Effects of Tree Height and Spatial Layout on Thermal Comfort in a Residential Area Based on ENVI-Met: A Case Study of a Typical Hot Summer Day in Qingdao

Abstract

In coastal residential areas, the combined effects of high temperature, high humidity, and weak wind conditions during summer intensify outdoor heat exposure and reduce pedestrian thermal comfort. To investigate the influence mechanisms of tree height and spatial layout on pedestrian-level thermal comfort, this study selected a typical residential community in Chengyang District, Qingdao, as the research site. Based on field meteorological observations, an ENVI-met model was established and validated. Using the existing composite greening scenario as the baseline, three tree layout types (row, cluster, and free layouts) and four height scenarios (4 m, 6 m, 8 m, and 10 m) were configured to quantitatively compare variations in physiological equivalent temperature (PET) under different planting schemes. The results indicate that tree configuration significantly affects summer thermal comfort. Its regulatory mechanism is governed not only by air temperature reduction but also by shortwave radiation interception, longwave radiation accumulation, and shading continuity. Although low-to-medium height trees can reduce local air temperature through transpiration, their limited canopy height and shading continuity restrict their ability to effectively attenuate direct shortwave radiation at pedestrian level, and in some cases may even increase mean radiant temperature (Tmrt) and PET. In contrast, 10 m tall trees arranged in row and cluster layouts can form continuous shaded cores, with the 10 m cluster layout demonstrating the best overall performance by significantly reducing Tmrt and PET. The free layout, characterized by dispersed canopies and fragmented shading, provides relatively limited thermal comfort improvement. The findings suggest that residential greening optimization should strengthen the coordination between tree height, canopy structure, and activity spaces. Tall trees should be prioritized in children’s play areas, elderly resting areas, residential entrances, main pedestrian pathways, and west-facing sun-exposed zones, while integrating building shadows and road orientation to create a continuous yet not overly enclosed shading network, thereby enhancing summer thermal adaptability in residential areas.

1. Introduction

In recent years, global climate change and rapid urbanization have exacerbated the Urban Heat Island (UHI) effect, leading to frequent extreme summer heat events that severely threaten the outdoor thermal comfort and public health of urban residents [1,2]. Urban design plays a significant role in regulating microclimates. Among design-oriented urban cooling solutions, vegetation has emerged as a prominent strategy due to its capacity to mitigate thermal impacts through evapotranspiration, solar radiation shading, and the improvement of surface radiation and heat exchange processes [3,4]. Consequently, vegetation is widely regarded as a vital approach for enhancing pedestrian thermal comfort, and urban greening has become a key technical measure for mitigating the UHI effect. In residential settings, optimizing tree height and spatial layout can further enhance this cooling effect [5]. Wong et al. [6] identified three key mechanisms for vegetation-induced cooling: shading (interception of solar radiation), evapotranspiration, and albedo enhancement. Similarly, Hwang et al. [7] demonstrated that shading is a crucial factor influencing the urban thermal environment and long-term outdoor thermal comfort. However, excessively dense planting can obstruct airflow and reduce wind speed, thereby diminishing overall cooling benefits [8]. Therefore, rationally configuring vegetation to maximize cooling effects in residential environments is of critical importance. Furthermore, tree height and spatial layout patterns jointly determine the microclimate regulation potential of residential areas [9]. The three-dimensional characteristics of trees (e.g., trunk height) dictate their capacity to intercept both shortwave and longwave radiation; notably, the canopies of tall trees can significantly reduce the mean radiant temperature (Tmrt) and physiological equivalent temperature (PET) at the pedestrian level [10]. Cai et al. [11] quantified the vertical gradient cooling effects of different canopy morphologies, revealing the significant impact of canopy structure on air temperature distribution and thermal comfort. Focusing on small-scale planting patterns, Abdi et al. [12] revealed how different tree layouts vary in regulating wind speed and local heat distribution. However, such studies predominantly focus on the vertical morphological characteristics of individual trees, paying less attention to the spatial organization of trees within actual residential settings. Furthermore, the tree height and canopy structure parameters in their research scenarios remained relatively fixed, leaving the coupling mechanism between height variations and layout patterns largely unexplored.

