Exploring Floor-Sitting as Adaptive Behavior in Tropical Apartment Residents: Regional and Indoor Climatic Influences in Indonesia

Abstract

In the tropical climates of Southeast Asia, the growing reliance on air conditioning (AC) for space cooling not only increases household energy consumption but may also diminish the role of culturally rooted adaptive behaviors such as floor-sitting. This study aims to explore the interaction between climatic factors, including regional and indoor climates, and thermally adaptive behaviors in Indonesian apartments, with a focus on floor-sitting. First, a large-scale questionnaire was conducted to analyze these interactions among different regional climates. Second, in-depth indoor climate measurements and a point-in-time questionnaire were conducted among the residents in the hotter regions. In the hotter regions like Jabodetabek (Jakarta metropolitan area) and Surabaya, floor-sitting was primarily conducted without using AC, often alongside fans in low-rise housing. In the cooler region of Bandung, floor-sitting was a common adaptive behavior with window openings in both high-rise and low-rise buildings. The in-depth measurement showed that low-rise buildings using higher thermal mass materials maintained stable indoor conditions for both air and floor temperatures even in the hotter region. The respondents could obtain coolness and remain thermally comfortable through a floor-sitting posture without using AC, especially when air and floor temperatures were both less than 31 °C. These results demonstrated that floor-sitting is a vital behavior that adapts to regional and indoor climatic conditions in the tropics while achieving thermal comfort and relying less on AC devices.

1. Introduction

1.1. Building Typology, Energy Use, and Adaptive Behavior in Southeast Asia

In Southeast Asia, populations are continuously increasing with economic growth, forming rapid urbanization in large cities. Owing to urbanization, building typologies are currently experiencing a significant transition, shifting from predominantly landed to vertical housing. In Indonesia, vertical housing generally consists of low-rise and high-rise multi-story apartments, which are increasingly promoted as solutions to rapid urbanization, limited land availability, population growth, and the demand for affordable housing in large cities. The earliest form of public vertical housing—Rumah Susun (walk-up or multi-story flats)—was introduced in the 1980s through government-supported housing programs [1,2]. Among them, Rusunawa (Rumah Susun Sewa) consists of rental units fully owned and managed by the government for eligible low-income households. Rusunami (Rumah Susun Sederhana Milik) represents a specific category of simple/low-cost ownership flats that can be purchased by low-income groups, making it an attractive option for middle-income groups who are seeking affordable and livable housing. Rusunami was first built in the 2000s [2], and the number of high-rise apartments has continued to increase over the last two decades. This transformation in housing form is closely tied to changes in residents’ lifestyles [3,4,5], thermal adaptation practices (e.g., window opening and air conditioning usage) [6], and building energy demand [7,8].

Such economic growth and urbanization also led to cultural shifts from traditional to modern (i.e., Western) lifestyles for residents in Southeast Asia. Western influences during this transition often change the way people behave in their houses and cause an increase in household energy consumption, with a greater dependence on energy-consuming appliances [9]. Energy demand projections highlight the importance of this trend. According to Latha et al. [10], Indonesia’s annual electricity consumption is expected to rise sharply between 2000 and 2030, mainly driven by the rapid growth of residential sectors. Among household appliances, air conditioner (AC) ownership is projected to increase substantially, making it one of the most influential factors shaping household electricity consumption and its peak demand [11,12,13].

In residential environments, especially in tropical climates like Southeast Asia, occupant adaptive behavior plays a critical role in forming thermal comfort and energy usage patterns [14]. While much attention has been given to architectural and mechanical solutions for indoor climate control, the adaptive behaviors of occupants—particularly those related to posture and spatial use—remain underexplored. In Indonesia, floor-sitting is known as a culturally embedded practice, commonly performed throughout the day for activities such as eating, studying, relaxing, and gathering [15]. This means that floor-sitting has been discussed mainly in view of adaptive behaviors influenced by local and regional cultures [16,17]. Beyond its cultural significance, floor-sitting may also serve as a thermally adaptive behavior, providing increased contact with cooler surfaces [18,19] and promoting comfort in tropical climates without relying on AC usage [20]. Erwindi et al. (2025) revealed that the reasons for preferring floor-sitting included comfort, coolness, and habit [18]. However, the growing reliance on active thermal adaptation strategies, particularly AC, not only increases household energy consumption but may also diminish the role of culturally rooted adaptive behaviors such as floor-sitting. They have rarely been studied and should be explored further.

1.2. Regional and Indoor Climatic Factors

To discuss thermally adaptive behaviors, it is important to fully consider the climatic contexts, including regional climate and indoor climate. Indonesia is known as a tropical country; however, there are different climatic zones owing to the country’s vast size and complex terrain. Putra et al. [21] developed climate zones for determining appropriate passive cooling techniques in Indonesia and classified the entire Indonesian territory into eight climate zones. In particular, there are mainly two climate categories: hotter regions among lowland areas (e.g., Jakarta and Surabaya) and cooler regions among highland areas (e.g., Bandung and Bogor). Mori et al. (2020) indicated the differences in window-opening, AC usage, and fan usage patterns between hotter and cooler regions in Indonesia [22].

In addition, building types, design, and materials are dominant factors influencing indoor climates and should therefore be considered when discussing thermally adaptive behaviors. In Indonesia, low-rise apartments (≤5 floors) mainly apply conventional reinforced concrete with high thermal mass for the main structure and have a single-loaded corridor. In contrast, medium- and high-rise buildings (>5 floors) apply precast concrete with low thermal mass and a double-loaded corridor [23]. Alfata et al. (2015) revealed that different building types affected indoor climate and residents’ thermal comfort [24].

1.3. Research Questions and Gap

Erwindi et al. (2025) indicated that behavioral adaptations with floor-sitting remain viable in achieving thermal comfort through survey results in Jabodetabek (Jakarta metropolitan area), Indonesia [18]. In the region, thermal comfort across different building types was mainly reported as comfortable by residents in their units while showing floor-sitting behavior, even for respondents who did not own AC. Accordingly, this raises the question of how regional and indoor climates form thermally adaptive behaviors in tropical apartment residents, with a particular focus on floor-sitting. Empirical evidence is still limited on how floor-sitting, as a culturally rooted habit, relates to thermal adaptations and comfort across different regional and indoor climates. Sitting posture is an understudied form of thermal adaptation [14], and the combined effects of regional and indoor climates on such behavior have seldom been investigated in depth.

1.4. Research Objectives

This study explores the interaction between climatic factors, including regional and indoor climates, and thermally adaptive behaviors in Indonesian apartments. In particular, it investigates the role of floor-sitting practices as a thermal adaptation strategy and compares their occurrence with other adaptation methods—such as AC and fan usage—across different climates and building types. By analyzing large-scale questionnaire data alongside in-depth indoor climate measurements, this study identifies key factors that influence thermally adaptive behavior in both low-rise and high-rise apartment settings. Further, the paper is structured in two parts, with the first part presenting a comparative analysis among different regional climates using a large-scale questionnaire to examine regional differences in floor-usage behavior, thermal adaptation strategies, thermal comfort perception, and household energy consumption. The second part involves in-depth indoor climate measurements, including air and floor temperatures, and a point-in-time questionnaire to explore the relationship between building types, indoor climates, floor-usage behavior, thermal adaptations, and thermal comfort. Through this combined approach, the study will provide insights into how floor-sitting as a traditional practice can support sustainable living in modern housing environments while also addressing broader questions of energy-saving and occupant-centered design in tropical contexts.

2. Methodology

2.1. Large-Scale Questionnaire Survey

2.1.1. Survey Design

This study used a cross-sectional and nationwide questionnaire survey to investigate floor-usage behaviors, thermal adaptation methods, thermal comfort, and household energy consumption across five cities—Bandung, Jabodetabek, Surabaya, Makassar, and Medan. The questionnaire was structured to document naturally occurring practices in daily life, mainly focusing on floor-sitting behaviors in these cities [25,26]. This method provided systematic data collection and qualitative insights, offering a comprehensive view of behavioral adaptation to climate and housing conditions [27,28].

Respondents represented a broad range of educational and income levels from low- to high-income households. With prior approval from building managers, trained surveyors from a survey company (IPSOS Indonesia) visited residential units, explained the survey purpose, and obtained informed consent. The questionnaire was conducted face-to-face, ensuring clarity and reliability in responses, and respondents received an honorarium in recognition of their contribution.

Table 1. Large-scale survey questions.

