Introduction

The rapid urbanization process exacerbates the urban heat island phenomenon, leading to an increased thermal burden on open spaces within campuses1. This escalation in temperature significantly reduces student utilization of these areas, detrimentally affecting their physical and psychological welfare1. Thoughtfully designed campus open spaces, featuring a variety of facilities and substantial vegetation coverage, serve as effective mitigators of the heat island effect, thereby playing an essential role in promoting student outdoor activities2. Thermal comfort is identified as a key environmental factor in crafting open spaces that are inviting to students. Studies have shown that incorporating thermal comfort considerations into the design of campus open spaces can potentially elevate student usage and contribute to their overall well-being3.

Empirical research on thermal comfort in campus open spaces has elucidated diverse findings. Dong Yuqin et al.4 assessed thermal comfort in six specific spatial layouts at Fujian Agriculture and Forestry University, which is located in a subtropical climate. Their findings indicated that lawn areas provided optimal thermal comfort in winter due to the mild temperatures and increased solar exposure typical of this region. This contrasts with studies conducted in other climates, where different environmental features may be more suitable for achieving thermal comfort during colder months. Further analysis by Xi Tianyu and Lei Yongsheng5 on thermal comfort among park users demonstrated a linear correlation between thermal sensation votes and PET in Harbin across summer and winter, with the neutral PET value in summer significantly exceeding that of winter. Their research indicated a quadratic relationship between thermal comfort and thermal sensation votes in summer, whereas a linear relationship was observed in winter, alongside a quadratic correlation between heat acceptance rates and PET in summer, with the PET range for a 90% acceptance rate extending from 11.6 to 35.9 °C. Niu Jiaqi6 explored thermal comfort among students engaging in various intensities of physical activities on campus during summer, concluding that elevated activity levels prompted preferences for cooler air temperatures and increased wind speeds, with negligible variations in humidity and solar radiation preferences. Neutral PET values were determined at 26.1 °C for low, 22.1 °C for medium, and 11.9 °C for high activity intensities, highlighting a rise in thermal discomfort with increased activity intensity. Chen Xin7 investigated thermal comfort at the Harbin Institute of Technology, noting an augmented sensitivity to outdoor conditions amid significant indoor and outdoor temperature disparities. Xie et al.8 explored thermal comfort in overhead public spaces at Shenzhen University, a region characterized by a hot and humid climate. The study highlighted that the acceptable physiologically equivalent temperature (PET) range varied significantly between the hot and cold seasons, and respondents exhibited different thermal adaptation behaviors based on seasonal variations. Similarly, Liu et al.9 conducted a field survey on local thermal comfort at a university campus in Shanghai. Their research demonstrated that the PET values varied significantly across different spatial zones within the campus, influenced by the surrounding spatial morphology and environmental conditions. Ning et al.10 took a novel approach by investigating the relationship between sunlight perception and outdoor thermal comfort on college campuses. Their findings revealed that people’s thermal comfort and desired temperatures were closely linked to their sunlight perception, with variations observed between different seasons. In cold regions, Guo et al.11 studied the spatial differences in outdoor thermal comfort during the transition season in Dalian, Northeast China. Using the Universal Thermal Climate Index (UTCI) as an evaluation metric, they found significant differences in thermal comfort across different spatial types, such as sunlight, tree shade, and building shade environments. These studies collectively emphasize the need for a nuanced approach to designing university campuses that can accommodate diverse thermal comfort needs. While existing research has provided a solid foundation, there remains a gap in understanding how specific microclimates within university campuses affect thermal comfort. This study aims to address this gap by exploring microclimatic differences across various underlayments within a university campus, thereby contributing to the development of targeted design strategies that enhance thermal comfort in educational settings.

Scholars have provided solutions worth learning for the design of different outdoor open spaces. Wu et al.12 championed sustainable resource management in campus open spaces, employing zoning planning and vegetation selection as key strategies to this end. Ayse et al.13 directed their planning and design efforts towards the open spaces on campuses with an intent to advance sustainable development objectives. Campos et al.14 articulated how campus planning and architectural design can be seamlessly integrated with the local cultural narrative, accentuating the value of cultural integration in design processes. Through the application of deep learning technology, Loo et al.15 deciphered tourists’ movement patterns, offering insights into enhancing outdoor engagement and enjoyment via strategic design interventions. Investigating the open spaces in Portugal, Vidal et al.16 probed into the interplay between urban public green spaces and user behavior, proposing design modifications aimed at optimizing the effectiveness and utilization of these areas.

