Whole-Body Human Thermal Models: E. H. Wissler

8
Whole-Body Human
Thermal Models
E. H. Wissler
Contents
8.1 Introduction 258
8.2 Evolution of Human Thermal Models 260
8.3 Model Structure 262
8.3.1 Geometry 263
8.3.2 Composition 264
8.4 Physiological Control Functions 265
8.4.1 Regulation of Blood Flow to Muscle 265
8.4.2 Regulation of Skin Blood Flow 267
8.4.2.1 Measurement of Skin Blood Flow 267
8.4.3 Algorithm for Computing Skin Blood Flow 268
8.4.3.1 Active Vasodilation: Thermal Factors 269
8.4.3.2 Active Vasodilation: The Effect of Exercise 269
8.4.3.3 Active Vasodilation: Other Factors 272
8.4.3.4 Cutaneous Vasoconstriction (CVCM): The Reflex Effect
of Mean Skin Temperature 273
8.4.3.5 Effect of Ts on Skin Blood Flow (SkBF) (CVCL) 273
8.4.3.6 Effect of Exercise on Cutaneous Vasoconstriction
(CVCE) 275
8.4.3.7 Combination of AVD, CVCL, CVCM, and CVCE
to Define qs 275
8.4.3.8 Comparison of Computed and Measured Forearm Blood
Flow (FBF) Data 276
8.4.4 Regulation of Sweating 277
8.4.4.1 Measurement of Sweat Rate 278
8.4.4.2 An Algorithm for Computing the Sweat Rate 279
8.4.4.3 Effect of Acclimation to Heat and Fitness
on Sweating 282
8.4.4.4 Effect of Exercise on Sweating 284
257
2009 by Taylor & Francis Group, LLC

258 E. H. Wissler
8.4.4.5 Hidromeiosis 285

8.4.4.6 Comparison of Computed and Measured Whole-Body
Sweat Rates 287
8.4.5 Shivering 288
8.4.5.1 An Algorithm for Computing the Rate of Shivering
Metabolism 288
8.5 Model Performance 290
8.5.1 Passive Exposure to Heat 292
8.5.2 Exercise at Three Intensities and Two Temperatures 293
8.5.3 Passive Exposure to Cold Air 296
8.5.4 Immersion in Cold Water 298
8.6 Conclusions 300
References 301
8.1 INTRODUCTION
The expression human thermal model can be defined in various ways. A broad
definition includes any relationship between one or more bodily temperatures and
environmental and metabolic variables, such as ambient temperature, humidity,
and intensity of exercise. Algorithms included under a broad definition are often
part of schemes for evaluating thermal comfort under various conditions. Such
models usually ignore the geometry and composition of the human body, and
treat physiological variables, such as skin blood flow, shivering, and sweating, in
a highly empirical manner. Narrower definitions of human thermal model limit
use of the term to models in which the temperature field is computed for a reason-
ably faithful representation of the human form and relevant physiological factors.
Although we will mention briefly models included under a broad definition, the
principal focus of this chapter will be on more rigorously defined models.
The usefulness of any mathematical model depends on many factors, the
most important of which is the accuracy and completeness of underlying fun-
damental equations. While the accuracy and speed of computational methods
used to obtain numerical values of various quantities are important, computa-
tional elegance cannot make up for the shortcomings of a fundamentally flawed
model. With a few exceptions [13], human thermal models have been developed
by engineers whose knowledge of, and interest in, human physiology appears to
be limited. In this chapter, we will review in detail physiological phenomena that
are important to human thermal regulation.
Human thermal models serve several useful purposes. Perhaps the most
important purpose is to incorporate diverse physiological and physical phenom-
ena into an internally consistent representation of human thermal regulation.
The difficulty of accomplishing that goal was described by Stolwijk and Hardy in
1966 [4]. They wrote in describing their model,

Whole-Body Human Thermal Models 259
In fact, it can be argued that no completely satisfactory model of a (physiologi-

cal) subsystem such as temperature regulation can be formulated by itself and that
inclusion of all thermal and non-thermal physiological data is essential for a solu-
tion. While recognizing this difficulty as basic in the study of physiology generally,
the separate analyses involved in each subsystem and component are essential in
each case and are therefore justified within limits for the present study.
An important example of interaction between systems is provided by the effect of

exercise on skin blood flow, which is discussed in some detail later.
Another purpose of models is to predict human behavior under potentially
life-threatening conditions. For example, expected survival time during acciden-
tal immersion in cold water is an important factor in managing search and rescue
missions, both in terms of the response time required for a successful outcome
and the maximum time a search should be continued. Laboratory studies involv-
ing human subjects provide useful information about initial cooling rates during
immersion in cold water, but such studies are never carried to life-threatening
limits, and predicting survival time involves significant extrapolation of the
resulting data. A good human thermal model can be very useful for that purpose,
although it also inevitably involves important assumptions, such as the condi-
tions under which a severely hypothermic individual no longer shivers.
A third purpose is to apply results obtained using human thermal manikins to
human performance. For example, protective garments worn by military person-
nel invariably impose considerable thermal stress on the wearer. Instead of test-
ing each garment on 10 human subjects under various conditions, it is preferable
to measure the properties of the garment on a manikin and use those results to
predict human behavior under various field conditions.
There are also important medical applications of human thermal models,
although not much has been done in that regard. The greatest interest to date has
been in preventing hypothermia during surgery, and commercial systems have
been developed for that purpose. On the other hand, patients are routinely cooled
during open-heart surgery to alleviate undesirable effects of cardiopulmonary
bypass. There is also considerable current interest in possible beneficial effects of
cooling victims of heart attack and stroke. It is reasonable to expect additional
medical applications of thermal models.
The final purpose to be mentioned is the prediction of human comfort under
various conditions. Although relating individual perceptions of comfort to phys-
ical variables, such as skin and central temperatures, sweat rate, and skin wetted-
ness, is a rather imprecise art, there is considerable interest in such applications.
Automotive engineers appear to be particularly interested in establishing com-
fortable conditions for passengers.
All previously developed human thermal models define the regulation of
skin blood flow, sweating, and shivering in terms of a central temperature (pre-
sumably representative of hypothalamic temperature) and mean skin tempera-
ture. Relevant control functions are usually defined to achieve good agreement

260 E. H. Wissler
between computed and observed temperatures under various conditions, but

that process does not necessarily yield results consistent with results obtained
in other physiological studies. An important objective of this chapter is to dem-
onstrate that control functions derived from independent physiological studies
yield a model with good thermal properties.
8.2 Evolution of human thermal models

Before we delve into the details of model formulation, it is worthwhile to review
briefly the history of human thermal models. Two of the earliest models are still
in use in modified form by those whose principal concern is thermal comfort
under moderately stressful conditions. The evolution of human thermal models
outlined in this section reflects the authors prejudices, and others might con-
struct a rather different version.
Probably the first model was proposed by A. C. Burton in 1934 [5]. His very
simple model represented the human body as a single homogeneous cylinder
with uniform metabolic heat generation. It suggested that the steady-state tem-
perature profile is parabolic in agreement with the experimental observations of
Bazett and McGlone [6]. Subsequently, Burton and Bazett [7] used the transient
solution of the heat conduction equation for the same system to interpret data
obtained with their bath calorimeter.
Since the results of experimental studies are nearly always reported as central
and mean skin temperatures, it followed quite naturally that a two-node model
would be developed. Machle and Hatch [8] described a core and shell model in
their extensive 1947 review paper summarizing the results of research conducted
during World War II at the Armored Medical Research Laboratory in Fort Knox,
Kentucky, and at the Pierce Laboratory at Yale University. Application of that
model was not discussed and it was probably not very useful, because it neglected
heat transfer between core and shell, assuming instead that core and shell tem-
peratures were linearly related.
A significant milepost in the development of human thermal models occurred
in 1948 with the publication of Penness classic paper [9] on the effect of blood
flow on tissue temperature. Pennes advanced the notion that heat transfer
between blood in small vessels and adjacent tissue is proportional to the product
of the perfusion rate and the difference between blood and tissue temperatures.
Although computed temperature profiles in the forearm had a slightly different
shape from profiles measured by drawing a fine thermocouple through the fore-
arm, the difference was eventually attributed to the way data for arms of different
size were analyzed [10]. Most current detailed models employ the Pennes model
(often referred to as the bioheat equation) to describe heat transfer in perfused
tissue. Implicit in the Pennes model is the assumption that no heat transfer occurs
between blood in small arteries and veins connected to capillaries and surround-
ing tissue. That assumption has been criticized and analyzed in considerable

detail (for a summary of early work, see Charney [11]), with the resulting conclu-
sion that Penness original model probably overestimates the rate of heat transfer
between blood and tissue. Multiplication of the Pennes expression by a factor that
depends on the perfusion rate provides a more accurate result [1214].
Wissler [15] applied concepts developed in Penness paper to develop the first
steady-state, multielement, human thermal model. His model consisted of six
homogeneous cylindrical elements representing the head, trunk, two arms, and
two legs, which were connected by circulating blood. The thermal energy balance
for blood made allowance for the effect of countercurrent heat transfer between
arterial and venous blood. When reasonable values were assigned to metabolic
and perfusion rates in the six elements and allowance was made for respiratory
heat loss, acceptable agreement between computed and measured temperatures
was obtained. That model did not consider the effect of thermal state on blood
flow, sweat secretion, or shivering.
Information gained from studies conducted at the Pierce Laboratory during
the next 5 years contributed greatly to our understanding of human thermoregu-
lation. Transient changes in central (usually rectal) and mean skin temperatures
were recorded during exposure of seated, lightly clad, male subjects to air tem-
peratures ranging from 13C to 48C. Rates of metabolic heat generation and
sweat secretion were determined by partitional calorimetry. The resulting data
are still invaluable for testing human thermal models.
In 1966, Stolwijk and Hardy [4] significantly advanced the art of human ther-
mal modeling with their publication of a theoretical study in which the concepts
of feedback control were applied to human thermoregulation. Although those
concepts were not original with Stolwijk and Hardy, the thoroughness of their
analysis enhanced the validity of human thermal modeling.
The 1966 Stolwijk and Hardy model consisted of three cylindrical elements
representing the head, trunk, and extremities. That model was implemented on
an analog computer, which undoubtedly limited the number of elements and
amount of detail it could contain. A total of seven regions represented the head
core (brain), trunk muscle, trunk core (viscera), extremity core, and a 2 mm thick
layer of skin on each element. The radius and length of each cylinder were defined
so that it had a mass and surface area appropriate to the anatomical region rep-
resented. Heat transfer by conduction occurred between adjacent regions in pro-
portion to the difference in regional temperatures. Blood contained in a central
pool exchanged heat with tissue located in each region. The principal contribu-
tion of that model was that it incorporated control functions for skin blood flow,
sweating, and shivering into a physically reasonable model. Subsequent models
have all employed the approach introduced in that paper.
Two notable models evolved from the 1966 Stolwijk and Hardy model. In 1970,
Stolwijk [1] developed a six-element model that ran on an early digital computer.
Each element was subdivided into four regions representing a central core sur-
rounded by muscle, subcutaneous fat, and skin. The regions were perfused with

262 E. H. Wissler
blood drawn from a central pool. Heat transfer by conduction between adjacent
regions occurred at a rate proportional to the temperature difference between
them. Using physiological control functions for skin blood flow, sweating, and
shivering based on the best available information, Stolwijk and Hardy signifi-
cantly advanced the art of human thermal modeling. Their model was used in
the design and operation of the Portable Life Support System for the Apollo mis-
sions, and it is still employed in several thermal comfort models. For example, the
Berkeley multinode comfort model [16] is based on the 25-node Stolwijk model,
although it has been augmented in many ways.
Also in 1970, Gagge et al. [2] developed a simple two-node model for the pur-
pose of evaluating thermal stress imposed by a given environment. Their objec-
tive was to develop an effective temperature scale that would allow engineers
and environmental scientists to compare thermal environments on the basis of
energy exchange. That simple model still finds application, although one must
question the virtue of extreme simplicity when powerful computational facilities
are readily available. Moreover, a recent study by Jay et al. [17] reaffirmed that
using core and mean skin temperatures to estimate the internal energy content of
the human body is quite inaccurate [18]. However, for the record, we mention an
example of that approach provided by Bruses individual thermal comfort model
[19], which is based on the two-node Gagge model.
The summary presented above is by no means complete. Other variants of
human thermal models evolved from the Stolwijk and Hardy model, but their
existence was often transient, and limitations of space and time preclude includ-
ing them in this document.
8.3 Model structure

Factors that affect human thermal regulation are both physical and physiological
in nature. Physical factors include the geometry and composition of the body.
Heat transfer from exposed skin or clothing is also a purely physical phenom-
enon, as is evaporation of sweat. Ambient conditions, such as fluid velocity,
temperature, humidity, and incident radiant flux, are also physical factors. In
general, the effect of physical factors is well understood and amenable to rational
analysis, although applications to human modeling are often rudimentary. For
example, modern computational fluid dynamic (CFD) techniques should permit
accurate computation of local heat and mass transfer coefficients for the human
body, but that is seldom done. Another physical factor normally treated rather
approximately is heat and mass transfer in clothing.
Under moderate environmental conditions, human beings can achieve
acceptable temperature control by regulating three physiological variables: blood
flow to skin, sweating, and shivering. While those variables respond primarily
to thermal stimuli, usually defined in terms of a central temperature assumed
to be the temperature of the hypothalamus, mean skin temperature, and local

skin temperature, they are also affected by

nonthermal factors. An example is that cir-
culatory responses to thermal conditions may
be modulated by exercise, because the pri- Qr
mary purpose of the circulatory system is to
satisfy the metabolic requirements of tissue,
especially the requirements of critical organs
and active muscle.
8.3.1 Geometry
Nearly all human thermal models approxi-
mate individual elements of anatomy as cir-
cular cylinders, although several represent
the head as a sphere. An early arrangement
is shown in Figure8.1. That one-dimensional
model developed in 1964 [20] assumed that
physical properties and temperature are func-
tions of radial position and time and required
a large central computer for execution. As
more powerful computers became common-
place, additional cylindrical elements were
added to models, and time-dependent physi-
cal properties and temperature were allowed
to vary with both r and q. Axial conduction Figure 8.1 A typical representation
of the human geometry.
is still generally neglected. For example, Fiala
et al. [21,22] developed a 15-element model,
and Qi [23] developed a 54-element model. The author has recently developed
a new 21-element model in which the head is represented by two elements, the
trunk is represented by three elements, and each arm and leg is represented by
four elements. All of the models mentioned allow a reasonably good representa-
tion of major organs, muscles, arteries, and veins.
Allowing physical properties and temperature to vary with angular position
removes a serious limitation of early one-dimensional models, which could not
represent nonuniform boundary conditions on a given element. For instance,
conditions on the anterior surfaces of an individual seated in an air-conditioned
automobile are markedly different from conditions on the posterior surface, and
there is no rational basis for defining average conditions in a one-dimensional
model.
One might expect that allowing temperature to vary with q greatly complicates
numerical analysis, but that is not necessarily true. According to Fiala et al. [21]
two considerations suggest that conduction in the q direction can be neglected
without seriously affecting accuracy. Within the very thin skin region, the prod-
uct of the temperature gradient parallel to the surface and the area through

