Thermal Comfort and Restorative Benefits of Waterfront Green Spaces for College Students in Hot and Humid Regions

Abstract

Global climate change presents a serious threat to the sustainable development of human society, highlighting the urgent need to develop effective adaptation strategies to mitigate the impact of climate-related disasters. Campus waterfront green spaces, integral to the blue-green infrastructure, have been demonstrated to facilitate stress recovery. However, in hot and humid regions, severe outdoor thermal conditions may impair students’ mental and physical health and cognitive function, leading to symptoms such as increased stress, anxiety, and depression. This study examined the influence of outdoor thermal environments on health recovery by selecting three different waterfront green spaces in this climate: Space A (medium water body, sky view factor (SVF) = 0.228), Space B (large water body, SVF = 0.808), and Space C (small water body, SVF = 0.292). The volunteers’ thermal comfort and the restorative benefits of these spaces were evaluated via the perceived restorativeness scale (PRS), heart rate (HR), and electrodermal activity (EDA). We found variations in the neutral physiological equivalent temperature (PET) across the spaces, with values of 28.1 °C (A), 28.9 °C (B), and 29.1 °C (C). The lowest skin conductance recovery rate (R_SC) at 0.8811 was observed in Space B, suggesting suboptimal physiological recovery, despite higher scores in psychological recovery (fascination) at 15.23. The level of thermal comfort in this hot and humid region showed a negative correlation with the overall PRS score, the “being away” dimension, and heart rate recovery (R_HR). At a lightly warm stress level, where PET increased from 31.0 to 35.7 °C, R_SC peaked between 1.45 and 1.53 across all spaces. These insights provide guidance for urban designers and planners in creating waterfront green space designs that can improve the urban microclimate and promote thermal health, achieving sustainable health.

1. Introduction

As global climate change intensifies, extreme temperatures and heat waves are becoming more frequent, significantly affecting outdoor activities and causing substantial physical and psychological health impacts [1]. In hot and humid climates, the prolonged periods of high temperatures pose increased health challenges [2]. University students face considerable psychological issues due to academic, social, and other pressures [3] Surveys conducted by the World Health Organization revealed that at least one-third of the students from 19 universities across eight countries reported one or more mental disorders [4]. Campus landscapes play a crucial role in fostering student health [5] and blue-green spaces are particularly important for physiological and psychological health [6], alleviating stress and anxiety [7], aiding in attention restoration [8], and providing other restorative benefits. Improving the environmental quality of campus waterfront green spaces can effectively enhance student well-being [9].

However, the mechanisms through which blue-green spaces affect health are complex, owing to their diverse environmental characteristics and the involvement of multiple mediating factors [10]. Therefore, exploring the specific mechanisms by which blue-green spaces affect physiological and psychological health is crucial.

Both theoretical and empirical studies have shown that blue-green spaces positively impact cognitive abilities and physiological and psychological health. In terms of restoration, the most influential theories include the stress reduction theory (SRT) and attention restoration theory (ART). Ulrich et al. [11,12] proposed the SRT, demonstrating the stress-reducing effects of natural environments (including water environments) on an individual’s physiological and psychological stress through measurements such as electroencephalography (EEG), electrodermal activity (EDA) [13], and heart rate (HR). Kaplan et al. [14,15] introduced ART, emphasizing the positive role of natural environments in alleviating mental fatigue and restoring attention, as decreased attentional capacity can lead to mental fatigue. Subsequent researchers have conducted extensive empirical studies on the restorative benefits of blue-green spaces, including coastal environments [16], urban rivers, and urban canals [17], showing the significant restorative potential of blue spaces for individuals. This research focus has shifted from early urban residents to children, adolescents, the elderly, and working populations, exploring the positive effects of natural environments on the psychological health and quality of life of different populations [18]. Regarding assessment methods, studies typically combine subjective and objective indicators. Subjective assessment tools for mental health include the profile of mood states (POMS) [19,20], perceived restorativeness scale (PRS) [21], and positive and negative affect schedule (PANAS) [22]. Objective physiological assessments cover variables such as heart rate (HR), blood pressure, electroencephalography (EEG) [23], and electrodermal activity (EDA) [24]. Research variables mainly focus on aspects such as plant diversity, color, and biodiversity, as well as the auditory effects of natural sounds.

Existing research confirms that natural environments can directly or indirectly promote restorative benefits, yet unresolved issues persist. Although studies have highlighted the independent effects of blue-green spaces on mental health, further evidence is required to elucidate the specific exposure characteristics of these spaces and their association with restorative benefits. Researchers have predominantly investigated the restorative benefits of landscapes from visual and auditory perspectives [25]. However, in outdoor settings, thermal sensation is equally crucial, and its role in the recovery process is often neglected.

Amidst escalating global climate change and urbanization, the relationship between thermal comfort and human health remains a focal point. Historically, research in the field of thermal comfort has primarily focused on parameters such as air temperature, relative humidity, wind speed, and mean radiant temperature, incorporating human factors such as clothing insulation and metabolic rate into the evaluations [26] with rare measurements of physiological data. Common outdoor thermal comfort models include the predicted mean vote (PMV), physiological equivalent temperature (PET), and universal thermal climate index (UTCI) [27]. Recent shifts toward multidimensional assessments of the thermal environment aim to more comprehensively explore the interplay among environmental features, cognition, psychology, physiology, and thermal sensation [28,29].

Urban climate models have great potential in mitigating climate change. Over the past few decades, scientists have developed a variety of planning models, such as COSMO [30], ENVI–met [31], PALM [32], SOLWEIG [33], and MITRAS [34]. Among them, ENVI–met is widely used in outdoor microclimate research and has become a major research tool for urban green and blue infrastructures [35]. ENVI–met, a computational fluid dynamics (CFD)-based model, can simulate surface–plant–air interactions in urban environments with a particularly detailed plant model [36]. Because there may be differences between experimental results and simulation results, scholars often use ENVI–met to verify the reliability of actual measurements and simulations. In order to predict real thermal comfort conditions more accurately, Yuchun Zhang and colleagues [37] emphasized the importance of integrating physiological and psychological factors into heat index models to accurately assess outdoor thermal comfort conditions.

Current research on the restorative benefits of thermal environments has predominantly focused on buildings and semi-outdoor spaces, often assessed using images and virtual environments [29,38]. With the advent of wearable technologies, field experiments in various outdoor settings have become increasingly common. For instance, Elsadek et al. [39] examined street trees for the impact of different urban roadside trees and their microclimates on human stress reduction. Song et al. [24] compared the restorative benefits of grasslands and forests. Zhou et al. [40] investigated the effects of heatwaves, air pollution, and green and blue spaces on hypertension. Our study highlights a lack of studies on the specific impact of thermal environments within blue-green spaces on restorative benefits. Further, there is a lack of focus on environmental psychology and thermal comfort.

