Content uploaded by Andrew Grundstein
Author content
All content in this area was uploaded by Andrew Grundstein on Mar 02, 2015
Content may be subject to copyright.
AFFILIATIONS: Gr u n d s t e i n a n d Me e n t e M e y e r —Department of
Geography, University of Georgia , Athens, Georgia; do w d —
Depar tment of Geology, University of Georgia, Athens, Georgia
CORRESPONDING AUTHOR : Andrew Grundstein, Department
of Geography, Climatology Research Laboratory, University of
Georgia, Athens, GA 30602
E-mail: andrewg@uga.edu
The abstract for t his article can be found in this issue, following the
table of contents.
DOI:10.1175/ 2010B AM S2912.1
In final form 9 April 2010
©2010 American Me teorological S ociet y
A table of maximum rates of temperature change inside motor vehicles should be useful in
educating the public about the dangers of vehicle-related hyperthermia.
The danger of leaving young children unattended
in vehicles has been well documented. There are
no unique codes for identifying vehicle-related
hyperthermia deaths in the International Classifica-
tion of Diseases (ICD) or in any U.S. federal or state
data source (Guard and Gallagher 2005). However,
vehicle-related hyperthermia deaths in children in
the United States have been constructed from news
accounts. Guard and Gallagher (2005) observed an
average of 29 deaths per year during the years 1995–
2002, while a more extensive dataset by Null (2009)
observed an average of 37 deaths per year during the
years 1998–2009. Most cases (54%) involve caregivers
simply forgetting their children; however, more than a
quarter of vehicle-related hyperthermia deaths (27%)
involve children that were intentionally left in the car
(Guard and Gallagher 2005). In some cases, parents
did not want to disturb a sleeping child but were
unaware of how quickly the car could heat up. Such
behavior indicates a clear lack of understanding by
parents and caregivers about the dangers of leaving
children unattended in vehicles.
The interior of a car, along with the particular case
of a child strapped into a child safety seat, represents
a unique environment that may create particularly
dangerous conditions. Multiple studies have investi-
gated how ventilation, shading, and different meteo-
rological conditions may affect maximum cabin tem-
peratures and rates of temperature change (Table 1).
With the car in direct sunlight and no ventilation,
maximum temperatures may reach values exceeding
70°C (Table 1). These stunningly high temperatures
are caused by a greenhouse effect, where the windows
are transparent to solar radiation but opaque to long-
wave radiation. As a result, a positive net radiation
balance occurs that leads to heating. In addition, the
lack of ventilation from closed windows reduces the
transport of energy via convection and further con-
QUANTIFYING THE
HEAT-RELATED HAZARD
FOR CHILDREN IN MOTOR
VEHICLES
b y an d r e w Gr u n d s t e i n , Jo h n do w d , a n d Ve r n o n Me e n t e M e y e r
1183
SEPTEMBER 2010AMERICAN METEOROLOGICAL SOCIETY |
tributes to heating. Zumwalt and Petty (1976) note
that exposure to high environmental temperatures
only leads to a large rise in body temperature when
the temperature-regulating mechanisms are not oper-
ating efficiently. In a hot vehicle without ventilation,
physiological mechanisms typically used for cooling,
including longwave radiation and convection, would
be ineffect ive . Furtherm ore, the eff ici ency of e vapor a-
tive cooling would be reduced as evaporated perspira-
tion accumulated in the vehicle.
The meteorological community, along with local
healt h offic ial s and t he me dia , has b een ac tive i n dis-
seminating information about heat-related hazards.
The National Weather Service (NWS), for instance,
will issue “excessive heat warnings” or “heat advi-
sories” depending on the severity of the conditions
(NWS 1994). Many communities worldwide have
adopted synoptic-based heat watch–warning systems
where “heat alerts” and “heat warnings” are issued
(Sheridan and Kalkstein 2004). Layered on top of
these general heat alerts are more specific warnings
by public and private entities about the dangers of
leaving children unattended in motor vehicles. While
most of these warnings include some statements
about how hot a car might get, vehicle temperature
data from many early studies were often obtained
with small datasets and questionable methodologies,
such as placing the temperature sensor directly on
the car seat.
