Available via license: CC BY 4.0
Content may be subject to copyright.
atmosphere
Article
Microclimate Variation and Estimated Heat Stress
of Runners in the 2020 Tokyo Olympic Marathon
Eichi Kosaka 1, *, Akiko Iida 1ID , Jennifer Vanos 2, Ariane Middel 3, Makoto Yokohari 1
and Robert Brown 4
1Department of Urban Engineering, School of Engineering, University of Tokyo, Tokyo 113-8656, Japan;
iida@epd.t.u-tokyo.ac.jp (A.I.); myoko@k.u-tokyo.ac.jp (M.Y.)
2Climate, Atmospheric Science & Physical Oceanography, Scripps Institution of Oceanography,
University of California at San Diego, La Jolla, CA 92093, USA; jkvanos@ucsd.edu
3Department of Geography and Urban Studies, Temple University, Philadelphia, PA 19122, USA;
ariane.middel@temple.edu
4
Department of Landscape Architecture and Urban Planning, College of Architecture, Texas A&M University,
College Station, TX 77843, USA; rbrown@arch.tamu.edu
*Correspondence: ek.at.uttf@gmail.com; Tel.: +81-90-6106-0621
Received: 31 March 2018; Accepted: 10 May 2018; Published: 17 May 2018
Abstract:
The Tokyo 2020 Olympic Games will be held in July and August. As these are the hottest
months in Tokyo, the risk of heat stress to athletes and spectators in outdoor sporting events is a
serious concern. This study focuses on the marathon races, which are held outside for a prolonged
time, and evaluates the potential heat stress of marathon runners using the COMFA (COMfort
FormulA) Human Heat Balance (HBB) Model. The study applies a four-step procedure: (a) measure
the thermal environment along the marathon course; (b) estimate heat stress on runners by applying
COMFA; (c) identify locations where runners may be exposed to extreme heat stress; and (d) discuss
measures to mitigate the heat stress on runners. On clear sunny days, the entire course is rated as
‘dangerous’ or ‘extremely dangerous’, and within the latter half of the course, there is a 10-km portion
where values continuously exceed the extremely dangerous level. Findings illustrate which stretches
have the highest need for mitigation measures, such as starting the race one hour earlier, allowing
runners to run in the shade of buildings or making use of urban greenery including expanding the
tree canopy.
Keywords:
urban microclimate; heat stress; COMFA Human Heat Balance Model; Tokyo 2020
Olympic Games; marathon games
1. Introduction
Heat stress from summer heat is expected to be a significant issue at the 2020 Tokyo Olympics [
1
].
In recent years, temperatures have risen in the world’s major cities due to both climate change and
the urban heat island effect, as caused by the increased longwave radiation from road and building
surfaces [
2
]. These factors have led to an intensification of health hazards posed by heat stress [
3
,
4
].
In addition, as Tokyo is located in a temperate zone and has a humid climate, its summers are hot and
humid; the Tokyo Olympics are scheduled to be held in July and August in central Tokyo, where the
average maximum daily temperature reaches above 30 ◦C and the average relative humidity reaches
over 70%. Consequently, the summer heat and humidity are expected to exert tremendous heat stress
on the human body [
5
–
7
], and it is estimated to have the poorest conditions among any previous host
cities [8,9].
In the outdoor sporting events, heat is expected to impact athletic ability. Of the Olympic events
that are held outdoors, marathons require participants to engage in 2–3 h of physical activity in summer
Atmosphere 2018,9, 192; doi:10.3390/atmos9050192 www.mdpi.com/journal/atmosphere
Atmosphere 2018,9, 192 2 of 16
heat; hence, there is a greater chance for negative impacts to the athletes, particularly if less prepared
and/or acclimatized. When holding marathons within dense and heterogonous cities, it is necessary
to account for the microclimate conditions occurring in urban areas [
10
–
12
]. Here, we focus on the
conditions on the marathon course and formulate strategies to combat heat stress, such as decreasing
solar and longwave radiation by shading with buildings, trees and shrubs [13].
Previous studies investigated the response of marathon runners under serious thermal conditions.
Helou et al. [
14
] studied six marathons in Europe and the United States and found a negative correlation
with runners’ performance and air temperature, humidity and air pollution. Solar radiation also causes
a progressive drop in runners’ comfort perception [15], endurance and time to exhaustion [16].
In a study targeting the marathon competition of the Tokyo Olympic Games,
Yamazaki et al.
[
17
]
assessed the thermal environment on the marathon course using stationary data from the Japan
Meteorological Agency. Kashimura et al. [
18
] measured the thermal environment on the marathon
course over three days with a Wet-Bulb Globe Temperature (WBGT) index meter. However,
these studies only investigated thermal conditions and did not assess the potential negative effects of
high metabolic activities. Because marathon runners continuously perform intense exercise, the high
metabolism increases the heat loads on the human body. Therefore, it is necessary to consider these
factors to estimate the risks of heat illness for marathon runners.
This study quantifies the thermal comfort and potential heat stress of marathon runners during
the 2020 Tokyo Olympic marathon, including human metabolic rate, using the COMFA (COMfort
FormulA) human heat budget model [
19
–
21
], and proposes strategies for improving microclimate
conditions on the marathon course to mitigate the heat load and hence the potential heat stress
to runners.
2. Methods
2.1. Study Area
This study focuses on the scheduled marathon course for the 2020 Tokyo Olympics (see Figure 1).
Runners will start the marathon at the New National Stadium, then pass the Imperial Palace Square,
Tokyo Tower, Ginza district and other landmarks in Central Tokyo. They will then reach a turning
point at Asakusa, the heart of the old downtown district, where they will double back along the
racecourse, finishing the race at the New National Stadium.
Figure 1. Proposed marathon course for the 2020 Tokyo Olympics.
Atmosphere 2018,9, 192 3 of 16
2.2. Research Methodology
2.2.1. Thermal Environment Measurements along the Marathon Course
Cars and bicycles equipped with weather observation instruments made return trips along the
planned marathon course to measure heat conditions along the marathon route. Measurements
were conducted in the left-hand traffic lane because runners must run in the left-hand traffic lane
in principle, according to the Japan Association of Athletics Federations. This study measured five
meteorological elements: air temperature, humidity, solar radiation, road surface temperature and
wind speed using the sensors shown in Figure 2. Table 1provides a list of the weather observation
instruments used in this study. POTEKA is a compact weather observation instrument developed
by Meisei Electric Co., Ltd. (Isesaki, Japan) that measures temperature, humidity, solar radiation
and wind speed. Surface temperature data were collected by infrared camera, Thermo GEAR G100
(
Nippon Avionics Co.
