Experimental investigation of the effect of surgical masks on outdoor thermal comfort in Xiamen, China

Abstract

The COVID-19 pandemic has significantly changed people's lifestyles, and wearing surgical masks in outdoor public spaces has become commonplace. However, few studies have explored the impact of wearing masks on outdoor thermal comfort in different seasons. From May 2021 to February 2022, a series of longitudinal experiments were conducted in Xiamen, China to examine the effect of wearing surgical masks on outdoor thermal comfort. Forty-two participants took part in the experiments with and without masks. During the experiments, the thermal perceptions of the subjects and environmental thermal parameters were collected. Differences in outdoor thermal comfort between subjects wearing masks and those not wearing masks were determined in summer, autumn, and winter. Results showed that 1) the subjects wearing masks had lower neutral temperatures, and this difference was particularly pronounced in summer and exacerbated by walking; 2) in warm environments, masks reduced thermal comfort, and discomfort associated with masks was worse when walking than when sitting; 3) wearing masks significantly worsened facial comfort and increased chest discomfort, as summer turned to winter, the impact of masks on facial comfort decreased; 4) radiation and air temperature were the environmental parameters with the greatest impact on outdoor thermal sensation. Subjects who wore masks preferred lower temperatures, radiation, and humidity, and higher wind speeds.

Graphical abstract

1. Introduction

Thermal comfort in outdoor environments is an important factor affecting quality of life. In recent years, an increasing number of studies on outdoor thermal comfort have been performed [1]. Previous studies on outdoor thermal comfort have been conducted in different climatic regions to investigate types of human thermal sensation and activities in various outdoor spaces [[2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15]]. These studies have shown that people's thermal comfort in outdoor environments is influenced by several factors, including combinations of thermal environment factors, regional climate, seasonal variations, individual clothing status, metabolic rate, and various psychological factors. However, the COVID-19 spread in late 2019 has brought new issues to the study of outdoor thermal comfort.

COVID-19 has had an unprecedented impact on the world and has significantly changed the way people live their lives. Masks effectively prevent COVID-19 infection and have thus become essential items in people's daily activities, such as travel, work, and study [16]. However, studies have shown that masks can cause discomfort. For example, long-term mask use can cause breathing problems [17], and the temperature inside a mask has significant effects on the thermal sensation of the human body [18]. Wearing a mask has been shown to cause slight changes in some physiological indicators but does not produce excessive heat load on a wearer [19].

Numerous studies have examined the impact of masks on thermal comfort. Nielsen et al. [18] conducted an experiment on subjects during exercise and found that masks had a significant effect on respiratory heat loss in humans, and air temperature primarily determined thermal sensation. Roberge et al. [19] discovered that wearing a surgical mask at a low-to-moderate work rate slightly increased heart rate, facial skin temperature, core temperature, fatigue, and thermal perception. Through indoor experiments, Lin et al. [20] discovered that wearing a mask in a humid atmosphere increases heat stress and impacts metabolic heat loss. Tang et al. [21] conducted a thermal comfort study enrolling mask-wearing and mask-free borrowers in a university library in Guangzhou during summer and discovered that mask-wearers had higher environmental comfort requirements and slightly lower neutral temperature. Li et al. [22] compared surgical masks and N95 masks in terms of thermal, and the results showed significant difference between temperature and humidity inside surgical and N95 masks, and the degree of discomfort caused by wearing N95 masks was higher than that of surgical masks. Zhang et al. [23] examined the thermal comfort of mask-wearers in a climate chamber. The results indicated that wearing surgical masks would decrease summer neutral temperature by 1.5 °C and optimal temperature by 1.4 °C. In addition, some studies [20,24,25] have demonstrated that masks with breathing valves provide higher degree of thermal comfort than masks without breathing valves. However, masks with breathing valves protect the wearer but not the surrounding population, so these masks were not recommended during the COVID-19 pandemic.

However, current studies on the effects of masks on human thermal comfort have some limitations. First, most existing studies have been conducted indoors. Few studies have examined the effects of masks on thermal comfort in outdoor environments despite that the number of people wearing masks in outdoor public spaces has increased dramatically. Second, the durations of previous studies were limited, and seasonal variations in the effects of masks on thermal comfort were not analyzed. Third, the vast majority of previous studies have only examined the effect of masks on thermal comfort in a single state, such as sitting or working, whereas the effects of wearing masks in different activity states on neutral temperature, thermal comfort, and preference for thermal environmental factors are rarely discussed.