Campuses, as the living environment for college students, should regard environmental comfort as a necessary condition for outdoor activities. Creating comfortable campus open spaces is beneficial for improving students’ physical and mental health, enhancing attention, alleviating stress, and thus improving the learning and cognitive environment. Therefore, improving the environmental comfort of campus open spaces plays an important role in providing students with a balanced and healthy life. Our study selected four iconic open spaces on a campus in Xi’an, China, a region characterized by a cold climate. The purpose is to investigate the thermal perceptions of college students during outdoor activities and to carry out adaptive redesigns. The main objectives of this study are: (1) to identify the key factors affecting students’ thermal perceptions in campus open spaces; (2) to formulate an optimal design strategy for campus open spaces based on meteorological data, students’ commuting patterns, and thermal perceptions, in order to enhance thermal comfort and increase student attendance. The research findings can provide a theoretical basis and technical guidance for creating comfortable campus open spaces based on thermal comfort.

Methods

Study area

Xi’an is located on the boundary between a semi-arid (BSk) and humid subtropical climate (Cwa)1. Xi’an is hot and humid in summer and cold and dry in winter. According to Xi’an’s meteorological data statistics from 2020 to 2023, it is known that the annual average high temperature is 22 °C and the maximum temperature is 41 °C. The lowest average yearly Ta is 10 °C, and the lowest is − 11 °C.

As shown in Fig. 1, this study selected four open spaces at a university in Xi’an as outdoor test locations, including Entrance Square (ES), Life Square (LS), Leisure Garden (LG), and Cultural Square (CS). The selection criteria As shown in Table 1. This selection process underscores the scientific approach employed in identifying optimal locations for examining thermal comfort and environmental interaction within campus open spaces.

Fig. 1
figure 1

Source: PPTX (2016). Retrieved from [https://kdocs.cn/l/cdMMti2DIj5d].

Site location and measured space.

Full size image
Table 1 Description of the four open spaces.
Full size table

Experimental design

Meteorological measurement

The research was planned from April 19, 2023, to July 19, 2023, with surveys conducted from 8:00 to 17:00 each day. Meteorological data including air temperature (Ta), relative humidity (RH), wind speed (Va), solar radiation (G), and black globe temperature (Tg) will be accurately recorded every minute. All equipment will be positioned at a height of 1.1 m above the ground to ensure uniform data collection. This experiment would be carried out only under conditions that are not affected by wind or rain. In scenarios of adverse weather, such as significant wind or heavy rainfall, the experiment will be paused. Details of the instruments used are provided in Table 2.

Table 2 Instrument parameter table.
Full size table

Questionnaire survey and activity records

During the experiment, university students who were free to move around in the measurement space were randomly selected to complete the questionnaire survey17. The first part of the questionnaire collected demographic information of the participants, including biological sex, age, height, weight, clothing, and main activity type. The thermal resistance of the clothing materials of the respondents was referenced to ASHRAE Standard 55 and ISO 773018,19. The activity type of the respondents was recorded to determine the metabolic rate of the subjects.

The second part of the questionnaire investigated the respondents’ thermal perception, including sensation, preference, comfort, acceptability and satisfaction. Thermal sensation was expressed using the ASHRAE 7-point scale, thermal preference was recorded using a 3-point scale, and thermal acceptability was recorded using a 2-point scale. In order for the respondents to accurately select their thermal perception, they were allowed to select any scale based on their actual feeling (Appendix).

The third part of the questionnaire asked the respondents to choose their preferred heat adaptation behavior (go under a shaded building, go to the shade of a building, go to the shade of a tree to cool off, reduce outdoor activities/exercise, go indoors and turn on a fan, use an umbrella, reduce clothes, go indoors and turn on air conditioning/cool air, drink iced beverages). The number of questionnaires conducted every 30 min was not fixed but was determined by the number of eligible respondents who participated in outdoor activities in each measurement space.

During the experiment, video cameras were used to observe and record student activities in the four areas20,21. The investigators stood in a position where they could observe the entire space and recorded in detail the number of students in each area and the type of their activities. To ensure the accuracy of the data, the investigators repeated the counting procedure at intervals of 60 min (such as 13:00, 14:00, 15:00, etc.). This study only counted college students who were active in each space for more than 5 min, and these students stayed for a long time (such as sitting or exercising) to reflect their true intention of using the space. Those who just passed by or stayed for a short time were not included in the statistics.

Respondent attributes

The questionnaire survey undertaken for this study was aimed at students attending a university in Xi’an, successfully gathering 545 “valid” responses. Among these participants, the gender distribution was 61% male and 39% female. The information of the respondents is shown in Table 3. Analysis of the participant’s activities in the 20 min preceding the survey revealed diverse engagement: 28% of the students were observed walking at a standard pace; 27% were found to be seated, either reading or interacting with their mobile phones; 19% stood while chatting or utilizing their mobile phones; 18% were in resting positions, such as sleeping or lying down; and a minor 4% were involved in vigorous physical activities, including brisk walking, cycling, running, or playing sports. It is noteworthy that all respondents had been residing in Xi’an for a period exceeding seven year, indicating their full acclimatization to the local climate. This demographic detail adds a layer of credibility and objectivity to their assessments of the outdoor thermal environment.

