 41-51. OLESENS., FANGER P. O., JENSEN P. B., and NIELSEN 0. J. (1972), Comfort limits for man exposed to asymmetric thermal radiation, Proc. o f CIB Symposium on Thermal Comfort, London, Building Research Station. P. 0.(1972), Physiological conrfort conditions at sixteen combiOLESENS., BASING3. J., and FANGER nations ofactivity, clothing, air velocity and ambient temperature, A S H R A E Trans. 78, 199-206. OSTERGAARD J., FANGER P. O., OLESEN S., and LUNDMADSEN Th. (1974), The effect on man’s comfort of a uniform airflowfuom diferent directions, ASHRAE Trans. 80, 142-157. ROHLESF. H., Jr., WOODSJ. E., and NEVINSR.G. (1973), The influence of clothing and temperature on sedentary comfort, ASHRAE Trans. 79, 71-80. F. H., Jr., WOODS J. E., and NEVINS R. G. (1974), The effects of air movement and temperature ROHLES on the thermal sensations of sedentary man, ASHRAE Trans. 80, 101-119. SCHBLER M. (1 9761, Vertikale temperaturgradienters indflvdelse p d menneskets termiske komfort, (The influence of vertical temperature gradients on human comfort), M. S. Thesis, Laboratory of Heating and Air Conditioning, Technical University of Denmark.
This Page Intentionally Left Blank
Chapter 15
THE DEPENDENCE OF COMFORTABLE TEMPERATURES UPON INDOOR AND OUTDOOR CLIMATES* M. A. HUMPHREYS Building Research Establishment, Building Research Station, Garston, Watford WD2 7lR, Great Britain.
CONTENTS Introductory discussion Available information Thermal comfort surveys Climatic data Free-running and heated or cooled buildings Free-running buildings Comfort temperatures Mean indoor temperatures during the surveys Discussion of free-running buildings Heated and cooled buildings Comfort temperatures Mean indoor temperatures Discussion of heated and cooled buildings Relating the comfort temperature and indoor mean temperature Discussion of relation between comfort temperature and indoor mean temperature Conclusions 1NTRODUCI'ORY DISCUSSION
A building and its occupants can be regarded, from the point of view of thermal performance, as a single self-regulating system whose objective is to achieve and maintain comfortable conditions for the occupants. The regulating processes include : (a) the careful design of the building for the climate, either by the deliberate application of modern building technology (GIVONI,1969) or by adherence to types of structure traditional to the locality (WULSIN,1949);
* The work described has been carried out as part of the research programme at the Building Research Establishment and is contributed by permission of the Director. British Crown Copyright.
230
M. A. HUMPHREYS
(b) the provision, and use if necessary, of heating or cooling plant; (c) the provision and use of blinds or shutters to regulate the admission of solar radiation ; (d) the use of doors, windows and other vents, or the provision of fans, to control ventilation (NICOLand HUMPHREYS, 1973); (e) the choice of clothing appropriate to the prevailing temperatures and activities (HUMPHREYS, 1973, 1977); (f) the choice of activity level appropriate to the warmth of the environment, for example, resting during the heat of the day; (g) the physiological changes associated with acclimatization to heat or cold by the occupants. The indoor temperature and the preferred temperature indicate the response of the regulatory system to the external disturbance caused by the weather. Success can be judged in terms of the thermal comfort achieved by the occupants, and this is indicated by comparing the temperature occurring within the occupied building with the temperatures preferred by the occupants. AVAILABLE INFORMATION THERMAL COMFORT SURVEYS
Much of the information needed for a comparison of this kind is available from published surveys of thermal comfort. A fieId study of thermal comfort is a survey of the occupants' reports of their subjective warmth, accompanied by objective measurements of the thermal environment within the building. Such a study, if conducted in normal circumstances, will indicate the temperatures prevailing in the buildings, the temperatures preferred by the occupants, and hence the degree 0s discomfort arising from temperatures which differ from those preferred. Numerous studies of this kind, of varying levels of detail, have been conducted in variouf seasons and climates during the last 50 or so years. A recent review of the techniques and an analysis of their principal results has been published (MUMPHREYS, 1976). Some conclusions relevant to the present discussion were as follows. (a) A comparison among the various indices of thermal comfort tested in the fieid studied did not yield an index of universal application. Simple indices such as air temperature or the temperature of a globe thermometer seemed to be as good as more complex indices. However, in warm humid climates it was an advantage to use an index which incorporated terms for both air velocity and humidity, and in hot dry climates air movement affected comfort. Table 15.1 indicates the degree of complexity necessary in different circumstances. (b) The mean indoor temperatures in these studies varied from 17 "C for the elderly in their homes during a n English winter (Fox et al., 1973) to 36 "C for clerical workers in Baghdad in summer (NICOL, 1974). The preferred temperature, obtained by analysis of the subjective responses, was closely related to the mean indoor temperature, except for a few studies where non-typical thermal environments had been selected for study, for example, by confining observations to the heat of the
Comfortnble temperntiires
231 T a b l e 15.1
Measurement of the thermal environment ~~
Set of measurements
Restrictions upon their use
1. Air temperature
Difference between air temperature and mean radiant temperature is small (< 2 “C). Air velocity is very slight (< 0.2 m s-l). Sweat is freely evaporated from the skin.
2. Globe temperature
Air velocity is very slight (< 0.2 ms-’). Sweat is freely evaporated from the skin
3. Globe temperature* Sweat is freely evaporated from the skin Air velocity 4. Globe temperature*
Air velocity Wet-bulb temperature
None
The table shows the simplest sets ofmeasurementsnecessary to specify the warmth oftheenvironment. They apply to indoor conditions where there is no great thermal asymmetry. * If the difference between air-temperature and mean radiant temperature is small, the air-temperature can be used instead of the globe temperature.
day (GOROMOSOV, 1963) or by choosing a single day of high solar radiation (SIB, 1967). The relation between the preferred temperature and the mean indoor temperature is shown in fig. 15.1.
16
20
25
30
35
38
Mean Temperature IT)
Fig. 15.1. Comfort temperature against indoor temperature. Each point (or pair of points) represents
a survey of t‘hermal comfort listed in table 15.3. The vertical axis is the “neutral” or “comfortable” temperature, and the horizontal axis is the mean temperature within the accommodation when the comfort votes were obtained
232
M. A. H U ~ H R E Y S Too cool and much too cool
a
mo 80
es 60 8
b a
LO
20
O - L
-2
0
2
L
Departure from neutral temperature (“C)
Departure frwn neutral temperature (“Cl
Too cool and much too cool
b 100
100
80
80
%m
€4
L
C
2
U
LO
LO
20
20
O - L
-2
0
2
0
L
Departure from neutral temperature
PCI
-L
-2
0
2
L
Deperture from neutral temperature (“C’
Too cool and much too cool
C
-$
100
100
80
80
60
60
LO
LO
20
20
C
a“2
0 - L
-2
0
2
L
Departure from neutral temperature i0C)
0 - 4
-2
0
2
L
Departure from neutral tefnpwoiure 1°C)
Fig. 15.2. The figures are distributions of comfort votes at temperatures around that found to be “neutral” or “comfortable”. The upper distributions (a) apply to adults in normal accommodation, the centre pair (b) to adults in air conditioned accommodation and the lower pair (c) to children aged about 10 years at school
Comfortable ternperatw e s
233
(c) The rating scales of thermal comfort in common use (e.g. BEDFORDor ASHRAE scales) behaved in general as equal interval scales, each step in the scale occupying approximately 4 "C.The "edges" of these steps were not sharp, but have'a standard deviation of approximately 3 "C (or 2 "C from surveys in air-conditioned accommodation). This standard deviation is attributable about equally to differences between people and to variation from time to time for the same person. This behaviour of the scales results in the distributions of thermal comfort shown in fig. 15.2 for adults in non air-conditioned accommodation, for adults in air-conditioned rooms, and for young children at school. These curves enable us to assess the effects of departures from the preferred temperature. The surveys contributing to HUMPHREYS' review are listed in table 15.2 which also summarises much of the information available from them. CLIMATIC DATA
Variation of the temperatures found to be comfortable may be understood better by considering the effects of climate. For this purpose HUMPHREYS (1978) rearranged the information from the field studies listed in table 15.2. Wherever possible the studies were broken down into subgroups for different seasons of the year, and a few further studies were included. The comfort temperature was obtained for each new grouping, usually either by applying probit analysis (fitting lines by eye) or by applying regression analysis. Where both methods could be used, the average of the two values was taken to be the comfort temperature. Different surveys used different indices of thermal comfort, and so where necessary and possible, the results were adjusted to an equivalent globe temperature for low air velocity. The resulting comfort temperatures are shown in table 15.3, which also shows the rearranged information from the comfort surveys. These comfort temperatures were related to the outdoor temperatures, obtained chiefly from the HMSO World Meteorological Tables (HMSO, 1958-1972). The mean daily maximum temperature and the mean daily minimum temperature for the months of the survey were abstracted for the appropriate geographical location, together with the values for the hottest and coldest months of the year, and the annual mean temperatures. However, if outdoor temperatures for the actual location, month and year of the survey were known, these were used in preference to the values from tables. Humidity was not included because it has little effect upon comfort temperature ; rather its effect is to increase the discomfort associated with temperatures above the comfort level. FREE-RUNNING AND HEATED OR COOLED BUILDINGS
The principal results of this analysis are shown in fig. 15.3. There is a'clear difference between free-running building (those where heating or cooling plant was absent or not being used) and heated or cooled buildings. The comfort temperatures in the free-running buildings (the filled circles on fig. 15.3) were linearly related to the outdoor mean temperature, while the comfort temperatures in the heated or cooled buildings (the open circles on fig. 15.3) fitted better a curved relation to the
231
M. A. HUMPHREYS Summary
-__ First author
Date
2
1
No. of obs. included
3
Wanner
1967
1973
Respondents
{i:;
{
666 542 498
2172 5349 -616
flZ [::;
Nicol
1974
Pepler
1971
SA
1938
Wyon
1968
1-'000 1-1000 >2500
Wyndham
1963
21 1
per resp.
Mean Mean temper- resature ponse oc
5
6
1 1
7
8
0.12 0.29 0.12
As
M/W M/W M/W
I.
22.9 22.7 23.2
M/W M/W
1 1
21.1 (24.1)
0.20 (0.67)
M/W MIW MIW
1
21.7 23.9 23.5
0.15
13M, IW 13M. 5W M 9M 7M 9M/W 9M/W 8
1 1
Webb
1959
393
11
-25 -25 -60
18.8 19.0
-
-0.05 -0.11 -
22.2 21.0 21.1
0.15 0.14
19.8 -19.0
19.8
7
0.24 0.26 0.23
21.1 22.8 23.4
21.2 22.9 23.2
0.23 0.23
18.9 19.4
19.1 19.2 21.0 (ET) 23.6
-0.03 -0.06 -
-60 -2, -1, 0, +1, +2, + 3 -2000 %I 19.7 60 0.74 0.25 28.3 - 1 -0.01 18.1 -0.02 40 0.42 1.27 31.0 -100 28.2 -100 28.6 1 21.4 0.71 0.24
-
-
22.9 21.4 28.0 35.9 33.4 24.4 22.1 24.7 28.8 22.4
0.06 0.24 0.59 0.66 0.43
-6OM/W
-150 -300 >l -150 -130 70 -50 -100 -100 -3
16M
-10
27.0 (ET) 17.2
8
col
23.3 20.8 21.5
-
-
0.02 0.08 0.20 0.22 0.14
-
-
-
-
-
-
-
-
0.24 0.20 0.31
-
-
-
19.3
0.16 0.20
22.6
-
0.19 0.19
-
0.50
0.02 0.33 0.02
0.33 0.33 0.21
0.64
0.31
0.23
0.05 1.oo
25.3 18.2 26.9
0.22
-
-
25.5 32.5 31.1 23.5 22.3 24.6 25.8
-
18.9 25.4 18.6 26.8 26.1 26.7 19.6 22.8 20.3 -
32.9 31.7
-
-
20.5
-
-
24.2 (ET) 17.5
0.70
0.21
22.4
23.0
0.25
0.48
23.2 (ET)
23.3 (ET)
-0.02
(ET) -3000
10
- 1, 0, +1, +2
-
{i;:
9
0.18 0.16 (0.07)
0.27 0.2
Fox 1973 796 1 - 0.06 M/W (d) 4-category asymmetrical scale, numbered -1, 0, +1, +2 Goromosov 1963 5902 28.8 1.40 (el scales with 4 categories above the neutral category (see col. 19 for numbers) Ambler 1955 176 6M several 25.2 1.00 Mookerjee
Neutral Neutral dard Regression temp. temp. mean (reg) b o b . ) rescoeff. "C "C uonse
+1
{:;;:
1962 1970 1955
obs.
4
- 1 , 0,
(b) 5-category scales, symmetrical categories, numbered -2, M Black lgS4 W 172M, McConnell 1940 -40000 573w Robley 275M/W 1947 -20000 (c) 7-category scales, symmetrical categories, numbered -3, Ambler -2000 1M 1966 Angus 1957 1261 M/W Ballantyne 1967 34M/W 1992 Bedford 1936 2571 M/W Ellis 1952 M 1829 1247 13W Ellis 21M 1953 2238 Hickish 1955 1537 M/W Hindmarsh Humphreys Malliotra
No. of
-
M
-
31.4
1.47
0.37
0.43
29.1
-
14M
-20
28.8
0.65
0.16
0.46
27.2
27.5
-300
30.3
1.86
0.31
0.44
26.0
26.3
(f) scale with 6 categories above the neutral (-6 to +G)
Rao
1952
1626
-
Comfortable temperatures
235
of information
Table 15.2 Stand-
Category widths
12
md
%
devia-
Comfortable
13
14
15
16
5.0 5.6 (13.2)
-
2.1 2.3 (5.1)
76 78 80
-
Country
Notes
18
19
17
England Switzerland Switzerland Sweden Sweden
Su, A/C, OW mostly in single offices Wi, OW, various types of heating Su, OW data pooled from A/C & non A/C rooms Wi, T,an average January day SP, T, A sunny spring day Wi, A/C, students in lecture theatre Su, A/C, students in lecture theatre ,,cool days” Su, A/C, students in lecture theatre ,,warm days”
I
-
-
1.9 2.2
90 100
3.1 2.8 4.3
-
1.6 1.4 1.9
66 68 74
Switzerland Switzerland Switzerland
49 41 67
England England U.S.A.
Wi, OW responses twice daily Wi, OW responses twice daily Su, A/C, OW, responses thrice daily
Su, A/C, OW. responses thrice daily
6.5
I
I
4.4 3.4 1.9
5.4 4.4 -
3.3 3.1 1.2
4.7
-
2.1
75
USA.
4.4
2.2
-
-
4.4 (7.1) 3.2 2.2 2.7 (4.6)
4.6 (2.2) 3.4
(2.6)
2.6 3.0 2.4 2.1 2.5 2.3
56 79 50 41 41 68
India (UP) England Papua England
8.3 4.6
4.1 4.5
3.5 3.4
76 50
-
-
-
5.1 4.6
5.0 3.7
4.5 3.6
-
-
-
46 58 -70 -80
8.3
8.3
3.1
-
Iraq India (UP) U.S.A. USA. Brazil (Rio) Brazil (Rio) England
-
4.4
2.2
63
Australia (N)
-
-
-
-
England
Indoors in the shade Wi, students, several races, lecture theatre Jan.-March in their own homes Wi, sedentary light industry, mostly women Voyage from Singapore to Hong Kong, crew Europeans on shore, home and office Exopeans on shore, home and office Su, mostly sedentary. Light industry, mostly men Ow, 11 am daily for a year (Sydney) Ow, observations over 15 months Voyage, trapical a x l temperate regions OW, hot dry weather, includes overnight obs. OW, hot dry weather, includes overnight obr.] Au, T, school lessons non A/C Sp, T, school lessons non A/C Wi, lightly active or sedentary Su, lightly active or sedentary) Surgeons, nurses, andesthetists, operating theatres Feb.-May manual work, daily obs. after lunch break Wi, elderly at home 9 am or 5 pm
6.8
1.7
2.0
90
USSR (S)
Su, dwellings obs. chiefly during heat of day
2.5
1.9
2.4
40
Nigeria
Scale - 2 to +4, includes some outdoor obs.
