Constantly Tracking and Investigating People’s Physical, Psychological, and Thermal Responses in Relation to Park Strolling in a Severe Cold Region of China—A Case Study of Stalin Waterfront Park

Abstract

It is important for engineering applications that we evaluate the thermal environment based on long-term tracking and investment. Methods merging environmental, physiological, and psychological domains to implement a human-centered approach were applied in this study to assess the outdoor thermal environment in a park. The constant influence of humans in the outdoor environment can change people’s physiological, psychological, and thermal responses. Additionally, the relationship between human physiological, psychological, and thermal factors was explored in this study. The results of this study provide the following findings: (1) In summer, subjects’ skin temperature increased by 0.35 to 2.83 °C during a one-hour outdoor test without shelter, while when tree shade was provided, subjects’ skin temperature dropped by 0.50 to 1.87 °C (except for motion segments). (2) In winter, if subjects stayed outdoors for 1 h, their body segments’ skin temperature dropped by a maximum of 7.93 °C. (3) When subjects went outside, in the early stage, their thermal responses fluctuated for a long time. Therefore, TSV, TCV, and TAV should be measured after they stay outdoors for 45 to 55 min in future studies. (4) Different body segments show different sensitivities to hot or cold. Considering this, a new group of formulas for mean skin temperature calculation are proposed with high accuracy (winter: 0.95; summer: 0.89). (5) Data for the one-hour change in different assessment indicators provide a good viewpoint for park design considering multiple aims such as comfort (TCV), pleasure (EVI), and increasing energy (PFI). Overall, this study took Stalin Waterfront Park as a case study, and some suggestions involving landscaping nodes, space types, and facilities are offered. Moreover, this study provides a novel theory and reasonable method that can be referred to in urban planning and landscape design.

1. Introduction

It has been forecast that 60% of the world’s population will live in cities by 2025 [1]. The steadily increasing urban population has intensified demand for outdoor public spaces (Lai et al., 2020) [2]. In order to improve their physical condition and stay healthy, more and more people participate in a range of outdoor exercises, such as square dancing, doing Tai Chi, cycling, running, and so on [3]. As for children, outdoor entertainments are helpful for both their physical and psychological development during their growth [4]. For example, they can attain more opportunities to meet and socialize with friends and acquaintances in parks [5,6]. Furthermore, with increasing pressure relating to work and life, it is common for individuals to travel outside to appreciate beautiful scenery, get in touch with nature, and relax. However, urban expansion leads to dramatic declines in public green space and results in an increase in urban heat islands [7,8,9,10,11]. Due to the UHI effect, heat stress nowadays is at an all-time high and becoming hazardous for human beings. To adapt to this increasing climate temperature, the urban population remains indoors most of the time, and uses air-conditioned spaces to maintain thermal comfort. Unfortunately, staying indoors leads to a sedentary lifestyle, which harms individuals’ health and increases energy consumption [12].

In order to create comfortable urban open spaces, promote the construction of livable cities, and improve the quality of life and happiness of citizens, scholars have launched a series of related studies [13,14,15,16,17], including research on public spaces [18,19], city squares [20,21,22], streets [23], campuses [24,25,26,27,28,29], central business districts [30], and residential communities [31,32,33]. Jianlin Liu et al. found that the architectural feature ‘lift-up’ (or elevated) design could be an effective design to improve weak pedestrian-level wind (PLW) conditions in a subtropical high-density city and thermal comfort at the pedestrian level [34]. José Abel Rodríguez Algeciras et al. found that the spatial distribution of thermal conditions at the street level depends strongly on aspect ratio and street orientation: high aspect ratios are favorable in summer, and low aspect ratios are suitable for winter [23]. Jelena Djekic et al. concluded that different ground materials significantly influence the thermal comfort of pedestrians, and suggested that special attention should be paid to the usage of light colors and smooth materials [35,36].

In addition to the above research, it cannot be ignored that urban parks are one of the most critical components of outdoor open spaces, and are usually regarded as indispensable places for citizens. Diverse natural resources in parks, such as water, trees, and grass, can provide attractive views for park visitors to decrease their levels of stress, evoke positive emotions, and promote restorative benefits [37,38,39,40,41]. Meanwhile, park greenery and water can help mitigate the urban heat island (UHI) effect, improve thermal comfort [42,43,44], attenuate noise levels [45,46], and improve air quality. However, urban planners and designers need be more well-acquainted with human comfort, while the visual aesthetic quality of the landscape, including the beauty and exuberance of vegetation, has always been one of the primary focuses in the majority of park designs [47]. As a result, the design and planning of parks have seldom taken thermal comfort into consideration, which may potentially undermine the utilization efficiency of these spaces [20,48,49]. Knes and Thorsson found that when the level of thermal comfort is within an acceptable range, people tend to stay outdoors longer (average 19–21 min) [16]. In contrast, when they feel uncomfortable, they stay outdoors for shorter periods (average 11 min). Thus, other than visual aesthetic quality, thermal comfort is also considered to be one of the crucial design considerations for the use and acceptance of outdoor spaces [47]. Nowadays, more and more scholars have begun to pay attention to the design of urban parks in order to create a pleasant microclimate environment for cities [47,50,51]. Some scholars have assumed that the species, configuration, location, number, and density of trees are factors that decide the degree to which trees influence the microclimate [52,53,54]. In addition, shade infrastructure such as pavilions, ramada, pergolas, and canvas has been shown to be useful for improving thermal comfort and the urban microclimate, as it can reduce incident solar radiation [47,55]. Water bodies have also been shown to improve thermal comfort by retaining warm air on cold days and cooling the air during hot weather [47]. Meanwhile, since thermal comfort is a complex cognitive response involving various factors, many researchers indicate that people’s thermal comfort is indeed affected by some non-thermal parameters. Haiying Wang proved that emotional states do have an impact on TSV in light activities, and individuals in a joyful emotional state demonstrate an improved TSV [56].

Furthermore, previous studies of thermal environment evaluation are usually based on transient thermal comfort surveys. However, citizens’ outdoor activities such as resting, meeting with friends, recreation, physical exercise, etc., often last for longer periods of time. This means that the outdoor environment has a constant impact on people rather than a short-term one. For example, in some cases, the outdoor thermal conditions are acceptable for a short time, while long-term exposure can result in a range of health problems, including dehydration, heat stroke and even death (hot conditions) [57,58,59,60,61] and trigger disorders such as nasal discharge, sneezing, shivering, aching, tingling, and numbness (cold conditions) [62]. Therefore, this transient acceptability may not imply that the public space is thermally acceptable for outdoor activities in reality. Pui Kwan Cheung’s research suggested that most respondents (94.6%) accepted the thermal environment at the moment of the interview, whereas only 69.0% of them found it acceptable to stay there for another hour [63].

In summary, parks have many benefits for cities. The thermal environment of a park is affected by built structures, vegetation, and water bodies in the park, as well as non-thermal factors, such as mood. Previous studies have shown that the quality of the thermal environment and the comfort level of individuals can be modified by small design details [23,64,65,66,67]. This study investigates the utilization of a park in a severely cold region of China to explore the interaction between humans and the thermal environment of the park. During the experiment, the changes in park climate parameters and the physiological skin temperature of pedestrians were traced constantly, and at the same time subjects’ trends regarding their psychological states and thermal responses were recorded continually. These results could serve as a reference for the design and planning of bioclimate parks in China.