Our study aims to fill this gap by conducting a comprehensive investigation of the effects of masks on outdoor human thermal sensation and thermal comfort in different seasons and at different activity levels. This investigation was conducted at a university in Xiamen, China to collect thermal comfort questionnaires from subjects and measure environmental parameters. Analyses were performed on the effect of surgical masks on thermal comfort in outdoor thermal environments. The main goals of this study are as follows:

• (1)To compare the difference in neutral temperature between wearing masks and not wearing masks during summer, autumn, and winter.
• (2)To evaluate the impact of masks on thermal comfort in outdoor environments.
• (3)To compare the effects of masks on the thermal comfort of various body parts in different seasons.
• (4)To quantify the effects of temperature, radiation, wind speed, and relative humidity on the thermal perception of groups with and without masks.

2. Methods

2.1. Experimental location

This research was carried out in Xiamen, China (N24° 43′, E118° 10’). According to the Chinese standard of climatic regionalization for architecture, this area belongs to the hot summer and warm winter climate regions (Fig. 1 (a)), with a humid and hot climate. In the meteorological records of Xiamen (Fig. 1 (b)), the average monthly outdoor temperature ranges from 10.7 °C to 32.7 °C for over a year. The hottest month of the year is July (28.5 °C), and the coldest month is January (13.3 °C). High relative humidity persists throughout the year, fluctuating between 66% and 84%. This study lasted from May 2021 to February 2022, spanning the hottest and coldest months.

Fig. 1.

(a) Climate classification for building design in China and the location of Xiamen. Source: [26]. (b) Monthly variation in temperature and relative humidity in Xiamen (2000–2021). Source: China Meteorological Data Sharing Service System.

The experimental site was in the rooftop garden of the Architecture Experimental Building of Huaqiao University. This is a purpose-built thermal comfort experiment site for outdoor use. It contains ponds, small structures, shrubs, and ground cover plants, and has a weather station that monitors exterior thermal ambient factors and a room where temperature and humidity can be controlled. Experiments conducted in this location can reduce the impact of outdoor traffic, changes in scenery, changes in acoustic environment, and other perturbing factors on experimental outcomes.

2.2. Subjects

A total of 42 subjects participated in this study, all of whom were daily users of the rooftop garden at the experimental site and were familiar with the experimental site. All subjects were acclimatized to the climate of Xiamen, having lived there for at least 1 year. During the experiment, all subjects were in good health and did not take any prescription medication, so they can represent the general healthy adult population. The subjects' physical characteristics are listed in Table 1 . All subjects were informed of the experiment's purpose and significance. Before participating in the experiment, volunteers were required to eat a balanced diet and get adequate rest 24 h before the experiment. Subjects participating in the same experiment were asked to wear the same clothes. The subjects' dressing and thermal insulation values for clothing ensembles are shown in Table 2 , and the insulation values are based on ISO 9920 (2007).

Table 1.

Basic information of subjects.

Gender	Amount	Age, years	Height, cm	Weight, kg	BMI
Male	20	25.5 ± 3.5	175 ± 8	67.5 ± 9.5	22.05 ± 3.12
Female	22	24.5 ± 3.5	165.5 ± 5.5	53 ± 7	19.94 ± 2.35

Table 2.

The subjects’ dressing and thermal insulation values for clothing ensembles.

Season	Condition	Garment	Thermal insulation value（clo）
Summer	–	Underpants, shirt with short sleeves, light trousers, light socks, shoes	0.5
Autumn	Ta≥24 °C	Underpants, shirt, light trousers, socks, shoes	0.6
	Ta＜24 °C	Panties, shirt, trousers, jacket, socks, shoes	1
Winter	–	Underwear with long sleeves and legs, shirt, trousers, sweater, jacket, socks, shoes	1.3

2.3. Surgical masks

The mask used in this study was a disposable surgical mask. According to the standards provided by the manufacturer, the inner dimensions of the mask are 17.5 cm × 9.5 cm (±3%). The mask facepiece is divided into three layers: inner, middle, and outer, of which the inner and outer layers are made of polypropylene spunbonded nonwoven fabrics and the middle layer filter material is polypropylene melt-blown nonwoven web [27]. The rubber bands at both ends of the mask are suspended from both ears, and then the mask nasal bar is pressed to better fit the nasal bridge and cover the nose, mouth, and jaw. Related studies have shown that surgical masks can effectively block the transmission of viruses and bacteria [28,29].Experimental process.