264 E. H. Wissler
which heat is conducted in the q direction is much smaller than the correspond-
ing product for heat transfer normal to the skin and, therefore, is negligible. In
deeper regions, temperature gradients are small, heat transport by convection is
dominant, and tissue temperature is determined largely by the rates of metabolic
heat generation and perfusion by blood. While those arguments are intuitively
appealing, they appear to be untested.
Several recent papers describe in somewhat nonspecific terms models in which
finite element methods are used to represent accurately the human form [2426].
Those models appear to describe faithfully external features of human geometry,
but it is unclear how accurately they represent internal structural features.
8.3.2 Composition
The temperature field within a region depends on its composition, because ther-
mal conductivity, thermal diffusivity, rate of heat generation, and perfusion rate
are all functions of tissue type. One of the more important tissue components is
subcutaneous fat, which has a relatively low thermal conductivity and low perfu-
sion rate, and is located directly under skin where it can limit heat transfer to cool
environments.
The local thickness of subcutaneous fat has been determined for more than
50 years by measuring the skinfold thickness with a standard calipers. A cor-
relation relating percent body fat determined by underwater weighing to skin-
fold thicknesses measured at four sites was developed in 1974 by Durnin and
Womersley [27], and is still widely used. However, when local subcutaneous fat
thickness derived from skinfold measurement at the site has been compared with
thickness determined from magnetic resonance imaging (MRI) or X-ray images,
the skinfold-derived value has usually been found to be quite inaccurate, often
underestimating the actual fat thickness by 50% [28,29].
In one of the more comprehensive studies, Hayes et al. [30] compared sub-
cutaneous fat thickness determined from MRI images with corresponding
thicknesses derived from skinfold measurements at 89 sites on 20 male and
20 female subjects. They found that skinfold measurement significantly underes-
timated the subcutaneous fat thickness. They also reported that the distribution
of subcutaneous fat varied markedly as a function of percentage of body fat and
gender, which is often not taken into consideration.
Since that report is not generally available, selected data are summarized in
Table8.1. Groups are defined in terms of mean subcutaneous fat thickness (SCF)
as determined by MRI measurement. For females: in Group 1, 12.1 mm < SCF;
in Group 2, 12.1 mm SCF < 15.2 mm; and in Group 3, 15.2 mm SCF. For
males: in Group 1, 4.8 mm < SCF; in Group 2, 4.8 mm SCF < 9.0 mm; and
in Group 3, 9.0 mm SCF. The body segments are as follows: 1 = upper trunk;
2 = lower trunk; 3 = head; 4, 5, and 6 = proximal, medial, and distal segments of
the legs; and 7, 8, and 9 = proximal, medial, and distal segments of the arms. The

Table8.1 Ratio of Local Subcutaneous Fat Thickness to Mean Fat Thickness Determined
by MRI
Females
Body Segment 1 2 3 4 5 6 7 8 9
Group 1 0.64 1.56 0.92 1.45 0.92 0.72 0.76 0.63 0.64
Group 2 0.72 1.70 0.62 1.36 0.85 0.61 0.81 0.55 0.47
Group 3 0.79 1.67 0.72 1.36 0.84 0.60 0.75 0.55 0.45
Males
Body Segment 1 2 3 4 5 6 7 8 9
Group 1 0.77 1.94 0.64 1.53 0.99 1.08 0.39 0.13 0.01
Group 2 0.96 1.67 0.66 1.23 0.84 0.78 0.78 0.25 0.09
Group 3 1.07 1.63 0.68 1.05 0.73 0.50 0.75 0.55 0.27
Source: Hayes et al. [30].
following relationships define true mean SCF in terms of the mean skinfold
thickness (SkF4) measured by calipers at four sites. For females,
SCF = 0.445 SkF4 1.013 (8.1)

and for males,
SCF = 0.721 SkF4 + 7.300 (8.2)
8.4 Physiological control functions

Human thermal models are often defined in terms of passive and active systems.
The passive system refers to the physical system in which the transient tem-
perature field is defined by the bioheat equation with appropriate boundary and
initial conditions, while the active system defines control functions for blood
flow (especially to skin and muscle), sweat secretion, and shivering. As we noted
previously, definition of the passive system ranges from very simple to very
sophisticated. Although a simple system may provide an acceptable description
of resting individuals in a warm environment, a detailed description is usually
required for cold environments when appreciable temperature gradients exist
within the body. We will assume that readers are quite knowledgeable about
methods available to construct the passive system and will focus attention on
the active system.
8.4.1 Regulation of Blood Flow to Muscle

Since convection by circulating blood is an effective transport mechanism for heat
transfer, the manner in which local perfusion rates are computed is an important
attribute of human thermal models. Highly variable blood flow to muscle and

266 E. H. Wissler
skin is especially important. Definition of suitable control functions for the cir-
culatory system is complicated by the fact that blood transports vital chemical
species, as well as heat, and both functions must be performed adequately. The
heart is a two-stage positive displacement pump that supplies a vascular system
in which the resistance of various branches is regulated in a manner that assures
adequate flow to vital organs, such as the heart and brain. Consequently, thermo-
regulatory control of blood flow cannot be treated as though it were an indepen-
dent system. That is discussed in great detail by Rowell [31,32].
The perfusion rate of active muscle affects the temperature increase owing to
enhanced metabolic heat generation during exercise. Typically, the steady-state
temperature of active muscle is 1C higher than the local arterial blood tem-
perature. Although the fundamental mechanisms that determine muscle blood
flow remain obscure, empirical data firmly establish that the perfusion rate of
active muscle increases promptly as the local rate of oxygen consumption (VO2 )
increases. Values of VO2 above the resting rate are determined by the rate at which
external work is done, the mechanical efficiency of the body (typically about
25%), and the relative involvement of various muscles.
The perfusion rate (q) is related to the arteriovenous oxygen difference (O2,av)
by the relationship
VO2
q= (8.3)
DO2 ,av

where O2,av in resting muscle is about 5 ml O2/100 ml of blood, and increases very
rapidly to approximately 13 to 17 ml O2/100 ml blood during exercise. Proctor et al.
[33] observed that the reduction of femoral venous oxygen content occurs at sur-
prisingly low exercise levels (for example, at 20 W), and Nielsen et al. [34] showed
that O2,av in active muscle is not a strong function of temperature.
An important question is whether blood flow to inactive muscle varies with
local tissue temperature. That question was especially pertinent a few years ago,
when venous occlusion plethysmography (VOP) was the predominant method
for measuring skin blood flow in the human forearm during rest and exercise in a
warm environment. That technique involves measuring the volume of a section of
forearm when venous outflow is blocked by applying appropriate pressure at the
wrist and below the elbow. Pressure applied at the wrist blocks both arterial and
venous blood flow, while pressure applied below the elbow blocks only venous
outflow from the forearm. Several studies [3537] established conclusively that
blood flow to inactive muscle does not increase with increasing temperature. On
the other hand, other studies strongly suggest that the perfusion rate of inactive
muscle decreases when muscle is cooled, although the precise nature of thermally
induced vasoconstriction in muscle remains obscure.

8.4.2 Regulation of Skin Blood Flow

Extensive investigation of skin blood flow has been motivated by several con-
siderations. One is that the rate of blood flow to maximally dilated skin is the
order of 7 l/min, which is an appreciable fraction of maximal cardiac output.
Moreover, since cutaneous vasculature is rather compliant, a significant transfer
of blood from the central venous pool to skin occurs during heat stress, which
may adversely affect cardiac performance. Skin is also the only organ in which
blood flow can be measured noninvasively.
Skin blood flow is an important factor in human thermoregulation. Most
models define skin blood flow simply as a function of central and mean skin tem-
peratures. While those temperatures are undoubtedly important determinants
of skin blood flow, more than 50 studies carried out during the past half century
have established that local skin temperature and several nonthermal factors are
also important. The results of those studies can be summarized as follows:
1. Two branches of the sympathetic nervous system affect efferent neural

mechanisms that control cutaneous vasculature of the head, trunk, and
limbs. One branch is an active vasodilator system, and the other branch
is an active vasoconstrictor system.
2. Modulation of skin blood flow owing to cold stress, moderate heat stress,
and exercise can be attributed to variable vasoconstrictor tone. As skin
and central temperatures increase, active vasodilation becomes effec-
tive, and the large increase in skin blood flow seen during severe heat
stress is caused primarily by active vasodilation.
3. Active vasoconstriction responds reflexively to mean skin temperature.
4. Local skin temperature affects local cutaneous vasoconstrictor tone.
5. Active vasodilation increases linearly with increasing central tempera-
ture above a threshold temperature that is modulated by mean skin tem-
perature, exercise, and posture.
8.4.2.1 Measurement of Skin Blood Flow

Prior to 1980 the only quantitative measure of skin blood flow was provided by
forearm blood flow (FBF) measurement using VOP. Around 1980, laser-Doppler
(LD) techniques that detect the velocity of red blood cells as they pass through
cutaneous capillaries were developed for sensing skin blood flow in a small area.
Choosing a laser of appropriate wavelength allows incident light to be almost
completely absorbed or reflected within the skin so that blood flow within under-
lying muscle is not detected [38]. When Johnson et al. [39] compared skin blood
flow measured using an LD instrument with flow measured by VOP, they found

268 E. H. Wissler
that LD flow rates were linearly related to flow rates measured plethysmographi-
cally, although the relationship was specific to each subject. LD instruments offer
the advantages that they can be used on any skin area, they respond rapidly to
changing flow rate, and the temperature of skin in the area of measurement can
be closely controlled.
Since LD measurements do not provide absolute values for skin blood flow
(SkBF), they are usually reported as a percentage of some reference value, usually
either a thermally neutral initial value or a maximal value recorded at a local
skin temperature of 42C. In addition, LD measurements are usually reported as
cutaneous vascular conductance (CVC), which is the ratio of the flow rate to the
mean arterial pressure.
The use of chemical agents to block centrally mediated changes in skin blood
flow is a time-honored technique. When combined with LD measurement of
SkBF, it allows investigators to differentiate clearly responses of the active vaso-
dilator system from responses of the vasoconstrictor system. Kellogg et al. [40,41]
very successfully employed bretylium blockade of active vasoconstriction to iso-
late the effect of active vasodilation on SkBF during exercise.
8.4.3 Algorithm for Computing Skin Blood Flow

Experimental results from studies reported during the past half century can be
represented reasonably well by the following relationship:
qs = qs ,r AVD CVCM CVCL CVCE (8.4)
where the quantity qs is the local cutaneous perfusion rate (typically reported as
ml blood/[100 ml tissue min]), and qs,r is the perfusion rate under reference con-
ditions, which we define as Ts = 34C, Ts = 34.5C, and Tc < the threshold central
temperature (Tc,th) for active vasodilation. AVD defines the centrally mediated
drive for active vasodilation, CVCM accounts for the reflex effect of Ts on CVC,
CVCL defines the locally mediated effect of Ts on CVC, and CVCE accounts for
the direct effect of dynamic exercise on vascular conductance and the increase in
mean arterial pressure (MAP) that occurs during exercise. By definition, each of
the functions has a value of unity at the reference condition.
With the exception of active vasodilation (AVD), definitions of the component
functions are based on CVC data. Forearm blood flow data are generally consis-
tent with CVC data when reasonable assumptions are made about muscle blood
flow. If we assume that plethysmography detects only skin and muscle blood
flows, we have
FBF = X s qs + Xm qm (8.5)

in which Xs and Xm are the volume fractions of skin and muscle, respectively, in
the forearm; and qs and qm are the respective perfusion rates. Cooper et al. [42]

reported that Xs = 0.086 and Xm = 0.636 in five forearms obtained postmortem

from different subjects.
8.4.3.1 Active Vasodilation: Thermal Factors

Defining the separate and collective contributions of central temperature and
Ts to regulation of skin blood flow has been very challenging, largely because it
is difficult to change one temperature without affecting the other. Central tem-
perature can be increased either by raising the skin temperature, or by increas-
ing metabolic heat production through exercise. A complicating factor associated
with the second approach is that moderate to heavy exercise causes a decrement
in skin blood flow under certain conditions.
Several papers published from 1974 through 1975 [43,44] established that FBF
increases linearly with Tc and Ts over certain temperature ranges. If FBF varies
linearly with Tc, qs also varies linearly with Tc when qm is constant. Accordingly,
we define AVD as follows: when Tc < Tc,th, AVD = 1.0, and when Tc,th< Tc,
AVD= 1.0 + (Tc Tc,th) (8.6)
in which is a gain constant. Tc,th depends on Ts and several nonthermal factors.