2. Methodology

2.1. Study Area and Assessment Sites

Harbin (45°41′ N, 126°37′ E), the capital of Heilongjiang province, used to be an international city with 16 consulates, and it earned the nickname “Paris in the Orient” [68]. Nowadays, it is one of the most popular tourism destinations in China, with a large number of pieces of European-style architectural and cultural heritage and various kinds of famous activities such as the “Harbin International Ice and Snow Festival” [69,70]. This festival is known as one of the world’s four major snow festivals, together with Japan’s Sapporo Snow Festival, Canada’s Quebec Winter Carnival, and Norway’s Oslo Ski Festival.

Harbin belongs to the Chinese severe cold zone [71]. The monthly mean air temperature during 2009–2018 was lowest in January (−23 °C) and highest in July (28 °C); the monthly relative humidity fluctuated between 49.0% and 77.0%; and the monthly mean wind speed fluctuated in the range of 2.5–3.8 m/s, with the dominant wind direction being southwesterly both annually and in the winter [72].

This study was conducted in a significant scenic spot called Stalin Waterfront Park (shown in Figure 1). It is very attractive to tourists because of its long history and special importance. The park is in the shape of a belt, spanning about 1750 m. The chief components of this park are three leisure paths, because they are the primary recreation venues in the park. The main pedestrian path is called the “Boulevard Path”, along which many deciduous trees are planted. The second path, located near the river, is called the “Hydrophilic Path”. There are few trees to provide any shade for people in summer, but tourists can come close to the water on this path. The third path, named the “Sightseeing Path”, has deciduous trees planted on just one side so that the bare side can offer a wide field of vision for appreciation of the view of the river.

In this study, we selected the “Boulevard Path” and the “Hydrophilic Path” as the research targets, because there are many more people on these paths and the differences of their thermal environment features are significant when compared with each other.

2.2. Measurement and Instruments

The data integration method was utilized to merge environmental, physiological, and psychological domains to implement a human-centered approach to monitoring and assessing the microclimate and outdoor comfort in the park.

First, thermal environment parameters such as air temperature, relative humidity, wind speed, and black globe temperature were measured. The weather station was positioned 1.5 m above the ground according to the ISO 7726 standard [73]. In addition, BES air temperature and relative humidity sensors were installed in a high-reflectivity aluminum box that resisted solar irradiation and ensured smooth self-ventilation on both sides. Wind speed data were recorded every 10 s. Air temperature, relative humidity, and globe temperature data were acquired at 1 min intervals. Indoor air temperature and relative humidity in the chamber were also monitored at 1 min intervals. The specifications and models of the sensors are listed in

Table 1. Metrological properties of the measuring instruments used.

Type of Probe	Variable Measured	Range	Accuracy
BES-02B air temperature sensor	Air temperature	−30~50 °C	±0.01 °C
BES-02B relative humidity sensor	Relative humidity	0~100%	±0.1% RH
Kestrel5500 anemometer	Wind speed	0.4~40 m/s	±0.1 m/s
BES-05 black globe thermometer	Globe temperature	−50~125 °C	±0.5 °C
DS1922L iButton	Skin temperature	−40~85 °C	±0.5 °C

.

Second, human physiological parameters such as skin temperature were recorded using wireless temperature data-loggers, namely a device named “iButton”. This device is so tiny (17.33 mm × 5.89 mm) and portable that it can be fixed on human skin easily and will not limit subjects’ normal activities. Eight local temperature sites such as the forehead, back, chest, forearm, upper arm, hand, thigh and calf were measured according to the EN ISO 9886:2004 Standard recommendation [74]. Earlier studies have shown that certain body segments are significantly sensitive to specific environments. Hui Zhang found that the foot is more sensitive to cold environments [75]. Meanwhile, Yingdong He proved that foot thermal sensation exerted a strong influence on overall thermal sensation (OTS) in response to a hot environment, too [76]. M. Nakamura concluded that the neck is an area susceptible during heat exposure [77]. We also found that the ear is extremely sensitive and causes people to feel pain during the cold season experiments. This is may be due to the ear protruding from the head, and both sides being exposed to the cold air. Therefore, in our experiments, we added the foot as a new supplementary skin measuring point. Meanwhile, the neck and ear were measured in the hot season and cold season, respectively.

This is not unusual, and some previous studies have chosen the foot, neck, and ear as necessary target parts previously [78,79,80,81,82]. All of the measurement points of human skin temperature were recorded at 1 min intervals, and these segments are shown in Figure 2.

In order to obtain thermal responses and psychological conditions, paper questionnaires were given to 15 subjects (Figure 3). Normally, thermal sensation rates are on a 7-point scale based on the ASHRAE standard. However, previous studies extended this to a 9-point thermal scale to cater to actual climate conditions, in areas such as Guangzhou (13.3~28.4 °C), Changsha (8.6~30.2 °C), and Israel (8~30 °C) [83,84,85]. Harbin is an exceptionally cold climate city and has an extremely cold winter, with the recorded lowest temperature being −37.7 °C [86]. Therefore, considering this particular outdoor thermal environment, the ASHRAE 7-point scale needs to be extended to a greater-precision scale (11-point scale) [62,86,87]. The 11-point scale of thermal sensation (LTS/OTS) that we used was as follows: −5: unbearably cold, −4: very cold, −3: cold, −2: cool, −1: slightly cool, 0: neutral, 1: slightly warm, 2: warm, 3: hot, 4: very hot, and 5: unbearably hot. The thermal acceptability vote was evaluated by using a 4-point scale (−2: completely unacceptable, −1: unacceptable, 1: acceptable, 2: completely acceptable), and thermal comfort was assessed by using a 7-point scale (−3: very uncomfortable, −2: uncomfortable, −1: slightly uncomfortable, 0: neutral, 1: slightly comfortable, 2: comfortable, 3: very comfortable), as in ordinary research.

In addition, psychological parameters such as people’s physical fitness index (PFI), which reflects whether the subject is tired or energetic, and emotion valence index (EVI), which reflects whether the subject feels bored or pleased, were necessary to evaluate, because tourists who their spend time in recreation usually take part in enjoyable activities: taking photos, appreciating attractive scenery, performing exercises, skating, and so on. Their physical and emotional states may have changed during these activities and these changes could have impacted people’s thermal responses and their perception of the thermal environment. The physical fitness index (PFI) ranged from −3 to +3 (−3: very tired, −2: tired, −1: slightly tired, 0: neutral, 1: slightly energetic, 2: energetic, 3: very energetic). The emotional valence index (EVI) was based on a 7-point scale too, which was defined as −3: very bored, −2: bored, −1: slightly bored, 0: neutral, 1: slightly pleased, 2: pleased, 3: very pleased.

2.3. Experimental Procedure

According to the past relevant studies of Ghahramani (2016) [88], Dai et al. (2017) [80], Takada (2013) [89], Sim et al. (2016) [90], Choi and Yeom (2017) [79], Wang and Liu (2020) [56], and Lau et al. (2019) [90], which were based on 12, 11, 15, 8, 16, 18, and 14 subjects, respectively, this experiment enrolled a total of 15 healthy subjects (9 male, 6 female). The anthropometric information of the subjects is shown in

Table 2. Information of subjects.

Sex	Subjects	Age	Height	Weight	Skin Area
female	6	24.5 ± 1.5	1.645 ± 0.015	55 ± 4	1.590 ± 0.050
male	9	25.0 ± 1.0	1.790 ± 0.040	75 ± 10	1.905 ± 0.125

. All of them are healthy and have lived in Harbin for more than 2 years.

The peak number of tourists who travel to Stalin Waterfront Park is concentrated in summer (especially from June to August, lasting about 3 months) and winter (particularly from December to January the next year, lasting about 2 months). The average maximum air temperature during these periods is 27.7 °C and −12.9 °C in summer and winter, respectively. Thus, we conducted the hot season test from 22 June to 23 June 2019, when the air temperature was 27 °C, close to 27.7 °C, representing the typical summer climate. Similarly, in the cold season, we chose 9 January and 10 January 2020 as the test dates, when the air temperature was −14 °C and −16 °C, respectively, close to the −12.9 °C average. Additionally, two consecutive days were selected so that we could make sure the test was finished within 48 h every time and improve the accuracy of the results.