Notably, the disciplines of environmental psychology and thermal comfort have started to synergize and intersect, with extensive research underscoring the pivotal role of blue-green spaces in enhancing human cognitive capabilities, as well as physiological and psychological health. However, the influence of the thermal environment in these spaces on restorative outcomes remains underexplored, particularly concerning its complex interplay with human physiological and psychological health. Waterfront green spaces on campuses, which are critical blue-green environments, are vital for boosting student well-being. Thus, investigating how the thermal environment of these spaces affects the physiological and psychological health of students is crucial.

Focusing more on the thermal health of college students living in outdoor spaces in the humid and hot regions of China is imperative. This study selected three types of waterfront green spaces within campuses located in humid and hot regions. Volunteers were recruited for onsite experiences, and PRS, along with objective physiological signals, were used to assess the restorative benefits of these waterfront green spaces. Concurrently, by integrating subjective thermal comfort questionnaires and objective physical measurements, we probed the thermal environmental parameters of waterfront greenspaces. Our objectives were, as follows:

• To uncover the thermal comfort and restorative benefits for college students in different waterfront green spaces;
• To investigate the relationship between thermal comfort and recovery benefits for college students in waterfront green spaces located in humid and hot regions.

The anticipated outcomes of this research should provide empirical support for the optimization of blue-green space design and serve as a reference for the investigation of thermal health effects and the planning and design of health-centric campuses.

2. Materials and Methods

2.1. Research Site

Guangzhou (112°–114.2° E, 22.3°–24.1° N), the capital of Guangdong Province, is situated along the subtropical coastline of China. According to the Köppen climate classification, the city has a humid subtropical climate characterized by abundant sunshine, high humidity, and heavy rainfall [41]. The experimental site was located in a waterfront green space adjacent to the Guangzhou University Library on Xiaoguwei Island, Guangzhou. This location is conveniently near both the library and the educational precinct. Specific experimental areas (Figure 1) included Space A (medium water body, SVF = 0.228), Space B (large water body, SVF = 0.808), and Space C (small water body, SVF = 0.292).

2.2. Measurement Indicators and Instruments

The thermal environment parameters include air temperature, humidity, wind speed, and radiation. In this experiment, air temperature (T_a), globe temperature (T_g), relative humidity (RH), and wind speed (V_a) were measured using the thermal comfort meter (SSDZY-1) [42]. The instrument was placed at a height following ISO 7730 [43], located near the head of the subjects at approximately 1.1 m above the ground. Physiological signals were also measured. HR was measured using a fingertip pulse oximeter, which operates by utilizing optical technology to monitor the differences in the blood volume passing through the vessels. During measurement, the finger was fully inserted into the rubber channel, and data was read from the display screen after the readings stabilized. This method is widely used in both clinical and experimental research [44]. EDA was measured using a wearable skin conductance device developed by Beijing Jinfa Technology. EDA sensors were strategically placed on the inner sides below the second joints of the middle and index fingers, which are among the most sensitive parts of the body, to ensure data reliability [45]. Extensive research shows that EDA, as a physiological indicator, can effectively reflect an individual’s psychological and physiological stress states in different thermal environments.

Table 1. Comparison of performance parameters and accuracy of the different measuring instruments.

Instrument	Model	Parameter	Range	Accuracy
Thermal comfort level recorder	SSDZY-1	Air temperature	−20 °C ± 80 °C	±0.3 °C
		Relative humidity	0.01–99.9%	±2% (10–90%)
		Globe temperature	−20–80 °C	±0.3 °C
		Wind speed	0.05–5 m/s	5% ± 0.05 m/s
Fingertip pulse oximeter	YX102	HR	25–250 beats/min	1 beats/min
Wearable electrodermal activity device	Beijing Jinfa Technology ErgoLAB Biosensing	EDA	≥0 μs	0.01 μs

lists the precision and range of all of the measuring instruments used in this study. A black bulb thermometer (diameter (D) and surface emissivity (ε_g) of 0.15 m and 0.95, respectively) was used to ensure measurement accuracy. The mean radiant temperature, T_mrt, was calculated as follows:

$${\mathrm{T}}_{\mathrm{m}\mathrm{r}\mathrm{t}}={\left\{{\left({\mathrm{T}}_{\mathrm{g}}+273\right)}^4+\left[{\displaystyle \frac{\left(1.1\times {10}^8\times {\mathrm{V}}_{\mathrm{a}}^{0.6}\right)}{\left({\mathsf{\epsilon }}_{\mathrm{g}}\times {\mathrm{D}}^{0.4}\right)}}\right]\times \left({\mathrm{T}}_{\mathrm{g}}-{\mathrm{T}}_{\mathrm{a}}\right)\right\}}^{{\displaystyle \frac{1}{4}}}-273$$

(1)

where T_a represents the air temperature, T_g represents the black globe temperature, V_a represents the wind speed, D is the black globe thermometer diameter, and ε_g is the emissivity.

2.3. Data Processing

2.3.1. Heat Rate Recovery and Skin Conductance Recovery Rate

In this study, the activation of the sympathetic nervous system significantly enhanced the contractility of both the ventricles and the atria, leading to an increased HR [46]. Conversely, enhanced parasympathetic activity during the recovery phase resulted in a decreased HR. The EDA measurements, indicative of sympathetic nerve activity, were crucial as they influence sweat gland secretion via postganglionic fibers, increasing skin conductance levels. After the stimulus ended, the skin conductance gradually returned to the baseline, indicating an electrodermal response [47]. The raw EDA data were filtered to transform them into skin conductance (SC), which was processed on the ErgoLAB physiological testing cloud platform using filters to minimize noise.

The focus of this study was to analyze the heart rate (R_HR) and skin conductance recovery rate (R_SC) to evaluate emotional recuperation through comparative assessments before and after stimulus exposure. This methodology involved calculating the stress response rate (R) by dividing the difference in the average physiological indices (F1 during stimulation and F2 during recovery) by F1, as follows [29,48]:

$$\mathrm{R}=(\mathrm{F}1-\mathrm{F}2)/\mathrm{F}2$$

(2)

2.3.2. Metabolic Rate

The metabolic rate in humans can be inferred via alternative metrics, such as the heart rate. In this study, the heart rate was used as a surrogate marker of the metabolic rate. The correlation between the empirically obtained metabolic rate and the heart rate was determined as follows [49]:

$$\mathrm{M}\prime ={\displaystyle \frac{\mathrm{H}\mathrm{R}-{\mathrm{H}\mathrm{R}}_0}{\mathrm{R}\mathrm{M}}}+{\mathrm{M}}_0$$