If we let Tc,th = Tco,th + Tc,th in which Tco,th is the threshold central temperature for
a resting, supine individual when Ts = 33C, Tc,th/ Ts 0.05.
8.4.3.2 Active Vasodilation: The Effect of Exercise

Since exercise and heat both cause cardiac output to increase and modify the dis-
tribution of blood volume, it is reasonable to assume that exercise affects circula-
tory responses to heat. The fact that splanchnic and renal blood flow and blood
flow to inactive muscle decrease during exercise also suggests that skin blood
flow might decrease during exercise. However, that possibility must be tempered
by the fact that skin blood flow plays an important role in thermoregulation,
while splanchnic and renal and inactive muscle blood flows do not. Numerous
early studies stimulated by the possibility that exercise affects skin blood flow
yielded somewhat contradictory results, although a reasonably consistent picture
can now be constructed.
Studies of the effect of exercise on SkBF conducted before 1990 focused on
the vasodilator system, that is, on AVD in Equation (8.4). In that context, exer-
cise could modulate SkBF either by modifying the slope of the Tc SkBF relation-
ship or by changing Tc,th. Johnson [45] was the strongest proponent of the notion
that nonthermal factors, such as exercise and posture, affect SkBF by modulating
Tcth. Early support for that concept came primarily from studies conducted with
an elevated Ts of 38C to 38.5C established by means of a liquid-perfused suit.
Experiments usually consisted of comparing the values of Tc,th during rest with
values observed during exercise of moderate intensity. Tc,th was defined as the

270 E. H. Wissler
1.2
Threshold Central Temperature Increase: C 1
0.8
0.6
0.4
0.2
0
40 50 60 70 80 90 100
Oxygen Consumption Rate: % VO
2,max
Figure 8.2 Tc,th as a function ofVO2,r when Ts 33C. Filled and open circles identify plethys-
mographic and LD data, respectively, of Smolander et al. [48]; open triangles identify data of Taylor
et al., [47]; and open squares identify data from Table 2 of Kenny et al. [49]. The graph of Equation (8.8)
is also shown.
value of Tes at which FBF increased sharply with increasing Tc. In a typical study,
Johnson and Park [46] concluded that Tc,th increases 0.28C with exercise, and
0.11C with the change from supine to upright posture.
Contradictory studies that showed no effect of exercise on Tc,th were conducted
in cool room air at lower mean skin temperature. While it was not apparent at
the time, it now appears that the difference between results obtained by different
investigators can be attributed to the difference in Ts . A lack of awareness of the
importance of skin temperature during the 1970s is suggested by the fact that Ts
was often not reported when exercise was performed in air.
The murky picture of the effect of exercise on SkBF began to clear with pub-
lication of a study by Taylor et al. in 1988 [47]. Those investigators evaluated the
effect of five different intensities of supine cycling in 21.1C air on the cutaneous
vascular responses of four men. Data from their study indicated that exercise had
no effect on CVC at low intensities of exercise, but exercise at intensities above
125 W, which corresponds to VO2 of approximately 45% of VO2 ,max, caused a dec-
rement in CVC at a given value of Tes. Taylor et al. attributed the decrement in
CVC, which was proportional to the workload, to an increment in Tc,th. Values of
Tc,th derived from their data are plotted in Figure8.2.
In a subsequent study, Smolander et al. [48] employed VOP to measure FBF
in six men during 15 minutes of cycling exercise in 25C air. Measurements were

made at five workloads varying from 50% to 90% of VO2 , max . Mean skin tempera-
tures close to 33.5C were independent of exercise intensity. The principal finding
of this study was that SkBF during dynamic work was significantly attenuated
when a subjects oxygen uptake exceeded 80% of his maximum oxygen consump-
tion rate (VO2 ,max ). LD measurement of SkBF and VOP measurement of FBF pro-
vided comparable indications of Tc,th for active vasodilation.
Smolander et al. [48] also attributed the decrement in FBF at moderate work-
loads to an increase in Tc,th. If we assume that Tc,th for VO2 = 0.5 VO2 ,max is the
normal threshold temperature for enhanced SkBF, then threshold temperature
increments (Tc,th) for higher workloads can be computed from values reported
by Smolander et al. The values of Tc,th computed in that way are 0.00, 0.06, 0.17,
0.37, and 0.78C for VO2 = 0.5, 0.6, 0.7, 0.8, and 0.9 VO2 ,max , respectively. Those
values are also plotted in Figure8.2.
Another paper that helped to clarify the effect of exercise on SkBF was pub-
lished by Kellogg et al. [40], who (as mentioned in Section 8.4.2.1) employed
bretylium blockade of active vasoconstriction [41] to isolate the effect of active
vasodilation on SkBF during exercise. Figure 3 in that paper shows clearly that
exercise delays the onset of active vasodilation by raising Tc,th. For the conditions
of their study, Tc,th during exercise was 0.28C higher than during rest. An addi-
tional observation was that the initiation of exercise causes a reduction in CVC
before active vasodilation is initiated.
The studies by Taylor et al. [47] and Smolander et al. [48] established that SkBF
depends on the intensity of exercise under conditions that impose moderate to
high stress on the circulatory system. We have chosen to define the relationship
between intensity of exercise and SkBF in terms of the relative rate of oxygen
consumption,
VO2
VO2 ,r = (8.7)

VO2 ,max

Equation (8.8), derived from the data of Smolander et al. and Taylor et al.,
defines the increase in Tc,th owing to exercise in the absence of heat stress.
Tc,th = 1.41 V 2 0.194 V + 0.01C (8.8)
where DV = (VO2 ,r - VO2 ,crit )/(1.0 - VO2 ,crit ). For the cool mean skin temperatures
employed by Taylor et al. and Smolander et al., VO2 ,crit = 0.5. The relationship
defined by Equation (8.8) is plotted in Figure8.2.
Although Equation (8.8) accounts for the effect of exercise on Tc,th at mod-
erate skin temperatures, it fails to account for the effect of moderately heavy
exercise on Tc,th at high skin temperatures. For example, Equation (8.8) yields

272 E. H. Wissler
Tc,th = 0C when Ts = 38C and VO2 , r = 0.5, which is inconsistent with the
threshold increment of 0.28C observed under those conditions by Johnson
and Park [46] and by Kellogg et al. [40]. A possible solution for that dilemma
is to assume thatVO2 ,crit decreases with increasing Ts , which is not unreasonable,
because cutaneous blood flow and volume both increase when the skin tempera-
ture is high. We will assume that VO2 , crit decreases linearly from VO2 ,crit = 0.5 for
Ts 33C to VO2 ,crit = 0 for 38C Ts . Then, if Ts = 38C and VO2 , r 0.5, as in the
studies of Johnson and Park, and Kellogg et al., the exercise-mediated Tc,th pre-
dicted by Equation (8.8) is 0.26C, which is reasonably close to the observed value
of 0.28C.
Data on which Equations (8.7) and (8.8) are based were derived from stud-
ies in which the onset of active cutaneous vasodilation was observed during a
period of rising central temperature. A very important complementary study by
Kellogg et al. [50] investigated the effect of exercise on established active vasodi-
lation. Conditions of that study were similar to those in the first study by Kellogg
et al. [40], except that moderately heavy exercise did not commence until after
active vasodilation had been firmly established by 35 to 40 minutes of passive
heating. In both studies, CVC was determined at two sites, one treated with
bretylium tosylate to block vasoconstriction and the other untreated. Contrary
to implications of previous observations, exercise in the second study reduced
CVC at the untreated site, and not at the bretylium-treated site, which implied
that exercise modulated cutaneous vasoconstriction, but did not affect active
vasodilation.
Several mathematical solutions for the paradox presented in the preceding
paragraph are possible. One is to assume that the central thermoregulatory cen-
ter processes afferent signals by integrating the rate of change, instead of sub-
tracting Tc,th from Tc. That is, we can assume that
t
dTc
AVD = 1.0 + a
dt dt
to
(8.9)

in which to is the instant at which Tc = Tc,th. If we assume that integration con-

tinues as long as Tc,th < Tc, modifying Tc,th after active vasodilation has begun has
no effect on SkBF. Another possibility is that Tc,th owing to exercise is transient
in nature and diminishes as active vasodilation progresses. That possibility is
completely unexplored.
8.4.3.3 Active Vasodilation: Other Factors

Extrapolation of relationships representative of moderate conditions of heat
stress and exercise to more severe conditions may require cardiac output that
approaches, or exceeds, the capacity of the heart. Consequently, it is necessary to
limit SkBF to physically realizable values. Several studies have shown that during

moderate exercise, the slope of the Tc FBF curve decreases significantly for cen-
tral temperatures above 38C, with the slope above Tes = 38C being roughly one-
half the slope at lower central temperatures [51,52]. Kellogg et al. [53] established
that the reduced slope can be attributed to reduced active vasodilation, which in
our model amounts to attenuating AVD. Therefore, if 38C < Tc and 0.4 < VO2 , r <
0.9, AVD defined by Equation (8.6) is modified as follows:
AVD = AVD* 0.5 (Tc 38.0) (8.10)
where AVD* is the value computed using Equation (8.6).

Moreover, according to Smolander et al. [48], the slope of the Tc FBF curve
is greatly reduced over its entire range at exercise intensities above 0.9 VO2 ,max.
Therefore, when 0.9 < VO2 , r , AVD is computed as follows:
AVD =0.25 (Tc Tc,o Tc,th) (8.11)
8.4.3.4 Cutaneous Vasoconstriction (CVCM): The Reflex

Effect of Mean Skin Temperature
Studies by Charkoudian and Johnson [54], Stephens et al. [55], and Kenney et al.
[56] established that, in addition to shifting Tc,th, Ts has a reflex effect on cuta-
neous vasoconstriction. Since Ts in those studies was lowered from an initially
comfortable level, Tc remained below Tc,th, and centrally mediated vasodilation
was not a factor. Local skin temperature at the measurement site was maintained
constant in each study.
Results from that study are plotted in Figure8.3. All data have been normal-
ized so that CVC is unity at Ts = 32C. A shortcoming of those data is that there
are no experimental data to define an upper limit for the ratio CVC/CVC32C,
although we know that SkBF has an upper limit [57]. The reason for the lack of
data is probably that it is impossible to raise Ts to 42C without simultaneously
increasing Tc, causing active vasodilation. There is also a paucity of data for Ts
below 30C. The curve shown in Figure8.3 is a graph of the function
1.422 + tanh[ 0.275 (Ts - 32.0)]

CVCM = (8.12)
2.018
8.4.3.5 Effect of Ts on Skin Blood Flow (SkBF) (CVCL)

The least ambiguous aspect of SkBF regulation is the effect of Ts on SkBF, which
can be studied under comfortable conditions that do not involve active vaso-
dilation, and for which the effect of Ts is fairly constant. Direct LD measure-
ments of SkBF were made by Charkoudian et al. [58] in their study of the effect
of female reproductive hormones on local control of SkBF. Those data are plotted

274 E. H. Wissler
140
120
100
CVCM: Percent
80
60
40
20
0
29 30 31 32 33 34 35 36 37
Mean Skin Temperature: C
Figure 8.3 Normalized CVC plotted as a function ofTs . (Data were obtained from: open circles, CVC
from Stephens et al. [55] Figure 1 (saline); diamonds, CVC from Stephens et al., Figure 3a (saline);
pluses, CVC from Stephens et al., Figure 5 (saline); and triangles, CVC from Charkoudian and Johnson
[54], Figure 2.)
in Figure8.4. Forearm blood flow data reported by Barcroft and Edholm [35,59],
Brown et al. [60], and Wenger et al. [61] agree qualitatively with the CVC data of
Charkoudian et al. in that they show a strong withdrawal of vasoconstrictor tone
at local skin temperatures above 32C.
6
Normalized CVC
0
20 22 24 26 28 30 32 34 36 38 40 42
Local Skin Temperature: C
Figure 8.4 Values of relative CVC normalized to the value of unity at Ts = 34C. (Note: Also shown is
the graph of CVCL(Ts).) (From Charkoudian et al. [58].)

8.4.3.6 Effect of Exercise on Cutaneous Vasoconstriction (CVCE)

Data from several studies suggest that exercise causes a significant decrease in
CVC. Kellogg et al. [40,50] observed that strenuous exercise reduces CVC by
increasing cutaneous vasoconstriction in both normothermia and hyperthermia.
Taylor et al. [62] observed that moderate leg exercise causes a reduction in blood
flow of a locally heated arm. In that study FBF at four temperatures, 36, 38, 40,
and 42C, was measured by VOP. While there was some variation with Ts, the
mean reduction was the order of 3 ml/(100 ml min).
Taylor et al. [63] evaluated the effect of absolute and relative work, and the
kind of exercise on CVC. They found that the CVC decrement was proportional
to absolute workload, and not relative workload. In addition, they found that
isometric exercise had no effect. Their data are represented reasonably well by the
following relationship:
0.071
CVCE = 1.0 - W (8.13)
110
in which W is the workload in watts.

The decrease in CVC with exercise is offset to some extent by an increase in
MAP. We will assume that MAP increases during leg exercise according to the
relationship
MAP = 85 + 23 VO2 ,r Torr (8.14)

in which VO2 ,r is the fraction of the maximum oxygen consumption rate.
8.4.3.7 Combination of AVD, CVCL, CVCM, and CVCE to Define qs

The four factors described above, AVD, CVCL, CVCM, and CVCE, were derived
from experiments that provide little guidance about the manner in which they
should be combined to form a comprehensive model of SkBF regulation. For
example, experimental investigations carried out at a fixed elevated Ts provide
no information about the combined effect of Ts and Tc. Although several authors
have concluded that responses owing to central and cutaneous temperatures
combine additively, a careful analysis of relevant data suggest that a multiplica-
tive combination is more appropriate.
We note that since the quadruple product, AVD CVCM CVCL CVC, has
a value of unity for the reference state, Tc < 37.0C, Ts = 34.5C, Ts = 34C, and
VO2 , r < 0.5, it must be multiplied by qs,r. Our estimate of qs,r is 14.7 ml/(100 cc of
skin min), which is consistent with an FBF reference value of 3.0 ml/(100 cc min).
Hence, we have
ml blood
qs = 14.7 AVD CVCM CVCL CVCE (8.15)
100 cc of tissue min


276 E. H. Wissler
16
14
12
FBF: ml/(100 ml min)
10
0
37 37.1 37.2 37.3 37.4 37.5 37.6 37.7 37.8 37.9
Esophageal Temperature: C
Figure 8.5 FBF measured by Nadel et al. [52] for exercise at 40% of VO2,max and Ts = 32.0C (circles),
33.7C (squares), and 35.2C (triangles). Filled symbols identify measured values, and open symbols
identify computed values.
A study that supports our hypothesis concerning the combined effect of the
four factors was performed by Nadel et al. [52]. They used VOP to measure FBF in
three relatively fit young men who performed moderate and heavy cycling exer-
cise at three room temperatures: 20C, 26C, and 36C. Corresponding mean
skin temperatures were approximately 32.0C, 33.5C, and 35.5C. The esoph-
ageal temperature varied from 37.1C to 37.7C during exercise at 40% VO2 ,max ,
and from 37.1C to 38.8C during exercise at 70% VO2 ,max . Forearm skin tempera-
ture was not measured. Data from that study are plotted in Figures8.5 and 8.6.
Corresponding values computed using Equations (8.1) and (8.19) and assuming
that Ts = Ts are also plotted. R2 for the computed and measured values is 0.95. Two
important conclusions can be drawn from the data plotted in Figures8.5 and 8.6.
One is that the slope of the Tes FBF relationship increases with increasing Ts , and
the other is that FBF is suppressed by moderate exercise.
8.4.3.8 Comparison of Computed and Measured Forearm Blood Flow (FBF) Data
Although the definitions of CVCM, CVCL, and CVCE are based on relatively
recent CVC data, it would be imprudent to ignore earlier FBF data. FBF mea-
sured by VOP includes blood flow to both skin and muscle. While several stud-
ies have established rather conclusively that forearm qm does not increase with
bodily temperature, the possibility remains that it decreases under hypothermic
conditions, and FBF data tend to support that possibility.
In Figure8.7, values of FBF computed using the algorithm described above
are compared with measured values reported by various investigators. That com-
parison certainly lends credence to the validity of our approach.