-
1.1
(0.8)
54
India (UP)
1.4
1.2
2.3
24
Singapore
Su scale -2 to +4, sedentary work, includes experimenters Scale -4 to +4, OW, office, and home, includes night obs.
2.4
2.0
1.7
17
-
-
-
-
-
-
-
-
-
Singapore Singapore England Australia England
-
-
60
India
}
Calcutta, students indoors, all year round
M. A. H U ~ H R E Y S
236
1
2
4
3
(g) children 7-category symmetrical, numbered -3,
Auliciems Auliciems Davies
1969 1973 1972
2624 665 1700
Humphreys Humphreys
1973 1977
4479 4990 I4385
Pepler
1972
{
-40000 -40000
5
-2, - l , O ,
6
7
8
9
0.59 0.29 0.25
0.20 0.10 0.08
0.21 0.23 0.10
0.81
1 0 1 1
+1, +2, + 3
- -
37SM, 246W -5 M/W 13M, 14W -60
19.9 21.8 20.5
40M,40W M/W M/W
-50 -40 -30
23.9 19.3 21.2
0.51
0.27 0.17
0.49
0.16
M/W M/W
-100
24.4 22.1
-
22.2 25.0
-0.06 1.90
-100
-
17.1 20.6 18.0
18.1 21.1
0.20 0.20 0.15
19.9 18.0
(18.4) 16.8 18.2
-
-
23.6 22.6
-
-0.01 0.16
-
-
'
16.8
-
(h) children, 25-category scale, -12 to + I 2
3651 I3365
Lane Key: M - Men W - Women Su - Summer
.Sp - Spring Au - Autumn Wi - Winter
11M, 11W 11M, 11W
-170
-150
-
OW - Office Workers T - Teachers A/C - Air-conditioned
The numbers in brackets are in some respect usnusual, and are discussed by Humphreys (1976)
North USSR
UK
winter
January
+-
2LO
. 2 -22 ~
+LO+
-m -18
-16 -11,
-12 -10 -8
UK July +180f
-L -2 o 2 L 6 a 10 12 is 16 Monthly mean outdoor temperature ["Cl
-6
18
BagMd Singapore +27Of
20 22
2s 26
July
.+3L0f
20
30
32 3~
Fig. 15.3. Relation between the preferred indoor temperaturesobtained from thermal comfort surveys and the mean outdoor air temperature, obtained from meteorological tables, for the seasons at which the surveys were conducted 0 - free-runuing buildings, regression line A 0 - other buildings, regression line B
Comfortable temperatures
12
13
14
15
16
17
42 4.4
2.9 4.8
3.1 3.6
4.9 5.6
24 35
71 75
-
England England England
4.4 4.0
4.8 8.0
8.0 5.5
5.8 8.0
30 38
85 74
England England
5.7
8.8
5.2
8.6
40
75
England
-
-
-
-
-
U.S.A.
-
-
-
-
-
-
-
-
-
-50 -50
-
18
U.S.A.
U.S.A. U.S.A.
237
19
Wi,school lessons, age 11-16, several schools Su, school lessons, age 11-16 several schools All year round, age 13, one class, one obs. per week Su, school lessons, age 12, 3 classes Su, ('71) school lessona, age 7-10, several classes Su, ('72) school lessons, age 7-10, several classes Au, school lessons several classes. not A/C SP, school lessons, several classes, not A/ C Iowa, school lessons, age 10, not AIC Iowa, school lessons, age 10, not A/C
outdoor temperature. Discussion of the relation between the mean indoor temperature, the comfort temperature and the climate, must bear in mind the distinction between these two modes of operating buildings.
FREE-RUNNING BUILDINGS COMFORT TEMPERATURE
The open circles in fig. 15.4 show the relation between the comfort temperature and the outdoor temperature. The index of outdoor temperature is the average of the daily outdoor maximum and the daily outdoor minimum for the period of the survey. Each point is the result either of a complete comfort survey, or of a seasonal subgroup of a survey. This average temperature is a satisfactory predictor of the comfort temperature because it shows on a single straight line the observations from temperate, tropical, equatorial and arid climates. The correlation coefficient is 0.97; the residual standard error is 1.0 "C.The equation of the line B is T, = 0.53 To,+ 11.9,
(15.1)
where T, is the comfort temperature and To,,the outdoor temperature, both in "C. The slope of the line indicates a rise of just over half a degree in comfort temperature for every degree rise in the outdoor mean temperature. Statistical examination shows that the relation cannot be significantly improved by weighting the mean daily maximum and minimum temperatures differently
T a b l e 15.3 Subjective comfort temperature and outdoor temperature
Continent of origin and first author
Neutral temp. from survey
Average ma'* and min. for period
Meteorological data for
Months included
Mode of Type of building oper-
Comments
n-
ation
1
2
3
4
5
6
7
8
Officeandhome
F
AFRICA Ambler, 1955
25.0*
30.6
21.1
AUSTRALASIA Ballantyne, 1967 Ballantyne, 1967 Ballantyne, 1976
25.4 25.4 25.0
31.1 28.3 29.8 *
24.4 Port Moresby 23.0 Port Moresby 23.9" Port Moresby
Ballantyne, 1976
27.2
29.8*
23.9"
Port Moresby
Ballantyne, 1976 Ballantyne, 1976 Hindmarsh, 1962 Hindmarsh, 1962 Hindmarsh, 1962 Hindmarsh, 1962 Macpherson, 1964 Macpherson, 1964 Wong, 1967 Wong, 1967 Wyndham, 1963
22.8 20.8 24.2 23.9 22.3 21.4 26.2O 27.6O 21.0 23.0 26.2*
25.6 13.3 25.2 18.9 17.4 23.3 32.8 32.8 16.3 25.0 32.8
13.9 5.6 18.0 11.5 9.1 15.4 23.3 23.3 8.5 17.6 25.0
Melbourne (Vic.) Melbourne (Vic.) Sydney (N.S.W.) Sydney (N.S.W.) Sydney (N.S.W.) Sydney (N.S.W.) Darwin (N.T.) Darwin (N.T.) Sydney (N.S.W.) Sydney (N.S.W.) Mapoon (Q.)
Jan.-Mar. Homes July-Sept. Homes Nov., Feb., Homes Aug. Nov., Feb., Homes Aug. Offices Feb. Offices Jul. Offices Jan.-Mar. Offices Apr.-Jun . Jul.Sep. Offices Oct .-D~c. Offices All year All year Offices Jun.-Aug. Offices DE.-Feb. Canteen Mar.
26.4
30.6
23,3
Singaaporc
All year
Port Harcourt
All year
F F F
Caucasians Caucasians Caucasians
F
Melanesians
E E F E E F F F E E F
ASIA
Ellis, 1953
9
O ~ c e a n d h o m e xj
Data from Batch-,lor, near Darwin
is: ?
____
1
2
Goromosov, 1963 Goromosov, 1963 Goromosov, 1963 Goromosov, 1963 Mookerjee, 195213 Mookerjee, 195/23 Nicol, 1974 Nicol, 1974 Rao, 1952 Webb, 1959 AMERICAS Lane, 1965 McConnell. 1940 Gagge, 1976 Newton, 1938
22.7 21.5 19.0 17.5 29.1 * 21.0* 31.2" 30.1* 26.1 27.3
Partridge, 1935 Partridge, 1935 Pepler, 1971 Pepler, 1972 Pepler, 1971 Pepler, 1972 Sh, 1938 Sit, 1938 Tasker, 1938 EUROPE Angus, 1957 Auliciems, 1973 Auliciems, 1969 Bedford, 1936 Black, 1954
.
3
-
29.0
4
5
USSR South USSR North USSR Central USSR South New Delhi New Delhi Baghdad New Delhi Calcutta Singapore
.
6
7
Jun.-Aug. Dec.-Feb. Dec.-Fep. Dep.-Fep. Apr.-Jul. Nov.-Feb. JUl.-AUg. May-Jun. All year All year
Flats Flats Flats Flats Various Various Office and home Office and home Office and home Office and home
8
- 8.3 1.o 37.8 24.2 43.3 39.7 31.7 30.6
15.9 -26.6 -14.9 - 7.1 25.4 8.8 24.4 27.2 21.1 23.3
22.2 23.7* 23.9 23.6*
8.9 26.9 26.7 26.8
- 1.1 17.8 18.9 16.2
Des Moines (Iowa) New York New York Minneapolis
Nov. Ju1.-Sep. Aug. Jun.-Sep.
Classroom Office Office
E E E E
24.5* 22.5* 23.5 22.3 23.6 22.6 24.6 25.8 22.5
26.1 -0.6 16.7 18.9 16.7 18.9 24.4 27.9 27.3 *
15.0 - 8.1 8.3 8.3 8.3 8.3 18.1 21.7 16.7*
Toronto Toronto Portland (Oregon) Portland (Oregon) Portland (Oregon) Portland (Oregon) Rio de Janeiro Rio do Janeiro Toronto
Jul. Dec.-Feb. Oct. May Oct . May May-Oct. Nov.-Apr. July
Climate-room Climate-room Classroom Classroom Classroom Classroom College College Climate-room
E E E E E E F F E
18.9 20.9 17.6 18.4 19.2
6.8 18.5 9.2 6.8 9.2
2.6 9.9 3.8 2.6 4.1
London London London London London
Dec.-Feb. May-Jun. 0ct.-Mar. Dec.-Feb. 0ct.-Mar.
Lecture room Classrooms Classrooms Factories Offices
E F E E E
-20.1
(Kew) (Kew) (Kew) (Kew) (Kew)
Office
9
Meteorological data pooled from several
F F F F F F Children Nevins' data See also Rowley, 1947 Children Children Teachers Children Teachers Children
Children Children Light industry
1
2
3
4
Black, 1966 Davies, 1972 Fox, 1973 Grandjean, 1966 Grandjean, 1968
22.2 18.0 17.5 20.9 21.3
21.2 12.3 7.8 5.9 23.3
12.8 7.0 2.6 -1.5 12.4
Hickish, 1955 Humphreys, 1970
19.4 19.9
20.1
11.6
Humphreys, 1970 Humphreys, 1970 Humphreys, 1970 Humphreys, 1970 Humphreys, 1973 Humphreys, 1977 Humphreys, 1977 S I B, 1967 S I B, 19b7 Wanner, 1973 Wanner, 1973 Wyon, 1968
19.7 19.3 20.0 20.2 19.9 18.1 16.8 19.8 -19.0 21.2 23.1 20.5
Key:
+ Estimated for 0.1 m s-1
6.8* 11.9* 14.6* 18.8* 21.0* 16.8* 16.4* 16.8* -1.0 15.0 3.5 23.7 13.9
5
London (Kew) J~n.-Aug. All year Birkenhead Jan.-Mar. London mew) Jm.-Mar. Zurich Zurich/Basel/Bern Jun.Sep.
London (Kew) 0.8* Garston
3.4* Garston 6.9* Garston 10.0* Garston 1.7* Garston 8.9* Garston 7.8* Garston 8.9* Garston 5.3 -2.3 12.8 7.3
Several Swedish towns Zurich Zurich London (Kew)
air-speed from author's regression equations. Estimated from effective temperature upon available information. 0 Eslimated from the temperature of the neutrallwarm transition by substracting 2 OC. x Local meteorological data contemporary with comfort observations. F ,,Free-ming" building, no energy used for heating or cooling. E Building with heating or cooling plant using energy.
*
6
7
Offices Classroom Homes Offices Offices
May-Aug. Factories Feb., Nov., OffiW
DK. Mar.,Apr. Oct., May. Jun., Sep. Jul., Aug. Jun. (1971) Jun. (1971) Jun. (1972) Jan. May Dec.-Feb. Jm.-Aug. All year
8
E F E E E F E
Offices Offices
Offices Offices Classrooms Classrooms Classrooms Classrooms Classrooms Lecture theatre Lecture theatre Operating theatre
9
"solar" school Old-people Some air-conditioning Light industry
Garston is north of London F F F
El
E
E E E
E
Children BRS data on young children Teachers Teachers
241
Comfortable temperatures
Fig. 15.4. Indoor temperature and comfort temperature against outdoor temperature for the freerunning buildings. The open circles are the comfort temperature and the line B shows their regression upon the mean outdoor temperature (equation (15.1)). The filled circles are the mean indoor temperature and line A shows their regression upon the mean outdoor temperature (equation (15.2) in thetext)
,Ll
,
I
I
10
15
20
I
25
3l
35
Mean outdoor temperature i°C)
from the simple average or by including the temperatures of the hottest or coldest months of the year. MEAN INDOOR TEMPERATURE DURING THE SURVEYS
The filled circles in fig. 15.4 show the relation between the indoor mean temperature and the outdoor temperatures. Again it is found that the average outdoor temperature for the period of the survey is a good predictor of the mean indoor temperature. The data from different types of climate fall close to a single straight line; the correlation coefficient is 0.96, and the residual standard error is 1.2 "C. The equation of the line A is
qfl= 0.55 T0,+14.1,
(15.2)
where qflis the mean indoor temperature in "C. . The slope of the line indicates that the mean indoor temperature rises by just over a half degree for every degree rise in the mean outdoor temperature. Again the prediction cannot be improved by a different weighting of the maximum and minimum temperatures, or by including the temperatures of the hottest or coldest months of the year.
242
M. A. HUMPHREYS DISCUSSION OF FREE RUNNING BUILDINGS
The similar behaviour of the comfort temperature and the indoor mean temperature in this class of building is very striking. The rise of comfort temperature with increased outdoor temperature is virtually the same as that of the mean indoor temperature. This is very informative. It means that any discrepancy between the mean indoor temperature and the comfort temperature is virtually constant across the range of climates represented by the observations. This strongly suggests that the respondents had adapted equally well (or badly) to their indoor environment no matter whether it was hot, cold, or indifferent. The average discrepancy between the mean indoor temperature and the comfort temperature is 2.4 “C. If the processes of adjustment to the indoor temperature (choice of clothing, regulation of activity level, physiological acclimatisation) were complete, on average no discrepancy would be expected. Two explanations for the discrepancy found can be suggested. Firstly, people may prefer to be 2.4 “C warmer than the value corresponding to the centre of the comfort scale. This constant bias in comfort scales, independent of the climate, does not seem very likely. It seems particularly unlikely that people in hot climates would prefer to be on the warm side of “neutral” or “comfortable”. The second explanation is that the mean indoor temperature found in the surveys may not indicate the mean temperature encountered by the respondents. This is certainly the case. Most studies contain observations made during the working day only, and all contain predominantly daytime observations. The mean temperature of these occasions will, therefore, give more weight to the hottest part of the day, and will not be truly representative of the subjects’ daily experience of temperatures. Since the buildings considered had no heating or cooling services in operation, appreciable diurnal swings of indoor temperature would occur. A sinusoidal swing amplitude 3 “Cwould give a bias of more than 2 “C if measurements were made only during the warmest 8 hours of the cycle. Diurnal temperature swings of 3 “C or more are common in free running buildings, so this explanation of the bias seems likely. Equation (1 5.2) also says something about the usefulness of carefully designing a building for the climate and of providing effective controls for ventilation and sunshine. If the buildings were all similar unoccupied blocks, and there were no incident solar radiation, the indoor temperature would on average be equal to the outdoor temperature. Such conditions would yield a line of unit slope. But the radiation heat gain is larger in hotter climates, so, allowing for this, a slope substantially greater than unity would be expected. The slope of about 0.5, found in equation (15.2), indicates that the design and operation of the buildings more than halves the effect of the different climates. It is therefore easy to understand why it is unwise to build in one climate a structure originally designed for a different climate. Equation (15.1) could be useful for the thermal design of buildings. Designs could be analysed to discover whether the indoor temperature would on average fall near the line, and so assess whether it would be practicable to do without heating or cooling for all or part of the year.