Currently, numerical simulation has become a prevalent approach for evaluating urban microclimates. Compared to Computational Fluid Dynamics (CFD) software, which primarily focuses on wind environment simulation, the ENVI-met model can comprehensively couple the shortwave and longwave radiation exchange processes among buildings, vegetation, surfaces, and the atmosphere. Therefore, it possesses significant advantages in simulating thermal environments and thermal comfort [13,14,15], and has been validated as highly applicable for simulating green space scenarios in residential areas [16]. Existing studies have either focused on the morphological and height effects of individual trees, or paid attention to the impacts of planar layout patterns on microclimates. However, there is a lack of systematic quantitative comparison regarding the interactive effects between tree height and spatial layout. Particularly in the context of high-density coastal residential areas, the comprehensive impact of the synergistic effects of different tree heights and various layouts on the radiation environment and human thermal comfort needs to be further explored.

This study takes the Yilufa Sunshine Scenic Garden community in Chengyang District, Qingdao, as the research site for conducting field surveys. Using the microclimate simulation software ENVI-met 5.8.0, this study simulates the microclimate environment and human outdoor thermal comfort under various vegetation configurations. The research focuses on exploring the comprehensive effects of three tree layout patterns—row, cluster, and free-form—along with four tree heights (4 m, 6 m, 8 m, and 10 m) on the microclimate and thermal comfort. Based on the simulated microclimate parameters, including air temperature, relative humidity, wind speed, Mean Radiant Temperature (Tmrt), and Physiological Equivalent Temperature (PET), the summer thermal comfort of the residential area is evaluated. Finally, by comparing the simulation results, universal conclusions are synthesized to provide support for optimizing the thermal environment in residential areas. This research can offer valuable insights for the planning and optimization of residential green spaces, providing a scientific basis for achieving a comprehensive balance between shading-induced cooling and ventilation-driven heat dissipation.

2. Materials and Methods

2.1. Study Area

Qingdao is located in the southern part of the Shandong Peninsula, bordering the Yellow Sea to the east and south. It serves as a crucial port city and a major land–sea transportation hub along China’s eastern coast. The region falls within a temperate monsoon climate zone and, regulated by oceanic circulation, exhibits prominent maritime characteristics with four distinct seasons [17]. This study selected the Yilufa Sunshine Scenic Garden residential community in Chengyang District, Qingdao, as the field measurement site. Constructed in 2011, the community covers a total area of approximately 55,549 m² and consists of eight 18-meter-tall and five 33-meter-tall residential buildings. The overall architectural layout adopts a row configuration. The greening configuration primarily utilizes a composite tree-shrub-grass pattern, encompassing three core vegetation types: trees, shrubs, and groundcovers. Trees mainly include Yulan magnolia, Dragon juniper, Chinese tulip tree, midget crabapple, pomegranate, purple-leaf plum, and persimmon; shrubs mainly consist of littleleaf boxwood, Japanese spindle, Weigela, and convex-leaf Japanese holly; and groundcovers are predominantly Kentucky bluegrass and ryegrass.

2.2. Field Measurement Scheme

In this study, a WX-BXQX5 compact portable weather station (manufactured by Xiang Environmental Co., Ltd., Weifang, China) was utilized to acquire the background meteorological conditions of the study area, including air temperature, relative humidity, wind speed, and wind direction. Simultaneously, Kestrel 5500 handheld weather meters (Nielsen-Kellerman Co., Boothwyn, PA, USA) were used to conduct field measurements of air temperature and relative humidity at the pedestrian level. The height of the measurement points was set at 1.4 m above the ground, closely corresponding to the typical height of human activity, which effectively reflects the characteristics of the outdoor pedestrian thermal environment in the residential area. Previous microclimate studies in residential areas or urban green spaces typically deploy measurement points based on the representativeness of spatial typologies, such as shaded areas under trees, open paved areas, street canyons, and vegetation-covered zones [9,18]. Taking into account the spatial heterogeneity of building layouts and underlying surface types within the study area, as well as typical residential spatial typologies [19], a total of four measurement points were established: Point A is located in a partially shaded space jointly influenced by buildings and vegetation; Point B is situated in a vehicular road and inter-building pedestrian passage; Point C is positioned in a densely vegetated, heavily shaded area; and Point D is set in a relatively open, paved area with minimal shading (Figure 1).

Field measurements were conducted from 7:00 to 19:00 on 26 July 2025. Continuous hourly observation data were recorded at each point, and the average values during stable observation periods were taken as the measured results for the corresponding hours [20]. 26 July was selected as the reference day for model validation and baseline simulation. This selection was primarily based on the fact that the maximum air temperature on that day reached 35.2 °C, which exceeds the threshold for a “high-temperature day”, as defined by standard Chinese meteorological criteria. Furthermore, observations indicated that the local background meteorological conditions remained highly stable from 25–27 July. The consistency of these weather conditions ensures that the single-day observational data is highly representative of typical summer extreme heat characteristics. This approach also aligns with previous methodologies used to investigate thermal environment evolution under specific climatic conditions [21,22]. The selected day exhibited the typical hot-humid and weak-wind characteristics of coastal cities (relative humidity ranging from 52.8% to 83.8%, and an average wind speed of 1.6 m/s), thus serving as an ideal scenario for evaluating the regulatory effects of tree configuration on the thermal environment.

