See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/24032501

Maximum Vehicle Car Temperatures under Different Meteorological

Conditions

Article in International Journal of Biometeorology · March 2009

DOI: 10.1007/s00484-009-0211-x · Source: PubMed

CITATIONS READS

106 21,640

3 authors, including:

Andrew Grundstein
University of Georgia
121 PUBLICATIONS 2,328 CITATIONS

SEE PROFILE

All content following this page was uploaded by Andrew Grundstein on 17 June 2016.

The user has requested enhancement of the downloaded file.

Int J Biometeorol
DOI 10.1007/s00484-009-0211-x
ORIGINAL PAPER
Maximum vehicle cabin temperatures under different
meteorological conditions
Andrew Grundstein & Vernon Meentemeyer &
John Dowd
Received: 4 July 2008 / Revised: 11 December 2008 / Accepted: 24 January 2009
# ISB 2009
Abstract A variety of studies have documented the
dangerously high temperatures that may occur within the
passenger compartment (cabin) of cars under clear sky
conditions, even at relatively low ambient air temperatures.
Our study, however, is the first to examine cabin temper-
atures under variable weather conditions. It uses a unique
maximum vehicle cabin temperature dataset in conjunction
with directly comparable ambient air temperature, solar
radiation, and cloud cover data collected from April
through August 2007 in Athens, GA. Maximum cabin
temperatures, ranging from 41–76°C, varied considerably
depending on the weather conditions and the time of year.
Clear days had the highest cabin temperatures, with average
values of 68°C in the summer and 61°C in the spring.
Cloudy days in both the spring and summer were on
average approximately 10°C cooler. Our findings indicate
that even on cloudy days with lower ambient air temper-
atures, vehicle cabin temperatures may reach deadly levels.
Additionally, two predictive models of maximum daily
vehicle cabin temperatures were developed using common-
ly available meteorological data. One model uses maximum
ambient air temperature and average daily solar radiation
while the other uses cloud cover percentage as a surrogate
for solar radiation. From these models, two maximum
vehicle cabin temperature indices were developed to assess
the level of danger. The models and indices may be useful
A. Grundstein (*) : V. Meentemeyer
Department of Geography, Climatology Research Laboratory,
University of Georgia,
Athens, GA 30602, USA
e-mail: andrewg@uga.edu
J. Dowd
Department of Geology, University of Georgia,
Athens, USA
for forecasting hazardous conditions, promoting public
awareness, and to estimate past cabin temperatures for use
in forensic analyses.
Keywords Automobile . Temperature . Hyperthermia .
Weather
Introduction
Every year infants and young children die from hyperther-
mia after being trapped within cars. A study by Guard and
Gallagher (2005) found that, in the United States, 171
children died from being trapped in vehicles, either
intentionally or by accident, during the period from 1995–
2002. Since 1998, Null (2008) found an average of 36
deaths per year from vehicle-related hyperthermia. Often
parents leave children in the car so as to avoid waking them
but are unaware of how quickly temperatures may increase
to dangerous levels (Guard and Gallagher 2005). Indeed,
under ideal conditions with full sun exposure and no
ventilation, temperatures have been observed to increase by
22–27°C within an hour (McLaren et al. 2005; Gibbs et al.
1995; King et al. 1981) and reach maximum temperatures
up to 89°C (Marty et al. 2001).
The temperature increase in the car is caused by a
greenhouse effect associated with a radiation imbalance and
reduced ventilation. A net radiation imbalance occurs as
solar radiation can pass into the vehicle through the
windows but long-wave radiation emitted by the car is
“trapped” and prevented from escaping (Roberts and
Roberts 1976). In addition, a car without ventilation (e.g.,
open windows) blocks the loss of energy via convection.
Studies comparing vehicles in direct sunlight and shade
reveal that cabin temperatures reached values that were 8°C–