(3)

where $\mathrm{M}\prime$ is the metabolic rate (W/m²), ${\mathrm{M}}_0$ is the metabolic rate in the inactive state (W/m²), HR is the heart rate at that moment, ${\mathrm{H}\mathrm{R}}_0$ is the heart rate at rest under thermally neutral conditions, and RM is the increase in the heart rate per unit metabolic rate, calculated as follows:

$$\mathrm{R}\mathrm{M}={\displaystyle \frac{{\mathrm{H}\mathrm{R}}_{\mathrm{m}\mathrm{a}\mathrm{x}}-{\mathrm{H}\mathrm{R}}_0}{\mathrm{M}\mathrm{W}\mathrm{C}-{\mathrm{M}}_0}}$$

(4)

where ${\mathrm{H}\mathrm{R}}_{\mathrm{m}\mathrm{a}\mathrm{x}}$ is the maximum heart rate (${\mathrm{H}\mathrm{R}}_{\mathrm{m}\mathrm{a}\mathrm{x}}=205-0.62\times \mathrm{A}$) and MWC (W/m²) is the maximum working capacity described as follows:

$${\mathrm{M}\mathrm{W}\mathrm{C}}_{\mathrm{m}\mathrm{a}\mathrm{l}\mathrm{e}}=\left(41.7-0.22\times \mathrm{A}\right)\times {\mathrm{W}}^{0.666}\hspace{0.17em}\text{and}$$

(5)

$${\mathrm{M}\mathrm{W}\mathrm{C}}_{\mathrm{f}\mathrm{e}\mathrm{m}\mathrm{a}\mathrm{l}\mathrm{e}}=\left(35.0-0.22\times \mathrm{A}\right)\times {\mathrm{W}}^{0.666}$$

(6)

where $\mathrm{A}$ is the age in years and $\mathrm{W}$ is the weight in kg.

2.4. Questionnaire Design and Analysis

This study was conducted during 1–4 October, 6–8 October, and 11 October 2023, with survey hours ranging from 9:00 a.m. to 6:30 p.m. Volunteers were recruited from the entire student body, resulting in a total of 640 collected questionnaires. The questionnaire consisted of three sections.

The first section collected basic respondent information, such as gender, age, height, weight, clothing, location of the experience space, and physiological signals from the experience space.

The second section employed the perceived restorativeness scale (PRS) [50] (

Table A2. Perceived restorativeness scale (PRS).

Dimension	Number	Content
Being Away	1	Being here is an escape experience.
	2	Spending time here gives me a break from my day-to-day routine.
	3	It is a place to get away from it all.
	4	Being here helps me to relax my focus on getting things done.
Fascination	5	This place has fascinating qualities.
	6	I want to get to know this place better.
	7	There is much to explore and discover here.
	8	I want to spend more time looking at the surroundings.
	9	The setting is fascinating.
Coherence	10	There is too much going on.
	11	It is a confusing place.
	12	There is a great deal of distraction.

). Scores were derived for each dimension, serving as assessment metrics for “being away”, “fascination”, and “coherence”. Due to the experiment’s sole reliance on sedentary activities, the “compatibility” dimension was deemed less relevant and thus excluded. The responses were structured using a seven-point Likert scale.

The third section comprised a subjective thermal perception questionnaire incorporating a thermal sensation vote (TSV), humidity sensation vote (HSV), wind sensation vote (WSV), thermal comfort vote (TCV), and unacceptability rate vote (URV). Considering the potential significant weather variations during the survey period, an expanded nine-point thermal sensation scale was adopted. The TSV utilized the following scale: −4 for very cold, −3 for cold, −2 for cool, −1 for slightly cool, 0 for neutral, 1 for slightly warm, 2 for warm, 3 for hot, and 4 for very hot. Thermal comfort was evaluated on a three-tier scale (−1 representing discomfort, 0 as neutral, and +1 indicating comfort). Thermal acceptability was established on a dichotomous scale: −1 for unacceptable and +1 for acceptable. Additionally, the ASHRAE-endorsed tripartite scale was utilized to document the participants’ preferences regarding wind (WP), humidity (HP), and temperature (TP). Instances where participants provided invalid feedback due to excessive heat were filtered out, considering the experimenter’s observations of the participants’ physical state.

Questionnaires exhibiting inconsistencies between the participants’ physical conditions and the reported thermal sensations were deemed invalid. To ensure the participants’ comprehensive understanding, each questionnaire item was meticulously explained before the official survey and complemented by preliminary training. All participants were confirmed to have no history of illnesses, such as hypertension or cardiovascular disease, and had abstained from medication, caffeine, or other stimulants prior to the experiment. Interactions among participants were prohibited during the experiment to avoid alterations in their psychophysiological states due to social engagement.

2.5. Experimental Procedure

The experimental protocol comprised three stages: preparatory, stress-stimulatory, and recovery (Figure 2). Throughout the experiment, each participant was involved in one of the three scenarios.

Stage 1: Participants were greeted at the registration desk, where researchers clarified the study’s objectives, experimental protocol, types of data to be collected, and principles of confidentiality. After a brief period of rest to ensure physiological stability, baseline physiological signals (HR and EDA) were recorded after a 5-min seated rest period. Subsequently, participants were asked to complete the PRS and a thermal comfort questionnaire.

Stage 2: Participants performed three backward digit span tasks designed to induce cognitive fatigue. This method has been shown to be effective in stimulating stress [38,51]. The researchers then measured the participants’ physiological signals again and administered the PRS and thermal comfort questionnaires.

Stage 3: Participants were allocated to various spaces following a Latin square balanced order. They were then required to sit still, adapt to the environment, and observe the scenery for 5 min. After this duration elapsed, the researchers assessed the participants’ physiological indices post-recovery, and once again, the participants completed the PRS and thermal comfort questionnaires.

3. Results

3.1. Descriptive Statistics

3.1.1. Basic Information on the Survey Respondents

In total, 657 questionnaires were distributed and 640 valid questionnaires were returned, with 260 from men and 380 from women. Among them, 145, 181, and 167 were completed in Spaces A, B, and C, respectively. The average age, weight, and height for the men were 19.9 years, 65.3 kg, and 173.1 cm, respectively; for women, these averages were 20.2 years, 51.0 kg, and 162.2 cm, respectively.

Table 2. Anthropometric data for the subjects (SD: standard deviation).

Sex	Number	Anthropometric Data of Subject	Mean	SD	Minimum	Maximum
Men	260	Age in years	19.9	1.74	17	25
		Weight in kg	65.3	10.98	44	95
		Height in cm	173.1	5.39	160	182
Women	380	Age in years	20.2	1.42	18	25
		Weight in kg	51.0	6.79	39	80
		Height in cm	162.2	5.41	152	179

presents more details about the respondents’ attributes.

3.1.2. Statistics of the Thermal Parameters, Total Clothing Insulation, Metabolic Rate, and Thermal Index in Different Spaces

Table A1. The thermal environment parameters, clothing thermal resistance, metabolic rate, and PET for the three spaces.

