17

Adaptive approaches to
enhancing resilient thermal
comfort in Japanese offices
Hom B. Rijal, Michael A. Humphreys and J. Fergus Nicol

Introduction

Overview
Most traditional vernacular buildings are adapted to their own climate and culture. They
have enough thermal mass, shading, ventilation and other adaptive features to dampen tem-
perature swings in the indoor thermal environment. Adaptive thermal comfort depends on
behavioural, physiological and psychological adaptations ( Brager and de Dear 1998; Hum-
phreys and Nicol 1998). Workers in many offices use a variety of ‘adaptive opportunities’
to regulate their indoor thermal environment and secure their thermal comfort. If the out-
door environment is acceptably comfortable, occupants in traditional buildings with opening
windows will behave differently according to the season and time of day, so maintaining
their thermal comfort. They do this by adjusting features like external shades, internal blinds
or by opening windows to naturally ventilate their spaces. The free running mode is a term
used to describe buildings with no heating or cooling systems in use. In well- designed,
passive buildings, as many traditional buildings are, people are free to use the various pas-
sive controls to manage their thermal environments without using any fossil fuel energy
to remain comfortable at different times of year. Such adaptable buildings not only help to
mitigate climate change, but also enable people to become accustomed to a wider range of
temperatures, extending their ability to survive in more varied environments, and enhancing
their personal resilience.
In all countries today, modern buildings are more or less dependent on mechanical heat-
ing, ventilating and air- conditioning (HVAC) systems to provide comfort all year, so wasting
large amounts of energy, particularly when outdoor temperatures are comfortable. Indoor
spaces in many HVAC buildings are controlled to constant temperatures throughout the
year, not only using large amounts of energy, but also reducing the need for the thermo-
regulatory systems of the body to adapt to variations of indoor temperature. Lack of thermal
exposure outside limited temperature ranges reduces the physiological ability of the body to
adapt to a broader range of temperatures, making individuals more vulnerable to thermal
stress ( Van Marken Lichtenbelt et al., 2017). What is of particular interest is: What exactly

DOI: 10.4324/9781003244929-22 279

Hom B. Rijal et al.

are the temperature ranges that people find comfortable in adaptable buildings in different
cities and countries of the world?
A resilient building is one which can withstand natural or systemic changes without seri-
ous cost to comfort or energy use. HVAC buildings which rely heavily on the use of energy
for mechanical cooling are often not resilient for thermal comfort and energy saving, and we
need to design passive and low- energy buildings to limit global warming within 2 degrees of
preindustrial level ( Roaf and Beckmann, 2017; Lau and Fong, 2021).

Japanese offices
Japanese office buildings in and around Tokyo are mainly large high-rise buildings, with
indoor climates controlled by HVAC throughout the year. The design temperature for sum-
mer was kept around 26–27°C. In winter, it was 21.1°C in 1955, 18°C from 1956 to 1963,
and 20–22°C since 1964 ( Yamamoto et al., 2010). They have glazed façades and no openable
windows. However, some smaller, low-rise office buildings are able to operate with AC, or
just with natural ventilation, using mixed mode systems (CBE, 2016). The use of windows
and pedestal fans controls indoor in the shoulder or middle seasons, or when the outdoor air
temperature is comfortable at other times of year. Mechanical cooling and heating are used
only during summer and winter peaks. This provides a life- saving flexibility of options that
proved so essential after two world- class catastrophic events: the Fukushima nuclear disaster
and COVID-19 pandemic. The Great East Japan Earthquake in 2011 caused enormous dam-
age, including the destructions of the Fukushima nuclear plant, leading to a 15% peak-power
reduction imposed to address power shortages (Tanabe et al., 2013). And widening the com-
fort setting in offices was an immediate response: the Japanese government recommended
that room air temperature should be 28°C in the cooling season, as part of the Cool Biz
scheme. Awareness of power savings was increased, with more than 90% of people accepting
poorer indoor working conditions because of the national emergency ( Tanabe et al., 2013).
During the COVID-19 pandemic, guidance has been issued by Japanese authorities that
windows must be left open in many building types, even when the AC is in use, to purge
spaces of the virus, because many HVAC systems recycle air from room to room, so increas-
ing the risk of cross transmission of the pathogen indoors (Hayashi et al., 2020; SHASE and
AIJ, 2020; Kurabuchi et al., 2021). Such incidents will increasingly happen in a future with
more extreme climates, but the key point is: what are the safe limits for indoor environments
in Japan in shared spaces like offices? This is where the establishment of an adaptive model
for Japanese offices, providing comfortable and safe limits for indoor temperatures, is of such
importance.

An adaptive model for Japanese offices

Rijal et al. (2017) proposed an adaptive model for Japanese offices that could provide such
safe temperatures, but it rested on the small quantity of data available then, so further work
was needed. Previous work had shown that various adaptive mechanisms made important
contributions that support the broader thermal preferences found in plotting the data for that
adaptive model. A later study begun in 2017 took both thermal measurements and surveyed
for occupant thermal comfort and behavioural responses. The work took more than two
years, in 23 office buildings in the Tokyo and Yokohama areas of Japan. The objectives then
were to explore seasonal differences in the comfort temperatures experienced in these build-
ings, on which could be based a more solidly grounded adaptive model for Japanese offices,

280
 Adaptive approaches

reinforced by a range of thermally mapped adaptive behaviours. How do people react to

hotter or colder temperatures and can that knowledge be used to inform policies on how to
extend the limits of indoor working environments, while not threatening the health, safety
or comfort of building users? The study provided the potential to map behaviours against
the thermal landscapes inside buildings, and offered insights into new strategies for designing
buildings that provide more resilient comfort in more extreme weather.

Methods

Investigated buildings
Two comfort surveys underpin the proposed model. In survey 1, 11 office buildings were
investigated in the Tokyo and Kanagawa areas of Japan from August 2014 to October 2015
( Rijal et al., 2017; Rijal et al., 2019b). In survey 2, 16 office buildings were investigated in
the same areas from August 2017 to November 2018. These contemporary buildings are
broadly typical of those available in the area ( Rijal et  al., 2020). The numbers of HVAC
and mixed mode buildings were 6 and 17, respectively (four mixed mode buildings were
in both surveys). The mixed mode buildings are of the change- over type (CBE, 2016), in
which the building can be in the free running mode, or air- conditioned (AC), depending on
the season and time of day. All mixed mode buildings have openable windows, and most of
them ( B4–B8) are university office buildings. Generally, people open and close windows in
spring and autumn to control the room temperature. In summer, people close the windows
and use split cooling. In winter, they close the windows and use heating. Two of the HVAC
buildings have both manual and automatic natural ventilation openings. The surveyed floors
were selected according to the willingness of the occupants to take part in the surveys, and
with the permission of the building managers.