19°C greater in the sun (Roberts and Roberts 1976; Surpure
1982). In addition, cars with greater ventilation had
substantially lower cabin temperatures. Maximum cabin
temperatures in vehicles with open windows were over
28°C cooler than those with closed windows but only 8°C–
15°C cooler with the windows “cracked” (open 50 mm)
(Roberts and Roberts 1976; King et al. 1981; Surpure 1982).
Most of these studies collected data on a single day,
typically a clear one in the summer when both air and car
temperatures would be maximized. Two studies, however,
investigated how variations in meteorological conditions
may affect cabin temperatures. McLaren et al. (2005)
quantified the heating within cars under days with a range
of ambient air temperatures. They collected data for 16
clear sunny days with air temperatures ranging from 22°C–
36°C. The maximum cabin temperature of the vehicle,
which depended on starting ambient air temperatures,
increased by an average of 22°C over ambient air temper-
atures within an hour. Even on the coolest day in the study,
with an ambient temperature of 22°C, the cabin temperature
reached a hazardous value of 47°C. Marty et al. (2001)
examined car temperatures under a variety of meteorolog-
ical conditions and in different seasons during the 1995–
2000 period. Exposure to direct sunlight lead to dramatic
increases in car temperatures but the degree of heating
varied by season. They offer an approximation that under
direct sun exposure maximum cabin temperatures may
reach 30°C in the winter, 60°C in the spring and fall, and
90°C in the summer. Their sample size of days with
inclement weather such as with rain were insufficient to
make any generalizations about cabin temperatures. The
authors also attempted to develop a model of cabin
temperatures based on the meteorological conditions but
were limited by the temporal resolution of their dataset.
Many United States state governments have recognized
the danger to children posed by high cabin temperatures.
Presently, 14 states have laws that prohibit leaving children
unattended in motor vehicles (Null 2008). However, even
states such as California, Florida, and Texas with such laws
still experience vehicle-related hyperthermia deaths, indi-
cating that the legislation does not provide a complete
deterrent. As mentioned above, many parents and care-
givers are simply not aware of how quickly vehicle
temperatures can reach hazardous levels. Thus, raising
awareness by documenting the degree of heating may be
useful in reducing deaths. The objective of this research,
then, is to build upon the work of McLaren et al. (2005)
and Marty et al. (2001) by examining maximum vehicle
cabin temperatures under a variety of meteorological
conditions during the spring and summer in Athens,
Georgia.
High temporal resolution meteorological and vehicle
cabin temperature datasets are used in this study and will
Int J Biometeorol
allow for a quantification of the influence of different
meteorological conditions upon cabin temperatures. In
addition, maximum cabin temperature models based on
commonly available meteorological data will be developed.
It is hoped that the models will be of use in forecasting
dangerous conditions to promote public awareness. In
addition, Marty et al. (2001) noted that there is a gap in
the forensic literature on identifying time of death under
nonstandard conditions such as bodies retrieved from
automobiles. Thus, the maximum cabin temperature model
may be useful for reconstructing past car temperatures for
forensic analyses.
Data and methods
The study was conducted from 1 April through 31 August
2007 in Athens, GA. A metallic gray 2005 Honda civic
with gray cloth seats was used. It was parked in an open lot
with direct exposure to sunlight. The windows were closed
during data collection because the objective was to assess
maximum possible temperatures. In addition, the lack of
ventilation represents typical conditions when children are
left in cars (Roberts and Roberts 1976).
The car temperature data were collected by HOBO
temperature sensors (Onset Computer Corporation; http://
www.onsetcomp.com/) that recorded temperatures every 5
min. The sensor was suspended approximately 150 mm
from the ceiling in a manner similar to that used by King et
al. (1981) and was never exposed to direct sunlight.
Ambient air temperature and solar radiation data at 5-min
resolution were obtained from an adjacent weather station
operated by the Department of Geography Climatology
Research Laboratory, located on the roof of the building
approximately 125 m from the parking lot. Cloud cover data
at hourly resolution were obtained from a National Weather
Service (NWS) Automated Surface Observing Station
(ASOS) located approximately 5 km away at Athens Ben
Epps Airport. The cloud cover data were presented in text
format as clear, few (>0/8–2/8), scattered (3/8–4/8), broken
(5/8–7/8), and overcast (8/8). Each category consists of a
range in coverage and is provided, if available, at different
levels in the atmosphere. For each hour, the average
coverage for the range was used and the level with the
greatest cover was set as the cloud cover value.
The study built a daily dataset consisting of maximum
cabin temperature, maximum ambient air temperature, and
average daytime [9:00 A.M.–5:00 P.M. eastern daylight time
(EDT)] solar radiation and cloud cover percentage. Only
days where the car was parked in the lot from at least noon
to 3:00 P.M. EDT and in the same parking spot were
retained. A total of 58 days were included in the study.
With this dataset, maximum cabin temperatures were
Int J Biometeorol