In each experiment, two or four participants were evenly divided into experimental and control groups. The subjects in the experimental group were required to wear masks, whereas the subjects of the control group were not. During each experiment, the subjects of both groups wore identical clothing. Every experiment lasted for 1 h. Each subject was instructed to first sit in a controlled indoor environment for 30 min (with an indoor air temperature of 24.5 °C–26.5 °C and a relative humidity of 45%–55%). The purpose was to stabilize the subject's physical state and reduce errors due to difference in thermal environment experience among subjects. The subjects then entered the outdoor space for the experiment, which lasted 30 min (Fig. 3). During the experiment, the subjects were asked to fill out a thermal comfort questionnaire: one questionnaire every 5 min and six questionnaires per experiment. The outdoor experiment was divided into two exercise states, sitting and walking. Throughout the day, each subject was only permitted to take part in one sitting experiment or one walking experiment. Walking was performed on a treadmill at a speed of 1.2 m/s.The same subject was instructed to participate in the experimental and control groups an equal number of times per season. When a subject decided to withdraw from the experiment while it was in progress, the data obtained from the experiment were invalidated. We conducted the experiment on a non-rainy day from 9:30 to 11:30 or 14:30 to 16:30 to prevent the effects of eating breakfast and lunch on the subjects' thermal sensation. Each month, 12–20 experiments were conducted. The experiment was approved by the Biomedical Experimental Ethics Review Committee of Huaqiao University.

Fig. 3.

(a) Experimental procedure. (b) Preparation before the experiment. (c) Sitting experiment. (d) Walking experiment.

2.4. Questionnaire survey

The questionnaire (i.e., electronic questionnaire) was distributed to the subjects during the experiment through a cellular phone. The e-questionnaire automatically recorded the time of submission. The questionnaire was divided into two sections: personal information and subjective perception survey. The first part of the questionnaire recorded the subject's name. Given that the subjects' physiological parameters were pre-recorded, they were not required to provide information on gender, age, height, and weight. The second part included the subjects' thermal sensation vote (TSV), thermal comfort vote (TCV), discomfortable body parts vote, radiation preference vote, temperature preference vote, wind speed preference vote, and humidity preference vote. The discomfortable body parts vote is multiple choice, whereas others are single choices. The thermal comfort questionnaire was designed by ISO 105511 (2019), and the specific options are shown in Table 3 .

Table 3.

Subjective voting scale.

Thermal Sensation	Thermal Comfort	Discomfort body parts	Radiation preference	Temperature preference	Air speed preference	Humidity preference
·-4 Very cold	·0 Comfortable	·Face	·Higher	·Higher	·Higher	·Higher
·-3 Cold	·1 Slightly uncomfortable	·Chest	·No change	·No change	·No change	·No change
·-2 Cool	·2 Uncomfortable	·Abdomen	·Lower	·Lower	·Lower	·Lower
·-1 Slightly cool	·3 Very uncomfortable	·Back
·0 Neutral	·4 Extremely uncomfortable	·Upper limbs
·+1 Slightly warm		·Lower limbs
·+2 Warm		·No discomfort
·+3 Hot
·+4 Very hot

2.5. Data collection

A weather station equipped with a global temperature sensor, a wind speed recorder, and an air temperature and humidity recorder was installed at the experiment site (Fig. 2 , Table 4 ). All sensors uploaded the collected data to a cloud server. The sensors were placed at a distance of 1.1 m from the ground and recorded environmental parameters around the clock with a recording interval of 1min.

Fig. 2.

Experimental site and thermal meteorological parameter instruments.

Table 4.

Specifications of measurement devices.

Equipment	Model	Measurement parameter	Range	Accuracy
Air temperature and humidity recorder	PTS-3	Ta（°C）;	−40-80 °C;	±0.1 °C;
		RH（%）	0-100%RH	±2%RH
Global radiation sensor	TBQ-2	G（W/m2）	0–2000W/m2	＜5%
Wind speed recorder	EC-9S	Ws（m/s）	0–70 m/s	±0.3 m/s

2.6. Thermal comfort indices

In this study, the physiologically equivalent temperature (PET) [30] was selected as an objective thermal comfort index. PET is widely used in outdoor thermal comfort studies and provides a convenient way to make comparisons among studies. In this study, the software Rayman Pro [31,32] was used to calculate PET. The six parameters required in calculating PET were air temperature (T_a), relative humidity (RH), wind speed (Ws), global radiation (G), clothing thermal resistance value, and metabolic rate. T_a, RH, Ws, and G were measured directly from the weather station. The thermal resistance of the garment is shown in Table 2. Metabolic rate was estimated based on ISO 8996 (2004). The mean PET within 5 min prior to the submission of the questionnaire was used as the PET corresponding to the questionnaire results.

