Assessing the Integrated Impacts of Outdoor Thermal Environment and Air Quality on Thermal Comfort in Residential Areas: A Case Study of Wuhan

Abstract

Climate change and rapid urbanization are intensifying global environmental challenges, particularly the nexus of urban heat stress and air pollution, which collectively impact human health and outdoor livability. This study investigates the spatiotemporal distribution of the outdoor thermal environment and PM2.5 concentrations in residential areas of Wuhan, a city with a hot-summer–cold-winter climate, and evaluates their combined effects on outdoor thermal comfort. Field measurements of microclimatic parameters and PM2.5 levels were conducted in two typical residential communities during winter and summer, supplemented by 582 valid questionnaires to assess residents’ subjective thermal responses. Key findings include (1) residents’ satisfaction decreases by approximately 1 unit for every 90-unit increase in AQI, and PM2.5 concentrations can effectively substitute for AQI in characterizing the impact of air quality on satisfaction. (2) The explanatory power of the UTCI-MTSV model (R²: 0.3152–0.7763) is pollutant-dependent; winter thermal comfort adheres to the linear law of UTCI, while dispersed summer votes indicate a non-linear effect of PM2.5 on comfort. (3) Wuhan residents show high thermal tolerance across AQIs. The lower limits of the Thermal Acceptability Range (TAR) are 7.1 °C (AQI-I) and 8.8 °C (AQI-II), both below neutral ranges, while the summer TAR-UTCI is 30.9 °C (above the neutral range). Better air quality improves the reliability of the thermal acceptability–UTCI fit. (4) TCV peaks at approximately 16 °C, increasing then decreasing with UTCI; at identical UTCI levels, better air quality enhances comfort, particularly within the 0–10 °C range. This study provides empirical evidence to inform urban design strategies for mitigating heat stress and pollution in hot-summer–cold-winter regions.

1. Introduction

Climate change and urbanization have precipitated significant environmental challenges, including global warming, exacerbated air pollution (particularly PM2.5), and increased frequency of extreme weather events. Data from the IPCC Sixth Assessment Report indicate that the global mean surface temperature rose by 1.2 °C between 1901 and 2021, with heat-wave-related deaths reaching 26,800 in 2019 [1]. In China, PM2.5 pollution caused an average of 0.971 million premature deaths annually up to 2017, incurring economic losses of 2.75 trillion CNY [2]. Urban residential areas, as critical components of the living environment, face dual pressures of declining thermal comfort and deteriorating air quality, directly affecting residents’ health and activity patterns [3]. Research aimed at ameliorating these issues can enhance the urban environment and positively influence public well-being. Among environmental concerns, diminished comfort and poor air quality are particularly pressing. Improving air quality reduces residents’ exposure to pollutants, while enhancing outdoor thermal comfort encourages physical activity and informs urban design and renewal.

Field measurements based on meteorological stations are frequently employed to assess outdoor environmental performance. Van Hove et al. [3], using observations from Rotterdam, demonstrated a pronounced link between construction intensity and the urban heat island effect. Perini et al. [4] confirmed through parametric building experiments that density metrics are positively correlated with air temperature, whereas increasing building height can amplify the cooling capacity provided by vegetation and green spaces. Thermal comfort is typically evaluated via on-site measurements combined with questionnaire surveys, grounded in thermal environment characteristics and human physiological response.

Key physical indicators influencing thermal comfort include air temperature (T_a), the core determinant of human heat balance, while relative humidity (RH), wind speed (Va), and other variables regulate the local environment, with seasonal variations in their effects [5,6,7]. Subjective evaluation indices such as the Thermal Sensation Vote (TSV), Thermal Comfort Vote (TCV), and Thermal Acceptability Vote (TAV) are also considered, with corresponding metrics provided by ISO 10551 [8]. Evaluation frameworks often employ indices like the Physiologically Equivalent Temperature (PET) and the Universal Thermal Climate Index (UTCI). For example, Wang Yi [9] found that UTCI exhibits a strong correlation with summer thermal sensation votes in Shanghai.

In assessing outdoor pollutants, researchers often rely on indicators such as mean PM2.5 concentration and the Air Quality Index (AQI) [10,11,12]. Particulate matter (PM2.5) is the dominant pollutant affecting air quality in Wuhan’s urban areas [13]; thus, mean PM2.5 is frequently employed as a neighborhood-scale indicator of outdoor air quality. Parameters like the AQI (defined by the U.S. EPA) integrate concentrations of common pollutants to provide an intuitive snapshot of pollution levels [14]. The AQI also estimates health risks from short-term pollutant exposure, making it a widely adopted field metric. This clear delineation offers a reliable basis for evaluating pollution risks in residential outdoor environments, facilitating investigation of their impact on living conditions and residents’ health.

Numerous studies have demonstrated mutual influence between thermal comfort and air quality, as illustrated in Figure 1. Strong solar radiation and high-temperature conditions intensify photochemical reactions, resulting in elevated concentrations of photochemical pollutants such as ozone [15,16,17]. Nevertheless, existing research has paid limited attention to the spatiotemporal heterogeneity of thermal comfort and pollutants. Most prior studies on residential environments employed a single-perspective approach, focusing on how individual microclimatic parameters affect human thermal perception outdoors [18,19,20]. Approximately 60% of these studies concentrate on outdoor thermal comfort indices (e.g., UTCI, PET, SET), while 40% address air-quality parameters (e.g., PM2.5, pollutant concentrations), primarily using on-site measurements. However, outdoor thermal comfort is governed by the combined effects of multiple factors. Historically, thermal comfort research has considered key microclimatic elements—air temperature, wind speed, humidity, and thermal radiation—as primary determinants, examining their individual influences on human sensation. The impact of air quality on thermal comfort perception has often been overlooked. Variations in outdoor environmental factors are complex and occur over wide spatial and temporal scales, with interdependent parameters; consequently, a single microclimatic variable cannot fully account for changes in thermal comfort. Wind-tunnel experiments on street-canyon prototypes by researchers such as Xie [21], Lin [22], and Mei [23] have revealed the dual mechanism by which building aspect ratios and layout density simultaneously affect buoyancy-driven ventilation efficiency and pollutant dispersion capacity. Their findings indicate that as independent variables change, air temperature rises while pollutant residence times lengthen, as detailed in Table 1.

Moreover, the literature reveals a paucity of integrated investigations into outdoor thermal environment and air quality both domestically and internationally. The team led by Chunping Miao [24] demonstrated a linear negative relationship between summer street-tree canopy density and PET values in Shenyang, while the group headed by Liyan Rui [25] confirmed via modeling in Nanjing that green-space configuration exerts simultaneous effects on both thermal conditions and air quality, highlighting the potential for synergistic regulation through residential design elements. Empirical evidence shows that these design attributes play dual roles in thermal comfort and pollutant dispersion. In Tianjin, Bian et al. [26] quantified a 1-grade decline in air satisfaction for every 230-unit increase in AQI and identified a neutral PET threshold of 21.68 °C, significantly lower than that reported for hot-humid regions. Employing spatial overlay analysis, Sevgi Yilmaz et al. [27] uncovered a conflicting pattern: building frontage ratios of 60–75% mitigate heat stress yet simultaneously reduce PM2.5 dispersion rates. At the mesoscale, Wowo Ding’s team [28] used large-scale gridded simulations to quantify the trade-off between street-canyon aspect ratios and improvements in thermal comfort versus prolonged PM2.5 residence times, further developing a coupled UTCI-AQI model. Collectively, these findings indicate that current research has yet to establish a multi-parameter framework for assessing outdoor thermal comfort that incorporates the impact of varying air pollution levels. Reconciling thermal environment and air-quality conflicts is therefore required to guide sustainable environmental design in high-density urban residential districts.