80
examined under a variety of cloud cover conditions and by
season. The seasons were defined as spring (April and
May) and summer (June, July, and August). Only the warm 70
season was studied because 75% of hyperthermia deaths
related to cars occur in the summer (Guard and Gallagher

Temperature (oC)
2005). The cloud cover conditions were categorized as clear 60

if cloudiness was ≤ 30% and cloudy if greater than 30%.

The partly cloudy (0.4–0.7) and cloudy (0.7–1.0) categories
in Changnon (1981) were grouped together in this study to
provide a sufficient dataset.
Maximum cabin temperature models were developed using
meteorological data. The first model uses maximum ambient
air temperature and average daytime solar radiation (9:00 A.M.–
5:00 P.M. EDT) as independent variables. These variables
were selected because previous studies have documented the
importance of ambient temperature (McLaren et al. 2005) and
solar radiation (Roberts and Roberts 1976; Surpure 1982) on
maximum cabin temperatures. The temporal resolution of the
data (i.e. maximum daily or average daily) were selected to
make the input data as accessible as possible for application.
In addition, solar radiation data are often not available.
Therefore, a second model was developed that uses average
daily cloud cover percentage (9:00 A.M.–5:00 P.M. EDT) as a
surrogate for solar radiation.
Results
Meteorological conditions during the study
The 58 study days include a wide variety of meteorological
conditions (Table 1). The study period extends from a few
weeks after the Spring Equinox and past the Summer
Solstice, which provide for a wide variety of solar angles and
thus solar radiation values. Average daily solar radiation
ranged from just over 250 W m−2 to 800 W m−2. Cloud
cover conditions varied from clear to overcast, with about
38% of the days classified as cloudy (cloud cover > 30%).
Finally, there was a wide variety in maximum ambient air
temperatures, ranging from mild (21°C) to extremely hot
(41°C).
Maximum cabin temperatures
The maximum cabin temperature data were grouped by
season and cloudiness to illustrate the range in values
Table 1 Summary statistics on
meteorological conditions Cabin temperature
during the study (°C)
50
40
N=13 N=5 N=29 N=11
30
spring-clr spring-cld summer-clr summer-cld
Fig. 1 Maximum cabin temperatures by season and cloudiness. The
box shows the lower and upper quartiles with the median marked by
the dark line. The whiskers mark the 5th and 95th percentiles with
outliers plotted as points. Clr Clear (cloudiness < 30%), cld cloudy
(Fig. 1). Cabin temperatures were highest during clear days,
with the greatest values in the summer. Clear days during
the spring averaged 61°C compared to 68°C in the summer.
Cloudy days had lower cabin temperatures, averaging 50°C
in spring and 58°C in summer.
To identify the particular conditions favorable to the
highest cabin temperatures as well as to quantify the
differences, anomalous values (±1 standard deviation) were
examined (Table 2). The difference in average maximum
cabin temperatures between the lowest and highest days
vary from 48°C to 74°C. It should be noted that even the
lowest cabin temperatures in the dataset are still sufficiently
high to pose a significant danger. The high temperature
days had greater ambient air temperatures by 10°C, average
solar radiation values over 200 W m−2 greater, and very
little cloud cover.
Maximum vehicle cabin temperature model
Two regression models of maximum cabin temperature
were developed from the 58 days of data. The first model
uses maximum daily ambient air temperatures and average
solar radiation as input variables. The high temporal
resolution solar radiation data allowed for the computation
of average daily solar radiation. The second model was
constructed using maximum daily ambient air temperatures
Ambient temperature Solar radiation Clear/cloudy
(°C) (W m−2) days