, Ltd., Shinagawa, Japan). These meteorological elements were linked to ground
position using an eTrex 30 GPS (Garmin Ltd., Olathe, USA).
Figure 2. Vehicle and bicycle equipped with weather observation instruments.
Table 1.
Summary of mobile meteorological data collection. Meteorological instruments and the
accuracy of each use for the mobile data collection. All were measured from an ~1.5-m height apart
from the surface temperature.
Instrument Parameter Model Accuracy Interval
Thermometer Air Temperature (◦C) POTEKA ±0.3 ◦C 1 s
Hygrometer Relative Humidity (%) POTEKA ±5% 1 s
Pyrometer Solar Radiation (W/m2)POTEKA ±10% 1 s
Surface thermometer Surface Temperature (◦C) ThermoGEAR G100 ±2◦C 3 s
Anemometer Wind + Activity Speed (m/s) POTEKA ±1.0 m/s 1 s
GPS Latitude and Longitude eTrex30 3 to 6 m 1 s
As shown by d’Ambrosio Alfano et al. (2007), d’Ambrosio Alfano et al. (2013) and Palella et al.
(2016), the use of instruments with higher quality provides more accurate and precise results [22–24];
however, among various requirements for the instruments for the given study, the compactness was
crucial for the mobile monitoring setups and because the measurements had to be conducted along
roads with busy traffic in the 2–3 h morning marathon window. Larger instruments with higher
accuracy were not compact enough to be placed on bicycles. The POTEKA, whose accuracy is certified
by the Meteorological Agency of Japan, is the best instrument to balance the conflicting requirements
of compactness and accuracy.
As shown in Table 2, the measurements were taken for nine days in the summer of 2016, starting
at the marathon’s scheduled starting time of 7:30 a.m., and for two days starting at 6:30 a.m., one hour
earlier than the scheduled start time, for a total of 11 days of measurements. The measurements
were taken under a variety of weather conditions: seven clear sunny and four overcast cloudy days.
The hottest day was August 9, when the temperature reached a high of 37.9
◦
C. Furthermore, August 9,
Atmosphere 2018,9, 192 4 of 16
2016 was the hottest day of the year and is the day of the men’s marathon during the 2020 Tokyo
Olympics. The average high on sunny days was 33.4
◦
C and on cloudy days was 29.2
◦
C. According to
observational data from the Meteorological Agency, the average temperature, humidity and sunshine
hours in 2016 matched the climate norms.
Table 2. List of measurement days, time of measurement and sky conditions.
Date Time Sky Condition
Jul 29, 2016 7:30–10:29 sunny
Aug 5, 2016 7:30–10:30 sunny
Aug 6, 2016 7:30–10:31 sunny
Aug 9, 2016 7:30–10:33 sunny
Aug 11, 2016 7:30–9:57 cloudy
Aug 12, 2016 7:30–10:15 cloudy
Aug 13, 2016 7:30–10:16 sunny
Aug 14, 2016 7:30–10:10 cloudy
Aug 23, 2016 6:30–9:20 cloudy
Aug 25, 2016 6:30–9:25 sunny
Aug 26, 2016 7:30–10:25 sunny
Additionally, to ascertain the degree of heat stress risk during a marathon held in the summer in
Tokyo, a “worst-case scenario” (representing the potentially warmest environment) and “best-case
scenario” (representing the potentially coolest environment) were hypothetically created to supplement
the day-by-day data. To create each scenario, the highest and lowest values of each climate factor,
excluding solar radiation, were selected from the nine measurement days and synthesized as the data
for each scenario. When synthesizing the data, the absolute humidity was calculated with the relative
humidity and temperature on the wettest/driest day, and the hypothetical relative humidity was then
re-calculated according to the temperature of the warmest/coolest day. Because solar radiation is
affected by cloudiness, the highest and lowest values in each section throughout the nine measurement
days were used as the data for each scenario.
2.2.2. Estimated Heat Balance of Runners: Application of the COMFA Human Heat Balance Model
Using the measured data, the heat load exerted on the runners from the environment along
the marathon route was calculated using the COMFA human heat balance model for each day and
scenario. The wind speed was first corrected using the two steps provided in Tokura et al. [
25
]. First,
the stationary wind speed was calculated by subtracting the moving speed of the vehicle or bicycle
from the field-measured speed. The moving speed was calculated by positional data of GPS. Second,
the wind speed felt by runners (or the relative wind velocity) was calculated by taking the square root
of the sum of the squares of the stationary wind speed and runners’ average race speed (20 km/h),
as in ISO (2007) [26].
To better visualize the distribution of the estimated heat stress along the route, the results for each
meteorological factor in the measurement data were divided into 100-m sections; the data recorded
in each section for each factor were then averaged and rendered as representative values used in the
analysis along the race route.
The COMFA model combines climate factors of temperature, humidity, wind and radiation,
with the personal factors of metabolic rate and clothing insulation to estimate thermal comfort and
potential heat stress [27–29]. The model defines the energy budget in W m−2as follows:
B=M+Rabs −C−E−Lemit (1)
where Bis the net energy budget, Mis metabolic heat production, R
abs
is the absorbed radiation by a
cylindrical human, Cis convective heat loss, Eis evaporative heat loss (sum of insensible and sensible
Atmosphere 2018,9, 192 5 of 16
heat losses [
29
]) and L
emit
is the emitted longwave radiation by a human. All fluxes in Equation (1) are
in W m−2.
The runners were estimated to have a low clothing insulation (I
cl
) value of 0.1 clo (~18.6 s m
−1
)
based on minimal, thin and low insulation running uniforms and a high amount of exposed skin.
We used the sum of the clo values for the individual garments (i.e.,
Icl =∑Iclu )
from ISO (2007) based
on panties/briefs (0.03 clo) and tank top (0.07 clo).
The COMFA model accounts for the activity of the athlete [
20
] when calculating thermal comfort
and heat stress; however, we are not able to account for inter-individual variability in the current study.