The schedule of the test on each day was executed based on an earlier study. Sun revealed that the number of visitors in Stalin Park increased beginning from 9 o’clock to 11 o’clock, then remained steady until 13 o’clock, when it grew a second time [91]. Moreover, above 50% of these tourists stayed in the park for 1~3 h [91]. The test time in this study was designed to consist of the actual use of the park. Subjects needed to arrive at the air-conditioned room at 9:30 for 30 min of preparation. This first stage aimed to ensure that the subjects’ physical and psychological conditions achieved steadiness [92]. Then, they went into the park and took a walk along the path at 10:00, and this stage lasted 1 h. After stage 2, at about 11:00, subjects went back to the air-conditioned room for a rest; this was the last stage in the morning and lasted about 30 min. The experiments in the afternoon started at 12:30 and ended at 14:30. The whole test lasted a total of about 2 h each time, and during this period, all of the subjects were asked to fill out the questionnaire every fifth minute. Figure 4 shows the specific procedure and details of the experiments.

It is worth noting that all of the subjects were informed that they should wear unified clothes in order to avoid the effect of clothing insulation differences. Subjects were asked to wear trousers and short-sleeved shirts for summer tests. In the winter tests, they wore long sleeves, sweaters, trousers, flannel trousers, thick long socks, boots, and gloves. The above clothing collocation reflected the typical outfits of people in Harbin, and the clothing level was 0.57 clo in summer and 1.85 clo in winter.

3. Results

3.1. Thermal Environment

The hydrophilic path data show that the average air temperature at this spot is 27.23 °C, the relative humidity is 42%, and the wind speed is 1.29 m/s in summer. The boulevard path results show that the average air temperature at this spot is 27.70 °C, the relative humidity is 40%, and the wind speed is 0.7 m/s in summer. The air-conditioned room’s (chamber) average indoor air temperature is about 24 °C, and the relative humidity is almost 45% in summer. The air temperature and relative humidity of the hydrophilic path are close to the values for the boulevard path. The wind speed around the hydrophilic path is 0.6 m/s higher compared with that around the boulevard path, and this may result from the trees surrounding the boulevard path.

In winter, after tree leaves fall, the thermal environment of the boulevard path was similar to that of the hydrophilic path. Thus, we evaluated the outdoor path thermal environment in winter by only carrying out tests on the boulevard path. The outdoor average air temperature is −14.3 °C, the relative humidity is 70%, and the wind speed is 0.81 m/s. In the heated room (chamber), the air temperature is 18.07 °C and the relative humidity is only 24%. The gap between the indoor and outdoor climate parameters is much more prominent in winter than in summer. The average air temperature outdoors is 32.32 °C lower than indoors, and the relative humidity is 45.65% higher than indoors in winter. However, in summer, the average air temperature outdoors is only 3.8 °C higher than indoors at most, and the outdoor relative humidity is 5.2% lower than indoors. Above data about thermal environment of the experiment spots are shown in Figure 5.

3.2. Local Skin Temperature (T_local)

3.2.1. Hydrophilic Path (Summer)

As Figure 6a shows, subjects’ skin temperature was stable during stage 1, and their local skin temperature values were in the range of 32.42 to 34.47 °C. The forehead skin temperature was the highest, while the hand skin temperature was the lowest. During stage 2, during the first 10 min, most parts of the human body skin temperature values decreased. The temperature of the forehead, hand, and chest reduced by more than 1 °C. After the short-term decrease, all measured body segments’ skin temperature began to rise. Calf temperature rose significantly and ultimately reached 36.07 °C, which was 2.83 °C higher compared with stage 1. In addition, for body segments such as the thigh, foot, and neck, the skin temperature reached its peak at 55 min in stage 2 and increased by more than 1 °C compared with stage 1: 2.51 °C, 1.90 °C, and 1.09 °C, respectively. During stage 3, the skin temperature of most body segments began to drop down after the subjects returned to the air-conditioned chamber. The temperature of the calf and thigh changed remarkably and decreased by 1.86 °C and 1.20 °C, respectively, compared with stage 2. The neck, foot, back, and upper arm temperature changed slightly and reduced by 0.35 to 0.77 °C.

Overall, subjects’ skin temperature was higher when they stayed on the hydrophilic path outdoors without trees as shelter than in the air-conditioned room. Meanwhile, subjects’ skin temperature dropped as soon as they returned to the air-conditioned chamber. Additionally, it is worth noting that when subjects left the room, their skin temperature decreased initially; this may be because of the wind’s impact. These results suggest that outdoor wind could be helpful in maintaining a low body temperature. However, the influence of air temperature and solar radiation would be more significant and lead to higher skin temperature if humans stay in a hot environment outdoors for a long time. However, when the subject returned to the air-conditioned chamber without solar radiation, their skin temperature soon dropped. Therefore, shelter is needed and helpful for humans in hot summers.

3.2.2. Boulevard Path (Summer)

Figure 6b reveals that all parts of the subjects’ body skin temperature were kept stable during stage 1. The hand skin temperature was the lowest, at only 31.62 °C. The skin temperature of other parts was distributed from 32.67 to 34.55 °C. During stage 2, almost all of the skin temperature values decreased. Neck skin temperature reduced dramatically and decreased by about 1.87 °C at T = 55 min of stage 2. The skin temperature of the forehead, hand, lower arm, and chest dropped to the lowest point within 30 min, and these decline values were more than 1 °C, at 1.82 °C, 1.77 °C, 1.63 °C, and 1.14 °C, respectively. The back and upper arm skin temperature decreased slowly. They dropped to the lowest point at T = 50 min of stage 2, and the decline values were only 0.5 °C. However, the calf, foot, and thigh showed an upward trend during stage 2, increasing in temperature by 1.34 °C, 0.93 °C, and 0.60 °C, respectively. This may be because the lower limbs (motion segments) generated heat when the subjects were required to keep walking during stage 2. During stage 3, the skin temperature of the lower limbs began to decline, while that of other parts increased when subjects returned to the air-conditioned chamber.

Overall, subjects’ skin temperature dropped when they stayed on the boulevard path surrounded by trees outdoors, except for the motion parts (lower limbs). Therefore, humans in outdoor spaces with trees as a shelter may be healthier than those staying in a room, even though the outdoor air temperature is higher than indoors. This suggests that an outdoor space with wind and shade in hot summers could help to maintain human skin temperature at a lower level, meaning that citizens do not have to stay at home for cooling purposes and can spend more time taking part in healthier outdoor activities.

3.2.3. Boulevard Path (Winter)

All parts of the body skin temperature were kept stable during stage 1. The skin temperatures of the ear and foot were 24.99 °C and 28.67 °C, respectively, and both values were lower than 30 °C; the skin temperature of other body parts was distributed from approximately 30.17 to 35.63 °C. During stage 2, as soon as subjects went outdoors, their whole-body skin temperature dropped immediately. The forehead temperature change was the most significant. It dropped to the lowest point at T = 45 min of stage 2 and was 9.05 °C less than in stage 1. The temperature of other parts continued to reduce until the end of stage 2, and the temperature drop value was distributed from 2.03 to 7.93 °C. During stage 3, skin temperature began to increase after subjects returned to the heated chamber. The skin temperature of the chest, back, forearm, upper arm, calf, and foot increased quickly in the first 15 min and then rose slowly. The other parts continued to increase in temperature during stage 3 (Figure 6c).