Table 3 Respondent attributes.
Full size table

Thermal comfort index

The Physiological Equivalent Temperature (PET), serving as the evaluative benchmark for human thermal comfort in this inquiry, assimilates the cumulative effects of ambient temperature and humidity on the human thermal condition. Endorsed by extant literature as a robust and functional metric, PET facilitates the appraisal of thermal comfort under varying environmental milieus. The comparative analysis of PET values across disparate settings allows for an insightful delineation of comfort levels under different climatic circumstances.

In the ambit of this research, a comprehensive set of variables including air temperature (Ta), relative humidity (RH), wind speed (Va), and solar radiation (G), alongside individual-specific determinants like metabolic rate and clothing thermal resistance, were meticulously inputted into the Rayman software (RayMan 1.2, https://www.urbanclimate.net/rayman/index.htm), facilitating the derivation of personalized PET estimations.

Results

Descriptive analysis

Meteorological parameters

Table 4 reveals that the LG has the apex mean Ta and RH at 25.9 °C and 62.7%, respectively, contrasting sharply with the ES, which records the minimum mean Ta at 23.8 °C. The LS and CS are noted for the lowest average RH, pegged at 54.1% and 54.3% respectively. ES distinguishes itself with the maximal average Va of 0.7 m/s, while LG stands out for the highest G at 302.8W/m2, with LS marking the lowest in solar radiation at 120W/m2. The Tg displays minimal fluctuations across the venues. The notable high maximum Ta at ES is attributed to its expansive landscape lake, a dearth of street trees, and the absence of shading constructs, which also account for the peak maximum and average Va observed at this site.

Table 4 Meteorological parameter values at various locations.
Full size table

Conversely, LG, with its unique “back” configuration, accentuated by fountains, water elements, and a dense canopy of shrubs and shade trees, manifests the lowest Va and the supreme average RH among the evaluated sites.

The architectural and natural designs of LS and CS, incorporating elements like small fountains, large trees, spray water sculptures, and shaded walkways, respectively, afford partial shading, thereby moderating their meteorological parameters Ta and Va to intermediate levels among the assessed spaces.

Thermal comfort vote and thermal sensation vote

According to the thermal sensation votes depicted in Fig. 2, the ES experienced a significant majority, 50%, of participants identifying as feeling “very hot” or "hot," with an additional 18.75% categorizing their sensation as "hot," contrasting sharply with a mere 18.75% reporting sensations of "cold," "very cold," or "cold."

Fig. 2
figure 2

TSV proportion distribution in four locations.

Full size image

In contrast, the LG saw the “moderate” sensation prevailing among 25% of its respondents, indicating a generally agreeable and comfortable temperature within this area. A further 37.5% of participants reported feeling “colder” or “hotter”.

For LS, the “hot” sensation was most prevalent, reported by 25% of the participants, suggesting that LS harbors a relatively high temperature that tends to make individuals feel somewhat warm; nonetheless, a considerable fraction, 18.75%, experienced a “moderate” sensation.

Similarly, in the CS, 25% of respondents perceived a “hot” sensation, indicating higher temperature levels leading to discomfort, while a combined total of 37.5% of participants described feeling “hot” or “very hot”.

Figure 3 delineates the thermal comfort survey results. In the ES zone, a predominant fraction of respondents, 25%, reported relative discomfort, indicating potential over-high ambient temperatures or other adverse environmental factors. Both moderately comfortable and uncomfortable categories garnered 18.75% of responses respectively, whereas a mere 6.25% felt comfortable or very comfortable, highlighting a crucial need for environmental quality enhancements in this locale.

Fig. 3
figure 3

TCV proportion distribution in four locations.

Full size image

The LG vicinity exhibits a quarter of respondents with moderate thermal sensations, reflecting the adequacy and stability of temperature and environmental conditions, catering well to the majority’s preferences. The counts of more comfortable and less comfortable respondents followed, with very small proportions, 6.25% each, feeling uncomfortable or very uncomfortable, suggesting an overall satisfactory environmental quality that fulfills most requirements.

In LS, 25% of participants indicated moderate comfort, signaling that the area’s environmental conditions are generally stable and meet the majority’s expectations. An 18.75% segment feeling less comfortable points to occasional fluctuations in environmental factors leading to discomfort for some.

The CS region demonstrated a balanced distribution among comfort levels, suggesting varied perceptions towards the area’s temperature and other environmental influences, underlining the subjective nature of thermal comfort.