This research will focus on providing infor-
mation that may aid public off icials, child safety
advocates, and the media in better educating the
public about the dangers of leaving children unat-
tended in vehicles. Results from this study may also
be used as part of a public health response to a heat
health warning to emphasize the extreme danger
of vehicle-related hyperthermia in children during
those unusually hot periods. The first portion of the
study determines maximum temperature change at
different time intervals using carefully positioned
high-temporal-resolution temperature sensors. An
extension from previous work involves placing the
results in an easy-to-use table of vehicle temperature
changes that shows conditions that may occur under
the most severe circumstances. In addition, previous
studies have discussed but have not quantified how
the environmental conditions in a car would affect
the energy budget of a child (e.g., Zumwalt and Petty
1976; King et al. 1981). Thus, a human heat balance
model will be used to investigate the energy budget of
Ta b l e 1. Summary of rates of temperature change within passenger vehicles from different studies. Values
were estimated from figures and tables presented in the various studies and selected to represent cases
when the car had minimum ventilation and was in direct sunlight. Max is the maximum temperature
reached while the car was parked. All values are rounded to the nearest degree Celsius. The asterisk
means that day temperatures reached 89°C , and the car was parked for 12 h.
Study
5
min
10
min
30
min
60
min Max
Instrument type
and location
Parked
(h) Location and dates
Gibbs et al. (1995) 7 16 24 27 60 Electronic; placed
on front seat 1.5 New Orleans, LA ;
27 Jul 1995; 1430–1600 LT
Grundstein et al. (2009) 76 Electronic; 15 cm
below roof center 6Athens, GA;
1 Apr–31 Aug 2007
King et al. (1981) 19 21 25 25 66 Electronic; 15 cm
below roof center 2
Brisbane, Queensland,
Australia; summer 1978 and
1979; 1100–1300 LT
Mart y et al. (2001) 89 Electronic 12*
Zurich and Chur,
Switzerland;
Jan 1995– Mar 2000
McLaren et al. (2005) 4–10 7–13 17–18 22–23 47 Electronic; 38 cm
above rear seat 1Freemont, CA ; 16 days;
15 May–8 Aug 2002
Roberts and Roberts
(1976) 15 45
Liquid in glass;
15 cm above front
seat cushion
0.75 Baltimore, MD;
Sep 1975; afternoon
Surpure (1982) 78
Liquid in glass;
suspended from
driver’s seat
8
Oklahoma City, OK;
first week, Jul 1980;
080 0–1600 LT
Zumwalt and Petty
(1976) 58 Liquid in glass;
back seat 5Dallas, TX; Jun– Oct 1975;
1200–1700 LT
1184 SEPTEMBER 2010
|
a child in a hot car and the influence of variations in
humidity and sun exposure on levels of heat stress.
DATA AND METHODS. Ai r temperatu res
within a vehicle were measured on 58 days—from
April through 31 August 2007—in Athens, Georgia
(33.95°N, 83.32°W). Measurements were taken within
a metallic gray 2005 Honda Civic with gray cloth
seats. Approximately 67% of vehicle-related hyper-
thermia deaths in children occur in cars as opposed
to larger vehicles, such as minivans or SUVs; there-
fore, a car provides a representative vehicle for the
study (Guard and Gallagher 2005). In addition, tem-
peratures increase more rapidly in smaller vehicles
(Surpure 1982), thus providing better estimates for
“worst case” scenarios. The car was parked in an open
asphalt-covered lot with direct exposure to sunlight.
Additionally, the windows were closed during data
collection to limit ventilation and maximize heating
within the vehicle.
The vehicle temperature data were collected by an
Onset Computer Corporation HOBO temperature
sensor (H008-003-002; resolution = 0.4°C, accuracy
= ±0.7°C) that recorded temperatures every 5 min.
The sensor was attached to a string and suspended
approximately 15 cm from the ceiling to avoid direct
exposure to sunlight and to be sufficiently far from
surfaces to accu rately measure air temperature.
Ambient outdoor air temperature, dewpoint tempera-
ture, and solar radiation
data at a 5-min resolution
were obtained from an ad-
jacent weather station op-
erated by the Department
of Geography Climatology
Research Laboratory (CRL),
located on the roof of the
bu i lding approx i mat ely
12 m higher in elevation
tha n t he car and 125 m
from the parking lot. Solar
radiation was measu red
with a Davis 6450 silicon
phot odiod e s enso r, and
temperature and humid-
it y were measu red with
a shielded and aspi rated
Davis 6382 temperat ure
and humidity sensor. Cloud
cover data at hourly resolu-
tion were obtained from a
NWS Automated Surface
Observing Station (ASOS)
located approximately 5 km away at Athens–Ben
Epps Airport.