In winter, there is a huge gap between the indoor and outdoor air temperature, and human skin temperature inevitably declined when the subjects left the heated chamber. This suggested that solar radiation in winter is not as helpful, and the impact of air temperature and wind is more significant than solar radiation, differing from summer. However, subjects’ skin temperature improved as soon as they entered the warm chamber. This means that chambers with commercial functions, such as hot-drink bars and souvenir shops, could be distributed around the park for human heating purposes.

3.3. Thermal Responses

3.3.1. Thermal Sensation (TS)

The hydrophilic path summer test results are shown in Figure 7a. Subjects’ thermal sensation votes (TSV) fluctuated around 0.5 during stage 1. According to the ASHARE55 standard, subjects’ thermal sensation was at the neutral level during this period. During stage 2, in the first 10 min, TSV decreased slightly, and this phenomenon is consistent with the results of skin temperature change (3.2.1). Then, TSV began to rise and fluctuate around one. After that, TSV rose again and peaked at a value of two at T = 50 min of stage 2. Subjects felt warm at this time. During stage 3, when subjects returned to the air-conditioned chamber, their TSV dropped and stayed below 0 for 15 min afterwards.

The boulevard path summer test results are shown in Figure 7b. Subjects’ thermal sensation votes fluctuated between −1.5 and −1 during stage 1. Subjects’ thermal sensation was at a “slightly cool–cool” level during this period. During stage 2, in the first 35 min, TSV fluctuated around a value of −0.3. Then, the thermal sensation vote increased to 0 and then remained stable. After that, TSV began to reduce at T = 50 min of stage 2. During stage 3, after subjects returned to the air-conditioned chamber 15 min later, their TSV dropped and stayed at −1. Subjects felt “slightly cool” at this moment.

The boulevard path winter test results are shown in Figure 7c. Subjects’ TSV fluctuated around a value of 0.5 during stage 1. Subjects’ thermal sensation was “slightly warm” during this period. During stage 2, in the first 30 min, TSV fluctuated between −2 and −1. Subjects felt “slightly cool” during this period. After that, subjects felt “cool”. At T = 45 min of stage 2, the TSV dropped to the lowest point, and subjects felt “cold” at this moment. During stage 3, once subjects returned to the heated chamber, their TSV increased quickly, but the value was still below 0, even 30 min later.

The above research results show that the thermal environment of the boulevard path is beneficial in offering the appropriate thermal experience for people undertaking leisurely activities. Subjects’ TSV on the boulevard path only rose slightly in this experiment, and the TSV value remained below 0 throughout the entire 60 min. This means that people feel “cool” in this place in summer. However, people on the hydrophilic path feel a little warm, and with the exposure time becoming longer, they feel hotter. People on the boulevard path in winter feel cold. Moreover, as time increases, they feel colder and colder. Stage 3 data showed that improving boulevard path subjects’ TS would be easy if an appropriate thermal environment could be offered, and then human TSV would improve immediately in summer. However, hydrophilic path subjects’ TSV required 20 min to decrease to a value below 0 in summer in order to feel neutral or cool. Additionally, the boulevard path subjects’ TSV increasing above 0 in order to feel neutral or warm in winter was much more complicated and required more than 30 min.

Overall, the outdoor thermal environment has a constant impact on human thermal sensation (TS); therefore, constant investments are necessary to evaluate the outdoor thermal environment accurately via the TSV indicator rather than transient investments. Hydrophilic and boulevard path tests show it would be more useful and practical to record subjects’ TSV after they have stayed outside for at least 50 min in summer. Boulevard path tests in winter show that TSV can be measured after the subjects have stayed outdoors for at least 45 min.

3.3.2. Thermal Comfort (TC)

The hydrophilic path summer results are shown in Figure 7d. During stage 1, subjects’ thermal comfort vote (TCV) fluctuated around the value of one. Subjects felt “slightly comfortable” during this period. During stage 2, in the first 15 min, TCV remained above 0. After 20 min, TCV fell below 0. Then, TCV reduced continually and dropped to the lowest point, and the value of TCV was close to −2 at T = 50 min during stage 2. People felt “uncomfortable” at this moment. During stage 3, when subjects returned to the air-conditioned chamber, their TCV increased to zero 10 min later and they felt “slightly comfortable” (TCV > 1) 20 min later.

The boulevard path summer results are shown in Figure 7e. During stage 1, subjects’ TCV fluctuated around one, and they felt “slightly comfortable” during this period. During stage 2, in the first 20 min, TCV remained stable around one, until T = 25 min of stage 2, when the TCV fell below one. At T = 50 min of stage 2, TCV reduced to the lowest point and the value was 0 at this moment. Overall, the measured value of TCV during the boulevard path test always remained above 0 and subjects were at a comfortable level throughout the entire outdoor experiment time in this location.

The boulevard path winter results are shown in Figure 7f. During stage 1, subjects’ TCV fluctuated around 0.5 and above 0. During stage 2, after subjects went outdoors, TCV fell below 0 immediately. At T = 50 min of stage 2, the TCV dropped to the lowest point of −1.17, and subjects felt “uncomfortable”. During stage 3, TCV rose immediately when the subjects returned to the heated chamber. Twenty minutes later, TCV rose above 0, and people returned to the “neutral” level of TSV.

Overall, the thermal environment of the boulevard path could offer a comfortable thermal experience for people in summer. Although the TCV measured at the boulevard path reduced slowly, the value always remained above 0 throughout the entire outdoor test. This implied that moving people on this path could stay neutral or slightly comfortable. However, people felt neutral or slightly comfortable only for 15 min when moving on the hydrophilic path. After that, they felt “uncomfortable” due to exposure to solar radiation. People who moved on the outdoor path in winter felt uncomfortable, and as time increased, they felt increasingly uncomfortable. After stage 2, subjects’ TCV after walking on the hydrophilic path in the summer required 10 min to increase above 0. Additionally, subjects’ TCV after walking on the boulevard path in winter required 20 min to increase above 0.

Overall, the outdoor thermal environment has a constant impact on human thermal comfort (TC). Therefore, constant investments are necessary to evaluate the outdoor thermal environment accurately via TCV indicators rather than transient investments. It would be more valuable and practical to record subjects’ TCV after they have stayed outside for at least 50 min in summer as well as winter.

3.3.3. Thermal Acceptance (TA)

The hydrophilic path summer results are shown in Figure 7g. During stage 1, the thermal acceptance vote (TAV) fluctuated around one, and subjects’ thermal perception of the thermal environment was “acceptable”. During stage 2, in the first 30 min, the thermal acceptance vote had no markable change and remained around one. Then, the TAV dropped below one and fell to the lowest point at T = 55 min of stage 2; the value of TAV at this point was below 0. Subjects’ thermal perception of the thermal environment was at a “neutral–not acceptable” level. During stage 3, once people returned to the air-conditioned chamber, the TAV value increased above 0 immediately. Then, TAV returned to the same level as stage 1 after 15 min.

The boulevard path summer results are shown in Figure 7h. During stage 1, the thermal acceptance vote fluctuated around one. Subjects’ thermal perception of the environment was “acceptable”. During stage 2, the thermal acceptance vote had no change. During stage 3, TAV increased after people returned to the air-conditioned chamber for 20 min.

The boulevard path winter results are shown in Figure 7i. During stage 1, the thermal acceptance vote fluctuated around one. Subjects’ thermal perception of the thermal environment was “acceptable”. During stage 2, in the first 30 min, TAV fluctuated between 0 and 0.5. Subjects’ thermal perception of the environment was at the “neutral–acceptable” level. At T = 35 min of stage 2, the TAV dropped below 0, then the TAV fell to the lowest point (TAV = −0.78) at T = 45 min of stage 2.The thermal environment was “not acceptable” at that moment. During stage 3, TAV increased to the same level as stage 1 immediately, as soon as people returned to the heated chamber.