Main categories	Sub-categories (Number of questions)
I. General	Basic Information (4)	Monthly income and expenses (1)
	Building information (1)	Weekly working time (3)
II. Energy—Lifestyle	Thermal adaptation during activity (AC and Fan usage schedule) (5)	Electronic appliances owned (2)
	Hot water usage (3)	Energy-saving action and passive behavior indicator (2)
	Lifestyle: weekend and weekdays (2)	Lifestyle and living environment—Including contact cooling (4)
III. Activity	Daily activity area during: mealtime, gathering, studying, relaxing, sleeping (5)
IV. Foot and floor cover	Floor cover during daily activity area (surface cover) (5)	Foot cover during daily activity (5)
V. Posture and preferences	Common sitting posture (2)	Common laying posture (2)
	Common sleeping posture (2)	Preference (like/dislike) sitting, lying, sleeping on the floor, and reason (3)
	Common surface for sitting, lying, and sleeping posture (6)	Frequency of sitting, lying, or sleeping on the floor (6)
	Common area for sitting, lying, and sleeping posture (6)	Psychological factors: thermal comfort, satisfaction, sensation (6)
VI. Body parts in contact with the floor	Body parts in contact during daily activity (5)	Preference (like/dislike) on foot sole floor contact (4)
	Common clothing during contact with the floor (2)

lists the categories and question items [18]. The questionnaire comprised 86 questions organized into six main categories and 24 sub-categories of questions. Category I covers individual and residential information. Category II addresses energy usage in daily activities, including AC use, residents’ lifestyle patterns, and psychological factors. Category III includes questions related to daily activities such as gathering, mealtime, studying, relaxing, and sleeping in their units. Category IV covers questions about foot coverings and floor-covering types at daily activity areas. Category V includes questions on postures during daily activities and thermal comfort levels. Category VI includes questions aimed at identifying specific body parts in direct contact with the floor and preferences for floor contact during the above-mentioned daily activities.

2.1.2. Building Typology

In the Indonesian context, vertical housing is formally classified into Rusunawa, Rusunami, and condominiums. Rusunawa are government-owned rental apartments provided for low-income households. Rusunami refers to ownership-based housing mainly for low- and middle-income groups subsidized by government assistance [29]. Condominiums, on the other hand, are privately developed high-end residences primarily accommodating high-income residents [30]. Beyond this economic distinction, vertical residences are also differentiated into high-rise and low-rise developments based on their building characteristics, which depend on their structural scale and form [23]. Based on the building design and materials, as described in Section 1.2, the classification of Low-rise (≤5 floors) and High-rise (>5 floors) was used for the study analysis. Rusunawa is predominantly Low-rise, following the government design guidelines. Rusunami ranges from Low-rise to High-rise buildings. A condominium is mainly characterized as High-rise. Further, building types are then classified into Low-rise and High-rise based on the survey results. In contrast, socio-economic factors are examined separately by grouping respondents into different income levels.

2.1.3. Sampling Strategy and Data Collection Process

As of 2023, there are 60,511 subsidized apartment units in Indonesia, comprising both Rusunawa and Rusunami [31]. The total number of housing units was treated as the finite population size (N) for the purpose of sample size determination. The adequacy of this sample size was evaluated by calculating the margin of error (MoE) at a 95% confidence level ($Mo{E}_{95}$), incorporating a finite population correction to account for the limited population size [32]. The margin of error was calculated as follows:

$$Mo{E}_{95}= z_{0.95}\sqrt{{\displaystyle \frac{{\mathsf{\sigma }}_p^2}{n}}}\sqrt{{\displaystyle \frac{N-n}{N-1}}}$$

where N is the total population size (60,511 housing units), n is the defined sample size, ${\sigma }_p^2$ is the variance of the population proportion with a conservative maximum value of 0.25 (p = 0.5), and $z_{0.95}$ = 1.96 corresponds to the 95% confidence level. Using this formulation, the margin of error for the present survey was estimated to be ±1.81%, indicating that the survey estimates are expected to fall within ±1.81 percentage points of the true population values for the subsidized apartment housing population. With this method, an initial target sample size of 3000 respondents was defined for the representativeness of residents living in Rusunawa and Rusunami in Indonesia. In addition, 800 respondents living in condominiums were considered as a supplementary sample for comparative purposes; these condominium samples were not included in the finite population used for margin-of-error estimation.

Based on the 2010 national census, Indonesia contained ten metropolitan areas with populations exceeding one million residents. The present study selected five representative cities among these metropolitan areas that capture distinct climatic conditions, urban structures, and housing characteristics. The surveyed cities were Bandung (2.5 million), Jabodetabek (10.6 million), Surabaya (2.9 million), Medan (2.4 million), and Makassar (1.4 million). Taking into account the relative population sizes and the distribution of apartment housing across the selected cities, the total target sample of 3800 respondents was allocated as follows: 1800 for Jakarta, 600 each for Surabaya and Bandung, and 400 each for Medan and Makassar, thereby ensuring sufficient representation of diverse urban contexts [18].

To discuss regional climatic factors on thermally adaptive behaviors, these cities were classified into two groups: the cooler region (i.e., Bandung: highland tropical climate) and hotter regions (i.e., Jabodetabek: Sub-monsoonal, Surabaya: Sub-savanna, Medan: Equatorial, and Makassar: Monsoonal) based on the Indonesian climatic zones for passive cooling by Putra et al. [21] (

Table 2. Samples for the large-scale questionnaire survey.

Climate Categories	City	Detailed Climatic Zones [21]	Annual Average Temperatures [°C], Higher (75 Percentile), Lower (25 Percentile)	Building Type (Number of Samples)
Hotter regions	Jabodetabek	Sub-monsoonal (3B) (Jakarta)	28.8, 30.5, 27.0	High-rise (871) Low-rise (962)
	Surabaya	Sub-savanna (4B)	28.3, 30.3, 26.0	High-rise (22) Low-rise (352)
	Medan	Equatorial (1A)	27.5, 30.0, 25.1	High-rise (1) Low-rise (142)
	Makassar	Sub-monsoonal (3B)	28.3, 30.2, 26.3	High-rise (0) Low-rise (292)
Cooler region	Bandung	Highland tropical (2A)	26.2, 28.5, 24.0	High-rise (174) Low-rise (315)

). Based on typical meteorological year data by the Meteorology, Climatology, and Geophysical Agency of Indonesia (Badan Meteorologi, Klimatologi, dan Geofisika) [33], outdoor temperatures are cooler in Bandung, with an annual average temperature of 26 °C. Conversely, Jabodetabek, Surabaya, Medan, and Makassar show higher outdoor temperature ranges with annual average temperatures of around 28–29 °C.

Data collection was conducted by the survey company between September and November 2022. A probability-based random sampling approach targeting residents of apartment-type housing, including Rusunawa, Rusunami, and condominiums, was employed to ensure balanced representation across gender, age, and ethnic groups. In total, 3383 apartment residents in the selected cities completed the questionnaire. The sample distribution included Rusunawa (48%), Rusunami (41%), and condominiums (11%). Building height or floor-level distribution was not employed as a sampling criterion during data collection. Instead, information on the total number of floors was recorded during the survey and subsequently used as an analytical variable. Consequently, for analytical purposes, the sample was classified into Low-rise (63%) and High-rise (37%). No post-survey weighting was applied, as the primary objective of the sampling design was to ensure representativeness of apartment housing residents as a whole, rather than representativeness by building height information. During data processing, several responses contained missing information regarding the unit floor level. Therefore, after excluding these cases, 3131 valid samples were retained for the final analysis. Household participation included 24% single-member, 44% two-member, 19% three-member, and 13% multi-member (four or more) households, with all respondents aged 18 years or older. A detailed demographic distribution is presented in

Table A1. Description of building categories.