3. Results

3.1. Thermal parameters

Based on temperature fluctuations in Xiamen and the seasonal division of similar climate zones [33,34], the period from May 1, 2021 to September 30, 2021 was categorized as summer, October 1, 2021 to November 30, 2021 as autumn, and December 1, 2021 to February 28, 2022 as winter. We collected 2184 valid questionnaires in total; 864 were collected in summer, 648 in autumn, and 672 in winter. The thermal environment parameters recorded during the experiment in the three seasons are shown in Table 5 .

Table 5.

Outdoor thermal parameters.

Season	Meteorological factors	Min	Max	Mean	Standard deviation
Summer	Ta（°C）	22.3	35.6	30.77	2.86
	G（W/m2）	54	935	284.22	311.75
	Ws（m/s）	0	3.5	0.59	0.85
	RH（%）	37.4	92.5	68.65	12.54
	PET（°C）	20.2	51.2	33.9	7.0
Autumn	Ta（°C）	18.5	31.7	25.56	3.24
	G（W/m2）	35	858	314.32	302.45
	Ws（m/s）	0	3.1	0.51	0.78
	RH（%）	25.3	79.5	54.78	12.1
	PET（°C）	14.5	48.3	28.51	8.96
Winter	Ta（°C）	10.1	26.4	18.04	3.99
	G（W/m2）	28	782	323.07	298.1
	Ws（m/s）	0	2.2	0.38	0.56
	RH（%）	20.6	81.4	45.6	10.16
	PET（°C）	7.2	39.9	21.4	8.1

3.2. Thermal sensation and neutral temperature

The TSV scatter plots for the subjects while sitting and walking are shown in Fig. 4, Fig. 5 , respectively. Linear regressions were performed separately. PET at TSV = 0 was defined as the neutral temperature, and PET within the interval (−1,1) for TSV was the neutral range. The results are shown in Table 6 . In the sitting position, the neutral temperatures were 25.7 °C (summer), 24.3 °C (autumn), and 22.2 °C (winter) for subjects wearing masks and 24.5 °C (summer), 23.2 °C (autumn), and 21.6 °C (winter) for subjects without masks. In the walking condition, the neutral temperatures were 23.8 °C (summer), 22.1 °C (autumn), and 19.5 °C (winter) for subjects with mask and 22.1 °C (summer), 20.8 °C (autumn), and 18.7 °C (winter) without masks. In the sitting position, differences in neutral temperature between subjects with masks and those without masks were 1.2 °C in summer, 1.1 °C in autumn, and 0.6 °C in winter. In the walking condition, the differences in neutral temperature between subjects with masks and those without masks were 1.7 °C in summer, 1.3 °C in autumn, and 0.8 °C in winter.

Fig. 4.

Correlation between TSV and PET in different seasons while sitting: (a) Summer - Without Masks; (b) Summer - With Masks; (c) Autumn - Without Masks; (d) Autumn - With Masks; (e) Winter - Without Masks; (f) Winter - With Masks.

Fig. 5.

Correlation between TSV and PET in different seasons while walking: (a) Summer - Without Masks; (b) Summer - With Masks; (c) Autumn - Without Masks; (d) Autumn - With Masks; (e) Winter - Without Masks; (f) Winter - With Masks.

Table 6.

Neutral temperature and neutral temperature range while sitting.

Condition	Season	Without Masks		With Masks
		Neutral temperature(°C)	Comfort temperature range (°C)	Neutral temperature(°C)	Comfort temperature range (°C)
Sitting	Summer	25.7	＜31.2	24.5	＜29.6
	Autumn	24.3	18.4–30.2	23.2	17.7–28.8
	Winter	22.2	17.6–26.8	21.6	16.9–26.5
Walking	Summer	23.8	＜29.5	22.1	＜27.4
	Autumn	22.1	15.9–28.2	20.8	15.1–26.4
	Winter	19.5	14.9–24.2	18.7	13.9–23.5

3.3. Effect of masks on thermal comfort

Fig. 6 shows change in mean thermal sensation vote (MTCV) with TSV in summer, autumn, and winter with and without masks. In summer, difference in MTCV between the groups with and without masks was the most obvious. When TSV >0, the MTCV of the mask-wearing group was higher than that of the mask-free group under the same thermal sensation, and the difference gradually increased as thermal environment deteriorated. In the walking condition, difference in MTCV between the groups under the same thermal sensation increased, and significant differences were found at TSV of 1, 2, or 3（p＜0.05）. In autumn and winter, when TSV <1, difference in MTCV between the mask-wearing and mask-free groups was generally small. When TSV >1, the MTCV of the mask-wearing group was greater.