243
Comfortable temperntuves HEATED AND COOLED BUILDINGS COMFORT TEMPERATURES
The open circles in fig. 15.5 show the comfort temperatures (Tco)and the mean for buildings where heating or cooling plant is in operation. outdoor temperatures (To*) The absence of surveys of thermal comfort in air-conditioned buildings in hot climates is unfortunate, because it leaves a gap in our knowledge of an area important
-
0
0
.
I
0 0
0
0 0
-
0
I
I
I
I
15
I
I
I
Fig. 15.5. Indoor temperature and comfort temperature against outdoor temperature for heated or cooled buildings. The open circles are the comfort temperature and the curve B shows their regression upon the square of the outdoor mean temperature (equation (15.3) on the text). The filled circles are the mean indoor temperature and the curve A shows their regression upon the square of the mean outdoor temperature, (equation (15.4) in the text)
for comfort. A quadratic equation is fitted over the range of observations. Polynomials of higher degree produce a better fit, but with some sacrifice of robustness, These forms cannot be safely extrapolated. These and possible alternativc mathematical models are discussed elsewhere (HUMPHREYS, 1978). The quadsztic equation best fitting the observations is T,,
= 0.0077
19.8.
(15.3)
The fit for this equation has a correlation coefficient of 0.66 and the standard error of the residuals is 1.6 "C. The equation can be improved to a statistically significant degree by including the mean daily maximum temperature of the hottest month of the year as predictor variable. The equation then becomes T,,
= 0.0065 T&+0.32
T,+12.4,
(15.4)
where Th is the mean daily maximum temperature of the hottest month in "C. Including this extra term raises the correlation coefficient to 0.77 and seduces the standard error of the residuals to 1.4 "C. A further interesting fact emerges from the analysis. The comfort temperature depends to some extent upon the country of origin, independently of the climate.
244
M. A. HUMPHREYS
Data from the USSR tend to lie on average 2 "C below the line, while data from the USA tend to fall 2 "C above the line, and English data 1 "C below it. Data from other countries are too few to show a significant departure from the line. No dependence upon the date of the survey has been found, in spite of the commonly held belief that comfort temperatures have been rising during the last 50 years. MEAN INDOOR TEMPERATURE
The filled circles in fig. 15.5 show the relation between the mean indoor temperature during the survey and the mean outdoor temperature, again for buildings with the heating plant or cooling plant in operation. Unfortunately, the mean indoor temperature is unknown for the surveys from buildings in the coldest climates. A quadratic reIation has again been fitted to the observations, and yields the equation
q,,= 0.0075 T:u+20.5.
(15.5)
The equation has a correlation of 0.61 and the standard error of the residuals is 1.5 "C. Again, the prediction is improved by including the mean daily maximum temperature of the hottest month of the year as a predictor variable. The equation is Tin= 0.0056 T:,,+O.29T,+13.8.
(15.6)
Including this term raises the correlation coefficient to 0.69 and reduces the standard deviation of the residuals to 1.4 "C. This indicates that the mean indoor temperature maintained in buildings depends upon the outdoor summertime temperature. Again, the indoor temperatures show a dependence upon the country of origin of the data, but because of the limited sample it is not possible to be specific beyond saying that the indoor temperatures in the U.S.A. surveys are about 2 "C higher than average. No dependence on the date of the survey can be demonstrated. DISCUSSION OF HEATED AND COOLED BUILDINGS
The correlation between comfort temperature and the mean temperature found in these buildings is very striking. It suggests that the same factors are contributing to their control. The discrepancy between the mean temperature and the comfort temperature averages 0.6 "C,the mean temperature being the higher, and the difference is virtually independent of the outdoor temperature. It was stated above, that if people successfullylearn to cope with their thermal environment, whether by adjusting their clothing or by adjusting the temperature, no systematic difference would be expected between the mean temperature and the neutral temperature. The fact that a constant discrepancy is found, different from that found in the free-running buildings, again suggests that there is a bias in the measurement of the mean temperature. This could occur for the same reason as in the free-running buildings, but would be expected to be much smaller, because the controlled operation of heating and cooling plant tends to reduce the diurnal swing of temperature within
245
L’omfortable temperatures
the accommodation. Thus measurements concentrated about the hottest third of the day will generally be only slightly greater than the true mean temperature. RELATING THE COMFORT TEMPERATURE AND THE INDOOR MEAN TEMPERATURE
The similarity of the variation of the comfort temperature and the mean indoor temperature with varying climate, both for the free-running buildings and for the heated or cooled buildings, suggests that the two temperatures are closely related. Fig. 15.6 is a scatter diagram of the mean indoor temperature plotted against the comfort temperature. Free-running buildings (open circles) are distinguished from the heated or cooled buildings (filled circles). Although the two sets of points overlap
15
20
25
33
35
Fig. 15.6. The relation between the comfort temperature and the indoor mean temperature for freerunning buildings (open circles) and for heated or cooled buildings (filled circles). The lines indicate the structural relations and have unit gradient. Line A is for the observations from heated and cooled buildings, and line B for the free-running buildings
to some extent, a clear difference exists. This difference can be demonstrated by regression analysis, including in the prediction of the comfort temperature a dummy variable D, taking the value of unity if the building is free running, or zero if it is heated or cooled. The equation is
T,, = 0.94 Tin-1.60+0.6.
(15.7
246
M.
A. HUMPHREYS
The correlation coefficient is 0.97 and the standard error of the residuals is 0.9". The value of the coefficientofD is significant at the 0.001 level of probability. A further small but statistically significant improvement in correlation can be obtained by including a term for outdoor temperature in the regression equation. However, this effect is of dubious validity because of the high correlation between the indoor and the outdoor climate. It has already been pointed out that the estimate of the mean indoor temperature is not without error, having a bias which depends upon the diurnal variation of temperature in the accommodation and upon the time distribution of the observations. The regression assumes that all the errors of prediction occur in the comfort temperature. The errors inherent in the mean indoor temperature are very similar in magnitude to those of estimating the comfort temperatures. Prediction of comfort temperature from the outdoor temperature, for people in free-running accommodation, has a residual error of 1.0 "C while predicting the mean temperature on the same basis leaves a residual error of 1.2 "C;for the heated or cooled accommodation the figure is 1.4 "C for both temperatures. The regressions of the indoor mean temperature upon the comfort temperature and of the comfort temperature on the indoor mean temperature therefore have equal status, and the geometric means of the pairs of regressions are reasonable estimates of the underlying relation. The gradient of such a line is simply the ratio of the standard deviations of the variables. For the heated or cooled accommodation, the standard deviation of the comfort temperature is 1.93 "C and that of the mean temperature 1.87 "C. The gradient of the line is therefore 1.03. For the free-running accommodation the standard deviation of the comfort temperature is 4.28 "C and of the indoor mean temperature 4.39 "C. The gradient of the line is, therefore, 0.97. Neither gradient differs significantly from their mean value of unity. A pair of lines of unit gradient, passing through the group-means in fig. 15.6, indicate the relation between the comfort temperature and the indoor mean temperature. DISCUSSION OF THE RELATION BETWEEN COMFORT TEMPERATURE A N D INDOOR MEAN TEMPERATURE
If people were on average completely adjusted to their thermal environment, if the mean temperature of the accommodation in the surveys represented the thermal environment to which they had adjusted, and if the scales of thermal comfort were without bias, then fig. 15.6 would be a single line of unit slope such that the comfort temperature and the indoor mean temperature were always equal. The observations yield a gradient of unity. This strongly suggests that, within the limits of the accuracy and ranges of the observations, the adjustment of people is complete; a unit change in the comfort temperature being associated with unit change in the mean temperature of the indoor environment. The lateral displacement of the lines thus indicates the presence of bias, either in the mean temperature or in the subjective scale which yields the comfort temperature. To suppose that the bias lies in the scale leads to an unlikely conclusion; it would be necessary to assert that comfort scales behaved differently,
Comfortable temperatures
241
in equivalent indoor mean environments, depending upon the mode of operation of the building. On the other hand, to suppose that the bias lies in the measurement of the mean temperature is consistent with the times of acquisition of the observations and the probable diurnal swings of temperature in the buildings. Although the results indicate that the mean indoor temperature is a biased indicator, they do not suggest a way of obtaining a better measure. Thermal experience can be divided into three broad categories: sleep in bed, up and about indoors, and up and about outdoors. In warm climate the distinction between the last two may be rather blurred, but in general the three conditions employ quite different modes of clothing, and they generally entail different degrees of physical activity. It could therefore by hypothesised that separate adaptive solutions exist for the three conditions, and that they affect each other only slightly. For example, a person may be comfortably warm in bed in a room at 5 "C or less, if he uses suitable bed clothing. However, the bedroom temperature would probably not greatly affect the temperature he would prefer in the living room during the day. Again, by the choice of good outdoor clothing, a person can be comfortable outdoors when it is freezing, but this would probably not greatly affect the temperature he would prefer in his living room. It may well be, therefore, that the temperature to which people "up and about indoors" adapt is the mean indoor temperature they are experiencing throughout their time "up and about indoors". This information is not usually available from surveys of thermal comfort because in most surveys the time sequence of observations has not received much attention. CONCLUSIONS
1. For buildings which are not consuming energy for heating or cooling (free -running buildings) the comfort temperature is predictable from the outdoor mean temperature for the appropriate season of the year. The relation is
T,, = 0.53 T,,+11.9, where T,, is the comfort temperature and To, the mean outdoor temperature in "C. The prediction has a standard error of 1.O "Cand applies to the range 10 " < To < 34 "C. 2. For buildings which are heated or cooled the comfort temperature cannot be predicted so well. The fitted relation is curvilinear, giving the equation
T,, = 0.0065 T u: f0.32 Th+12.4, where Th is the average daily maximum temperature for the hottest months of the year. The prediction has a standard error of 1.4 "Cand applies to the range -24" < To, < 23 " C ,18" < Th < 30 "C. The comfort temperature further depends, to a lesser extent, upon the country of origin in a way which appears to be unrelated to the climate ; Russian and English comfort temperatures are below average while NorthAmerican ones are above average. 3. The mean indoor temperatures are found to lie on lines parallel to the comfort temperatures. For free-running buildings they are on average 2.4 "C higher than the
248
M. A. HUMPHREYS
comfort temperature. For heated or cooled buildings they are on average 0.6"C higher than the comfort temperatures. It is argued that these differences are attributable to selective sampling of the temperature during the diurnal temperature cycles. 4. The comfort temperature is highly correlated with the mean indoor temperature, the structural relation having unit gradient. It is argued that this relation indicates excellent adaptation to the average experience of indoor temperature while people are up and about indoors. 5. Further studies are necessary for reliable determination of the dependence of comfort temperatures in cooled buildings on the outdoor temperature in hot climates. REFERENCES AMBLER H. R. (1955), Notes on the climate ofNigeria with reference to personnel, Journal of Tropical Medicine and Hygiene 58, 99-112. H. R. (1966), Conditions of thermal comfort in North India, Journal of Tropical Medicine AMBLER and Hygiene 69, 275-281. ANGUST. C. and BROWN J. R. (1957), Thermal comfort in the lecture room, an experimental study of winter requirements, JIHVE 25, 175-182. AULICIEMS A. (1969), Thermal requirements of secondary school children in winter, J. Hyg. Cambridge 67, 59-65. AULICIEMS A. (1973), Thermal sensations of secondary school children in summeriime, J. Hyg. Cambridge 71,453-458. BALLANTYNE E. R., BARNED J. R., and SPENCER J. W. (1967), Environment assessment oj'ucclimatised Caucasian subjects at Port Moresby, Papua, Proceedings 3 rd Australian Building Research Congress. BALLANTYNE E. R., AIRAHM., HILLR. K., SP SPENCER J. W. (1976), Probit analysis of thermal sensation assessments, Paper to Annual Meeting Australian Inst. Rfrig. Air Cond. and Heating, Melbourne. BEDFORD T. (1936), The warmth factor in comfort at work, MRC Industrial Health Board, Report No. 76, HMSO London. BLACK F. W. (1954), Desirable temperatures in offices, JIHVE 22,319-328. BLACK F. W. and MILROY E. A. (1966), Experience OJ air-conditioning in offices, JlHVE 34,188-196. DAVIESA. D. M. (1972), Subjective ratings of the classroom environment: a sixty-two week study of St. George's School, Walhsey, University of Liverpool. ELLISF. P. (1952), TIermal comfort in warm humid atmospheres - observations in a warship in the tropics, J. Hyg. Cambridge 50, 415432. ELLISF. P. (1953), Thermal conijort in warm humid atmospheres - observations on groups and individuals in Singapore, J. Hyg. 51, 386-404, Cambridge. Fox R. H., WOODWARD P. M., EXTON-SMITH A. N., GREENM. F., DONNISON D. V., and WICKS M. H. (1973), Body ternperutures in the elderly: a national study ofphysiological, social and environmental conditions, Br. Med. J. 1, 200-206. GAGGE A. P. (1976), Summer survey of thermalpreferences, Report to the Department of Interior Federal Energy Office (prepared by A. P. GAGGE for the late R. G. Nevins), J. B. Pierce Foundation, New Haven. GIVONIB. (1969), Man, Climate and Architecture, Elsevier, Amsterdam. GOROMOSOV M. S. (1963), The microclimate in dwellings, Moscow State Publishing House for Medical Literature, (in English: BRS Library Communication No. 1325, 1965). G RANDJEAN E. (19661, Ra~~niklimatischeWirkungen verschiedener Heizsysteme in Buros, Schweiz. BI. Luft. Heiz. 3, 18-23.