Variable	Overall	High-rise	Low-rise
	N = 3131 1	N = 1068 1	N = 2063 1
Age	38 (10)	36 (11)	38 (10)
Age category
+55	199, (6.4%)	57, (5.3%)	142, (6.9%)
10s	76, (2.4%)	33, (3.1%)	43, (2.1%)
20–24	294, (9.4%)	143, (13%)	151, (7.3%)
25–29	436, (14%)	163, (15%)	273, (13%)
30–34	444, (14%)	141, (13%)	303, (15%)
35–39	488, (16%)	148, (14%)	340, (16%)
40–44	452, (14%)	145, (14%)	307, (15%)
45–49	381, (12%)	113, (11%)	268, (13%)
50–54	361, (12%)	125, (12%)	236, (11%)
Gender
Female	2258, (72%)	676, (63%)	1582, (77%)
Male	873, (28%)	392, (37%)	481, (23%)
Number of family members living together
1	740, (24%)	427, (40%)	313, (15%)
2	1384, (44%)	437, (41%)	947, (46%)
3	588, (19%)	127, (12%)	461, (22%)
4	288, (9.2%)	54, (5.1%)	234, (11%)
5	94, (3.0%)	16, (1.5%)	78, (3.8%)
6	32, (1.0%)	7, (0.7%)	25, (1.2%)
7	5, (0.2%)	0, (0%)	5, (0.2%)
Occupation category
Army/police	9, (0.3%)	6, (0.6%)	3, (0.1%)
Business owner/Executive level or above	32, (1.0%)	31, (2.9%)	1, (<0.1%)
Educator	47, (1.5%)	16, (1.5%)	31, (1.5%)
Entrepreneur/Store owner	613, (20%)	332, (31%)	281, (14%)
Housewife with a side job	258, (8.2%)	52, (4.9%)	206, (10.0%)
Housewife without a side job	968, (31%)	150, (14%)	818, (40%)
Laborer (farmer, stone mason, maid)	90, (2.9%)	5, (0.5%)	85, (4.1%)
Laborer with a license (driver, mechanic, carpenter)	138, (4.4%)	16, (1.5%)	122, (5.9%)
Normal employee	766, (24%)	371, (35%)	395, (19%)
Not working/unable to search for work	26, (0.8%)	2, (0.2%)	24, (1.2%)
Professional	39, (1.2%)	28, (2.6%)	11, (0.5%)
Retiree	32, (1.0%)	8, (0.7%)	24, (1.2%)
Student	83, (2.7%)	44, (4.1%)	39, (1.9%)
Unable to work	6, (0.2%)	1, (<0.1%)	5, (0.2%)
Workers with a license (nurse)	24, (0.8%)	6, (0.6%)	18, (0.9%)
Building structure
Ceramic	1, (<0.1%)	1, (<0.1%)	0, (0%)
Concrete brick	6, (0.2%)	0, (0%)	6, (0.3%)
Galvalume	19, (0.6%)	0, (0%)	19, (0.9%)
Reinforced concrete frame	2859, (91%)	887, (83%)	1972, (96%)
Steel frame	240, (7.7%)	180, (17%)	60, (2.9%)
Wall	1, (<0.1%)	0, (0%)	1, (<0.1%)
Wood	5, (0.2%)	0, (0%)	5, (0.2%)
Level of the unit	5.0 (5.1)	9.7 (6.4)	2.6 (1.3)
Level of the building	9.9 (8.1)	20.5 (5.2)	4.5 (0.6)
Unit type
Family 1 bedroom (1BR)	876, (28%)	158, (15%)	718, (35%)
Family 2 bedrooms (2BR)	1488, (48%)	736, (69%)	752, (36%)
Family 3 bedrooms (3BR)	11, (0.4%)	4, (0.4%)	7, (0.3%)
Studio	756, (24%)	170, (16%)	586, (28%)
Area for gathering
Bed room	59, (1.9%)	40, (3.8%)	19, (0.9%)
Corridor/Hallway	1, (<0.1%)	0, (0%)	1, (<0.1%)
Dining room	21, (0.7%)	11, (1.0%)	10, (0.5%)
Dining room; Corridor/Hallway	1, (<0.1%)	0, (0%)	1, (<0.1%)
Kitchen/Pantry	1, (<0.1%)	1, (<0.1%)	0, (0%)
Kitchen/Pantry; Bedroom	3, (<0.1%)	3, (0.3%)	0, (0%)
Living room	2860, (92%)	947, (89%)	1913, (93%)
Living room; Bedroom	42, (1.3%)	14, (1.3%)	28, (1.4%)
Living room; Corridor/Hallway	3, (<0.1%)	0, (0%)	3, (0.1%)
Living room; Corridor/Hallway; Outdoor area	1, (<0.1%)	0, (0%)	1, (<0.1%)
Living room; Dining room	13, (0.4%)	9, (0.8%)	4, (0.2%)
Living room; Dining room; Bedroom	1, (<0.1%)	1, (<0.1%)	0, (0%)
Living room; Dining room; Corridor/Hallway; Kitchen/Pantry; Bed room	1, (<0.1%)	1, (<0.1%)	0, (0%)
Living room; Kitchen/Pantry	2, (<0.1%)	2, (0.2%)	0, (0%)
Living room; Outdoor area	28, (0.9%)	21, (2.0%)	7, (0.3%)
Living room; Outdoor area; Bedroom	1, (<0.1%)	1, (<0.1%)	0, (0%)
Living room; Terrace/Veranda/Balcony	20, (0.6%)	1, (<0.1%)	19, (0.9%)
Living room; Terrace/Veranda/Balcony; Bed room	3, (<0.1%)	0, (0%)	3, (0.1%)
Living room; Terrace/Veranda/Balcony; Corridor/Hallway	2, (<0.1%)	0, (0%)	2, (<0.1%)
Living room; Terrace/Veranda/Balcony; Corridor/Hallway; Outdoor area	1, (<0.1%)	0, (0%)	1, (<0.1%)
Living room; Terrace/Veranda/ Balcony; Outdoor area	2, (<0.1%)	0, (0%)	2, (<0.1%)
Outdoor area	9, (0.3%)	8, (0.8%)	1, (<0.1%)
Outdoor area; Bedroom	2, (<0.1%)	2, (0.2%)	0, (0%)
Terrace/Veranda/Balcony	45, (1.4%)	3, (0.3%)	42, (2.0%)
Terrace/Veranda/Balcony; Corridor/Hallway	1, (<0.1%)	0, (0%)	1, (<0.1%)
Total energy consumption [GJ/year]	12 (7)	16 (7)	9 (5)
Electricity energy consumption [GJ/year]	7.6 (5.5)	11.3 (5.8)	5.7 (4.1)
Gas energy consumption [GJ/year]	4.29 (2.70)	4.96 (3.22)	3.92 (2.29)
House type
Condominium	341, (11%)	314, (29%)	27, (1.3%)
Rusunami	1278, (41%)	625, (59%)	653, (32%)
Rusunawa	1512, (48%)	129, (12%)	1383, (67%)
Income
Under Rp.900,000	34, (1.1%)	6, (0.6%)	28, (1.4%)
Rp.900,001–Rp.1,250,000	63, (2.0%)	8, (0.8%)	55, (2.7%)
Rp.1,250,001–Rp.2,500,000	497, (16%)	20, (1.9%)	477, (23%)
Rp.2,500,001–Rp.4,000,000	1066, (34%)	129, (12%)	937, (46%)
Rp.4,000,001–Rp.7,500,000	749, (24%)	295, (28%)	454, (22%)
Rp.7,500,001–Rp.12,000,000	437, (14%)	359, (34%)	78, (3.8%)
Rp.12,000,001–Rp.20,000,000	198, (6.4%)	186, (18%)	12, (0.6%)
Rp.20,000,001–Rp.40,000,000	51, (1.6%)	48, (4.5%)	3, (0.1%)
Over Rp.40,000,001	4, (0.1%)	4, (0.4%)	0, (0%)
Preference for sitting on the floor
Dislike sitting on the floor	1039, (33%)	649, (61%)	390, (19%)
Like sitting on a floor	1991, (64%)	374, (35%)	1617, (78%)
Not sure	101, (3.2%)	45, (4.2%)	56, (2.7%)
Thermal sensation in the daytime
Hot	49, (1.6%)	3, (0.3%)	46, (2.2%)
Warm	734, (23%)	147, (14%)	587, (28%)
Slightly warm	276, (8.8%)	84, (7.9%)	192, (9.3%)
Neutral	1255, (40%)	425, (40%)	830, (40%)
Slightly cool	664, (21%)	343, (32%)	321, (16%)
Cool	143, (4.6%)	61, (5.7%)	82, (4.0%)
Cold	10, (0.3%)	5, (0.5%)	5, (0.2%)
Thermal sensation at nighttime
Hot	8, (0.3%)	2, (0.2%)	6, (0.3%)
Warm	239, (7.6%)	24, (2.2%)	215, (10%)
Slightly warm	180, (5.7%)	28, (2.6%)	152, (7.4%)
Neutral	964, (31%)	298, (28%)	666, (32%)
Slightly cool	1106, (35%)	508, (48%)	598, (29%)
Cool	524, (17%)	159, (15%)	365, (18%)
Cold	110, (3.5%)	49, (4.6%)	61, (3.0%)
Thermal comfort in the daytime
Very uncomfortable	34, (1.1%)	4, (0.4%)	30, (1.5%)
Uncomfortable	349, (11%)	70, (6.6%)	279, (14%)
Slightly uncomfortable	493, (16%)	115, (11%)	378, (18%)
Comfortable	2255, (72%)	879, (82%)	1376, (67%)
Thermal comfort at nighttime
Very uncomfortable	18, (0.6%)	2, (0.2%)	16, (0.8%)
Uncomfortable	104, (3.3%)	21, (2.0%)	83, (4.0%)
Slightly uncomfortable	340, (11%)	83, (7.8%)	257, (12%)
Comfortable	2669, (85%)	962, (90%)	1707, (83%)
Thermal satisfaction in the daytime
Very unsatisfied	45, (1.4%)	8, (0.7%)	37, (1.8%)
Unsatisfied	153, (4.9%)	20, (1.9%)	133, (6.4%)
Slightly unsatisfied	234, (7.5%)	36, (3.4%)	198, (9.6%)
Neither of them	154, (4.9%)	36, (3.4%)	118, (5.7%)
Slightly satisfied	505, (16%)	193, (18%)	312, (15%)
Satisfied	1952, (62%)	725, (68%)	1227, (59%)
Very satisfied	88, (2.8%)	50, (4.7%)	38, (1.8%)
Thermal satisfaction at nighttime
Very unsatisfied	92, (2.9%)	21, (2.0%)	71, (3.4%)
Unsatisfied	56, (1.8%)	12, (1.1%)	44, (2.1%)
Slightly unsatisfied	140, (4.5%)	36, (3.4%)	104, (5.0%)
Neither of them	113, (3.6%)	32, (3.0%)	81, (3.9%)
Slightly satisfied	374, (12%)	83, (7.8%)	291, (14%)
Satisfied	2227, (71%)	832, (78%)	1395, (68%)
Very satisfied	129, (4.1%)	52, (4.9%)	77, (3.7%)

¹ Mean (SD); N, (%).

(Appendix A).