2.3. ENVI-Met Microclimate Simulation

2.3.1. Study Area Model Construction and Parameter Setting

A simulation area model with 140 × 103 × 33 grids and a resolution of 2 m × 2 m × 2 m was built in ENVI-met (Figure 2). At the same time, five nested grids were set [17] to improve the accuracy of on-site environment simulation.

Based on the plant height and canopy spread characteristics obtained through field surveys, and referencing empirical values from existing ENVI-met simulation studies [23], parameterization was conducted for representative plants across different height classes and life forms. To ensure the validity of the 4 m and 6 m evergreen tree configurations in the model, they were designated as small Dragon juniper (4 m) and Dragon juniper (6 m), respectively. The specific vertical distribution of their Leaf Area Density (LAD) and geometric characteristics are detailed in

Table 1. Field-surveyed dominant species and corresponding representative height classes.

Plant Species	Scientific Name	Plant Type	Height (m)	Crown Spread (m)	LAD (m2/m3)
Kentucky bluegrass	Poa pratensis L.	Grass	0.2	-	0.30, 0.30, 0.30, 0.30, 0.30, 0.30, 0.30, 0.30, 0.30, 0.30
Littleleaf boxwood	Buxus sinica var. parvifolia M. Cheng	Shrub	1	-	0.60, 0.60, 0.15, 0.32, 0.48, 0.69, 0.98, 1.21, 0.78, 0
Small Dragon juniper	Juniperus chinensis ‘Kaizuka’	Low tree	4	1.5–3	0.75, 0.75, 0.75, 0.36, 0.84, 0.92, 1.23, 1.21, 0.78, 0
Dragon juniper	Juniperus chinensis ‘Kaizuka’	Medium tree	6	2–3	0.75, 0.75, 0.75, 0.36, 0.84, 0.92, 1.23, 1.21, 0.78, 0
Yulan magnolia	Magnolia denudata Desr.	Tall tree	8	4–8	0.075, 0.075, 0.075, 0.075, 0.25, 1.15, 1.06, 1.05, 0.92, 0
Chinese tulip tree	Liriodendron chinense	Large tree	10	8–10	0.04, 0.04, 0.07, 0.11, 1.10, 1.10, 1.10, 1.10, 0.10, 0

.

The simulation boundary conditions for this study are detailed in

Table 2. Boundary condition settings for ENVI-met 5.8.0 simulation.

Parameter Category	Parameter	Input Value
Geographic parameters	Location	Qingdao, China
	Coordinates	36°31′ N, 120°40′ E
Time and date	Start date	26 July 2025
	Start time	4:00
	Total simulation time	14 h
Meteorological conditions	Max/Min temperature	35.2/27.8 °C
	Max/Min relative humidity	83.8%/52.8%
	Wind direction	197.57°
	Wind speed at 10 m	1.60 m/s
Surface materials	Brick road (KK)	Roughness: 0.01, Albedo: 0.3, Emissivity: 0.9
	Asphalt road (ST)	Roughness: 0.01, Albedo: 0.12, Emissivity: 0.9
	Loam soil (LO)	Roughness: 0.015, Albedo: 0, Emissivity: 0.9
	Grey concrete (PG)	Roughness: 0.01, Albedo: 0.3, Emissivity: 0.9
Initial soil conditions	0–20 cm depth	20 °C/65%
	20–50 cm depth	20 °C/70%
	50–200 cm depth	19 °C/75%
	Below 200 cm	18 °C/75%

. The wind speed at 10 m above ground level was calculated based on the wind profile equation [24], while parameters at the 2500 m upper boundary, such as atmospheric specific humidity and cloud cover, were set to the software’s default values [25]. PET was selected as the outdoor thermal comfort index for this study. It is defined as the air temperature in a typical indoor setting that maintains the human core and skin temperatures in equilibrium with the human energy balance experienced under complex outdoor conditions [26]. Within the ENVI-met software, the Bio-met module is utilized to assess thermal comfort, calculating PET values using the simulated wind speed, air temperature, and humidity, alongside specific personal parameters [27].