The measurements, thermal comfort survey and occupant behaviour survey

The indoor air temperature, globe temperature, relative humidity and air movement were
measured 1.1m above floor level, away from direct sunlight, using a data logger. Outdoor air
temperature and relative humidity were obtained from the nearest meteorological station.
The thermal comfort survey was conducted in Japanese. The thermal sensation scale used is
shown in Table 17.1. We distributed a paper questionnaire for survey 1, while a web-based
survey was mostly used for survey 2. We conducted both transverse and longitudinal surveys
but for this chapter, we analyse only the data from the transverse surveys.
Transverse surveys were conducted one day each month by researchers visiting each build-
ing with measurement instruments and with questionnaires for the subjects. On each visit,
one set of responses was collected from each subject. The instruments were set up on an office
table, and questionnaire responses were requested from subjects seated near the instruments
(within 2–3 metres). While people were responding to the questionnaire, the researcher re-
corded the environmental controls in use, and the thermal environmental data. After com-
pleting the data collection for that group, the instruments were moved to the next group, and
so on. This process was repeated for all groups each month. A few people did not provide re-
sponses because of their busy schedule, and some were not in the office at the time of the visit.
Window opening, heating use and cooling use were recorded in binary form at the time of
completing the questionnaire (0 = window closed or heating/cooling off, 1 = window open
or heating/cooling on). We collected 4,660 votes in survey 1 and 2,637 votes in survey 2.

281
Hom B. Rijal et al.

Table 17.1 Scale used for the thermal comfort surveys

No. Modified thermal sensation vote (mTSV)

1 Very cold
2 Cold
3 Slightly cold
4 Neutral (neither hot nor cold)
5 Slightly hot
6 Hot
7 Very hot

Calculation methods

Griffiths method
The comfort temperature is estimated by using the Griffiths’ method (Griffiths, 1990; Nicol
et al., 1994; Humphreys et al., 2013; Rijal et al., 2017).

Tc = T g + ( 4 − C ) / a (17.1)

Tc: The comfort temperature by Griffiths’ method (°C); Tg: Globe temperature (°C);
C: Thermal sensation vote; a: The rate of change of thermal sensation with room
temperature.
When the thermal sensation vote is 4 (neutral), the comfort temperature equals the globe
temperature; otherwise, it differs from it. The comfort temperature was estimated using a
value of 0.50 for the rate of change as in previous studies (Humphreys et al., 2007; Hum-
phreys et al., 2013; Rijal et al., 2017).

Running mean outdoor temperature

The exponential running mean outdoor temperature is calculated using the following equa-
tion (McCartney and Nicol, 2002).

Trm = αTrm −1 + (1 − α )Tod −1 (17.2)

where Trm-1 is the running mean outdoor temperature for the previous day (°C), and Tod-1
is the daily mean outdoor temperature for the previous day (°C). So, if the running mean
has been calculated (or assumed) for one day, then it can be readily calculated for the
next day, and so on. Į is a constant between 0 and 1, which defines the speed at which
the exponential running mean responds to the outdoor air temperature. ASHRAE Stan-
dard (2013) recommends using a value of Į between 0.60 and 0.90. The correlation
between comfort temperature and outdoor temperature is almost constant in this range
( McCartney and Nicol, 2002), so the value chosen is not critical. We have chosen Į to
be 0.8, as evaluated by McCartney and Nicol (2002) and as used in the derivation of the
CEN standard (CEN, 2007).

282
 Adaptive approaches

Logistic regression analysis

To predict the proportion of windows open, heating and cooling used, logistic regression
analysis was conducted (see, for example, Nicol and Humphreys, 2004). The relationship
between the probability ( p) of that control being used, and the spot outdoor temperature
during the voting (To) is of the form:

{ }
logit ( p ) = log p / (1 – p ) = bTo + c (17.3)

where p = exp(bTo + c ) / {1 + exp(bTo + c ) } (17.4)

where exp (exponential function) is the base of the natural logarithm, b is the regression
coefficient for To and c the constant in the regression equation.

Results and discussion

Data was divided into three groups for analysis. If heating was used at the time of the survey
visit, the data were classified as being in the heating mode ( HT). If cooling was in use at the
time of the visit, the data were classified as being in the cooling mode (CL). If neither heating
nor cooling was in use, the data were classified as being in the free-running mode (FR). The
CL and HT modes are distinct groups of data (generally CL occurs in summer and HT in
winter), and they need to be analysed separately. This classification differs from that used in
the CIBSE Guide (CIBSE, 2006), and in ASHRAE Standard 55. In CIBSE guide, the data
was combined for the heating and cooling modes. For the ASHRAE standard, buildings
were classified as naturally ventilated ( NV) or AC.

Outdoor and indoor temperatures

The Köppen climate classif ication of the investigated areas is humid subtropical (Cfa).
The mean outdoor air temperatures at the voting time for survey 1 were 20.7°C,
10.4°C and 24.9°C for FR, HT and CL modes, respectively. The values were similar
for survey 2.
The indoor globe temperature correlates strongly with the indoor air temperature
in both surveys ( Rijal et al., 2020), and so the results can be presented using the globe
temperature alone. The mean globe temperatures during the voting for survey 1 were
25.0°C, 23.8°C and 25.9°C for FR, HT and CL modes which are 0.1°C, 0.2°C and
0.7°C lower than survey 2. Yamamoto et al. (2010) also found that Japanese typical of-
fices were operated at around 27°C in summer and 23°C to 24°C in winter. The Japanese
government recommended that indoor temperature should be 20°C in winter ( HT) and
not higher than 28°C in summer (CL) ( Nakashima, 2013). ( In 2005, they recommended
the temperature setting and later they recommended the indoor temperature.) The re-
sults show that the mean indoor temperatures during heating and cooling were quite
different from those recommended. The seasonal range of the indoor temperature was
quite small, while there was a wide seasonal range of outdoor temperature. During the
surveys, respondents felt hot in HT mode and cold in CL mode, and thus seasonal vari-
ation of temperature setting is required for resilient comfort, which can also contribute
to energy saving.