Correlation between TSV and TCV

Figure 4 elucidates the correlation between TCV and TSV within the ES, LG, LS, and CS areas. In ES, an increase in TSV precipitates an initial increase in TCV, followed by a subsequent decline, a trend that intensifies with higher thermal sensation votes. This pattern underscores a significant sensitivity to thermal comfort in ES, where variations in TSV are intimately linked to changes in TCV.

Fig. 4
figure 4figure 4

Correlation diagram between TCV and TSV.

Full size image

LG, while following a similar pattern to ES of initial increase and subsequent decrease in both TSV and TCV, demonstrates a greater responsiveness in TCV, to changes in TSV than observed in ES. This suggests that in LG, temperature variations have a diminished impact on the subjective thermal experiences of individuals.

Figure 4 elucidates the relationship between TSV and TCV across different locales. For ES, a TSV ranging from − 1.26 to 0.63 correlates with a positive TCV, indicating a particular sensitivity to thermal comfort. In LG, TSV values span from − 2.50 to 2.15, also resulting in a positive TCV, reflecting a more stable thermal comfort perception among its inhabitants. LS’s TSV ranges from − 2.18 to 1.59, and CS from − 1.87 to 1.33, both leading to TCVs greater than zero. These findings underscore the significant influence of environmental characteristics and personal thermal sensations on thermal comfort vote, highlighting the variance in thermal comfort sensitivity and stability experienced by residents of ES and LG, respectively.

Meteorological variables preferences

Figure 5A illustrates the temperature preferences among residents of ES, LG, LS, and CS. In the ES, there is a strong preference for a reduction in temperature, with the majority of respondents expressing a desire for cooler conditions. In contrast, the LG shows a preference for maintaining the current temperature, indicating a general satisfaction with the existing climate. The LS and CS also show a preference for cooler temperatures, with a smaller proportion of respondents favoring an increase.

Fig. 5
figure 5

Preference map of meteorological variables at each observation location.

Full size image

Figure 5B depicts respondents’ preference for G in four locations. In the ES, there is a marked preference for a decrease in solar radiation, indicating a desire for reduced sun exposure. The LG shows a preference for maintaining the current level of solar radiation, reflecting satisfaction with the present sun exposure levels. The LS and CS also show a preference for lower solar radiation, consistent with the high solar intensity characterizing these locations.

Figure 5C illustrates that in the ES, the majority of respondents anticipate a decrease in relative humidity, indicating a prevalent expectation for reduced humidity levels. In the LG, the predominant expectation is for stable humidity levels, reflecting a general anticipation to preserve the current humidity. In the LS and CS, a greater number of individuals expect a decrease in humidity, suggesting a collective aspiration for lower humidity levels.

Figure 5D presents that within the ES region, a predominant segment of the population expects an augmentation or constancy in Va, with a minimal fraction anticipating a reduction. In the LG, a substantial majority anticipates an increase in wind speed, reflecting widespread anticipation for increased wind velocity. In the LS and CS, a larger proportion of the populace expects an increase in wind speed, reflecting a collective expectation for heightened wind velocity.

The analysis reveals distinct preferences for meteorological parameters across different regions. Specifically, there is a pronounced inclination for the reduction of Ta, G, and RH in the ES, LS, and CS regions, contrasted with a predilection for preserving existing conditions in LG. A unanimous expectation for an increase in Va is observed across all regions. The strategic reduction of Ta coupled with an elevation of Va is aimed at promoting sweat evaporation and improving convective heat exchange with the surrounding environment, thus contributing to the reduction of thermal discomfort.

Environmental parameters with the highest influencing factors

Figure 6 indicates that respondents in the LS region identified temperature as the most crucial factor, followed by Va and G. This indicates a tendency towards higher Ta in LS, with Va and G also exerting considerable influence. Furthermore, variations in G could significantly affect residents’ thermal perception. The absence of shading from buildings on one side, combined with the dominance of hard paving, minimal water, and sparse hedges in the lower storefront, likely contributes to the lower comfort levels in this area. These elevated temperatures may result from increased thermal radiation associated with the extensive hard-paved surfaces.

Fig. 6
figure 6

Environmental parameters with the highest influencing factors.

Full size image

CS’s environmental dynamics are primarily dominated by Ta, with G as a secondary influence, underscoring the prominence of high Ta and the consequential effects of G. The area is shaded on both sides, but the shading is not complete due to the layout and orientation of the buildings. This allows for some direct solar radiation to reach the ground, especially during certain times of the day. Additionally, the hard paving and other reflective surfaces in the area contribute to the overall solar exposure. The limited vegetation in the CS area also means that there is less natural shading to mitigate the effects of solar radiation, thereby contributing to the observed low comfort levels.

Respondents in the ES region predominantly identified Ta and G as the most significant factors. This can likely be explained by the absence of surrounding buildings, the extensive subsurface water, limited shading from street trees, and the high traffic volume in the area.