To address this empirical gap, the present study focuses on residential communities in Wuhan, China, as the study area and employs field measurements to investigate the spatiotemporal patterns of outdoor thermal comfort and PM2.5, thereby exploring their relationship. The study compares associations between different air pollutant concentrations and thermal comfort, and develops regression models relating UTCI to TCV under varying air pollutant concentration classes. The main goal is to gain an in-depth understanding of the distributional characteristics of outdoor thermal comfort and air quality in Wuhan, provide scientific evidence for optimizing the development of healthy residential districts, and offer a reference design system.

Table 1. Literature review on outdoor thermal comfort and outdoor air pollutants.

Author	Core Focus	Research Content	Study Morphology	Evaluation Parameters	Indicators and Methods
Miao C, He X, Gao Z, et al. [29]	Thermal Comfort and Air Quality	Vertical distribution of thermal comfort in relation to PM2.5 and O3 concentrations	street canyon	PET, PM2.5, O3, NO2	Temperature, humidity, pollutant concentration (field measurement)
Xie X, Liu C H, Leung D Y C [21]	Air Quality	Influence of street-canyon morphology on air quality	prototype/street canyon	pollutant concentration	Air temperature, wind speed (wind-tunnel experiment)
Martilli A, Nazarian N, Krayenhoff E S, et al. [30]	Thermal Comfort	Bulk Richardson number, ambient wind parameters, and surface-heat distribution	prototype/building layout	air temperature	Air temperature, wind speed (wind-tunnel experiment)
Lin L, Hang J, Wang X, et al. [22]	Air Quality	Street-canyon morphology, bulk Richardson number, ambient wind parameters, and surface-heat distribution	prototype/street canyon	pollutant concentration	Wind speed, turbulent kinetic energy (wind-tunnel experiment)/pollutant concentration (wind-tunnel experiment)
Mei S J, Liu C W, Liu D, et al. [23]	Air Quality	Effects of street-canyon morphology on wind environment and air quality	prototype/street canyon	air temperature/air exchange rate/pollutant residence time/pollutant concentration	Wind speed, air temperature (empirical formula)
Bian G, Gao X, Zou Q, et al. [26]	Thermal Environment and Air Quality	Combined impacts of urban-park thermal environment and PM2.5 on thermal comfort	urban park	PET, AQI, TSV, TCV	Temperature, wind speed, PM2.5 (field survey)
Yilmaz S, Sezen I, Irmak M A, et al. [27]	Thermal Comfort and Air Quality	Differences in thermal comfort and SO2/PM10 levels between urban centers and suburbs	urban center/suburb	PET, SO2, PM10, wind speed	Temperature, elevation, building density (GIS analysis)
Rui L, Buccolieri R, Gao Z, et al. [31]	Microclimate and air quality	Regulation of thermal environment and PM10 by different residential vegetation layouts	residential area	PM10, temperature, wind speed	Green-space coverage ratio, isolation index (ENVI-met simulation 5.6)

Table 1. Literature review on outdoor thermal comfort and outdoor air pollutants.

Author	Core Focus	Research Content	Study Morphology	Evaluation Parameters	Indicators and Methods
Miao C, He X, Gao Z, et al. [29]	Thermal Comfort and Air Quality	Vertical distribution of thermal comfort in relation to PM2.5 and O3 concentrations	street canyon	PET, PM2.5, O3, NO2	Temperature, humidity, pollutant concentration (field measurement)
Xie X, Liu C H, Leung D Y C [21]	Air Quality	Influence of street-canyon morphology on air quality	prototype/street canyon	pollutant concentration	Air temperature, wind speed (wind-tunnel experiment)
Martilli A, Nazarian N, Krayenhoff E S, et al. [30]	Thermal Comfort	Bulk Richardson number, ambient wind parameters, and surface-heat distribution	prototype/building layout	air temperature	Air temperature, wind speed (wind-tunnel experiment)
Lin L, Hang J, Wang X, et al. [22]	Air Quality	Street-canyon morphology, bulk Richardson number, ambient wind parameters, and surface-heat distribution	prototype/street canyon	pollutant concentration	Wind speed, turbulent kinetic energy (wind-tunnel experiment)/pollutant concentration (wind-tunnel experiment)
Mei S J, Liu C W, Liu D, et al. [23]	Air Quality	Effects of street-canyon morphology on wind environment and air quality	prototype/street canyon	air temperature/air exchange rate/pollutant residence time/pollutant concentration	Wind speed, air temperature (empirical formula)
Bian G, Gao X, Zou Q, et al. [26]	Thermal Environment and Air Quality	Combined impacts of urban-park thermal environment and PM2.5 on thermal comfort	urban park	PET, AQI, TSV, TCV	Temperature, wind speed, PM2.5 (field survey)
Yilmaz S, Sezen I, Irmak M A, et al. [27]	Thermal Comfort and Air Quality	Differences in thermal comfort and SO2/PM10 levels between urban centers and suburbs	urban center/suburb	PET, SO2, PM10, wind speed	Temperature, elevation, building density (GIS analysis)
Rui L, Buccolieri R, Gao Z, et al. [31]	Microclimate and air quality	Regulation of thermal environment and PM10 by different residential vegetation layouts	residential area	PM10, temperature, wind speed	Green-space coverage ratio, isolation index (ENVI-met simulation 5.6)

2. Methodology

2.1. Cases of the Urban Areas

Wuhan is located at the confluence of the Yangtze and Han Rivers (geographical coordinates 29°58′–31°22′ N, 113°41′–115°05′ E) and serves as the capital of Hubei Province. In recent years, Wuhan’s urban development has advanced steadily. According to the 2022 Wuhan Statistical Yearbook, urban construction continues alongside population growth, with about 80.49% of Wuhan’s population residing in urban areas by the end of 2022. Consequently, issues such as pollution and thermal environment problems have increased, posing threats to residents’ physical and mental health. Meteorological data analysis results are as follows: The Wuhan Statistical Yearbook 2022 [32] data are shown in Figure 2, showing that August is the hottest month in summer, with temperatures reaching up to 40 °C, while January is the coldest month in winter, with average temperatures around 3 °C and lows below zero. Previous studies on the annual distribution of PM2.5 pollutants in Wuhan, based on measurement times in the Wuhan Environmental Air Quality Monthly Report (Wuhan Ecological Environment Monitoring Center, 2023; Wuhan Ecological Environment Monitoring Center, 2024), indicate that the average PM2.5 concentration in July 2024 was 13 μg/m³, while in December 2023, it was 60 μg/m³.