283
Hom B. Rijal et al.

Adaptive thermal comfort

Distribution of thermal sensation

Table 17.2 shows the distribution of thermal sensation vote by building types and modes. In
mixed mode buildings, about 1/3 of data are distributed in each mode. It is important from
the viewpoint of energy use that 32% of data are in FR mode in the mixed mode buildings.

Table 17.2 Distribution of thermal sensation votes (1–7) by building types and modes

Building

MM HVAC

Mode N % N %

FR 949 32.2 164 3.8

HT 949 32.2 1507 34.6
CL 1048 35.6 2680 61.6
All 2946 100 4351 100

Figure 17.1 Percentage of thermal sensation vote. A: Each survey, B: All

284
 Adaptive approaches

A mixed mode building may save about 30% of heating and cooling energy, which might be
important for resilient building design. However, only 3.8% of votes came from the two fully
HVAC buildings. One building used an automatic natural ventilation system and another
building did not use heating or cooling during the time of the survey. The percentage of
votes for CL mode is 61.6%, which is almost twice than the HT mode, and thus the duration
of cooling use is much longer than heating use in HVAC buildings.
The percentage of thermal sensation responses is shown in Figure 17.1. The percentages of
thermal sensation votes are 15.3% in FR mode, 33.7% in HT mode and 51.1% in CL mode,
and thus 85% responses are from heating and cooling modes. The distributions were much the
same in both surveys. Occupants sometimes felt hot (greater than 5) in CL mode and some-
times felt cold ( less than 3) in HT mode, despite the use of heating or cooling. 12.5% of over-
heating ( hot side votes) in HT mode and 11.6% of over cooling (cold side votes) in CL mode
are found. It indicated that the excessive energies for heating and cooling are being used to cre-
ate uncomfortable indoor thermal environments, which are not suitable from the viewpoint of
the resilient thermal comfort and energy use. Perhaps because of the small monthly variations
of temperature in the offices, and the tendency of people to adapt to the temperatures they en-
counter, there are many ‘4 neutral’ votes in each mode (55–58%). It is conventional to consider
as comfortable responses that fall in categories 3, 4 and 5 ( Fanger, 1972; Humphreys and Nicol,
1970; Nicol et al., 1999; ASHRAE, 2013). These percentages are very high (95%). Thus, it
can be said that occupants were generally satisfied in the thermal environment of their offices.

Comfort temperature
Figure  17.2 shows the percentage of comfort temperature as calculated by the Griffiths’
method. The distributions for the two surveys are similar. We calculated the comfort tem-
perature for each thermal sensation vote. Each value is of low accuracy, but is unbiased. ( The
low accuracy of the individual values contributes to the quite large standard deviations of the
distributions. Nevertheless, the standard errors of the means in Figure 17.2 are all less than
0.1 K.) The mean comfort temperature obtained from both surveys is 24.9°C (FR mode),
24.3°C (HT mode) and 25.6°C (CL mode) which are very similar to indoor globe tempera-
ture. We found that in these buildings, the comfort temperature was 4.3°C higher in HT
mode and 2.4°C lower in CL mode than the recommended values by Japanese government
( Nakashima, 2013). The range of comfort temperature in free running mode is 21. 3–28.5°C
(Mean ± 2SD). The acceptable range of room temperatures can be extended from 18 to 30°C
by a task-ambient heating and cooling system (Zhang et al., 2010).
Figure 17.3 and Table 17.3 show the relationship between the comfort temperature and
globe temperature for raw data and binned data. To reveal trends in the scattered plots of
the raw data, we have also analysed the binned data, as performed by other studies (Gautam
et al., 2019; Jowkar et al., 2020). The correlation between comfort temperature and globe
temperature is high. The regression coefficients of raw data and binned data are very similar.
The results showed that fundamentally the people had adapted to a large extent to the tem-
peratures that were provided. Had more seasonal drift of indoor temperature been provided,
it is likely that people would have adapted to it (see Humphreys et al., 2016, chapter 27). If the
indoor temperature is low, the comfort temperature is on average above the indoor tempera-
ture, and if the indoor temperature is high, the comfort temperature is on average below the
room temperature. People on average slightly under-adapt to changes in room temperature,
as has been found in previous studies (Humphreys et al., 2016). It has been found that people
adapt better to gradual drifts in room temperature than to sudden changes. This indicates

285
Hom B. Rijal et al.

Figure 17.2 Percentage of comfort temperature. A: Each survey (Rijal et al., 2017), B: All

that sufficient ‘thermal mass’ should be included in the design, to reduce sudden changes in
room temperature that may lead to thermal stress for occupants. This shows that how much
change can be tolerated depends on the rate of change in temperatures, as well as the extent
of such changes. This is relevant to making occupants vulnerable to thermal hazards, and
thus of increasing or decreasing their thermal resilience to outdoor temperatures.

Monthly and seasonal differences in temperature

Figure 17.4 shows the monthly and seasonal variation of the comfort temperature. The results
are similar for both surveys. It is evident that the comfort temperature closely tracks the mean
indoor globe temperature over the year, the difference between them being about 1K in any
month or season. The comfort temperature and the indoor globe temperature both show
rather little monthly and seasonal variation. The comfort temperature is 22.1°C in January,
25.9°C in August and September in FR mode. Thus, the monthly variation of the mean
comfort temperature is 3.8 K ( Figure 17.4A). The comfort temperature is 24.9°C in spring,
25.7°C in summer, 24.7°C in autumn and 23.6°C in winter in FR mode, and thus seasonal

286
 Adaptive approaches

Figure 17.3 Relation between the comfort temperature and globe temperature. A & B: Raw
data; C (Rijal et al., 2017) & D: Binned data. (Each point indicates the monthly
mean temperature by building.)