Comfortable temperatures
249
GRANDJEAN E. (1968), Raumklimatische Untersuchungen in Biiros wahrend der warmen Jahreszeit, Heiz. Luft. Haustechn. 19, 118-123. HICKISHD. E. (1955), Thermal sensations of workers in light industry in summer. A $eld study in Southern England, J. Hyg. Cambridge 53, 112-123. HINDMARSH M . E. and MACPHERSON R. K. (19621, Thermal comfort in Australia, Australian Journal of Science 24, 335-339. HMSO, Tables of Temperature, Relative Humidity, Precipitation and Sunshine @ the World. Parts 1-6, HMSO London 1958-1972. HUMPHREYS M . A. and NICOLJ. F. (1970), An investigation into thermal comjort of office workers, JIHVE 38, 181-189. HUMPHREYS M. A. (1973), Classroom temperature, clothing and thermal comfort - a study of secondary school children in summertime, JIHVE 41, 191-202, (also available as Building Research Establishment paper CP 22/74). M. A. (1976), Field studies of thermal comfort compared and applied, JIHVE 44,5-27, HUMPHREYS (available as Building Research Establishment paper CP 76/75). HUMPHREYS M. A. (1977), A study of the thermal comfort of primary school children in summer, Building and Environment 12, 231-239, (available as Building Research Establishment paper CP 17/78). HUMPHREYS M. A. (1977), Clothing and the outdoor microclimate in summer, Building and Environment 12, 137-142. HUMPHREYS M. A. (1978), Comfortable indoor temperatures related to the outdoor air temperature, Building Research and Practice (in press). LANEW. R. (1965), Education, Childrenand Comjort, University of Iowa, Iowa. MACPHERSON R. K. (1964), Between group diyerence@athermal comfort standards, Proceedings of the Lucknow Symposium, Arid Zone Research 24,213-219, Paris, UNESCO. MALHOTRA M. S. (1955), Environmental comfort zone in warm and humid atmosphere, Journal of Scientific and Industrial Research of India 14, 469-374. MCCONNEL~ V J . J. and SPIEGELMAN M. (1940), Reactions of 745 clerks to summer air-conditioning, Heating Piping and Air Conditioning 12, 317-322. MOOKERJEE G. C. and MURGAI M. P. (1952), A preliminary report on the determination of romjbrt zone of Zndiaiz subjects during North Indian summer, Journal of Scientificand Industrial Research of India l l a , 14-16. MOOKERJEE G. C. and SHARMA R. N. (1953), A report on environmental comfovt zone in dry tropics, Journal o f Scientific and Industrial Research 12a, 283-287, India. NEWTON A. B., HOUGHTON F. C., GUTBERLET C., and QUALLEY R. W. (1938), Summer cooling requirements of 275 workers in an air-conditionedoffice, ASHVE Trans. 44,337-356. NICOLJ. F. and HUMPHREYS M. A. (1973), Thermal comfort as part of a self-regulating system, [In:] Thermal Comfort and Moderate Heat Stress, HMSO, London. NICOL J. F. (1974), An analysis of some observationsof thermal comfort in Roorkee, Zndia andBaghdad, Iraq, Annals of Human Biology 1, 411426, (available as Building Research Establishment paper CP 4/75). PARTRIDGE R. C. and MACLEAN D. L. (1935), Determination of the comfort zone for schoolchildren, Journal of Industrial Hygiene 17, 66-71. PEPLERR. D. (1971), The thermal comfort of teachers in climate controlled and non-climate controlled schools, ASHRAE Trans. 77, 43-51. PEPLER R. D. (1972), The thermal comfort of students in climate controlled and non-climate controlled schools, ASHRAE Trans. 78, 97-109. RAOG. N. (1952), Comfort range in tropical Calcutta - a preliminary experiment, Indian Journal of Medical Research 40, 45-52. F. B., JORDAN R. C ,and SNYDER W. E. (1947), Comfort reactions of workers during occupancY ROWLEY of air-conditioned oJfice, ASHVE Trans. (No. 1321), 357-368. SA P. (1938), Conforto Termico, Departmento de Estatistica e Publicidade, Rio de Janeiro.
250
M. A. HUPMHREYS
S. 1. B. (Anon)., Teachers’ opinions of classroom climate - a questionnaire survey, Statens Institute for Byggnadsforkskning, Report No. 31, Stockholm 1967. C . (1938), Cooling requirements for summer comjort air conditioning in Toronto, ASHVE TASKER Trans. (No. 1100), 549-558. H. U. (1973), Comfort and air quality in air-conditioned rooms, [In:] Thermal Comfort and WANNER Moderate Heat Stress, H M S O , London. WEBBC. G. (1959), An analysis of some observations of thermal comfort in an equatorial climate, Brit. J. Industr. Med. 16, 297-310. WONGF. G. (1967), The significance of work comfort in architecture, Architectural Science Review 10, 119-130. F. R., (1949), Adaptations to climate among non-European peoples, [In:] Physiology of Heat WULSIN Regulation and the Science of Clothing, ed. : L. H. NEWBURGH, Saunders, Philadelpbia. WYOND. P., LIDWELL 0. M., and WILLIAMS R. E. 0. (1968), Thermal comfort during surgical operations, J. Hyg. 66, 229-248, Cambridge. WYNDHAM C . H . (1963), Thermal comfort in the hot humid tropics of Australia, Brit. J. Industr. Med. 20, 110-117.
Chapter 16
THE EFFECTS OF MODERATE HEAT STRESS ON MENTAL PERFORMANCE” D. P.
WYON
National Swedish Institute for Building Research, Box 785, 801 29 GBvle, Sweden.
I. ANDERSEN and G. R. LUNDQVIST Institute of Hygiene, University of Arhus, Universitetsparken, 8000 Arhus C, Denmark.
CONTENTS Introduction Method Experimental conditions Performance tests Results Comfort Skin temperature variation Performance analysis Discussion Conclusions INTRODUCTION
PROVINS (1966) has suggested that the effect of moderate heat stress is to lower levels of arousal, while higher levels of heat stress tend to increase arousal. A similar suggestion had earlier been made by POULTON and KERSLAKE (1965) in the contex, of rapidly rising body temperatures. The effect is believed to be a largely unconscioust adaptive behavioural response : as thermal stress rises beyond the range in which vasodilatory control of the heat balance is effective, sweating must progressively increase and take over control of the heat balance. However, as the onset of sweating takes time and tends to be aversive for clothed, sedentary workers, the immediate conscious or unconscious response is to relax as much as possible, thus immediately reducing bodily heat production and perhaps postponing or avoiding the onset of sweating. One way to achieve this aim is to intentionally work less hard, i.e. to
* This study was supported by Statens almindelige Videnskabsfond, Statens laegevidenskabelibe Fond, Statens teknisk Videnskabelige Fond, Copenhagen, Denmark and by Statens Institut for Byggnadsforskning, Stockholm, Sweden.
252
D. P. WYONet al.
exert less effort. In the experiment of PEPLERand WARNER (1968), subjects reported that they exerted least effort at 27 "C - the temperature at which they were most comfortable - and worked least well a t this temperature. Their rate of working at constant accuracy, was higher both above and below 27 "C. They worked best at 20 "C.WYON(1974) has shown that up to 50 yoless work was performed in a typewriting task at 24 "C than at 20 "C under experimental conditions, and the performance of school work was shown by HOLMBERG and WYON(1969) to be worse at 27 "C that at 20" or at 30 "C. The above evidence suggests that moderate heat stress affects the performance of mental work, by reducing arousal to a minimum just before sweating becomes effective as a means of thermoregulation. Well-motivated subjects can maintain mental performance in moderate heat stress, as shown by numerous published and unpublished studies indicating no effect of heat upon mental performance of temperatures below 36 "C (e.g. the studies reviewed by WING, 1965). WYON(1969) showed that effort, as indicated by a measure of cardiac variability, increased significantly at 27 "Cwhen subjects maintained their performance in spite of the moderate heat stress. These experiments were carried out under constant temperature conditions. Subjects were thus able to perceive the need for increased effort to maintain performance not only by means of their sensations during the exposure but also by experiencing the contrast, as they entered the exposure chamber, between external and exposure conditions. The present experiment was designed to investigate whether moderate heat stress would have a more marked effect on mental performance if this latter stimulus to compensatory effort were removed by exposing subjects to slowly rising temperatures. The temperature-time profile was chosen to be typical of conditions in a room occupied by a number of people on a hot day, the temperature rising by 3-4 "C in the course of an hour before being reduced temporarily, as if by flush ventilation or by a short break in which the room was unoccupied. A variety of mental tasks was used to investigate the nature of moderate heat stress effects on mental performance in more detail. METHOD EXPERIMENTAL CONDITIONS
The experimental conditions and an analysis of the thermal comfort sensation of the subjects have been given by WYONet al. (1972). The subjects were 72 healthy Danish high-school pupils of high ability, 36 male and 36 female; they were volunteers, but paid by the hour. Each subject attended only once. None of them were heat-acclimatized. They were randomly assigned in single-sex groups of 4 to one of three temperature conditions, shown in the lower part of fig. 16.1. The control condition (0) was a saw-tooth variation of temperature rising from 20 to 23 "C in the course of 50 min., then decreasing to 20 "C in the following 10 min. break, for 3 identical cycles. In condition I a base-line increase of 1 "C per hour was imposed on the saw-tooth variation of condition 0 , with the result that the temperature rose 4 "C in 50 min. and fell 3 "C in the break, starting as for condition 0 at 20 "C and thus ending at 26 "C after 3 hours. In condition 2 the same variation occurred but
Heat stress and mental performance
153
at a higher level: the temperature was initially 23 "C and thus became 29 "C after 3 hours. The temperature was measured as the arithmetic mean of air temperature and radiation temperature and will be referred to as temperature only. The climate chamber, described by ANDWENand LUNDQVIST (1970), was designed for dynamic studies and the walls inside the insulation had a very low thermal capacity. The absolute humidity of the air was constant at 2.53 kPa (corresponding to 50% RH at 20 "C),the air velocity was 12&3 cm s-l and the air supply rate was 600 m3 h-l with no re-circulation.
g!m
570
-M) ZSLl LO
6xl
Fig. 16.1. The two upper sets of curves show the variation with time of the mean dial comfort vote. The boundaries between the cold and comfortable zone and between the comfortable and the hot zone are shown as horizontal lines at 33 and 67 respectively, on the 0-100 scale. The lower set of curves show the variation with time of ambient temperature during the three experimental conditions 0, 1 and 2. The hatched columns represent the rest periods
2
2o
b.'
'
8" 23
1;i 22
$0
8;
,& ,;o
;1
l;o
Time (mini
The subjects wore standard cotton-drill suits. Cotton shorts and a cotton T-shirt were worn underneath, together with cotton socks and light canvas shoes. The total insulation value of this assembly was 0.11 m2K W-' (0.7 clo). As the suits are very similar to the denim jeans and jackets that are frequently worn in school by Danish pupils of this age, both male and female, they can be considered a fairly typical "basic school clothing" with deviations in practice being towards higher insulation values, as when a woollen sweater is worn. The exposures took place after lunch each day after a normal morning at school. PERFORMANCE TESTS
Three different test assemblies were used, each taking 50 minutes. Each group of 4 subjects encountered all three test assemblies in the course of their 3-hour exposure. The six possible orders of presentation were replicated three times in a balanced Latin-square design, so that each test assembly was performed by 8 different subjects in each hour of the three temperature conditions, i.e. by one male and one female group of 4. Independent-measure comparisons are therefore possible on each performance test for sex, time and tkmperature condition.
254
D. P. WYONet a]. Sentence comprehension
A translated version of the BADDELEY (1968) test of grammatical reasoning was used. Subjects worked through a typed list of statements, marking them as true or false. In the original version the statements are of the kind: A follows B : BA B is not preceded by A : AB In Danish the passive voice is d e d far less often than in English. The present version therefore used the following sentence constructions, literally translated before A stands B, B stands before A , after B stands A , A stands after B - together with theiI negated from, e.g. “B star ikke efter A” ( B does not stand after A ) or “foran B star ikke A” (before B does not stand A ) . Note that all of these examples correctly describe BA, but were equally often used to (incorrectly) describe AB. The 8 different statements could contain the letters A and B in either order and could describe AB or B A ; the 32 different sentences thus formed were printed in pseudo-random order with a rolling restriction on their recurrence. Subjects attempted as many as possible in 5 minutes, yielding measures of both speed and error frequency. Multiplication
Each unit of this test consisted of two 3-digit numbers, without zeros, printed one below the other, to be multiplied together. There was space below each unit to set out the calculation in the conventional way. Subjects were provided with several different sheets, each bearing 5 columns and 5 rows of these units. Subjects attempted as many units as possible in 15 minutes, yielding measures of both speed and error frequency. Word memory
A list of 25 common words, chosen at random from a much longer list, was presented visually to the subjects, who read each word aloud in unison with the experimenter. The list was removed and subjects then worked through a list of 50 common words including the 25 “target” words. The task was to classify each word as seen/unseen and each answer as certainluncertain. After a further 20 minutes, during which a reading task was performed, the same list of 50 words was again presented, arranged in a different random order, and was to be similarly classified. This method has been used again in moderate heat stress by one of the present authors (D.P.W.) in an experiment described by LANGKILDE et al. (1973). Cue-utilisation
Subjects performed three parallel version of the TSAI-PARTINGTON test, desciibed by AMMONS (1955), at intervals during one 50 minute period. Twenty-five circles containing letters or numbers, placed at random on a page, must be linked by drawing a line through 1-A-2-B- and so on. The score was taken as the number of correct links made in 40 seconds. Performance of this test is known to be adversely affected by a high level of motivation (EYSENCK and WILLETT, 1962) and moderate heat stress has been shown by WYON(1969) to improve performance of the test, probably by lowering arousal and motivation.
Heat stress and mentaf performmice
255
Other perforinance tests
Only brief mention will be made of the remaining performance tests. SpelIing ability was tested by presenting a list of short sentences, each including one commonly mis-spelt test word. In half of the sentences the test word was incorrectly spelled. The spelling in each sentence was to be classified as correct/incorrect and each such answer as certain/uncertain. Vocabularywas tested by presenting a list of definitions of words, half of which were wrong. Each definition was to be classified as correct/ incorrect and each such answer as certain/uncertain. Spelling and vocabulary performance were scored in the same way as the word memory test, described below. A 15-minute test of reading speed and comprehension was used, in which subjects had to select the correct word from 3 alternatives at regularly occurring choice-points in the text. The choice could be made only if the test had been read and fully understood at the choice-points in question. The method has been used in a Swedish version by HOLMBERG and WYON(1969) and in an EngIish version by W ~ o ~ ( 1 9 6 to 9 ) show the effect of moderate heat stress on 8-1 1 year-old children. Two open-ended tests of creativity were used; subjects had to list as many novel uses as possible of some familiar objects, and as many instances as possible of certain simple kinds of objects, e.g. red objects, square objects. These tests have also been used in moderate heat stress by WYON(1969). Finally, a 15 minute test of manual dexterity and perseverance was included: using an IBM "Port-a-Punch", subjects had to enter eight 5-digit random cards. These cards look like standard SO-column punch cards, but on every other column the punch holes are pre-scored so that it is possible to punch out the required hole with a blunt point. A transparent plastic cover with perforations ensures that the point can only be pressed through at the right places. Since the cards can be read into a computer, the test can be automatically scored in terms of speed and errors.
RESULTS COMFORT
Thermal comfort was recorded using a dial-voting technique. It may be seen from the average comfort-vote curves in the upper part of fig. 16.1 that the conditions were judged to be comfortable throughout except that themale subjectsbecame slightly too hot towards the end of the second and third hours of condition 2, i.e. at temperatures above 27 "C. The distribution of comfort votes about the mean was such that no more than 5 yo of the male subjects were too hot at 23 "C.It could be shown that subjects were more often distracted by their thermal sensations when they were uncomfortable - they adjusted the voting dial more often. Twenty eight percent of the female subjects were too cold at 23 "C, but in practice they would be able to wear more clothing. 23 "C was therefore recommended as a suitable maximum temperature for school classrooms on the basis of the comfort analysis (WYON et al., 1972).
D. P. WYONet al.
256
SKIN TEMPERATURE VARIATION
In fig. 16.2 the variation of mean skin temperature (10 points) with time is shown for the three temperature conditions. It may be seen that there was no appreciable hysteresis: mean skin temperature was close to a linear function of temperature, regardless of the time or foregoing temperature variation. In fig. 16.1 it is apparent
0
' c -
Condition
Symbol
0
0
1
a
2.
n
31-
30 I
I
I
I
I
I
22
23
2L
25
26
27
9 -
;
3L-
F 8 E
33-
c
c Y lJl
33
Air
temperoture
Air
temwmture
I 28 (OC)
-
31 -
30
I 22
23
21
26
2s
27
28
("C)
Fig. 16.2. Variation with temperature of the mean skin temperature. The points shown are the mean values at the start and end of each 50 minute rising-temperature condition for the three experimental conditions 0, l and 2
that comfort votes also followed temperature without hysteresis. In the following analysis of performance it is therefore assumed, unless a time effect is present in the control condition, that interactions between temperature condition and time may be interpreted as being due to the rising temperatures, i.e. that performance was also a direct but not necessarily linear function of the prevailing temperature at any given time.