Fig. 6.

Correlation between MTSV and PMV in different seasons: (a) Summer-Sitting; (b) Summer-Walking; (c) Autumn-Sitting; (d) Autumn-Walking; (e) Winter-Sitting; (f) Winter-Walking. T-Test significative difference with: *, p < 0.05; **, p < 0.01.

3.4. Discomfort in different parts of the body

A higher proportion of subjects felt discomfort in the chest and abdomen than in other parts of the body in the mask-free group (Fig. 7 ). The face was the most unpleasant portion in the mask-wearing group, and the percentage of the subjects who experienced discomfort in the face was higher than that in the group without masks (p＜0.001). Approximately 75.9%, 68.5%, and 53% of mask wearers experienced facial discomfort in the summer, autumn, and winter, respectively, when they were seated; and 92.8%, 79.6%, and 66.7%, respectively, when they were walking. The percentage of facial discomfort decreased as the season changed from warm to cool, but the facial discomfort was exacerbated by walking. In addition, the group that wore masks had a slightly higher proportion of chest discomfort than the group that did not, but no significant difference was found in winter (p＞0.05). Differences in the degrees of discomfort in the abdomen, back, and upper and lower extremities were found between the groups.

Fig. 7.

Percentages of participants who voted that they experienced discomfort in various body parts: (a) Summer-Sitting; (b) Summer-Walking; (c) Autumn-Sitting; (d) Autumn-Walking; (e) Winter-Sitting; (f) Winter-Walking. Chi Square Test significative difference with: *, p < 0.05; **, p < 0.01; ***, p < 0.001.

3.5. Meteorological factors

Random forest was used in analyzing the relative importance of thermal environment parameters in TSV. Random forest refers to a classifier that uses multiple trees to train and predict samples [35]. The classifier was first proposed by Leo Breiman and Adele Cutler. A decision tree is a primary classifier that generally classifies features into two classes. A forest with a series of decision trees is created. The constructed decision tree has a tree-like structure and can be considered a collection of if then rules, and the main advantages are readability of the model and fast classification [36].

Two random forests were created for each season to predict the relative importance of T_a, RH, Ws, and G in TSV without and with masks. The results are shown in Table 7 . The TSV for each season was mainly predicted by G and T_a. In winter, G was slightly more important than T_a, and the opposite was true in summer and autumn. The influence of RH on thermal sensation was relatively low. The effect of Ws on thermal sensation was lower than the effects of G and T_a but higher than the effect of RH. The importance of Ws in thermal sensation increased gradually from summer to winter. The temperature had a greater effect on the thermal sensation of mask-wearing subjects than those on subjects without masks, whereas Ws had a smaller effect on thermal sensation. This difference was pronounced in summer.

Table 7.

Random forest modeling results for summer, autumn, and winter.

Season	Condition	Ta (°C)	RH (%)	Ws(m/s)	G (W/m2)
Summer	Without Masks	0.501	0.045	0.106	0.348
	With Masks	0.613	0.031	0.051	0.305
Autumn	Without Masks	0.467	0.051	0.085	0.397
	With Masks	0.54	0.033	0.047	0.38
Winter	Without Masks	0.307	0.059	0.152	0.482
	With Masks	0.354	0.073	0.123	0.45

Fig. 8 depicts the seasonal thermal preferences of subjects wearing masks and those without masks for environmental parameters. Overall, subjects wearing masks preferred higher Ws and lower G, T_a, and RH. In summer, more than half of the subjects hoped to reduce G, T_a, and Ws, especially those who were wearing masks. The highest proportion of subjects who expected no change in G and T_a was observed in autumn. In winter, difference in perception of meteorological parameters between the groups was the smallest, suggesting that masks have less effect on thermal comfort in winter than in summer or autumn. G, T_a, and Ws were the most important meteorological parameters affecting thermal sensation. In summer and autumn, the proportion of subjects who wore masks and preferred G, T_a, and Ws to remain constant was lower than that of subjects who did not wear masks, and the opposite result was obtained in winter. This finding indicated that subjects wearing masks were more likely to be satisfied with outdoor thermal environment factors in winter. Among the four meteorological factors, the proportion of preferred RH to remain unchanged was always the highest. The main reason was that the subjects had strong adaptability to high humidity in the south, and the satisfaction of subjects wearing masks or not wearing masks with humidity was the highest in winter.