Overall, the thermal environment of the boulevard path could offer a constantly acceptable thermal experience for people in summer. People felt that the boulevard path’s thermal environment throughout the experiment was acceptable (60 min). However, regarding the thermal environment of the hydrophilic path in summer, subjects’ TAV dropped below 0 after 50 min. Additionally, the thermal environment of the boulevard path in winter was “acceptable” only in the first 30 min. After that, TAV dropped below 0 and people perceived the thermal environment as more and more unacceptable as time increased.

According to the results, the TAV in the hydrophilic path summer test and boulevard path winter test during stage 2 took 55 min and 45 min to fall to the lowest point, respectively. Therefore, TAV should be recorded when subjects stay outdoors for at least 50 min in summer and 45 min in winter.

3.4. Psychological Responses

3.4.1. Emotional Valence (EV)

The hydrophilic path summer results are shown in Figure 8a. During stage 1, subjects’ emotional valence vote increased slowly and then remained stable at a value of one. They felt “slightly pleased” at that moment. During stage 2, subjects’ EVI decreased in the first 5 min. Then, EVI began to increase, lasting for 10 min. At T = 20 min of stage 2, EVI dropped again and fell below 0 at T = 30 min. At T = 35 min, EVI fell to the lowest point (−1), and subjects felt “slightly bored”. During stage 3, the emotional valence vote increased to 0 after ten minutes since they returned to the air-conditioned chamber. Then, EVI continued to increase and peaked at two at the end of this stage. Subjects felt “pleased” at that moment.

The boulevard path summer results are shown in Figure 8b. During stage 1, the subjects’ emotional valence vote fluctuated between 1 and 3. Their emotional state was distributed from “slightly pleased” to “very pleased”. During stage 2, when subjects went out, their emotional valence vote decreased but still remained above 0 for a long time. Then, EVI fell to the lowest point with the value of −0.67 at T = 45 min of stage 2. During stage 3, the EVI of the subjects grew to 0 immediately as soon as they returned to the air-conditioned chamber. Additionally, the EVI reached the highest value of three at the end of this stage. They felt “very pleased” at that moment.

The boulevard path winter results are shown in Figure 8c. Subjects’ emotional valence vote stayed above 0 during stage 1. During stage 2, in the first 20 min, the value of EVI was still above 0. Then, EVI dropped below 0 and fell to the lowest point (−0.89) at T = 55 min during stage 2. They felt “slightly bored” at that moment. During stage 3, EVI increased, but with fluctuations. People felt more and more pleased until 15 min later.

Overall, the thermal environment of the boulevard path in summer was beneficial in helping people stay pleased consistently. The hydrophilic path could improve people’s emotional valence only at the beginning, due to the hydrophilic space’s entertainment function. However, after that, subjects’ EVI began to reduce and then fell below 0. This happened 15 min earlier compared with the situation for the boulevard path. People on the boulevard path in winter could only stay pleased for 20 min, and the time is much shorter. The stage 3 results indicated that boulevard path subjects’ EVI could increase above 0 immediately in summer, while the hydrophilic path subjects required 10 min for this to happen. In winter, the subjects needed to spend 25 min outdoors in order to increase the EVI above 0.

3.4.2. Physical Fitness (PF)

The hydrophilic path summer results are shown in Figure 8d. During stage 1, subjects’ physical fitness vote was above one, and this implied that they were “slightly energetic”. During stage 2, the subjects’ physical fitness vote began to decrease, and then the PFI dropped below 0 at T = 25 min. At T = 40 min of stage 2, PFI fell to the lowest point and the value was −1.33. Subjects felt “slightly tired” at that moment. During stage 3, the subjects’ physical fitness vote increased to 0 after 15 min once they returned to the air-conditioned chamber.

The boulevard path summer results are shown in Figure 8e. PFI during the first stage was above one. During stage 2, in the first 40 min, subjects were on a “neutral–energetic” level. When T = 55 min, PFI decreased to the lowest point (−1.67), and subjects felt tired at that moment. However, during stage 3, PFI increased above 0 quickly as soon as they returned to the air-conditioned chamber.

The boulevard path winter results are shown in Figure 8f. During stage 1, subjects’ physical fitness was on a “slightly tired–neutral” level. Additionally, PFI gradually increased at the end of stage 1. During stage 2, in the first 10 min, subjects’ physical fitness remained at the “neutral” level. After that, PFI began to decline. At T = 50 min, PFI dropped to the lowest point and the value was −0.94. During stage 3, PFI increased back to 0 by the end of the stage.

Overall, it is easier to feel tired outdoors in summer, but the thermal environment of the boulevard path could alleviate this tired feeling. People feel less tired in winter compared with summer when they move outdoors. However, it took 15 min longer to recover after the outdoor boulevard path experiment in winter than after the hydrophilic path experiment in summer.

4. Discussion

4.1. Psychological Responses

4.1.1. Thermal Sensation and Local Skin Temperature

Table 3. Linear regression analysis between thermal sensation and local skin temperatures in summer.

Regions	Independent Variable (X)	Equation	R-Square	Sig.
Forehead (FH)	Skin temperature of forehead: 32.73 ≤ X ≤ 34.48	TSV = 0.16X − 4.09	0.007	NS
Hand(H)	Skin temperature of hand: 29.85 ≤ X ≤ 32.91	TSV = 0.61X − 18.58	0.58	0.000 ***
Thigh(T)	Skin temperature of thigh: 32.69 ≤ X ≤ 35.36	TSV = 0.92X − 30.59	0.71	0.000 ***
Upper arm (UA)	Skin temperature of upper arm: 32.22 ≤ X ≤ 34.11	TSV = 1.06X − 34.87	0.74	0.000 ***
Calf (CA)	Skin temperature of calf: 32.62 ≤ X ≤ 36.07	TSV = 0.74X − 24.89	0.71	0.000 ***
Back (B)	Skin temperature of back: 32.87 ≤ X ≤ 34.55	TSV = 1.32X − 44.20	0.62	0.000 ***
Forearm (FA)	Skin temperature of forearm: 31.63 ≤ X ≤ 33.23	TSV = 0.92X − 29.34	0.27	0.010 **
Neck (N)	Skin temperature of neck: 32.38 ≤ X ≤ 35.21	TSV = 0.74X − 24.52	0.66	0.000 ***
Chest (C)	Skin temperature of chest: 32.11 ≤ X ≤ 33.29	TSV = 1.23X − 42.01	0.34	0.003 **
Feet (F)	Skin temperature of foot: 33.87 ≤ X ≤ 35.91	TSV = 1.11X − 38.07	0.60	0.000 ***

*** p < 0.001, ** p < 0.01, NS—not significant (p > 0.05).

shows the correlation between the skin temperature of various body segments and thermal sensation (TS) in summer. The forehead skin temperature was not linearly correlated with thermal sensation. Other body parts’ skin temperatures had a close relationship with TS. This relationship is described by a regression equation. The temperatures of the upper arm, thigh, and calf were correlated with thermal sensation, with high R-square values, exceeding 70% of the variance. The temperatures of the neck, back, foot, and hand was correlated linearly with thermal sensation, with moderate R-square values (0.5~0.7). Chest and forearm skin temperatures correlated with thermal sensation, with weak R-square values (below 0.5), and the skin temperature of these body parts was positively correlated with thermal sensation. Thus, if the skin temperature increased, then the thermal sensation was warmer. The back, chest, foot, and upper arm had a high slope coefficient of their regression equation and the value exceeded one. So, the back, chest, foot, and upper arm are more sensitive to thermal sensation.

