Creating a Thermally Comfortable Environment for Public Spaces in Coastal Villages Considering Both Spatial Genetics and Landscape Elements

Abstract

Thermal comfort is an important criterion affecting the comfort evaluation of public spaces in villages. However, related studies remain scarce because of the intricate climates of seafront villages. In this study, the effect of landscape elements on thermal comfort within public spaces in seafront villages was examined. The spatial gene method was employed to extract the layout characteristics of typical public spaces and identify villages with the most comprehensive spatial elements as simulation subjects to enhance our understanding. The Physiological Equivalent Temperature (PET) was selected to quantitatively assess the effect of landscape elements on thermal comfort. The analysis results revealed varying thermal mitigation capacities across different types of landscape elements. Plants, notably in plazas and courtyards, along with buildings on beaches, emerged as the most significant contributors to thermal comfort. Moreover, a diurnal variation in the influence of landscape elements on thermal comfort was observed, attributed to the unique climatic conditions of seafront villages. During daytime, structural elements exerted the most substantial effect on PET in public spaces, accounting for more than 60%, whereas their influence waned in the evening. In contrast, as the sea breeze intensified in the evening, the planting method contributed over 71% to PET.

1. Introduction

Surveys indicate that nearly half of China’s current population resides in rural areas [1], where public spaces hold significant functional importance in the villagers’ daily lives. With increasing demand for quality in village public spaces [2], thermal comfort is considered a paramount criterion for assessing their utility and villager satisfaction [3]. Inadequate thermal comfort in these spaces deters villagers’ participation in activities and intensifies the heat island effect, increasing energy consumption within the village [4]. Hence, addressing thermal comfort needs in public spaces is essential for enhancing their usage and, consequently, the overall quality of life of the local population.

Villages in China’s coastal regions are distinguished by traditional features encapsulated by local lifestyles and customs [5]. Influenced by the hot and humid climate, maritime culture, and belief systems prevalent in western Guangdong, villages in that area contain public spaces deeply ingrained in regional attributes. Public spaces in seafront villages exhibit complex and diverse thermal and environmental characteristics shaped by various factors, including climate, topography, architecture, and vegetation. Consequently, developing targeted strategies to mitigate thermal discomfort in coastal village public spaces represents a challenge that merits further exploration and resolution.

The study of thermal comfort in public spaces has yielded numerous scholarly contributions encompassing both urban and rural environments. Extensive research has been conducted on thermal comfort in various urban public spaces, including squares [6], parks [7,8], and streets [9,10]. Other studies have investigated the thermal comfort of city public spaces at diverse scales [11,12,13]. In terms of rural public spaces, most thermal comfort research originated in China. Xiao et al. synthesized the characteristics of traditional village spaces and subsequently utilized the ENVI-Met5.1.1 software to assess thermal comfort and devise pertinent strategies [14]. In contrast, Fan et al. conducted a statistical and spatial quantification of public spaces in Henan’s traditional villages using ArcGIS10.8, leading to detailed thermal comfort evaluations and planning using ENVI-met5.1.1 [15]. Nevertheless, thermal comfort research has predominantly focused on inland areas, with islands and coastal regions receiving less attention, although in recent years, researchers have gradually focused on this part of the world. For instance, Tang et al. combined field research with average radiant temperature estimation methods to determine the thermal preferences of beachgoers and residents and explored the distinctive thermal comfort of beach areas [16]. Similarly, Bare investigated thermal comfort perceptions related to landscape elements among the inhabitants of Qingdao’s coastal villages through interviews [17]. These are active explorations of thermal comfort in coastal villages, but they are still at the primary stage of investigation and empirical measurement. The lack of comprehensive research in this domain can be attributed to the following factors:

(i) The diversity and complexity of public spaces in rural coastal areas present significant challenges in establishing a universal thermal environmental framework [18].

(ii) The intricate climatic conditions of coastal regions, characterized by the ocean’s slow thermal absorption and release, result in more gradual temperature variations and minimal seasonal differences [19].

(iii) Maritime influences such as humidity and sea breezes play notable roles in shaping the thermal environment [20].