20
16
FBF: ml/(100 ml min)
12
0
37 37.2 37.4 37.6 37.8 38 38.2 38.4 38.6 38.8 39
Esophageal Temperature: C
Figure 8.6 FBF measured by Nadel et al. [52] for exercise at 70% of VO2,max and Ts = 32.0C (circles),
33.6C (squares), and 35.5C (triangles). Filled symbols identify measured values, and open symbols
identify computed values.
8.4.4 Regulation of Sweating

Sweating is essential for survival in a warm environment. Various studies have
established that the neutral range of skin temperature for resting individuals is
approximately 33C to 35C. In other words, 35C is the maximum mean skin
25
Computed FBF: ml/(100 ml min)
20
15
10
0
0 5 10 15 20 25
Measured FBF: ml/(100 ml min)
Figure 8.7 Comparison of computed and measured FBF.

278 E. H. Wissler
temperature for which heat generated in metabolically active organs at their nor-
mal temperatures can be transported to the skin. If 80 W of metabolically gener-
ated heat are transferred from skin to air by convection and radiation, and the
combined heat transfer coefficient is 10 W/(m2C), the mean skin-to-air tem-
perature difference is roughly 4C. It follows that resting individuals must sweat
when the air temperature is above 31C.
Basic attributes of sweating are summarized below:
1. Local sweat rate (SR) is affected by Tc , Ts , and Ts.

2. Sweating at a given thermal state appears to be greater during exercise
than during rest.
3. Local sweat rate varies over the surface of the body. In general, the slope
of the Tc SR relationship is greater on surfaces where sweating begins
early as Tc increases.
4. Fitness and heat acclimation both facilitate sweating.
5. Prolonged wetting of skin reduces sweat secretiona process known as
hidromeiosis.
Although sweating is absolutely essential to survival in a hot environment, it

is actually a rather crude thermoregulatory mechanism. Sweat is secreted copi-
ously during severe heat stress, but the rate of evaporative cooling is often limited
by the rate of diffusion of water from skin to air, rather than by the rate of sweat
secretion. In a humid environment, sweating may have little effect on thermal
response.
8.4.4.1 Measurement of Sweat Rate

Continuous weighing of a subject, with suitable allowance for respiratory loss of
mass, is the preferred technique for determining the transient whole-body rate of
sweat evaporation. Local rates of sweat secretion are usually determined by mea-
suring the rate of evaporation under a capsule ventilated with a steady stream of
dry air with resistance hygrometry used to measure the humidity of the effluent
air stream [64]. While that technique is useful for assessing the effect on sweating
of various factors, such as heat acclimation, it is difficult to determine the whole-
body rate of sweat secretion from local values [65].
Sweat rates measured on several skin areas in a single study indicate that
the slope of the Tc SR curve and Tc,th for sweating vary from region to region.
Figure8.8 shows how the local sweat rate at six sites increased with time dur-
ing exhaustive exercise (90% of VO2 ,max ) in 30C air [66]. Since the esophageal
temperature during those trials increased at the rate of 0.2C/min, the sweat-
ing sensitivity for the chest was roughly 0.1 mg/(min cm2C). Sweating began
almost simultaneously on the abdomen, chest, and back. Sweating on the fore-
arm began about 2.25 minutes later, which corresponds to a 0.45C higher Tc,th.

mg/(cm2 min)
0.8 8
Rate of Regional Skin
Sweating
0.6 Abdomen Back 6
Chest Thigh
0.4 4
0.2 2
Forearm Up. arm
0 1 2 3 4 5 6 7 8 9 10
Time Min
Figure 8.8 Change in sweating over different skin areas at the start of heavy exercise in 30C room
air. The solid line is an average of five observations and the other five curves are single observations plot-
ted in time and magnitude relative to the chest curve. (Saltin et al. [66], Figure 7. With permission.)
Central threshold temperatures for sweating on the thigh and upper arm were
0.6C and 0.8C higher, respectively, than Tc,th for the chest. Moreover, as Nadel
et al. [67] noted, there appears to be a tendency for lower sweating sensitivity on
areas that have higher Tc,th.
8.4.4.2 An Algorithm for Computing the Sweat Rate

The discussion in this section focuses on the dependence of sweating on thermal
factors, acclimation to heat, fitness, and exercise. Two papers by Nadel and his
colleagues [67,68] defined the principal characteristics of sweating. In the fol-
lowing discussion of those papers, central temperature refers to the esophageal
temperature, and sweat rate is defined as grams of sweat produced per square
meter of skin per second. Since the latent heat of vaporization of water at 32C is
2425 J/gm, evaporation at the rate of 0.1 gm/(m2 s) removes heat at the rate of
242.5 W/m2. For a person whose body surface area is 1.8 m2, an evaporation rate
of 0.1 gm/(m2 s) corresponds to a whole-body evaporation rate of 0.65 kgm/hr.
The first paper by Nadel et al. [68] described experiments in which they con-
tinuously measured the weight of supine, seminude subjects supported on 5 cm
fish netting suspended from an aluminum frame on a sensitive beam balance. A
subjects anterior skin temperature was rapidly varied by controlling the radiant
heat flux from lamps located overhead. Esophageal temperature was increased
by imposing brief periods of exercise between observation intervals. This study
yielded a canonical relationship between sweat rate and relevant body tempera-
tures. That relationship has the following form:
dT T - Ts ,o
SR = a (Tc - Tc,o ) + b (Ts - Ts,o ) + g s + ro exp s (8.16)
dt d


280 E. H. Wissler
in which the term that involves dTdt

s is included only if dTs < r C/min. A specific
dt o
example for whole-body sweating (derived by Nadel et al. from mean data for six
resting subjects with Tty < 37.5C) is
0.08 (Tc - Tc ,o ) + 0.01 (Ts - Ts ,o )

T - 34.0 gm
SR = exp s (8.17)
dTs 10.0 m2 s
+ 0.05 + 0.2
dt

The second paper by Nadel et al. [67] refined and extended concepts presented
in the first paper. Three relatively fit male subjects performed 10-minute bouts of
cycling exercise at 80% of VO2 ,max . Ambient conditions were an air temperature of
26C and a relative humidity of 40%. A sensitive beam balance was used to mea-
sure whole-body weight loss. Data from two sweat collection capsules placed on
various skin areas established that sweat secretion was initiated on different areas
at different esophageal temperatures. In general, early sweating was associated
with high local sensitivity (i.e., with a larger slope of the Tes SR relationship).
Nadel et al. interpreted Equation (8.17) as the product of a central drive
that increases linearly with central and mean skin temperatures, and a periph-
eral modifying factor that increases exponentially with local skin temperature.
They also concluded that the increase in central drive per degree increase of Tes
is approximately 10 times the increase per degree of Ts. In addition, they con-
cluded that the natural logarithm of the peripheral factor increases 0.1 per degree
increase of local skin temperature.
Nadel et al. [67] correlated their data for the steady-state rate of sweating on
the chest with the following equation:
T - 34 gm
SR = 0.21[ (Tes - 36.6) + 0.02 (Ts - 34 )]exp s - 0.033 2 (8.18)
10 m s

They hypothesized that the last term [0.033 gm/(m2 s)] represents the mini-
mum rate of glandular secretion for which sweat is expelled from sweat ducts;
in other words, they assumed that sweat secreted from a gland at less than
0.033 gm/(m2 s)] is reabsorbed through the surface of the duct and does not reach
the skin surface.
Other research studies clarified certain aspects of sweating without chang-
ing appreciably basic tenets presented above. Wurster and McCook [69], Saltin
et al. [66], and Wyss et al. [43] also observed that sweating is strongly attenuated
by falling mean skin temperature. Wyss et al. [43] measured sweat rate under a
capsule on the forearm during direct heating and cooling of four male subjects.
Skin temperature was rapidly varied by controlling the temperature of water in a

tube suit worn by the subjects, as shown in Figures8.1 and 8.2 of their article [43].
Data for their subjects were well represented by the following equation:
dTs
SR = a (Tes - 36.5) + b (Ts - 33.0) + d +g (8.19)
dt
in which = 0.12 (0.002) gm/(m2sC), = 0.013 (0.0005) g/(m2sC), and =

0.11) g/(m2 s). The parameter, , varies depending on the sign of dT dt
s . When dTs
dt
is positive, = 0.026 (0.005); and when dTs
is negative, = 0.015 (0.002). Values
dt
within parentheses are standard deviations. It is not surprising that parameters
derived from this study are different from those reported by Nadel et al. [67],
because sweat rate on the forearm was measured in this study while whole-body
sweat rate was measured by Nadel et al., and it is well known that sweat secretion
is not uniform over the skin surface.
We will adopt Equation (8.16) as the basic relationship defining the local rate
of sweating in terms of Tc, Ts, dTs , and Ts. Application of Equation (8.16) requires
dt
specification of seven parameters, which calls for some rather arbitrary decisions.
First, we assume that (0.05 to 0.1) , which is consistent with results from sev-
eral studies [66,67]. Accepting the values reported by Nadel et al. [68] for , , and
ro yields the following relationship between sweat rate and thermal factors:
dT T - Ts ,o
SR = a[ Tc - Tc ,o + 0.1 (Ts - Ts ,o )] + 0.05 s + 0.2 exp s (8.20)
dt 10

If we also assume that Ts,o = Ts ,o, Equation (8.20) then contains three arbitrary
parameters: the sweating sensitivity, , and two reference temperatures, Tc ,o and
Ts ,o. The reference temperature Ts ,o can be specified arbitrarily, because it appears
in combination with and Tc ,o, at least during heating, which is our primary
concern. Setting Ts ,o = 34.0C leaves two parameters to be specified, and Tc,o.
Results from several experimental studies are summarized in Table8.2.
Unfortunately, the wide range of values in Table8.2 makes assigning appro-
priate values to and Tc,o rather difficult. For young resting individuals who are
neither especially fit nor heat acclimated, we will use Equation (8.11) from the
paper by Saltin et al. [66], expressed as follows:
0.046 (Tes - 36.4 ) + 0.00285 (Ts - 34.0)

Ts - 34.0 gm
SR = exp 10 (8.21)
dTs m2 s
+ 0.02
dt
The parameters shown in Equation (8.21) might appear to be odd choices because
= 0.046 gm/(m2sC) is certainly on the low end of the values in Table 8.2.

282 E. H. Wissler
Table8.2 Threshold Central Temperature for Sweating and Sweat Sensitivity

Determined in Various Studies
Sensitivity
First Author Activity Skin Area Tc,th (C) (gm/(m2sC))
Nadel et al. [68] Rest Thigh 37.2 0.06
Mack et al. [70] Rest Chest 36.8 0.30
Kondo et al. [71] Rest Forearm 36.8 0.19
Kondo et al. [71] Exercise Forearm 36.8 0.12
Nadel et al. [68] Exercise Chest 36.6 0.21
Nadel et al. [72] Exercise Chest 37.4 0.21
Roberts et al. [73] Exercise Chest 37.5 0.178
Pandolf et al. [74] Exercise Upper arm 37.0 0.083
Cotter et al. [75] Exercise Mean of eight sites 36.9 0.58
Patterson et al. [76] Exercise Mean of five sites 37.0 0.7
Nadel et al. [68] Rest Whole body 36.7 0.08
Hardy [77] Rest Whole body 36.6 0.07
Stolwijk [78] Rest Whole body 37.0 0.038
Saltin et al. [66] Exercise Whole body 36.4 0.046
However, those values were chosen because they are based on whole-body data
collected for both resting and exercising subjects over a considerable period of
time at the Pierce Foundation Laboratory. By comparison, data published in sev-
eral papers by Nadel and colleagues indicate that local sweat rates measured with
aspirated capsules tend to be higher than whole-body rates measured under simi-
lar conditions [66,67].
8.4.4.3 Effect of Acclimation to Heat and Fitness on Sweating

Since the purpose of sweating is to facilitate heat removal from skin when con-
vective and radiative heat transfer are inadequate, the rate of sweat secretion is
determined primarily by thermal factors. However, that does not preclude the
possibility that nonthermal factors also influence sweating. We will consider
four possible nonthermal factors: acclimation to heat, fitness, exercise, and skin
wettedness.
Various studies have shown that enhanced sweating is the principal effect
of acclimation to heat, which can be realized either by reducing the threshold
central temperature for sweating or by increasing the slope of the Tes SR rela-
tionship. In their third paper, Nadel et al. [72] investigated the separate effects
of physical training and heat acclimation on sweating. Six relatively unfit men
underwent physical training by cycling 1 hour per day for 10 consecutive days
in a cool environment (Ta = 22C). Physical training was followed by a similar
period of acclimation to heat effected by exercise in a warm, dry environment.
The sweat sensitivity of each subject was determined by measuring the rate of

Table8.3 Effect of Training and Heat Acclimation on Sweating Parameters

Pretraining Posttraining Postacclimation
Slope: gm/(m2sC) 0.16 0.22 0.24
Threshold temperature: C 37.5 37.4 37.3
Source: Roberts et al. [73].
sweat secretion on the chest as a function of esophageal and local skin tempera-
tures. The authors concluded that physical training and heat acclimation both
enhance the rate of sweat secretion at given values of esophageal and mean skin
temperature, although they do so by entirely different mechanisms. Physical
training increases the slope of the Tes SR relationship; that is, it increases the
value of in Equation (8.16). Heat acclimation, on the other hand, lowers the
threshold central temperature for initiating the central drive for sweating. These
studies were conducted with each subject serving as his own control, and while a
majority of the subjects responded as described above, some exhibited little
response, or even responded negatively.
Results from a subsequent study by Roberts et al. [73] were consistent with the
conclusions of Nadel et al. [72]. Roberts et al. also employed a two-step procedure
(10 days of exercise training followed by 10 days of heat acclimation) to assess
the effect of training and acclimation on chest sweating. Mean values of the slope
and threshold temperature are shown in Table8.3. Although those values suggest
that the slope increases with training and heat acclimation, differences between
the three conditions were not statistically significant.
Pandolf et al. [74] studied the effect of 10 days of heat acclimation on the rela-
tionship between rectal temperature and whole-body sweat rate for two groups of
subjects, one young and the other middle-aged. Their results are summarized in
Table8.4. The difference in slopes and threshold temperatures between this study
and the study of Roberts et al. [73] can probably be attributed to the fact that
Roberts et al. measured local sweating on the chest, while Pandolf et al. measured
whole-body sweating. Otherwise, results from the two studies are consistent in
that they show a small increase in slope and a 0.3C decrease in threshold tem-
perature with acclimation.
Table8.4 Effect of Heat Acclimation on Sweating Parameters

Slope: gm/(m2sC) Threshold Temperature: C
Day 1 Day 10 Day 1 Day 10
Young 0.083 0.098 36.99 36.67
Middle-aged 0.078 0.105 36.79 36.64
Source: Pandolf et al. [74].