Space	Parameters	Abbreviation	Mean	Standard Deviation	Minimum	Maximum
A	Air temperature	Ta (°C)	33.09	4.92	21.42	39.68
	Globe temperature	Tg (°C)	31.65	3.14	24.25	35.66
	Mean radiant temperature	Tmrt (°C)	32.61	3.44	24.22	37.85
	Relative humidity	RH (%)	69.35	4.53	59.00	78.40
	Air velocity	Va (m/s)	1.11	1.17	0.00	6.46
	Total clothing insulation	Icl (clo)	0.36	0.08	0.30	0.79
	Metabolic rate	W (W/m2)	102.46	34.37	60.00	200.99
	Physiological equivalent temperature	PET (°C)	33.1	4.9	21.4	39.7
B	Air temperature	Ta (°C)	34.56	4.52	24.18	40.49
	Globe temperature	Tg (°C)	38.11	6.38	24.14	47.93
	Mean radiant temperature	Tmrt (°C)	47.24	13.34	23.94	85.77
	Relative humidity	RH (%)	64.69	4.80	57.20	77.90
	Air velocity	Va (m/s)	1.71	1.11	0.03	5.97
	Total clothing insulation	Icl (clo)	0.36	0.07	0.30	0.79
	Metabolic rate	W (W/m2)	111.08	38.17	60.00	243.65
	Physiological equivalent temperature	PET (°C)	38.4	7.6	20.6	49.4
C	Air temperature	Ta (°C)	31.73	3.06	24.29	37.16
	Globe temperature	Tg (°C)	32.85	3.51	24.22	38.78
	Mean radiant temperature	Tmrt (°C)	34.77	4.75	24.03	49.55
	Relative humidity	RH (%)	69.29	5.71	58.80	78.70
	Air velocity	Va (m/s)	0.81	0.58	0.00	3.73
	Total clothing insulation	Icl (clo)	0.36	0.07	0.30	0.79
	Metabolic rate	W (W/m2)	100.53	31.40	60.00	194.97
	Physiological equivalent temperature	PET (°C)	33.5	4.5	22.1	40.5
All	Air temperature	Ta (°C)	32.75	3.93	24.18	40.49
	Globe temperature	Tg (°C)	34.71	5.41	24.14	47.93
	Mean radiant temperature	Tmrt (°C)	39.25	10.73	23.94	85.77
	Relative humidity	RH (%)	67.62	5.52	57.20	78.70
	Air velocity	Va (m/s)	1.23	1.05	0.00	6.46
	Total clothing insulation	Icl (clo)	0.36	0.07	0.30	0.79
	Metabolic rate	W (W/m2)	104.97	35.12	60.00	243.65
	Physiological equivalent temperature	PET (°C)	35.2	6.4	20.6	49.4

presents a comprehensive summary of the empirical data for the outdoor thermal parameters, including air temperature (T_a), globe temperature (T_g), mean radiant temperature (T_mrt), relative humidity (RH), wind speed (V_a), and derived physiological equivalent temperature (PET). For the collective assessment of waterfront green spaces, the mean (maximum) values for T_a, T_g, RH, and V_a were recorded as 32.8 °C (40.5 °C), 34.7 °C (47.9 °C), 68.43% (81.5%), and 1.23 m/s (6.46 m/s), respectively. The average clothing insulation (I_cl) among the subjects was 0.36 clo, dropping to a low of 0.3 clo in warmer weather conditions. During cooler weather on the 8th and 11th of October, I_cl peaked at 0.79 clo. The calculated PET and metabolic rates exhibited extremes of 49.4 °C (20.6 °C) and 243.73 W/m² (60 W/m²), respectively.

Among the three spaces, Space B had the highest mean T_a (34.6 °C), T_g (38.1 °C), T_mrt (47.2 °C), and PET (38.4 °C). The openness and maximum SVF of 0.808 in Space B likely contributed to these high values. Conversely, the highest RH was observed in Space A (69.35%) while Space B had the lowest RH (64.69%), likely due to differences in vegetation coverage and transpiration. Additionally, the highest average wind speed in Space B (1.71 m/s) could be due to its proximity to a larger body of water and the scarcity of tall trees.

3.1.3. Thermal Sensation Vote

Figure 3 illustrates the TSV in the three different spaces. In Space A, 20.4% of the university students felt neutral (TSV = 0), 21.1% felt slightly warm (TSV = 1), and 23.8% felt warm (TSV = 2). In Space B, 31.4% of the university students felt warm (TSV = 2) and 24.9% felt very warm (TSV = 3), which was the highest proportion. In Space C, 23.8% of the university students felt neutral (TSV = 0) and 25.6% felt slightly warm (TSV = 1). This indicates that most of the respondents perceived the outdoor thermal environment as generally warm, especially in Space B, where the thermal sensation of the university students was most pronounced.

3.2. Relationship between the Thermal Parameters and TSV

Figure 4 illustrates the relationship between the mean TSV (MTSV) per 1 °C increment and the corresponding T_a, RH, T_mrt, and V_a values in the three spaces. The results indicated that in all three spaces, T_a, T_mrt, RH, and V_a were positively correlated with the MTSV. Specifically, Figure 4a shows the relationship between the MTSV and T_a, with slopes of 0.3483 (A), 0.3190 (B), and 0.3370 (C), and the most significant correlations with R² values of 0.9093 (A), 0.9336 (B), and 0.9111 (C). This indicates that Ta is the most influential meteorological parameter affecting the MTSV. Figure 4b shows the relationship between the MTSV and T_mrt, with slopes of 0.1932 (A), 0.0977 (B), and 0.1867 (C), and R² values of 0.6616 (A), 0.9336 (B), and 0.9111 (C). Figure 4c shows the relationship between the MTSV and RH, with slopes of 0.2301 (A), 0.2480 (B), and 0.1615 (C), and R² values of 0.7763 (A), 0.6988 (B), and 0.6765 (C). This may be because high humidity levels impair evaporative cooling, leading to an increased thermal sensation. Compared with that of T_mrt, university students in waterfront green spaces in hot and humid areas were more sensitive to RH. Figure 4d shows the relationship between the MTSV and V_a, with R² values of 0.1680 (A), 0.5721 (B), and 0.5570 (C), indicating that the instability of V_a in Guangzhou during this period made the university students less sensitive to the changes in wind speed.

3.3. Outdoor Thermal Benchmarks

3.3.1. Neutral PET and Neutral PET Range

Figure 5 illustrates the MTSV corresponding to PET computed in 1 °C increments. These results demonstrated a significant positive correlation between the MTSV and PET across the three spaces. The slopes of the linear regression equations for all spaces and each specific space were 0.2120 (all), 0.2368 (A), 0.2154 (B), and 0.2300 (C), corresponding to PET/MTSV values of 4.7 °C (all), 4.2 °C (A), 4.6 °C (B), and 4.4 °C (C). This suggests that during the transitional seasons in humid and hot regions, university students showed greater thermal sensitivity in Space A.