Heat stress nad mental perjormance
257
PERFORMANCE ANALYSiS
Data fulfilling the necessary assumptions were examined using analysis of variance in a 3 x 3 x 2 independent measures design, with temperature condition, hour and sex as independent variables, respectively. F-values are stated when significant and the raw data is tabulated in these cases. A simpler non-parametric analysis, making fewer assumptions of the data, has been used instead when the data appears to be not normally distributed, and to confirm the influence of single variables when interactions are indicated by the analysis of variance. Sentence comprehension
The analysis of variance indicates a significant effect of time, F (2,54) = 3.7 ( P < 0.05), and a significant interaction between temperature condition and time, F(4,54) = 2.77 (P< 0.05). Raw scores are set out in table 16.1. Kruskal-Wallis analysis (SIEGEL,1956) for the effect of time in the control condition, pooling male and female data, yields a chi-square on 2 degrees of freedom (d.f.) of 4.061 (0.10 < < P < 0.20). The trend, shown in the upper part of fig. 16.3, is the usual tendency to work less hard at difficult or boring tasks in the middle period of an exposure. The effect of temperature must therefore be tested for each hour separately. KruskalWallis analysis for differences between temperature conditions yields chi-square values on 2 d.f. of 2.029 (NS), 1.805 ( N S ) and 7.726 (P < 0.05) for the first, second and third hour, respectively. This test was performed between the 24-th and 29-th Table 16.1 Sentence comprehension: units attempted in 5 min -
~~~
Condition 0
Hour 1
Mean Hour 2
Mean
Hour 3
Mean
Condition 1
Male 46 47 50 34 44.3
Female 47 20 29 58 38.5
Male 43 37 36 42 39.5
36 38 22 37 __ 33.3
51 38 37 26 -38.0
49
46
56 32 48
47
33 41
46
40
-
-
23 60 38
40 35.3
Condition 2
Female 37 25 40 46 37.0
-
-
I
41 50 37 54 45.5
53 22 24 36 33.8
Male 29 41 44 41 38.8
Female 42 42 27 21 34.5
34 52 42 35 40.8
29 41 24 41 33.8
44
54 55 54 49 53.0
_
38 41 68 47.8
_46.3_ 51 - - -3938.3_ _ _ _ 47.5
minute of the hour, by which .time the temperature had risen to within 0.5 "C of the final temperature for each hour (fig. 16.1). The comparisons were therefore between 23,24,27 "C in the first hour, 23,25,28 "C in the second hour and 23,26,29 "C 17
- Bioengineering
D. P. WYONet al.
258
in the third hour. The significant third-hour effect is shown in the lower part of fig. 16.3. Subjects performed this test over 20% more slowly at the intermediate temperature of 26 "C.There were no significant effects upon the error frequency, which was on average 5%. Sentence comprdmon
(010
I
Hour 1
Hour 2
Hour 3
TWOtUre
50-
(P
P
E+
0 L1
5
Fig. 16.3. Performance of the sentence comprehension test. The mean number of units attempted in 5 minutes is shown as a function of time at 23 "C in the upper part of the figure and as a function of temperature in the third hour in the lower part. Each point
LO-
30
I
Mulriplication
The analysis of variance indicates a significant interaction between temperature condition and sex, F(2,54) = 3.24 (P< 0.05). Repeating the analysis for each sex separately indicates a significant effect of temperature condition for the male subjects, F(2,27) = 4.37 (P < 0.05), but no significant effect for the female subjects. The raw scores are set out in table 16.2. This test was performed between minutes 32 and 47 of the hour. The male subjects worked more slowly in condition 2. In fig. 16.4the scores are shown for each sex, as a function of the average final temperature in conditions 0 and 1, but as a function of the temperature experienced by each group in condition 2 to show in more detail the trends around the critical temperature. Kruskal-Wallis analysis of this grouping, with cell frequencies of 12, 12, 4, 4 and 4, yields a chi-square on 4 d.f. of 10.605 for the male subjects (P < 0.05) and 3.644 for the female subjects (NS). The test was performed most slowly by the male subjects at 28 "C. Although no significant effect of temperature can be shown for the female subjects, fig. 16.6 shows that their performance tended to a maximum
Heat styess and rnenfnl performonce
259 Table 16.2
Multiplication: units attempted in 15 min Condition 1
Condition 0
Hour 1
Mean
Hour 2
Mean Hour 3
Mean
Male 31 21 15 -23 _ 22.5
Female 22 13 34 _13 _ 20.5
28 24 22 21 23.8
22 22 18 17 19.8
Male 18 26 20 _15 19.8 17 25 17
Condition 2
Female 19 19 13 -18 17.3
Male 17 12 21 18 17.0
Female 19 25 17 25 21.5
17 19 29 24 22.3
12 12 11 17 13.0
24 19 26 28 24.3
16 27 14 19 _ . 19.0
10 20 26 20 19.0
16 17 29 16 19.5
20 - - - - 19.8
20 18 20 18 19.0
_
26 21 13 24 _ _ 21.0
21 18 17 31 21.8
_
Mu1t i pl I cat ion
3 Ea
4
(NS)
25-
/A
c
c.
0
test by male and female subjects as a function of temperature. Each point represents the mean number of units attempted in 15 minutes
/ \
(P < 005) I
I
I
I
rather than a minimum at this temperature. There were no significant effects on the error frequency, which was on average 23%.
260
D. P. WYONet al. Word memory
Using the method outlined by GREEN and SWETS (1966) the rating scale replies were used to construct for each subject the ROC (Receiver Operating Characteristic) curve of signal detection theory, on the analogy that recognition memory involves the detection of a “signal” (memory trace) in noise. The ordinate is taken to be the probability of the subject correctly recognising any given word and the abscissa is taken to be the probability of false-positive recognition, i.e. that he will incorrectly “recognize” the word. Both can take any value between 0 and 1, depending on the subjects’ ability to remember words and his willingness to risk saying he does in error. These two aspects of the subjects’ performance can vary independently: with a given ability to remember, the subject can appear lo recognize more words correctly by saying he does more often, i.e. by increasing the probability of false-positive recognition commensurately. The relationship between the two probabilities defines the ROC curve. A straight diagonal through the origin represents straightforward guesswork. If subjects had no memory of the list, they would achieve on average as many false-positive recognitions as correct recognitions, regardless of how certain they said they were of their replies on the rating scale. The cumulative frequency of correct and false-positive recognitions for each category of the rating scale defines three points on the ROC curve, which must also pass through the origin and the top right-hand corner. The area under the ROC curve thus defined by five points can be calculated and is a non-parametric, dimensionless measure of recognition memory performance, regardless of criterion. The slope of the tangent at the point defined by the centre of the rating scale can also be estimated, and is a measure of the criterion used by the subject. In signal detection theory it is called the likelihood ratio “Beta”. The value of Beta increases if the subject is reluctant to make falsepositive recognitions, i.e. it is a measure of his degree of caution in recall. In the present case Beta was taken to be the slope of the line joining the 1st and 3rd points defined by the rating scale reply frequencies. Where subjects did not use all categories of the rating scale, no value of Beta was therefore derived. However, since the approximate locus of the ROC curve can be determined from a single point, the area under the curve was obtained in every case. Having read the target list 20 minutes from the start of the hour, subjects attempted to remember the words twice, between minutes 22 and 21 and between minutes 43 and 48. The ROC curve analysis was carried out for each occasion separately, but the trends were found to be the same and the final analysis was performed on the average for each subject of the values obtained on the two occasions. Memory
Analysis of variance indicates a significant interaction between temperature condition and time, F(4,54)= 2,81 (P < 0.05). No significant effects of time or sex can be shown under the control condition 0, but taking rising temperature conditions f and 2 together, there are significant interactions between temperature condition and sex, F(1,36) = 4.98 (P < 0.05). The ROC curve areas are set out in table 16.3 and the trends with temperature for each sex are shown in fig. 16.5.
Heat stress and mental performance
261
T a b l e 16.3 Word memory :Area under ROC curve Condition 0
Hour 1
Mean : Hour 2
Mean : Hour 3
Mean :
Condition 2
Condition 1
Male 0.851 0.726 0.804 0.803 0.796
Female 0.833 0.912 0.705 0.790 0.810
Male 0.768 0.750 0.769 0.716 - _ 0.751
0.823 0.906 0.792 0.768 0.822
0.740 0.682 0.848 0.763 __ 0.758
0.883 0.848 0.751 0.858 ___ 0.835
0.838 0.849 0.716 0.616 __ 0.755
0.801 0.843 0.813 0.727 -0.796
0.678 0.808 0.846 0.553 0.721
0.795 0.750 0.731 0.859 __ 0.784
0.881 0.821 0.859 0.950 0.878
0.953 0.883 0.816 0.710 0.840
0.772 0.733 0.659 0.783 0.737
-
-
Female Male 0.782 0.676 0.788 0.851 0.796 0.841 0.841 _ - _0.789 _ 0.775 0.816
I
8
Word
80 .
Female 0.761 0.918 0.850 0.843 0.843 0.811 0.901 0.794 0.770 0.819
I _
0.842 0.809 0.843 0.875 0.842
memory
4-
kg
70
5
s
60
8 Fig. 16.5.' Performance of the word memory test: recall as a function of temperature for male and female subjects. Each point represents the mean of four different subjects except at 23 "C,at which temperature 12 subjects of each sex performed the test. A11 subjects performed the test twice. performance is shown as the percentage improvement on chance performance 0.5, so that 50% represents a Receiver Operating Characteristic (ROC) curve area of 0.75
2
2
11
1 25
23
Air
29
27
temperature
('C)
D. P.
262
WYONet al.
Memory performance increased with rising temperatures up to a maximum at 26 "C, the reduction at temperatures above 26 "C being most marked for the male subjects, The increase from 24 to 26 "C can be shown to be significant: Kruskal-Wallis analysis, pooling male and female data, yields a chi-square on 2 d.f. of 6.041 (P< 0.05). In the same way, the decrease from 26 to 29 "C yields a chi-square on 3 d.f. of 5.135 (0.10 < P < 0.20). Criterion
Nine of the 72 subjects did not use all 4 reply categories on the rating scale, so no Beta value could be calculated for them. Analysis of variance, carried out using the unweighted cell means and a within-cell mean square with reduced degrees of freedom as recommended by WINER(1962) for missing data, indicates a significant interaction between temperature condition and time, F(4,45) = 3.19 (P < 0.05). There were no significant effects of time or sex under the control condition, but taking rising temperature conditions 1 and 2 together, the interaction between temperature condition and time is even more significant F(2,34) = 6.15 (P< 0.01). The Beta values are set out in table 16.4 and the trends with temperature in fig. 16.6. Beta decreased Table 16.4 Word memory: uncertainty criterion (Beta) Condition 0 Hour 1
Mean Hour 2
Mean __ Hour 3
Mean
Male -
Condition 1
Condition 2
Female
Male
Female
Male
Female
0.930 0.940 1.500 1.260 -1.158
0.430
0.815 1.050 0.530 __ 0.798
0.750 0.545 1.200 0.635 __ 0.783
1.000 1.430
0.820 0.620 0.785 0.670
0.590 0.350 0.750 0.540
0.953
0.724
0.558
0.435 0.41 5 0.515 0.705 __ 0.518
0.725 1.000 0.630 0.375 __ 0.683
0.500 0.850 0.795 __ 0.715
1.005 0.605 0.885 0.875
0.945 0.605 0.835
0.815 1.035
0.780 0.550 1.660 0.870
1.080 1.100 0.425 0.905
0.710 0.410 1.060 0.635
0.620 0.465 1.025
0.704
0.703
-
0.965
0.878
-
-
-
0.843
-
-
-
-
0.860 0.770 1.100 0.960 __ 0.923
0.570 0.945
0.795
0.925
-
0.758
to a minilnuin at 27 "C.The decrease from 24 to 27 "C can be shown to be significant: Kruskal-Wallis analysis, pooling male and female data, yields a chi-square on 3 d.f. of 7.561 (P< 0.06). The subsequent increase of Beta from 27 to 29 "C can similarly be shown to be significant: Kruskal-Wallis rznalysis yields a chi-square on 2 d.€. of 6.157 (P< 0.05). Subjects were most confident of their replies at 27 "C;they were less confident and took more care to avoid incorrect recognition at temperatures both above and below 27 "C.
263
Heat stress and mental performance
Degree of caution in recall
Fig. 16.6. Performance of the word memory test, as in fig. 16.5. Degree of caution in recall, Beta, as a function of temperature. The points at 23 "C are for the control condit ion
I I
8 23
1
I
I
25
27
29
Air temperature ('C)
Cue-utilization
The first version of the Tsai-Partington test was performed in the first 5 minutes of the hour, before the temperature had risen from its initial value. Analysis of variance indicates a significant main effect of temperature condition, F(2,54) = 3.82 (P < 0.05), and a significant interaction between temperature condition and time, F(4,54) = 2.91 (P < 0.05). The scores are given in table 16.5 and the trends with temperature in fig. 16.7. Note the good agreement between the average performance of the 24 subjects in the control condition, who performed the test at 20 "C,and that of the 8 subjects who performed the test in the first hour of condition I , also at 20 "C. Performance of this test in the range 20 to 25 "C shows a clear minimum at 23 "C. The performance of the two parallel versions of this test showed marked learning and either very complex patterns of interaction, mainly between time and sex, or no significant effect at all. The results cannot be interpreted unequivocally in terms of temperature. Other performance tests
No significant effects could be shown on the remaining performance tests except that the female subjects performed the spelling test significantly better than did the male subjects (P< 0.05). The spelling, vocabulary, card-punching and first creativity test were carried out in the first 20 minutes of the hour, before temperatures had risen to any extent, but the reading test and the second creativity test were both carried out after the 30-th minute of the hour, when temperatures had effectively reached their final values.
D. P. WYONet al. Cue - utilization 0 -
o-Contrdl
Y v7
c -
D
2 ffl -
7 -
Q
G
0
6-
5 -
-
1
I
I
21
23
25
("C)
Air temperoture
represents the mean of 8 different subjects tested at each temperature, except that 24 subjects were tested at 2OoC in the control condition T a b l e 16.5
Tsai-Partington: links made in 40 s ____
Condition 0 Hour 1
Male 8 11
Female 3
9 5
Mean
Hour 2
Mean Hour 3
Mean
8.3
3 2 7 3.8
7 8 12 11 12 11 4 11 -~ 8.8 10.3 _.