This study selected two typical residential communities in Wuhan as measurement sites: Paris Haoting Community and Jindi Sun City Community. Paris Haoting Community is located at No. 2 Luoyu East Road, Hongshan District, Wuhan, China; Jindi Sun City Community is located at No. 369 Guanshan Avenue, Guanshan, Hongshan District, Wuhan, China. As shown in Figure 2, both communities are the typical post-2000 residential areas in urban Wuhan, which have similar spatial and morphological structures, representing prevalent medium-to-high-density residential forms developed during China’s rapid urbanization. More broadly, Wuhan is situated within China’s vast hot-summer–cold-winter region, a strategic climatic zone that encompasses many of the country’s major metropolitan areas and serves as a critical hub for population and economic activity. Consequently, both communities, surrounded by traffic arteries in the central urban area, are subject to characteristic urban heat island effects and anthropogenic pollution, making them relevant and strategic cases for investigating the research problem.

2.2. Research Methodology

This study employs a combination of microclimate measurements and thermal comfort surveys to investigate the microclimate and thermal comfort in residential areas of Wuhan.

2.3. Microclimate Measurement

The study selected representative meteorological days in winter (20–23 January 2024) and summer (27–30 July 2024) to conduct research in two representative communities. To eliminate interference from extreme weather, measurement dates avoided periods of strong convective weather such as rain or snow, and were deliberately chosen on sunny or cloudy days. Considering the potential impact of traffic volume differences between weekdays and weekends on pollutants, measurement times covered two weekdays and weekends in both the coldest months (December–January) and the hottest months (June–August) to enhance data representativeness. Furthermore, considering human activity patterns, the measurement period was set from 9:00 to 17:00, during which outdoor thermal comfort and pollutant distribution were investigated.

To capture the spatiotemporal heterogeneity of microclimate and air quality, 12 fixed measurement points were systematically established across the two sites. Points were selected to cover a spectrum of common outdoor space types, stratified by key morphological variables known to influence the physical environment. Selection accounted for variations in ground cover (pavement, lawn) and the presence or absence of tree shading; actual locations are shown in Figure 2. Additionally, a monitoring point was set up on the rooftop of the residential area to record background data. At each measurement point, a PC-8/8A mini weather station and a TM-188D handheld comprehensive heat index meter were installed to obtain meteorological parameters at a height of 1.1 m above ground, including ambient temperature (T_a, °C), relative humidity (RH, %), globe temperature (T_g, °C), wind speed (V, m/s), and wind direction. Data were automatically recorded at 1 min intervals. The layout of measurement points and instruments used are shown in Appendix A.

2.3.1. Questionnaire Survey

The questionnaire survey consisted of three main parts. Detailed information is presented in Appendix B.

(1)

Personal Information

The first part collected basic personal information from respondents, including fundamental physical characteristics, clothing thermal resistance, and activity status. According to ASHRAE data [33], half of the respondents had a seated metabolic rate of 1.0 met, while the other half were in a standing slow-walking state with a metabolic rate of 1.6 met.

(2)

Thermal Environment Evaluation and Air Quality Assessment

Subjective thermal responses were obtained through the questionnaire, reflected by thermal sensation votes. Thermal sensation was assessed using the traditional ASHRAE 7-point scale (−3: very cold; −2: cold; −1: slightly cool; 0: neutral; 1: slightly warm; 2: warm; 3: very warm). Thermal comfort was evaluated using a 5-point scale (−2: very uncomfortable; −1: uncomfortable; 0: neutral; 1: comfortable; 2: very comfortable). Thermal acceptability was represented on a 5-point scale (−2: completely unacceptable; −1: unacceptable; 0: neutral; 1: acceptable; 2: completely acceptable). Preferences for meteorological conditions such as air temperature, wind speed, relative humidity, and solar radiation were recorded using a 5-point scale. Perceptions of pollutants were expressed on a 5-point scale (−2: very poor; −1: poor; 0: neutral; 1: good; 2: very good).

(3)

Additional Information

Beyond basic physiological information, researchers sought a more accurate understanding of individuals’ outdoor conditions and comfort perceptions. This involved statistical analysis of residents’ outdoor thermal comfort and sensitivity to pollutants, as well as recording the criteria by which residents assess air quality. Furthermore, other activity information of the subjects was documented, with detailed inquiries into typical residential activity specifics.

2.3.2. Thermal Comfort and Air Quality Indicators

(1)

Outdoor Thermal Environment Indicators

Previous studies have demonstrated the high applicability of the UTCI in hot-summer–cold-winter regions. Therefore, this study employs UTCI to comprehensively assess the outdoor thermal environment, with specific adjustments for Wuhan. The study utilizes RayManPRO3.1 to calculate UTCI, inputting air temperature (T_a), relative humidity (RH), wind speed (V), mean radiant temperature (T_mrt), and individual human parameters. Among meteorological parameters, T_a and RH are directly measured, while wind speed V at 10 m is calculated.

T_mrt, as a crucial factor in human energy balance and thermal comfort, is commonly calculated using the black globe temperature (T_g). According to ISO7726 standards [34], the formula for calculating T_mrt is given as Equation (1), and the formula for calculating wind speed at 10 m is as follows Equation (2).

$$\begin{array}{c}{T}_{mrt}={\left[(T_g+273.15{)}^4+{\displaystyle \frac{1.10\times 10^8{V}^{0.6}}{\epsilon D^{0.4}}}\left(T_g-T_a\right)\right]}^{0.25}-273.15\end{array}$$

(1)

where T_g is the black globe temperature (°C), V is the wind speed (m/s), T_a is the air temperature (°C), D is the sphere diameter (0.05 m), and ε is the black globe emissivity (0.95).

$$\begin{array}{c}{v}_x=v_m\times {\displaystyle \frac{\mathit{log}(\frac{z}{z_0})}{\mathit{log}(\frac{m}{z_0})}}\end{array}$$

(2)

where $v_x$ represents the wind speed at height z, m/s, $v_m$ is the wind speed at the measured height, m/s, and $z_0$ is the roughness parameter, with a value of 0.1.

In addition to UTCI, the Thermal Acceptability Range (TAR) is another key metric for quantifying outdoor thermal comfort, illustrating the dynamic adaptability of residents to thermal conditions in their environment. ASHRAE Standard 55 [33] specifies that acceptable thermal conditions are those in which at least 80% of respondents find the conditions acceptable.

(2)

Air quality indicators

The Air Quality Index (AQI) is a quantitative measure used to specifically assess air quality levels. According to the NAAQS definition [35], air quality can be categorized into six classes, as shown in

Table 2. Air pollutant concentrations and corresponding air quality AQI classification evaluation standard.

PM2.5 Concentration (μg/m3)	PM10 Concentration (μg/m3)	AQI	Air Quality Level	Air Quality Grade
0.0–12.0 (24 h)	0–54 (24 h)	AQI ≤ 50	Excellent	I
12.1–35.4 (24 h)	55–154 (24 h)	50 < AQI ≤ 100	Good	II
35.5–55.4 (24 h)	155–254 (24 h)	100 < AQI ≤ 150	Mild Pollution	III
55.5–150.4 (24 h)	255–354 (24 h)	150 < AQI ≤ 200	Moderate Pollution	IV
150.5–250.4 (24 h)	355–424 (24 h)	200 < AQI ≤ 300	Heavy Pollution	V
250.5–350.4 (24 h)	425–504 (24 h)	300 < AQI ≤ 400	Severe Pollution	VI
350.5–500.4 (24 h)	505–604 (24 h)	400 < AQI ≤ 500	Severe Pollution	VI

.