An average metabolism of ~700 W m
−2
was determined by allowing for energy to be converted to
mechanical work during running, where the net gain from metabolism, M, is 0.8M
total
+ 0.2M
rest
[
30
],
where M
rest
is 1 MET (58 W m
−2
) and M
total
is based on running at an average speed of 18 km h
−1
for an elite runner, or 14.5 METs [
31
]. The influence of an increased relative wind velocity of the
runner to promote cooling via decreased resistance to vapor and heat transfer from higher wind is also
determined by the COMFA model (see [21,28,32]).
The absorbed radiation was determined from the combined influence of short- and long-wave
radiation experienced by a human, based on an effective area of 0.78 for a standing human and average
surface (Acyl), using the following equation.
Rabs =Ae f f Kin(abs)+Ku p(abs)+La(abs)+Lg(abs)
Acyl !(2)
where K
in(abs)
and K
up(abs)
are the absorbed solar radiation from the sky and ground, respectively, while
L
a(abs)
and L
g(abs)
are the incoming and outgoing longwave radiation, respectively. The full detailed
methods are found in Hardin and Vanos (2017) and Kenny (2008) [
33
,
34
]. Below, we outline any
changes made from studies [
21
,
28
,
32
] for the given study. The outgoing longwave (L
out
) radiation was
determined from the measured ground temperature and the use of the Stephan–Boltzmann equation.
Incoming longwave radiation (Lin) was determined based on the summation of infrared energy from
the sky or from the buildings. To determine the portion of the sky visible in the upper 180
◦
hemisphere
(Sky View Factor (SVF)) in each location, we acquired images via Google Street View and used an
SVF calculation methodology by Middel et al. [
35
–
37
]. Ninety-degree field of view images were
retrieved in four cardinal directions and upwards for all Street View locations that are part of the race
course [
36
]. The images were then segmented into sky versus non-sky pixels using a convolutional
neural network [
37
] and converted into hemispherical fisheye views using equiangular projection.
Finally, the SVF value (0–1) was calculated using the modified Steyn method [35].
The absorbed L
in
(L
in(abs)
) was determined based on the SVF weighting of the longwave emitted
from objects in the sky hemisphere (Lo) and longwave from the sky (La), as follows:
La=213 +5.5Ta·SVF (3)
Lo=εhσ(To+273.15)(1−SVF)(4)
Lin(abs)=0.5εh(Lo+La)(5)
Equation (2) is based on Monteith and Unsworth (2008) [
38
], and the T
o
is the temperature of
objects in the sky hemisphere, estimated from vertical surface temperature data using an infrared
camera (Nippon Avionics, ThermoGear G100, Shinagawa, Japan).
Finally, as per Harlan [
39
], the calculated human energy budgets were classified into four stages:
>340 W m
−2
are extremely dangerous; 200–340 W m
−2
are dangerous; 120–200 W m
−2
are caution;
and <120 W m−2are safe.
Atmosphere 2018,9, 192 6 of 16
Convective heat losses were calculated based on a heat flux rate, determined based on the
temperature gradient between the skin and air, divided by the resistances (s m
−1
) to the flow of energy
from one surface to the other:
C=ρCpTsk −Ta
rc+ra(6)
where
ρ
C
p
is the volumetric heat capacity of air, T
sk
is the skin temperature [
21
,
40
], r
c
is the clothing
resistance (s m−1) and rais the boundary air resistance (s m−1) [28].
The evaporative heat loss (E, W m−2) was determined as follows:
E=wρLvqsk −qa
rcv +rav (7)
where q
sk
and q
a
are the specific humidity (kg of water vapor per kg of moist air) at T
sk
and T
a
,
ρ
is the
density of air (kg m
−3
), L
v
is the latent heat of vaporization (J g
−1
), r
cv
is the clothing vapor resistance
(s m
−1
) and r
av
is the aerodynamic vapor resistance (s m
−1
). Vapor resistance calculations can be
found in [20,27]. The skin wettedness (w) is calculated as [41]:
w=0.42(M−58)
Emax
+0.06 (8)
where E
max
is the maximum amount of evaporative heat loss from the skin (controlled by environmental
parameters) (W m−2), calculated as in Equation (7), while holding wat saturation.
2.2.3. Identifying Hot Locations along the Marathon Course
The calculated human energy budget was mapped using ArcGIS 10.2, graphed and visualized to
identify locations on the marathon route where runners may be subjected to heat stress. This study
primarily examined the results of four cases: August 9, the hottest day of the year in 2016 and the day
the men’s marathon will be held during the 2020 Tokyo Olympics, the worst-case scenario, the best-case
scenario and August 25, a clear, sunny day when measurements were taken one hour early. In addition,
the calculated results for each day and each of the two scenarios were compared to estimate the
expected degree of heat stress risk during the Tokyo marathons in 2020 and identify locations with
high heat loads that lead to poor thermal comfort and heat stress.
The model is not a heat strain model (which refers to the effects on body physiology, such as
core temperature or sweat rate leading to heat stroke) that occur due to overall heat loads. Rather,
the model focuses on conditions that increase body temperature, including heat loads from physical
exertion, ambient temperature and environmental factors (i.e., leading to heat stress) [
42
]. The COMFA
model is a rational model, which is more complex than direct models (such as the Wet-Bulb Globe
Temperature) because rational models allow one to change clothing, metabolism, height, weight, etc.,
internally within the model, which is not possible with direct approaches like the WBGT, humidex or
the heat index [43].
2.2.4. Approaches to Mitigate Heat Stress
Based on the locations rated as dangerous for each day and scenario, this study investigated the
feasibility of possible methods to mitigate heat stress on the runners, such as by changing the route to
increase shade and expanding the tree canopy along the marathon route, greening and the installation
of sunshades.
3. Results
3.1. Mean and Standard Deviation of Meteorological Variables
Table 3shows the average and standard deviation of measurements for all dates and scenarios.