For each day with vehicle temperature data, the
maximum 5-, 10-, 30-, and 60-min temperature
changes were computed. The focus of the study is on
the most severe possible conditions with the greatest
temperature changes. Thus, maximum temperature
changes were examined only for clear days. In total,
14 days—ranging from 12 April to 24 July—were
examined (Table 2). This period encompasses dates
near the spring equinox and the summer solstice, pro-
viding for a variety of solar angles and solar radiation
values. Maximum rates of temperature change were
assessed from 1100 to 1300 EDT to capture the pe-
riod in the vicinity of solar noon (around 1300 EDT)
when solar heating would be most intense. Graphs of
car temperatures (Figs. 1 and 2) show that the rates
of temperature change are reduced approaching
1300 EDT, as the hot car emitted greater amounts of
longwave radiation relative to incoming solar radia-
tion, thereby reducing net radiation.
The second portion of the study examines the en-
ergy budget of a child in a hot car using a human ther-
mal exchange model called the man–environment
heat exchange model (MENEX). It has been employed
in several experimental studies of the human thermal
budget (e.g., Katavoutas et al. 2009; Tuller 1997) and
is capable of computing the various energy budget
components, including absorbed solar radiation,
Ta b l e 2. Interior vehicle air temperature change by time inter val
and average solar radiation (1100–1300 EDT) for study days in 2007.
Temperature is in degrees Celsius, and the solar radiation is in watts per
square meter.
Month Day 5 min 10 min 30 min 60 min Solar
413 1.7 3.3 9.6 17.4 794
416 2.0 3.3 9. 3 16.8 8 21
417 2.5 4.7 12.3 21.0 850
425 3.4 6.4 12 .9 20.4 810
430 3.0 5.5 14.7 24.0 840
501 3.6 6.8 14.1 21.0 829
521 3.0 5.6 15. 6 25.8 860
522 2.1 3.8 10.8 18.6 794
530 2.6 5.1 13 . 5 24.0 790
618 1.8 3.2 8.7 15.0 776
621 3.5 5.8 13.8 25.8 884
622 2.8 5.0 13. 5 24.0 878
712 2.5 4.5 12 .6 22.2 850
724 3.0 5.9 15.6 24.6 813
1185
SEPTEMBER 2010AMERICAN METEOROLOGICAL SOCIETY |
net longwave radiation, turbulent fluxes of sensible
and latent heat, as well as metabolic heat production
and heat loss through respiration (Błażejczyk 1994).
Further, MENEX accounts for physiological factors
such as skin temperature, skin wetness, and clothing
albedo and insulation. The human heat balance equa-
tion is defined as
S = M + Q + H + LE + C + Res, (1)
where S is the net heat storage or change in body
heat content; M is metabolic heat production; Q is
the radiation balance of the person; H and LE are
convective transfers of energy via sensible and latent
heat, respectively; C is conduction; and Res is the heat
loss by respiration. Positive (negative) fluxes indicate
a gain (loss) in net heat storage. Changes in the body
heat content will be used to quantify the heat stress on
the child. The degree to which changes in body heat
content relate to particular health outcomes such as
heat stroke or death, however, is not
well established and may vary with
the age and health of the child.
Model simulations were per-
formed of a child seated inside the
car as well as one outside the car
to serve as a reference. The human
thermal exchange model was modi-
fied slightly for the simulations of a
child within the car. First, longwave
radiation emitted by the interior of
the vehicle was determined using the
average interior surface temperature
and the Stefan–Boltzmann equation
with an emissivity of 0.97. Second, a
conduction term was computed to
account for the fact that a child would be strapped in
a child safety seat as follows:
C = K(Tcar − Tskin)A, (2)
where K is the heat transfer coeff icient t hrough
clothing as computed by MENEX, Tcar − Tskin is the
temperature gradient between the child’s skin Tskin
and safety seat Tcar, and A is a constant that accounts
for the portion of the child’s body that is in contact
with the seat. Here, the simulation is performed for
a 2-yr-old toddler sitting in a forward-facing child
safety seat. A contact value of 0.24 is used, which
represents a child’s torso, legs, and head in contact
with the seat (Raja and Nicol 1997). The degree of
contact, however, may vary somewhat with the par-
ticular child safety seat used. For instance, the degree
of contact may be higher in an infant safety seat that
is designed to cradle the child. Nevertheless, the re-
sults should approximately represent conditions for
children 3 yr old or younger who are placed in child
safety seats.
Several input values were adjusted to account for
the physiology of the child and the climate conditions
within the car (Table 4). The physiological charac-
teristics of an average 2-yr-old toddler were used in
modeling. The metabolic rate of the child was esti-
ma ted at 61 W m−2, which is consistent with the caloric
needs of 700 kcal day−1 (Durnin 1981). Ambient air
temperatures during the study period were generally
high; therefore, it is assumed that the child is dressed
in summer attire with clothing insulation of 0.6 clo1
(Błażejczyk 1994). Also, the child is assumed to be
Fi g . 2. Time series of interior vehicle air temperatures,
ambient outdoor air temperatures, and solar radiation
from 0900 to 1700 on 22 Jun 2007.