Specifically, based on the concentration of PM2.5, the corresponding sub-item relative index is derived according to a specified calculation formula and is designated as the Air Quality Index (AQI) value for this pollutant. Meanwhile, sub-item indices for other pollutants, such as PM10 and CO, are incorporated into the reference framework. If the sub-item relative indices of these pollutants are greater than or equal to the sub-index corresponding to PM2.5, the larger value is selected as the final AQI. This approach ensures that the AQI accurately reflects the comprehensive impact of pollutants on air quality. The corresponding relationship between the AQI limit values of PM2.5 and its mass concentration limits is presented in Table 2. Accordingly, in this study, meteorological and pollution data obtained from field measurements were converted into the corresponding AQI values using the prescribed calculation formula. The computational method for pollutant values adopts an exponential interpolation calculation, as expressed in Equation (3):

$$\begin{array}{c}AQI={\displaystyle \frac{I\mathrm{high}-I\mathrm{low}}{\rho \mathrm{high}-\rho \mathrm{low}}}\left(\rho -\rho \mathrm{low}\right)+I\mathrm{l}\mathrm{o}\mathrm{w}\end{array}$$

(3)

where I is the Air Quality Index (AQI), which is the output value, ρ is the daily average mass concentration of PM2.5, which is the input value, I_low is the index limit corresponding to ρ_low, a constant, I_high is the index limit corresponding to ρ_high, a constant, ρ_low is the mass concentration limit less than or equal to ρ, a constant, and ρ_high is the mass concentration limit greater than or equal to ρ, a constant.

Among them, the constants of the index limits and mass concentration limits are shown in

Table 3. Air quality index limits and mass concentration limits for PM2.5.

Air Quality Grade	AQI	Mass Concentration Limit ρ/(μg/m3)
I	0	0
II	50	35
III	100	75
IV	150	115
I	200	150
VI	300	250
VI	400	350
VI	500	500

.

(3)

Overall Comfort Vote

The Overall Comfort Vote (OCV) is an index employed to comprehensively evaluate human overall comfort in a given environment, encompassing multiple perceptual dimensions such as thermal comfort, air quality comfort, visual comfort, and acoustic comfort [36]. Unlike indicators focusing on a single environmental factor, OCV emphasizes the combined effects of multiple environmental factors and their collective influence on human perception. In studies of the built environment, OCV is typically obtained through questionnaires or subjective evaluation scales, whereby participants are asked to rate their overall comfort according to their personal perceptions. Quantification is generally conducted using a Likert scale, such as a 7-point or 5-point scale [37].

3. Results

3.1. Measured Physical Environmental Characteristics of Outdoor Spaces

On-site measurements of the outdoor environment were conducted over 8 days. See

Table 4. Winter physical environment measurements.

		A1	B1	C1	D1	E1	F1	A2	B2	C2	D2	E2	F2
Ta	Max	2.4	1.5	2.3	10.1	2.7	3.2	8.5	8.6	8.5	9.3	7.7	7.4
	avg.	1.0	−0.1	0.8	1.6	−0.1	0.7	5.0	4.8	5.4	5.3	4.4	4.6
	Min	−2.6	−3.6	0.6	−3.6	−3.2	−3.0	1.3	0.3	0.6	0.7	0.5	1.1
RH	Max	47.8	47.6	47.6	51.5	46.5	45.3	53.5	56.5	52.4	54.3	57.9	55.3
	avg.	38.4	38.4	38.8	38.8	37.9	36.4	37.0	35.7	35.1	34.2	35.5	34.8
	Min	27.3	26.5	27.3	27.0	27.7	24.9	28.8	27.9	28.7	26.5	28.7	28.2
Va	Max	1.9	2.2	1.6	5.9	1.6	1.5	2.3	1.2	1.3	1.5	1.0	0.8
	avg.	0.7	1.3	0.5	1.3	0.6	0.6	0.4	0.5	0.5	0.7	0.5	0.4
	Min	0.0	0.1	0.0	0.0	0.0	0.0	0.0	0.0	0.1	0.0	0.0	0.0
G	Max	685.9	788.1	4441.8	147.7	149.3	618.0	95.7	693.2	72.8	615.2	400.8	465.5
	avg.	138.1	139.9	550.4	15.3	67.4	211.7	51.2	269.6	41.2	271.9	94.9	55.2
	Min	6.3	3.3	0.0	0.0	7.4	0.0	18.7	24.6	4.4	0.0	20.0	0.0
Tg	Max	4.9	3.1	19.5	18.4	2.8	20.2	17.9	17.0	23.4	23.6	17.9	17.9
	avg.	1.4	1.5	2.8	1.9	1.5	9.9	6.4	8.1	12.4	13.9	6.4	6.4
	Min	0.2	0.0	0.0	0.0	0.0	1.4	1.7	1.3	2.3	4.0	1.7	1.7
UTCI	Max	9.1	6.4	23.8	19.1	3.7	24.3	23.8	20.8	27.6	26.4	21.5	19.6
	avg.	3.2	0.1	2.8	−5.5	1.3	5.6	6.3	8.9	14.2	16.1	6.6	6.7
	Min	−8.4	−11.3	−7.2	−23.2	−6.1	−6.2	−1.5	0.5	1.4	1.6	1.8	2.1
PM2.5	Max	31.0	34.0	35.0	49.0	35.0	33.0	58.7	64.7	64.7	64.1	66.1	62.5
	avg.	18.6	17.3	18.3	14.5	19.0	19.8	47.7	49.1	49.8	49.3	51.8	51.4
	Min	1.0	4.0	0.0	1.0	1.0	8.0	38.4	41.1	41.1	40.6	43.4	43.0

Note. Definitions for the environmental parameters in this table: T_a (Air Temperature, °C); RH (Relative Humidity, %); Va (Wind Speed, m/s); G (Global Solar Radiation, W/m²); T_g (Black Globe Temperature, °C); UTCI (Universal Thermal Climate Index, °C); PM2.5 (μg/m³).

. The definitions for the environmental parameters in this table are as follows: T_a (Air Temperature, °C); RH (Relative Humidity, %); Va (Wind Speed, m/s); G (Global Solar Radiation, W/m²); T_g (Black Globe Temperature, °C); UTCI (Universal Thermal Climate Index, °C); PM2.5 (μg/m³).

Table 5. Summer physical environment measurements.