Mean 64 32 621 42/16

Max 76 41 802
Min 41 21 256
 Int J Biometeorol

Table 2 Average cabin temperatures and meteorological values for factors. First, cloud cover in this study is a surrogate for solar
days with cabin temperatures that are ±1 standard deviation from the
radiation and is therefore a less direct way of inferring solar
mean. Values rounded to nearest whole numbers
heating. Second, the observing station where the cloud data
+1 SD −1 SD were collected is several kilometers from the study site.
Ambient temperature (°C) 37 27 Thus, it is possible to have some inconsistencies, particularly
Cabin temperature (°C) 74 48 on partly cloud days where it may be clear (cloudy) at the
Solar radiation (W m−2) 664 437 observing station but cloudy (clear) over the study site. This
Cloud cover (%) 13 46 is exactly what happened on 20 July, where the model
N (days) 6 10 provided for a large overestimate of the maximum cabin
temperature. The adjacent weather station recorded a drop in
solar radiation during the afternoon indicating cloud cover but
the NWS cloud data indicated clearer conditions, leading the
and cloud cover as independent variables. Cloud cover data model to overestimate the cabin temperature. If this point is
are often more commonly available and provide a surrogate removed, the model performance improves considerably with
for solar radiation. the explained variance increasing to 66%. Finally, the use of
The results show that the model using ambient air average daily cloud cover, which is insensitive to the timing of
temperature and solar radiation does an excellent job the cloud cover, affected model performance. The largest
simulating maximum daily car temperatures (Fig. 2a). The residuals, particularly overestimates, tended to occur on partly
equation is:
a
Cabin temperature ¼ 0:036ðsolarÞþ1:02ðambientÞþ8:8 80

ð1Þ
70
Modeled Temperatures ( oC)

where “solar” refers to average daily solar radiation

(W m−2) and “ambient” to maximum ambient air temper-
atures (°C). It explains 80% of the variability in the data 60

and has an root mean squared error (RMSE) of 3.7. Some

of the larger residuals, particularly overestimates at the 50
lower observed temperatures, are the result of using average
daily solar radiation values. On some of the days, there are
40
clear conditions for much of the day but cloud cover during
midday. As a result, the average solar radiation may be high
(leading to higher cabin temperature estimates) but since 30
30 40 50 60 70 80
the car is exposed to less solar radiation during peak
Observed Temperatures (oC)
heating times, it may have a lower actual interior
temperature. The opposite situation occurred when there b 80
were clouds most of the day but clear conditions during
midday. Overall, however, the model is capable of
70
Modeled Temperatures ( oC)

capturing the variability among days with a wide range in

meteorological conditions.
Not unexpectedly, the second model using ambient 60

temperatures and cloud cover, simulates the observed

maximum cabin temperatures less successfully than the 50
one using solar radiation but still is capable of capturing
much of the magnitude and variability of maximum car
40
temperatures (Fig. 2b). The model is as follows:

Cabin temperature ¼
0:141ðcloudÞþ0:932ðambientÞþ37:1 30

30 40 50 60 70 80
ð2Þ
Observed Temperatures (oC)
where “cloud” refers to average daily cloud cover percentage
Fig. 2 Comparison of observed and modeled maximum cabin
and “ambient” to maximum ambient air temperatures (°C). It temperatures for models using a maximum ambient air temperature
explains 60% of the variance and has a larger RMSE of 5.2. and average solar, and b maximum ambient air temperature and cloud
The lower model performance may be explained by three cover. The solid line is the one-to-one line
Int J Biometeorol

Table 3 Maximum vehicle cabin temperature index using solar radiation and maximum ambient air temperature. Light shading Interior vehicle
heat advisory, dark shading excessive vehicle interior heat warning

40 57 59 60 62 64 66 68 69 71 73 75 77 78
38 55 57 58 60 62 64 66 67 69 71 73 75 76
Maximum Ambient Air

36 53 55 56 58 60 62 64 65 67 69 71 73 74
Temperature (ºC)