Table 4. Linear regression analysis between thermal sensation and local skin temperature in winter.

Regions	Independent Variable (X)	Equation	R-Square	Sig.
Forehead (FH)	Skin temperature of forehead: 23.42 ≤ X ≤ 32.30	TSV = 0.15X − 5.94	0.69	0.010 **
Hand (H)	Skin temperature of hand: 22.24 ≤ X ≤ 30.15	TSV = 0.16X − 6.34	0.94	0.000 ***
Thigh (T)	Skin temperature of thigh: 25.70 ≤ X ≤ 31.72	TSV = 0.22X − 8.21	0.84	0.000 ***
Upper arm (UA)	Skin temperature of upper arm: 31.34 ≤ X ≤ 34.27	TSV = 0.44X − 16.59	0.89	0.000 ***
Calf (CA)	Skin temperature of calf: 26.64 ≤ X ≤ 30.27	TSV = 0.32X − 10.85	0.57	0.004 **
Back (B)	Skin temperature of back: 31.74 ≤ X ≤ 35.40	TSV = 0.35X − 13.58	0.84	0.000 ***
Forearm (FA)	Skin temperature of forearm: 31.54 ≤ X ≤ 34.34	TSV = 0.46X − 17.21	0.89	0.000 ***
Ear (E)	Skin temperature of ear: 19.46 ≤ X ≤ 24.92	TSV = 0.25X − 7.58	0.89	0.000 ***
Chest (C)	Skin temperature of chest: 33.61 ≤ X ≤ 35.66	TSV = 0.63X − 23.62	0.83	0.000 ***
Feet (F)	Skin temperature of foot: 23.41 ≤ X ≤ 28.59	TSV = 0.27X − 8.93	0.86	0.000 ***

*** p < 0.001, ** p < 0.01.

shows the correlation between the skin temperature of various body segments and thermal sensation (TS) in winter. The significance level for the forehead and calf was 0.05 (p < 0.05), and the significance level for the other parts was 0.001 (p < 0.001). The skin temperature of the hand correlated linearly with thermal sensation, with a very high R-square value, accounting for 94% of the variance. The temperature of the upper arm, forearm, ear, foot, thigh, and back correlated linearly with thermal sensation with an R-square value exceeding 0.80. Additionally, the shin temperature of all of these parts was positively correlated with thermal sensation. Thus, if the skin temperature increased, the thermal sensation was warmer too. The chest had a high slope coefficient of its regression equation and the value exceeded 0.5. So, the chest is the most sensitive part to thermal sensation in winter.

Overall, the winter skin temperature correlated linearly with thermal sensation, with a higher R-square compared with summer. This may be due to the air temperature difference between indoors and outdoors in different seasons. The gap in temperature between indoors and outdoors was extremely large in winter, as the value was 32.4 °C, while the gap value was only 3.8 °C in summer. The most important factor that affects thermal sensation in winter is temperature. Additionally, thermal sensation in summer was affected by other factors. In addition, the slope coefficient of the regression equation in summer is higher than in winter. Therefore, body skin temperature was more sensitive to thermal sensation in a hot environment than in a cold environment in this study. This was due to the thermal experience of the locals living in a cold environment for a long time in the severe cold region of China.

4.1.2. Mean Skin Temperature and Overall Thermal Sensation

In many previous studies, the relationship between physiological and thermal responses has been quantitatively analyzed. These studies found that the thermal response of the human body under non-uniform conditions depends on the temperature of various local body segments. So, skin temperature of several body parts was considered and integrated as the mean skin temperature (MST) by summing these local skin temperatures using corresponding area weighting factors. Normally, the traditional seven-point method by Hardy and Dubio (1938) is applied to calculate the MST [62,93]. Meanwhile, recently, in order to improve the accuracy of the result, a nine-point method was proposed, involving more body measuring parts [94]. However, with the research procedure becoming more and more complex, more skin measurement points mean more devices need to be attached to the skin. This could impede subjects’ normal activities and have a bad impact on the experiment result. So, a three-point method is recommended, especially for subjects required to perform some specific activities [95,96].

Though these equations are used widely, they may not be very precise because of the different climates all over the world. Peel concluded that distinct equations are necessary for each climatic zone and each season for the prediction of thermal response [96]. This study was carried out in a severely cold area of China; the climate in winter is extremely cold. Previous studies in this location show that the R-square of the regression equations for TSV and MST (calculated by Hardy and Dubio’s seven-point method) is small (0.54~0.65) [62]. Therefore, a new method was used for MST to determine and explore the relationship between MST and TSV.

Table 5. Correlation matrix for body part temperatures with thermal sensation (TSV) in summer.

Correlation	TSV	FH	H	T	UA	CA	B	FA	N	C	F
TSV	1
FH	0.085	1
H	0.760 **	0.576 **	1
T	0.844 **	−0.061	0.607 **	1
UA	0.859 **	0.349	0.896 **	0.825 **	1
CA	0.840 **	−0.121	0.639 **	0.959 **	0.844 **	1
B	0.790 **	0.322	0.832 **	0.718 **	0.945 **	0.735 **	1
FA	0.515 **	0.665 **	0.644 **	0.496 *	0.606 **	0.334	0.495 *	1
N	0.812 **	0.516 **	0.885 **	0.720 **	0.929 **	0.660 **	0.928 **	0.760 **	1
C	0.580 **	0.740 **	0.807 **	0.539 **	0.746 **	0.439 *	0.627 **	0.946 **	0.833 **	1
F	0.771 **	−0.207	0.486 *	0.955 **	0.681 **	0.934 **	0.525 **	0.383	0.531 **	0.411 *	1

** p < 0.01, * p < 0.05,. Note: TSV = thermal sensation vote, FH = forehead, H = hand, T = thigh, UA = upper arm, CA = calf, B = back, FA = forearm, N = neck, C = chest, F = foot.

shows the correlation between the thermal sensation and the skin temperature of different body parts in summer. The correlation matrix indicates that the temperatures of all of selected body parts are significantly correlated with thermal sensation, except for the forehead. Additionally, the correlation between different body parts’ skin temperatures is shown in Table 5. The upper arm had the strongest relationship with thermal sensation (correlation = 0.859). Meanwhile, the upper arm also had a remarkable relationship with the hand and back, and the correlation value was around 0.90. The thigh had the second strongest relationship with thermal sensation (correlation = 0.844). Additionally, the thigh correlated with the calf and foot, and the correlation value was above 0.95. In addition, the neck is an important body part that cannot be ignored. The correlation matrix indicates that the neck was significantly correlated with the other body parts. Thus, it is reasonable to utilize neck data to represent other body parts that are not selected.

As explained in

Table 6. Determination of weighting factors for selected positions in summer.

	Sensitivity (S1 = R2)	Skin Surface Area (S2)	a = S1/∑S1	b = S2/∑S2	(a × b)/∑(a × b)	Weighting Factor
Thigh	0.844	0.163	0.336	0.629	0.627	0.62
Upper arm	0.859	0.081	0.342	0.313	0.318	0.32
Neck	0.812	0.015	0.323	0.058	0.056	0.06
Sum ∑	2.515	0.259	——	——	——	——

, these selected parts’ weighting factors were derived by taking into account not only the relative surface areas represented by each measurement, but also the sensitivity of thermal sensation to skin temperature at each point [93]. The R-square shows the correlation between skin temperature and thermal sensation (shown in Table 5). S2 is the proportion of skin surface area that was determined according to the previous study [97]. Additionally, the weighting coefficients were determined by rounding down (a × b)/∑(a × b) to one decimal place. The proposed formula for computing MST in summer is thus:

MST _(summer) = 0.62T_thigh + 0.32T_upperarm + 0.06T_neck

(1)

Table 7. Correlation matrix for body temperature parts with thermal sensation (TSV) in winter.