		A1	B1	C1	D1	E1	F1	A2	B2	C2	D2	E2	F2
Ta	Max	38.6	35.8	36.1	36.7	37.0	36.2	35.3	38.5	35.2	36.6	36.0	36.3
	avg.	36.8	35.2	35.0	35.3	35.6	35.4	34.5	35.9	34.5	35.8	33.8	35.3
	Min	33.5	34.1	33.2	33.6	32.3	34.0	32.2	31.7	32.5	33.8	29.9	32.9
RH	Max	63.9	61.8	65.3	60.6	63.3	55.8	67.7	71.2	66.3	67.0	73.1	57.3
	avg.	53.1	56.1	56.9	54.5	53.4	50.7	57.4	47.2	57.4	58.4	47.8	50.2
	Min	48.2	54.3	52.7	51.1	49.1	47.4	53.2	38.5	53.8	55.9	37.3	46.4
Va	Max	1.0	1.1	0.8	2.1	0.7	0.9	0.1	0.8	0.8	0.6	0.5	1.2
	avg.	0.7	0.7	0.6	1.3	0.6	0.7	0.1	0.6	0.7	0.4	0.3	0.9
	Min	0.4	0.5	0.4	0.9	0.3	0.6	0.0	0.4	0.5	0.3	0.2	0.7
G	Max	582.7	671.9	96.1	557.1	609.7	561.1	33.4	761.5	361.5	411.0	320.6	508.8
	avg.	373.3	299.6	25.3	357.2	477.0	311.5	22.1	524.9	137.9	168.1	139.9	71.3
	Min	46.9	32.2	8.2	79.6	40.3	0.0	6.9	93.2	0.0	11.4	46.5	0.0
Tg	Max	49.8	44.0	54.4	48.3	42.8	41.3	35.5	40.0	36.8	37.8	37.2	38.0
	avg.	41.5	38.9	42.0	43.1	36.3	38.9	34.7	37.5	35.5	36.0	34.0	36.5
	Min	33.6	35.5	35.6	37.2	31.8	37.1	32.3	32.0	32.5	33.9	30.0	32.9
UTCI	Max	49.3	44.3	50.0	47.7	42.1	41.5	41.2	45.5	41.0	43.0	42.8	43.5
	avg.	42.8	40.2	41.9	43.6	38.2	39.7	39.8	42.3	39.7	41.5	39.0	41.2
	Min	36.4	37.7	38.0	38.9	34.2	38.5	37.0	35.5	36.8	38.5	34.0	37.0
PM2.5	Max	36.4	37.7	38.0	38.9	34.2	38.5	38.4	8.1	8.3	7.3	34.1	24.0
	avg.	36.4	37.7	38.0	38.9	34.2	38.5	26.9	6.9	6.8	6.5	29.6	18.9
	Min	36.4	37.7	38.0	38.9	34.2	38.5	18.7	5.9	5.2	5.4	25.2	9.9

Note. Definitions for the environmental parameters in this table: Ta (Air Temperature, °C); RH (Relative Humidity, %); Va (Wind Speed, m/s); G (Global Solar Radiation, W/m²); T_g (Black Globe Temperature, °C); UTCI (Universal Thermal Climate Index, °C); PM2.5 (μg/m³).

presents the average, maximum, and minimum values of air temperature, relative humidity, wind speed, global solar radiation, black globe temperature, UTCI, and PM2.5 concentration at each measurement point in (1) Paris Haoting and (2) Jindi Sun City, respectively.

Regarding air temperature, the mean value in winter was 2.83 °C (maximum: 10.1 °C, minimum: −3.63 °C). The average temperature in Paris Haoting (1) was 3.32 °C lower than in Jindi Sun City (2), indicating a significant temperature difference in Wuhan between cold spell periods and non-cold spell periods in winter. The summer average temperature was 35.27 °C (maximum: 38.6 °C, minimum: 29.9 °C). The average temperature in Paris Haoting was 1.47 °C higher than in Jindi Sun City, reflecting noticeable inter-community temperature differences during summer.

In terms of relative humidity, the winter average was 38.16%, with a maximum of 51.5% and a minimum of 24.9%, showing relatively small variation. The summer average was 53.75%, with a maximum of 73.1% and a minimum of 37.3%, showing larger variation, indicating that humidity has a more significant impact on thermal comfort in summer. The winter average black globe temperature was 1.43 °C, with a maximum of 19.5 °C and a minimum of 0 °C; the summer average was 41.75 °C, with a maximum of 54.4 °C and a minimum of 30.0 °C, reflecting the significant influence of solar radiation on the thermal environment. The winter average wind speed was 0.68 m/s, with a maximum of 5.9 m/s and a minimum of 0 m/s; the summer average was 0.69 m/s, with a maximum of 2.1 m/s and a minimum of 0 m/s. Variations in wind speed significantly impact thermal comfort, especially in summer, where higher wind speeds help alleviate thermal discomfort.

For relative humidity, the winter average was 38.16% (maximum: 51.5%, minimum: 24.9%), showing relatively limited variation. In contrast, the summer average was 53.75% (maximum: 73.1%, minimum: 37.3%), exhibiting greater fluctuation, which suggests a more significant impact of humidity on thermal comfort in summer. The average black globe temperature was 1.43 °C in winter (maximum: 19.5 °C, minimum: 0 °C) and 41.75 °C in summer (maximum: 54.4 °C, minimum: 30.0 °C), underscoring the substantial influence of solar radiation on the thermal environment. The average wind speed was 0.68 m/s in winter (maximum: 5.9 m/s, minimum: 0 m/s) and 0.69 m/s in summer (maximum: 2.1 m/s, minimum: 0 m/s). Variations in wind speed significantly impact thermal comfort, with higher speeds providing greater relief from thermal discomfort in summer.

The average UTCI was 3.21 °C in winter (maximum: 23.8 °C, minimum: −23.2 °C), indicating generally poor thermal comfort conditions during the humid-cold winter. The summer average was 41.25 °C (maximum: 50.0 °C, minimum: 34.0 °C), reflecting the strong stress of high temperatures on thermal comfort. The average PM2.5 concentration was 18.43 μg/m³ in winter (maximum: 68.7 μg/m³, minimum: 0 μg/m³) and 26.6 μg/m³ in summer (maximum: 38.9 μg/m³, minimum: 5.2 μg/m³). Elevated PM2.5 concentrations in winter may be associated with heating activities and traffic flow, whereas lower summer concentrations are likely related to increased rainfall and vegetation coverage.

3.1.1. Analysis of Measured Thermal Environment Results

Based on the measurement results (see Figure 3), winter temperatures are characterized by lower values and greater fluctuations, while summer features higher temperatures with significant humidity variations, both seasons markedly affecting thermal comfort. Distinct seasonal differences in black globe temperature and wind speed further influence the thermal environment. UTCI results indicate poor thermal comfort in winter and significant thermal discomfort in summer. Seasonal variation in PM2.5 concentration reflects changes in air quality, with high winter concentrations linked to heating and traffic emissions, and lower summer levels correlated with rainfall and vegetation cover. These measurements provide a crucial scientific foundation for understanding seasonal variations in thermal comfort and pollutant status within typical Wuhan residential areas, facilitating the evaluation of outdoor environmental quality.

Further analysis of Figure 3 elucidates key spatiotemporal patterns driving thermal discomfort. Winter UTCI (Figure 3a) shows high sensitivity to solar access and urban morphology, with exposed sites (e.g., A1, F1) experiencing brief daytime warming while shaded areas remain cold. In stark contrast, summer (Figure 3b) is characterized by prolonged periods of severe heat stress (UTCI > 40 °C), with minor relief provided only by dense shading. This persistent thermal stress directly shapes the subjective thermal perceptions reported by residents.