284 E. H. Wissler
Similar results were reported by Armstrong and Kenney [79], who studied
the effect of age and heat acclimation on the mean body temperature (defined as
0.8 Tre + 0.2 Tsk )sweating relationship during 90 minutes of passive heating. In
both groups, 9 days of acclimation lowered the threshold mean body temperature
about 0.4C.
8.4.4.4 Effect of Exercise on Sweating

The effect of exercise on sweating is mildly controversial. Van Beaumont and
Bullard [80] were the first to investigate carefully whether exercise directly affects
sweating. They used resistance hygrometry to measure the rate of sweating under
7 cm2 capsules located on the forearm and calf. Subjects exercised at 160 W on a
cycle ergometer after an initial, 1-hour period of rest in the test chamber. When
the chamber temperature was 37.5C, light sweating occurred before exercise
began. Under those conditions, definite increases in sweat rates on both the rest-
ing forearm and the working leg were observed within 2 seconds of the onset of
exercise, before any change in central or skin temperature occurred. When the
chamber temperature was 30C, preexercise sweating was absent and sweating
began after 30 to 80 seconds of exercise.
In a subsequent investigation, Gisolfi and Robinson [81] also observed rapid
changes in the rate of sweat secretion during intermittent dynamic exercise when
subjects had a warm core (Tc above 38C), a warm skin (Tsk above 35.5C), or
both. Robinson [82] and Robinson et al. [83] reported that, for a given thermal
state defined by central temperature, mean skin temperature, and body heat load,
the rate of sweating is higher during exercise than during rest.
Similar results were reported by Yanagimoto et al. [84], whose subjects cycled
for 80 seconds at 30, 50, and 70% ofVO2 ,max in a chamber where the tempera-
ture was 35C and the relative humidity was 50%. Those conditions caused light
sweating during the 1-hour preexercise period. Increased sweating on the chest,
thigh, and forearm was observed to begin during the first 2 to 16 seconds of exer-
cise. The latent period decreased and the rate of sweating increased with increas-
ing intensity of exercise.
Yamazaki et al. [85] measured esophageal and mean skin temperatures and
forearm sweat rates as six male subjects performed cycling exercise that varied
sinusoidally in intensity from 10% to 60% of VO2 ,max . Four periods (1, 3, 4, and
8 seconds) were employed. Since sweat rate varied sinusoidally during the 1- and
3-second oscillations even though there was little variation in esophageal or
mean skin temperature, Yamazaki et al. [85] concluded that exercise modulates
sweat rate.
Tam et al. [86] studied two subjects, one normal and the other paraplegic,
during passive heating and during exercise. During exercise in a warm envi-
ronment, a sharp increase in whole-body sweat rate (evaluated as a weighted
mean of four local sweat rates) [65] was observed before there was a measurable
increase in aural or mean skin temperature. The initial sharp increase in sweat

rate was followed by a slower increase with increasing central temperature. The
slope during the second phase was similar to the slope for passive heating, but an
apparent threshold central temperature obtained by extrapolating the tympanic
temperatureSR curve to zero sweat rate was about 1C lower than the threshold
tympanic temperature for rest.
Although a considerable body of evidence exists indicating that sweating is
enhanced during dynamic exercise, results from other studies fail to support that
concept. For example, Johnson and Park [46] observed that Tes forearm sweating
relationships for supine rest and upright exercise were virtually identical. In both
cases, subjects wore a liquid-perfused suit that maintained a mean skin tempera-
ture of 38.0 to 38.5C. One difference between those experiments and the experi-
ments of Van Beaumont and Bullard [80] and Gisolfi and Robinson [81] is that
light sweating was present in the latter two studies, which exhibited an enhance-
ment of sweating when work commenced, while that apparently was not true of
the study by Johnson and Park.
Saltin et al. [66] performed a series of experiments with one subject to define
more clearly how exercise affects the onset of sweating. Their subject cycled at
90% of VO2 ,max in a chamber where the temperature was 30C and the humidity
was low. No sweating was observed during the first minute of exercise, but by the
end of the second minute, sweating occurred on at least 50% of the skin.
8.4.4.5 Hidromeiosis
Another nonthermal effect is hidromeiosis, which defines the progressive sup-
pression of sweating when the skin is wet. An early description of hidromeiosis
was published by Taylor and Buettner [87], who presented previously unpublished
data collected in 1932, 1942, and 1946. Although their analysis is somewhat lim-
ited by the assumption that the thermal drive for sweating is proportional to the
mean skin temperature, their data indicated clearly that sweat secretion depends
on environmental factors, such as humidity, wind speed, and air pressure, which
they called the environmental effect. They concluded that any environmental
factor that facilitates evaporation of sweat promotes sweat secretion at a given
mean skin temperature. The presence of water on the skin, whether caused by
intense sweating or by immersion, inhibits sweat secretion [88,89].
Brown and Sargent [90] also conducted an extensive study of hidromeiosis,
from which emerged the following six characteristics:
1. In an environment of constant high ambient temperature, a person

exhibits initially an outburst of sweating that reaches a peak in 1 to
2 hours, after which the rate progressively declines.
2. The rate of decline is steeper in moist heat than in dry heat.
3. The degree of sweat depression appears to be related to the magnitude of
the initial maximum rate, suggesting that a threshold rate of sweating
must be reached to initiate hidromeiosis.

286 E. H. Wissler
4. Hidromeiosis does not appear to be an adaptive process, since frequently

the rectal temperature rises in the face of the declining sweat rate.
5. Dehydration may accelerate hidromeiosis.
6. The duration of exposure to thermal stress, rather than the intensity of
work, appears to be a factor in the occurrence of hidromeiosis.
Brown and Sargent [90] concluded that there is a threshold rate of sweating
below which hidromeiosis does not occur, and that rate is sufficient to maintain
a fully wetted skin. Since hidromeiosis usually becomes apparent only after 1 or
2 hours of intense sweating, it is not normally a factor in experiments of shorter
duration.
Two very interesting early studies conducted by Gerking and Robinson [91,92]
were consistent with the first four points of Brown and Sargent [90]. Male subjects
in the Gerking-Robinson studies walked for 6 hours with a 5-minute break for
weighing at the end of each hour. The predominant metabolic rate was 220 W/m2,
although that was reduced to 150 W/m2 for several subjects. Experiments were
carried out under two different ambient conditions: humid, and hot and dry.
Humid conditions involved an air temperature of 31.9C to 38C and rela-
tive humidity of 95% to 51%. In the hot and dry conditions, air temperature
ranged from 40.0C to 50.1C with a relative humidity of 38% to 18%. Subjects
either wore shorts, shoes, and socks, or a poplin tropical uniform, shoes, and
socks.
In 50 experiments, the average rate of sweating during the first 2 hours was
1400 gm/hr. The rate of sweat secretion during the sixth hour varied from to 10%
to 80% of the initial rate, depending on environmental conditions, with high
humidity resulting in a pronounced decline in the rate of sweating. Subjects were
able to maintain a steady sweating rate of 780 gm/hr in moderate conditions.
The Robinson-Gerking studies [91,92] are unique in that they evaluated the
effect of clothing on sweat secretion. As expected, clothing inhibited evapora-
tion of sweat, and excess sweat accumulated on the skin and in the clothing.
Consequently, the rate of sweat secretion by clothed subjects decreased continu-
ously after the second hour, and by the fourth hour, the rate of evaporation was no
longer sufficient to prevent skin and rectal temperatures from increasing. Sweat
deficiency during the last 3 hours of work was partially alleviated by evaporation
of sweat that had accumulated in the cotton uniform during the first 3 hours (as
much as 800 gm in one case).
Nadel and Stolwijk [93] studied hidromeiosis as one part of the series of inves-
tigations discussed above. In the hidromeiosis study they reduced the amount of
sweat accumulated on skin either by wiping the skin with a towel, or by increas-
ing the wind speed to facilitate evaporation. In both cases, they observed that
drying the skin promptly increased the rate of sweat secretion. Qualitatively,
their observations were similar to those of Brown and Sargent [90].

Candas et al. [94] investigated the effect on hidromeiosis of acclimation to

humid heat. Their study took place over 10 consecutive days during which eight
subjects clad only in shorts reclined for 2 hours and 45 minutes on a web bed
in a humid environment. The initial rate of sweat secretion for seven subjects
increased linearly with days of acclimation; the eighth subject showed no effect of
acclimation. Candas et al. found that the rate of decline of sweat secretion owing
to hidromeiosis was proportional to the rate of sweat secretion during the prehi-
dromeiosis period. Their data are well represented by the following relationship:
SR = SRo [1 - 0.009 (t - to )] (8.22)
where SRo is the maximum rate of sweat secretion before the onset of hidromeio-
sis at to. Time is measured in minutes. The degree of sweat secretion attenuation
observed by Candas et al. is several times larger than was observed by Robinson
and Gerking [92]. Based on their observations of sweat secretion (deduced from
the rate of sweat drippage), Candas et al. concluded that hidromeiosis does not
occur on a given region until it is fully wet.
From the limited point of view of thermoregulation of nude subjects, hidromei-
osis is unimportant, because it does not occur until an area is fully wet, and then
the rate of evaporation is determined by environmental conditions. Nevertheless,
hidromeiosis cannot be completely neglected for two reasons. One is that the rate
of sweat secretion affects ones level of hydration when adequate water replace-
ment does not occur, and appreciable dehydration has thermoregulatory con-
sequences. Another reason is that excess sweat accumulates in garments and
changes their physical properties. For example, as noted above, Robinson and
Gerking noted that one subject accumulated 824 grams of sweat in his poplin
uniform during the first 3 hours of walking at a metabolic rate of 220 W/m2
under hot and dry conditions. During the last 3 hours of the 6-hour experiment,
his sweat rate was below the rate of evaporation from his clothing, and one-third
of the previously accumulated sweat evaporated.
8.4.4.6 Comparison of Computed and Measured Whole-Body Sweat Rates

Values computed using Equation (8.21) with suitable modification to account
for the effect of training, heat acclimation, and exercise are compared with 42
experimentally observed values in Figure8.9. Tes,o was assigned a value of 36.7C
for resting nonacclimated individuals. We assumed that acclimation plus train-
ing reduce Tes,o by 0.3C, and exercise reduces Tes,o by 0.9C; the two effects are
additive. Represented are three combinations of acclimation and exercise: unac-
climated rest, unacclimated exercise, and acclimated exercise. Although agree-
ment between computed and measured values might be improved by modifying
the factors given above, which are supported by experimental data, those factors
provide acceptable agreement between computed and measured values.

288 E. H. Wissler
0.28
0.24
Computed SR: gm/(sq m sec)
0.20
0.16
0.12
0.08
0.04
0
0 0.04 0.08 0.12 0.16 0.20 0.24 0.28
Measured SR: gm/(sq m sec)
Figure 8.9 Comparison of sweat rates computed using Equation (8.21) with corresponding measured
values. Filled circles denote acclimated exercising subjects, open circles denote unacclimated exercising
subjects, and open triangles denote unacclimated resting subjects.
8.4.5 Shivering
During exposure to extreme cold, metabolic heat generation owing to shivering
can approach five times the resting metabolic rate [95]. While shivering helps to
prevent hypothermia under most conditions, the net benefit may be small during
immersion in cold water, because peripheral blood flow also increases. Shivering
responds to signals generated both centrally and peripherally.
8.4.5.1 An Algorithm for Computing the Rate of Shivering Metabolism

Our basic representation of shivering is derived from the study of Hayward
et al. [96], who measured the oxygen consumption rate together with tympanic
and rectal temperatures in two groups of subjects during immersion in 10C
water. One group had a mean skinfold thickness of 7.8 mm, while the mean
skinfold thickness of the other group was 10.4 mm. That study is particularly
interesting because it demonstrates that shivering, unlike skin blood flow and
sweating, exhibits a rather clear dependence on the rate of change of mean skin
temperature.
Two other studies suggest that shivering is also stimulated by rapidly falling
central temperature. Piantadosi et al. [97] attributed enhanced shivering to rapid
central cooling (defined in terms of changing rectal temperature) in subjects who
breathed hyperbaric heliox at pressures corresponding to depths of 1400 and
1800 ftsw. In the other study, Nadel et al. [98] induced a rapid 0.6C decrease in
tympanic temperature by having subjects ingest 500 gm of ice cream during a
10-minute interval. They attributed the ensuing sharp, brief interval of enhanced

shivering to the transient decrease in central temperature. However, we will

not include this factor in the shivering algorithm because there are not enough
experimental data to define the effect, and it is unlikely to be important in most
situations to which thermal models are applied.
Consequently, we assume that the rate of shivering can be represented as the
sum of two parts:
M sh = M sh ,1 + M sh ,2 (8.23)

M sh,1 , which accounts for the effect of rapidly decreasing mean skin temperature,
is defined as follows:
dM sh ,1 dT (8.24)
= F1 s - b1 M sh ,1
dt dt

where F1 = 0 if 1.5C/min < dT s ; F = A ( - dTs - 1.5) if 3.5C/min < dTs < 1.5C;
dt 1 1 dt dt
and F1 = 2.0 A1 if dT
dt
s < 3.5C/min. According to this model, slowly decreasing T
s
does not stimulate shivering, and a maximum stimulus occurs when Ts decreases
more rapidly than 3.5C/min. Values of the two parameters are A1 = 120 W/C
and 1 = 0.5 m1.
Under most conditions, the largest contribution to shivering is provided by
M sh,2, which is defined as follows:
dM sh ,2
= b 2 [F2 (Tc , Ts ) - M sh ,2 ] (8.25)
dt
in which F2 (Tc, Ts) is the rate of shivering normally observed under steady-state
conditions. The value of 2 is 0.177 m1. Specific relationships for F2 can be derived
from papers by Hayward et al. [96] and Tikuisis and Giesbrecht [99]. We use the
function defined by Tikuisis and Giesbrecht with slightly modified parameters.
155.5 (36.6 - Tc ) - [1.57 (28.0 - Ts ) - 47.0](30.0 - Ts )

F2 = (8.26)
BF
where BF is the percentage of body fat.

Benzinger [100] concluded from his extensive early study of shivering that
shivering metabolism decreases when the mean skin temperature is below 20C.
Although several of his hypotheses were refuted by subsequent investigators, a
recent study by Tikuisis and Giesbrecht [99] seems to support his hypothesis con-
cerning the effect of skin temperature on shivering. Accordingly, when Ts < 20C,
we reduce the shivering rate computed according to Equation (8.26) by adding
the multiplicative factor,1.0 - 0.004 (20.0 - Ts )2.