2.1.4. Estimation of Annual Household Electricity Consumption

Annual household electricity consumption was quantified in units of gigajoules per year (GJ/year) based on monthly energy use data collected through a structured questionnaire survey. Respondents were primarily asked to report on their household electricity consumption using official utility billing records, covering at least the preceding 12 months, where available. When billing records were unavailable, respondents were requested to provide monthly electricity consumption values or payment amounts based on their recollection. Monthly electricity consumption values originally reported in kilowatt-hours (kWh) were converted into energy units using a standard conversion factor (1 kWh = 3.6 MJ) and subsequently aggregated to annual electricity consumption in gigajoules per year (GJ/year). The reported electricity consumption represents total household electricity use, including but not limited to AC, lighting, appliances, and other electrical loads. AC-specific electricity consumption was not isolated in this study. The detailed procedure for this estimation is shown in Appendix B,

Table A2. Electricity using charge for each contract power in 2022 [60].

Contract Power [VA]	Usage Charge [IDR/kWh]
900	1352.00
1300, 2200	1444.70
3500–5500, 6500~	1699.53

,

Table A3. Tax rates of public street light by city [49].

City	Public Street Light Tax Rates [%]
Jakarta	3
Surabaya	8
Bandung	6
Makassar	10
Medan	7

and

Table A4. Stamp fee by electricity bill [49].

Electricity Bill [IDR]	Stamp Cost [IDR]
x ≤ 200,000	0
200,000 < x ≤ 1,000,000	3000
1,000,000 < x	6000

.

2.1.5. Statistical Analysis

To ensure data reliability for the annual electricity consumption, outliers were screened using standardized Z scores, and only observations with absolute Z scores of 3 or less were retained for the final analysis (n = 2932). To check the statistical differences in the questionnaire results between regional climates and building types, a statistical test was conducted. Comparative variables were classified into two or three categories: AC usage vs. Non-AC in thermal adaptation, Floor-sitting vs. Non-floor-sitting in main posture, Cool-side vs. Neutral vs. Warm-side sensations in daytime and nighttime thermal sensation, and Comfortable-side vs. Uncomfortable-side sensations in daytime and nighttime thermal comfort. Due to unequal sample sizes and small expected counts in some categories, the Fisher–Freeman–Halton exact test was applied instead of the chi-square test. Fisher–Freeman–Halton tests with Monte Carlo simulation (B = 100,000) were conducted to perform pairwise comparisons of occupant behaviors and thermal responses between high-rise and low-rise buildings in both hotter and cooler regions (seed = 123). To account for multiple comparisons, the Benjamini–Hochberg adjustment (BH-adjusted) was applied in addition to raw p-values. This approach allowed us to identify statistically significant differences in behavioral and thermal adaptation patterns across building types and climatic contexts. The stats package (ver. 4.5.0) of R version 4.5.0 was employed for all of the statistics.

2.2. In-Depth Measurement

2.2.1. Details of the Measurement and Questionnaire

The second part of this study comprises indoor climate measurements and a questionnaire survey to explore the relationship between building types, indoor climates, floor-usage behavior, thermal adaptations, and thermal comfort experienced during respondents’ daily activities. Our sample size was limited; however, these relationships can be discussed through detailed on-site measurement and a point-in-time questionnaire. The survey was conducted in Surabaya, selected as a sample location in a hotter region, from late July to mid-September 2023.

Data collection was carried out using two complementary approaches. First, on-site measurements were conducted to capture indoor climate conditions within the respondents’ dwellings, including indoor temperature and floor temperature. Second, point-in-time questionnaire surveys were administered online to capture respondents’ floor-usage behaviors, postures, thermal adaptations, thermal comfort, thermal sensation, and temperature demand during floor contact.

For indoor climates, air temperature and floor surface temperature were monitored using thermistor sensors with data loggers (RTR-503B, T&D Corporation, Matsumoto, Japan) and button-type temperature sensors (iButton, Maxim Integrated, San Jose, CA, USA), respectively. Prior to field deployment, all sensors were cross-checked during pilot measurement under controlled indoor conditions to verify consistency and to minimize inter-sensor bias. Air temperature sensors were installed at a height of 110 cm above the floor, corresponding approximately to the breathing zone for seated occupants, which is commonly adopted in thermal comfort field studies (

Table 3. Measurement equipment list.

Measured Items	Equipment	Accuracy and Resolution	Installment Location
Indoor air temperature, outdoor air temperature, humidity	RTR503B, T&D Corporation, Matsumoto, Japan	Measurement accuracy: ±0.5 °C Resolution: 0.1 °C	Measured at 110 cm height, in the center of the living room (indoor measurement), balcony (outdoor measurement)
Floor temperature	iButton, Thermochron, Analog Devices, San Jose, CA, USA	Measurement accuracy: ±0.5 °C Resolution: 0.0625 °C	Living room area, 4 points at the main daily sitting posture activity area

). The sensors were suspended using a non-conductive holder (camera tripod with non-conductive covering) and positioned away from direct solar radiation, electrical appliances, and immediate airflow sources like fans (Figure 1). Radiation shields were used for measuring indoor air temperature, and care was taken to avoid exposure to direct sunlight through windows.

Floor surface temperature was measured using iButton sensors. It was selected as the sensor due to its very small size (thickness < 0.01 m) and wireless operation, allowing it to be installed discreetly. Covered with aluminum tape, the sensors were placed in direct contact with the finished floor surface at locations frequently used for sitting by the residents.

All sensors recorded data at 5-min intervals, which provides sufficient temporal resolution to capture short-term indoor thermal fluctuations and occupant adaptive actions. Sensor clocks were synchronized prior to installation to ensure temporal alignment across datasets. Data collection was continuous throughout the measurement period. During data processing, missing data were neglected for the analysis. Because outlier values were minimal, they were included in the analysis.

Measurement locations within each dwelling were selected to represent typical living zones rather than boundary conditions. Sensors were generally placed in the main living area, avoiding immediate proximity to windows, exterior walls, or cooling devices unless these features characterized the primary occupied zone. (Appendix C,

Table A5. Details of the in-depth measurement unit plan and equipment settings.

Building	Unit	Unit Plans and Equipment Settings
HRA	HRA-1
HRB	HRB-1
	HRB-2
	HRB-3
LRA	LRA-1
	LRA-2
LRB	LRB-1
	LRB-2
	LRB-3

Legend:  Temperature and Humidity; Floor temperature.

). This setup ensured minimal intrusion, enabling continuous temperature recording without interrupting the daily activities of residents inside the unit (Figure 1).

To capture variations throughout the day, the questionnaire survey was divided into four time periods: morning (05:00–10:30), noon (10:30–14:30), afternoon (14:30–18:00), and night (18:00–24:00) (see Figure 2). This allowed for the assessment of daily thermal adaptation patterns in relation to floor usage. For each respondent, the survey was conducted for six days with continuous indoor climate measurements. The respondents were asked to conduct daily behaviors and postures (e.g., sitting on a floor, sitting on a chair/sofa, and standing) at their selected location in the unit during the four time periods while answering the questions. As shown in

Table 4. Question items in the point-in-time questionnaire.

Categories	Question items (Number of questions)
I. General	Basic Information (6)	Health condition (1)
II. Clothing condition	Description of clothing during experiment session (1)
III. Foot sole and floor cover condition	Description of foot cover during experiment session (1)	Description of floor cover during experiment session (1)
IV. Thermal adaptation	Description of thermal adaptation operating during the experiment session (1)
V. Activity and place	Description of activity during floor contact experiment session (1)	Description of location during floor contact experiment session (1)
VI. Posture and preferences	Main floor contact posture during experiment (1)	Current floor contact posture during experiment (1)
VII. Body parts in contact with the floor	Specify body parts in contact during experiment (1)
VIII. Thermal comfort	Floor contact thermal comfort (1)	Whole-body thermal comfort (1)
IX. Thermal sensation	Floor contact thermal sensation (1)	Whole-body thermal sensation (1)
X. Thermal control	Floor temperature control possibility (1)	Indoor air temperature control possibility (1)

, the respondents were asked to answer 24 questions in 10 categories. The 10 categories include general questions, clothing condition, foot sole and floor cover condition, thermal adaptation, activity and place, posture and preference, body parts in contact with the floor, thermal comfort (whole body and body parts that contact with the floor), thermal sensation (whole body and body parts that contact with the floor), and thermal control. Each questionnaire required approximately 10 min to complete across time periods.

In this stage of the study, respondents were selected based on their housing type (High-rise and Low-rise) and AC-usage behaviors (AC usage and Non-AC) to maintain consistency with the large-scale questionnaire, which denotes actual AC operation during specific daily activities as reported. The survey was designed to record thermal adaptation methods, including AC usage and main sitting posture, separately for five daily activities: sleeping, studying or working, resting and relaxing, mealtime, and gathering. Based on these responses, AC usage during gathering was defined as a binary variable indicating whether AC was operated during shared living activities. The gathering activity was selected for focused analysis, similar to the large-scale questionnaire analysis, because it occurred in common sharing spaces where posture-related behaviors and thermal adaptation strategies were most clearly expressed across households.