3.1.2. Analysis of Measured Pollutant Results

Further analysis of PM2.5 concentrations across different measurement points within the same residential communities (see Figure 4) revealed significant spatial heterogeneity. Specifically, point F2 exhibited the highest average PM2.5 concentration (36.5 μg/m³), whereas B1 and D2 had relatively lower averages (13.8 μg/m³ and 22 μg/m³, respectively). Diurnal variations were also observed, with concentrations peaking around 9:00 and reaching a trough at 15:00, potentially related to patterns in human activities and traffic flow.

Considering combined spatiotemporal variations, PM2.5 concentrations at both sites exhibited periodic characteristics. Human activities were identified as a significant influencing factor. Overall, PM2.5 concentrations were higher in Jindi Sun City than in Paris Haoting, potentially attributable to differences in population density, traffic flow intensity, and surrounding anthropogenic emission sources. Additionally, the diurnal fluctuation range of PM2.5 concentration was greater in Jindi Sun City, a phenomenon possibly related to specific climatic conditions and traffic pollution levels during the measurement period. Analysis of other meteorological data over time showed a weak correlation between air temperature and pollutant distribution within the two communities.

The spatiotemporal patterns in Figure 4 reveal the direct impact of local sources and sinks on PM2.5. The consistently high concentrations at point F2 are attributable to its proximity to major traffic arteries and limited green space, inhibiting dispersion. In contrast, points with better ventilation and vegetation (e.g., B1, D2) maintained lower levels. The characteristic diurnal cycle—peak during morning rush hour and afternoon minimum—closely follows traffic emissions and atmospheric mixing conditions. Notably, winter (Figure 4a) exhibits both higher concentrations and greater variability than summer (Figure 4b), consistent with seasonal emission patterns and meteorological stagnation. These findings provide the physical context for the residents’ air quality satisfaction votes and its interplay with thermal sensation.

Distribution statistics for all PM2.5 concentration data from both communities (see Figure 5) indicate that on the measurement day (20 January 2024), the overall median concentration was 30 μg/m³ (Paris Haoting: 24 μg/m³; Jindi Sun City: 32 μg/m³). The two studied communities share similar orientation and layout forms and are surrounded by roads with relatively consistent traffic flow, enhancing the representativeness of the distribution results.

On the other hand, the statistical summary in Figure 5 quantifies the significant inter-community difference in PM2.5 exposure. Jindi Sun City’s higher median, larger interquartile range, and elevated maximum concentration demonstrate a more polluted and variable environment compared to Paris Haoting. This validated contrast between the two sites is fundamental to our subsequent investigation, as it creates a natural experiment for assessing how different AQI levels modulate the relationship between thermal environment and comfort.

3.2. Outdoor Voting Characteristics of Wuhan Residents

3.2.1. Basic Data Analysis of Wuhan Resident Respondents

(1)

Respondent Age and Gender Distribution

A total of 597 questionnaires were distributed across winter and summer, with 582 valid responses retained after screening. The survey controlled for gender and age factors, as shown in

Table 6. Gender ratio of respondents in residential areas and proportion of age of respondents in residential areas of Wuhan.

Gender	Percentage	Age	Percentage
		<18	10%
male	51%	18–35	30%
		35–50	16%
female	49%	50–60	13%
		>60	10%

. Males accounted for 51% of respondents and females 49%, indicating a balanced gender ratio. Minors were excluded from the analysis as their environmental perceptions differ significantly from adults and they constituted a small proportion of the sample.

(2)

Characteristics of Residents’ Activity Time Distribution

The primary outdoor activity times for residents in each season are shown in Figure 6. In summer, peak activity concentrated in the evening (18:00–20:00), with the proportion of votes particularly high at 19:00 (13%). In winter, the peak period occurred in the afternoon (15:00–17:00), with the highest vote proportion at 16:00 (13%). This pattern is likely related to sunlight availability, temperature variations, and resident lifestyles. Summer evenings are relatively comfortable, encouraging outdoor activities, while winter afternoons offer ample sunshine and warmer temperatures, making them more favorable for going out.

The characteristics of residents’ daily outdoor activity duration are shown in

Table 7. Statistics on residents’ daily average activity duration preferences in different seasons.

Average Activity Duration (Min/Day)	Percentage of Summer Votes	Percentage of Winter Votes
0–14	6.61%	6.22%
15–29	14.02%	10.27%
30–60	34.92%	22.16%
>60	44.44%	61.35%

. Analysis of vote distribution across different time intervals (min/day) revealed differences in the concentration of activity times between seasons. In summer, the highest proportion of votes (34.92%) was for the 30–60 min/day interval; in winter, the highest proportion was for the >60 min/day interval. This suggests that summer activity times are more concentrated, typically within half an hour to an hour, while winter activity times are more dispersed, with a larger proportion exceeding one hour.

Based on these findings and considering the impact of seasonal differences, 9:00 and 15:00 were selected as typical analysis periods for both winter and summer.

3.2.2. Relationship Between Thermal Sensation Vote (TSV) and Thermal Comfort Vote (TCV)

Regression analysis was employed to fit curves and examine the quantitative relationship between the Thermal Sensation Vote (TSV) and Thermal Comfort Vote (TCV) for residents in Wuhan residential areas (see Figure 7). The results reveal a non-linear relationship. As the TSV value increases, the TCV value first increases and then decreases, exhibiting a clear threshold. When TSV is between −1.35 and 1.88, TCV ≥ 0, with TCV reaching its maximum at TSV = 0.26. This indicates that Wuhan residents experience the highest comfort within the “slightly cool” to “slightly warm” thermal sensation range, peaking under nearly neutral thermal conditions.

Further comparative analysis incorporating different pollutant concentrations and seasons (see Figure 8) found that while the pollutant concentration gradient did not drastically alter the overall shape of the TSV-TCV curve, differences in data distribution between winter and summer suggest an indirect influence on the comfort threshold. For instance, the peak TCV value corresponding to TSV = 0.26 in summer was slightly lower than in winter, implying that hot and humid conditions may attenuate the comfort experienced in neutral thermal states. The model’s overall explanatory power (R² = 0.6393) is moderate, indicating that besides thermal sensation, other environmental variables such as air pollution and wind speed exert superimposed effects on thermal comfort perception, warranting further quantification.

3.2.3. Relationship Between Air Quality Index (AQI) and Air Satisfaction Vote (ASV)

First, according to the AQI calculation method introduced earlier in Section 2.3.2, the measured PM2.5 concentration data were converted into the corresponding air quality categories shown for the pollutant indicator section and averaged. Then, by fitting AQI and ASV data, Equation (4a) for the correlation between ASV and PM2.5 concentration and Equation (4b) for the correlation between ASV and AQI were ultimately obtained through statistical methods, as shown below:

$$\begin{array}{c}ASV=1.10061{PM2.5}^2+1.6087PM2.5+29.494\left(R^2=0.9227\right)-273.15\end{array}$$

(4a)

$$\begin{array}{c}ASV=-0.0061AQI+0.8723\left(R^2=0.9791\right)\end{array}$$

(4b)

Within a certain range, there is a good linear relationship between Air Satisfaction Vote (ASV) and the Air Quality Index (AQI). Simultaneously, the difference in the correlation coefficient (R²) compared to that between PM2.5 concentration and ASV is small, proving that the effect of these two indicators on ASV is extremely similar. PM2.5 concentration can largely substitute for AQI’s role in characterizing air quality. From Figure 9, it can be seen that for every increase of approximately 90 in AQI, satisfaction decreases by one unit, indicating that Wuhan residents perceive air quality accurately. Better air quality leads to higher satisfaction. When ASV = 0, the predicted neutral AQI is calculated as 131, meaning no tendency towards satisfaction or dissatisfaction with the air environment. Even when AQI is 0, satisfaction with air quality does not reach a very high level, implying that for Wuhan residents, acceptance of air quality within their residential communities is relatively high, and also that factors other than PM2.5 may influence residents’ judgments.