2.3.2. Model Evaluation

According to the method of reference [17], the accuracy of the model is evaluated by using root mean square error (RMSE). The specific formula is as follows:

$$RMSE=\sqrt{{\displaystyle \frac{1}{n}}{\displaystyle \sum _{i=1}^n}{\left(y_i^{\prime }-y_i\right)}^2}$$

(1)

$y_i$ is the observation value, $y_i^{\prime }$ is the model simulation value, and n is the number of observations.

2.4. Simulation Scenarios

To exclude the confounding effects of total greenery volume on the thermal environment, this study implemented a strict controlled-variable approach for all optimization scenarios. The green space ratio of the original field site (baseline) was approximately 30%. Accordingly, the green space ratio for all optimization scenarios was uniformly set to 30%, with a tree-to-shrub ratio of 1:2, thereby establishing a comparable vegetation base featuring a three-tiered tree-shrub-grass structure. Based on common greening configuration patterns in residential areas and the findings from the field survey (Figure 3), the spatial layout of trees was conceptualized into three typologies: row, cluster, and free. The specific arrangements are as follows: In the row layout, trees were planted in rows along the east–west axis with a 6 m spacing, while shrubs were planted in strips or patches between the tree rows and along the periphery. In the cluster layout, three irregular planting clusters were established according to the site’s spatial characteristics (each cluster covering an area of approximately 4 m² with a 6 m spacing between clusters), where trees and shrubs were densely grouped within these patches. In the free layout, the plants were randomly distributed across the site, deliberately avoiding paved roads and utility infrastructure areas.

Vegetation configuration requires a comprehensive consideration of multiple factors. Drawing upon field surveys and a literature review, this study analyzed thermal comfort with varying tree heights [28]. Considering that excessively tall trees can negatively impact daylight access for buildings [29], the tree heights were set at 4 m, 6 m, 8 m, and 10 m, while keeping other parameters, such as the root zone area and foliation period, constant. The subsequent optimization scenarios are denoted by abbreviations (e.g., C10 refers to the cluster layout with 10 m trees; R10 refers to the row layout with 10 m trees; and F10 represents the free layout with 10 m trees) (Figure 4). The specific plant positions for the free layout are detailed in Table S1. Notably, the reference baseline established in this study represents the actual site conditions, retaining the neighborhood’s original composite tree-shrub-grass greenery. These simulations aim to explore enhancement strategies building upon the existing green infrastructure, specifically to verify whether the proposed optimization scenarios can deliver superior thermal comfort regulation compared to the current conditions.

2.5. Statistical Analysis

This study utilized the variations (ΔTa, ΔTmrt, and ΔPET) of each optimization scenario relative to the baseline as the analytical variables to examine whether different tree configuration schemes have a statistically significant impact on the outdoor thermal environment. The Δ values were defined as the differences between the simulated values of the optimization scenarios and those of the existing baseline. A negative value indicates a reduction in the corresponding thermal environment indicator, whereas a positive value represents an increase.

One-way Analysis of Variance (ANOVA) was employed to test the overall statistical differences among the twelve tree configuration schemes. When significant differences were detected (p < 0.05), Tukey’s Honestly Significant Difference (HSD) post hoc test was conducted to determine pairwise differences between scenarios [30]. All statistical analyses were performed based on data extracted at a pedestrian height of 1.4 m during the period from 09:00 to 18:00.

3. Results

3.1. Model Accuracy Verification

To verify the simulation accuracy of ENVI-met model for microclimates in a coastal residential area in Qingdao, this study compared the simulated and measured data of typical sunny and hot days in summer (9:00–18:00) (Figure 5). In complex urban microclimate simulations, root mean square error (RMSE) is the key index for measuring the absolute accuracy of the model [27]. The results showed that the RMSE of Ta was 1.75 °C, and the RMSE of RH was 5.32%. Among them, the temperature error is lower than the high-precision standard of 2.0~2.5 °C, and the humidity error is also within the acceptable range of 10% [13], indicating that the model has achieved high accuracy in the core indicators, which is enough to support the subsequent in-depth evaluation of PET of different greening layout schemes.