The neutral temperature is the temperature at which the human body feels neither cold nor hot, i.e., MTSV = 0.

Table 3. Neutral PET and neutral PET ranges in the different spaces.

Space	Regression Formula	R2	Neutral PET(°C)	Neutral PET Ranges (°C)
All	MTSV = 0.2120PET − 6.0756	R2 = 0.9377	28.66	26.30–31.02
A	MTSV = 0.2368PET − 6.6436	R2 = 0.9296	28.06	25.94–30.17
B	MTSV = 0.2154PET − 6.2151	R2 = 0.8924	28.85	26.53–31.18
C	MTSV = 0.2300PET − 6.6874	R2 = 0.8951	29.08	26.90–31.25

summarizes the regression equations, neutral PET values, and neutral PET ranges for all spaces and each individual space. The calculated neutral PET values were 28.7 °C (all), 28.1 °C (A), 28.9 °C (B), and 29.1 °C (C). The neutral temperature range, where −0.5 < MTSV < 0.5, was as follows: the neutral range of PET for all spaces was 26.3–31.0 °C; for Space A, it was 25.9–30.2 °C; for Space B, it was 26.5–31.2 °C; and for Space C, it was 26.9–31.3 °C. These data indicate that in the three spaces, university students experienced a wider range of thermal stress in Space B compared to Spaces A and C, suggesting that a larger water view may assist in adapting to thermal stress, potentially offering greater flexibility and adaptability in temperature regulation.

3.3.2. Thermal Acceptability

The ASHRAE Standard 55 specifies that the acceptable thermal condition is acceptable to at least 80% of respondents [43]. At 1 °C intervals, the percentage of unacceptability corresponding to the PET was calculated for university students in the three spaces, and a polynomial fitting was performed. As shown in Figure 6, although the correlations in Spaces A and C were strong (R² = 0.8250 and 0.8369, respectively), the correlation between the thermal index and the percentage of thermal acceptability in Space B was more significant (R² = 0.8899). This indicates that the thermal environment in Space B has a greater impact on the thermal sensation of university students, possibly due to the larger SVF in Space B, an important factor affecting thermal tolerance. The acceptable PET values for 80% of university students were 34.63 °C (A), 33.68 °C (B), and 36.77 °C (C).

3.4. Psychological Perception Recovery

The Comparison of PRS for the Different Types of Various Waterfront Spaces

Table 4. Comparison of PRS and its different dimensions.

Pairwise Comparison	Spatial Dimension	Mean	Z	p
A–B	A-Being away	11.36	ns	ns
	A-Fascination	14.01	−3.126	0.002 **
	A-Coherence	9.85	ns	ns
	A-PRS	35.19	ns	ns
B–C	B-Being away	11.08	ns	ns
	B-Fascination	15.23	−2.427	0.015 *
	B-Coherence	9.23	ns	ns
	B-PRS	35.24	ns	ns
A–C	C-Being away	11.12	ns	ns
	C-Fascination	14.14	ns	ns
	C-Coherence	9.11	−2.085	0.037 *
	C-PRS	34.40	ns	ns

* p < 0.05, significant difference from the Wilcoxon signed-rank test; ** p < 0.01, highly significant difference from the Wilcoxon signed-rank test; ns, not significant.

presents an analysis of the perceived restorative qualities of the various waterfront spaces, conducted using the Wilcoxon signed-rank test. The total PRS scores across the three types of waterfront green spaces did not differ significantly. Pairwise comparisons within the dimensions of these spaces revealed that in the “being away” dimension, no significant differences were found between Spaces A, B, and C. “Fascination” scores were notably higher for Space B (15.23) than for Spaces A (14.01) and C (14.14). For the “coherence” dimension, Space A (9.85) achieved significantly higher scores than Spaces B (9.23) and C (9.11), suggesting that certain spatial qualities may contribute more significantly to the perception of the coherence of waterfront green spaces.

3.5. Physiological Signal Recovery

3.5.1. Physiological Signal Parameters during the Stimulation and Recovery Phases

Table 5. Physiological signals during the stimulation and recovery phases in the three spaces.

Parameters	Abbreviation (Units)	Stage	Space	Mean	Standard Deviation	Minimum	Maximum
Heart rate	HR (beats/min)	Stimulation	A	98.75	9.03	76.00	117.00
			B	98.60	9.24	76.00	117.00
			C	98.51	9.44	76.00	117.00
		Recovery	A	85.10	13.86	49.00	123.00
			B	87.55	14.86	45.00	120.00
			C	84.84	12.70	50.00	115.00
Skin conductance	SC (μs)	Stimulation	A	7.65	4.68	1.50	22.05
			B	7.28	4.78	0.90	22.05
			C	7.49	4.52	1.19	22.05
		Recovery	A	3.77	2.87	0.49	14.62
			B	5.09	4.11	0.50	22.03
			C	4.34	3.43	0.39	19.28

presents the HR and SC data of the university students during the stimulation and recovery phases in the three spaces. Paired-sample t-tests and Wilcoxon signed-rank tests were used to compare the physiological parameters between the stimulation and recovery phases. During the recovery phase, the average HR in Spaces A, B, and C were 85.05, 87.72, and 84.72 beats per minute, representing a significant reduction from the stimulation phase (p < 0.001) by 13.69, 11.05, and 13.67 beats per minute, respectively. The average SC values for Spaces A, B, and C were 3.77, 5.09, and 4.34 μs, also showing a significant reduction from the stimulation phase (p < 0.001) by 3.88, 2.19, and 3.14 μs, respectively.

3.5.2. Skin Conductance Recovery Rate

Figure 7 presents the details of the calculated R_SC for the university students in the three spaces. The independent-samples t-test analysis indicated significant disparities in the average R_SC values among students in Spaces A, C, and B (p < 0.001). Space B exhibited the lowest R_SC (0.88), potentially due to its highest SVF of 0.808, resulting in thermal discomfort and suboptimal physiological recovery. Conversely, Space A recorded the highest R_SC (1.74), suggesting superior physiological recovery, while Space C maintained a moderate R_SC (1.41). These findings underscore the influence of spatial characteristics, such as the SVF, on the restorative potential of landscape environments.

3.5.3. Heart Recovery Rate

Figure 8 presents the details from the analysis of the university students’ R_HR in the three waterfront spaces, conducted via an independent-samples t-test. The analysis revealed no significant variances in R_HR among students within the different waterfront spaces (p < 0.05). Specifically, the mean R_HR values (with maximum values in parentheses) for Spaces A, B, and C were 0.19 (0.98), 0.15 (1.11), and 0.18 (0.79), respectively.