Table 17.3 Regression equations of comfort temperature and globe temperature

Survey Data Equation n R2 S.E. p

Survey 1 Raw Tc = 0.57Tg + 10.6 4658 0.28 0.014 <0.001

Binned* Tc = 0.56Tg + 10.9 138 0.85 0.020 <0.001
Survey 2 Raw Tc = 0.65Tg + 8.7 2637 0.42 0.015 <0.001
Binned* Tc = 0.65Tg + 8.6 206 0.81 0.022 <0.001
All Raw Tc = 0.61Tg + 9.7 7,295 0.34 0.010 <0.001
Binned* Tc = 0.60Tg + 9.9 344 0.83 0.015 <0.001

Tc: Comfort temperature (°C), Tg: Indoor globe temperature (°C), n: Number of samples, R 2: Coefficient of
determination, S.E.: Standard error of the regression coefficient, p: Significance level of regression coefficient, *:
Weighted regression analysis.

difference in comfort temperature is 2.1°C (Figure 17.4B&C). These seasonal variations are

smaller than we had found for dwellings: 9.4°C ( Rijal et al., 2019a).
The summer comfort temperature might be low due to low air movement. The mean air
movement and standard deviation in summer were 0.18±0.14 m/s in FR mode and 0.15±0.09
m/s in CL mode. Rather than turning on the cooling in summer to bring down the room
temperature, consideration should be given to providing efficient fans to increase the air
movement. When the occupants control the fans, air speeds of up to 1.2 m/s are permissible
according to the ASHRAE standard (ASHRAE, 2013). Such air speeds are equivalent to
cooling the room temperature by 3 or 4 degrees ( Humphreys, 1970; Humphreys and Nicol,
1995; Nicol, 2004). The limit to elevated air speed is because papers are disturbed at higher

287
Hom B. Rijal et al.

Figure 17.4 Monthly and seasonal variation of comfort temperature and globe temperature
with 95% confidence intervals. A: FR (Rijal et al., 2017), B: Each survey (Rijal et al.,
2017), C: All. ( The mean outdoor temperature is also shown in A.)

288
 Adaptive approaches

speeds. Since so much office work is now done on computer rather than on paper, perhaps
the need for this limit is passing. People accept far greater seasonal variations of room tem-
perature in other countries, and when working at home. It may be concluded that permitting
greater seasonal variation in office temperatures, with the provision of effective fans, could
save energy and decrease the vulnerability of occupants to experiencing discomfort when
temperatures vary from the norm.

The adaptive model

An adaptive model relates the indoor comfort temperature to the outdoor air temperature
(Humphreys, 1978; de Dear and Brager, 1998; ASHRAE, 2004; CEN, 2007; Nicol and
Humphreys, 2010). Figure 17.5 and Table 17.4 show the relation between the comfort tem-
perature (calculated by the Griffiths’ method) and the running mean outdoor temperature.
However, the mean comfort temperatures in the HT and CL modes were not very different
(see Figure 17.2), and thus we have combined them, as is done in the CIBSE guide. The data
are binned at 1°C interval of running mean outdoor air temperature.
Broadly similar trends were demonstrated in both surveys for raw and binned data. The
relations are useful, because the regression lines indicate a substantial rise in the temperature
for comfort indoors with rising outdoor temperature, over the range for which heating or
cooling is likely to be used.
The regression coefficient and the correlation coefficient in the FR mode are higher than
in the HT and CL modes. As shown in Table 17.5, the regression coefficient is lower than that
in the CEN standard (FR = 0.33) and it is similar to the CIBSE guide (CL&HT = 0.09). It
is lower than found for Japanese dwellings ( Rijal et al., 2013, 2019a).
It is probable that the low gradients that are found for these ‘adaptive models’ just reflect
the small seasonal trends of the indoor temperatures in our sample of office buildings. 85%
of data of this study is from the CL&HT mode, and thus we need to increase the sample size
of the FR mode to obtain a reliable estimate. In view of the climatic variation across Japan,
field surveys are to be conducted in various parts of Japan.
The equations can be used to predict the indoor comfort temperature for these buildings.
For example, when the running mean outdoor temperature is 25°C, the comfort tempera-
ture would be 26.1°C for the FR mode. Similarly, when the running mean outdoor tempera-
ture is 10°C and 28°C in the CL&HT mode, the comfort temperature would be 24.5°C and
26.3°C, respectively. The results indicate that the range of the mean comfort temperature for
CL&HT mode is small – less than 2K – probably because the indoor environment is designed
based on narrow range of PMV/PPD and the occupants adapted to the small seasonal vari-
ation of the temperature in these offices. Had the government-recommended temperatures
been provided, the occupants would perhaps have adapted to the summer or winter environ-
ment of the office without compromising comfort or productivity, but further surveys would
be needed to explore the question.
The comfort temperature is different in FR and CL&HT modes for a given outdoor tem-
perature which might be one of the important issues. If they can adapt to the outdoor tempera-
ture in the FR mode, logically they can also adapt to outdoor temperature in CL&HT mode.
If we apply the adaptive model to the temperature setting of the HVAC building in middle
seasons, it may be possible to save a significant amount of energy in HVAC buildings.

289
Hom B. Rijal et al.

Figure 17.5 Relation between the comfort temperature and the running mean outdoor tem-
perature. A (Rijal et  al., 2017), B, C  & D: Raw data; E  & F: Binned data. ( The
dashed lines indicate equality of indoor comfort temperature and outdoor run-
ning mean temperature.)

290
 Adaptive approaches

Table 17.4 Regression equations of adaptive model

Mode Survey Data Equation n R2 S.E. p

FR Survey 1 Raw Tc = 0.21 Trm + 20.8 422 0.42 0.012 <0.001

Binned* Tc = 0.20 Trm + 20.9 11 0.79 0.035 <0.001
Survey 2 Raw Tc = 0.15 Trm + 22.5 689 0.14 0.014 <0.001
Binned* Tc = 0.15 Trm + 22.4 20 0.60 0.030 <0.001
All Raw Tc = 0.15 Trm + 22.3 1,111 0.21 0.009 <0.001
Binned* Tc = 0.15 Trm + 22.2 31 0.62 0.022 <0.001
HT & CL Survey 1 Raw Tc = 0.07 Trm + 23.9 4,236 0.10 0.003 <0.001
Binned* Tc = 0.06 Trm + 23.9 40 0.55 0.009 <0.001
Survey 2 Raw Tc = 0.10 Trm + 23.5 1,948 0.20 0.005 <0.001
Binned* Tc = 0.10 Trm + 23.5 36 0.62 0.014 <0.001
All Raw Tc = 0.08 Trm + 23.5 6,184 0.14 0.003 <0.001
Binned* Tc = 0.08 Trm + 23.7 76 0.56 0.008 <0.001

Tc: Comfort temperature (°C), Trm: Running mean outdoor temperature (°C), n: Number of samples, R 2: Coefficient
of determination, S.E.: Standard error of the regression coefficient, p: Significance level of regression coefficient, *:
Weighted regression analysis.