1 ANSI/ASHRAE (1992) defines a “clo” as a unit to express
the thermal insulation provided by garments and clothing
ensembles, where 1 clo = 0.155 m2*C / W.
Fi g . 1 . Interior vehicle air temperatures between 1100 and 1300 EDT
on 14 different clear days.
1186 SEPTEMBER 2010
|
unacclimatized to hot conditions with skin that is
initially 33°C and dry (Fanger 1972; Hoppe 1998).
Physiological changes are simulated by increasing
the sk in t empe rat ure and wetn ess . Mod eled s kin tem-
peratures were not used, as Katavoutas et al. (2009)
found the empirical equation used in MENEX may
not be appropriate to use for unacclimatized people.
Rather, skin temperatures were varied from 33°C to
a maximum of 37°C in 20 min based on observations
from Fiala et al. (2001), where unacclimatized sub-
jects were exposed to high temperatures. All model
simulations used the same skin temperatures, so that
the influence of different environmental conditions
could be isolated. Wetness is computed as a function
of skin temperature, reaching complete wetness at
temperatures >36.5°C (Błażejczyk 1994). Wind speed
in the car without ventilation (i.e., windows rolled
up) is minimal. A nominal value of 0.1 m s−1 is used,
which is similar to values used in studies of indoor
climates (Hoppe 1998).
Model simulations were performed for 22 June
2007 from 1300 to 1400 EDT using data collected
from the study vehicle and the nearby CRL weather
station. This was a clear day, falling near the summer
solstice and providing suitable conditions for a case
study of a worst-case scenario. Input solar radiation
and humidit y data were obtained from the CRL
weather station. Initial humidity levels in the vehicle
were assumed to be similar to outside values, as air
was entrained in the vehicle while placing the child in
the car. Over time, evaporated perspiration from the
child would increase humidity within the vehicle. All
perspiration was assumed to evaporate, and dewpoint
temperatures were iteratively increased each time
step. The windows of the vehicle attenuate some of
the incoming solar radiation. Measurements of solar
radiation taken inside and outside the car indicate
that values within the vehicle were reduced by ap-
proximately 50%. Thus, only half of the measured
solar radiation was input into the model. The interior
surface temperatures of the car were obtained using
an Omega OS530HR handheld infrared thermom-
eter. Measurements were taken of the seats, floor,
ceiling, and windows and averaged to provide a rep-
resentative interior surface temperature. The average
interior surface temperature was used in computing
longwave radiation emissions from the car, and the
seat temperature was used in calculating the conduc-
tion term.
RE S ULTS . Maximum rates of temperature change.
The 14 days utilized in this study provide a repre-
sentative sample for assessing ideal conditions for
maximum heating (Table 2). The sample data are
distributed over the entire study period, with days in
each month from April through July. Overall, there
were 36 days with clear skies and 30 of those days
with solar radiation data. The average solar radiation
during peak heating periods (e.g., 1100–1300 EDT)
for those 30 days was 824 W m−2 , with a range from
758 to 884 W m−2. The data used for this study had
average solar radiation that was slightly greater at
828 W m−2 and a range from 776 to 884 W m−2. The
sample dataset includes the two days with the greatest
average solar radiation and three other days among
the top 10 in solar radiation.
There were different initial ambient air tempera-
tures and rates of temperature change for the 14 peak
heating periods studied (Fig. 1). Initial ambient air
temperatures at 1100 EDT ranged from 15° to 34°C,
with temperatures reaching 43°–62°C by 1300 EDT.
Temperature changes were computed for 5-, 10-, 30-,
and 60-min periods (Table 2). The average (maxi-
mum) temperature change over each time interval
is 2.7°C (3.6°C) for 5 min, 4.9°C (6.8°C) for 10 min,
12.6°C (15.6°C) for 30 min, and 21.5°C (25.8°C) for
60 min. The comparatively lower rates of temperature
change for cloudy days is indicated by looking at
values on two days (4 May and 2 July) with complete
cloud cover. Average temperature changes are 1.7°C
for 5 min, 3.0°C for 10 min, 7.6°C for 30 min, and
8.9°C for 60 min.