Fig. 8.

Percentage distribution of thermal preference for environmental parameters with or without masks：（a）Summer; （b）Autumn; （c）Winter.

4. Discussion

4.1. Neutral temperature

Table 8 shows the results of several outdoor thermal comfort studies on neutral PET. Most studies were conducted before the COVID-19 epidemic; thus, the subjects can be assumed to wear masks rarely at that time. Previous studies have shown that neutral PET is generally higher in summer than in winter [5,37,40,[47], [48], [49]]. The neutral PET in different seasons in Taichung [7], Hong Kong [5], and Guangzhou [10,40], which belong to the hot-summer-warm-winter climate region as Xiamen, were close to the neutral PET under the mask-free condition. The winter neutral PET in Harbin [49], Mianyang [41], Tianjin, and Chengdu [46] were lower than that in this study, because of regional climate differences.

Table 8.

Neutral PET in different outdoor thermal comfort studies.

Authors	Place	Season	Condition	Neutral PET	Linear regression equation
This study	Xiamen	Summer	Sitting-Without Masks	25.7	TSV = 0.1807·PET-4.6452
			Sitting-With Masks	24.5	TSV = 0.1945·PET-4.7628
			Walking-Without Masks	23.8	TSV = 0.1765·PET-4.3692
			Walking-With Masks	22.2	TSV = 0.1867·PET-4.2997
		Autumn	Sitting-Without Masks	24.3	TSV = 0.1697·PET-4.1321
			Sitting-With Masks	23.2	TSV = 0.1783·PET-4.1425
			Walking-Without Masks	23.2	TSV = 0.1733·PET-4.029
			Walking-With Masks	21.8	TSV = 0.1785·PET-3.8982
		Winter	Sitting-Without Masks	22.2	TSV = 0.2194·PET-4.8746
			Sitting-With Masks	21.6	TSV = 0.2092·PET-4.5389
			Walking-Without Masks	20.5	TSV = 0.2129·PET-4.3727
			Walking-With Masks	19.7	TSV = 0.2064·PET-4.067
Lin [7]	Taichung	Cool Season	–	23.7	TSV = 0.1991·PET-4.7229
		Hot Season	–	25.6	TSV = 0.1184·PET-3.0253
Cheng [5]	Hong Kong	Summer	–	25.0	TSV = 0.1372·PET-3.4335
		Winter	–	21.0	TSV = 0.119·PET-2.4638
Lai [37]	Tianjin	Summer	–	15.6	TSV = 0.101·PET-1.571
		Winter	–	9.2	TSV = 0.188·PET-1.73
Zeng [38]	Chengdu	Summer	–	24.8	TSV = 0.105·PET-2.6
Chen [39]	Shanghai	Autumn	–	24.5	TSV = 0.037·PET-0.98
		Winter	–	22.4	TSV = 0.071·PET-1.59
Li [40]	Guangzhou	Spring	–	25.6	TSV = 0.25·PET - 6.40
		Summer	–	–	TSV = 0.07·PET - 0.36
		Winter	–	15.6	TSV = 0.18·PET - 2.81
Cheng [41]	Mianyang	Summer	–	30.0	TSV = 0.1114·PET - 3.0032
		Autumn	–	23.2	TSV = 0.1048·PET - 2.4274
		Winter	–	17.3	TSV = 0.0735·PET - 1.2726
Fang [12]	Guangzhou	Summer and autumn	1∼2Met	21.9	TSV = 4.11ln(PET)-12.65
			2.6Met	19.4	TSV = 4.79 ln(PET)- 14.247
Leng [42]	Harbin	Winter	Static activities	21	TSV = 0.127·PET-2.667
			Dynamic activities	18.2	TSV = 0.112·PET-2.037
Niu [43]	Xi'an	Summer	light intensity activities	26.1	TSV = 0.0822·PET-2.142
			moderate intensity	22.1	TSV = 0.0661·PET-1.4632
			vigorous intensity	11.9	TSV = 0.0571·PET-0.679
Nasrin [44]	Kuala Lumpur	April to June	–	25.6	TSV = 0.248·PET-6.341
Tang [45]	Guangzhou	Summer	Before physical training	19.3	TSV = 0.15·PET-2.9
			After physical training	4.6	TSV = 0.09·PET-0.41
Wei [46]	Chengdu	Summer	–	15.1	TSV = 0.1046·PET-1.5816
		Winter	–	12.8	TSV = 0.0828·PET-1.062