When the air temperature exceeds 30
◦
C, the value rises rapidly from the start of measurement, and the
Atmosphere 2018,9, 192 7 of 16
standard deviation is large on many sunny days. In contrast, on cloudy days, the temperature did not
reach 30
◦
C, and the standard deviation was small. Relative humidity was >60% on the majority of
days. However, the humidity was lower on August 9 (average of 27%), as the temperature was the
highest. The wind speed was nearly constant across all measurement dates. Regarding solar radiation,
the maximum value was approximately 1000 W m
−2
on all sunny days, yet the average value and
standard deviation varied. On days with clouds, there were greater variations in standard deviation,
maximum and average values of solar radiation. A positive association was found between solar
radiation and road surface temperature.
Table 3. Mean and standard deviation (SD) of meteorological variables for each day and scenario.
Day or Scenario Ta(◦C) RH (%) Wind Speed Solar Radiation Road Surface
Temperature
(m s−1) (W m−2)(◦C)
29 July Mean ±SD 29.3 ±1.26 58.9 ±4.59 2.0 ±0.79 427.5 ±317.76 35.3 ±5.64
Range 27.0–32.8 49.1–66.4 0.5–4.4 46.9–1024.1 27.5–49.5
5 August Mean ±SD 30.8 ±1.27 58.2 ±4.65 2.0 ±0.82 281.1 ±215.59 36.4 ±4.43
Range 28.7–33.9 48.1–65.6 0.5–5.7 51.7–1052.8 29.8–51.4
6 August Mean ±SD 30.6 ±1.08 62.2 ±3.35 1.9 ±0.69 349.8 ±238.76 37.7 ±4.31
Range 28.2–32.7 55.5–69.8 0.7–4.5 56.1–999.6 30.9–49.0
9 August Mean ±SD 34.3 ±1.82 37.0 ±5.87 2.0 ±0.88 350.8 ±249.13 38.6 ±5.06
Range 29.6–37.6 29.5–50.3 0.6–8.3 43.2–949.1 27.8–54.8
11 August Mean ±SD 28.2 ±0.45 46.2 ±1.86 1.9 ±1.19 125.7 ±95.83 33.4 ±1.66
Range 27.1–29.5 42.9–51.2 0.4–10.4 28.3–375.4 30.5–39.4
12 August Mean±SD 27.3 ±0.82 59.0 ±4.41 1.7 ±0.87 231.7 ±83.16 34.2 ±2.24
Range 25.9–29.2 49.7–64.6 0.4–10.3 94.4–450.0 29.9–40.6
13 August Mean ±SD 27.8 ±0.90 53.5 ±3.32 1.7 ±0.82 347.4 ±193.51 36.1 ±4.05
Range 26.3–30.0 48.3–59.9 0.3–6.0 60.3–1066.4 29.5–50.4
14 August Mean ±SD 26.3 ±0.85 54.8 ±3.19 1.7 ±0.87 253.4 ±119.17 34.3 ±3.04
Range 24.7–28.2 49.1–59.9 0.4–10.7 61.3–727.8 29.0–43.9
23 August Mean ±SD 27.8 ±0.96 69.0 ±3.79 1.9 ±0.78 101.4 ±70.31 27.7 ±1.61
Range 26.0–29.5 62.2–75.2 0.5–5.0 12.7–399.4 25.1–32.3
25 August Mean ±SD 29.0 ±1.07 61.0 ±4.88 2.0 ±0.67 203.0 ±190.22 31.0 ±4.15
Range 26.9–31.6 50.4–70.1 0.5–5.4 14.6–824.0 25.4–45.3
26 August Mean ±SD 30.1 ±1.09 57.3 ±3.55 2.0 ±0.73 311.9 ±227.17 35.5 ±5.00
Range 27.4–32.6 51.0–65.8 0.6–5.1 27.3–1015.9 27.7–50.4
Worst Reference
day August 9 August 5 August 5 Max of each
section August 9
Mean ±SD 34.3 ±1.82 47.4 ±5.55 2.0 ±0.82 514.4 ±276.26 38.6 ±5.06
Range 29.6–37.6 38.5–62.1 0.5–5.7 109.8–1066.4 27.8–54.8
Best Reference
day August 14 August 11 August 13 Min of each
section August 11
Mean ±SD 26.3 ±0.85 51.8 ±3.35 2.0 ±0.88 97.5 ±62.94 33.4 ±1.66
Range 24.7–28.2 45.6–59.6 0.6–8.3 27.3–341.4 30.5–39.4
Ta: air temperature, RH: relative humidity.
3.2. Maps and Graphs of Human Energy Budget
3.2.1. August 9, 2016
The calculated human energy budgets for August 9 are illustrated in Figures 3and 4. Given the
relatively high air temperatures and extreme relative humidity, conditions considered dangerous
occur immediately near the start of the race. One section, which includes a stretch of east-west
highway (15.8–17.2 km), runs toward the Sun and had high values of solar radiation and road surface
temperatures, so the energy budget exceeded the benchmark for the extremely dangerous level.
Values on the homeward route exhibited higher energy budget values as compared to the outward
route due to increases in temperature, solar radiation and road surface temperatures as the day
progressed. The same stretch of east-west highway on the return path (21.3–22.7 km) also showed an
Atmosphere 2018,9, 192 8 of 16
energy budget that exceeded the extremely dangerous threshold. From 22.8–27.4 km lies a stretch of
highway running north-south that was lined with high-rise buildings and is therefore shaded. Thus,
the intensity of the solar radiation and road surface temperatures were relatively low compared to
other locations, and the energy budget here was almost 100 W m
−2
lower than for other locations and
rated as dangerous. However, from the 27.5-km mark outward, many locations were illuminated by
direct solar radiation. Therefore, the road surface temperatures and solar radiation exhibited higher
values than in other locations, which resulted in an increase in accumulated temperature and intense
longwave radiation, and almost all locations were identified where values exceeded the extremely
dangerous level, all the way to the 35.8-km mark. For example, the energy budget in front of the
Imperial Palace Square (30.4–31.2 km) ranged from 342.8–489.5 Wm
−2
, because there were no tall trees
or buildings that could provide shade for runners.
Figure 3.
Energy budget for August 9, 2016 mapped on the marathon route (left: outward;
right: homeward).
Figure 4. Energy budget for August 9, 2016 graphed as a function of distance.
3.2.2. Worst-Case Scenario
Figures 5and 6display the results for the worst-case scenario. In comparison to August 9,
the worst-case scenario has higher humidity and lower wind speed. Therefore, the energy budget is
higher throughout the entire race, with stretches considered extremely dangerous beginning at the
start of the outward transect. The levels of solar radiation in the worst-case scenario show the most
intense solar conditions, however, still providing variability due to building and vegetation shading.