3.2. Thermal Comfort Responses to Different Tree Configuration Schemes

3.2.1. Results of Significance Testing

To ascertain the magnitude of the impact of different tree configuration schemes on thermal environment indicators, Kruskal–Wallis non-parametric tests were performed on the microclimate variations across all observation periods (09:00–18:00). The results indicate that across all observation periods, the different tree configuration schemes exhibited highly significant statistical impacts on ΔTa, ΔTmrt, and ΔPET (p < 0.001). Detailed statistical test results for each hour are provided in Table S2. In terms of effect size (η²), the tree configuration schemes demonstrated the highest explanatory power for air temperature (η² ranging from 0.71 to 0.85, indicating a large effect). Conversely, the effect sizes for Tmrt and PET were relatively moderate (with η² ranging from 0.16 to 0.24 and 0.17 to 0.30, respectively). This further corroborates that while different spatial layouts can significantly alter air temperature, their effectiveness in improving overall human thermal comfort (PET) diverges significantly. This divergence is primarily driven by the spatial heterogeneity of the local wind environment and shading continuity.

3.2.2. Responses of ΔTa, ΔTmrt, and ΔPET Under Various Schemes During Typical Peak Heat Hours

As shown in

Table 3. Comparison of microclimate indicator differences among different tree configuration schemes at 14:00.

Scenario	ΔTa (Mean ± SD)/°C	ΔPET (Mean ± SD)/°C	ΔTmrt (Mean ± SD)/°C
R4	0.22 ± 0.23 f	0.72 ± 1.75 cde	2.17 ± 4.63 c
R6	0.14 ± 0.20 ef	0.43 ± 1.69 cd	1.10 ± 4.26 bc
R8	0.12 ± 0.19 e	0.22 ± 1.66 bc	0.46 ± 4.10 b
R10	−0.13 ± 0.21 cd	−1.21 ± 1.87 a	−3.06 ± 4.29 a
C4	−0.08 ± 0.23 d	1.27 ± 1.47 e	3.85 ± 3.87 d
C6	−0.19 ± 0.19 bc	0.79 ± 1.34 cde	2.34 ± 3.60 cd
C8	−0.24 ± 0.18 b	0.38 ± 1.39 cd	1.21 ± 3.73 bc
C10	−0.61 ± 0.26 a	−1.69 ± 2.21 a	−3.47 ± 5.09 a
F4	0.61 ± 0.20 i	1.21 ± 1.62 e	2.38 ± 4.50 cd
F6	0.55 ± 0.17 hi	0.97 ± 1.50 de	1.52 ± 4.04 bc
F8	0.53 ± 0.17 h	0.79 ± 1.40 cde	0.94 ± 3.77 bc
F10	0.33 ± 0.18 g	−0.33 ± 1.28 b	−1.95 ± 3.23 a

Data in the table represent the differences in indicators for each scenario relative to the baseline at 14:00 (Δ = scenario − baseline). Negative values indicate a reduction, while positive values indicate an increase. Data are presented as mean ± standard deviation (SD). Different superscript letters (a, b, c, etc.) denote the results of post hoc multiple comparisons using the Tukey HSD test (α = 0.05). Values not sharing a common letter differ significantly from each other, whereas those sharing a letter show no significant difference. Letters are assigned in ascending order of the mean values, with ‘a’ representing the lowest value (i.e., the optimal cooling effect).

, different tree configuration schemes had statistically significant impacts on microclimate indicators at 14:00 (Tukey HSD, p < 0.05). An analysis of the differences reveals that C10 achieved the maximum reductions in both ΔTa (−0.61 ± 0.26 °C) and ΔPET (−1.69 ± 2.21 °C), whereas low trees (≤6 m) exhibited negative effects on thermal comfort. Post hoc comparison results indicate that the free layout underperformed compared to the cluster and row layouts in terms of ΔTa. Furthermore, variations in tree height between 4 m and 8 m did not produce a distinct gradient change in the indicator differences. Overall, among the 12 tree configuration schemes, only C10, R10, and F10 yielded negative ΔPET values, while the remaining low- to medium-height tree schemes exacerbated heat stress to varying degrees. Notably, scenarios C4, C6, and C8 exhibited a phenomenon of decreased air temperature but increased PET, whereas F10 showed an increase in air temperature alongside a decrease in PET. Therefore, the subsequent sections will first analyze the spatial heterogeneity and daytime evolution patterns of C10, F10, and R10, followed by an in-depth discussion on the difference gradients and the specificities of certain scenarios.