290 E. H. Wissler
8.5 MODEL PERFORMANCE

As we mentioned earlier, an important objective of this chapter is to investigate
whether physiological control functions determined in their own right allow an
acceptable description of human thermal regulation when incorporated into a
human thermal model. Before we consider specific cases, it is worthwhile to dis-
cuss several factors relevant to comparison of computed and measured values.
We start with the following quotation attributed to W. I. Beveridge (1908), which
appears in the last paragraph of Rowells book on human circulation: No one
believes an hypothesis except the originator, but everyone believes an experiment
except the experimentor [31].
Probably the greatest frustration encountered by a physical scientist engaged in
human thermal modeling is the biological variability of individuals. Physiological
experiments usually involve a small number of subjects whose physical charac-
teristics may vary considerably. For example, the physical characteristics of eight
subjects in the cold-immersion experiments performed by Hayward et al. [96] are
summarized in Table8.5.
The mean rectal temperature of the first three less fat subjects decreased to
34.0C during 45 minutes of immersion in 10C water, while the mean rec-
tal temperature of the other five fatter subjects remained above 34.5C during
60 minutes of immersion at the same temperature. Clearly, subcutaneous fat
thickness affects an individuals response to severe cold, but can we assume that
the response of an individual with mean physical characteristics is the same
as the mean response of a group with diverse characteristics?
In general, models easily account for two variable physical factors, body mass
and subcutaneous fat thickness, although, as we saw earlier, fat thickness deter-
mined using calipers may be quite inaccurate. Degree of acclimation to heat and
Table8.5 Physical Characteristics of Subjects in the

Cold-Immersion Study of Hayward et al.
Age Weight Height Skinfold
Subject (years) (kgm) (cm) (mm)a
1 21 71.4 188.4 7.4
2 21 75.9 179.1 7.8
3 22 80.9 181.3 8.2
4 21 64.5 174.2 9.0
5 20 92.7 193.0 9.8
6 20 76.4 181.6 10.3
7 22 74.1 178.5 11.0
8 26 76.8 185.4 12.0
a Mean of thicknesses at four sites.
Source: Hayward et al. [96].

38
36
Temperature: C
34
32
30
28
0 20 40 60 80 100 120 140 160 180 200 220 240
Time: Min
Figure 8.10 Measured and computed temperatures for seminude subjects resting for 60 minutes at
18C, 120 minutes at 42C, and 60 minutes at 18C. Plotted are the measured tympanic (filled circles),
rectal (filled triangles), and mean skin temperatures, and the computed esophageal (open circles) and
mean skin temperatures (open triangles).
physical fitness can also be important factors under particular conditions, but
they are not usually reported. In addition, factors such as the threshold tempera-
tures for active vasodilation and the onset of sweating vary considerably between
individuals and from day to day for a given individual.
Another important consideration is lack of precision in measured values.
For example, the central temperature for thermoregulatory control is usually
approximated by the rectal, esophageal, or tympanic temperature, none of which
provides a consistently accurate approximation for the temperature of the inac-
cessible hypothalamus. Measured central temperatures may differ from one to
another by as much as 1C during transient periods. The disparity is illustrated
in Figure8.10 and discussed more fully following Figure8.11.
Mean skin temperature typically determined as the weighted mean of six to
ten skin temperatures is only an approximation for the true mean skin tempera-
ture. In addition to the problem of definition, we note that simply taping a ther-
mocouple on the skin alters the temperature of the skin. Moreover, there is very
little experimental evidence pertaining to the integrated cutaneous input when
conditions vary considerably over the surface of the body, for example during
partial immersion in cold water.
In summary, differences between computed and measured values should not
automatically be attributed to deficiencies in the model; valid reasons for the dif-
ferences may be attributable to the manner in which the experiment was per-
formed, or the results are interpreted.

292 E. H. Wissler
120
100
Rate of Evaporation: W/sq m
80
60
40
20
0
0 20 40 60 80 100 120 140 160 180 200 220 240
Time: Min
Figure 8.11 Measured (filled circles) and computed (open circles) rates of evaporative cooling for the
conditions shown in Figure 8.10.
We will compare computed and measured values for four cases. The first case
involves extended exposure of resting subjects to a hot environment, the sec-
ond case involves exercise at three intensities in cool and warm environments,
the third case involves passive exposure to cold air, and the fourth case involves
immersion in cold water. In each case, subjects were seminude, which eliminates
uncertainty associated with modeling clothing and provides a clearer evaluation
of the physiological model. These particular cases were chosen to test the model
because we have experimentally measured values for one or two central tempera-
tures, a mean skin temperature, the metabolic rate, and the rate of evaporative
cooling (when sweating is an important factor).
8.5.1 Passive Exposure to Heat

The first case is one of many reported by Hardy and Stolwijk [77,78]. Their
experiments carried out in the partitional calorimeter of the Pierce Foundation
Laboratory at Yale University were fairly unique in that the rate of sweat evapo-
ration was measured continuously using a sensitive beam balance to record sub-
ject weight. Seminude subjects sat quietly in a chair during the experiment, with
minimal air movement in the chamber. The rate of oxygen uptake, rectal temper-
ature, tympanic temperature, and mean skin temperature were also measured.
Plotted in Figures8.10 and 8.11 are results for a 4-hour experiment in which
three subjects spent the first hour in a chamber where Tair = 18C, the next
2 hours in a chamber where Tair = 42C, and the fourth hour back in the first

chamber. Computed values of the esophageal temperature, mean skin tempera-

ture, and rate of evaporative cooling are plotted.
Several points are worth mentioning about the results of this case. First, a sig-
nificant difference exists between measured rectal and tympanic temperatures.
The difference is largest during relatively steady conditions at the beginning
and end of the experiment, when the rectal temperature is approximately 0.5C
higher than the tympanic temperature. That difference is fairly typical of steady-
state rectal and tympanic temperatures. During the initial heating period, the
rectal temperature decreased for 20 minutes (presumably owing to an increase
in peripheral circulation), while the tympanic temperature increased. Because
the tympanic temperature increased more rapidly than the rectal temperature
immediately after subjects moved into the warm room, the difference between
the two temperatures was negligible during much of the heating period. It should
be noted that part of the increase in tympanic temperature was probably due
to the influence of ambient air temperature on tympanic temperature. Several
studies have established that tympanic temperature is affected by ambient tem-
perature, although the cause is not clear. One likely cause is that venous drainage
from the face affects temperature in the ear through countercurrent exchange of
heat with blood in the carotid artery. Since the actual change in arterial blood
temperature was probably intermediate between changes in the rectal and tym-
panic temperatures, the computed change in esophageal temperature may be
slightly too large.
Agreement between measured and computed mean skin temperatures is rea-
sonable given the problems associated with such measurements. The reason for
the sizeable temperature disparity during the third phase of the experiment is
unknown. Agreement between measured and computed rates of evaporative
cooling is excellent.
8.5.2 Exercise at Three Intensities and Two Temperatures

Experiments simulated in the second case were also carried out in the Pierce
Foundation partitional calorimeter [101]. This case differs from the previous case
in that this study involved a seminude subject who performed cycling exercise at
three intensities (25%, 50%, and 75% ofVO2 ,max ). Measured results for this case
are particularly interesting because they are for one subject, instead of being
mean values for five or six subjects. Hence, we are simulating the behavior of a
subject for whom all of the physical parameters are known.
Rectal, active muscle, and mean skin temperatures were measured together
with the metabolic and evaporative cooling rates. Measured and computed
results are plotted in Figures8.12 and 8.13 for 10C ambient conditions, and in
Figures8.14 and 8.15 for 30C ambient conditions. Results for 20C were consis-
tent with those two conditions.

294 E. H. Wissler
40
38
36
34
Temperature: C
32
30
28
26
24
0 20 40 60 80 100 120 140 160 180
Time: Min
Figure 8.12 Measured and computed temperatures for a seminude subject during three periods of
cycling exercise in 10C air. Plotted are the measured rectal (filled circles), active muscle (filled squares),
and mean skin (filled triangles) temperatures, and computed esophageal (open circles), muscle (open
squares), and mean skin (open triangles) temperatures.
600
500
400
Rate: W/sq m
300
200
100
0
0 20 40 60 80 100 120 140 160 180
Time: Min
Figure 8.13 Measured and computed metabolic (circles) and evaporative cooling (triangles) rates
during cycling exercise in 10C air. Filled symbols denote measured values and open symbols denote
computed values.

41
40
39
38
Temperature: C
37
36
35
34
33
32
31
0 50 100 150 200 250
Time: Min
Figure 8.14 Measured and computed temperatures for seminude subjects during three periods of
cycling exercise in 30C air. Symbols have the same meaning as in Figure 8.12.
Agreement between measured and computed values for this rather demand-
ing case is reasonable. The greatest discrepancies are between measured and
computed mean skin temperatures in the 30C chamber, and between measured
and computed active muscle temperatures in both chambers. The discrepancy
between computed and measured mean skin temperatures could be caused by
600
500
400
Rate: W/sq m
300
200
100
0
0 50 100 150 200 250
Time: Min
Figure 8.15 Measured and computed metabolic and evaporative cooling rates during cycling exercise
in 30C air. Symbols have the same meaning as in Figure 8.13.

296 E. H. Wissler
several factors, including uncertainty in the heat and mass transfer coefficients,
improper distribution of sweat secretion, and errors in the computed local blood
flow rates. The smaller than expected increase in computed muscle temperature
might be attributable to our assumption that the entire mass of leg muscle par-
ticipates equally in the work of cycling, which is probably not true, especially at
lower intensities of exercise.
8.5.3 Passive Exposure to Cold Air

The last two cases involve exposure to conditions sufficiently cold to invoke shiv-
ering. Experimental data for the third case were obtained from four studies [102
105] performed at the Centre de Reserches du Service de Sante des Armees in
Lyon, France. In those experiments, recumbent, seminude subjects were exposed
for 2 hours to 1C, 5C, and 10C air. Computed results are compared with mea-
sured values in Figures8.16, 8.17, and 8.18.
Although there was some disparity between results from different studies, two
rather interesting conclusions emerged from the observations. One is that shiv-
ering and strong vasoconstriction are able to prevent a large decrease in central
temperature during several hours of cold-air exposure. The second, somewhat
surprising conclusion is that a smaller decrease in central temperature occurs
during exposure to 1C air than during exposure to 10C air, as is illustrated by
the temperatures plotted in Figure8.16. While that behavior can be attributed in
part to the higher metabolic rate observed during exposure to 1C air, as shown
38
37.5
Central Temperature: C
37
36.5
36
35.5
35
0 20 40 60 80 100 120
Time: Min
Figure 8.16 Measured rectal temperature [103] (filled symbols) and computed esophageal tempera-
ture (open symbols) during exposure of seminude subjects to cold air. Ambient temperatures were 1C
(squares), 5C (triangles), and 10C (circles).

36
Mean Skin Temperature: C 34
32
30
28
26
24
22
20
0 20 40 60 80 100 120
Time: Min
Figure 8.17 Measured [103] and computed mean skin temperature for seminude subjects during
exposure to cold air. Symbols have the same significance as in Figure 8.16.
200
180
160
140
Metabolic Rate: W/sq m
120
100
80
60
40
20
0
0 20 40 60 80 100 120
Time: Min
Figure 8.18 Measured [103] and computed mean metabolic rate for seminude subjects during expo-
sure to cold air. Symbols have the same significance as in Figure 8.16. Measured values on the dashed
lines are mean values for nine subjects [102,105].

298 E. H. Wissler
in Figure 8.18, enhanced vasoconstriction induced by rapidly decreasing skin

temperature during exposure to 1C air also seems to be involved.
It is difficult to compare computed and measured mean skin temperatures
during the first few minutes of cooling because the measured values reported at
time = 0 were apparently recorded after the cold exposure began. Nevertheless,
it is encouraging to note that computed mean skin temperatures during the last
half of the experiment are close to measured values.
The sizeable discrepancy between computed and measured metabolic rates
might indicate that the shivering control function used in the model is inad-
equate, which is not totally unreasonable because it was derived from experi-
mental studies in which subjects were exposed to rather cold water, while these
experiments involved less severe exposure to cold air. The experimental data plot-
ted in Figure8.18 suggest that shivering began almost simultaneously with expo-
sure to cold air, but the data in Table8.2 of the paper from which those data were
taken [104] reported that shivering began after cold exposures of 6.1 minutes,
15.7 minutes, and 17.5 minutes for air temperatures of 1C, 5C, and 10C, respec-
tively. In addition, a graph of the mean metabolic rates for nine subjects during
exposure to 1C and 10C air, represented by the broken lines in Figure 8.18,
suggests that shivering develops more slowly in 10C air than the data plotted in
Figure8.18 indicate. Although determination of shivering metabolic rate appears
to be straightforward, that may not be completely true. Interested readers may
want to read a discussion by Buskirk et al. [106] of possible artifacts that can
affect such measurements during mild cold exposure.
8.5.4 Immersion in Cold Water

The fourth case provides a more severe test of the models ability to simulate
responses to cold exposure. In these experiments, eight young men whose physi-
cal characteristics are shown in Table 8.5 were immersed to the neck in 10C
water. One criterion for termination of the experiment was a rectal temperature
of 35C. The three subjects with the lowest mean skinfold thicknesses reached
that condition after 45 minutes of immersion, while five fatter subjects tolerated
60 minutes of immersion. Measured rectal temperatures and computed esopha-
geal temperatures for the two groups are plotted in Figure8.19. Measured and
computed metabolic rates are plotted in Figure8.20.
The computed change in esophageal temperature for the thin subjects (1.5C)
is significantly less than the measured change in tympanic temperature (2.8C).
On the other hand, the computed change in esophageal temperature for the fat-
ter subjects (1.8C) is close to the measured change in tympanic temperature
(1.9C). While the second result is encouraging, it may be somewhat fortuitous
because of differences that can occur between tympanic and esophageal tem-
peratures during periods of rapid change. It is disturbing that the model does
not appear to account properly for the effect of subcutaneous fat thickness
on response during immersion in cold water. Another important difference

37.5
37
36.5
36
Temperature: C
35.5
35
34.5
34
33.5
33
0 10 20 30 40 50 60 70 80 90
Time: Min
Figure 8.19 Measured tympanic temperature (filled symbols) and computed esophageal temperature
(open symbols) during immersion in 10C water. Circles denote heavy subjects (mean skinfold thickness =
10.4 mm) and triangles denote thin subjects (mean skinfold thickness = 7.8 mm).
between computed and measured central temperatures for both groups is that
the model does not account for the after-drop in central temperature, which is
often observed in cold-immersion experiments. Unfortunately, the physical and
physiological reasons for after-drop are unclear, and there is no obvious reason
for the difference between simulated and observed responses.
300
250
Metabolic Rate: W/sq m
200
150
100
50
0
0 10 20 30 40 50 60 70 80 90
Time: Min
Figure 8.20 Measured and computed metabolic rates during immersion in 10C water. Symbols have
the same significance as in Figure 8.19.