Table 17.5 Regression equations in this study and previous studies

References Buildings Mode Equation

Rijal et al. (2017) Offices FR Tc = 0.21 Trm + 20.8

CL&HT Tc = 0.07 Trm + 23.9
CIBSE (2006) Offices CL&HT Tc = 0.09 Trm + 22.6
CEN (2007) Offices FR Tc = 0.33 Trm + 18.8
ASHRAE (2004) Offices NV Tc = 0.31 Tom + 17.8
Humphreys (1978) All types FR Tc = 0.534 Tom + 11.9
Rijal et al. (2019a) Dwellings FR Tc = 0.480 Trm + 14.4
CL Tc = 0.180 Trm + 22.1
HT Tc = 0.193 Trm + 18.3
All Tc = 0.432 Trm + 15.4
Rijal et al. (2013) Dwellings FR Tc = 0.531 Trm + 12.5
CL Tc = 0.297 Trm + 18.8
HT Tc = 0.307 Trm + 16.5

FR: Free running, CL: Cooling, HT: Heating, NV: Naturally ventilated, Tc: Comfort temp. (°C), Trm: Daily running
mean outdoor air temp. (C), Tom: Monthly mean outdoor air temp. (°C).

Adaptive mechanisms
The comfort temperature indoors was found to vary with the outdoor temperature, because
people are adapting well in their offices using various behavioural, physiological and psycho-
logical adaptations ( Brager and de Dear, 1998; Humphreys and Nicol, 1998). The following
section focuses on adaptive mechanisms the occupants used to regulate their resilient thermal
comfort.

291
Hom B. Rijal et al.

Window opening behaviour

The overall mean value of ‘open window’ for all data is 0.17 (n = 2918). When we compared
by building, the mean value ranged from 0.00 to 0.45. The mean window opening is 0.42
(n = 949), 0.03 (n = 944) and 0.07 (n = 1025) for FR, HT and CL modes, respectively. Inter-
estingly, the mean window opening in the UK office buildings was 0.70 in NV mode and
0.04 in AC mode ( Rijal et al., 2007). As the window opening is very low in the CL and HT
modes, we shall limit the analysis to the FR mode.
Derived equations predicted the window opening behaviour in these Japanese offices.
Such predictions are needed for the thermal simulation of buildings. The following regres-
sion equations were obtained for all data in between the windows open and the outdoor air
temperatures:

Survey 1 : logit ( p) = 0.507To − 10.0 (n = 399, R 2* = 0.27, S.E. = 0.059, p < 0.001) (17.5)

Survey 2 : logit ( p) = 0.251To − 6.4 (n = 550, R 2* = 0.17, S.E. = 0.030, p < 0.001) (17.6)

All: logit ( p ) = 0.350To − 7.6 (n = 949, R 2* = 0.26, S.E. = 0.027, p < 0.001) (17.7)

where R 2* is Cox and Snell R 2. The regression coefficient of the survey 1 is higher than that
of survey 2. Survey in the UK ( Rijal et al., 2007) returned a regression coefficient of 0.181,
so the regression coefficient of this research is higher than previously found due to different
climate and culture. This trend is similar with the Japanese dwellings (Majima et al., 2007;
Rijal et al., 2013, 2018). Perhaps window opening behaviour varies widely from building to
building, depending how easy it is to open a window. As shown in Figure 17.6, the propor-
tion of windows open rises as the outdoor temperature rises.
Window opening is important in mixed mode buildings. Manual window opening might
be better in small size office buildings. Constraints on window opening should be avoided
for thermal comfort ( Rijal et al., 2018). Window opening should be easy to control, with
good access, with screens to prevent insects and so on. The adaptive approach depends on
the actions of the building occupants and thus the windows could be readily used by office
workers. In open-plan offices, it is better to decide who is responsible for window opening.
Automatic window opening based on adaptive algorithms might be better in large offices.

Figure 17.6 Relation between the window opening and outdoor air temperatures in FR mode

292
 Adaptive approaches

Window opening may not be sufficient in deep floor plan offices, and thus a combination of
window opening and ceiling fans might be advantageous.

Clothing adjustments
The mean clothing is 0.73 clo, 0.89 clo and 0.60 clo in FR, HT and CL modes, respectively.
The results show that people adjusted their clothing considerably in each mode. In order to
predict the clothing insulation, a regression analysis of the clothing insulation and outdoor
air temperature was conducted. Figure 17.7 and Table 17.6 show the relation between the
clothing insulation and outdoor air temperature for raw data and binned data. The data are
binned at 1°C interval of outdoor air temperature.
Similar results have been found for both surveys for raw data and binned data. The regres-
sion coefficients are negative for all equations. It shows that the clothing insulation decreases
when outdoor air temperature is increased. This trend is similar to that is found for Japanese
dwellings ( KC et al., 2018; Rijal et al., 2019a).
Due to the introduction of the Cool Biz in 2005, clothing insulation is now much lower
in office buildings. However, it is still heavier than clothing in dwellings ( KC et al., 2018;
Rijal et al., 2019a). If clothing insulation was closer to that in dwellings, energy could be
saved in office buildings by increasing the temperature setting for turning on the cooling.
According to Nicol et al. (1994), a seasonal difference of 0.50 clo corresponds to a difference

Figure 17.7 Relation between the clothing insulation and outdoor air temperature A: Raw
data, B: Binned data

293
Hom B. Rijal et al.

Table 17.6 Regression equations of clothing insulation and outdoor air temperatures

Mode Survey Data Equation N R2 S.E. P

FR Survey 1 Raw Icl = −0.030 To + 1.2 419 0.27 0.002 <0.001

Binned* Icl = −0.030 To + 1.3 16 0.78 0.004 <0.001
Survey 2 Raw Icl = −0.025 To + 1.2 676 0.23 0.002 0.001
Binned* Icl = −0.025 To + 1.2 30 0.70 0.003 <0.001
All Raw Icl = −0.027 To + 1.2 1095 0.27 0.001 0.001
Binned* Icl = −0.027 To + 1.3 46 0.73 0.002 <0.001
HT & CL Survey 1 Raw Icl = −0.014 To + 0.9 4180 0.30 <0.0001 <0.001
Binned* Icl = −0.014 To + 1.0 67 0.78 0.001 <0.001
Survey 2 Raw Icl = −0.017 To + 0.9 1922 0.44 <0.001 <0.001
Binned* Icl = −0.017 To + 1.1 62 0.85 0.001 <0.001
All Raw Icl = −0.015 To + 1.0 6102 0.35 <0.0001 <0.001
Binned* Icl = −0.015 To + 1.0 129 0.79 0.001 <0.001

Icl : Clothing insulation (clo), To: Outdoor air temperature (°C), n: Number of samples, R 2: Coefficient of
determination, S.E.: Standard error of the regression coefficient, p: Significance level of regression coefficient, *:
Weighted regression analysis.

of 3.5– 4.0°C in the comfort temperature, and thus it plays an important role for resilient
comfort and energy saving.