Correlation	TSV	FH	H	T	UA	L	B	FA	E	C	F
TSV	1
Forehead	0.833 **	1
Hand	0.968 **	0.819 **	1
Thigh	0.917 **	0.949 **	0.938 **	1
Upper arm	0.941 **	0.888 **	0.978 **	0.985 **	1
leg	0.755 **	0.987 **	0.749 **	0.925 **	0.848 **	1
back	0.915 **	0.947 **	0.937 **	0.998 **	0.984 **	0.923 **	1
forearm	0.941 **	0.901 **	0.973 **	0.989 **	0.999 **	0.863 **	0.989 **	1
ear	0.942 **	0.803 **	0.974 **	0.903 **	0.937 **	0.726 **	0.891 **	0.929 **	1
chest	0.910 **	0.935 **	0.938 **	0.998 **	0.988 **	0.913 **	0.996 **	0.992 **	0.899 **	1
feet	0.926 **	0.935 **	0.954 **	0.994 **	0.988 **	0.901 **	0.988 **	0.990 **	0.934 **	0.993 **	1

** p < 0.01. Note: TSV = thermal sensation vote, FH = forehead, H = hand, T = thigh, UA = upper arm, CA = calf, B = back, FA = forearm, N = neck, C = chest, F = foot.

shows the correlation between the thermal sensation and the skin temperature of different body parts in winter. The correlation matrix indicates that the temperatures of all body parts were significantly correlated with thermal sensation. Hand temperature had the strongest relationship with thermal sensation, and the correlation value was 0.968. In addition, the ear was the only body part that was exposed to cold air except for the forehead in winter. However, the ear had the second strongest relationship with thermal sensation, and the correlation value was 0.942, much higher than the forehead. In addition, the thigh was also an important body part that could not be ignored. The correlation matrix indicates the thigh was significantly correlated with the other nine body parts and the correlation value was 0.903 at least. Thus, it is reasonable to utilize thigh data to represent other body parts that are not selected.

The proposed formula for computing MST in winter is thus:

MST_(winter) = 0.75T_thigh + 0.24T_hand + 0.01T_ear

(2)

The weighting coefficients were determined in the same way as for summer, and the procedure is shown in

Table 8. Determination of weighting factors for selected positions in winter.

	Sensitivity (S1 = R2)	Skin Surface Area (S2)	a = S1/∑S1	b = S2/∑S2	(a × b)/∑(a × b)	Weighting Factor
Thigh	0.917	0.163	0.324	0.762	0.752	0.75
Hand	0.968	0.049	0.342	0.229	0.239	0.24
Ear	0.942	0.002	0.333	0.009	0.009	0.01
Sum ∑	2.827	0.214	——	——	——	——

.

In order to verify the accuracy of the MST calculated using this new model approach, we compared the method with the previous models. The equations of the models are shown in

Table 9. Comparison of MST calculated with the different models.

MST Model	Formula	Relative Value	TSV Model	Reference
9 points	MST = 0.07Tfh + 0.05Th + 0.19Tt + 0.07Tua + 0.13Tl + 0.175Tb + 0.07Tfa + 0.175Tc + 0.07Tf	0.92 (winter)	TSV(winter) = 0.31MST – 11.38 (R2 = 0.84)	[90]
		0.88 (summer)	TSV(summer) = 1.4MST – 46.47 (R2 = 0.78)
7 points	MST = 0.07Tfh + 0.05Th + 0.19Tt + 0.14Tua + 0.13Tl + 0.35Tc + 0.07Tf	0.92 (winter)	TSV(winter) = 0.33MST – 12.13 (R2 = 0.84)	[62]
		0.87 (summer)	TSV(summer) = 1.39MST – 45.72 (R2 = 0.76)
3 points	MST = 0.40Tfh + 0.40Tc + 0.20Tf	0.88 (winter)	TSV(winter) = 0.25MST – 9.5 (R2 = 0.77)	[98]
		0.57 (summer)	TSV(summer) = 1.42MST – 46.85 (R2 = 0.33)
3 points	MST(winter) = 0.74Tt + 0.24Th + 0.02Te MST(summer) = 0.62Tt + 0.32Tua + 0.06Tn	0.95 (winter)	TSV(winter) = 0.21MST – 7.8 (R2 = 0.90)	This study
		0.89 (summer)	TSV(summer) = 1.05MST – 34.64 (R2 = 0.80)

. All models showed more precise results in winter than in summer. The nine-point method and the seven-point method showed almost the same level of accuracy. The correlation value of the relationship between MST calculated using the nine-point method and thermal sensation was 0.92 in winter, the same that was calculated using the seven-point method. The MST–TS correlation value was 0.88 and 0.87 when MST was calculated using the nine-point method and the seven-point method, respectively, in winter. MST calculated using the previous three-point method showed a relatively weak relationship with TS. The correlation value was only 0.88 and 0.57 in winter and summer, respectively. However, MST calculated using the new three-point method in this study showed relatively high accuracy in practice. The correlation value of the relationship between MST calculated using the new three-point method was 0.95 and 0.89 in winter and summer, respectively. Additionally, the regression equations for TSV and MST were:

TSV_(summer) = 1.05MST − 34.64

(3)

TSV_(winter) = 0.21MST − 7.8

(4)

According to Formulas (3) and (4), the thermal sensation is hotter when the mean skin temperature increases both in summer and winter. The slope value of Formula (3) was higher than that of Formula (4). Thus, MST has a more powerful impact on thermal sensation in summer than in winter. However, people prefer low thermal sensation in hot summers, and they prefer high thermal sensation in winter. Therefore, it is essential to take measures to control the increase in MST or even decrease the MST in summer. On the contrary, MST should be improved in winter.

4.2. Thermal Sensation and Thermal Comfort

4.2.1. Thermal Acceptance and Thermal Comfort

The thermal acceptance vote showed no significant correlation with the thermal comfort vote in the summer boulevard path test. The thermal comfort vote of the boulevard path in summer was distributed between 0 and 1.33, as shown in Figure 9a. Additionally, the thermal acceptance vote was stable and the value of TAV was one. Because the subjects’ thermal comfort vote value was above 0, they felt relatively comfortable. So, subjects’ thermal perception of the comfort level would be an “acceptable” level. According to the above phenomenon, it is obvious that different degrees of thermal comfort were acceptable for subjects. Meanwhile, there may be some other potential factors that could affect thermal sensation. In addition, other tests on the hydrophilic path in summer and the boulevard path in winter showed a linear relationship between the thermal acceptance vote and thermal comfort vote. People would be more comfortable if the environment had a higher acceptance vote.

4.2.2. Thermal Sensation and Thermal Comfort

The thermal sensation vote showed no significant correlation with the thermal comfort vote in the summer boulevard path test. The value of the thermal sensation vote was distributed in a negative interval, and that of the thermal comfort vote was distributed in a positive interval. This means that people will feel comfortable (TCV > 0) in summer if their thermal sensation tends to be cool (TSV < 0). The summer hydrophilic path test showed a negative linear correlation between thermal comfort and thermal sensation. The value of thermal sensation measured in the test was positive and subjects felt hot (Figure 9b). That was why the value of thermal comfort vote concentrated in the negative interval. Thus, it is essential to take measures to decrease the thermal sensation of subjects on the hydrophilic path to improve thermal comfort in summer. In the winter boulevard path test, there was a positive linear correlation between thermal comfort and thermal sensation. Thermal comfort can be improved by increasing thermal sensation. TSV measured in the boulevard test in winter was below −1, and subjects felt cold during this test. So, the subjects felt uncomfortable and the TCV value was below 0. It would be helpful to take measures to increase thermal sensation to improve thermal comfort in winter.