8 7 6 6 6.8
11 6 6 11 8.5
Condition 1 Male 10 6 7
Female 7 7 5 9
Condition 2 Male 1 7 6
Female 4 5
7
5 -
9 8.0
-
6 8 5 6 6.3
5 7 4
5.3 5 7 5
8 -
8 -
9 8 5 7 7.3
6 5 7 10 __ 7.0
5 6
4 4 4.8
__
7.0
6.0
-
-
6.3
7
5.3
6 5 9 _- 6 6.5
2 8 5 9
__
6.0
DISCUSSION
The present results indicate that the effects of moderate heat stress on mental performance are marked but by no means straightforward, even when comfort sensation and skin temperature vary as closely linear functions of the prevailing temperature. Asystematic relationship between heat stress and mental performance
Heat stress and mental performance
265
has been demonstrated at intervals of only 1 "C, despite the low numbers of subjects exposed to each temperature. Perfomance of the sentence comprehension test demands concentration and clear thinking. POULTON et al. (1978) have recently shown that the performance of a card-sorting version of this test by hospital doctors was made significantly worse by a sleep-debt of only 3 hours. Performance was worse in the third hour of the present experiment, at 26 "C, in comparison with 23 or 29 "C. This corresponds to the decline in the reading speed and comprehension of children found at 27 "C, in comparison with 20 and 30 "C, by HOLMBERG and WYON(1969). Then, as now, the effect can be interpreted as a lowering of arousal at levels of moderate heat stress, corresponding to the limit of vasodilatory control just below the onset of sweating. The multiplication task also demands concentration, but is very well practised and familiar. The male subjects performed least well at 28 "C,but the female subjects were unaffected. This may reflect the distracting effect of thermal discomfort, as the female subjects were not nearly so uncomfortable. The classroom behaviour of 8-year old boys was observed to be more adversely affected by a room temperature of 27 "Cthan by one of 20 or 30 "C in the experiment of WYONand HOLMBERG (1973). Girls were not affected in the same way by the intermediate temperature. Recognition memory was improved by moderate heat stress. Subjects remembered the words better at 26 "C than at temperatures below or above this value. They were (justifiably) more confident of their answers at this intermediate temperature. Using the same test, LANGKILDE et al. (1973) also found that subjects were significantly more confident of their replies (i.e. Beta took lower values) at a temperature 4 "C above their individual optimum temperature. The comparison was between an average optimum temperature of 22.6 "C, and the raised temperature of 26.5 "C. In their case confidencewas misplaced, as their recognition memory had not improved. This supports the reality of the effect as an independent consequence of moderate heat stress. Similarly, WYON(1969) found that his subjects became less critical of their own replies in the creativity test at 27 "C in comparison with 20 or 23.5 "C. EASTERBROOK (1959) sets out the reasons for believing that raised arousal reduces cue-utilisation and can therefore be expected to impair memory. The converse, that lowered arousal improves memory, is consistent with the arousal interpretation of the present results. The effect is quite large: male subjects improved their ROC curve area from about 0.75 at 24 and 29 "C (which is only 50 % better than chance performance 0.50) to 0.88 at 26 "C, which is over 75 % better than chance performance. Since caution and self-criticism demand and even augment concentration it is plausible that lowered arousal would tend to make subjects over-confident and less cautious. The cue-utilisation test was performed in the range 20-25 "C. It was performed least well at 23 "C. Unlike tasks demanding concentration, this test is adversely affected by high levels of motivation: it is performed best by subjects who are not trying very hard. The inference must be that subjects were trying their hardest at 23", which coincides almost exactly with the temperature at which they were optimally comfortable, 23.8 "C. Performance improved at temperatures above
266
D. P. WYONet al.
and below this value as they tried less hard. Performance of similar versions of this test by 11 year-old boys was significantly improved at a constant temperature of 26 "C in comparison with 20 or 23.5 "C (WYON,1969). The same effect has been demonstrated in the present experiment, but in a more restricted temperature range. Reading is such a well-practised activity for 17 year-olds that it is hardly surprising that it was unaffected by the heat. Although a temperature of 27 "C impaired the reading speed and comprehension of 8 year-old Swedish children (HOLBERG and WYON,1969), 11 year-old English children were able to maintain performance a t the same temperature at the cost of increased effort (WYON,1969). The creativity test proved insensitive to the effects of heat for 17 year-olds, although significant effects have been shown at 27 "C in the same test when performed by 11 year-olds: they returned more unique answers, but were less critical of their own responses, as mentioned above. The remaining tests were performed in a very restriced range of temperature and might have been affected by heat stress if performed later in the hour. ANDERSEN et al. (1976) exposed patients with known cardiac defects to the conditions 0 and 2 of the present experiment. Their performance of a numerical addition test, the card-punching test and the creativity test was significantly worse at temperatures of 27 and 28 "C than that of matched control subjects although no differences were found in the control condition. The control subjects, 50 year-old Danish men, performed these tests equally well in the heat and in the control condition. The pattern of response indicated that the patients responded to the heat stress by becoming more tense whereas their controls and the subjects of the present experiment appear to have become more relaxed at raised temperatures up to 27 "C. Thermal comfort is clearly a poor guide to the probable effects of heat on mental performance. In the present experiment subjects appear to have exerted most effort at their temperature for optimum comfort, 23 "C; this was beneficial for the performance of some tasks and detrimental for others. In the experiment of PEPLERand WARNER (1968), subjects exerted least effort at their temperature for optimum comfort, 27 "C; this adversely affected their performance of a programmed learning task. This reflects the difficulty of making subjective estimates of thermal comfort in a meaningful way when comfort is regarded as an end in itself. Instead, thermal comfort should be regarded as a by-product of the interaction between the effects, of the task and the environment on the behaviour of the subject.
CONCLUSIONS
Moderate heat stress, only a few degrees Centigrade above the optimum, has a marked effect on mental performancewhen temperatures rise slowly. Tasks demanding concentration and clear thinking are adversely affected, but memory and cueutilisation can be improved by temperatures up to 26 "C,declining rapidly thereafter. In hot weather, concentrated work should therefore be carried out early in the day. Memory and tasks requiring an increased breadth of attention can with
Heat stress and mental perjormance
267
advantage be postponed until the temperature has risen, provided that it does not rise above 27 "C. The positive and negative effects of these moderate levels of heat stress are likely to be greater for men than for women. REFERENCES AMMONS C. H. (1955), Task for study of perczptual learning and performance variables, Perceptual and Motor Skills 5, 11-14. I., JENSEN P. L., JUNKER P., THOMSBN A.,and WYOND. P. (1976), The efecfs of moderate ANDERSEN heat stress on patients with ischemic heart disease, Scandinavian Journal of Work, Environment and Health 4, 256-272. I. and LUNDQVIST G. R. (1970), Design and performance of an environmental chamber, ANDERSEN Int. J. Biometeor. 40, 402-405. BADDELEY A. D. (1968), A 3-min. reasoning test basedongrammaticalreasoning,PsychonomicScience 10, 341-342. I. A. (1959), The effect of emotion on cue-utilisation and the organization of behaviour, EASTERBROOK Psychological Review, 66, 183-201. EYSENCK H. J. and WILLETTR. A. (1962), Cue-utilization as a junction of drive: an experimental study, Perceptual and Motor Skills 15, 229-230. J. A. (1966), Signal Detection Theory and Psychophysics, John Wiley and GREEN0. M. and SWBTS Sons, New York. I. and WYOND. P. (1969), The dependence of performance in school on classroom temHOLMBERG perature, Educational and Psychological Interactions, No. 31, 20 p., School of Education 20045 Malmo 23, Sweden. K., WYOND. P, and. FANGER P. 0. (1973), Mental performance LANGKILDE G., ALEXANDERSEN during slight cool or warm discomfort, Archives des Sciences Physiologiques27, 511-518. R. D. and WARNER R. E. (1968), Temperatureand learning: an experimental study, ASHRAE PEPLER Trans. 74, 211-219. POULTON E. C., HUNTG. M., CARPENTER A., andEDWARDs R. S. (1978), Theperjormance ofjunior hospital doctors following reduced sleep and long hours of work, Ergonomics 21,279-295. E. C. and KERSLAKE D. McK. (1965), Initial stimulating effect of warmth upon perceptual POULTON efficiency, Aerosp. Med. 36, 29-32. K. A. (1966), Environmental heat, body temperature and behaviour: an hypothesis, Aust. PROWINS J. P S Y C ~18, O ~118-129. . SIEGELS. (1956), Nonparametric Statistics, McGraw-Hill, New York. WIMR B. J. (1962), Statistical Principles in Experimental Design, Mc Graw-Hill, New York. WINQJ. F. (1965), A review of the effects of high ambient temperature on mentalperformance, AMRLTR-65-102, 34 p. WYOND. P. (1969), The efects of moderate heat stress on the mentalperformanceof children, National Swedish Institute for Building Research, D 8/69, 83 p. WYON D. P. (1974), The efects of moderate heat stress on typewriting performance, Ergonomics 17, 309-3 18, WYON D. P., ANDERSEN I., and LUNDQVIST G. R. (1972), Spontaneous magnitude estimation of thermal discomfort during changes in the ambient temperature, J . Hyg. Camb. 70,203-221. I. (1973), Systematic observation of classroom behav our during moderate WYOND. P. and HOLMBERG M. A. HUMPheat stress, [In:] Thermal Comfort and Moderate Heat Stress, eds. :F. J. LANGDON, H R E Y ~and J. F. NICOL,Building Research Establishment, Report 2. 284 p. London : HMSO.
This Page Intentionally Left Blank
V. SUMMING-UP
This Page Intentionally Left Blank
Chapter 17
PHYSICS, PHYSIOLOGY AND PSYCHOLOGY K. CkNA Environmental Physics, Institute of Building Science, Technical University of Wroctaw, 50-370 Wroclaw, Poland.
J. A. CLARK Environmental Physics, Department of Physiology and Environmental Studies, University of Nottingham. Sutton Bonington, Loughborough, LE12 5RD. Great Britain.
CONTENTS Introduction Physics Problems of measurement The units of insulation Physiology Terminology Thermoregulation, survival and acclimatization Thermoregulation and comfort Psychology The measurement of comfort Clothing and comfort Comfort in practice INTRODUCTION
The preceding chapters of this book were prepared before the related School at Karpacz, Poland, held in early September 1978. The matters raised in discussion during the School are incorporated in this chapter. They fall naturally into three categories: Physics - the regulation of man’s thermal environment, including clothing. Physiology - the mechanisms of thermoregulation and acclimatization and their variation with health and age. Psychology - the perception of comfort and discomfort and its relation to other competing stimuli. PHYSICS
There is no difficulty in writing physical equations which correctly describe the energy balance of a human being, and common principles lie behind the different et al., NISHI, GAGGE and VOGTet approaches evident in the chapters by J. A. CLARK al. Though each represent the energy balance in different ways, they can be understood
272
K. CENAand J. A. CLARK
easily by others working in the same field. The reasons for the divergence of treatment are twofold : First, each group of workers is concerned with a different problem. For example, the solution of the energy balance for a man working at high metabolic rates can be formulated almost entirely in terms of sweating heat loss and tolerable heat storage, whereas an energy balance for survival in the Arctic need consider sensible heat loss only. The equations, when simplified, therefore appear different in each case. Second, the heat balance equation must be formulated in terms of measurable variables, and it is problems of measurement which preclude exact solutions, except in theory. There is also a third factor which makes an exact solution of the energy balance for a human impractical; physiological variability. R. P. CLARK(chapter 4) describes the distribution of surface temperature over men, measured by thermography. Evidently real people do not approximate at all well to the assemblage of isothermal cylinders and spheres usually used as the basis of “engineering” models of human heat exchange (HOUDAS,chapter 7). Though informative, the results obtained from such models must obviously be treated with caution. PROBLEMS OF MEASUREMENT
The main difficulty in environmental physics is, however, that of measurement Some measurement problems are intractable: it is not possible to adequately map the distribution of air flow even in quite a small inhabited room, except in a research context, but knowledge of the velocity is required to estimate convective heat transfer. However, other measurements, which currently present problems to those interested in human comfort, could be made easily with commercially available instruments. HARDY (1934) long ago developed a thermopile radiometer for surface temperature measurements in physiological studies, but his instrument was not developed for other applications. The net radiometer (FUNK, 1964), which is based on similar principles, has been a standard instrument in the surface-layer meteorology of vegetation for some years. However, it has rarely been applied to study the thermal physiology of animals, though it allows direct measurement of radiant heat exchanges. Despite this, the “globe temperature”, the temperature of a black 6 inch (15 cm) diameter globe, is still used frequently in the specification of radiant fields. Conversion from the “globe temperature” to radiation fluxes is inaccurate :as it requires a simultaneus knowledge of the air temperature and the mean air velocity (for insertion in the heat balance equation for the globe). The globe thermometer also presents problems associated with its long time constant; it is easier to measure the radiation field directly. The net radiometer consists of a thermopile with the sensitive elements exposed on the two opposite faces of a blackened disc. If the disc is held “above” and parallel to a surface which is receiving radiation, the two faces will be differentially heated by the incident thermal and short-wave radiation falling on the “upper” face, and the emitted thermal and reflected short-wave radiation falling on the “lower” face. The thermopile output is therefore directly proportional to the net radiation flux, provided the element is shielded from draughts. Air-inflated domes of thin poly-
Physics, physiology and psychology
27 3
ethylene, which is virtually transparent at all the wavelengths of interest, provide appropriate draught protection (fig. 17.1). The application of net radiometers to measure the radiative energy balance of human subjects has been described by FUNK(1964), CLARK, GENAand MONTEITH (1973) and by CENAand CLARK (1978a)
Fig. 17.1. Net radiocxter. Thc instrument mcasuiies the net radiative energy flux normal to the faces of the thermopile disc
Substitution of 2 black metal cavity of known tempersture for one of the domes (fig. 17.2) also allows direct measurement of the “Effective Radiant Field” which is closely related to the “Isothermal Net Radiation” in meteorological literature (MONTEITH, 1973). There are further points of ambiguity in the physics, again associated with a problem of measurement. For a wet bulb thernixneter the ratio of the transfer coefficients for latent heat, he, and Convection. h,, is the reciprocal of the psychrometric constant, y , 1.e. h,lh, = y - I 5 (17.1)
he can then be written as y-l h,, so that only one transfer coefficient need be used in the energy balance equ-t‘ u 1011. Similarity theory suggests that the ratio will be the same for other surfaces, such as a man’s skin, provided that both transfers are driven by the same mechanisms. However, for surfaces other than an ideal wet bulb, it is difficult to separat: heat transfer by convection from that by thermal radiation. Therefore, when the combined coefficient for sensible heat transfer is used in equation (17.1) an empitical quantityJI ,? must be substituted for y (GAGGE, chapter 5 ; NIsHr and GAGGE,1977). The problem is exxerbated when clothing adds to the resistances to heat transfer, since the 18
- Bioengineering
274
K. CEXAand J. A. CLARK
Fig. 17.2. Net radiometer fitted with a reference cavity. In this form the instrument may be used to measure the radiation field incident from the hemisphere above the exposed face of the thermopile (courtesy of Solar Radiation Instruments, Altona, Victoria, Australia)
importance of the radiative component is likely to differ with fabric structure (CENA and CLARK,1978b). THE UNITS OF INSULATION
We must here digress to consider a problem more relevant to the psychology of the research worker than to that of human comfort. Scientists of all disciplines find it easiest to “visualize” measured quantities when they are presented in units which result in simple integer numbers for commonly experienced values. In the study of the thermal physiology of man one particular unit has continued in widespread use for this reason despite the change to S . I. units. GAGGE, BURTONand BAZETT(1941) defined the empirical clo unit of clothing insulation as that of the everyday clothing required to maintain a sitting subject in comfort in an ambient temperature of 21 “C.However, as in the previous chapters by J. A. CLARK et al. and GOLDMAN, conversion to S . I. units is necessary for the estimation of rates of sensible heat loss HD from clothed persons. The conversion involves an inconvenient multiple, since 1 clo = 0.155 m2 K W-I , therefore, as in equation (3.1 b)
AT
HD
=0.1551’ ’
(1 7.2)
where I’ is the insulation in clo. There seem to be two possible ways of preserving the integer multiple convenience in measurements of clothing insulation. The first is to ascribe the units to the constant
Physics, physiology and psychology
275
term in equation (17.1), so that clothing is measured as a dimensionless number which is a multiple or fraction of “everyday” clothing. The second possibility is to adopt a different form of S. I. unit which also gives convenient values. Quite fortuitously there is such a unit available. Thermal insulation in m2 K W-’ is defined by equation (3.1 a) HD = ATII.