3.6. Relationship between the Thermal Indices and Restorative Benefits

3.6.1. Relationship between PET and PRS

As depicted in Figure 9, to ascertain the correlation between the PRS and PET values, the PRS scores for “being away”, “fascination”, and “coherence” corresponding to PET at 1 °C intervals were calculated. A significant negative correlation was observed between PET and PRS, as well as the distance across all spaces (R² = 0.6506, R² = 0.6491). However, “fascination” and “coherence” were not significantly correlated (R² = 0.3678, R² = 0.1450). Among the three specific types of waterfront spaces, “being away”, “fascination”, and “coherence” were negatively correlated with PET, albeit not significantly, with Space B showing a moderate correlation between PET and PRS (R² = 0.5693).

3.6.2. Relationship between PET and R_SC

Figure 10 presents the polynomial fitting analysis used to investigate the correlation between PRS and R_SC, with the R_SC values determined for PET at 1 °C increments. The outcome demonstrated an inverted U-shaped relationship between PET and R_SC, characterized by an R² of 0.6023, suggesting a more complex interaction than a straightforward negative correlation. This implies that while the R_SC initially increases with increasing PET to a certain point, it subsequently decreases.

3.6.3. Relationship between PET and R_HR

As shown in Figure 11, the R_HR values corresponding to PET at 1 °C intervals were calculated and subjected to a linear fitting. The results indicated a significant negative correlation between the PET and R_HR across all spaces. The slopes of the linear regression equations for all spaces and each individual space were 0.0119 (all), 0.0163 (A), 0.0122 (B), and 0.0100 (C), with corresponding PET/R_HR values of 84.03 (all), 61.35 (A), 81.97 (B), and 100 (C), respectively. This suggests that as the PET increases, the R_HR decreases, highlighting the importance of thermal comfort in the recovery process within landscape environments.

3.7. Recovery Benefits Corresponding to Different Thermal Stress Categories

Table 6. Relationship between PET and the physiological and psychological recovery indices in all spaces.

Space	Regression Formula	R2
All	PRS = 0.4295PET + 50.1017	R2 = 0.9377
All	Being away = 0.2258PET + 19.3879	R2 = 0.6491
All	Rsc = 0.0053PET2 + 0.3375PET − 3.8402	R2 = 0.6023
All	RHR = –0.0119PET + 0.5969	R2 = 0.7521
All	Fascination = −0.1546 PET +20.5066	R2 = 0.3678
All	Coherence = −0.0708 PET +11.2078	R2 = 0.1450

presents the relationship between PET and the physiological and psychological recovery indices in all spaces.

Table 7. Different categories of thermal stress and the corresponding recovery benefits.

Space	Stress Category	Strong Cold Stress	Cold Stress	Cool Stress	Slightly Cool Stress	No Thermal Stress	Lightly Warm Stress	Warm Stress	Hot Stress	Strong Hot Stress
All	TSV	<−3.5	−2.5 to −3.5	−2.5 to −1.5	−1.5 to −0.5	−0.5 to 0.5	0.5 to 1.5	1.5 to 2.5	2.5 to 3.5	>3.5
All	PRS	/	/	42.86 to 40.83	40.83 to 38.81	38.81 to 36.78	36.78 to 34.75	34.75 to 32.73	32.73 to 30.70	<30.70
All	Being away	/	/	15.58 to 14.51	14.51 to 13.45	13.45 to 12.38	12.38 to 11.32	11.32 to 10.25	10.25 to 9.19	<9.19
All	RSC	/	/	0.34 to 0.98	0.98 to 1.37	1.37 to 1.53	1.53 to 1.45	1.45 to 1.14	1.14 to 0.59	<0.59
All	RHR	/	/	0.40 to 0.34	0.34 to 0.28	0.28 to 0.23	0.23 to 0.17	0.17 to 0.12	0.12 to 0.06	<0.06
All	Fascination	/	/	17.90 to 17.17	17.17 to 16.44	16.44 to 15.71	15.71 to 14.98	14.98 to 14.25	14.25to 13.52	<13.52
All	Coherence	/	/	10.01 to 9.68	9.68 to 9.35	9.35 to 9.01	9.01 to 8.68	8.68 to 8.34	8.34 to 8.01	<8.01

presents the calculated values of PRS, the “being away” dimension, R_SC, and R_HR for various thermal stress categories, based on the regression equations in Table 3 and Table 6. The levels of cold stress were not considered in this study due to the scarcity of respondent evaluations under such conditions. At the “cool” stress level, university students exhibited peak PRS, “being away”, and R_HR values across all environments. As PET increased from 16.9 to 21.6 °C, the values for PRS, “being away”, and R_HR decreased from 42.86 to 40.83, 15.58 to 14.51, and 0.40 to 0.34, respectively, indicating that cool stimuli can enhance recovery benefits for students in humid-warm regions during autumn. When the thermal stress level was within the “lightly warm” stress range with PET rising from 31.0 to 35.7 °C, the students’ R_SC peaked across all environments, from 1.45 to 1.53. This finding suggests that lightly warm stress can boost recovery in university students.

4. Discussion

4.1. Factors Affecting Thermal Comfort

A variety of factors affect thermal comfort. This study analyzed thermal comfort using the comprehensive physiologically equivalent temperature (PET) index. Table 8 presents the comparison of the neutral PET among different space types and climate zones. The results show that the neutral PET for university students in waterside green spaces is 28.7 °C. Compared to the same region and space type, the neutral PET was 1.86 °C higher than that of a summer waterside camping site in Guangzhou [31], possibly due to seasonal variations. Compared to regions with different climates, Guangzhou’s neutral PET is 5.2 °C higher than that of Harbin in the same season [52], suggesting that the university students in Guangzhou have adapted to the hot and humid climate. Neutral PET differences result from seasonal adaptations that affect thermal comfort, primarily reflecting psychological adaptation.

Water features play a crucial role in the thermal comfort of university students in blue-green spaces. In specific spaces, the average PET values for the three spaces were 33.1 °C (A), 38.4 °C (B), and 33.5 °C (C), respectively. Space B is adjacent to a larger water body where solar radiation accelerates water evaporation, increasing surface and nearby air temperatures, while Space C is near a smaller water body. However, a study by Jingcheng Xu et al. [53] found that as the water body area increases, the heat index decreases, with the best thermal comfort observed at 10–20 m away from the water’s edge. This contradicts the results from this study, possibly due to the distance between the space and the water body. Besides water features, the visibility and openness of the sky, which are determined by SVF, are important for thermal comfort.