4.3. Psychological and Thermal Comfort

4.3.1. Emotional Valence and Thermal Comfort

Figure 10a shows that thermal comfort has a positive linear correlation with emotional valence in different environmental tests. This means if the subjects are in a good mood, the perception of thermal comfort will be better. The results of the boulevard path test in summer showed that although subjects always felt comfortable throughout the entire process, their comfort level was lower when EVI was below 0, and their comfort level was higher when EVI was above 0. The value of EVI in both the hydrophilic path summer test and boulevard path winter test was almost entirely distributed in the negative interval, and TCV was distributed in the negative interval too. Thus, it is necessary to take some measures such as designing a beautiful scene and providing an interesting activity space to improve people’s mood.

4.3.2. Physical Fitness and Thermal Comfort

Figure 10b shows that thermal comfort has a positive linear correlation with physical fitness in different environmental tests. That means if the subjects are energetic, the perception of thermal comfort will be better. Additionally, it is more important to improve the thermal comfort by increasing physical fitness when the TCV is in a worse condition, as shown in Figure 10b. The PFI value of both the hydrophilic path in the summer test and the boulevard path in the winter test was concentrated in the negative interval, while TCV was concentrated in the negative interval too. So, to improve thermal comfort, rest spaces with relevant infrastructures should be considered in park design.

5. Conclusions

1. Outdoor climate parameters were stable during the 1-hour outdoor test. However, the differences between the indoor and outdoor climate parameters were large. The gap value of the air temperature between the indoors and outdoors was 32.4 °C and 3.8 °C in winter and summer, respectively.

2. The thermal environment of the boulevard path has a positive effect on skin temperature, which helps to limit skin temperature growth and even decrease skin temperature in summer (drop by 0.50 to 1.87 °C, except for the motion segments). The thermal environment of the park’s hydrophilic path was beneficial in decreasing human skin temperature in the short term. However, with an increase in exposure time, skin temperature rose because of solar radiation (increase by 0.35 to 2.83 °C). In winter, there was a huge gap between the indoor and outdoor air temperature, and it is inevitable that human skin temperature declines when subjects stay outdoors (7.93 °C maximally).

3. Although the thermal condition of the environment was stable during the 1-hour outdoor test, subjects had a different level of thermal responses with the accumulation of their exposure time. Therefore, it would be more practical and accurate to evaluate the outdoor thermal environment based on long-term tracking and surveys rather than previous transient investments. The TSV, TCV, and TAV results in this study show that thermal response assessments should be collected only if subjects stay outdoors at least for 50 (TSV), 50 (TCV), and 55 (TAV) minutes, respectively, in summer and 45 (TSV), 50 (TCV), and 45 (TAV) minutes in winter, respectively.

4. There is a positive correlation between local skin temperature and overall thermal sensation. Different segments of the human body have different sensitivities to thermal sensation in different thermal environments. During hot summers, the back, chest, foot, and upper arm are more sensitive to human thermal sensation, so when humans feel hot, it would be possible to adopt some approaches to decrease these segments’ skin temperature in order to obtain an overall cool feeling. During cold winters, the chest, forearm, and upper arm are more sensitive to thermal sensation. Therefore, some approaches could be adopted to keep these segments’ skin temperature high or decrease it as slowly as possible so that they feel warm and prolong their outdoor activity time.

5. Mean skin temperature calculated using the skin temperature of appropriate parts of the body can predict thermal sensation accurately. Previous formulas have usually selected seven or nine body segments, and this is not convenient when subjects are required to keep moving during the experiment, or the tests must last for a long time. Additionally, some three-point formulas were proposed previously, not considering the seasonal differences, so the results are not very accurate in summer. This study proposed a group of new three-point MST calculation formulas (one for the hot season and one for the cold season), which involve only three body segments and have more accurate performance considering the different body segments’ sensitivity to hot or cold environments.

6. It is possible to improve thermal comfort by adjusting human thermal sensation. In hot summers, reducing the thermal sensation level could improve thermal comfort. Additionally, in winter, increasing the thermal sensation level could provide higher thermal comfort. Furthermore, methods for adjusting human thermal sensation are shown in the third point in this section.

7. TAV has a positive linear correlation with TCV, but in summer, the relationship is not significant. Because TAV in summer is not as sensitive as TCV, the value of TAV remains stable or only changes slightly. However, in winter, TAV has a positive linear correlation and the relationship is significant. Furthermore, the TAV results showed that it is possible to ensure that subjects’ thermal responses are acceptable by limiting the outdoor exposure time to less than 30 min in winter, although they inevitably feel cold and uncomfortable, according to the TSV and TCV results.

8. There is a positive linear relationship between thermal comfort and emotional valence. So, outdoor leisure paths could be designed with beautiful landscape nodes and entertainment facilities. In order to make sure that the value of subjects’ emotional valence is positive, EVI results show that the entertainment spaces with facilities should be planned within 30 min intervals.

9. The physical fitness vote has a positive linear relationship with the thermal comfort vote. In summer, resting facilities could be placed in shady spaces for physical fitness and thermal comfort purposes. In winter, resting spaces could be designed with a heated room such as a hot-drink bars or souvenir shops with air-conditioners in the park. Additionally, in order to make sure that the value of subjects’ physical fitness is positive, PFI results show that the rest spaces with relevant facilities should be planned within 25 min intervals.

6. Recommendation

Walkability, comfort, and health are the most substantial indicators of the quality of a public space. It is essential to design an outdoor space based on people’s physical, psychological, and thermal state. Exposure time should be considered in the thermal environment assessment. Transient investments cannot ensure that the public space is thermally comfortable for long-term outdoor activities. This study revealed the tourists’ change in physical responses, psychological responses, and thermal responses during constant 1-hour exposure to a park leisure path. According to the results of the tests, there are some suggestions for the leisure path design of Stalin Waterfront Park.

The boulevard path has an excellent performance in improving the health and experience of tourists in summer. People’s local skin temperatures stay at the appropriate value or increase very slowly. Additionally, people feel cool and comfortable for at least 60 min of traveling. However, they may become bored and tired when the travel time exceeds 40 and 45 min, respectively. It is therefore important to design beautiful landscape nodes or attractive activity spaces to attract people’s interest and to improve their mood within a walking distance of 40 min. This can also be achieved by planning entertainment spaces with rest facilities to improve people’s vigor within a walking distance of 45 min.

Tourists travel along the hydrophilic path in summer without trees as shelter, and their physical skin temperature increases gradually. People will feel hot and uncomfortable when they walk on the hydrophilic path for more than 10 min and 20 min, respectively. Based on the change in thermal responses, it is helpful to provide a shady place within a walking distance of 10 min. The results of the psychological responses showed that people may feel bored and exhausted when they travel for more than 30 and 25 min, respectively. So, some multiple-function spaces should be designed within a walking distance of 25 min—for example, an open space with beautiful views and activity spaces with shaded resting facilities.

Tourists travel along the boulevard path during winter, and their physical skin temperature decreases gradually. Additionally, because of the extremely low air temperature in this very cold region, people will inevitably feel cold and uncomfortable as soon as they go outdoors. Additionally, the change in the psychological responses implied that people feel exhausted when their travel time exceeds 15 min. Additionally, their positive mood may only be maintained for 25 min. Thus, there should be some rest places for a short-term break within a walking distance of 15 min and some beautiful views or landscape nodes within a walking distance of 25 min. Additionally, more importantly, heated rooms should be set along the leisure path, providing combined indoor retail or exhibition functions, within a walking distance of 35 min.