The effect of different activity metabolic rates on outdoor neutral temperature was significant. Studies in Harbin [42], Xi'an [43], and Guangzhou [12,45] showed that outdoor neutral temperature decreases with increasing activity intensity (Table 7). Our research indicated that the neutral temperature in the sitting condition was higher than that in the walking condition. The differences between the two in the absence of masks were 1.9 °C, 2.2 °C, and 2.7 °C in summer, autumn, and winter, respectively. The differences between the two were 2.3 °C, 2.4 °C, and 2.9 °C in summer, autumn, and winter, respectively, in the mask-wearing group. Wearing masks increased the neutral temperature difference between sitting and walking. During exercise, the body's metabolic rate and oxygen demand increase, a higher amount of heat needs to be removed through breathing, and the frequency of breathing through the mouth increases. Oral breathing expels more heat than nasal breathing [50], promotes heat accumulation in masks and results in a heightened perioral skin thermal sensation. In addition, some studies found that breathing through the nose cools the venous blood leading to the skull and reduces brain heat sensation, whereas breathing through the mouth diminishes this cooling mechanism [51]. Owing to breathing resistance caused by the mask, people who wear masks during exercise breathe more frequently through the mouth than those who do not wear a mask. The neutral temperature gap between mask-wearing subjects who walked and those who sit increased.

Zhang et al. [23] discovered that masks lowered the neutral temperature of indoor workers by 1.5 °C. According to the study conducted by Tang et al. [21] in a library in Guangzhou, indoor borrowers with masks had a 0.5 °C higher neutral operative temperature and 0.3 °C higher neutral standard effective temperature than those without masks. The indoor studies did not identify seasonal differences in the effects of masks on thermal comfort. The results of our study in an outdoor environment showed that neutral PET differences between subjects wearing masks and those without masks were 1.2 °C, 1.1 °C, and 0.6 °C while sitting in summer, autumn, and winter, respectively. The differences were 1.7 °C, 1.3 °C, and 0.8 °C in summer, autumn, and winter, respectively, during walking.

Under normal conditions, facial skin temperature is influenced by the temperature and humidity of the surrounding air [18]. Wearing a mask generates a microenvironment in the vicinity of the perioral region, creating a mask dead space that functions as the mask wearer's breathing environment. This microenvironment has a substantial effect on heat exchange in the face. In hot outdoor environments during summer, the temperature gradient between the surrounding environment and the environment inside the mask decreases, and the heat in the dead space of the mask is less likely to dissipate [22], which can worsen the microenvironment inside the mask and increase the thermal sensation of a mask wearer. The reason for this seasonal difference may be related to the local thermal sensitivity of the human skin. Perception of thermal changes in the environment varies by human surface part [52]. The face has the highest sensitivity to thermal changes in a warm environment [53]. In a cool environment, the thermal sensitivity of the face is relatively low, and the thermal sensitivity of the chest and abdomen increases [54,55]. In a warm environment in summer, a mask increases thermal stress on the overall thermal sensation of the human body. In a cool environment in winter, the effect of wearing a mask on the overall thermal sensation decreases due to the decreased thermal sensitivity of the face.

4.2. Thermal comfort votes

As the outdoor thermal environment deviates from the neutral temperature, people's discomfort and TCV increase. A survey conducted by Huang et al. [56] in Mianyang, China showed that the percentage of people who felt comfortable outdoors was lower in summer (36.5%) than in winter (44.1%). In humid and hot subtropical regions, such as Taichung [7], Guangzhou [40,57], and Belo Horizonte [48], surveys showed that summer was the least comfortable season, whereas winter was mild; outdoor respondents were more likely to feel comfortable in the latter season. In winter, the degree of comfort was significantly higher in warm environments than it was in summer. In the present study, when TSV = 1, 2, 3, MTCV was lower in winter than in summer, similar to the results of studies in the aforementioned regions.