In the LG region, the equitable distribution of preferences for Va, RH, G, and Ta indicates a well-balanced equilibrium of environmental factors. This equilibrium is likely due to the area’s abundant vegetation, grass-covered terrain with minimal water presence, and surrounding buildings that offer shade, thereby enhancing the overall environmental comfort.

In summation, the determinant environmental factors across each region are uniquely characterized, with ES and CS being predominantly influenced by temperature and solar irradiation. LS follows in impact, with LG experiencing the least effect. These differences are likely due to variations in building layouts, underlayment materials, and sky view factors (SVF) in each area. Shading from buildings and vegetation tends to reduce temperatures, whereas extensive hard surfaces in the underlayment can increase temperatures, contributing to thermal discomfort.

Space use and attendance

To elucidate the impact of varying thermal environments across campus open spaces on student activities, an hourly log of student presence and their engaged activities was maintained during the observation period. This meticulous recording facilitated the construction of a spatiotemporal activity participation map, illustrated in Fig. 7. Within the ES zone, ambulatory activities predominated, whereas literary activities such as reading registered the lowest engagement. A discernible decline in student presence was recorded between 11:00 and 14:00, a window potentially coinciding with academic schedules or meal times, thereby attenuating visible activity patterns. In contrast, the LS vicinity, situated near dining amenities, exhibited a scant engagement in physical activities and a more uniform distribution of dining-related activities. The CS area, benefiting from its proximity to dormitories and dining halls, displayed an equitable distribution of activity participation, indicative of constant student traffic.

Fig. 7
figure 7

Space usage and attendance.

Full size image

Predominantly, ambulation was the most embraced activity across all locales. The interval between 8:00 and 10:00 saw an inclination towards stationary activities like sitting or standing, likely favored due to the morning’s cooler conditions, facilitating leisurely pursuits. The midday hours of 11:00–14:00 showcased a contraction in the diversity of activities, aligning with academic or nutritional engagements. Ultimately, the examination underscores the temporal and spatial variability in space usage and student engagement, profoundly shaped by thermal environments and the selected activity modalities, with students adapting their activity preferences to the prevailing environmental dynamics.

Outdoor thermal benchmarks

NPET signifies the average vote count for thermal sensation prediction, and NPETR encapsulates the thermal comfort evaluation of an environment, amalgamating NPET with a thermal sensation grading assessment.

Figure 8 delineates a correlation between PET and MTSV, employing a linear model to ascertain the TSV’s weighted average per 1 °C PET increment. The relationship between PET (Physiologically Equivalent Temperature) and MTSV (Mean Thermal Sensation Vote) was analyzed for different campus locations using linear regression models. The sensitivity coefficient (represented by slope a in the equation y = ax + b) varies across these locations, reflecting differences in how environmental conditions and building layouts influence thermal sensation. Table 5 summarizes the fitted equations and R2 values.

Fig. 8
figure 8figure 8

Correlation diagram between PET and MTSV.

Full size image
Table 5 Fitted equations and R2 values for PET and MTSV across campus locations.
Full size table

ES: The sensitivity coefficient of 0.124 indicates that thermal sensation in this area increases significantly with rising PET values. The relatively high sensitivity suggests that the open layout and minimal shading at ES make it more susceptible to temperature changes, causing a pronounced impact on thermal comfort.

LG: With a sensitivity coefficient of 0.0924, the relationship between PET and MTSV is less pronounced compared to ES. This lower sensitivity may be attributed to the dense vegetation and effective air circulation in LG, which help moderate the impact of temperature variations on thermal sensation.

LS: The sensitivity coefficient of 0.0992 indicates a strong response of thermal sensation to changes in PET, similar to ES but slightly lower. This suggests that while the environment at LS is influenced by temperature, factors such as high pedestrian flow and mixed-use activities may buffer or amplify this effect, depending on specific conditions.

CS: The sensitivity coefficient of 0.1065 is slightly lower than ES but higher than LG and LS, indicating a moderate to high response to PET changes. The high R2 value of 0.9308 further supports a strong correlation between PET and MTSV, reflecting how the built environment and human activity at CS contribute to significant variations in thermal sensation.

These variations in sensitivity coefficients across different locations highlight the importance of considering specific environmental and architectural factors when assessing thermal comfort. The differences suggest that areas with less shading and more open space (like ES) are more sensitive to temperature changes, whereas areas with more vegetation and complex layouts (like LG) have a buffering effect, leading to lower sensitivity.