(17.3)
However, we may use the similarity of heat and mass transfer to present thermal resistance in an alternative form. The Lewis relation (EDE, 1967) states that when the Sherwood and Nusselt number are approximately equal, the mass transfer coefficient, hm, (which has the dimensions of velocity) is equal to the heat transfer coefficient,h,, divided by the volumetric specific heat of the fluid, QC,, i.e. (17.4) If in equation (17.4) we replace h, by I - ’ , the reciprocal of the thermal insulation, and h, by an equivalent resistance (where h, = r-l), we obtain the equality (17.5) where the resistance r has units of s in S.I., the inverse of velocity (CLARK, CENAand MONTEITH,1973; MONTEITH, 1973; CAMPBELL, 1977). The equation for sensible heat transfer, with insulation cxpressed in these units, becomes : HD = Qc,AT/r.
(17.6)
CENAand CLARK (1978~)have recently pointed out that the choice of the S.T.P. value for the volumetric specific heat of air, as the constant ec, in equation (17.6), gives a simple integer multiple conversion with the clo, with 200 s m-l = 2s cm-’ = 1 clo within 0.4%. Table 17.1 shows representative values of total insulation for man, expressed in the three alternative units. As previously suggested by the authors the use of the s m-l unit based on the Lewis relation (or rather its scm-‘ submultiple) has obvious merits.
T a b l e 17.1 Representative thermal resistances in m2K W-I, the empirical clo unit, and the alternative based o n the Lewis relation Case Best bed or sleeping bag Warmest practicable Arctic clothing Heavy clothing for outdoors in winter “Normal” clothing Light summer and hot climate clothing Shorts and singlet for exercise
Thermal resistances mz KW-’ clo s cm-1 1.55 0.17
10 5
0.31
2 1
20 10 4 2
0.155 0.077
0.5
1
0.037
0.3
0.6
K. CENAand J. C. CLARK
276
PHYSIOLOGY TERMINOLOGY
There are disturbing ambiguities between the terminology of thermal physiology and that of studies of human comfort. In particular, thermal physiologists use the term “thermal neutrality” (J. A. CLARK et al., chapter 1 ; MOUNT,1974) to define the condition in the zone of least metabolism of the metabolic diagram, the “thermoneutral zone”. There is an obvious danger of confusion with the use of the words “thermally neutral” to describe that environment which produces no sensation of discomfort. The two are different. The “thermoneutral zone”, which corresponds to zone 2, in fig. 17.3, has been the subject of internationally recognized definition .(BLIGH and JOHNS~N,1973). Those concerned with the measurement of comfort should therefore beware of employing similar words to describe the preferred comfort zone, which is a much narrower band of temperatures (2, in figure 17.3) at the upper end of Z , , which MOUNT(1974) termed the “zone of least thermoregulatory effort”. The definition of the “comfort zone” is itself ambiguous since part of the uncertainty in its measurement and definition lies in physiological variability (but we will return to this again later). The variety of words used to define the comfort zone
g!
e
0
Environmental ternperat ure Fig. 17.3. Metabolic diagram showing the thermoneutral zone, Z1, comfort zone Z2, and zone of “least thermoregulatoryeffort”, Z 3 . Line M, is net metabolicheat production, line E the evaporative heat loss from the skin andline HD the total sensible heat loss from the skin (i.e. M,+E+&= = 0, where losses are given a negative sign)
may also have different meanings (FANGER, 1970; MCINTYRE, chapter 13). Perhaps we should regard the comfort zone (2, in figure 17.3) as a single temperature (FANGER’S Preferred Temperature) with associated limits of uncertainty in measurement.
Physics, physiology and psycliology
277
THERMOREGULATION, SURVIVAL AND ACCLIMATIZATION
ROBERTSHAW (chapter 11) has described the severe strain which may be imposed on the very young and very old by “normal” environments. Thermoregulatory strain of comparable severity, though usually voluntary, may also be experienced by healthy adults, during recreation and work (GONZALEZ, chapter 8; HOLM^ and BERGH,chapter 9; VOGTet al., chapter 6). In such circumstances man depends on physiological mechanisms for thermoregulation, just like any other homeotherm. The available strategies are limited in number and we will consider them separately, though they may by applied together. For illustration we will use the response to cold alone. The four sections of figure 17.4 (HILLand RAHIMTULLA, 1965) represent the effects of the four basic responses to cold on the variation of metabolic rate with temperature. There are two indicators of success ;lowering of the critical temperature T* and reduction of the metabolic rate at temperatures below the lower critical temperature. The first strategy is essentially that of the poikilotherm (fig. 17.4a). If the body core temperature is allowed to fall as ambient temperature falls, the lower critical temperature and the energy costs of homeothermy are both reduced. Because of the problems of maintaining high metabolic rates at low body temperatures, this strategy is not usually adopted by advanced homeotherms; the risks of hypothermia are too great. However, there is some evidence that it may be employed by the neonate human during the first few days of life (HEY,1974), perhaps as a device to conserve energy. In contrast, the lowered body temperature sometimes observed in the old is almost certainly associated with hypothermia (ROBERTSHAW, chapter 11). In the healthy adult the risk of hypothermia is greatest in water immersion, when high rates of heat transfer, which may be accelerated by swimming, can produce rapid body cooling (HOLMBRand BERGH,chapter 9). The second strategy is of transient importance only. Immediately after birth the cuticle of a baby is moist, and for several hours rates of evaporation greatly exceed diffusion through the adult skin, i.e. when not sweating (HEYand KATZ, 1969). Drying out of the skin and the consequent reduction in evaporative heat loss, lowers the critical temperature, as in fig. 17.4b. Because of this the thermoregulatory competence of the neonate improves greatly even within the first day postpartum. Burn patients who have lost large areas of cuticle will have a raised lower critical temperature for similar reasons. In consequence they may suffer “cold stress” in an environment which is comfortable for the nursing staff attending them (who have intact skin and, probably, a higher metabolic rate). There is also new evidence (RUTTER, personal communication) that the cuticle of the very premature baby may allow high rates of evaporation for some weeks after birth, presumably due to incomplete development. In consequence it may behave rather like a wet bulb thermometer, and can have a “lower critical temperature” which is above the normal body temperature. In practical terms, because of evaporation at the skin the incubator temperature must be above 37 “C to allow the baby to maintain a body temperature of 37°C.
K. CENAand J. A. CLARK
278
a
b
t
I
T;
I
I
TY T,
aJ
c
P
0 0
L l 0
Environmental temperature
----B
Fig. 17.4. Diagram of the lower part of the metabolic diagram, to shown the effects of: (a) body core temperature T,; (b):evaporative heat loss E; (c) minimum metabolic rate M,, and (d) tissue resistance u,, on the lower critical temperature of the thermoneutral zone, T* (from HILLand RAHUITULLA, 1965)
The third strategy is that indicated in figure 17.4c, regulation of the rate of metabolism within the thermoneutral zone. This has no effect on the cost of homeothermy below the lower critical temperature, but raising of the minimum metabolic rate lowers T* , and may also allow a decrease in the lower lethal temperature. Conversely, low rates of net metabolism per unit surface area are the cause of the high critical temperature of the neonate and elderly. The healthy adult may produce the opposite effect by exercise or by increasing his food intake - though both cost energy. MOUNT(1968), for example, found that the lower critical temperatures of pigs of mass comparable to a man were 25 "C,23 "Cand 17 "C when starved, and fed at maintenance and four times maintenance, respectively. The effects of activity
Physics, physiology and psychology
279
are similar but geater: a man can maintain an activity of about 180 W m-2 almost indefinitely while awake, about three times the "standard" metabolic rate of the resting, but fed, zdult. The corresponding values of T* when naked are about 7 "C and 27 "C. An increase in food intake is a recognized acclimatization response to chronic cold exposure (FOLK,chapter 10). A complicating factor in thermal physiology is that, because of acclimatization, the responses to an acute and a chronic stress may differ. It is also difficult to demonstrate acclimatization properly unless the subjects of measurement are na'ive. Most citizens of Northern Europe and America have had some experience of cold stress but they are unlikely to possess the extremes of adaptation demonstrated by the Ama of Japan or the Australian aborigine. Acclimatization to cold is mainly through the fourth strategy, represented in figure 17.4d. If the naive subject has a maximum body resistance rsl, and this is increased to r,, on acclimatization, then the lower critical temperature is reduced in proportion, so that (17.7) Moreover, at any temperature below T;"the energy costs of homeothermy are also reduced, again in proportion to the change in conductance when below q. Such a change in conductance accompanies complex physiological and biochemical chapter 8; FOLK,chapter 10) but has largely physical consechanges (GONZALEZ, quences. Acclimatization to cold and heat have long term effects as well as those demonstrable over a few weeks or days. Changes in food intake may contribute both in the short term, through their influence on metabolic heat production, and over the longer term, via changes in morphology which may require months or years. There is evidence, for example, that people who habitually expose themselves to the stress of swimming in cold water develop increaseed subcutaneous fat HOLM^ and BERGH,chapter 9). In the animal kingdom it is well known that, where a species has a wide climatic distribution, the individuals from cold climates tend to be more compact and better insulated than those from the tropics (Bergman's Rule). MOUNT (1968) demonstrated that similar morphological difference may be generated by the exposure to chronic cold and heat stress of pigs from the same litter. Perhaps differences in house heating, due to wealth, explain the differences between the lean aristocrat and the short peasant. Such a controlled experiment is obviously impossible with human subjects, but we may conclude that, because of adaptations of this type, the range of competence of the healthy adult is probably greater than can be conveniently demonstrated in the laboratory. There remain, however, some unresolved problems; for ethical reasons we are unable to perform experiments on those most at risk from cold and heat exposure, the old and infirm, the malnourished and the premature neonate. Our knowledge of their thermoregulatory capacity and its changes on acclimatization therefore remains incomplete. One way of approaching this problem is by "detective work"; determining the conditions which have actually caused incidents of death from heat or cold stress (STOLWIJK, personal communication).
280
K. CENAand J. A. CLARK THERMOREGULATION AN D COMFORT
The sensationsof “comfort ”are related to man’s thermoregulatory responses and to his thermal environment in a complex way. The autonomic mechanisms of thermoregulation always try to maintain a certain “set” core temperature. When the core temperature is artificially driven away from the “set” level it is relatively easy to show that any input assisting its restoration is pleasurable (CABANAC, chapter 12) even when the set point is itself displaced from its normal level by fever or exercise. The physiological determinants of comfort are more difficult to define when the set point is not disturbed. Here we may again refer to the “metabolic diagram” (MOUNT,1974) presented here in abbreviated form as figure 17.3. Obviously the sensation of cold discomfort, which is the drive for behavioural thermoregulation, may be associated with cooling of the peripheries when they are vasoconstricted. At the opposite extreme, in the zone of evaporative regulation, discomfort due to heat is often associated with sweating. However, this is only true at rest. During exercise or work sweating may be necessary for comfort (GONZALEZ, chapter 8). For example, BERGLUND and STOLWIJK (1978) present a graph (originally from FANGER,1970) showing the dependence of comfortable (they use the words “neutral thermal sensation”) skin temperature and sweat rates on metabolic rate, redrawn here as figure 17.5. This suggests that at the maximum sustainable work rate, about three times the standard rate of 60 W m-2, the rate of sweating sensed as comfortable has increased from roughly 5 to 55 W m-2, while the corresponding skin temperature drops from almost 34 “Cto 31 “C.The latter value would be shown as uncomfortably cool in studies of comfort in sedentary subjects, though corresponding to a point within the “zone of least thermoregulatory effort”. Within the thermoneutral zone we can consider that physiologicalcontrol of body temperature operates by “tuning” of the mechanisms of heat loss to match a constant metabolic production. In particular, the narrow range of optimal comfort is associated with a gradation of sensation chapter 2). over a very narrow range of mean body temperature (NISHI, PSYCHOLOGY THE MEASUREMENT OF COMFORT
The problem of defining comfort is in part implicit in the way comfort is measured. MCINTYRE (chapter 13) states succinctly that, by definition, a comfortable environment is one in which the occupants say they are comfortable. This presumes not only that the environment is measured, but that the judgement of the occupants must be communicated to the experimenter. Therefore, no matter how precise the measurement of the physical variables which specify the environment, the link between that environment and its rating as comfortable or uncomfortable is through questions put to people. The psychology of both the questioned and the questioner can therefore influence the results of measurements in this field. For example, MCINTYRE (1977) himself has shown that the answers may be different when people are asked “are you comfortable” or “is factor X making you uncomfortable”. Workers attempting to measure comfort should therefore beware of the bias which may be introduced by seemingly innocent forms of question and answer.
Physics, physiology and psychology
281
Where a comfort scale of some sort is used there may be other problems. Even the number of steps on the scale can be significant, and the words used to define them can cause uncertainty. A three step scale is the least ambiguous, with the
!
c
31
30k
29
50
100
150
ZOC
Metabolic rate [ W m - * j
Fig. 17.5. Mean skin temperature and sweat rate for psrsons with a neutral thermal sensation, a various activity levels or rates of net heat production. (Redrawn from BERGLWD and STOLWIJK 1978; originally from FANGER, 1970)
categories “too hot”, “comfortable”, and “cold”, but this is too coarse for normal use. In some circumstances people prefer to be “comfortably cool” or “comfortably warm”. Comfort scales with up to fourteen points have been used. However, we should remember that human sensitivity to gradations of the intensity of stimuli is limited, and that we can detect only about seven steps in most. The scale of loudness instructions on musical scores is a good example. The seven point “ASHRAE” and “Bedford” scales of comfort are consistent with this concept, and have been shown to produce reliable results in the hands of many workers when more elaborate scales have proved unreliable. Additional problems may be encountered when the seven point scales are translated from the original English language for applications in countries with a different native tongue. It may be difficult to find an exact equivalent of the original meaning, and both the technical knowledge of the translator and his ability in both languages may be crucial. There also exist cultural differences which can negate the whole comfort questionnaire if it is applied naively in a “strange” reminded that whereas the people of Northern Europe and environment. CABANAC America look forward to a “warm welcome” when visiting friends, those of tropical Africa use exactly the opposite wording.