Consistent with this activity-based definition and with the categorization applied in the large-scale questionnaire, a total of ten respondents for the in-depth measurements were selected, representing contrasting building and adaptation conditions: (i) High-rise Non-AC (Rusunami units without AC ownership or usage, n = 2), (ii) High-rise AC (Rusunami units with AC ownership and usage, n = 3), and (iii) Low-rise Non-AC (Rusunawa units without AC ownership or usage, n = 5) (

Table 5. Information on units and respondents.

Building Characteristics	High-Rise		Low-Rise
Thermal adaptation (number of units)	Non-AC (1)	AC (3)	Non-AC (5)
Building and Unit ID (number)	HRA-1	HRB-1 HRB-2 HRB-3	LRA-1 LRA-2	LRB-1 LRB-2 LRB-3
Floor levels (unit orientations)	7 (NE)	12 (E), 19 (S), 16 (S)	2 (W), 2 (W)	3 (E), 2 (W), 1 (S)
Structure	Reinforced concrete	Reinforced concrete	Reinforced concrete	Reinforced concrete
External wall material	Lightweight concrete	Lightweight concrete	Reinforced concrete	Reinforced concrete
Floor material	Ceramic tile	Ceramic tile	Ceramic tile	Ceramic tile
Floor cover	Without carpet	With carpet (partial respondent)	Without carpet	Without carpet
Corridor	Inside	Inside	Inside	Outside
Shading on windows	Balcony (partial)	Balcony (partial)	Overhang	Outside corridor
Window to wall ratio (WWR) **	27%	23%	32%	14%
Respondents (age)	Female (40s), Male (20s)	Female (20s), Female (20s), Female (20s)	Female (40s), Female (50s)	Female (20s), Female (50s), Male (50s)
Household composition	Family of 5 (Parents and 3 young adults)	Single (2), Living with younger brother (1)	Family of 4 (Parents with 2 kids) (2)	Family of 4 (Parents with 2 kids) (1), Middle-aged couples (2)
Unit plans *
Building photos

* See details in Appendix C; ** See details in Appendix D.

).

Information on the units and respondents is shown in Table 4. All four surfaces were finished with ceramic tiles.

2.2.2. Summary of Units and Respondents’ Characteristics

Table 5 summarizes the characteristics of the residential units and respondents included in the in-depth measurement stage. The information is grouped by building height into the High-rise and Low-rise types to reflect differences in architectural configuration and indoor environmental exposure. For each point, the table reports the thermal adaptation condition of the unit (AC or Non-AC), along with a unique unit ID used to anonymize participants and ensure traceability across analyses.

Building- and unit-level attributes are presented to document the physical context of the measurements, including floor level, building structure, external wall material, floor material, floor covering, corridor type, presence of window shading, and window-to-wall ratio (WWR, Appendix D, Figure A1). These parameters are reported because they influence solar exposure, heat storage, ventilation potential, and surface thermal conditions, which are particularly relevant for posture-related thermal adaptation. In addition, the table includes key respondent characteristics, namely age, gender, and household composition. Spatial information is further complemented by unit plans and photographs of the building.

3. Results

3.1. Questionnaire-Based Analysis of Regional Climate and Adaptive Behaviors

3.1.1. Overall Characteristics for Regional Climates and Building Types

Table 6. Overall characteristics of the questionnaire results (a) with statistical pairwise comparison tests (b) among regional climates and building types.

(a) Questionnaire Results Variables					Hotter Regions				Cooler Region
					High-Rise, N = 894		Low-Rise N = 1748		High-Rise, N = 174		Low-Rise, N = 315
Thermal adaptation during gathering:    AC usage					554, (62%)		115, (6.6%)		27, (16%)		1, (0.3%)
Non-AC					340, (38%)		1633, (93%)		147, (84%)		314, (99.7%)
Main posture during gathering:    Floor-sitting					268, (30%)		1458, (83%)		116, (67%)		299, (95%)
Other postures					626, (70%)		290, (17%)		58, (33%)		16, (5.1%)
Daytime thermal sensation:    Cool-side					341, (38%)		307, (18%)		68, (39%)		101, (32%)
Neutral					368, (41%)		727, (42%)		57, (33%)		103, (33%)
Warm-side					185, (21%)		714, (41%)		49, (28%)		111, (35%)
Nighttime thermal sensation:    Cool-side					588, (66%)		802, (46%)		128, (74%)		222, (70%)
Neutral					258, (29%)		602, (34%)		40, (23%)		64, (20%)
Warm-side					48, (5.4%)		344, (20%)		6, (3.4%)		29, (9.2%)
Daytime thermal comfort:    Comfortable-side					759, (85%)		1147, (66%)		120, (69%)		229, (73%)
Uncomfortable-side					135, (15%)		601, (34%)		54, (31%)		86, (27%)
Nighttime thermal comfort:    Comfortable-side					825, (92%)		1424, (81%)		137, (79%)		283, (90%)
Uncomfortable-side					69, (7.7%)		324, (19%)		37, (21%)		32, (10%)
Annual electricity consumption [GJ/year]					12.4 (5.6)		5.9 (4.0)		6.1 (3.9)		4.3 (4.5)
(b) Pairwise fisher/wilcoxon test, BH-adjusted	Cooler High-Rise (N = 174) vs. Cooler Low-Rise (N = 315)		Hotter High-Rise (N = 894) vs. Cooler High-Rise (N = 174)		Hotter Low-Rise (N = 1748) vs. Cooler High-Rise (N = 174)		Hotter High-Rise (N = 894) vs. Cooler Low-Rise (N = 315)		Hotter Low-Rise (N = 1748) vs. Cooler Low-Rise (N = 315)		Hotter High-Rise (N = 894) vs. Hotter Low-Rise (N = 1748)
	Raw p Vakue	Adjusted p Value	Raw p Vakue	Adjusted p Value	Raw p Vakue	Adjusted p Value	Raw p Vakue	Adjusted p Value	Raw p Vakue	Adjusted p Value	Raw p Vakue	Adjusted p Value
	Effect Size		Effect Size		Effect Size		Effect Size		Effect Size		Effect Size
Thermal adaptation during gathering	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
	0.313		0.344		0.098		0.543		0.098		0.603
Main posture during gathering	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
	0.377		0.282		0.125		0.571		0.116		0.531
Daytime thermal sensation	0.191	0.191	0.043	0.052	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
	0.082		0.077		0.156		0.149		0.132		0.256
Nighttime thermal sensation	0.049	0.059	0.132	0.132	<0.001	<0.001	0.002	0.003	<0.001	<0.001	<0.001	<0.001
	0.108		0.063		0.168		0.102		0.178		0.223
Daytime thermal comfort	0.404	0.404	<0.001	<0.001	0.402	0.404	<0.001	<0.001	0.014	0.021	<0.001	<0.001
	0.04		0.154		0.02		0.139		0.054		0.204
Nighttime thermal comfort	0.001	0.002	<0.001	<0.001	0.362	0.362	0.193	0.231	<0.001	<0.001	<0.001	<0.001
	0.153		0.167		0.02		0.039		0.08		0.144
Annual electricity consumption	<0.001	<0.001	<0.001	<0.001	0.127	0.127	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
	−0.43		0.72		−0.07		0.82		0.38		−0.70

n (%) Classifications of the regional climates and building heights were applied at the analysis stage and were not used as sampling strata. For categorical variables: Pairwise test: Fisher test with BH-adjusted methods; Effect size: Cramer’s V; For numerical variable (Annual electricity consumption): Pairwise test: Wilcoxon test with BH-adjusted methods; Effect size: Cliff’s delta.

shows the overall characteristics of the questionnaire results between regional climates and building types. The table also presents pairwise comparison test results using Fisher–Freeman–Halton tests between the groups that combine regional climates and building types (e.g., comparison between the Hotter High-rise group and the Cooler Low-rise group).

In the results of thermal adaptation, the proportion of AC usage was larger in the Hotter High-rise than in other groups. In the statistical tests, all comparisons showed statistical differences (<0.1%). The results for the main posture showed a similar pattern: the proportion of floor-sitting was lower (30%) in Hotter High-rise, while it became the majority in other groups. All comparisons also showed statistical differences. For daytime thermal sensation, the proportion of cool-side sensations was larger in Hotter High-rise, Cooler High-rise, and Cooler Low-rise. In Hotter High-rise, AC use was considered the main contributing factor. The statistical tests indicated no statistical difference between Cooler High-rise and Cooler Low-rise, or between Hooter High-rise and Cooler High-rise. Nighttime thermal sensation also showed a similar tendency, but the proportion of cool-side sensation was larger for all groups compared to the daytime results. Thermal comfort showed a large proportion for all groups in both daytime and nighttime; nighttime results were even larger. Average electricity consumption was larger in Hotter High-rise (12.4 GJ/year), due to AC usage. Electricity consumption in the other groups was less than half that of Hotter High-rise, with statistically significant differences among all pairwise comparisons, except for Hotter Low-rise vs. Cooler High-rise.

3.1.2. Thermal Adaptations and Main Posture

Figure 3 shows the proportions of thermal adaptations and main postures during gathering activities. Household income is an important socio-economic indicator of residents and one of the factors influencing their lifestyle [34]; therefore, the results are classified into each monthly household income level. Indonesians view gatherings as an essential aspect of their lifestyle, reflecting the importance of maintaining social connections [35]. Herein, gathering activities were selected as the primary focus within residential units for the activity analysis.