Heating and cooling use

Heating and cooling use in mixed mode building was analysed, also providing thermal limits
for the simulation of thermal behaviours in such buildings. The following regression equa-
tions were obtained for heating use or cooling use and the outdoor air temperature:
Heating use

Survey 1 : logit ( p ) = −0.839To + 13.6 ( n = 1241, S.E. = 0.055, R 2* = 0.66, p < 0.001) (17.8)

Survey 2 : logit ( p ) = −0.497To + 6.8 ( n = 1688, S.E. = 0.028, R 2* = 0.54, p < 0.001) (17.9)

All: logit ( p ) = −0.589To + 8.8 ( n = 2929, S.E. = 0.024, R 2* = 0.59, p < 0.001) (17.10)

Cooling use

Survey 1 : logit ( p ) = 0.359To − 8.5 ( n = 1241, S.E. = 0.024, R 2* = 0.37, p < 0.001) (17.11)

Survey 2 : logit ( p ) = 0.499To − 11.7 ( n = 1698, S.E. = 0.026, R 2* = 0.59, p < 0.001) (17.12)

All: logit ( p ) = 0.437To − 10.3 ( n = 2939, S.E. = 0.017, R 2* = 0.52, p < 0.001) (17.13)

The regression coefficient for heating use for survey 1 is higher than for survey 2. There is
less difference between the coefficients for cooling use. The regression coefficient for heating
use is higher than previously found in Europe and Pakistan ( Rijal et al., 2009). This trend
is similar to that found for Japanese dwellings ( Imagawa et al., 2019). These equations are
presented in Figure 17.8. Note that the proportion of the heating use rises as the outdoor

294
 Adaptive approaches

Figure 17.8 Relation between the heating use or cooling use and outdoor air temperature in
mixed mode building. A: Heating use, B: Cooling use

temperature decreases, and the proportion of the cooling use rises as the outdoor temperature
increases.
People start to use heating about 17°C and cooling about 20°C of outdoor air tempera-
ture, and about half of them used heating at 15°C and cooling at 24°C of outdoor tempera-
ture in mixed mode building. However, HVAC buildings used heating or cooling all year
round. Thus, mixed mode buildings can save energy significantly in comparison to HVAC
buildings. If we design the mixed mode building properly, it may be possible to push the
heating use curve in left side and cooling use curve in right side. By considering passive
heating methods and internal heat storage, it may be possible to increase the temperatures at
which heating starts to be used. Similarly, the installation of ceiling fans might be helpful to
increase temperature at which cooling starts to be needed. Generally, reducing the heating
temperature setting or increasing the cooling temperature setting by 1°C, around 10% of
energy used can be saved (Hoyt et al., 2015; Rijal et al., 2021).

Towards thermal resilience

This research project has again demonstrated the power of human adaptation to provide
comfortable conditions for occupants of offices. People secured their comfort by choosing
clothing suitable for the temperature, by opening or closing windows (if the building had
openable windows), by turning heating on or off, and by switching on cooling when desired.
The result of these processes ensured that the comfort temperature tracked the indoor tem-
perature month by month throughout the year, so achieving a high level of thermal comfort,
such that 95% of the subjective responses were within the central three categories of the scale
(the ‘comfort zone’). The FR mode of a mixed mode building is resilient for thermal comfort
and can save a significant amount of energy compared to an HVAC building. The seasonal
difference of indoor temperature is small in offices, and it needs to increase to be similar to
that found for dwellings, so improving office buildings by providing the various adaptive
opportunities for resilient thermal comfort.
However, the results from the two surveys included in the project differed in some re-
spects, although showing the same broad trends. In survey 1, the use of heating increased more
steeply as the temperature fell ( Figure 17.8), the window opening increased more steeply as
the temperature rose (Figure 17.6) and the comfort temperature (in the free-running mode)

295
Hom B. Rijal et al.

rose more steeply as the outdoor temperature rose. These differences are not entirely unex-
pected, because adaptation rests on the interaction between the building and its occupants,
and is further influenced by social constraints. But the differences mean that it is not possible
to provide a single definite relation between, say, window opening and outdoor temperature.
Relationships of this kind are averages from the buildings included in the survey. This limits
their usefulness for the thermal simulation of buildings but offers the prospect of a consider-
able degree of flexibility in responses to indoor temperature changes, and it is that flexibility
that denotes the extent to which occupants in buildings can adapt thought the own actions,
in some way, the degree of their resilience to change.

Conclusions
We summarise the principal findings of these surveys of comfort and occupant behaviour in
mixed mode buildings in Tokyo and Kanagawa ( Japan) as follows:

1 The occupants were highly satisfied with the thermal environment of their offices, as
indicated by the high proportion of ‘neutral’ responses, and the large proportion of ther-
mal sensation votes in the three central categories of the scale.
2 Cooling and window opening started to be used when the outdoor temperature rose
above about 20°C and heating when the outdoor temperature dropped below 17°C. The
average comfort temperature was found to be 25.6°C when cooling was used, 24.3°C
when heating was used and 24.9°C when neither heating nor cooling were used (the free
running mode). The comfort temperature for heating mode is surprisingly high and it is
not resilient for energy saving.
3 The seasonal difference in comfort temperature in offices (2.1°C) is significantly less
than had been found for Japanese dwellings (9.4°C). The comfort temperatures tracked
the concurrent mean indoor globe temperatures. We need to permit more seasonal vari-
ation in indoor temperature in office buildings, as was found in dwellings for resilient
thermal comfort.
4 Adaptive models are given to estimate the probable comfort temperature from outdoor
air temperature. It can be used not only for the design and temperature control of the
FR mode but also HVAC buildings during the comfortable season.
5 Behavioural adaptations (window opening, clothing adjustments, heating/cooling use)
are related to the outdoor air temperature. They underlie the adaptive models, which are
important for resilient thermal comfort and energy saving.
6 Based on these finding, it is possible to assume that Japanese offices that can operate ef-
fectively in free running mode can provide generally safe limits for indoor temperatures
in the region of 21.3°C and 28.5°C, provided there are enough adaptive opportunities.

Because people tend to adapt to the indoor climate they encounter, there can be no sin-
gle universal adaptive relation between comfort temperature and the outdoor temperature.
The design and construction of the building and its services intervene between the climate
and the indoor comfort temperature. Nevertheless, a thermal comfort survey of the type
described can shed light on the success of the design of the buildings and its services. The
general level of thermal comfort in these offices was uncommonly high, indicating successful
thermal design. For minimal energy consumption, consistent with thermal comfort, the
building should be capable of flexibly providing comfortable indoor temperatures over a
wide range of outdoor temperatures, without the need for heating or cooling.