34 51 52 54 56 58 60 61 63 65 67 69 70 72
32 49 50 52 54 56 58 59 61 63 65 67 68 70
30 47 48 50 52 54 56 57 59 61 63 65 66 68
28 45 46 48 50 52 54 55 57 59 61 63 64 66
26 43 44 46 48 50 52 53 55 57 59 61 62 64
24 40 42 44 46 48 49 51 53 55 57 58 60 62
22 38 40 42 44 46 47 49 51 53 55 56 58 60
20 36 38 40 42 44 45 47 49 51 53 54 56 58
200 250 300 350 400 450 500 550 600 650 700 750 800
-2
Average Daily Solar Radiation (W m )

cloud days. Cloud cover plays a more significant role in
affecting cabin temperatures when it occurs during midday.
However, in some cases there was heavy cloud cover for
much of the day but less during midday. This would lead to
higher cloud cover averages and underestimates of cabin
temperatures. Clear conditions during most of the day but
clouds during midday would lead to the reverse situation.
Maximum vehicle cabin temperature index
Maximum vehicle cabin temperature indices were devel-
oped to aid in advising the public about the dangers of high
cabin temperatures (Tables 3, 4). The indices were derived
from the solar and cloud-based models respectively and are
designed to be similar to the well-known heat index, where
an apparent temperature is computed from the ambient air
temperature and relative humidity (Steadman 1979). To be
consistent with NWS terminology in their heat/health
warning systems, categories that are directly comparable
to “heat advisory” and “excessive heat warning” are used
(NWS 2005). The NWS issues a “heat advisory” when the
heat index is 41°C–46°C degrees for less than 3 hours. An
“excessive heat warning” is issued when the heat index is
≥41°C for more than 3 hours or exceeds 46°C for any
period of time. This study will define a “vehicle interior heat
advisory” when the maximum cabin temperature is 41°C–
46°C and “excessive vehicle interior heat warning” when the
cabin temperature exceeds 46°C.
The range of input values (e.g., maximum ambient air
temperature, solar radiation, and cloud cover) for computing

Table 4 Maximum vehicle cabin temperature index using cloud cover and maximum ambient air temperature. Light shading Interior vehicle heat
advisory, dark shading excessive vehicle interior heat warning. Ovc is overcast

40 74 73 71 69 67 66 64 62 60
38 73 71 69 67 65 64 62 60 58
Maximum Ambient Air

36 61 60 58 56 54 53 51 49 47
Temperature (ºC)

34 69 67 65 64 62 60 58 56 55
32 67 65 63 62 60 58 56 55 53
30 65 63 62 60 58 56 54 53 51
28 63 61 60 58 56 54 53 51 49
26 61 60 58 56 54 53 51 49 47
24 59 58 56 54 52 51 49 47 45
22 58 56 54 52 51 49 47 45 44
20 56 54 52 50 49 47 45 43 42
0 1/8 2/8 3/8 4/8 5/8 6/8 7/8 8/8

clear few scattered broken ovc

Average Daily Cloud Cover (octaves)