On August 9, according to data from the Japan Meteorological Agency, at the time of the data collection
at the 35.9-km mark, it was overcast, and the Sun was obscured. Therefore, on August 9, the energy
budget values after the 35.9-km mark were low. However, the worst-case scenario virtualized clear
sunny days, and therefore, the values did not decrease after the 35.9-km mark and exceeded the
extremely dangerous threshold from 27.5 to the goal. Based on these results, this section of the route is
considered as have a particular need for improvements of the thermal conditions.
Atmosphere 2018,9, 192 9 of 16
Figure 5.
Energy budget for the worst-case scenario mapped on the marathon route (left: outward;
right: homeward).
Figure 6. Energy budget for the worst-case scenario graphed as a function of distance.
3.2.3. Best-Case Scenario
Figures 7and 8show the results of the best scenario. In this scenario, it is assumed that the route
will be constantly overcast, so large changes along the route are not observed, in contrast to August 9
and the worst-case scenario; instead, the energy budget rises gradually with increasing temperature.
On the outward route, there is a series of locations evaluated as in the safe range. On the homeward
route, the ratings waver between safe and caution, and there are several dangerous spots in the latter
half. No spots on the course are evaluated as extremely dangerous.
Figure 7.
Energy budget for the best-case scenario mapped on the marathon route (left: outward;
right: homeward).
Atmosphere 2018,9, 192 10 of 16
Figure 8. Energy budget for the best-case scenario graphed as a function of distance.
3.2.4. August 25, 2016: Early Start
On August 25, measurements began at 6:30 a.m., an hour earlier than usual. The highest
temperature during the measurements was 31.5
◦
C. It was clear and sunny throughout the
measurement period. Therefore, in terms of solar radiation, August 25 was not subject to great
weather variability, such as temporary overcast conditions, and each location was considered as subject
to the individual maximum solar radiation given the time of day and the presence of vegetation
or buildings.
Figures 9and 10 show the results for August 25. There is a series of dangerous sections on the
outward route, but no spots on the outward course are evaluated as extremely dangerous. On the
homeward route, just as on August 9 and in the worst-case scenario, high energy budget values are
observed from the 27.5-km mark onwards. However, from 28.2–29.9 km and 33.4–36.4 km, the energy
budget was lower than those in the surrounding areas. Therefore, as expected, the length of the
extremely dangerous area on August 25 is shorter than that of the worst-case scenario. This is largely
due to the fact that measurements began an hour earlier than usual, and thus, the solar radiation was
at a lower intensity, as well as a significant amount of shade from buildings and trees lining the course
remained present.
Figure 9.
Energy budget for August 25, 2016 mapped on the marathon route (left: outward;
right: homeward).
Figure 10. Energy budget for August 25, 2016 graphed as a function of distance.
Atmosphere 2018,9, 192 11 of 16
3.3. Comparison of the Results for Each Case
Table 4compares the percentage of the overall number of locations rated as safe, caution,
dangerous, and extremely dangerous for August 9 and 25 and the worst and best scenarios. For August
9, 29.4% of the total route was rated as extremely dangerous and 70.3% of the total route as dangerous.
In the worst-case scenario, 54.9% of the route, almost half the length, was rated as extremely dangerous.
In contrast, in the best-case scenario, 70.8% of the route was rated as being in the safe range, with 28.6%
of the route rated as dangerous. On August 25, 44.8% was rated as caution, 50.5% as dangerous and
3.9% as extremely dangerous.
Table 4.
The percentages of the route rated as safe, caution, dangerous and extremely dangerous for
each day and scenario.
August 9, 2016 Worst Case BEST CASE August 25, 2016
Safe 0% 0% 70.8% 0.8%
Caution 0.3% 0% 28.6% 44.8%
Dangerous 70.3% 45.1% 0.6% 50.5%
Extreme Dangerous 29.4% 54.9% 0% 3.9%
Based on these findings, and particularly as demonstrated by the worst-case scenario, sunny,
high-temperature and high-humidity conditions produce dangerous conditions throughout the entire
route. Consequently, there are some variables that can be modified to improve the thermal environment;
nonetheless, heat stress prevention measures are required for the entire course. As with the best-case
scenario, even when the skies are overcast and temperatures are not very high, many stretches remain
rated as caution or dangerous. Therefore, we believe that a strategy for dealing with heat stress is
necessary regardless of the sky conditions, as the temperature and humidity are expected to be at
levels considered extreme, and wind speeds are generally low and difficult to alter.
Figure 11 summarizes the heat budget graphs for each measurement date and scenario along
the marathon route. Additionally, the course was divided into 25 sections according to the street
direction, which is shown on the top bar of Figure 11. Eight sections have a higher energy budget than
before and after the sections in the worst-case scenario and on August 9 and 25: (A) 15.8–17.2 km and
21.3–22.7 km, (B) 27.5–27.9 km, (C) 28–29.9 km, (D) 30.4–31.2 km, (E) 31.4–33.3 km, (F) 33.4–34.4 km,
(G) 34.5–36.7 km, and (H) 36.8–37.7 km.
Figure 11. Graphs comparing the heat budget for each day and scenario, including street direction.
Atmosphere 2018,9, 192 12 of 16
3.4. Characteristics of Sections Considered Dangerous
3.4.1. Individual Characteristics of Each Section
Section A (15.7–17.2 km, 21.3–22.7 km)
In these spots, the streets run east-west. Therefore, even though the streets are lined with buildings
and trees, they were not able to provide much shade, and the entire street will be in direct sunlight at
race time. Therefore, solar radiation and road surface temperatures will be high and the energy budget
relatively high in the surroundings.
Section B (27.5–27.9 km)
As is the case with Section A, the streets run east-west. Therefore, even though the streets are
lined with buildings and trees, the entire street will be in direct sunlight at race time.
Section C (28–29.9 km)
From 28–28.1 km, there are trees partially blocking the Sun, but the shade does not reach the left
lane where the route runs. After that section, there are buildings that partially block the Sun. Based on
data from the worst-case scenario and August 9, the Sun will be high in the sky at race time, and the
building shadow will not reach the left lane; hence, this area has a high energy budget. However, if the
race time is an hour earlier, e.g., as on August 25, the building shade reaches the left lane. Therefore,
it has a low energy budget along most of this stretch.