300 E. H. Wissler
A final comment about results for the third and fourth cases is in order.
Although a strong objective of this development has been to establish that the
physiological control functions for skin blood flow, sweating, and shivering
determined without regard to thermoregulation per se yield a model with accept-
able thermoregulatory characteristics, it was necessary for the two cold-exposure
cases to make some ad hoc adjustments. Although those adjustments are not sup-
ported by independent physiological experiments, neither do they violate such
observations.
We found that the model responded properly to cold-air exposure only after
it was modified by assuming that rapidly falling mean skin temperature causes
vasoconstriction in inactive muscle and venoconstriction in the arms and legs
(which increases arteriovenous countercurrent heat transfer). Unfortunately,
changes required to conserve bodily heat during exposure to cold air also pre-
vented cooling during immersion in cold water, and, therefore, it was necessary
to assume that those changes are reversed when the skin temperature falls below
20C. Hence, vasoconstriction in inactive muscle and venoconstriction were
stronger during exposure to 1C air than during exposure to either 10C air or
10C water. Although independent experimental evidence supporting the sec-
ond assumption was reported by Cannon and Keatinge [107], the validity of both
responses warrants further experimental investigation.
8.6 conclusions
To develop a useful human thermal model, one must be able to formulate physi-
cal concepts in mathematical terms and have a good grasp of scientific litera-
ture dealing with thermal and circulatory physiology. Unlike physical scientists,
physiologists do not express their findings in quantitative terms that are easily
incorporated into complex models. For example, there are only two or three early
papers that propose a quantitative relationship between skin blood flow and cen-
tral and mean skin temperatures, and those relationships have been ignored by
subsequent investigators. Consequently, there are no sources to which modelers
can turn for quantitative relationships between physiological variables. The prin-
cipal purpose of this chapter is to present relationships that were derived from an
extensive examination of the relevant literature.
The relationships presented in this chapter seem to work well when incorpo-
rated into a human thermal model. Although that model is not described, it is
similar to other models developed during the last 30 or 40 years. The important
thing is that the model provides reasonable agreement between computed and
measured thermal responses to conditions involving rest and exercise during
exposure to hot and cold ambient conditions.
Challenges remain for those who are fascinated by the functioning of the
human body, as I have been for 50 years. It seems to me that one of the biggest

challenges is to discover how the body responds to distributed stimuli. For exam-
ple, if a soldier walks immersed to the waist in cold water, do skin blood flow and
shivering respond to mean skin temperature computed for the entire body, or do
they respond preferentially to cold stimuli coming from wet skin? Answers to such
questions will allow the development of better models, and well-designed human
thermal models can play an important role in answering such questions.
REFERENCES
1. J. A. J. Stolwijk, A mathematical model of physiological temperature regulation in
man, NASA-9-9531, September 1970.
2. P. A. Gagge, J. A. J. Stolwijk, and Y. Nishi, An effective temperature scale based on a
simple model of human physiological regulatory response, ASHRAE Trans., vol. 70
(Part 1), pp. 247262, 1970.
3. K. K. Kraning and R. R. Gonzalez, A mechanistic computer simulation of human
work in heat that accounts for physical and physiological effects of clothing, aerobic
fitness, and progressive dehydration, J. Therm. Biol., vol. 22, pp. 331342, 1977.
4. J. A. J. Stolwijk and J. D. Hardy, Temperature regulation in man: a theoretical study,
Pfluegers Archiv, vol. 291, pp. 126162, 1966.
5. A. C. Burton, The application of the theory of heat flow to the study of energy metab-
olism, J. Nutrition, vol. 7, pp. 497533, 1934.
6. H. C. Bazett and B. McGlone, Temperature gradients in the tissues in man, Am. J.
Physiol., vol. 82, pp. 415451, 1927.
7. A. C. Burton and H. C. Bazett, A study of the average temperature of the tissues, of
the exchanges of heat and vasomotor responses in man by means of a bath calorim-
eter, Am. J. Physiol., vol. 117, pp. 3654, 1936.
8. W. Machle and T. F. Hatch, Heat: mans exchanges and physiological responses,
Physiol. Rev., vol. 27, pp. 200227, 1947.
9. H. H. Pennes, Analysis of tissue and arterial blood temperature in the resting human
forearm, J. Appl. Physiol., vol. 1, pp. 93122, 1948.
10. E. H. Wissler, Pennes 1948 paper revisited, J. Appl. Physiol., vol. 85, pp. 3541,
1998.
11. C. K. Charney, Mathematical models of bioheat transfer, in Advances in Heat Transfer,
vol. 22, pp. 19155, Y. I. Cho, J. P. Hartnett, and T. F. Irvine Jr., editors, New York:
Academic Press, 1992.
12. H. Brinck and J. Werner, Efficiency function: improvement of classic bioheat
approach, J. Appl. Physiol., vol. 77, pp. 16171622, 1994.
13. S. Weinbaum, L. X. Xu, L. Zhu, and A. Ekpene, A new fundamental bioheat equation
for muscle tissue, part I: blood perfusion term, ASME J. Biomech. Eng., vol. 119, pp.
278288, 1997.
14. L. Zhu, L. X. Xu, Q. He, and S. Weinbaum, A new fundamental bioheat equation for
muscle tissue: part II: temperature of SAV vessels, ASME J. Biomech. Eng., vol. 124,
pp. 121132, 2002.
15. E. H. Wissler, Steady-state temperature distribution in man, J. Appl. Physiol., vol. 16,
pp. 734740, 1961.
16. H. Zhang, C. Huizenga, Z. Hui, E. Arens, and T. Yu, Considering individual physi-
ological differences in a human thermal model, J. Therm. Biol., vol. 26, pp. 401408,
2001.

302 E. H. Wissler
17. O. Jay, F. D. Reardon, P. Webb, M. B. DuCharme, T. Ramsay, L. Nettlefold, and G. P.

Kenny, Estimating changes in mean body temperature for humans during exercise
using core and skin temperatures is inaccurate even with a correction factor, J. Appl.
Physiol., vol. 103, pp. 443451, 2007.
18. A. L. Vallerand, G. Savourey, and J. H. M. Bittel, Determination of heat depth in the
cold: partitional calorimetry vs. conventional methods, J. Appl. Physiol., vol. 72, pp.
13801385, 1992.
19. M. Bruse, ITCM: a simple dynamic 2-node model of the human thermo-regulatory
system and its application in a multi-agent system, Ann. Meteorol., vol. 41, pp. 398
401, 2005.
20. E. H. Wissler, A mathematical model of the human thermal system, Bull. Math.
Biophysic., vol. 26, pp. 147166, 1964.
21. D. Fiala, K. J. Lomas, and M. Stohrer, A computer model of human thermoregulation
for a wide range of environmental conditions: the passive system, J. Appl. Physiol.,
vol. 87, pp. 19571972, 1999.
22. D. Fiala, K. J. Lomas, and M. Stohrer, Computer prediction of human thermoregula-
tory and temperature responses to a wide range of environmental conditions, Int. J.
Biometeorol., vol. 45, pp. 143159, 2001.
23. Y. Qi, A new two-dimensional human thermal model and study of heat transfer in
living tissue, unpublished doctoral dissertation, University of Texas at Austin, 1994.
24. J. P. Rugh and D. Bharathan, Predicting human thermal comfort in automobiles, Paper
2005-01-2008, Presented at the Vehicle Thermal Management Systems Conference and
Exhibition, Toronto, May 2005.
25. N. Gao, J. Niu, and H. Zhzng, Coupling CFD and human body thermo-regulation
model for the assessment of personalized ventilation, HVAC&R Research, vol. 12, pp.
497518, 2006.
26. Q. Xu and X. Luo, Dynamic thermal comfort numerical simulation model on 3D
garment CAD, Appl. Math. Computation, vol. 182, pp. 106118, 2006.
27. G. V. A. Durnin and J. Womersley, Body fat assessed from total body density and its
estimation from skinfold thickness: measurement on 481 men and women aged 16
to 72 years, Brit. J. Nutrition, vol. 32, pp. 7797, 1974.
28. J. H. Himes, A. F. Roche, and R. M. Siervogel, Compressibility of skinfolds and mea-
surement of subcutaneous fatness, Am. J. Clinical Nutrition, vol. 32, pp. 17341740,
1979.
29. P. A. Hayes, P. J. Sowood, A. Belyavin, J. B. Cohen, and F. W. Smith, Subcutaneous fat
thickness measured by magnetic resonance imaging, ultrasound, and calipers, Med.
Sci. Sports Exer., vol. 20, pp. 303309, 1988.
30. P. A. Hayes, J. B. Cohen, and P. J. Sowood, Subcutaneous fat distribution of adult
males and females measured by nuclear magnetic resonances, in IAM Report No.
655, Farnborough, UK: RAF Institute of Aviation Medicine, 1987.
31. L. B. Rowell, Human circulation regulation during physical stress, New York: Oxford
University Press, 1986.
32. L. B. Rowell, Human cardiovascular adjustments to exercise and thermal stress,
Physiol. Rev., vol. 54, pp. 75159, 1974.
33. D. N. Proctor, S. C. Newcomer, D. W. Koch, K. U. Le, D. A. MacLean, and U. A.
Leuenberger, Leg blood flow during submaximal cycle ergometry is not reduced in
older normally active men, J. Appl. Physiol., vol. 94, pp. 18591869, 2003.
34. B. Nielsen, G. Savard, E. A. Richter, M. Hargreaves, and B. Saltin, Muscle blood flow
and muscle metabolism during exercise and heat stress, J. Appl. Physiol., vol. 69, pp.
10401046, 1990.

35. H. Barcroft and O. G. Edholm, The effect of temperature on blood flow and deep
temperature in the human forearm, J. Physiol., vol. 102, pp. 520, 1943.
36. H. Barcroft, K. D. Bock, H. Hensel, and A. H. Kitchin, Die Muskeldurchblutung des
Menschen bei indirekter Erwrmung und Abkhlung. Pflg. Arch. ges. Physiol., vol.
261, pp. 199210, 1955.
37. J.-M. R. Detry, G. L. Brengelmann, L. B. Rowell, and C. Wyss, Skin and muscle com-
ponents of forearm blood flow in directly heated man, J. Appl. Physiol., vol. 32, pp.
506511, 1972.
38. J. L. Saumet, D. L. Kellogg Jr., W. F. Taylor, and J. M. Johnson, Cutaneous laser-
Doppler flowmetry: influence of underlying muscle blood flow, J. Appl. Physiol.,
vol. 65, pp. 478481, 1988.
39. J. M. Johnson, W. F. Taylor, A. P. Shepherd, and M. K. Park, Laser-Doppler measure-
ment of skin blood flow: comparison with plethysmography, J. Appl. Physiol., vol. 56,
pp. 786803, 1984.
40. D. L. Kellogg, J. M. Johnson, and W. A. Kosiba, Control of internal temperature
threshold for active vasodilation by dynamic exercise, J. Appl. Physiol., vol. 71, pp.
24762482, 1991.
41. D. L. Kellogg, J. M. Johnson, and W. A. Kosiba, Selective abolition of adrenergic
vasoconstrictor responses in skin by local iontophoresis of bretylium, Am. J. Physiol.,
vol. 257 (Heart Circ. Physiol., vol. 26), pp. H11599H1606, 1989.
42. K. E. Cooper, O. G. Edholm, and R. F. Mottram, The blood flow in skin and muscle
of the human forearm, J Physiol., vol. 128, pp. 258267, 1955.
43. C. R. Wyss, G. L. Brengelmann, J. M. Johnson, L. B. Rowell, and M. Niederberger,
Control of skin blood flow, sweating, and heart rate: role of skin vs. core temperature,
J. Appl. Physiol., vol. 36, pp. 726733, 1974.
44. C. R. Wyss, G. L. Brengelmann, J. M. Johnson, L. B. Rowell, and D. Silverstein, Altered
control of skin blood flow at high skin and core temperatures, J. Appl. Physiol., vol.
38, pp. 839845, 1975.
45. J. M. Johnson, Nonthermoregulatory control of human skin blood flow, J. Appl.
Physiol., vol. 61, pp. 16131622, 1986.
46. J. M. Johnson and M. K. Park, Effect of upright exercise on threshold for cutaneous
vasodilation and sweating, J. Appl. Physiol.: Resp. Env. Exercise Physiol., vol. 50, pp.
814818, 1981.
47. W. F. Taylor, J. M. Johnson, W. A. Kosiba, and C. M. Kwan, Graded cutaneous vas-
cular responses to dynamic leg exercise, J. Appl. Physiol., vol. 64, pp. 18031809,
1988.
48. J. Smolander, J. Saalo, and O. Korhonen, Effect of work load on cutaneous vascular
response to exercise, J. Appl. Physiol., vol. 71, pp. 16141619, 1991.
49. G. P. Kenny, J. Periard, W. S. Journeay, R. J. Sigal, and F. D. Reardon, Cutaneous active
vasodilation in humans during passive heating postexercise, J. Appl. Physiol., vol. 95,
pp. 10251031, 2003.
50. D. L. Kellogg, J. M. Johnson, and W. A. Kosiba, Competition between cutaneous
active vasoconstriction and active vasodilation during exercise in humans, Am. J.
Physiol., vol. 261 (Heart Circ. Physiol., vol. 30), pp. H1184H1189, 1991.
51. G. L. Brengelmann, J. M. Johnson, L. Hermansen, and L. B. Rowell, Altered control
of skin blood flow during exercise at high internal temperatures, J. Appl. Physiol., vol.
43, pp. 790794, 1977.
52. E. R. Nadel, E. Cafarelli, M. F. Roberts, and C. B. Wenger, Circulatory regulation
during exercise in different ambient temperatures, J. Appl. Physiol: Resp. Env. Physiol.,
vol. 46, pp. 430437, 1979.