At longer time intervals, these results are consis-
tent with other studies that observed hourly tempera-
ture increases ranging from 22° to 27°C (McLaren
et al. 2005; Gibbs et al. 1995; King et al. 1981). There
are some differences with other studies at shorter
intervals that are likely related to the positioning of
the sensors. Large rates of temperature change may
have been related to the exposure of the sensor to
direct sunlight (King et al. 1981) or the location of
the sensor on the car seat (Gibbs et al. 1995), which
would be influenced by the seat temperature and not
be representative of vehicle air temperatures. In all
cases, the maximum temperature changes occurred
around noon. This timing is tied to the radiation
balance of the vehicle and can be illustrated using
measurements of incoming solar radiation and ve-
hicle air temperatures for 22 June 2007 (Fig. 2). The
cabin and ambient air temperatures are relatively
similar early in the day but diverge rapidly between
1130 and 1300 EDT. Between 1100 and 1300 EDT, the
solar angles are high (61°–79°C, respectively), leading
to intense solar heating of the car seats, which in turn
warms the overlying air. Any increase in temperature
of the car, however, will lead to increases in longwave
1187
SEPTEMBER 2010AMERICAN METEOROLOGICAL SOCIETY |
emissions to the fourth power, as indicated by the
Stefan–Boltzmann law. Thus, the enormous rise in
car temperature during this time increases longwave
emissions relative to incoming solar radiation and
slows the subsequent rate of heating.
A table of maximum vehicle temperature changes. A
table of maximum passenger compartment tem-
perature changes was developed to aid in advising
the public about the dangers of leaving children un-
attended in cars (Table 3). The table considers initial
ambient air temperature when the car is parked, and
the temperature changes for 5-, 10-, 30-, and 60-min
intervals. The table is designed to show the maximum
possible changes in temperature over each interval by
using the greatest observed temperature changes from
Table 2. Of course, factors such as whether the car is
in direct sunlight, the time of day (i.e., different solar
angles and solar radiation), the amount of cloudiness,
an d ventil ation (i.e., win dows rolle d dow n) wil l inf lu-
ence the actual amount of temperature change.
Table 3 shows that the thermal hazard is a func-
tion of both the initial ambient air temperature and
the time interval over which the heating occurs. One
way to characterize the meaning of these tempera-
tures in terms of a health hazard is to place them in
the context of heat health warnings
provided by the National Weather
Service (NWS 1994). For example, a
heat advisory is issued by the NWS
when the heat index is 41°–46°C
for less than 3 h. An excessive heat
warning is issued when the heat
index is ≥41°C for more than 3 h
or exceeds 46°C for any period of
time. Thus, one could say that if the
outside air temperature is 34°C, then
the vehicle could reach the level of a
heat advisory within 10 min and an
excessive heat warning within 30
min. Of course, the temperatures
listed in the table only provide an
indicator of the level of danger. One
must consider the age and health
of the child when assessing danger
as well as the fact that children in
general are particularly susceptible
to heat-related illnesses (Hoffman
2001). Children’s smal l size gives
them a high surface-area-to-mass
ratio that a llows them to absorb
more energy from the environment
than an adult, and their ability to
cool through perspiration is less efficient (Hoffman
2001). In addition, young children are not able to
adjust their behavior in response to the heat, such as
removing clothing or exiting the car (McLaren et al.
2005; Hoffman 2001)
Modeling the energy budget of a child in a hot car. A
human heat balance model was used to examine the
influence of humidity and full sun exposure on the
heat stress of a child in a hot vehicle. The modeling
study was performed using data collected from a 2005
Honda Civic that was parked for approximately 4 h.
During the study period from 1300 to 1400 EDT, the
sky was clear; outdoor air and dewpoint temperatures
averaged 33° and 11°C, respectively; and solar radia-
tion was 954 W m−2 . Outside wind speeds, adjusted
from the roof to 1.5 m using a logarithmic wind
profile, were low at approximately 1 m s−1. The air
temperature within the vehicle averaged 65°C, and
the average temperature of interior surfaces includ-
ing the ceiling, floor, seats, and windows was 69°C.
Model simulations were conducted in 5-min time
steps during the course of the hour.
Fou r si mulat ions were pe rformed for a ch ild wit h-
in the car, including a default simulation and simula-
tions representing conditions with high humidity, low
Ta b l e 3. Maximum interior vehicle air temperature reached for
different time intervals. The values are rounded to the nearest
degree, so that the car heats by 4°C in 5 min, 7°C in 10 min, 16°C
in 15 min, and 26°C in 60 min.