Section D (30.4–31.2 km)
The 30.4–31.1-km section passes through the square in front of the Imperial Palace. There are no
trees or buildings present in the surroundings to provide shade, so the solar radiation intensity and
road surface temperatures will be high at race time. The energy budget is also high compared to the
surroundings, on both the outward and homeward legs. From 31.1–31.2 km, there are trees partially
blocking the Sun, but the shade does not reach the left lane. Therefore, it has a high energy budget.
Section E (31.3–33.3 km)
In this section, the Sun is positioned in the southeast, and the streets are almost entirely unshaded
at race time, so solar radiation and road surface temperatures will be high.
Section F (33.4–34.4 km)
As is the case with Section A, the streets run east-west. Therefore, even though the streets are
lined with buildings and trees, the entire street is in direct sunlight at race time.
Section G (34.5–36.7 km)
In this section, trees are located to partially block the Sun. Data from the worst-case scenario
and August 9 and 25 indicate that the trees provide partial shade, so the energy budget will be a little
lower than in the surrounding areas. However, where tree growth is poor and tree heights are low,
the energy budget will be high.
Section H (36.8–37.7 km)
As is the case with Section A, the streets run east-west. Therefore, even though the streets are
lined with buildings and trees, the entire street will be in direct sunlight.
Atmosphere 2018,9, 192 13 of 16
3.4.2. Summary
From the analysis in Section 3.4.1, the energy budget is affected by the locations of buildings and
trees that block the Sun relative to the street direction.
Using these data, high energy budget spots on the marathon course were classified into five groups,
as shown in Table 5. Group I (A, B, F, H) has shade from trees and buildings, and the road direction
is W-E. Group II (C) has shade from trees and buildings, and the road direction is NE-SW. Group III
(D, G) has shade only from trees, and the road direction is NE-SW. Group IV (D) has neither trees nor
buildings that provide shade, and the road direction is NE-SW. Group V (E) has shade provided by
trees and buildings, and the road direction is NW-SE.
In the next section, mitigation approaches are discussed for each group.
Table 5.
Classification by the existence of buildings and street trees that block the Sun based on
street direction.
Both Trees and Buildings Only Trees Nothing
W-E A B F H (I)
NE-SW C (II) D G (III) D (IV)
NW-SE E (V)
4. Approaches to Heat Stress Mitigation
The Sun will be facing the flow of traffic; therefore, these sections will be difficult to shade. On the
homeward leg, a small amount of shade is present from trees near the road. Therefore, adjusting the
tree pruning and expanding the tree canopy so that it projects over the side of the road is a possible
measure to consider. It is also hypothesized that installing temporary sunshades to cover the sky above
the road could also be effective.
4.1. Group II
The left traffic lane, where the runners will be running, will be unshaded, although there will be
some shade in the right traffic lane. Therefore, in this section, the runners should also be allowed to
run in the right traffic lane, which will enable them to reduce heat stress.
4.2. Group III
The streets remain unshaded by buildings, but are partially shaded by trees; therefore, expanding
the tree canopy will provide more shade. Restricting pruning should expand the canopy and produce
more shade on the streets, possibly reducing heat stress on the runners.
4.3. Group IV
As in Group I, there is no shade from buildings or trees. Therefore, it would be desirable to create
shade by proactively planting trees along the streets. However, less than three years remain until the
Olympics, so more urgent action is necessary, such as placing temporary tree planters and sunshades.
4.4. Group V
The left traffic lane, where the runners will be traveling, will be unshaded, although there will
be some shade in the right traffic lane. Therefore, in this section, at first, the runners should also be
allowed to run in the right traffic lane. Second, adjusting tree pruning and expanding the tree canopy
and the installation of temporary sunshades to cover the sky above the road would be an improvement.
Atmosphere 2018,9, 192 14 of 16
5. Conclusions
This study estimated the potential heat load that runners would be exposed to during the planned
course of the marathon in the 2020 Tokyo Olympics. These findings were used to propose possible
methods for reducing that stress. The primary findings are as follows.
1.
On clear sunny days (the worst-case scenario), many stretches along the course were rated as
dangerous or extremely dangerous. In particular, the stretch from 27.5 km–37.7 km is dominated
by race segments in direct sunlight and had near-continuous sections that significantly exceeded
the extremely dangerous standard. This section is in particular need of measures to improve heat
stress conditions.
2.
Under cloudy weather (the best-case scenario), from the start of the race, there were stretches that
were rated as at the safe level. However, in the second half of the race, there were stretches rated
as caution. Therefore, even under conditions that are the coolest for the given time of year, it is
necessary to take measures against heat stress.
3.
Starting the race one hour earlier would decrease race temperatures and increase the number of
shaded sections on the return route. These results suggest that starting the race one hour earlier
would be an effective measure against heat stress by shortening the continuous stretches rated as
extremely dangerous, as demonstrated by comparing data gathered using the race’s scheduled
start time.
4.
Based on the classified sections along the course, we propose: (1) allowing runners to run in
the shade of buildings, (2) making use of urban greenery, such as expanding the tree canopy,
and (3) placing temporary tree planters and sunshades as three effective strategies for reducing
the heat stress from the Sun and longwave radiation.
To summarize, the current marathon course is considered extremely high-risk for heat stress,
particularly on the homeward course, with its high percentage of stretches in direct sunlight.
In addition, the risk of heat stress is present even during overcast weather. Therefore, there is a
need for measures to reduce the overall heat load caused by the sun and longwave radiation, tailored
to the characteristics of individual segments; measures to reduce the effective temperature on runners
throughout the entire course and an earlier race start time than currently scheduled would be effective
at reducing heat stress.