4. Quantitative Evaluation of Outdoor Thermal Comfort Under Different Air Quality Levels

4.1. Thermal Sensation Vote (TSV)

By analyzing the thermal acceptability range under different AQI levels, the Mean Thermal Sensation Vote (MTSV) was plotted for each 1 °C interval of the Universal Thermal Climate Index (UTCI) and compared with a quadratic curve. The influence of air quality levels corresponding to different PM2.5 concentrations on thermal comfort evaluation and the associated patterns were analyzed. Based on the AQI classification method described in Table 2, thermal sensation regression models were developed for winter, summer, and annual conditions under varying pollutant levels (see Figure 10 and Figure 11).

The results indicate seasonal differences in the linear relationship between UTCI and MTSV (see

Table 8. Regression equations of UTCI and mean thermal sensation (MTSV) under different seasons and AQI levels.

Thermal Comfort Index	Classification		Regression Model of Universal Thermal Climate Index (UTCI) and Mean Thermal Sensation
UTCI	AQI-I		$\begin{array}{c}MTSV=0.0486UTCI-1.1822\left(R^2=0.7763\right)\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\left(5\mathrm{a}\right)\end{array}$
	AQI-II		$\begin{array}{c}MTSV=0.0377UTCI-0.7744\left(R^2=0.6261\right)\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\left(5\mathrm{b}\right)\end{array}$
	Summer	AQI-I	$\begin{array}{c}MTSV=0.0982UTCI-3.5146\left(R^2=0.5058\right)\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\left(5\mathrm{c}\right)\end{array}$
		AQI-II	$\begin{array}{c}MTSV=0.0542UTCI-1.9597\left(R^2=0.3152\right)\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\left(5\mathrm{d}\right)\end{array}$
	Winter	AQI-I	$\begin{array}{c}MTSV=0.0697UTCI-1.2123\left(R^2=0.5748\right)\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\left(5\mathrm{e}\right)\end{array}$
		AQI-II	$\begin{array}{c}MTSV=0.0268UTCI-0.65\left(R^2=0.4232\right)\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\left(5\mathrm{f}\right)\end{array}$

), with model explanatory power varying with pollutant levels. Under winter conditions, the regression model MTSV = 0.0486UTCI − 1.1822 for the AQI-II level exhibits a high goodness-of-fit (R² = 0.7763), suggesting a strong correlation between thermal sensation and UTCI in low-temperature environments, where PM2.5 concentration has a stable impact on thermal comfort evaluation. Under summer conditions, the model MTSV = 0.0982UTCI − 3.5146 (Equation (5c)) corresponding to the AQI-I level has the largest slope, but the R² value decreases to 0.5058. This indicates that in high-temperature and high-humidity environments, the interfering effect of pollutant concentration on thermal sensation increases, reducing the model’s explanatory power. In the comprehensive annual analysis, the R² of MTSV = 0.0542UTCI − 1.9597 (Equation (5d)) under the AQI-II level is only 0.3152, further confirming that seasonal factors significantly modulate thermal comfort evaluation.

Under the same pollutant level, the R² values of winter models (e.g., Equations (5e) and (5f)) are generally higher than those of summer models (e.g., Equations (5c) and (5d)). This implies that thermal comfort evaluation in winter is more likely to follow the linear relationship with UTCI, while in summer, due to the influence of combined environmental factors, thermal sensation votes exhibit greater dispersion. Furthermore, the fluctuation of R² values (0.3152–0.7763) across different AQI levels reveals the non-linear influence mechanism of PM2.5 concentration on thermal comfort evaluation. Deteriorating air quality may weaken the statistical correlation between UTCI and MTSV by altering human thermal adaptation behaviors.

The obtained equations are as follows:

4.2. Thermal Acceptability Range (TAR)

To investigate human thermal comfort within specific temperature ranges, the thermal acceptability range is commonly used for evaluation. Following the Thermal Acceptability Range (TAR) defined by ASHRAE Standard 55 [33], this study adopted the UTCI temperature range acceptable to 80% of respondents. The thermal acceptability range under different AQI levels was analyzed, and the percentage of thermal unacceptability for each 1 °C interval of UTCI was plotted and compared with a quadratic curve.

Overall, residents in Wuhan exhibit relatively high acceptance of the winter thermal environment (see

Table 9. Acceptable UTCI ranges for residents in Wuhan in different seasons.

Season	Acceptable UTCI Range for Residents in Wuhan (°C)
Summer	19.17–33.98
Winter	7.12–20.15

). The quadratic curve fitted to the data in Figure 12 shows an R² of 0.5414 between thermal acceptability and UTCI. Based on the intersection points between the curve and the 20% unacceptability threshold, the overall 80% thermal acceptability range is determined to be 13.3–27.1 °C. Figure 13a,b plot the intersection points between two quadratic fitting curves and the 20% unacceptability threshold under different AQI levels (AQI-I and AQI-II) in different seasons. The R² values of the two fitted quadratic curves are 0.563 and 0.258, respectively, indicating that under the AQI-I air quality level, the relationship between thermal acceptability and UTCI is more reliable. The lower limit of TAR for AQI-I is 7.1 °C, and that for AQI-II is 8.8 °C, both lower than the lower limits of their respective overall neutral ranges. Similarly, the summer UTCI (30.9 °C) is higher than the upper limits of their respective neutral ranges. This indicates that residents in Wuhan exhibit considerable tolerance to changes in the outdoor thermal environment.

4.3. Comprehensive Impact of Air Quality and UTCI on Thermal Comfort (TCV)

In this study, UTCI was grouped at 1 °C intervals, and AQI was classified according to the “Ambient Air Quality Standard” (GB 3095-2012) [38]. This grouping clarifies the distribution between UTCI and Thermal Comfort Vote (TCV) and facilitates the identification of goodness-of-fit.

As shown in Figure 14, which presents the variation trend of TCV under different air quality levels and thermal environments, significant differences in human thermal comfort are observed at the same UTCI. Regardless of air quality level changes, TCV increases with UTCI, peaks at approximately 16 °C, and then decreases under overheated conditions. Meanwhile, under identical UTCI conditions, better air quality leads to higher thermal comfort evaluations. That is, improved air quality enhances the perceived comfort of the thermal environment for Wuhan residents, particularly within the 0–10 °C temperature range.