Initial ambient air temperature (°C)
50 54 57 66 76
48 52 55 64 74
46 50 53 62 72
44 48 51 60 70
42 46 49 58 68
40 44 47 56 66
38 42 45 54 64
36 40 43 52 62
34 38 41 50 60
32 36 39 48 58
30 34 37 46 56
28 32 35 44 54
26 30 33 42 52
24 28 31 40 50
22 26 29 38 48
20 24 27 36 46
5 min 10 min 30 min 60 min
Time interval
1188 SEPTEMBER 2010
|
humidity, and shade. Dewpoint temperatures where
varied by ±10°C about the mean observed dewpoint
temperature of 11°C to simulate high (21°C) and low
(1°C) humidity conditions. In the shade scenario,
the child is not exposed to direct beam radiation but
does receive some diffuse solar radiation, assumed
to be 30% of global solar radiation (Rosenberg et al.
1983). For comparison, a model simulation was also
performed for t he same day and times of a child
standing outside the car. Output from the model
included absorbed solar, net longwave, latent heat,
sensible heat, conduction, metabolic heat production,
respiratory heat losses, and net heat storage.
The average change in body heat content of the
child in the hot vehicle during the 1-h period was
250 W m−2 (Table 4). This is more than 3 times the
net storage gain for a child standing outside of the
car. The energy transfer mechanisms directed toward
the child were very different inside and outside the
vehicle. Within the vehicle, net longwave accounted
for 44%, conduction for 28%, sensible heat for 16%,
and solar for 12% of the exogenous energy transfers to
the child. Thus, the dominant energy transfers were
via conduction and longwave radiation. Outside of
the car, most of the energy transfers to the child were
from solar radiation, with approximately 89% from
solar radiation, 10% from net longwave radiation, and
1% from sensible heat.
The great difference in both the magnitude and
distribution of energy f luxes may be explained by
the unique environment of the interior of the hot
vehicle. The extremely high surface temperatures
in the vehicle direct t he vast majority of radiant
energy in the form of longwave radiation toward
the child as well as transferring large amounts of
energy v ia conduct ion t hroug h the c hild s afety seat.
The relatively small contribution from absorbed
solar radiation occurs because the wi ndows attenu-
ate some of the insolation, and the projected area
that strikes the body is small during periods with
high solar angles. The strong temperature gradient
between the air and child’s skin results in transfers
of sensible heat toward the child; however, the flux
is only 43 W m−2 because of the lack of turbulence.
Evaporative cooling from latent heat transfers away
from the child averages −80 W m−2 . The small value
is related to the low turbulence and the negative
feedback of increased humidity levels in the vehicle
from evaporated perspiration. Indeed, moisture
from the child increased the relative humidity (RH)
from 6% to 19% during the hour. The results support
the observations of Zumwalt and Petty (1976) that
mechanisms that are generally available for cooling
are either reduced (evaporation of perspiration) or
actually lead to heat gains for the child (longwave
radiation and sensible heat).
Ta b l e 4. Average input and output values for human heat balance model simulations of a child in a hot
car for 22 Jun 2007 from 1300 to 1400 EDT. All values except for clothing insulation and wind speed are
rounded to the nearest whole number.
Variable Outside
Initial
car
High-humidity
car
Low-humidity
car
Shaded
car
Biophysical inputs Skin temperature (°C) 33 36 36 36 36
Clothing insulation (clo) 0.6 0.6 0.6 0.6 0.6
Meteorological conditions Air temperature ( °C) 33 65 65 65 65
Average cabin surface
temperature (°C) 69 69 69 69
Initial dewpoint ( °C) 11 11 21 −1 11
Initial/final RH (%) 26/24 6 /19 11 / 2 0 3 /19 6/19
Wind speed (m s−1)1.0 0 .1 0.1 0.1 0.1
Energy fluxes Absorbed solar (W m−2)62 32 32 32 10
Net longwave (W m−2 ) 7 119 119 119 11 9
Sensible heat (W m−2) 1 43 43 43 43
Latent heat (W m−2)−53 −80 −60 −90 −80
Conduction (W m−2 ) 0 77 77 77 77
Metabolism (W m−2)61 61 61 61 61
Respiration (W m−2 )−5 −2 −2 −3 −2
Net heat storage (W m−2)78 250 270 239 228
1189
SEPTEMBER 2010AMERICAN METEOROLOGICAL SOCIETY |
Simulations were performed to examine how
shading and variations in humidity would affect the
change in the body heat content of the child. The first
set of simulations varied the initial dewpoint from −1°
to 21°C. As one would expect, the drier (more humid)
conditions are associated with greater (less) evapo-
ration. The high-humidity scenario resulted in 8%
gr eater net hea t storag e as a res ult of reduce d ev apor a-
tive cooling. In contrast, the greater latent heat fluxes
away from the child in the low-humidity simulation
reduced net heat storage by about 4% compared
with the default simulation. The relative differences
among the different scenarios were small because of a
negative feedback of evaporated perspiration; that is,
increased perspiration will lead to greater humidity
within the car, which will reduce humidity gradients
and subsequent evaporation rates. This feedback ex-
plains how relative humidity for both the high- and
low-humidity scenarios began at different values but
converged near 20%. The second set of simulations
involved the child being shaded from direct sunlight.