maximum cabin temperatures is based on the original data so
that the index values involve no extrapolation. As mentioned
earlier, the models are designed to maximize cabin temper-
atures. Of course, the actual maximum temperatures may
vary depending on the shading and the level of ventilation. In
addition, the hazard depends on the length of exposure to the
high temperatures, the humidity level which may affect the
amount of evaporative cooling from perspiration, and the age
and health of the child. Children are particularly sensitive to
heat illness because of their less effective thermoregulatory
systems and their inability to modify behavior such as
removing excess clothing or exiting the vehicle (McLaren et
al. 2005; Tsuzuki-Hayakawa and Tochihara 1995). Tables 3
and 4 show that, under almost any environmental condition
during the spring and summer, maximum vehicle cabin
temperatures reach the heat advisory or excessive heat
warning categories. These tables make clear that under no
circumstances should a child ever be left unattended in a
vehicle.
Discussion and conclusions
This is the first study to quantify the influence of different
ambient air temperatures and solar radiation values on
vehicle cabin temperatures. It uses a unique maximum
cabin temperature dataset that is linked with meteorological
data on ambient air temperatures, solar radiation, and cloud
cover. The study period included a range of ambient air
temperatures along with a variety of cloud cover and solar
radiation conditions.
The first portion of this study investigates maximum
cabin temperatures. These temperatures ranged from 41°C
to 76°C. The magnitude of maximum cabin temperatures
varied by both season and cloud coverage, which influ-
enced the amount of solar radiation. Clear days in the
spring average 61°C versus 68°C in the summer. The
higher values are related to the higher ambient air temper-
atures in the summer. This is consistent with McLaren et al.
(2005), who found that the maximum cabin temperature
was dependent on the initial ambient air temperature.
Maximum cabin temperatures on cloudy days averaged
approximately 10°C cooler than clear days with temper-
atures of 50°C in the spring and 58°C in the summer.
Cloudy days tended to have a wider range in maximum car
temperatures than clear days owing to the variability in
cover. The upshot of this research is to reinforce the
findings of McLaren et al. (2005) and demonstrate that
cabin temperatures can reach hazardous levels even under
relatively mild conditions (i.e., cloudy with mild ambient
air temperatures).
Using the cabin temperature and meteorological datasets,
two models of maximum car temperatures were developed.
The models were designed to maximize the potential
Int J Biometeorol
temperature and thus assume no ventilation and maximum
sun exposure. They were also designed to use average daily
meteorological values that may be more commonly
available. When solar radiation data are available, the
solar-based model (Eq. 1) is recommended because of its
superior performance. These models may be of use in
forecasting hazardous conditions and helping to inform the
public of the dangers of leaving children and infants
unattended in vehicles. For instance, weather forecasts
typically provide daily maximum temperatures and cloud
cover, which could be used to forecast cabin temperatures.
These forecasts may be used in conjunction with the
maximum vehicle cabin temperature indices that assign
danger levels—“vehicle interior heat advisory” and “exces-
sive vehicle interior heat warning”—that are consistent with
official NWS terminology. In addition, the models may be
used in forensic analyses to reconstruct cabin temperatures.
Guard and Gallagher (2005) found that over 25% of
vehicle-related hyperthermia deaths occurred when parents
or other caregivers intentionally left a child unattended in a
vehicle, indicating a clear lack of knowledge about the
heat-related hazard. Public advisories have been docu-
mented to raise awareness. Sheridan (2007), for instance,
noted that 90% of survey respondents from several urban
areas were aware of heat warnings. However, many of the
respondents did not actually modify their behavior because
they did not perceive a threat to their health. A warning of
the danger of high cabin temperatures may be successful at
both raising awareness and modifying behavior if it clearly
communicates the vulnerability of children and the distinct
health outcomes, including serious heat-related illnesses or
death.
References
Changnon S (1981) Midwestern cloud, sunshine and temperature
trends since 1901: possible evidence of jet contrail effects. J Appl
Meteorol 20:496–508
Gibbs LI, Lawrence DW, Kohn M (1995) Heat exposure in an
enclosed automobile. J La State Med Soc 147(12):545–546
Guard A, Gallagher SS (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
King K, Negus K, Vance JC (1981) Heat stress in motor vehicles: a
problem in infancy. Pediatrics 68(4):579–582
Marty WT, Sigrist T, Wyler D (2001) Temperature variations in
automobiles in various weather conditions: an experimental
contribution to the determination of time of death. Am J Forensic
Med Pathol 22(3):215–219
McLaren C, Null J, Quinn J (2005) Heat stress from enclosed
vehicles: moderate ambient temperatures causes significant
temperature rise in enclosed vehicles. Pediatrics 116:e109–
e112
National Weather Service (NWS, 2005) National Weather Service
Glossary. Available at: http://www.weather.gov/glossary
Int J Biometeorol
Null J (2008) Hyperthermia deaths of children in vehicles. Golden
Gate Weather Services; 2008. Available at: http://ggweather.com/
heat
Roberts KB, Roberts EC (1976) The automobile and heat stress.
Pediatrics 58(1):101–104
Sheridan (2007) A survey of public perception and response to heat
warnings across four North American cities: an evaluation of
municipal effectiveness. Int J Biometeorol 52(1):3–15
View publication stats
Steadman RG (1979) The assessment of sultriness. Part I: A
temperature-humidity index based on human physiology and
clothing science. J Appl Meteorol 18:861–873
Surpure JS (1982) Heat-related illness and the automobile. Ann Emerg
Med 11(5):263–265
Tsuzuki-Hayakawa K, Tochihara K (1995) Thermoregulation during
heat exposure of young children compared to their mothers. Eur J
Appl Physiol 72:12–17