Focusing on the hotter regions, there is a clear tendency that AC usage accounted for a larger proportion in High-rise than in Low-rise. AC usage proportion also increased with monthly household income. In High-rise buildings, AC usage exceeded 70% in the income group of Rp. 7500 K–Rp. 12,000 K, whereas in Low-rise buildings, it remained below 25% for the same income group. In Low-rise, fan usage was dominant for all income ranges. The proportions of “Neither of them” (i.e., not using any thermal adaptation method) were negligible for both building types. This indicates that thermal adaptations using active cooling methods, such as AC and fans, were necessary in the hotter regions.

In the cooler region, both High-rise and Low-rise showed a higher proportion of “opening windows” and “neither of them.” AC usage appeared only in the higher income range of High-rise; however, these proportions were smaller than those of High-rise in the hotter regions. These results indicate that thermal adaptation using active cooling methods, like AC and fans, was not necessarily needed in the cooler region.

Regarding the main posture during gathering activities (Figure 3b), the proportion of chair/sofa-sitting was larger for High-rise in the hotter regions. As for AC usage patterns (Figure 3a), the proportion of chair/sofa-sitting increased along with monthly household income. There was a higher correlation between AC usage and chair-sitting posture with income levels. In High-rise buildings of the cooler region, floor-sitting was also the majority, and this tendency was different from the hotter regions. In Low-rise, floor-sitting was the main posture for both climatic regions. Considering thermal adaptations (Figure 3a), respondents in Low-rise in the hotter regions primarily used fans for cooling while sitting on the floor. Similarly, respondents in the cooler region mainly relied on opening windows and sitting on the floor. This indicates that the combination of airflow (fan and natural ventilation through opening windows) and floor-sitting is necessary for Low-rise residents.

3.1.3. Floor-Sitting Preferences

Figure 4a shows the proportion of preference for sitting on the floor based on thermal adaptation methods. “Like sitting on a floor” was mainly dominant in Low-rise for both climatic regions. In High-rise in the hotter regions, “Dislike sitting on a floor” was the majority. This corresponds to the actual postures where chair/sofa-sitting was obvious, as shown in Figure 3b. However, respondents who selected “Fan” and “Opening front door” showed a higher proportion of “Like sitting on a floor”. In the High-rise in the cooler region, “Not sure” was also obvious.

Figure 4b shows the proportion of reasons for preferring sitting on a floor based on thermal adaptation methods. “Habit,” “Cool,” and “Comfortable” were major reasons for floor-sitting in both regions. In particular, the proportion of “Habit” was larger in Low-rise, and that of “Cool” was larger in High-rise. This shows that residents sat on the floor to enhance coolness even in High-rise in the hotter regions. “Comfortable” appeared in both High-rise and Low-rise. No clear differences in the reported reasons were observed between thermal adaptation methods. However, the proportion of “Cool” was largely shown in the High-rise in the cooler region under window-opening conditions. This indicates that thermal relief could be obtained through floor contact under window-opening conditions.

3.1.4. Thermal Sensation and Comfort

Figure 5 shows the proportion of thermal sensation and comfort with thermal adaptation methods during gathering activities. During daytime (Figure 5a), warm-side sensations mainly appeared in the Low-rise in both the hotter and cooler regions. During nighttime (Figure 5b), cool-side sensations became a larger proportion compared to daytime. A clear difference was shown in the High-rise in the cooler region, where larger proportions of warm sensation occurred during daytime and cold sensation during nighttime for the “Opening window(s)/door” conditions. In the cooler region, window opening resulted in distinct indoor thermal conditions, being warm during daytime and cold during nighttime. This is considered characteristic of the highland tropical climate with a large diurnal temperature range.

Figure 6 shows the proportion of thermal comfort with thermal adaptation methods during gathering activities. During daytime (Figure 6a) in the hotter region, a comfortable sensation was the majority for both High-rise and Low-rise. The proportion of comfortable sensation was slightly larger in High-rise than in Low-rise. This tendency corresponds to the thermal sensation results, with a relatively higher proportion of warm-side sensation in Low-rise. In the High-rise buildings of the cooler region, uncomfortable-side sensations were slightly larger for “Opening window(s)/door.” This corresponds to the thermal sensation, where the warm sensation was higher for “Opening window(s).” It can also be mentioned that the residents who answered “Opening window(s)” for thermal adaptations showed a higher proportion of “Cool” as the reason for floor-sitting (Figure 4b). This indicates that residents could obtain contact cooling effects during floor-sitting postures while opening windows. During nighttime (Figure 6b), most answers were comfortable for both building types and climatic conditions.

3.1.5. Electricity Consumption Patterns

Figure 7 shows annual electricity consumption for the AC usage group and the Non-AC group, based on the monthly household income level. In the hotter regions, there were clear trends that electricity consumption increased with income level in the AC usage group for both High-rise and Low-rise groups. In contrast, the trend was not obvious for the Non-AC group. In the cooler region, even for the AC usage group, electricity consumption did not change with income levels. This result indicates that AC usage was a dominant factor in electricity consumption across income levels in the hotter regions. The results of the Non-AC group support this finding, showing that income levels did not largely affect electricity consumption even in the hotter regions.

3.2. In-Depth Measurement Analysis of Indoor Climate and Adaptive Behaviors

3.2.1. Thermal Adaptation and Main Posture

This section focuses on indoor climatic factors and analyzes the relationship between building type, indoor climate, main posture, thermal adaptation, and thermal comfort using the in-depth measurement and point-in-time questionnaire data obtained in the apartments of Surabaya (i.e., High-rise (Non-AC), High-rise (AC), and Low-rise (Non-AC)).

Figure 8a shows the relationship between thermal adaptation, indoor air temperature, and floor temperature across time periods and building types. The detailed characteristic data and time-series data of indoor air and floor temperatures are shown in Appendix E and Appendix F, respectively. In High-rise (Non-AC), fan usage was reported for all time periods. In High-rise (AC), opening window(s) were reported in addition to AC usage except for nighttime. In Low-rise (Non-AC), fan usage and opening window(s) were reported for all time periods. There are characteristic trends of indoor and floor temperatures among these building types. In High-rise (Non-AC), indoor and floor temperatures were both higher from noon until night. During the afternoon, these temperatures were both higher than 32 °C. In contrast, in High-rise (AC), air and floor temperatures were both lower for all time periods, mostly less than 30 °C, owing to AC usage. In Low-rise (Non-AC), air and floor temperatures were within the ranges between High-rise (Non-AC) and High-rise (AC), showing intermediate characteristics in these temperature conditions. In particular, the floor temperature was more stable than the air temperature.

Figure 8b shows the results of the main posture during the gathering activities with the same format. For all time periods and building types, “Sitting on a floor” was the main posture. “Sitting on chair/sofa” and “Lying on the floor” were also reported in High-rise (AC) and Low-rise (Non-AC). Floor-sitting posture was reported regardless of thermal adaptation methods and building types in the obtained samples. Floor covers were not used in High-rise (Non-AC) and Low-rise (Non-AC), while in High-rise (AC), one respondent utilized a carpet during floor-sitting (see Table 5). Footwear was not used, and all respondents were barefoot in their units.

3.2.2. Thermal Sensation and Comfort

Figure 9 shows the results of thermal sensation and thermal comfort in relation to indoor air temperature and floor temperature across time periods and building types. Figure 9a,b presents the results of whole-body thermal sensation and comfort. Figure 9c,d shows thermal sensation and comfort for the body parts contacting the floor. In High-rise (Non-AC), “Hot/Warm” and “Uncomfortable” sensations for the whole body appeared from noon until night, when indoor and floor temperatures were both higher (over 31 °C), indicating that fan usage alone was not enough to obtain coolness and comfort. In High-rise (AC), “Hot/Warm” and “Slightly uncomfortable” sensations also appeared, especially when the respondents opened the windows. When they used AC, “Cool” and “Comfortable/Very comfortable” sensations were dominant. This result demonstrated a clear relationship between AC usage and thermal sensation and comfort. In Low-rise (Non-AC), thermal sensations remained on the cooler side, especially when air and floor temperatures were both lower (less than 30 °C), and a comfortable sensation was consistently reported throughout the day. “Slightly warm” sensations emerged during the afternoon as indoor air temperatures increased; however, thermal comfort remained predominantly “Comfortable”.

Regarding thermal sensation and comfort for the body parts in contact with the floors (Figure 9d), “Neutral/Slightly cool” and “Neutral/Slightly comfortable” were the majority for thermal sensation and comfort in High-rise (Non-AC). In High-rise (AC), “Neutral” and “Cool” were reported for thermal sensation, and “Slightly comfortable” and “Very comfortable” were reported for thermal comfort. These variations in floor-contact sensations correspond to those in whole-body sensations. This indicates the differences between window opening and AC usage; specifically, lower temperatures during AC usage periods provided cooler and very comfortable sensations. In Low-rise (Non-AC), “Cool/Cold” was dominant for thermal sensation, and “Comfortable” was dominant for thermal comfort. Thermal sensation in floor contact was mostly on the cooler/colder side compared to that of the whole body. This can be attributed to the thermal mass of the normal concrete floor with ceramic tile used in Low-rise, which maintained a lower surface temperature than the indoor air. The cooler floor surfaces, together with occupants’ adaptive behaviors and familiarity with floor-sitting, likely helped sustain coolness and comfort despite relatively higher air temperatures. This is further discussed in Section 4.2.