3.3. Analysis of Spatial Heterogeneity in PET Improvement

Figure 6 compares the spatial distribution of ΔPET at the pedestrian level (1.4 m) at 14:00 for the R10, F10, and C10 scenarios. The improvements in thermal comfort across all three layouts manifested as localized effects rather than uniform site-wide enhancements, with areas of PET improvement and deterioration interspersed across the site. Notably, the areas with reduced PET in the F10 scenario were the most fragmented and discontinuous, resulting in an unstable overall heat mitigation performance. In contrast, the extent of ΔPET improvement in the C10 scenario expanded significantly, with a higher concentration of blue areas distributed directly beneath the clustered tree canopies. However, localized red patches remained in the C10 map, indicating that even in the scenario with the most pronounced overall improvement, thermal comfort effects are still heavily influenced by localized spatial conditions. Overall, the impact of tree configuration on thermal comfort exhibits strong spatial heterogeneity. The improvements in PET underscore the variations in thermal comfort driven by the continuity of shading.

3.4. Daytime Variations in Extreme Heat Stress

This study referenced an outdoor thermal comfort standard empirically calibrated based on a typical northern Chinese city (Tianjin) [31]. This standard defines the ranges of PET corresponding to residents’ Thermal Sensation (

Table 4. Evaluation scale of physiological equivalent temperature (PET) in Northern China.

PET (°C)	Thermal Sensation	Physiological Stress Level
11~24	Neutral	No thermal stress
24~31	Slightly warm	Slight heat stress
31~36	Warm	Moderate heat stress
36~46	Hot	Strong heat stress
>46	Very hot	Extreme heat stress

). Given that this study focuses on the heat mitigation efficacy of trees under high-temperature summer environments, the evaluation intervals from this standard were extracted to serve as the grading benchmark for assessing the capability of trees to alleviate heat stress.

In this study, areas with PET > 46 °C were defined as extreme heat stress zones, and the variations in their areal proportions from 09:00 to 18:00 were statistically analyzed (Figure 7). The results indicate that the areal proportions of extreme heat stress for all scenarios exhibited a daytime pattern characterized by an initial increase followed by a decrease, maintaining high levels from noon to early afternoon. The differences among the various tree configuration schemes primarily manifested in two aspects: first, whether the peak area of high heat stress is effectively reduced, and second, whether the extreme heat stress zones dissipate more rapidly in the afternoon. The areal proportions of extreme heat stress in the low-tree scenarios were higher than that of the baseline, indicating that their thermal comfort benefits were insufficient to offset the adverse changes in the local thermal environment. In contrast, the tall-canopy scenarios were more capable of decreasing the proportion of PET > 46 °C areas during the afternoon.

Comparing the layout differences among C10, R10, and F10 reveals that C10 achieved the most pronounced and efficient reduction in the extreme heat stress area, followed by R10, while F10 showed the smallest improvement. This demonstrates that the mitigation of extreme heat stress does not rely solely on the quantity or height of trees. Only when the tree canopies form stable shading that geometrically matches the residents’ activity spaces and the open spaces between buildings can the extreme heat stress zones be effectively reduced. Conversely, when the canopy distribution is misaligned with high-exposure areas, the heat stress mitigation effects may remain limited even if the tree height is increased.

3.5. Daytime Variations in PET Improvement

As shown in Figure 8, the three 10 m tree layout scenarios exhibit dynamic differences in ΔPET throughout the daytime period from 09:00 to 18:00. Among them, C10 demonstrates the most outstanding and stable daytime cooling performance. Particularly during the high-temperature and high-radiation period from 13:00 to 16:00, it achieves the maximum decrease in ΔPET, exhibiting a robust afternoon heat mitigation capability. R10 ranks second, with its ΔPET generally fluctuating around −1 °C throughout the daytime. F10 shows the lowest efficacy in improving thermal comfort, with its ΔPET hovering around 0 °C for most of the daytime. Overall, the concentrated canopies in C10 provide continuous and overlapping effective shading, yielding the most stable overall effect. In contrast, the fragmented canopy of F10 is susceptible to sunlight penetration (forming sunflecks) under oblique solar angles, failing to maintain stable radiation shielding throughout the daytime.

4. Discussion

4.1. Regulatory Mechanisms of Tree Spatial Layouts on Radiation and Thermal Comfort

Figure 9 reveals the profound impact of C10, R10, and F10 on the radiant flux density on the human body surface and Tmrt at 14:00. First, the interception of direct shortwave radiation (Direct SW) acts as the primary defense against extreme heat stress [32]. As shown in Figure 9, both C10 and R10 can significantly and equivalently attenuate the direct shortwave radiation reaching the pedestrian level. In contrast, due to scattered planting and fragmented canopies, F10 fails to form continuous shaded protection zones, resulting in a drastic decline in its direct shortwave interception capacity [33,34].