This reduces absorbed energy by 22 W m−2 and the
total heat storage by approximately 9%.
That case study represents a particularly harsh
scenario. Some children are “forgotten” and left in
the vehicle all day, while others are placed in hot
cars in the middle of the day, in some cases with a
broken air conditioner blowing hot air, and have died
wit hin 15 mi n (Associated Press, 24 August 2005 and
9 August 2006). Thus, the energy balance of the child
and the rate at which the child would suffer a heat-
related illness will vary depending on the particular
conditions. Air and surface temperatures of the car,
for example, will differ depending on how long the
car has been parked, whether it is in direct sunlight,
and the degree of ventilation. Also, absorbed solar
radiation may actually increase earlier or later in the
day as lower sun angles lead to larger surface areas
on the body that are exposed to sunlight.
CONCLUSIONS. More than 2,500 children die
each year from unintentional injuries in the United
States (Borse et al. 2008). While the number of chil-
dren who die from vehicle-related hyperthermia is
only a small percentage of this total, it is a hazard
that is so easily preventable. Children should never
be left unattended in vehicles, regardless of ambient
air temperatures, because of risks such as abduction
or injury to the child from incidents such as being
asphyxiated from entrapment by vehicle windows
(NHTSA 2009). However, there is a clear pattern—
both seasonally and by temperature threshold—to
vehicle-related hyperthermia deaths that suggests
targeted “reminders” may be helpful as a warning
strategy. Approximately 75% of deaths occur during
the summer months (Guard and Gallagher 2005) and
data from 231 vehicle-related hyperthermia deaths
during the 2003–08 period show that more than 70%
occurred on days with maximum outdoor tempera-
tures ≥31°C (Null 2009). In direct sunlight and during
the course of an hour, temperatures within a car could
exceed 57°C on such days. Modeling results show that
the dominant transfers of energy toward the child
(longwave radiation and conduction) are driven by
temperature; therefore, passenger compartment air
temperatures serve as a good indicator of the heat-
related hazard. High humidity and exposure to direct
solar radiation will also increase the net heat storage
of a child; however, the inf luence of humidity varia-
tions is limited because of a negative feedback.
Unfortu nately, many people are u naware of
the dangers of leaving a child unattended in a car.
Thus, education may be an important component
in reducing vehicle-related hyperthermia deaths.
Sheridan (2007) documented that heat health warn-
ings are helpful in raising awareness of the dangers
of heat-related illness. An easy-to-use table of vehicle
temperatures changes, as presented here, may help
public off icials and the media communicate with
the public about the hazard of vehicle-related hyper-
thermia in children. Temperature thresholds used
by the National Weather Service for their heat health
warnings may be used to place the temperatures in
the context of healt h hazards. Importantly, these
thresholds should only be used to emphasize how
quickly temperatures in a car can reach hazardous
levels and therefore as a warning to never leave a child
unattended in a vehicle. We hope this characteriza-
tion linking temperatures with an explicit warning
about health dangers to children will modify the
behavior of caregivers and result in fewer tragedies.
Preventing injuries to children from vehicle-related
hyperthermia will ultimately require a multifaceted
approach including education, regulation, engineer-
ing, and legislation (Guard and Gallagher 2005).
REFERENCES
ANSI/ASHRAE, 1992: Thermal environmental condi-
tions for human occupa ncy. American National
Standards Institute/American Society of Heating,
Refrigerating, and Air-Conditioning Engineers
Standard 55-1992, 22 pp.
Błażejczyk, K., 1994: New climatological-and-physio-
logical model of the human heat balance outdoor
(MENEX) and its applications in bioclimatologi-
1190 SEPTEMBER 2010
|
cal studies in different scales. Bioclimatic Research
of the Human Heat Balance, K. Błażejczyk and B.
Krawczyk, Eds., Zeszyty IGiPZ PAN, 27–58.
Borse, N. N., J. Gilchrist, A. M. Dellinger, R. A. Rudd,
M. F. Ballesteros, and D. A. Sleet, 2008: CDC
childhood injury report: Patterns of unintentional
injuries among 0-19 year olds in the United States,
2000–2006. National Center for Injury Prevention
and Control, Centers for Disease Control and Pre-
vention, 116 pp.