296
 Adaptive approaches

In the climate of Tokyo and Yokohama, this requires effective shading in summer, and win-
ter warming by solar radiation from the lower angle winter sunshine. The overall mass of the
buildings helps maintain stable office temperatures, avoiding extreme temperature-differences
during the diurnal cycle, while the provision of openable windows under the control of the
occupants assisted ventilation and provided temperature control. The provision of heating and
cooling on demand ensured that comfort could be achieved during both winter and summer.
However, we need to provide seasonal variation in indoor temperature similar to that found in
dwellings by improving the office building and providing the various adaptive opportunities
for resilient thermal comfort and energy saving. People are more conscious of the need for
design for the mixed mode rather than fully HVAC buildings. So, we need to shift building
design policy towards the resilient thermal comfort in the coming days.

Acknowledgements
We are grateful to all those who participated in the survey, and to the students who helped
with the data- entry. We would like to thank Prof. Susan Roaf and Prof. Roberto Lam-
berts for providing useful comments. We also thank the following organisations for their
cooperation: Gotoh Educational Corporation, Hulic Co. Ltd., Nikken Sekkei Ltd., IDEA
Consultants Inc., IS Logistic, Panasonic Corporation, PS Company Ltd., Tokyo City Uni-
versity, Tokyo City University Todoroki Junior and Senior High School, Tokyu Fudosan
Next Generation Engineering Center Inc. and Tsuzuki Ward. This research was supported
by Grant-in-Aid for Scientific Research (C) Number 17K06681.

References
ASHRAE Standard 55, 2004. Thermal environment conditions for human occupancy. Atlanta, Georgia:
American Society of Heating Refrigeration and Air- conditioning Engineers.
ASHRAE Standard 55, 2013. Thermal environment conditions for human occupancy. Atlanta, Georgia:
American Society of Heating Refrigeration and Air- conditioning Engineers.
Brager, G.S., & de Dear, R.J., 1998. Thermal adaptation in the built environment: A literature review.
Energy and Buildings, 27, pp. 83–96.
CBE, 2016. Mixed mode case studies and project database, Center for the Built Environment, Uni-
versity of California, Berkeley, https://cbe.berkeley.edu/mixedmode/ (Accessed: 20 March, 2020)
CIBSE, 2006. Environmental Design. CIBSE Guide A, Chapter 1, Environmental criteria for design. Lon-
don: Chartered Institution of Building Services Engineers.
Comité Européen de Normalisation (CEN), 2007. EN 15251: Indoor Environmental Input Parameters
for Design and Assessment of Energy Performance of Buildings Addressing Indoor Air Quality,
Thermal Environment, Lighting and Acoustics, CEN, Brussels.
de Dear, R.J., & Brager, G.S., 1998. Developing an adaptive model of thermal comfort and preference.
ASHRAE Transactions, 104(1), pp. 145–167.
Fanger, P.O., 1972. Thermal comfort, analysis and application in environmental engineering. Danish Technical
Press, Copenhagen.
Gautam, B., Rijal, H.B., Shukuya, M., & Imagawa H., 2019. A field investigation on the wintry ther-
mal comfort and clothing adjustment of residents in traditional Nepalese houses. Journal of Building
Engineering, 26, 100886.
Griffiths, I.D., 1990. Thermal comfort in buildings with passive solar features: Field studies. Report to
the Commission of the European Communities. EN3S- 090 UK: University of Surrey Guildford.
Hayashi, M., Yanagi, U., Azuma, K., Kagi, N., Ogata, M., Morimoto, S., Hayama, H., Mori, T.,
Kikuta, K., Tanabe, S., Kurabuchi, T., Yamada, H., Kobayashi, K., Kim, H., & Kaihara, N., 2020.
Measures against COVID‐19 concerning summer indoor environment in Japan. Japan Architectural
Review, 3(4), pp. 423– 434.
Hoyt, T., Arens, E., & Zhang, H., 2015. Extending air temperature setpoints: Simulated energy savings
and design considerations for new and retrofit buildings. Building and Environment, 88, pp. 89–96.