4. Discussion

4.1. Questionnaire-Based Analysis of Regional Climate and Adaptive Behaviors

Table 7. Summary of the large-scale questionnaire for regional climate and adaptive behaviors. Summary of the large-scale questionnaire for regional climate and adaptive behaviors.

City	Temperature Condition	Building Type	Thermal Adaptation	Main Posture
Jabodetabek Surabaya Medan Makassar	Hotter (Annual average temperature: 28–29 °C)	High-rise	AC (higher income), Fan (Lower income)	Chair/sofa sitting (higher income), Floor-sitting (lower income)
		Low-rise	Fan	Floor-sitting
Bandung	Cooler (Annual average temperature: 26 °C)	High-rise	Neither of them, Window opening	Floor-sitting
		Low-rise	Neither of them, Window opening	Floor-sitting
City	Preference for Sitting on the Floor	Thermal Sensation (Daytime)	Thermal Sensation (Nighttime)	Thermal Comfort (Day-Night)
Jabodetabek Surabaya Medan Makassar	Not preferred (higher income) Preferred (lower income)	Slightly cool— Slightly warm	Cool—Neutral	Comfortable
	Preferred	Slightly cool—Warm	Cool—Neutral	Comfortable
Bandung	Preferred	Slightly cool—Warm	Cold—Neutral	Comfortable
	Preferred	Slightly cool—Warm	Cool—Neutral	Comfortable

shows the summary of the main findings from the large-scale questionnaire. Regarding thermal adaptations, the current study revealed that AC usage was dominant for higher-income residents in High-rise buildings of the hotter regions. In contrast, fan usage was predominant among lower-income residents of High-rise buildings and across all income ranges in Low-rise buildings in the regions. Although window opening was also reported in these regions, the proportions were small. Mori et al. (2020) clarified that window opening and fan usage were frequently reported in public rental apartments (Rusunawa) for low- and middle-income earners in hot-humid regions, including Surabaya [22]. Feriadi and Wong (2004) also revealed that occupants in naturally ventilated buildings in hot and humid regions tended to modify the living environments by creating air movement (fans and opening windows) [36]. These findings suggest the importance of airflow without using AC in the hotter regions. In the cooler region, the current study revealed that AC usage and window opening were both reported in High-rise, and this was consistent with the findings obtained for the same region (Bandung) by Mori et al. (2020) [22].

In addition to thermal adaptation strategies, clear differences were observed in main posture and floor-sitting preferences across these building types and regions. The hotter regional residents in High-rise more frequently chose chair-sitting while using AC, especially for the higher income brackets. Floor-sitting preference results among these residents showed that although the preference for floor-sitting was lower, obtaining “Cool” was a major reason for floor-sitting. The hotter regional residents in Low-rise chose floor-sitting while using fans. The cooler region was also highlighted for its high prevalence of floor-sitting in both High- and Low-rise buildings, despite minimal AC or fan usage, often paired with window opening as a thermal adaptation. High-rise residents in the cooler region reported a strong preference for floor-sitting, often citing “Cool” as the reason, especially while keeping windows open. Although the cooler climate imposes lower thermal stress, residents still selected floor-sitting to stay cool, indicating a moderate form of thermal adaptation. Interestingly, even higher-income High-rise residents in the cooler region, who faced no economic constraints, selected floor-sitting for cooling. For the Low-rise residents in both the hotter and cooler regions, “Habit”, “Cool”, and “Comfortable” were the main reasons for floor-sitting. This indicates that floor-sitting is a habitual and adaptive behavior to obtain coolness and comfort for Low-rise residents.

Thermal sensation and comfort results revealed further layers of insight. In the High-rise buildings of the hotter regions, cooler-side sensations were more frequent, reflecting AC usage. In High-rise buildings of the cooler region, hot-side sensations were reported during daytime—when window opening was common—indicating that residents relied on floor-sitting and natural ventilation for cooling. At night, cooler-side sensations increased, likely due to the temperature drop (see Table 2), suggesting incomplete adaptation to daytime warmth. In Low-rise apartments of the hotter regions, although neutral and warmer-side sensations were the majority during daytime, residents mostly felt thermally comfortable with floor-sitting and fan usage.

4.2. In-Depth Measurements of Indoor Climate and Adaptive Behaviors

Table 8. Summary of the in-depth measurements of indoor climate and adaptive behaviors.

Building Type	Indoor Air Temperature/Floor Temperature	Thermal Adaptation	Main Posture	Thermal Sensation
High-Rise (Non-AC)	Higher/ Higher	Fan	Floor-sitting	Hot— Cool
High-Rise (AC)	Lower/ Lower	Window opening/ AC	Floor-sitting	Hot— Cool
Low-Rise (Non-AC)	Moderate/ Slightly lower (Stable)	Fan/ Neither of them	Floor-sitting/ Lying on the floor	Cool
Building Type	Perceived Thermal Comfort	Thermal Sensation in Floor Contact	Thermal Comfort in Floor Contact	Shading/Thermal Mass
High-Rise (Non-AC)	Uncomfortable— Slightly comfortable—Uncomfortable	Neutral— Slightly cool	Neither of them— Slightly comfortable	Poor shading, low thermal mass
High-Rise (AC)	Comfortable—Slightly Uncomfortable— Very comfortable	Neutral— Cool	Slightly comfortable— Very comfortable	Poor shading, low thermal mass
Low-Rise (Non-AC)	Comfortable	Cool— Cold	Comfortable	With shading, high thermal mass

summarizes the findings from in-depth measurements. Indoor air and floor temperatures were influenced by the building types and AC usage patterns. In High-rise (Non-AC), both air and floor temperatures were relatively high. This can be attributed to its lightweight wall construction with low thermal mass and without shading devices (Table 5). The unit faced northwest and easily received solar radiation during the afternoon owing to the sun’s position in the measurement period (late July to mid-September). Alfata et al. (2015) indicated that modern high-rise apartments in Indonesia are poorly designed in terms of thermal insulation and solar shading, and most apartments are constructed of aerated light concrete walls with relatively low thermal mass [24]. By contrast, High-rise (AC) maintained lower air and floor temperatures. The building has a similar design, WWR, and materials to High-rise (Non-AC). However, the units faced south and east, receiving less solar radiation. Furthermore, lower air and floor temperatures were maintained through AC operation from morning until night. One respondent utilized a carpet during floor-sitting; this might be due to the lower air and floor temperatures using AC.

Low-rise (Non-AC) recorded lower air and floor temperatures than High-rise (Non-AC). The WWRs of the two Low-rise buildings were different (i.e., 32% for LRA and 14% for LRB). However, the two Low-rise buildings shared design features—such as shaded window eaves and normal concrete walls with higher thermal mass—that helped stabilize indoor thermal conditions. In addition, differences emerged between the two. One (i.e., LRB) exhibited relatively higher air and floor temperatures, likely because its living room (the primary occupied space during daytime and the measurement site) was located in a perimeter zone, making it more vulnerable to outdoor thermal fluctuations. These units faced different orientations (i.e., east, south, and west). In the other building (i.e., LRA), the living room faced an internal corridor, with bedrooms along the external wall facing west, helping maintain more stable indoor conditions.

These indoor climate differences were reflected in thermal adaptations, thermal sensation, and thermal comfort responses. High-rise (Non-AC) residents who utilized fans reported hotter and less comfortable sensations during afternoons when both air and floor temperatures peaked. High-rise (AC) showed similar hot-side and slightly uncomfortable sensations during periods of opening windows without AC operation. During the time using AC, a cool and comfortable sensation was obtained. As Alfata et al. (2015) indicated, it was difficult to achieve thermal comfort without relying on AC in new private apartments [24]. Thermal sensation and comfort were also on the cool and comfortable side in Low-rise buildings (Non-AC), where fan usage and window opening often appeared. These units consistently maintained cooler and more comfortable sensations for the residents, especially when air and floor temperatures were both less than 31 °C. The 80% upper comfort limit of the operative temperature (OT) can be estimated according to a previous study on naturally ventilated buildings in hot-humid climates [37]. The resultant OT was 30.3 °C, and the obtained air and floor temperatures in the conditions (31 °C) were slightly higher than this value. Among these conditions, thermal sensation and comfort in floor contact were cool/cold and comfortable. Notably, floor-contact thermal sensations were cooler/colder than those of the whole body. For Indonesians, the closest thermal descriptors of a feeling of thermal comfort were the cool and cold sides [38]. This was also shown in tropical residents, where comfort was associated with verbal anchors on the cold side [39]. Across all cases, floor-sitting remained the dominant posture. Contact cooling effects [19] might be obtained through floor-sitting postures; however, they were solely assumed from subjective reports combined with observed air and floor temperatures. Therefore, the mechanisms of contact cooling effects should be clarified through laboratory experiments with physical (floor temperatures and heat flux), physiological (skin temperatures), and psychological (thermal and comfort sensations) measurements in future studies.