In green space design, SVF is a critical consideration as it determines the availability of solar radiation. The average PET of Space B with a larger SVF is approximately 5 °C higher than Spaces A and C. SVF determines the heat transfer by radiation between the sky and open space. Tree shading is a primary means to achieve shade. Emilio Terrani et al. [54] found that the number of people claiming to feel thermally comfortable under tree shade was more than twice that in unshaded areas, with another important parameter being the presence of trees and vegetation. The foliage of trees can block solar radiation, reducing the amount reaching the ground [55]. Not only the quantity and type of vegetation, such as grass or trees, can affect the impact of urban vegetation on outdoor thermal comfort, but also the structural, optical, interception, and physiological characteristics of the vegetation [56,57]. In this study, the fully shaded planting method adopted in Space C may have been influenced by the improper arrangement of trees, which affected ventilation. Compared with other spaces, the average wind speed was the lowest, reaching 0.81 m/s. Therefore, while achieving shade through trees, particular attention is paid to tree layout to ensure good ventilation effects.

Table 8. Comparison of the neutral PET among different space types and climate zones.

Location	Type	Season	Climate	Neutral PET(°C)
This study	waterfront green spaces	transition season	subtropical monsoon climate (Cfa)	28.7
Guangzhou, China [31]	waterfront campsite	summer	subtropical monsoon climate (Cfa)	26.8
Harbin, China [52]	high-density urban center	winter, summer, transition season	temperate monsoon climate (Dwa)	24.8, 23.5
Xi’an, China [58]	urban river landscape	winter, spring, and summer	warm temperate semi-humid monsoon climate (Cwa/Bsk)	21.7, 20.5, and 18.9
Jijel City, Algeria [59]	square and public garden	winter	mediterranean climate with mild winters and hot, humid summers (Csa)	20.4
Isfahan, Iran [60]	blue-green infrastructure	winter	arid climate (Bwk)	26.2
Haryana [61]	urban park	summer	tropical savanna climate (Bsh)	30.8

Table 8. Comparison of the neutral PET among different space types and climate zones.

Location	Type	Season	Climate	Neutral PET(°C)
This study	waterfront green spaces	transition season	subtropical monsoon climate (Cfa)	28.7
Guangzhou, China [31]	waterfront campsite	summer	subtropical monsoon climate (Cfa)	26.8
Harbin, China [52]	high-density urban center	winter, summer, transition season	temperate monsoon climate (Dwa)	24.8, 23.5
Xi’an, China [58]	urban river landscape	winter, spring, and summer	warm temperate semi-humid monsoon climate (Cwa/Bsk)	21.7, 20.5, and 18.9
Jijel City, Algeria [59]	square and public garden	winter	mediterranean climate with mild winters and hot, humid summers (Csa)	20.4
Isfahan, Iran [60]	blue-green infrastructure	winter	arid climate (Bwk)	26.2
Haryana [61]	urban park	summer	tropical savanna climate (Bsh)	30.8

4.2. The Recovery Benefits of Different Types of Waterfront Green Spaces

This study explored the recovery of physiological signals, specifically the HR and skin SC. The average HR in the three distinct environments decreased by 13.69 (A), 11.05 (B), and 13.67 (C) beats/minute. Compared with prior research, where HR reductions were 1.92 and 3.14 beats/minute in front of small and large green walls, respectively [28], this study exhibited a more pronounced decline. This could be attributed to the stimulation phase incorporated in this study, which involved higher heart rate stimulation coupled with the restorative nature of the actual waterfront greenery setting. Previous studies have also verified that aquatic scenes consistently rank highest in terms of restorativeness [17]. The SC averages significantly dropped by 3.88 (A), 2.19 (B), and 3.14 (C) μs. This aligns with ART, as the presence of natural landscape elements can attenuate the effects of stimuli, activate the parasympathetic nervous system, and facilitate relaxation and stress reduction [11,12]. Moreover, while previous research has affirmed the restorative nature of natural settings, this study observed negative recovery rates among participants, potentially due to the high outdoor temperatures inducing thermal stress and discomfort, and thereby exacerbating stress levels [62]. This is in contrast to another study on college students in urban parks, which reported higher levels of perceived restorativeness [24]. Climate variations play a critical role in recovery outcomes, as evidenced by our study where temperatures up to 35 °C contributed to what we term ‘negative recovery rates’. Urban parks, less familiar to students than their usual campus environments, augment a sense of remoteness from everyday stressors [63]. Forests, with their dense foliage, not only shield against direct sunlight and help maintain moisture levels but also create secluded, layered environments that provide a sense of mystery and security. This feeling of mystery makes the vegetation more appealing [64]. In contrast, the expansive open fields of grasslands with their simple vegetation structure reduce visual and psychological clutter, further fostering a sense of restoration.

The recovery effects varied across spaces, which may have stemmed from the distinct environmental characteristics and measurement methodologies employed. Comparative results of the perceived recovery dimensions indicated that Space B achieved a higher fascination score (15.23), whereas Space A was superior in terms of coherence (11.36). This could be attributed to the large water body in Space B, which, as supported by previous studies, can mitigate stress. The visibility of water bodies has been shown to correlate positively with stress relief and is conducive to relaxation [65,66]. Conversely, Space A likely offered a sense of shelter due to its overhead tree coverage, creating a covered space akin to a refuge with one side open for external observation. This setup can stimulate exploratory desires and enhance the feelings of coherence and shelter [67].

4.3. Recommendations for Planning Long-Term Health Based on the Physiological and Psychological Differences among Students

Previous research has indicated that decreased skin conductance is positively associated with a higher PRS [67]. However, this study found that the lowest R_SC was recorded in Space B while the PRS score was the highest, particularly in the fascination dimension which reached 15.23. This discrepancy may stem from previous studies conducted under virtual reality (VR) conditions, whereas this study utilized on-site field observations. Although VR provides controlled experimental conditions, critical environmental variables may be overlooked [48]. This study considered microclimatic factors in actual settings, such as temperature, humidity, and wind speed, which play significant mediating roles in physiological recovery. Open spaces are often susceptible to solar radiation, potentially reducing physiological recovery effects. Despite the physical discomfort, individuals still experience psychological recovery and pleasure, highlighting the complex interactions between cognition, psychology, and physiology in dynamic environments. However, physiological discomfort does not necessarily indicate an unhealthy state. We observed that as the PET increased, the human body experienced thermal discomfort, leading to a decrease in HR, which was consistent with previous research [68]. On the subjective level of perceived recovery, earlier studies indicated that sunlight can activate the human thermosensory system [69]. Space B, with a high SVF of 0.808 and extensive water visibility, offers a sense of openness. While an increase in PET may enhance thermal discomfort, previous studies suggest that the presence of sunlight enhances the visual appeal of landscapes, thereby boosting psychological recovery [70,71]. Previous studies have confirmed that sunlight is closely associated with increased well-being and reduced depression.