304 E. H. Wissler
53. D. L. Kellogg, J. M. Johnson, W. L. Kenney, P. E. Pergola, and W. A. Kosiba,

Mechanisms of control of skin blood flow during prolonged exercise in humans,
Am. J. Physiol., vol. 265 (Heart Circ. Physiol. 34), pp. H562H568, 1993.
54. N. Charkoudian and J. M. Johnson, Reflex control of cutaneous vaso-constrictor sys-
tem is reset by exogenous female reproductive hormones, J. Appl. Physiol., vol. 87,
pp. 381385, 1999.
55. D. P. Stephens, K. Aoki, W. A. Kosiba, and J. M. Johnson, Non-norandrenergic
mechanism of reflex cutaneous vasoconstriction in men, Am. J. Physiol. Heart Circ.
Physiol., vol. 280, pp. H1496H1504, 2001.
56. W. L. Kenney and C. G. Armstrong, Reflex peripheral vasoconstriction is diminished
in older men, J. Appl. Physiol., vol. 80, pp. 512515, 1996.
57. W. F. Taylor, J. M. Johnson, D. OLeary, and M. K. Park, Effect of high local tempera-
ture on reflex cutaneous vasodilation, J. Appl. Physiol.: Resp. Env. Exercise Physiol.,
vol. 57, pp. 191196, 1984.
58. N. Charkoudian, D. P. Stephens, K. C. Pirkle, W. A. Kosiba, and J. M. Johnson,
Influence of female reproductive hormones on local thermal control of skin blood
flow, J. Appl. Physiol., vol. 87, pp. 17191723, 1999.
59. H. Barcroft and O. G. Edholm, Temperature and blood flow in the human forearm,
J. Physiol., vol. 104, pp. 366376, 1946.
60. G. M. Brown, J. D. Hatcher, and J. Page, Temperature and blood flow in the forearm
of the Eskimo, J. Appl. Physiol., vol. 5, pp. 410420, 1953.
61. C. B. Wenger, L. A. Stephenson, and M. A. Durkin, Effect of nerve block on response
of forearm blood flow to local temperature, J. Appl. Physiol., vol. 61, pp. 227232,
1986.
62. W. F. Taylor, J. M. Johnson, D. S. OLeary, and M. K. Park, Modification of the cuta-
neous vascular response to exercise by local skin temperature, J. Appl. Physiol.: Resp.
Env. Exercise Physiol., vol. 57, pp. 18781884, 1984.
63. W. F. Taylor, J. M. Johnson, and W. A. Kosiba, Roles of absolute and relative load in
skin vasoconstriction responses to exercise, J. Appl. Physiol., vol. 69, pp. 11311136,
1990.
64. R. W. Bullard, Continuous recording of sweat rate by resistance hygrometry, J. Appl.
Physiol., vol. 17, pp. 735737, 1962.
65. H.-S. Tam, R. C. Darling, J. A. Downey, and H.-Y. Cheh, Relationship between evap-
oration rate of sweat and mean sweating rate, J. Appl. Physiol., vol. 41, pp. 777780,
1976.
66. B. Saltin, A. P. Gagge, U. Bergh, and J. A. J. Stolwijk, Body temperatures and sweating
during exhaustive exercise, J. Appl. Physiol., vol. 32, pp. 635643, 1972.
67. E. R. Nadel, J. W. Mitchell, B. Saltin, and J. A. J. Stolwijk, Peripheral modifications to
central drive for sweating, J. Appl. Physiol., vol. 31, pp. 828822, 1971.
68. E. R. Nadel, R. W. Bullard, and J. A. J. Stolwijk, Importance of skin temperature in the
regulation of sweating, J. Appl. Physiol., vol. 31., pp. 8087, 1971.
69. R. D. Wurster and R. D. McCook, Influence of rate of change in skin temperature on
sweating, J. Appl. Physiol., vol. 25, pp. 237240, 1969.
70. G. W. Mack, L. M. Shannon, and E. R. Nadel, Influence of -andrenergic block-
ade on the control of sweating in humans, J. Appl. Physiol., vol. 61, pp. 17011705,
1986.
71. N. Kondo, M. Shibasaki, K. Aoki, S. Koga, Y. Inoue, and C. G. Crandall, Function
of human eccrine sweat glands during dynamic exercise and passive rest, J. Appl.
Physiol., vol. 90, pp. 18771881, 2001.

72. E. R. Nadel, K. B. Pandolf, M. F. Roberts, and J. A. J. Stolwijk, Mechanism of thermal

acclimation to exercise and heat, J. Appl. Physiol., vol. 37, pp. 515520, 1974.
73. M. F. Roberts, C. B. Wenger, J. A. J. Stolwijk, and E. R. Nadel, Skin blood flow and
sweating changes following exercise training and heat acclimation, J. Appl. Physiol.,
vol. 43, pp. 133137, 1977.
74. K. B. Pandolf, B. S. Cadarette, M. N. Sawka, A. J. Young, R. P. Francesconi, and R. R.
Gonzalez, Thermoregulatory responses of middle-aged and young men during dry-
heat acclimation, J. Appl. Physiol., vol. 65, pp. 6571, 1988.
75. J. D. Cotter, M. J. Patterson, and N. A. S. Taylor, Sweat distribution before and after
repeated heat exposure, Eur. J. Appl. Physiol., vol. 76, pp. 181186, 1997.
76. M. J. Patterson, J. M. Stocks, and N. A. S. Taylor, Humid heat acclimation does not
elicit a preferential sweat redistribution toward the limbs, Am. J. Physiol. Reg. Integ.
Comp. Physiol., vol. 286, pp. R512R518, 2004.
77. J. D. Hardy and J. A. J. Stolwijk, Partitional calorimetric studies of man during expo-
sures to thermal transients, J. Appl. Physiol., vol. 21., pp. 17991806, 1966.
78. J. A. J. Stolwijk and J. D. Hardy, Partitional calorimetric studies of responses of man
in thermal transients, J. Appl. Physiol., vol. 21, pp. 967977, 1966.
79. C. G. Armstrong and W. L. Kenney, Effects of age and acclimation on responses to
passive heat exposure, J. Appl. Physiol., vol. 75, pp. 21622167, 1993.
80. J. Van Beaumont and R. W. Bullard, Sweating: its rapid response to muscular work,
Science, New Series, vol. 141, pp. 643646, 1963.
81. C. Gisolfi and S. Robinson, Central and peripheral stimuli regulating sweating dur-
ing intermittent work in men, J. Appl. Physiol., vol. 29, pp. 761768, 1970.
82. S. Robinson, Temperature regulation in exercise, Pediatrics, vol. 32, pp. 691702,
1963.
83. S. Robinson, F. R. Meyer, J. L. Newton, C. H. Tsao, and L. O. Holgersen, Relations
between sweating, cutaneous blood flow, and body temperature in work, J. Appl.
Physiol., vol. 20, pp. 575582, 1965.
84. S. Yanagimoto, T. Kuwahara, Y. Zhzng, S. Koga, Y. Inoue, and N. Kondo, Intensity-
dependent thermoregulatory responses at the onset of dynamic exercise in mildly
heated individuals, Am. J. Physiol. Reg. Integ. Comp. Physiol., vol. 285, pp. R200
R207, 2003.
85. F. Yamazaki, R. Sone, and H. Ikegami, Responses of sweating and body temperature
to sinusoidal exercise, J. Appl. Physiol., vol. 76, pp. 25412545, 1994.
86. H.-S. Tam, R. C. Darling, H.-Y. Cheh, and J. A. Downey, Sweating response: a means
of evaluating the set-point theory during exercise, J. Appl. Physiol., vol. 45, pp. 451
458, 1978.
87. C. L. Taylor and K. Buettner, Influences of evaporative forces upon skin temperature
dependency of human perspiration, J. Appl. Physiol., vol. 6, pp. 113123, 1953.
88. D. F. Brebner and D. McK. Kerslake, The time course of the decline in sweating pro-
duced by wetting the skin, J. Physiol. London, vol. 175, pp. 295302, 1964.
89. D. F. Brebner and D. McK. Kerslake, The effects of soaking the skin in water at vari-
ous temperatures on the subsequent ability to sweat, J. Physiol. London, vol. 194,
pp. 111, 1968.
90. W. K. Brown and F. Sargent II, Hidromeiosis, Arch. Environ. Health, vol. 11, pp. 442
453, 1965.
91. S. D. Gerking and S. Robinson, Declines in the rates of sweating of men working in
severe heat, Am. J. Physiol., vol. 147, pp. 370378, 1946.
92. S. Robinson and S. D. Gerking, Thermal balance of men working in severe heat, Am.
J. Physiol., vol. 149, pp. 476488, 1947.

306 E. H. Wissler
93. E. R. Nadel and J. A. J. Stolwijk, Effect of skin wettedness on sweat gland response, J.
Appl. Physiol., vol. 35, pp. 689694, 1973.
94. V. Candas, J. P. Libert, and J. J. Vogt, Effect of hidromeiosis on sweat drippage during
acclimation to humid heat, Eur. J. Appl. Physiol., vol. 44, pp. 123133, 1980.
95. D. A. Eyolfson, P. Tikuisis, X. Xu, G. Weseen, and G. G. Giesbrecht, Measurement
and prediction of peak shivering intensity in humans, Eur. J. Appl. Physiol., vol. 84,
pp. 100106, 2001.
96. J. S. Hayward, J. D. Eckerson, and M. L. Collis, Thermoregulatory heat production in
man: prediction equation based on skin and core temperatures, J. Appl. Physiol., vol.
42, pp. 377384, 1977.
97. C. A. Piantadosi, E. D. Thalmann, and W. H. Spaur, Metabolic response to respira-
tory heat loss-induced core cooling, J. Appl. Physiol., vol. 50, pp. 829834, 1981.
98. E. R. Nadel, S. M. Horvath, C. A. Dawson, and A. Tucker, Sensitivity to central and
peripheral stimulation in man, J. Appl. Physiol., vol. 29, pp. 603609, 1970.
99. P. Tikuisis and G. G. Giesbrecht, Prediction of shivering heat production from core
and mean skin temperatures, Eur. J. Appl. Physiol., vol. 79, pp. 221229, 1999.
100. T. H. Benzinger, Heat regulation: homeostasis of central temperature in man, Physiol.
Rev., vol. 49, pp. 671759, 1969.
101. B. Saltin, A. P. Gagge, and J. A. J. Stolwijk, Body temperatures and sweating during
thermal transients caused by exercise, J. Appl. Physiol., vol. 28, pp. 318327, 1970.
102. J. H. M. Bittel, Heat debt as an index for cold adaptation in men, J. Appl. Physiol., vol.
62, pp. 16271634, 1987.
103. J. H. M. Bittel, C. Nonotte-Varley, G. H. Livecchi-Gonnot, G. L. M. J. Savourey, and
A. M. Hanniquet, Physical fitness and thermoregulatory reactions in a cold environ-
ment in men, J. Appl. Physiol., vol. 65, pp. 19841989, 1988.
104. G. Savourey and J. Bittel, Thermoregulatory changes in cold induced by physical
training in humans, Eur. J. Appl. Physiol., vol. 78, pp. 379384, 1998.
105. A. L. Vallerand, G. Savourey, and J. H. M. Bittel, Determination of heat debt in the
cold: partitional calorimetry vs. conventional methods, J. Appl. Physiol., vol. 72, pp.
13801385, 1992.
106. E. R. Buskirk, R. H. Thompson, and G. D. Whedon, Metabolic response to cold air
in men and women in relation to total body fat content, J. Appl. Physiol., vol. 18, pp.
603612, 1963.
107. P. Cannon and W. R. Keatinge, The metabolic rate and heat loss of fat and thin men
in heat balance in cold and warm water, J. Physiol., vol. 154, pp. 329344, 1960.

Whole-Body Human Thermal Models: E. H. Wissler

Uploaded by

Copyright:

Available Formats

Whole-Body Human Thermal Models: E. H. Wissler

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Whole-Body Human Thermal Models: E. H. Wissler

Uploaded by

Copyright:

Available Formats

8

2009 by Taylor & Francis Group, LLC

8.4.4.5 Hidromeiosis 285

2009 by Taylor & Francis Group, LLC

In fact, it can be argued that no completely satisfactory model of a (physiologi-

An important example of interaction between systems is provided by the effect of

2009 by Taylor & Francis Group, LLC

between computed and observed temperatures under various conditions, but

8.2 Evolution of human thermal models

2009 by Taylor & Francis Group, LLC

2009 by Taylor & Francis Group, LLC

8.3 Model structure

2009 by Taylor & Francis Group, LLC

skin temperature, they are also affected by

2009 by Taylor & Francis Group, LLC

2009 by Taylor & Francis Group, LLC

SCF = 0.445 SkF4 1.013 (8.1)

SCF = 0.721 SkF4 + 7.300 (8.2)

8.4 Physiological control functions

8.4.1 Regulation of Blood Flow to Muscle

2009 by Taylor & Francis Group, LLC

2009 by Taylor & Francis Group, LLC

8.4.2 Regulation of Skin Blood Flow

1. Two branches of the sympathetic nervous system affect efferent neural

8.4.2.1 Measurement of Skin Blood Flow

2009 by Taylor & Francis Group, LLC

8.4.3 Algorithm for Computing Skin Blood Flow

qs = qs ,r AVD CVCM CVCL CVCE (8.4)

2009 by Taylor & Francis Group, LLC

reported that Xs = 0.086 and Xm = 0.636 in five forearms obtained postmortem

8.4.3.1 Active Vasodilation: Thermal Factors

AVD= 1.0 + (Tc Tc,th) (8.6)

in which is a gain constant. Tc,th depends on Ts and several nonthermal factors.

8.4.3.2 Active Vasodilation: The Effect of Exercise

2009 by Taylor & Francis Group, LLC

Threshold Central Temperature Increase: C 1

2009 by Taylor & Francis Group, LLC

Tc,th = 1.41 V 2 0.194 V + 0.01C (8.8)

2009 by Taylor & Francis Group, LLC

in which to is the instant at which Tc = Tc,th. If we assume that integration con-

8.4.3.3 Active Vasodilation: Other Factors

2009 by Taylor & Francis Group, LLC

AVD = AVD* 0.5 (Tc 38.0) (8.10)

where AVD* is the value computed using Equation (8.6).

AVD =0.25 (Tc Tc,o Tc,th) (8.11)

8.4.3.4 Cutaneous Vasoconstriction (CVCM): The Reflex

1.422 + tanh[ 0.275 (Ts - 32.0)]

8.4.3.5 Effect of Ts on Skin Blood Flow (SkBF) (CVCL)

2009 by Taylor & Francis Group, LLC

2009 by Taylor & Francis Group, LLC

8.4.3.6 Effect of Exercise on Cutaneous Vasoconstriction (CVCE)

in which W is the workload in watts.

8.4.3.7 Combination of AVD, CVCL, CVCM, and CVCE to Define qs

2009 by Taylor & Francis Group, LLC

2009 by Taylor & Francis Group, LLC

8.4.4 Regulation of Sweating

Figure 8.7 Comparison of computed and measured FBF.

2009 by Taylor & Francis Group, LLC

1. Local sweat rate (SR) is affected by Tc , Ts , and Ts.