Previous studies have shown that wearing a mask increases discomfort. Zhang et al. [23] discovered that as the indoor air temperature increased, the thermal comfort of mask wearers decreased at an accelerated rate. Tang et al. [21] found that 73.9% of the indoor population wearing masks felt uncomfortable during summer, and facial discomfort accounted for 71% of cases of discomfort. These results were similar to the present study. In a warm environment, the MTCV of mask wearers was always higher than that of subjects without masks for the same thermal sensation, and this difference was more pronounced under the walking condition. The face was the most uncomfortable area for mask wearers, followed by the chest. As temperatures changed from warm to cool from summer to winter, the proportion of individuals experiencing facial discomfort while sitting or walking decreased. In a cool environment, this difference was nonsignificant. This result is due to the fact that the contribution of facial thermal comfort to overall thermal comfort is greater in warm environments and less in cool environments [54]. Therefore, difference in MTCV between subjects with and without masks under cold stress was small. The chest and overall comfort can be affected by a mask's restrictive effect on breathing. A portion of the exhaled air remains between the mask and a wearer's face [61]. Several studies have demonstrated that the level of carbon dioxide in a mask's dead space is significantly higher than the normal value [61]. The high concentration of CO₂ and low concentration of O₂ inhaled by mask wearers stimulate the sympathetic nervous system and increase the body's cardiopulmonary burden [58], increasing the degrees of increased chest and general discomfort. In an exercise state, the body's O₂ demand and breathing rate increase, increasing discomfort.

4.3. Meteorological factors

A number of investigations [37,49,59,60] have demonstrated that T_a is the most essential meteorological parameter for outdoor thermal comfort and has the highest association with thermal sensation. G and Ws are regarded as significant factors influencing outdoor thermal sensation [[61], [62], [63], [64], [65]]. Our study showed that T_a and G were the most important thermal environment parameters affecting thermal perception, in subjects with or without masks. This result is similar to the results of previous studies. The effect of T_a on thermal sensation was greater for mask-wearing subjects than subjects without masks. This result should be related to the fact that the temperature gradient inside and outside a mask affects heat dissipation in the mask [22].In winter, the effect of Ws on thermal sensation was greater than that in summer. The effect of Ws on thermal perception in the masked group was lower than that in the mask-free group. However, the mask-wearing subjects preferred increase in Ws and decrease in RH. This finding can be explained by the fact that wearing a mask for an extended period of time causes an increase in T_a and RH inside the mask, which hinders people's respiratory heat dissipation and facial skin heat dissipation [18,66]. Consequently, individuals with masks prefer to increase the level of heat dissipation by increasing Ws and decreasing RH.

4.4. Limitations

In this study, we reported the effect of masks on thermal perception during the first 30 min of outdoor exposure, but did not examine the situation after 30 min. Future research should investigate the effect of masks on outdoor thermal perception over longer time periods. The subjects selected for this study were all college students in good physical condition, and investigations involving other age groups or people with underlying diseases were lacking. This is a longitudinal experimental study and no cross-sectional investigation was conducted, but different types of outdoor spaces may also lead to differences in the effects of masks on thermal comfort.In addition, only surgical masks were used. The thermal comfort effects of other types of masks may be different from those of surgical masks, and thus such potential differences should be explored in future studies.

5. Conclusions

This study aimed to assess the impact of surgical masks on outdoor thermal comfort in Xiamen, China. The following conclusions were drawn.

• (1)Surgical masks reduced the neutral temperature of individuals outdoors. In the sitting position, surgical masks decreased the neutral PET by 1.2 °C, 1.1 °C, and 0.6 °C in summer, autumn, and winter, respectively. In the state of walking, wearing masks decreased the neutral PET by 1.7 °C, 1.3 °C, and 0.8 °C in summer, autumn, and winter, respectively. Masks had the greatest impact on neutral temperature in summer and the least in winter and had a greater impact on walking neutral temperature than on sitting neutral temperature.
• (2)Wearing masks in a warm environment decreased the subjective thermal comfort of individuals. When TSV >0, the MTCV of mask-wearing subjects was higher than that of subjects without masks for the same thermal sensation. In the walking condition, this difference was more pronounced. In cool environments, the effect of masks on thermal comfort was nonsignificant.
• (3)Surgical masks significantly increased facial discomfort. The discomfort under the walking condition was more intense. As the season changed from summer to winter, discomfort caused by masks tended to diminish.
• (4)Radiation and air temperature were the primary environmental parameters that influenced the outdoor thermal sensation in hot and humid regions, regardless of the use of masks. The influence of air temperature on thermal sensation was higher in subjects with masks than in those without masks and was more pronounced in summer. Subjects with masks preferred lower air temperature, radiation, and humidity and higher wind speeds.

CRediT authorship contribution statement

Zhiqiang Zhou: Writing – review & editing, Writing – original draft, Visualization, Methodology, Investigation, Formal analysis, Conceptualization. Liang Dong: Supervision, Resources, Project administration, Funding acquisition.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The authors do not have permission to share data.