297
Hom B. Rijal et al.

Humphreys, M.A., 1970. A simple theoretical derivation of thermal comfort conditions. Journal of the
Institution of Heating & Ventilating Engineers, 38, pp. 95–98
Humphreys, M.A., 1978. Outdoor temperatures and comfort indoors. Building Research and Practice
( Journal of CIB), 6(2), pp. 92–105.
Humphreys, M.A, & Nicol, J.F., 1970. An investigation into thermal comfort of office workers. J. Inst.
Heat. Vent. Eng. ( JIHVE), 38, pp. 181–189.
Humphreys, M.A., & Nicol, J.F., 1995. An adaptive guideline for UK office temperatures. In: Nicol,
F., Humphreys, M., Sykes, O., & Roaf, S. ( Eds.), Standards for thermal comfort. London: E FN Spon,
Chapman & Hall, pp. 190–195.
Humphreys, M.A.,  & Nicol, J.F., 1998. Understanding the adaptive approach to thermal comfort.
ASHRAE Transactions, 104(1), pp. 991–1004.
Humphreys, M.A., Nicol, J.F., & Raja, I.A., 2007. Field studies of thermal comfort and the progress of
the adaptive model. Advances in Building Energy Research, 1, pp. 55– 88.
Humphreys, M.A., Nicol, J.F.,  & Roaf, S., 2016. Adaptive thermal comfort: Foundations and analysis.
Routledge, Abingdon.
Humphreys, M.A., Rijal, H.B., & Nicol, J.F., 2013. Updating the adaptive relation between climate and
comfort indoors; new insights and an extended database. Building and Environment, 63, pp. 40–55.
Imagawa, H., Rijal, H.B., & Shukuya, M., 2019. Study on modelling of the consciousness, behavior
and desired information of occupants in relation to energy use. Journal of Environmental Engineering,
AIJ, 84( 763), pp. 855– 864 (in Japanese with English abstract)
Jowkar, M., Rijal, H.B., Montazami, A., Brusey, J., & Temeljotov- Salaj, A., 2020. The influence of
acclimatization, age and gender-related differences on thermal perception in university buildings:
Case studies in Scotland and England. Building and Environment, 179, 106933.
KC, R., Rijal, H.B., Yoshida, K., & Shukuya, M., 2018. An in- situ study on occupants’ behaviors for
adaptive thermal comfort in a Japanese HEMS condominium. Journal of Building Engineering, 19,
pp. 402– 411.
Kurabuchi, T., Yanagi, U., Ogata, M., Otsuka, M., Kagi, N., Yamamoto, Y., Hayashi, M., & Tanabe,
S., 2021. Operation of air- conditioning and sanitary equipment for SARS- CoV-2 infectious dis-
ease control. Japan Architectural Review, 4(4), pp. 608–620.
Lau, K.K.L., & Fong, K.F. 2021. Editorial: Smart and healthy within the two- degree limit. Architectural
Science Review, 64:1–2, pp. 1– 4.
Majima, M., Umemiya, N., Yoshida, H., & Rijal, H.B., 2007. Thermal comfort of traditional narrow
alleys in an urban area: Survey for Nishijin district in Kyoto. Journal of Environmental Engineering,
AIJ, 622: pp. 41– 48.
McCartney, K.J., & Nicol, J.F., 2002. Developing an adaptive control algorithm for Europe. Energy and
Buildings, 34 (6), pp. 623– 635.
Nakashima, Y., 2013. Climate change policies in Japan / What are COOL BIZ and WARM BIZ?
Japan Environment Quarterly, 3, pp. 2–3, Ministry of the Environment.
Nicol, F., 2004. Adaptive thermal comfort standards in the hot–humid tropics. Energy and Buildings,
36( 7), pp. 628– 637.
Nicol, J.F., & Humphreys, M.A., 2004. A stochastic approach to thermal comfort – occupant behavior
and energy use in buildings. ASHRAE Transactions, 110(2), pp. 554–568.
Nicol, F., & Humphreys, M., 2010. Derivation of the adaptive equations for thermal comfort in free-
running buildings in European standard EN15251. Building and Environment, 45(1), pp. 11–17.
Nicol, F., Jamy, G.N., Sykes, O., Humphreys, M., Roaf, S., & Hancock, M., 1994. A survey of ther-
mal comfort in Pakistan toward new indoor temperature standards. Oxford England: Oxford Brookes
University.
Nicol, J.F., Raja, I.A., Alludin, A., & Tamy, G.N., 1999. Climatic variations in comfortable tempera-
tures: The Pakistan projects. Energy and Buildings, 30, pp. 261–279.
Rijal, H.B., Honjo, M., Kobayashi, R., & Nakaya, T., 2013. Investigation of comfort temperature,
adaptive model and the window opening behaviour in Japanese houses. Architectural Science Review,
56(1), pp. 54– 69.
Rijal, H.B., Humphreys, M.A., & Nicol, J.F., 2009. Understanding occupant behaviour: the use of
controls in mixed-mode office buildings. Building Research and Information, 37(4), pp. 381–396.
Rijal, H.B., Humphreys, M.A., & Nicol, J.F., 2017. Towards an adaptive model for thermal comfort in
Japanese offices. Building Research & Information, 45( 7), pp. 717–729.

298
 Adaptive approaches

Rijal, H.B., Humphreys, M.A.,  & Nicol, J.F., 2018. Development of a window opening algorithm
based on adaptive thermal comfort to predict occupant behavior in Japanese dwellings. Japan Archi-
tectural Review, 1(3), pp. 310–321.
Rijal, H.B., Humphreys, M.A., & Nicol, J.F., 2019a. Adaptive model and the adaptive mechanisms for
thermal comfort in Japanese dwellings. Energy and Buildings, 202, 109371.
Rijal, H.B., Humphreys, M.A., & Nicol, J.F., 2019b. Behavioural adaptation for the thermal comfort
and energy saving in Japanese offices. Journal of the Institute of Engineering, 15(2), pp. 14–25.
Rijal, H.B., Humphreys, M.A., & Nicol, J.F., 2020. Adaptive model and the adaptive mechanisms for
thermal comfort in Japanese offices, Proceedings of 11th Windsor Conference: Resilient Comfort , Cum-
berland Lodge, Windsor, UK, 16–19 April 2020. London: Network for Comfort and Energy Use
in Buildings, pp. 759–773, http://nceub.org.uk ( ISBN: 978-1-9161876-3- 4)
Rijal, H.B., Tuohy, P., Humphreys, M.A., Nicol, J.F., Samuel, A., & Clarke, J., 2007. Using results
from field surveys to predict the effect of open windows on thermal comfort and energy use in
buildings. Energy and Buildings, 39( 7), pp. 823– 836.
Rijal, H.B., Yoshida, K., Humphreys, M.A., & Nicol J.F., 2021. Development of an adaptive ther-
mal comfort model for energy- saving building design in Japan. Architectural Science Review, 64:1–2,
pp. 109–122.
Roaf, S.,  & Beckmann, K., 2017. Chapter  4 Monitoring climate change adaptation: Lessons from
Scotland, In: Karyono, T.H., Vale, R., & Vale, B. ( Eds.), Sustainable buildings and built environments to
mitigate climate change in the tropics - Conceptual and practical approaches, Springer Nature, Switzerland,
pp. 39–51.
SHASE ( The Society of Heating, Air- Conditioning and Sanitary Engineers of Japan), (AIJ) Architec-
tural Institute of Japan, 2020. Role of ventilation in the control of the COVID-19 infection: Emer-
gency presidential discourse. http://www.shasej.org/recommendation/covid-19/2020.03.23%20
shase_COVID20200323_eng.pdf
Tanabe, S., Iwahashi, Y., Tsushima, S., & Nishihara, N., 2013. Thermal comfort and productivity in
offices under mandatory electricity savings after the Great East Japan earthquake. Architectural Science
Review, 56(1), pp. 4–16.
Van Marken Lichtenbelt, W., Hanssen, M., Pallubinsky, H., Kingma, B., & Schellen, L., 2017. Healthy
excursions outside the thermal comfort zone. Building Research and Information, 45( 7), pp. 819– 827.
Yamamoto, M., Nishihara, N., Kawaguchi, G., Harigaya, J., & Tanabe, S., 2010. Comparison of the
transition of thermal environment in office between Japan and the US. Clima 2010 WellBeing Indoors
Proceedings.
Zhang, H., Arens, E., Kim, D.E., Buchberger, E., Bauman, F.,  & Huizenga, C., 2010. Comfort,
perceived air quality, and work performance in a low-power task– ambient conditioning system.
Building and Environment, 45, pp. 29–39.

299