282
K. CENAand J. A. CLARK CLOTHING AND COMFORT
The temperature which is “comfortable” varies with the clothing worn; for sedentary persons from about 30 “Cwhen naked to 21 “Cin 0.155 m2 K W-l(2 s cm-l or 1 do). The practical upper limit is a little lower, about 27 “C for the minimum clothing acceptable at work. When comfort measurements are part of a survey, variation of clothing between the subjects is therefore a source of some uncertainty. It is difficult to specify clothing insulation with precision, and values given in the literature are generally estimated from a record of the clothing worn by applying empirical formulae, such as those given by SPRAGUE and MUNSON(1973). The origin of such formulae is almost always in sets of measurements of the insulation of similar garments, made on a heated manikin (GOLDMAN, chapter 3). However, it has long been recognized that the movement of human subjects during activity may cause substantial degradation of their insulation (BELDING et al., 1947). There will also be effects of posture and substrate, which may increase the effective insulation for seated persons but are almost uninvestigated. There is an obvious need for further in vivo measurements of clothing insulation (MITCHELL and VAN RENSBURG, 1973), which could allow comparison of the manikin measurements with the actual insulation afforded by the same clothing at various levels of activity. COMFORT IN PRACTICE
One of the most reliable measurements from studies of human comfort is thas of preferred temperature. In a long series of measurements P. 0. FANGER and hi” colleagues have shown that, whatever their race or place of origin, people “prefer : the same temperature (when wearing the same standard clothing). There is one snag. they are asked to select this temperature when they have little else to think about In a rather simiIar vein, if pigs are confined in a cold box they will quickly learn to operate a switch to obtain infra-red heat, and can be shown to keep themselves thermoneutral by this means (BALDWIN,1974). They too are presumably selecting a “preferred temperature”. Unfortunately, if they are presented with the choice between a sty with an infra-red heater and switch and a much colder and windier paddock they spend most of their time in the paddock. Presumably the paddock i s more “interesting” than the maintenance of thermoneutrality, or rather motivation overrides the requirements of thermoregulation. Similarly, the “preferred temperature” may tell us more about man’s physiology than about his everyday comfort. WYON(chapter 16) observes that man’s performance of work tasks is not necessarily optimal at his preferred temperature, and indeed may improve in conditions which are uncomfortable by conventional tests. There are obviously interactions between the task, the state of thermal comfort and other stimuli. These may affect arousal and work in a complex way, so that performance of some tasks may be iinproved and that of others worsened by the same stimulus. This is obviously an area which warrants much further work. Another conundrum remains : while there is a single “preferred temperature”, in practice, the environments people consider comfortable are essentially those they are used to (HUMPHREYS, chapter 15). This implies that man may acclimatize or be
Physics, physiology and psyclzology
283
habituated to the environment he experiences, over a much wider range than is usually considered desirable in air-conditioned buildings. For practical purposes it is unimportant whether this acclimatization is physiological or through changes in behaviour. The energy costs of air-conditioning are high. If people can be shown to successfully habituate to a wide range of working environments without depression of their performance at work, worthwhile energy savings could be obtained in the control of building climates. REFERENCES BELDING H. S., RUSSELL € D., I. DARLING R. C., and FOLK G. E.(1947), Amfysis of factors concerned in maintaining energy balance for dressed men in extreme cold; effects of activity on the protective value and comfort of Arctic clothing, Am. J. Physiol. 149, 223-239. BERGLUND L. G. and STOLWIJK J. A. J. (1978), The use of simulation models of human thermoregulation in assessing acceptability of complex dynamic thermal environments, [In:] Proceedings of the Symposium in honor of A. P. GAGGE, J. B. Pierce Foundation, New Haven, Connecticut. K. G. (1973), Glossary of terms for thermalphysiology, J . Appl. Physiol. 35, BLIGHJ. and JOHNSON 941-961. CAMPBELL G. S. (1977), An Introduction to Environmental Biophysics, Springer, New York, Heidelberg, Berlin. CENAK. and CLARKJ. A. (1978a), Thermal resistance units, J. Thermal Biol. 3, 173-174. CENAK. and CLARKJ. A. (1978b), Net radiation and heat transfkr through clothing: the effects of insulation and colour, Ergonomics 21, 691-696. CENAK . and CLARKJ. A. (1978c), Thermal insulation of animal coats and human clothing, Phys. Med. Biol. 23, 565-591. CLARKJ. A., CENA K. and M o m J. L. (1973), Measurements of the local heat balance of animal coats and human clothing, J. Appl. Physiol. 35, 751-754. EDE A. J. (1967), An Zntroduction to Heat Transfer, Pergamon Press, Oxford. FANGER P. 0. (1970), Thermal Comfort, Danish Technical Press, Copenhagen. FUNK J. P. (1964), Direct measurement of radiative heat exchange of the human body, Nature 201, 904-905. GAOGEA. P., BURTON A. C., and BAZETTH.C. (1941), A practical system of unitsfor the description of the heat exchange of man with his environments, Science 94, 428430. HARDYJ. D. (1934), The radiation of heat from the human body. I . An instrument for measuring the radiation and surface temperature of the skin, J. Clin. Invest. 13, 593-604. HEYE. N . (1974), Physiological control over body temperature, [In:] Heat Loss from Animals and Man, eds.: J. L. MONTEITH and L. E. MOUNT,Buttenvoths, London. HEYE. N . and KATZG. (1969), Evaporative water loss in the newborn baby, J. Physiol. London 200, 605-619. HILLJ. R. and RAHIMTULLA K. A. (1965), Heat balance and the metabolic rate of newborn babies in relation to environmental temperature; the efect of age and weight on basal metabolism, J. Physiol. (London) 180, 239-265. MCINTYRE D. A. (1977), Sensitivity and discomfort associated with overhead thermal radiation, Ergonomics 20, 287-296. MITCHELL D. and VANRENSBURG A. J. (1973), Assessment of clothing insulation: The problem and its solution using direct calorimetry on exercising men, Arch. Sci. Physiol. 27, A 149-A162. MONTEITH J. L. (1973), Principles of Environmental Physics, Arnold, London. MOUNTL. E. (1968), The Climatic Physiology of the Pig,Arnold, London. MOUNTL. E. (1974), The concept of thermal neutrality, [In:] Heat Loss from Animals and Man, eds.: J. L. MONTEITH and L. E. MOUNT,Buttenvorths, London. SPRAGUEC. H. and MUNSON D. M. (1974), A composite ensemble method for estimating therrnat insulating values of clothing, ASHRAE Trans. 80, 120-129.
This Page Intentionally Left Blank
SUBJECT INDEX Absorption, thermal radiation, 19, absorptivity, thermal radiation - of human skin, 20 - of clothing, 20 acclimatization, 157-167 - to cold, 161-164, 279 - to heat, 137-138, 159-160 - indices, 162 age and comfort, 205-206 air movement, 31 - measurement, 215-216 - thermal sensations, 197 air temperature, 30 - measurement, 215 - thermal sensations, 197, 212-213 alliesthesia, 184 ambient temperature, 30, (see also air temperature) ambient vapour pressure, 31 aquatic environment, 145-154 ASHRAE comfort scale, 197, 201, 281 asymmetric radiation, 208-209, 222-223
Bedford comfort scale, 197, 201, 281 behavioural thermoregulation, 21, 131, 184 black body radiation, 19 blood flow, thermoregulatory, 130 body heat storage, (see heat storage) body insulation, 17, 21, 172 body temperature - in aquatic environment, 146-147 - circadian cycle, 187-188 - heat balance, 18 - mean value, 136 - oestral cycle, 187-188 - regulation, 187 brown fat tissue, 161, 172 buildings - cooled, 243-245 - “free running”, 233, 237-242 - heated, 243-245
Cardiac output, 124-125 characteristic dimension, 22 circadian cycle - set point, 187-188 - thermal preference, 187 circulatory response to exercise, 1% climatic changes, 164-166 clo, 215, 274. clothing, 41-55 - cold environment, 48-49 - comfort, 91,282 - effect of drape and venting, 54 - heat transfer, 34, 82 - insulation, 17, 31,215, 274 - insulation units, 31, 42, 275 - materials, 49-53 - resistance to water vapour loss, 7 - thermal sensations, 197 cold environment - acclimatization, 161-164 - clothing, 48 - discomfort, 94 - elderly, 174 - infants, 170 - stress, 148 - survival, 153-154 comfort, 195-214 - age and sex, 205-206 - air movement, 212-213 - ASHRAE scale, 197, 201, 281 - Bedford scale, 197, 201, 281 - climate, 206 - clothing, 91, 282 - conditions, 203, 255, 280 - definitions, 80-81 - equation, 96 - humidity, 210-211 - measurement, 280-281 - motivation, 181-182 - pleasure, 189-190 - season, 206 - surroundings, 207
286
-
temperature, 196, 199, 233, 238-241, 243 245-248 - zone, 276 conductance, skin, 128-130, 133, 175 conduction, heat, 14, 18, 113 convection - blood heat transfer, 113 - cooling coeficient, 73-74 - effect of posture, 69 - forced, 22-23, 25, 70 - free, 22-23, 25, 66 - heat exchange, 21, 24, 30, 35-36, 43, 66-75 - heat transfer coefficient, 22, 34, 36 - mixed, 23 - in respiratory tract, 33 - in water, 148-149 core temperature, 15 - in water, 146 - (see also temperature epigastric, hypothalamic, internal, oesophageal, rectal) critical temperature, 15-17, 277-279 - lower, 16-17 - upper, 16-17
elderly, 174-177 - cold exposure, 174 - heat exposure, 176 energy balance, (see heat balance) evaporation - discomfort, 133 - heat exchange, 14,30,37-38,83-84,125-129 - respiratory tract, 33, 37 - sweat, 24, 102-108, 281 - water through skin, 24, 38 exercise - discomfort, 132-140 - effort, 141
Fever, 187 floors - temperature and comfort, 211-212,224-225 - materials and comfort, 225 forced convection, 22-23, 25, 70 free convection, 22-23, 25, 66 Fourier’s law, 113
Grashof number, 22, 67, 68 Diffusion of water vapour skin, 37-38 - respiratory tract, 37 discomfort asymmetric radiation, 208-210, 222-223 - body temperatures, 136, 138, 140 - cold, 94, 132, 139 draughts, 223-224 - energy requirements in buildings, 226 - exercise, 132-140 - Aoor temperature, 211-212, 224-225 - humidity, 210-211 - 10~11,221-226 - - radiation draughts, 207-208 - sweating, 135, 138 temperature variations, 210 - vertical temperature gradients, 225-226 - - WUIII, 92, 132-140 diving, cold discomfort, 152 draught, 213, 223-224 dry heat exchange, 33, 46 DuBois formula, 20, 30
~
Effective Radiant Field, 31, 34, 82, 273 Effective Temperature, 15, 85-89, 92-93, 95
Heat balance, 14, 29-30 heat exchange - conduction, 14, 18, 113 - convection, 21, 33, 35-36, 66-75 - dry, 33, 46 - evaporation, 24, 37, 83-84 - radiation, 14, 30, 31, 34-35 heat exposure - acclimatization, 159 - clothing, 49-51 - elderly, 176177 - infants, 173-174 - sweat rate, 176 heat loss - total, 84 - in water, 145-154 heat production, (see metabolic heat production) heat storage, 15,18,30,32,125 - and clothing, 52 heat stress - elderly, 176 - exercise, 132-140 - infants, 173
287
-
performance, 257-266
- sweat rate, 99-109
Heat Stress Index, 100 heat transfer coefficient - convection, 22, 34, 36 - effective insensible, 83, 84 - effective sensible, 83, 84 - radiation, 21, 33, 82 humidity - comfort, 210-211 - relative, 31 Humid Operative Temperature, 85, 95 hyperthermia, 16, 183, 188, 189 hypothalamic temperature, 131, 185-186, 188 hypothermia, 15, 16, 153, 183, 188, 189
Index of Thermal Stress, 101 indices of thermal comfort, 81-97 indices of thermal stress, 99-109 indoor temperature, 241-247 - free running buildings, 241-242 - heated and cooled buildings, 243-247 infants, (see newborn) insulation - body tissue, 17, 21, 172, 176 -- boundary-Payer of air, 21, 23 -- clothing, 17, 31, 215, 274 - external, 17 - measurement, 44-46 - skin, 21 - subcutaneous, 146 units, 274-275 irradiance, 217 Isothermal Net Radiation, 273 ~
Laminar flow, 67-68 Lewis number, 25
Mean radiant temperature, 21, 31, 34, 216 metabolic heat production, 14-17, 30, 32-33 125, 128, 171 - activity level, 215 - response to cold, 147,161-163,170-172,175 mixed convection, 23 models, thermoregulatory, 111-1 19 -- multi-element, 117 - multi-layer, 115 - partial, 114 - whole body, 114
Net radiometer, 272-273 neutral temperature - comfort, 197 - thermal experience, 204-205 newborn, 170-174 - cold exposure, 170 - evaporative water loss, 173-174 - heat exposure, 173 - temperature distributions, incubators, 62, 65 - tissue insulation, 172 New Effective Temperature, 92-96 Nusselt number, 22, 23, 25
Oesophageal temperature - aquatic environment, 136, 146-147 -- exercise, 130 - thermal sensation, 188 oestral cycle of set point, 187-188 Operative Temperature, 33, 81-82, 95
Panting, 129, 160 pendulum effect, 70 performance tests, 253-255 permeability index, 43 Permeation Efficiency Factor for clothing, 37, 83 Prandtl number, 22, 23 Predicted Four Hour Sweat Rate, 100, 106 Predicted Mean Vote, 96 Predicted Percentage Dissatisfied, 201-203 preferred temperature - - climate, 206 - comfort, 197, 282 - season, 206 -~ surroundings, 207 psychrometric constant, 273 Radiation - absorption, 19 - absorptivity, 20 -- area for heat exchange, 82 - asymmetric, 208-209, 222-223 - draughts, 207-208 - emission, 19 - emissivity, 19 - energy exchange, 14, 19-21, 30, 33, 34, 43 - heat transfer coefficient, 21, 33, 82 - incident, 19 - measurement, 272-273 - net, 19, 20 - shortwave, 20
288 - solar, 19
- thermal, 19 - vector, 217 Recommended Working Time, 107 rectal temperature - exercise, 128 - thermal sensation, 183 - water, 152 mflectivity, radiation, 20 Required Sweat Rate, 102-105, 107-109 Reynolds number, 23
Schlieren photography, 22, 66 sensible heat exchange, 16, 33 set point, 187 Sherwood number, 24, 25 shivering, 16,48, 148, 161-162 skin conductance, 128-130, 133, 175 skin temperature - distribution in different environments, 58-60 94 - infants, 62, 65 - mean value, 32, 65, 94, 128, 256, 281 - measurement, 58 - preferred, 184 - thermal sensation index, 94, 132-133 skinfold thickness 146, 153 skin wettedness - cold environment, 94 - dehition, 32 - evaporative heat loss, 37-38, 84 - warm discomfort, 133, 135 standard environment, 92 Standard Effective Temperature, 87, 91, 92 95-97 Standard Humid Operative Temperature, 87,95 Standard Operative Temperature, 86, 90, 97 Standard Operative Vapour Pressure, 86,87,97 subjective temperature, 198 surface area - DuBo~s,20, 30 - external surface, 20 - for convectiveheat exchange, 20 - for radiative heat exchange, 20, 82 survival in water, 153 sweating - acclimatization, 159 - evaporative heat loss, 24-25 - exercise, 62, 130 - heat strain, 107 - sex differences, 176-177
swimming, cold discomfort, 151
Temperature - ambient, 30 - air, 97, 215-216 - body, 32, 112-113, 126, 129, 135, (see also body temperature) - clothing, 197 - comfort, 196, 199, 203, 233, 238-248 - dew-point, 31 - effective, 15 - floor, 211, 22.4 - hypothalamic, 131, 185-186 - indoor, 241-247 - indices, 81-94 - internal, 126, 131, 133-134, 182-184 - mean radiant, 21, 34, 216 - neutral, 197, 204 - oesophageal, 130, 136, 146-147, 188 - operative, 33, 82 - plane radiant, 216 - preferred, 197, 206, 207, 282 - rectal, 128, 152, 183 - regulation, 182, 187 - skin, (see skin temperature) - subjective, 198 - surface, 19 - tympanic, 183 - vector radiant, 217 - virtual, 25 Temperature Humidity Index, 95 thermal balance equation, 114 thermal comfort, (see comfort) thermal discomfort, (see discomfort) thermal efficiency factor, 34, 83 thermal indices, (seeindices of thermal comfort) thermal neutrality, 222, 276 thermal sensations, 131, 132, 202, 207 t herniogenesis - nonshivering, 161, 162, 172 - shivering, 161, 162 thermographic techniques, 58-59 thermoneutral zone, 16, 17, 158-159, 276 thermoregulation - behavioural, 21, 131, 184 - comfort, 280 - exercise, 124-127 thermoregulatory changes - autonomic responses, 185 - sweating, 125-130
2 89
- vasomotor responses, 173 turbulent flow,67, 68, 71
Vapour pressure - ambient, 31 - saturated, 31
1s
- Bioengineering
124, 129-130, 157,
- units, 5, 24 vasoconstriction, 124, 129, 157, 175 vasodilation, 124, 129, 173 ventilation, pulmonary, and exercise, 125 virtual temperature, 25 Wet Bulb Globe Temperature index, 95, 100
.
This Page Intentionally Left Blank