Based on Figure 14 and Figure 15, polynomial fitting was conducted, and the resulting equations are shown in

Table 10. Regression equations of UTCI and TCV under different seasons and AQI levels.

Air Quality	Season	Thermal Comfort Index UTCI and Average Thermal Comfort Regression Model
AQI-I	Summer	$\begin{array}{c}TCV=-0.0081{UTCI}^2+0.7564UTCI-17.155\left(R^2=0.67\right) \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\left(6\mathrm{a}\right)\end{array}$
	Winter	$\begin{array}{c}TCV=-0.0015{UTCI}^2+0.0554UTCI-0.415\left(R^2=0.6572\right) \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\left(6\mathrm{b}\right)\end{array}$
	Throughout the year	$\begin{array}{c}TCV=-0.004{UTCI}^2+0.034UTCI-0.5275\left(R^2=0.3471\right)\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\left(6\mathrm{c}\right)\end{array}$
AQI-II	Summer	$\begin{array}{c}TCV=-0.0048{UTCI}^2+0.1552UTCI-1.2473\left(R^2=0.3273\right) \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\left(6\mathrm{d}\right)\end{array}$
	Winter	$\begin{array}{c}TCV=-0.0186{UTCI}^2+1.7321UTCI-39.758\left(R^2=0.4295\right) \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\left(6\mathrm{e}\right)\end{array}$
	Throughout the year	$\begin{array}{c}TCV=-0.003{UTCI}^2+0.0356UTCI-0.7103\left(R^2=0.3471\right)\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\left(6\mathrm{f}\right)\end{array}$

. A comprehensive analysis reveals that the relationship between UTCI and TCV under different air quality levels conforms to the presented equations. Air quality significantly impacts TCV, and residents provide more sensitive evaluations of thermal comfort under better air quality conditions. Conversely, in cold winter environments, the relationship between air quality and TCV weakens. When residents experience extreme discomfort, the influence of air quality factors on TCV decreases, and UTCI becomes the dominant factor. This phenomenon may occur because residents pay less attention to air quality perception and its impact under extreme cold discomfort, which aligns with empirical expectations. It can thus be inferred that good air quality may enhance mood, thereby improving thermal acceptability. Conversely, poor air quality with severely reduced visibility can negatively affect mood and the evaluation of the thermal environment, consistent with life experience and theoretical expectations. Due to space limitations, this study did not explore the specific physiological pathways through which air quality affects thermal comfort.

5. Discussion

The findings of this study both corroborate and extend existing knowledge on the interplay between thermal comfort and air quality, particularly in the context of high-density residential environments in hot-summer–cold-winter regions.

5.1. Thermal Tolerance and Regional Adaptability

A comparative analysis with Bian et al.’s work [26] in urban parks reveals a distinct adaptive capacity among residential inhabitants. Notably, the climatic and urban environmental contexts of the two studies differ substantially: while Bian et al.’s research was conducted in Tianjin, a city in the cold climate zone of northern China, the present study focuses on Wuhan, located in the hot-summer–cold-winter region of central China. These regional distinctions are characterized by stronger urban heat island effects in Wuhan, especially during the summer, due to its higher background temperatures, humidity levels, and intensive urban development. Specifically, the Thermal Acceptability Range (TAR) thresholds identified in our study extend significantly beyond the conventional neutral UTCI spectrum, suggesting that the habitual exposure and functional requirements characteristic of residential environments—coupled with the region-specific climatic stressors—may foster enhanced thermal tolerance compared to recreational park settings.

5.2. Limitations and Future Research Directions

This study has several limitations that should be considered. First, the quantitative models are specific to Wuhan’s climate and cultural context, limiting their generalizability; however, the core finding that the thermal environment and air quality exhibit significant interaction, modulated by seasonal variations and comfort thresholds, provides a valuable conceptual framework for other high-density cities in hot-summer–cold-winter regions. Validation across diverse geographical and climatic zones is necessary.

Furthermore, this study primarily relied on the integrative UTCI to characterize the thermal environment. While this provides a robust physiological assessment, future research with larger datasets could employ advanced statistical modeling, such as multi-variable regression with interaction terms or machine learning, to disentangle the complex and potentially non-linear interactions between individual meteorological parameters and air pollutants in shaping occupant perception.

Finally, the underlying physiological and psychological mechanisms by which air quality influences thermal perception were not directly measured. Interdisciplinary studies combining physiological monitoring with environmental surveys are recommended to elucidate these causal pathways.

6. Conclusions

This study selected typical residential areas in Wuhan to conduct research on outdoor thermal comfort in winter in the hot-summer–cold-winter region. Through questionnaire surveys and outdoor physical environment measurements, taking Wuhan as an example, it investigated the impacts of air quality and thermal environment on residents’ thermal comfort in this region and further explored the characteristics of outdoor thermal comfort and its correlation with microclimatic factors. Based on the obtained results, the following key conclusions have been drawn:

(1)

For every increase of approximately 90 in AQI, satisfaction decreases by one unit for the residence in the studied case. Additionally, this study found that PM2.5 concentration can largely substitute for AQI’s role in characterizing air quality for the satisfaction levels.

(2)

The explanatory power of the UTCI-MTSV model varies substantially with pollution levels (R2: 0.315–0.776), revealing significant modulation by seasonal and environmental context. Thermal comfort in winter largely adheres to the linear relationship with UTCI, whereas summer exhibits greater vote dispersion due to compounded environmental influences. The lower R2 in summer suggests behavioral adaptations (e.g., seeking shade, adjusting activity), and psychological effects of poor air quality introduce variability not captured by physiological indices alone. These findings highlight the non-linear influence of PM2.5 on thermal perception and underscore the need for integrated models that incorporate microclimatic, pollution, and behavioral data to better predict comfort in complex urban settings.

(3)

Residents in Wuhan have a stronger tolerance to changes in the outdoor thermal environment regardless of the AQI levels tested. The lower limit of TAR for AQI-I is 7.1 °C, and that for AQI-II is 8.8 °C, both of which are lower than the lower limits of their respective overall neutral ranges. For the summer, UTCI (30.9 °C) is higher than the upper limits of their respective neutral ranges. However, the fitting between thermal acceptability and UTCI has higher reliability with better air quality, further hinting the increased influence of air quality on the residence’s thermal comfort perception.

(4)

TCV increases with the increase in UTCI, reaches a peak at approximately 16 °C, and then decreases under overheated conditions. Under the same UTCI conditions, better air quality leads to higher thermal comfort evaluation by people, especially within the temperature range of 0–10 °C.

In summary, this study advances the field of outdoor environmental research through two key contributions. First, it establishes an integrated empirical methodology specifically tailored for high-density residential environments, combining synchronized microclimatic measurements, pollutant monitoring, and subjective comfort surveys. Second, it reveals PM2.5’s seasonally modulated interference with thermal perception—a non-linear effect particularly pronounced in summer, where behavioral adaptations and psychological factors amplify the dispersion of thermal sensation votes. Unlike previous studies, this research specifically captures the unique environmental dynamics of high-density urban neighborhoods in hot-summer–cold-winter regions, characterized by strong urban heat island effects and remarkable resident thermal tolerance. These findings provide a new framework for understanding and addressing the combined challenges of heat stress and air pollution in sustainable urban design.