Table 6 elucidates that the ES region registers the minimum NPET at 17.33 °C, accompanied by an NPETR interval of 13.25–21.36 °C. This data implies that the ES locale experiences comparatively lower temperatures, leading to suboptimal thermal experiences for its residents. This thermal sensation could be further impacted by additional environmental variables, such as increased humidity levels and reduced wind velocities, which may alter inhabitants’ thermal comfort and preferences. In contrast, the LG area exhibits the apex NPET of 21.84 °C, with an NPETR delineating 16.43–27.25 °C, suggesting a warmer thermal ambiance and possibly more satisfactory thermal experiences. The LS region’s NPET is noted at 20.95 °C, positioning it intermediate between ES and LG concerning temperature, with an NPETR ranging from 15.91 to 25.99 °C. The CS region demonstrates an NPET of 18.04 °C and an NPETR of 11.62–24.69 °C, mirroring ES in terms of lower perceived temperatures and adverse thermal sensations among its denizens.

Table 6 NPET and NPETR in various locations.
Full size table

To further explore the complex relationships between the microclimatic parameters and TSV, we conducted a stepwise regression analysis. The stepwise regression was performed using both forward and backward selection methods, with the Akaike Information Criterion (AIC) as the model selection criterion.

The stepwise regression results are shown in Table 7.

Table 7 Stepwise regression results.
Full size table

According to the stepwise regression results, Ta and Tg are the most important environmental parameters influencing the TSV at the four locations, which is consistent with the conclusions in Sect. 3.1.4. Meanwhile, the stepwise regression results suggest that:

For the ES, Va and RH should be increased, while Ta and G should be reduced. For the LG, Va should be increased, while G and Ta should be reduced. For the LS, Va and RH should be increased, while Ta and G should be reduced. For the CS, Va should be increased, while G and Ta should be reduced.

Thermal adaptations

Figure 9 delineates the adaptive strategies deployed by individuals to mitigate heat stress across different locales. In ES, "drinking iced beverages" was cited by 22% of the survey participants as the foremost method for thermal comfort adjustment, with "entering air-conditioned interiors" and “minimizing clothing” being preferred by 17% and 16% of respondents, respectively, as mechanisms for coping with the heat. Notably, only a small fraction (3% and 4%) opted for "utilizing architectural shade," indicative of a potential scarcity of shaded structures or pavilions within ES.

Fig. 9
figure 9figure 9

Heat adaptation behavior.

Full size image

Within the LG region, the predominant cooling strategy, embraced by 24% of individuals, involved "retreating to air-conditioned spaces." Conversely, the tactics of "utilizing shaded buildings" and "curtailing outdoor activities" were each favored by merely 5% of the populace.

For LS and CS, “attenuating clothing” was the principal method for heat adaptation, garnering 16% and 18% of participant preferences, respectively, closely followed by the choice of "seeking air-conditioned refuge," which accounted for 14% and 15%. Strategies such as "seeking architectural shade," "limiting outdoor ventures," and "finding solace under trees" were comparatively less favored.

Optimal designs for open spaces

Studies have shown that residents’ motivation for activities in open spaces is related to their thermal perception, usage patterns, and usage satisfaction22,23. Residents with different activity expectations may also have different expectations of the thermal environment of the space24. Therefore, providing a variety of microclimates allows residents to adapt to the current thermal environment, meeting their preferences and expectations. To meet the thermal comfort needs of campus students, landscape design should embody a series of physical and physiological effects to accommodate students’ outdoor activity needs.

This study’s outcomes suggest a significant correlation between the patterns of use in campus open spaces and the thermal perception, usage dynamics, and satisfaction of the respondents. Enhancing the microclimatic conditions of campus environments is thus identified as a crucial measure to enable students’ acclimation to the thermal milieu.

Diminishing G and Ta substantially increases engagement with open spaces during peak heat periods. Across the four evaluated spaces, a higher incidence of G was observed, with the CS area experiencing increased levels of RH and Ta. Analysis of the TSV from respondents indicates a collective inclination towards reduced Ta and augmented Va within these areas.

Activities within these open spaces predominantly encompass group or public interactions, including leisure, physical exercise, and socializing. Individuals taking respite in these settings demonstrate a higher acceptance of thermal conditions. Additionally, participants in outdoor sports activities, motivated by a desire to sustain physical health, report superior thermal comfort. Accordingly, the design of campus landscapes should extend beyond scenic considerations to include spaces tailored for relaxation, social interaction, and physical exercise, as delineated in Table 8 and Fig. 10.

Table 8 Optimal design strategies for campus open spaces.
Full size table
Fig. 10
figure 10

Optimized design of campus open space.