Physiological signals and psychological adaptations reveal the potential long-term effects of thermal discomfort on the college students’ overall health and well-being. Physiological signals exhibit sensitivity to external environmental stimuli. However, psychological adaptations might experience a certain degree of lag [72]. This delay in psychological adjustment, which only gradually changes with increased exposure time, can lead to adverse outcomes. If psychological adaptation fails to keep pace with physiological warning signals, the accumulation of negative effects might ensue, thereby posing long-term health risks to students.

From a physiological health perspective, studies have shown that the prolonged exposure to sunlight and excessive absorption of ultraviolet rays can increase the risk of skin cancer [73]. Additionally, high temperatures have been linked to an increased risk of myocardial infarction in the short term [74]. In terms of mental recovery, thermal discomfort can lead to irritability and anxiety [75]. Prolonged thermal discomfort may trigger emotional problems, impacting learning and social activities.

An individual’s thermal adaptation capabilities vary based on age, gender, geographical location, and health status [76]. Therefore, urban planners and campus developers should implement personalized alert systems for different student groups and various activities. Some scholars suggest that physiological signals from wearable biosensors can reliably and accurately estimate the likelihood of heat stress among onsite workers. For activities like outdoor club surveys and campus open-air assemblies, it is crucial to consider the characteristics of the space, the duration of the activity, and the number of participants, to provide appropriate warning measures.

Moreover, enhancing psychological adaptability through design and cognitive approaches can also mitigate thermal discomfort. Studies suggested that mitigating the adverse effects of a specific environmental factor, such as heat, could be achieved by enhancing other aspects, such as visual elements [77]. For example, employing visually cool colors, and orienting spaces towards water can collectively enhance heat adaptability. Cognitively, enhancing environmental education and knowledge about the thermal environment can help students cope with heat discomfort more effectively, and promoting health concepts can strengthen their heat adaptation capabilities.

In summary, recognizing the differences between physiological signals and psychological adaptations is essential for crafting effective interventions and enhancing student health and comfort on campus.

4.4. Paths to Optimize Thermal Comfort Models through Psychological Factors

Traditional thermal comfort models mostly used neutral thermal sensation to measure thermal comfort, which had limitations and failed to take psychological factors into account. In unstable outdoor settings, the individuals’ subjective psychological states fluctuate, leading to a misalignment between neutral and ideal thermal conditions. The alliesthesia theory suggests that in a thermal environment, whether any transient heat sensation detected by our skin was pleasant or unpleasant, depends on whether the subsequent regulation of our body’s overall thermal balance pushes it towards or away from the thermal neutral zone [78]. This indicates that emotions played an important role. Our study delineated the PET ranges that yield high recovery benefits under various thermal stress conditions. When the thermal stress level reached “lightly warm” stress (PET rising from 31.0 to 35.7 °C), the R_SC values peaked (1.45 to 1.53). Research has also suggested that exposure to thermally comfortable conditions, in which warm contact stimuli are applied to the hands and feet, can evoke pleasure [79]. This may be because the external spaces at the three study sites experienced “hot” and “very hot” thermal stresses, which contrasted sharply with the thermal stress levels within the study spaces. When the participants moved from the external space to the study spaces, as long as they did not enter the overheated areas, the difference in thermal stress contributed to a sense of physical and psychological satisfaction, fostering physiological recovery. Psychological expectations regulate people’s perception thresholds for thermal stimuli. Anticipating a warm environment may increase tolerance to slightly higher temperatures, while expecting a cool one heightens sensitivity to high temperatures.

Regarding the connection between psychology and thermal comfort, this research found that the TSV was especially influenced by perception and psychological adaptation factors. Future research can establish the relationship between them by quantifying psychological factors and building a heat-related model, determining the influence weights and elements of psychological factors on thermal comfort prediction. This will help establish the connection between them and improve the predictive ability of the thermal comfort model. Figure 12 proposes an integrated model that combines thermal, psychological, cognitive, and emotional factors, offering a potential research framework that benefits improvements in microclimate modeling, directing future research.

5. Limitations

This study had several limitations. Primarily, it focused exclusively on university students, a demographic that may not represent the broader population. Furthermore, the study was confined to short-term exposure to outdoor heat during sedentary activities, which limited the generalizability of the findings to other conditions and activities. Additionally, due to the composition of the participant pool in the university, the study did not consider gender differences, which could influence thermal perception and comfort. Future research should broaden the demographic scope, extend the duration of the studies, explore the impact of various activities, and integrate simulation tools. Incorporating these elements would provide a more comprehensive understanding of how different environments and activities influence psychological restoration and thermal comfort.

6. Conclusions

This study investigated the thermal comfort and restorative benefits for college students in waterfront green spaces in humid and hot regions. These results compared the thermal comfort and restorative benefits. Based on these findings, this paper proposes landscape optimization suggestions for waterfront green spaces. These findings are summarized, as follows:

• Thermal benchmark analysis: based on a linear regression model of PET and MTSV, the neutral PET of the campus waterfront green space was calculated to be 28.66 °C. The neutral PET of the three spaces were 28.1 °C (A: SVF = 0.228), 28.9 °C (B: SVF = 0.808), and 29.1 °C (C: SVF = 0.292), respectively. This indicated that the adaptability and tolerance of college students to the thermal environment were influenced by various factors, including the water body landscape, SVF, and the contrast of the surrounding thermal environment;
• Physiological signal recovery: the space with a high SVF value (B: SVF = 0.808) had a lower R_SC, indicating poorer physiological recovery effects due to thermal discomfort. However, the space with a low SVF value (A: SVF = 0.228) had a higher R_SC (1.7358), indicating better physiological recovery;
• Psychological restoration perception: Space B demonstrated enhanced fascination, whereas Space A showed superior coherence;
• Thermal comfort and restorative benefits: there was a significant negative correlation between the PET of waterfront green spaces and the PRS scores, particularly for the “being away” dimension and R_HR;
• The PET range for high restorative benefits in waterfront green spaces of hot and humid regions: at “cool” stress levels, the PRS, “being away”, and R_HR values peaked for the university students as the PET ranged from 16.9 to 21.6 °C, with decreases from 42.86 to 40.83, 15.58 to 14.51, and 0.40 to 0.34, respectively. As PET increased from 31.0 to 35.7 °C under “lightly warm” stress, the range for Rsc peaked from 1.45 to 1.53.

To enhance the restorative benefits, this study offers the following design guidelines for campus waterfront green areas:

• Enhance the spatial shelter: design enclosed spaces and select densely canopied plants to create shaded layouts;
• Optimize the relationship between the space and the water: emphasize visual water elements and appropriately design the distance between water bodies and spaces;
• Enrich thermal experiences and stimulate thermal pleasure: design a variety of microclimatic spaces using hard plazas, water bodies, and vegetation to optimize spatial layouts and enhance the contrast in the thermal environment.