Radiant heat load is a critical factor governing pedestrian thermal comfort [35], and canopy shading can reduce radiation absorption on the human body surface by blocking direct shortwave radiation [36]. As illustrated in the figure, although R10 and C10 possess equivalent shortwave shading capabilities, C10 can effectively attenuate downward longwave radiation. F10 exhibited an increase in downward longwave radiation (+9 W·m⁻²), indicating that isolated tree canopies rapidly heat up after absorbing solar shortwave radiation, which subsequently causes the trunks and branches to warm up [37] and release substantial amounts of sensible heat into the surrounding environment [38]. This provides a robust explanation for the previously mentioned phenomenon in F10, where air temperature increases while PET decreases.

4.2. PET Responses to Low-to-Medium-Height Clustered Trees

As shown in Figure 10a, at 14:00, the direct shortwave radiation (Direct SW) for both the C4 and C6 scenarios increased. As previously mentioned in Table 3, the Ta for these two scenarios decreased by 0.08 °C and 0.19 °C, respectively, while both Tmrt and PET increased. The slight decrease in Ta is primarily attributed to the cooling effect of plant transpiration, which absorbs radiant energy and releases latent heat [32,39]. Conversely, the increases in Tmrt and PET are due to the fact that the 4 m and 6 m plants (small Dragon juniper and Dragon juniper), constrained by their canopy shapes and limited spread, fail to provide a sufficiently shaded environment. An increase in tree height can enhance the overall canopy volume and the vertical shading projection, slightly strengthening the trees’ ability to intercept radiation and dissipate latent heat through transpiration. Therefore, although C4 and C6 were assigned the same LAD, the cooling magnitude of C6 is slightly greater than that of C4, resulting in a relatively smaller increase in PET.

C8 similarly exhibited the phenomenon of decreased Ta alongside increased PET. Compared to C4 and C6, C8 exerted a better shading effect, achieving a reduction in direct shortwave radiation. The calculation of Tmrt accounts for all spatial directions, whereas direct radiation is primarily oriented vertically [40]. Consequently, the ΔTmrt for C8 remained positive, indicating that although it may have attenuated a portion of the direct shortwave radiation, it did not lead to an overall decrease in the comprehensive radiant heat load at the pedestrian level.

Furthermore, the increase in humidity caused by vegetation transpiration is an important contributing factor to the elevated PET in C4 and C6. Figure 10b shows that the relative humidity for C4, C6, and C8 was consistently higher than that of the baseline, indicating that clustered trees enhanced the water vapor input into the near-surface layer. Under high-temperature conditions, elevated humidity weakens the evaporative cooling efficiency of human sweat, thereby exacerbating thermal discomfort [41]. Given that the wind speed differences between C4, C6, C8, and the baseline in this study were all less than 0.02 m/s, the wind environment is not treated as a primary factor for discussion herein. In summary, the transpiration cooling magnitude of C4, C6, and C8 is insufficient to offset the adverse effects of inadequate shading, elevated Tmrt, and increased relative humidity on the human heat balance, thus resulting in the inverse response of decreased Ta but increased PET. From this, it can be inferred that in high-density built-up areas, planting evergreen trees with tower-shaped canopies (such as Dragon juniper) intercepts less solar shortwave radiation and tends to cause a higher accumulation of longwave radiation, which paradoxically facilitates the emergence of localized high-temperature and high-humidity microclimates. In contrast, deciduous broad-leaved trees with lush summer foliage and expansive canopy spreads can provide superior, continuous shading. It is speculated that they are more advantageous for improving outdoor summer thermal comfort in residential areas.

4.3. Synergistic Strategies of Tree Height and Spatial Layout for Optimizing Thermal Comfort in Residential Areas

Regarding tree height, this study found uncertainties in the thermal comfort improvement effects of trees at heights of 4, 6, and 8 meters. Although short trees can produce a certain cooling effect through transpiration, their limited canopy height, crown spread, and shade projection make it difficult to form stable shade at the height of human activity. Only when the canopy height and shading range reach a certain threshold can human thermal comfort be stably improved. In terms of spatial layout, under the condition of tall trees, a row layout can enhance the continuity of shade along pedestrian paths, reducing residents’ thermal exposure while moving. A cluster layout is more suitable for concentrated activity spaces. While short trees in a cluster layout may suffer from insufficient shading, tall trees in a cluster layout can form a stable shadow core, thereby reducing the area of extreme heat stress. A free layout offers stronger landscape naturalness and spatial flexibility, but its shade distribution is relatively dispersed. Without targeted control considering the sun’s path and activity spaces, it may lead to fragmented local shadows and uneven heat exposure.