The following two points should also be considered as future research topics. First, this study
identified connected stretches evaluated as extremely dangerous for the runners during the planned
homeward course for the marathon competition in the Tokyo Olympics. However, whether these
stretches could potentially inflict a serious impact on the runners, and cause them to quit the race,
cannot be evaluated based solely on the results of this study. Although preceding literature refers
to core body temperature, which is correlated with the risks of heat stress, gradually increasing in
the dangerous thermal environment while skin temperature increases rapidly [
44
], the effect of heat
accumulation on core body temperature is still not well discussed. Further research that investigates
the effect of heat accumulation on human health is therefore suggested. Second, measures to reduce
heat stress were discussed in Section 4, but their degree of effectiveness on a person moving at a certain
speed and not stopping in any one place for a long period of time, such as a runner, remains unknown;
again, further research is necessary.
Author Contributions:
The authors declare that each have contributed equally to the production of this document,
whereby: (i) E.K., A.I. and M.Y. performed the measurement and analyzed the data; (ii) J.V. and R.B. contributed
to the calculation of the COMFA energy budget; and (iii) A.M. calculated the sky view factor. All authors aided in
the writing and reviewing of the manuscript.
Funding: This research received no external funding.
Acknowledgments:
Thermal environment measurements were conducted in collaboration with Shimizu
Corporation. The weather observation instrument “POTEKA” was provided by Meisei Electric Corporation.
Conflicts of Interest: The authors declare no conflict of interest.
Atmosphere 2018,9, 192 15 of 16
References
1.
Tokyo Metropolitan Government Environment Bureau. Available online: https://www.kankyo.metro.tokyo.
jp/climate/other/countermeasure/suishinkaigi.html (accessed on 16 June 2017). (In Japanese)
2.
Golden, J.S. The built environment induced urban heat island effect in rapidly urbanizing arid
regions—A sustainable urban engineering complexity. Environ. Sci. 2003,1, 321–349. [CrossRef]
3.
Tan, J.; Zheng, Y.; Tang, X.; Guo, C.; Li, L.; Song, G.; Zhen, X.; Yuan, D.; Kalkstein, A.; Li, F.; et al. The urban
heat island and its impact on heat waves and human health in Shanghai. Int. J. Biometeorol.
2010
,54, 75–84.
[CrossRef] [PubMed]
4.
Patz, J.A.; Campbell-Lendrum, D.; Holloway, T.; Foley, J.A. Impact of regional climate change on human
health. Nature 2005,438, 310–317. [CrossRef] [PubMed]
5.
Nakai, S.; Itoh, T.; Morimoto, T. Deaths from heat-stroke in Japan: 1968–1994. Int. J. Biometeorol.
1999
,43,
124–127. [CrossRef] [PubMed]
6.
Wetterhall, S.F.; Coulombier, D.M.; Herndon, J.M.; Zaza, S.; Cantwell, J.D. Medical care delivery at the 1996
Olympic Games. J. Am. Med. Assoc. 1998,279, 1463–1468. [CrossRef]
7.
Zhang, J.J.; Wang, L.D.; Chen, Z.; Ma, J.; Dai, J.P. Medical care delivery at the Beijing 2008 Olympic Games.
World J. Emerg. Med. 2011,2, 267–271. [CrossRef] [PubMed]
8.
Kakamu, T.; Wada, K.; Smith, D.R.; Endo, S.; Fukushima, T. Preventing heat illness in the anticipated hot
climate of the Tokyo 2020 Summer Olympic Games. Environ. Health Prev. Med.
2017
,22, 68. [CrossRef]
[PubMed]
9.
Smith, K.R.; Woodward, A.; Lemke, B.; Otto, M.; Chang, C.J.; Mance, A.A.; Balmes, J.; Kjellstrom, T. The last
Summer Olympics? Climate change, health, and work outdoors. Lancet 2016,388, 642–644. [CrossRef]
10.
Huang, L.; Li, J.; Zhao, D.; Zhu, J. A fieldwork study on the diurnal changes of urban microclimate in four
types of ground cover and urban heat island of Nanjing, China. Build. Environ. 2008,43, 7–17. [CrossRef]
11.
Miyazaki, H.; Moriyama, M.; Yoshida, A. Field study on green canopy as urban cool-spot. Nat. Hum. Act.
1996,1, 51–56.
12.
Middel, A.; Häb, K.; Brazel, A.J.; Martin, C.; Guhathakurta, S. Impact of urban form and design on
microclimate in Phoenix, AZ. Landsc. Urban Plan. 2014,122, 16–28. [CrossRef]
13.
Middel, A.; Selover, N.; Hagen, B.; Chhetri, N. Impact of shade on outdoor thermal comfort—A seasonal
field study in Tempe, Arizona. Int. J. Biometeorol. 2016,60, 1849–1861. [CrossRef] [PubMed]
14.
Helou, N.E.; Tafflet, M.; Berthelot, G.; Tolaini, J.; Marc, A.; Guillaume, M.; Hausswirth, C.; Toussaint, J.F.
Impact of environmental parameters on marathon running performance. PLoS ONE
2012
,7, e37407.
[CrossRef] [PubMed]
15.
Hodder, S.G.; Parsons, K. The effects of solar radiation on thermal comfort. Int. J. Biometeorol.
2007
,51,
233–250. [CrossRef] [PubMed]
16.
Otani, H.; Kaya, M.; Tamaki, A.; Watson, P.; Maughan, R.J. Effects of solar radiation onendurance exercise
capacity in a hot environment. Eur. J. Appl. Physiol. 2016,116, 769–779. [CrossRef] [PubMed]
17.
Yamasaki, Y.; Tsunemattsu, N.; Yokohama, H.; Umeki, K.; Honjo, T. Prediction of Thermal Environment in
Olympic Marathon Course by Thermal Comfort Map. Environ. Inf. Sci. 2016,30, 43–48. (In Japanese)
18.
Kashimura, O.; Minami, K.; Hoshi, A. Prediction of WBGT for the Tokyo 2020 Olympic Marathon.
Jpn. J. Biometeorol. 2016,53, 139–144.
19.
Brown, R.D.; Gillespie, T.J. Microclimatic Landscape Design: Creating Thermal Comfort and Energy Efficiency;
Wiley: Hoboken, NJ, USA, 1995; pp. 1–208.
20.
Vanos, J.K.; Warland, J.S.; Gillespie, T.J.; Kenny, N.A. Improved predictive ability of climate–human–
behaviour interactions with modifications to the COMFA outdoor energy budget model. Int. J. Biometeorol.
2012,56, 1065–1074. [CrossRef] [PubMed]
21.