Full size image

Discussion

Main factors affecting students’ thermal sensation in open spaces on different campuses

The outcomes of this study delineate variations in the determinants of students’ thermal perceptions across four unique campus open spaces, likely linked to diverse attributes such as vegetative coverage, spatial geometries enclosed by adjacent structures, the aspect ratios of these buildings, pavement material types, the presence and characteristics of water bodies, and the sky view factor (SVF)25,26. Yet, overarching these specific factors, ambient temperature is identified as the predominant influencer of students’ thermal sensation within this analysis, corroborating the consensus across multiple studies that temperature is the critical component affecting outdoor thermal comfort27,28,29. After temperature, the influences of solar radiation and wind velocity are also pronounced. Investigations in Chongqing have underscored the significance of solar radiation and air temperature in thermal perception30, while findings from Changsha, Hunan, China, emphasize the primacy of outdoor air temperature followed by solar radiation in impacting outdoor thermal sensation31. In contrast, a study assessing outdoor thermal comfort among residents of Nanjing ranked the impact of various factors from solar radiation, atmospheric pressure, peak temperature, and wind speed, to relative humidity. This disparity is ascribed to the consistent climatic conditions during the summer sampling period in Nanjing and the variable weather throughout the experiment, introducing atmospheric pressure as a notable factor. Nevertheless, in broader analyses incorporating extensive data sets, atmospheric pressure does not emerge as a principal factor.

Attendance pattern

The findings from this investigation affirm the profound impact of microclimatic conditions on student occupancy within campus open spaces. The escalation of Ta is directly linked to a significant reduction in the patronage of these areas, a trend that resonates with the outcomes of preceding research focused on open space attendance20,32,33. The study further highlights the critical role played by the architectural layout, the extent of shaded areas, and the inherent functionality of these spaces in attracting student usage. Notably, areas such as LG and LS, which boast a myriad of spatial features including water elements, shade-giving pavilions, and parasols, register higher attendance figures. This preference persists despite slightly less conducive microclimate conditions relative to LG, a discrepancy ascribed to the scarcity of activity and leisure facilities within the LG space.

Thermal comfort metrics in outdoor spaces

NPET

Our investigation across four distinct open spaces on a university campus demonstrated a range of NPET, with the ES at 17.3 °C, the LG at 21.8 °C, the LS at 21.0 °C, and the CS at 18.0 °C. This diversity signifies that students’ thermal comfort levels vary significantly, even within a confined geographic locale. Notably, ES exhibited the lowest thermal tolerance, while LG displayed the highest. The architectural and environmental configuration of ES, marked by its expansive openness and considerable paved surfaces, coupled with static water bodies that enhance solar reflectance and augment humidity, diminishes thermal tolerance. In contrast, the structural arrangement of LG, with buildings on either side facilitating ventilation and a substantial presence of vegetative cover providing shade, fosters a higher thermal tolerance.

Comparative analysis with other thermal comfort assessments in Xi’an across varied spatial contexts, such as residential zones and different campus areas, revealed NPET figures by Zhang et al.34 and Mi et al.35 as 19.4 °C and 19.7 °C, respectively. These figures suggest a relative consistency in thermal comfort perceptions, attributable to similar participant demographics and experimental timings. Contrarily, Niu, et al.36 summer-exclusive study in educational open spaces reported a higher NPET of 26.1 °C. As shown in Table 9, these findings, consistent with our campus exploration, illustrate the pronounced variability in NPET across disparate settings and investigative conditions. Furthermore, Ma et al.37 examination of the elderly in urban parks indicated a lower NPET for this demographic compared to a mixed-age cohort, possibly due to the survey’s temporal limitations and the elderly’s diminished thermal sensitivity and enhanced cold endurance. These outcomes emphasize the critical role of environmental elements in outdoor design considerations, as delineated by38.

Table 9 Thermal benchmarks in the study where the location is Xi’an.
Full size table

A multitude of investigations has leveraged portable experimental apparatuses for the collection of meteorological observations, evidencing the differential local microclimates across disparate locales. In the ambit of the Urban Senses initiative, Gallinelli et al. engineered a microclimate observation backpack, designed to assess air quality, acoustic pollution, radiative flux, ambient temperature, atmospheric moisture, and aerodynamic velocity, in addition to acquiring hemispherical imagery for Sky View Factor (SVF) estimation. Their research delineated the dynamic nature of the Physiological Equivalent Temperature (PET) across temporal and spatial dimensions, pinpointing urban form and structure as the principal determinants of locational discrepancies39. Concurrently, Speak et al.'s exploration in Florence, through the utilization of mobile sensing equipment, undertook a comprehensive assessment across a variety of urban open spaces, including gardens, parks, and plazas. Their findings underscored the significant role of arboreal canopy, urban canyon phenomena, and the architectural proportionality of height to width in modulating thermal comfort across these environments40.