Durnin, J. V. G. A., cited 2009: Basal metabolic rate in
man. Proc. Joint FAO/WHO/UNU Expert Consulta-
ti on on Ene rgy an d Protei n Requ irem ent s, Rome, Ita-
ly, FAO, Rep. FAO/WHO/UNU EPR/81/5. [Available
online at www.fao.org/DOCREP/MEETING/004/
M2845E/M2845E00.HTM.]
Fanger, P. O., 1972: Thermal Comfort: Analysis and
Applications in Environmental Engineering. McG raw-
Hill, 244 pp.
Fiala, D., K. J. Lomas, and M. Stohrer, 2001: Computer
prediction of human thermoregulatory and tem-
perature responses to a wide range of environmental
conditions. Int. J. Biometeor., 45, 143–159.
Gibbs, L. I., D. W. Lawrence, and M. Kohn, 1995: Heat
exposure in an enclosed automobile. J. La. State Med.
Soc., 147, 545–546.
Grundstein, A., V. Meentemeyer, and J. Dowd, 2009:
Maximum vehicle car temperatures under different
meteorological conditions. Int. J. Biometeor., 53,
255–261.
Guard A., and S. S. Gallagher, 2005: Heat related deaths
to young children in parked cars: An analysis of 171
fatalities in the United States, 1995-2002. Inj. Prev.,
11, 33–37.
Hoffman, J. L., 2001: Heat-related illness in children.
Clin. Pediatr. Emerg. Med., 2, 203–210.
Hoppe, P., 1998: Comfort requirements in indoor cli-
mate. Energy Build., 11, 249–257.
Katavoutas, G., G. Theoharatos, H. A. Flocas, and D. N.
Asimakopoulos, 2009: Measuring the effects of heat
wave episodes on the human body’s thermal balance.
Int. J. Biometeor., 53, 177–187.
King, K., K. Negus, and J. C. Vance, 1981: Heat stress
in motor vehicles: A problem in infancy. Pediatrics,
68, 579–582.
Marty, W. T., T. Sigrist, and D. Wyler, 2001: Tempera-
ture variations in automobiles in various weather
conditions: An experimental contribution to the
determination of time of death. Amer. J. Forensic
Med. Pathol., 22, 215–219.
McLaren, C., J. Null, and J. Quinn, 2005: Heat stress
from enclosed vehicles: Moderate ambient tempera-
tures causes significant temperature rise in enclosed
vehicles. Pediatrics, 116, e109–e112 , doi:10.154 2/
peds.2004-2368.
NHTSA, cited 2009: Not-in-traffic surveillance 2007—
Children. NHTSA Rep. DOT HS 811 116, 2 pp.
[Available online at http://www-nrd.nhtsa.dot.gov/
Pubs /811116 .pd f.]
Null, J., cited 2009: Hyperthermia deaths of children
in vehicles. [Available online at ht t p://gg w eat her.
com/heat.]
NWS, 1994: Excessive heat watch, warnings and advi-
sory heat index criteria. Regional Operations Manual
Letter E-5-94, Eastern Region, NWS, Bohemia, NY,
3 pp.
Raja, I. A., and F. Nicol, 1997: A technique for record-
ing and analysis of postural changes associated with
thermal comfort. Appl. Ergon., 28, 221–225.
Roberts, K. B., and E. C. Roberts, 1976: The automobile
and heat stress. Pediatrics, 58, 101–104.
Rosenberg, N. J., B. L. Blad, and S. B. Verma, 1983:
Microclimate: The Biological Environment. John
Wiley and Sons, 495 pp.
Sheridan, S., 2007: A survey of public perception and re-
sponse to heat warnings across four North American
cities: An evaluation of municipal effectiveness. Int.
J. Biometeor., 52, 3–15.
—
, and L. S. Kalkstein, 2004: Progress in heat watch–
warning system technology. Bull. Ame r. Met eor. So c.,
85, 1931–1941.
Surpure, J. S., 1982: Heat-related illness and the auto-
mobile. Ann. Emerg. Med., 11, 263–265.
Tuller, S. E., 1997: Climatic controls of the cool human
thermal sensation in a summertime onshore wind.
Int. J. Biometeor., 41, 26–33.
Zumwalt, R. E., and C. S. Petty, 1976: Temperature
in closed automobiles in hot weather. Forensic Sci.
Gaz., 7, 7–8.
1191
SEPTEMBER 2010AMERICAN METEOROLOGICAL SOCIETY |