4.3. Thermal Adaptation, Lifestyle, and Comfort in Tropical Residential Contexts

As discussed above, the obtained results suggest the importance of building design and materials (i.e., adequate shading and thermal mass). Li (2025) synthesized recent evidence showing that building design, envelope performance, and interior configurations jointly influence indoor thermal conditions and occupant comfort [40]. Architectural characteristics, such as window-to-wall ratio (WWR) and shading conditions, will influence solar heat gains, ventilation potential, and radiant temperature distribution, shaping indoor thermal exposure and adaptive behaviors [41,42,43]. Other previous studies have also shown that building design features regulating solar access and airflow can significantly modify perceived comfort and adaptive behavior, especially in naturally ventilated dwellings [44]. The high thermal mass of houses in a hot-humid climate showed that the air temperature in the spaces maintained relatively low values [45]. Behavioral studies indicate that occupants dynamically adjust posture, window opening, and fan use in response to these micro-environmental conditions rather than relying solely on AC systems [42]. These findings provide a design-based context for the observed differences between low-rise and high-rise buildings in this study, particularly the role of thermal mass and floor temperature stability in supporting floor-sitting as a viable adaptive strategy.

However, over recent decades, a marked shift has occurred toward active thermal adaptation, most notably the widespread adoption of AC. Rising household incomes, intensifying urban heat stress [46], and growing expectations for stable indoor environments have accelerated this trend [47]. While AC delivers immediate comfort, it also elevates residential electricity demand [48], straining both household budgets and national energy systems. Miyamoto et al. (2024) indicated that the use of AC had a significant impact on energy consumption in Indonesia [49]. In the current study, AC usage was a dominant factor for electricity consumption across income levels in the hotter regions (See Figure 7).

In contrast, other active methods, such as electric fans, consume less energy but are often perceived as inadequate in extreme heat. As AC use becomes more prevalent in high-rise housing, it displaces culturally rooted practices that historically enabled residents to adapt to local climatic conditions with minimal energy input [30,50]. Studies indicate that thermal adaptation in tropical climates cannot be explained purely in physiological terms, as it is also mediated by cultural traditions and lifestyle-related behaviors [51,52]. Furthermore, psychological dimensions of comfort play a critical role: residents in tropical regions often accept warmer indoor conditions than international benchmarks (e.g., ASHRAE 55), provided they retain adaptive opportunities such as posture adjustment, ventilation control, or cultural practices [26,53].

Socio-demographic and lifestyle variables, including household composition, occupation, and duration of spending time at home, also affect posture flexibility and cumulative thermal exposure. Empirical evidence suggests that time spent at home significantly influences thermal control behaviors and comfort expectations, reinforcing the role of daily occupancy patterns in shaping adaptive practices [54]. The current study could only discuss regional and indoor climatic factors influencing adaptive behaviors; therefore, lifestyle variables should be analyzed in further studies.

4.4. Posture-Based Thermal Adaptation in Relation to Recent Comfort Studies

Floor-sitting posture has historically been embedded in social and cultural practices across many Asian societies [16,17], where activities such as dining, gathering, or relaxing are often conducted on floor mats, carpets, or directly on the floor. Beyond its cultural importance, floor-sitting can also serve as a passive cooling strategy. Physical contact with the floor can enhance heat dissipation from the body [19], particularly when the floor is cooler than the ambient air. In tropical housing without AC, this simple adaptation can provide localized thermal relief. For instance, Erwindi et al. (2025) found that while floor-sitting persists among lower-income residents in publicly subsidized housing (Rusunawa), its prevalence declines among higher-income and High-rise dwellers who favor AC and standard furniture usage [18]. The current study revealed that floor-sitting behavior was also reported among higher-income and High-rise dwellers of the cooler regions. This divergence is closely tied to the climatic conditions, apartment design, cooling availability, and occupant socioeconomic status. In addition, this study indicated the importance of the combination of indoor air and floor temperatures to maintain cooler thermal sensation and comfort while conducting floor-sitting behavior without using AC in the hotter regions.

Occupant posture was increasingly recognized in recent thermal comfort research as not merely a neutral background condition, but as an active component of behavioral thermal adaptation, particularly in warm and humid environments. Systematic reviews demonstrated that adaptive behaviors—such as changes in sitting posture, use of floor surfaces, fan operation, and window opening—play a critical role in mediating comfort when indoor temperatures exceed conventional comfort ranges [55]. Further insights into thermal comfort optimization and heat resilience indicate that adaptive capacity at the occupant level mitigates thermal stress when mechanical cooling is limited or absent. Kajjoba et al. (2025) reported that in low-income tropical housing, thermal comfort is frequently maintained through passive means and behavioral flexibility rather than continuous AC use [56]. Jain et al. (2024) highlighted that resilient indoor environments emerge not only from advanced technologies but also from occupants’ ability to interact with building features and microclimatic conditions [57]. Furthermore, Apriliyanthi et al. (2025) identified patterns of occupants’ thermal adaptation, including fan and window use, clothing adjustment, and also AC set point, through cluster and regression analysis in hot-humid conditions of Indonesia, highlighting behavioral diversity across residential types [58]. Similarly, in 2024, a field study in naturally ventilated buildings demonstrated that acceptable thermal comfort can be achieved through adaptive behaviors under fluctuating indoor conditions, even with short monitoring periods [59]. These perspectives provide an important interpretive framework for the present findings, in which floor-sitting emerged as a prevalent strategy for achieving acceptable thermal comfort, and aligns with where residents in low-rise, non-air-conditioned (Non-AC) units achieved thermal comfort through floor contact, fan use, and window operation rather than reliance on air-conditioning (AC).

However, this study has several limitations. First, although the sampling design ensured representativeness of apartment housing residents in the large-scale questionnaire, building height was not used as a stratification variable at the sampling stage. Therefore, analyses by building height should be interpreted as comparative and exploratory within the sampled apartment population, rather than as population-weighted estimates. Second, while the consistency between observed posture choices, indoor climate conditions, and contemporary thermal comfort research strengthens the validity of the conclusions obtained from the in-depth measurement, the limited number of measured cases constrains generalization. Therefore, further studies with an increased number of samples for the measurements are important.

5. Conclusions

This study analyzed the influence of regional and indoor climatic factors on floor-sitting as an adaptive behavior and thermal comfort in Indonesian apartments. First, a large-scale questionnaire was employed to clarify the relationship between regional climates, floor-sitting behavior, thermal adaptations, household electricity consumption, and thermal comfort. Next, in-depth measurement was conducted to examine the influence of building type and indoor climate on thermal adaptations and thermal comfort with floor-sitting behavior in the hotter region (Surabaya).

Collectively, the evidence suggests that thermal comfort in tropical housing is shaped through the dynamic interaction of climatic conditions, building types, cultural practices, and technological interventions. The ongoing shift from passive, culturally embedded practices (floor-sitting with lifestyle adjustments) to energy-intensive solutions (AC) represents a redefinition of comfort expectations. Recognizing and integrating such adaptive behaviors as floor-sitting is critical for developing sustainable housing strategies that respect cultural practices while reducing energy demand in rapidly urbanizing tropical regions.

As a conclusion, this study reveals several key findings on floor-sitting as an adaptive behavior in Indonesian climates and building types:

• Climate-driven adaptation: In the hotter regions like Jabodetabek and Surabaya, floor-sitting was primarily conducted without the use of AC and often alongside the use of electric fans in Low-rise housing. In contrast, in High-rise housing of these hotter regions, AC usage and chair/sofa sitting were common adaptive behaviors. In the cooler region of Bandung, however, floor-sitting remained a frequent adaptive behavior in both Low-rise and High-rise housing and was often accompanied by window openings for both building types. Climatic conditions shaped thermal adaptations and floor-usage behaviors. In addition, floor-sitting consistently emerged as a habitual and adaptive behavior to obtain coolness and comfort for Low-rise residents in both climatic regions.
• Socio-cultural resilience: Even among higher-income individuals living in High-rise residences in Bandung, floor-sitting and window opening remained preferred over AC use, indicating non-economic motivations for this practice in the cooler region.
• Building performance matters: In the hotter region of Surabaya, the Low-rise apartment units consistently maintained cooler and comfortable sensations for the residents without using AC, especially when air and floor temperatures were both less than 31 °C, which was lower than those in the High-rise apartment unit without AC usage. The indoor climatic conditions in the Low-rise apartments were supported by building materials with high thermal mass and building design with adequate solar shading. Furthermore, thermal sensation in floor contact was cooler/colder than that of the whole body, and the residents could feel comfortable, suggesting the importance of a cooler floor for thermal comfort.

These findings indicated that cooler climates could provide cool and thermally comfortable living environments with floor-sitting in dwellings; however, even in hotter climates, creating stable indoor climates with adequate consideration for building design and materials might serve similar environments where residents could continue the culturally embedded adaptive behavior of floor-sitting. Overall, this study demonstrated that floor-sitting is a vital behavior that adapts to regional and indoor climatic conditions in the tropics while achieving thermal comfort and relying less on AC devices.