Thermal acceptable range

The overall TAR of the four-campus open spaces in this study is 20–38 (°C). The warm season in Tel Aviv, Israel is similar to Xi’an41, and its TAR is 19–26 (°C). Tel Aviv, Israel The lower limit of the TAR (19.0 °C) is similar to the lower limit of this study (20 °C). This may be because although the experiment in Israel is year-round, it has a Mediterranean climate with relatively warm winters, and its year-round research is different from that of this study. The range of climate change in the period is similar, so the lower bounds are close. The upper limit of TAR in Tel Aviv, Israel (26.0 °C) is very different from the lower limit of this study (38 °C). This is because the highest summer temperature in Tel Aviv is about 31 °C, while the highest summer temperature in Xi’an can even reach over 40 °C. Therefore, students in this study have stronger heat tolerance and a higher upper limit of thermal acceptability. The difference between the TAR of the four open spaces in this study may be due to the different shade areas, ventilation conditions, and cooling effect of the water feature in each environment.

Conclusions

This study investigated the thermal comfort of college students in four open spaces on campus in Xi’an, China, using meteorological measurements, questionnaires, and activity logs. The main findings are as follows:

  1. (1)

    Key Factors Influencing Thermal Perception: Ta and Tg were identified as the primary meteorological determinants of students’ thermal sensation, with Va and RH playing secondary roles. These factors varied across different open spaces, indicating distinct thermal experiences.

  2. (2)

    Thermal Comfort Baseline: The NPET and NPETR were established for each open space. For example, the NPET at the ES was 17.3 °C, with a range of 13.3–21.4 °C, while at the LG, it was 21.8 °C, with a range of 16.4–27.3 °C.

  3. (3)

    Impact of Microclimate on Space Utilization: The utilization of campus open spaces by students was significantly influenced by the outdoor microclimate, available facilities, and spatial configurations. This led to diverse behavioral adaptations for thermal regulation, such as seeking shade or reducing outdoor activities during hot periods.

  4. (4)

    Optimal Design Strategies: Based on the findings, an optimal design strategy for campus open spaces was proposed. This includes increasing vegetation cover, enhancing natural ventilation, and providing shading structures to improve thermal comfort. The goal is to create more inviting and comfortable outdoor spaces that encourage student interaction and engagement.

The study highlights the importance of considering microclimate conditions in the design and planning of urban open spaces. The findings provide valuable insights for creating more comfortable and sustainable environments in cities facing similar challenges of urban heat islands and climate change. By improving thermal comfort, these strategies can enhance the quality of life for urban residents and promote outdoor activities, contributing to public health and well-being. Future research could explore the application of these strategies in different urban contexts and further refine the design guidelines for thermal comfort in outdoor spaces.

Limitations and future direction

We propose several extensions to our study to enhance the comprehensiveness of subsequent research endeavors. First, given Xi’an’s distinctive climate—characterized by warm and moist summers, frigid and arid winters, and prolonged temperate interseasonal periods—a comprehensive examination of outdoor thermal comfort across all seasons is warranted. Expanding the temporal scope of data collection will allow for a more thorough understanding of how thermal comfort varies throughout the year, particularly about seasonal extremes. Second, our investigation currently focuses on student populations within university settings. To develop a more holistic understanding of thermal comfort, future studies should consider including a broader demographic range, encompassing different age groups and potentially different educational institutions or public spaces. This expansion would help in generalizing the findings to a wider population and diverse environments. Third, while our design recommendations offer a conceptual foundation aimed at ameliorating existing conditions, they are not accompanied by a detailed practical or cost–benefit analysis. We recognize that such an analysis is essential for assessing the feasibility and economic viability of the proposed strategies. Future research should incorporate these practical considerations to ensure that the design recommendations are both effective and sustainable in real-world applications. Fourth, our analysis primarily focused on environmental and behavioral factors such as temperature, humidity, wind speed, and student adaptation strategies like seeking shade or adjusting clothing. However, we did not delve into the specific effects of these adaptation strategies or the underlying psychological mechanisms that drive these behaviors. Exploring psychological factors, such as thermal perception, cognitive appraisal, and the process of thermal adaptation, would require additional physiological experiments and a more extensive research framework. While these aspects are crucial for a comprehensive understanding of how students interact with their thermal environment, they extend beyond the scope of the current study. Given the already substantial amount of data and analysis included in this manuscript, incorporating these additional dimensions would have risked making the content overly dense and potentially detracting from the primary focus of our research. Therefore, while we recognize the importance of these factors, they have been identified as areas for future research. In subsequent studies, we plan to explore the psychological and physiological dimensions of thermal comfort to provide a more holistic understanding of how students adapt to outdoor thermal environments. Finally, although we propose several design strategies for optimizing campus open spaces to enhance thermal comfort, these strategies have not been subjected to quantitative simulations or rigorous empirical testing. The next step involves conducting detailed quantitative simulations to rigorously verify the effectiveness and specificity of these proposed design strategies. Furthermore, practical and cost–benefit analyses of these strategies are necessary to ensure their feasibility and sustainability in real-world settings. Future research should aim to bridge the gap between theoretical design recommendations and practical implementation by providing comprehensive cost–benefit analyses that can guide decision-making in campus planning and development.