First, tall trees should be prioritized in spaces frequently used during summer to improve the continuity of shade in human activity areas. Special emphasis should be placed on covering children’s play areas, elderly resting areas, residential entrances, main pedestrian paths, and areas significantly exposed to the western sun in the afternoon. Second, when adopting cluster greening, dense congregations of short trees should be avoided; priority should be given to tree species with larger crown spreads, higher clear trunk heights, and stronger shading capacities. Third, the arrangement of trees should be integrated with building shadows, road orientations, and the locations of open spaces to form a continuous yet not overly enclosed shading network. This ensures that the landscaping not only reduces radiation exposure but also maintains good spatial openness and user comfort.

Furthermore, the selection of tree species should also be considered in synergy with the spatial layout. The LAD, crown spread, clear trunk height, and canopy light transmittance of different tree species will significantly affect shading efficiency and Tmrt. For tree species with relatively compact crowns and limited horizontal shade expansion capabilities, such as the Dragon juniper, their thermal comfort improvement effects may be restricted if planted in a cluster format at low-to-medium heights. In activity spaces that require large areas of continuous shade, it is more appropriate to select tall trees with larger crown spreads, higher clear trunk heights, and stronger canopy shading capacities. Therefore, optimizing thermal comfort in residential areas requires a comprehensive consideration of the matching relationships among tree height, canopy structure, LAD distribution, and spatial layout.

4.4. Research Limitations

This study still has certain limitations. First, the scenario simulation was limited to the daytime of a typical hot summer day, making it difficult to comprehensively reflect the thermal comfort effects under different weather types, continuous high-temperature processes, and seasonal changes. Meanwhile, the nocturnal thermal environment was not further analyzed, whereas trees might affect longwave radiation dissipation and near-surface ventilation at night. Future studies could introduce continuous multi-day observations and all-weather microclimate simulations. Second, the setting of model parameters was somewhat simplified. To control variables, this study parameterized the crown spread and LAD distribution of plants, which cannot fully capture the differences among actual tree species in crown morphology, light transmittance, and transpiration capacity. In addition, using default values may also introduce certain model uncertainties. Future research could combine measured data for a more refined calibration of the plant model. Finally, this study evaluated the thermal comfort improvement effect primarily from the perspective of the physical environment, without fully considering residents’ actual activity paths, stay durations, behavioral adaptations, and differences in individual thermal perception. Future studies could integrate behavioral observations and questionnaire survey data to construct residential greening optimization methods that are closer to actual usage needs.

Overall, the regulatory mechanisms revealed in this study still need further verification under more diverse climate backgrounds, residential morphologies, and tree species types.

5. Conclusions

This study employed an ENVI-met simulation approach validated with field measurement data to investigate the combined effects of tree height and spatial layout on pedestrian-level thermal comfort in coastal residential areas. Tall trees can improve the local thermal environment by enhancing canopy shading and reducing human radiant exposure [42,43], while clustered greening generally produces a more pronounced local cooling effect [44,45]. To some extent, this study extends the existing theoretical framework. Zhang et al. [46] indicated that the interception of shortwave radiation is the dominant mechanism affecting pedestrian thermal comfort in high-density urban areas. This study further quantifies how canopy height and layout continuity jointly regulate this radiation-dominated mechanism. In addition, Yan et al. [47] emphasized the importance of vegetation configuration in optimizing outdoor thermal comfort. By integrating tree height and spatial layout as two key variables, this study establishes a more systematic greening assessment framework at the residential community scale. The results reveal the following key findings:

• Tree height is a key factor influencing the improvement of summer thermal comfort. Compared with 4 m, 6 m, and 8 m trees, 10 m tall trees can significantly reduce Tmrt and PET during peak radiation periods, especially under extreme heat stress conditions. This indicates that the vertical development of the canopy plays a decisive role in mitigating extreme pedestrian-level heat load.
• Spatial layout modulates the cooling effect of tree height. The cluster layout shows the most stable and efficient capacity to reduce PET, followed by the row layout, whereas the free layout exhibits relatively unstable improvement due to fragmented shading and insufficient continuity. This suggests that shading continuity and spatial integrity are essential prerequisites for effective radiation control.
• Although low- and medium-height trees (≤6 m) in the cluster layout can reduce local air temperature through transpiration, their insufficient capacity to intercept shortwave radiation and the enhanced accumulation of longwave radiation may instead lead to higher PETs than the baseline. Therefore, greening optimization in residential areas should comprehensively consider the matching relationship among tree height, canopy structure, and the layout of activity spaces.