Brown, R.D.; Gillespie, T.J. Estimating outdoor thermal comfort using a cylindrical radiation thermometer
and an energy budget model. Int. J. Biometeorol. 1986,30, 43–52. [CrossRef] [PubMed]
22.
d’Ambrosio Alfano, F.R.; Palella, B.I.; Riccio, G. The role of Measurement Accuracy on the Heat Stress
Assessment according to ISO 7933: 2004. WIT Trans. Biomed. Health 2007,11, 115–124.
23.
d’Ambrosio Alfano, F.R.; Dell’Isola, M.; Palella, B.I.; Riccio, G.; Russi, A. On the measurement of the mean
radiant temperature and its influence on the indoor thermal environment assessment. Build. Environ.
2013
,
63, 79–88. [CrossRef]
Atmosphere 2018,9, 192 16 of 16
24.
Palella, B.I.; Quaranta, F.; Riccio, G. On the management and prevention of heat stress for crews onboard
ships. Ocean Eng. 2016,112, 277–286. [CrossRef]
25.
Tokura, H.; Nabeshima, M.; Nisioka, M.; Nakao, M. Mobile observation method for spatial variation of
direction and velocity of wind on grand level. Proc. Res. Meet. Arch. Inst. Jpn. Kinki Branch (Environment)
2007,47, 289–292. (In Japanese)
26.
ISO 9920: Ergonomics of the Thermal Environment: Estimation of Thermal Insulation and Water Vapour Resistance
of a Clothing Ensemble; ISO: Geneva, Switzerland, 2007.
27.
Kenny, N.A.; Warland, J.S.; Brown, R.D.; Gillespie, T.G. Part A: Assessing the performance of the COMFA
outdoor thermal comfort model on subjects performing physical activity. Int. J. Biometeorol.
2009
,53, 415–428.
[CrossRef] [PubMed]
28.
Kenny, N.A.; Warland, J.S.; Brown, R.D.; Gillespie, T.G. Part B: Revisions to the COMFA outdoor thermal
comfort model for application to subjects performing physical activity. Int. J. Biometeorol.
2009
,53, 429–441.
[CrossRef] [PubMed]
29.
Vanos, J.K.; Warland, J.S.; Gillespie, T.J.; Slater, G.A.; Brown, R.D.; Kenny, N.A. Human energy budget
modeling in urban parks in Toronto, ON and applications to emergency heat stress preparedness. J. Appl.
Meteorol. Climatol. 2012,51, 1639–1653. [CrossRef]
30.
De Freitas, C.R.; Dawson, N.J.; Young, A.A.; Mackey, W.J. Microclimate and heat stress of runners in mass
participation events. J. Clim. Appl. Meteorol. 1985,24, 184–191. [CrossRef]
31.
Ainsworth, B.E.; Haskell, W.L.; Herrmann, S.D.; Meckes, N.; Bassett, D.R., Jr.; Tudor-Locke, C.; Greer, J.L.;
Vezina, J.; Whitt-Glover, M.C.; Leon, A.S. Compendium of Physical Activities: A second update of codes and
MET values. Med. Sci. Sports Exerc. 2011,43, 1575–1581. [CrossRef] [PubMed]
32.
Vanos, J.K.; Herdt, A.J.; Lochbaum, M.R. Effects of physical activity and shade on the heat balance and
thermal perceptions of children in a playground microclimate. Build. Environ.
2017
,126, 119–131. [CrossRef]
33.
Hardin, A.W.; Vanos, J.K. The influence of surface type on the absorbed radiation by a human under hot, dry
conditions. Int. J. Biometeorol. 2018,62, 43–56. [CrossRef] [PubMed]
34.
Kenny, N.A.; Warland, J.S.; Brown, R.D.; Gillespie, T.G. Estimating the radiation absorbed by a human.
Int. J. Biometeorol. 2008,52, 491–503. [CrossRef] [PubMed]
35.
Middel, A.; Lukasczyk, J.; Maciejewski, R. Sky View Factors from synthetic fisheye photos for thermal
comfort routing—A case study in phoenix, Arizona. Urban Plan 2017,2, 19–31. [CrossRef]
36.
Middel, A.; Lukasczyk, J.; Maciejewski, R.; Demuzere, M.; Roth, M. Sky view factor footprints for urban
climate modeling. Urban Clim. 2018, under review.
37.
Lukasczyk, J.; Middel, A.; Zakrzewski, S.; Arnold, M.; Maciejewski, R. Urban Form and Composition of Street
Canyons: A Human-Centric Big Data and Deep Learning Approach. Landsc. Urban Plan. 2018, under review.
38.
Monteith, J.L.; Unsworth, M.H. Principles of Environmental Physics, 3rd ed.; Academic Press: New York, NY,
USA, 2008; pp. 1–418.
39.
Harlan, S.L.; Brazel, A.J.; Prashad, L.; Stefanov, W.L.; Larsen, L. Neighborhood microclimates and
vulnerability to heat stress. Soc. Sci. Med. 2006,63, 2847–2863. [CrossRef] [PubMed]
40.
Vanos, J.K.; Warland, J.S.; Gillespie, T.J.; Kenny, N.A. Thermal comfort modelling of body temperature and
psychological variations of a human exercising in an outdoor environment. Int. J. Biometeorol.
2012
,56, 21–32.
[CrossRef] [PubMed]
41.
Havenith, G.; Holmer, I.; Parsons, K.C. Personal factors in thermal comfort assessment: Clothing properties
and metabolic heat production. Energy Build. 2002,34, 581–591. [CrossRef]
42.
Vanos, J.K.; Warland, J.S.; Gillespie, T.J.; Kenny, N.A. Review of the physiology of human thermal comfort
while exercising in urban landscapes and implications for bioclimatic design. Int. J. Biometeorol.
2010
,54,
319–334. [CrossRef] [PubMed]
43.
Hosokawa, Y.; Grundstein, A.J.; Vanos, J.K.; Cooper, E.R. Environmental Condition and Monitoring. In Sport
and Physical Activity in the Heat; Casa, D.J., Ed.; Springer: Cham, Switzerland, 2018; pp. 147–162.
44.
Höppe, P. Different aspects of assessing indoor and outdoor thermal comfort. Energy Build.
2002
,34, 661–665.
[CrossRef]
©
2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).