The recent increase in scholarly attention towards the development of human habitats in remote villages underscores the significance of enhancing rural thermal comfort, which is pivotal for improving the residents’ quality of life. In response, extensive research has been conducted to devise strategies for improving thermal comfort. These strategies focus on modifying the impact of landscape elements on thermal comfort by optimizing plant arrangements [21], selecting appropriate tree species [22], employing various underlays [23], and incorporating water features [24]. Yu et al. furthered this field by conducting a quantitative analysis of thermal comfort in relation to common landscape configurations within plaza spaces, proposing 15 retrofit scenarios to examine the interactive effects of these configurations on thermal comfort [25]. These studies suggest that reconfiguring the coastal village landscape is an effective approach to optimizing the thermal environment of public spaces.

In terms of accelerating village construction, integrating climatological insights and technological advancements into rural landscape design is key to fostering more sustainable and climate-resilient rural environments, thereby enhancing the comfort and well-being of rural inhabitants. Consequently, this study focused on the public spaces of the coastal villages of Naozhou Island, western Guangdong Province, China. The flowchart for analyzing the effect of landscape elements on thermal comfort in a seaside village is shown in Figure 1. Using the spatial gene method, public spaces were categorized into three distinct types: temple-affiliated plazas, public courtyards, and open beaches. The planar characteristics and spatial components of each type were examined to establish baseline models for working conditions. The ENVI-met5.1.1 software was used to recalibrate the combination of landscape elements within the public spaces for thermal comfort assessment using the Physiological Equivalent Temperature (PET) metric and combined with an orthogonal experimental methodology with 81 test scenarios for each typical public space to facilitate the analysis of their impact on thermal comfort, and ultimately to derive the optimal combination of landscape elements for the thermal comfort significance of the waterfront villages. On this basis, it helps designers to use appropriate landscape configurations to minimize thermal discomfort.

2. Materials and Methods

In this study, the influence of landscape elements on the thermal comfort of different public space types (including square, beach, and public courtyard) is quantitatively investigated using the spatial genetic method. Taking the public spaces of a coastal village on Naozhou Island as an example, this study explores how the selection of landscape elements can be adapted to the public spaces of the coastal village as a way to promote the creation of a comfortable environment. The results of this study can provide some effective suggestions for the development of China’s coastal villages.

2.1. Study Area

The objects of this study were public spaces within the seafront villages of Naozhou Island, situated in a low-latitude area near the Tropic of Cancer in the western part of Guangdong Province, China. According to China’s Thermal Design Code for Civil Buildings [26], Naozhou Island is classified as a hot summer and warm winter zone, exhibiting a tropical and subtropical monsoon climate. Influenced by a year-round maritime climate, the island is characterized by mild winters and intensely hot summers that last for an extended period, with the summer period extending from early April to early November and the autumn and spring periods spanning from mid-November to early April.

The selection of seafront villages on Naozhou Island as the research subject was intentional because of their pronounced exposure to the marine climate (Figure 2). The study area encompassed 50 natural villages, inclusive of five administrative villages and two community neighborhood committees. A comprehensive documentation of the landscape elements was conducted, encompassing both spatial and physical aspects. Spatial records entailed the layout, scale, and enclosure of public spaces; physical records included vegetation, subsurface types, pathways, edifices, and structures. Subsequently, field photography of each public space was conducted, emphasizing the variance in thermal comfort across different public spaces.

2.2. Plane Characteristics of Public Spaces in Coastal Villages Based on the Spatial Genetic Approach

Village public spaces exhibit distinct and stable spatial patterns that encapsulate the fundamental principles that ensure the continuity of traditional villages and mirror local culture and its nuances [27]. The traditional spatial classification methods, such as those based on spatial morphology [28] or a typological approach [29], are more inclined to focus on isolated single-dimensional features and are insufficient for quantitative analysis of the overall spatial features inside and outside the village. In view of this, the spatial genetic method of synthesizing geometry and topology based on the principle of multidimensionality and spatial wholeness is chosen.

The spatial gene is derived from the theory of the urban complex system [30], which has an independent and relatively stable spatial combination pattern [31]. It is not only a copy of historical forms and symbols but also a continuation of the territorial combination pattern of spatial elements and their inherent generative mechanism, which forms morphological organization and the place-making method. The spatial gene has a hierarchical nature, which exists at each spatial level of the city, colony, neighborhood, street, courtyard, and so on, and it is composed of genetic components and diagrams [32]. The extraction method is elements, patterns, and texts. The morphological expression of space extracts patterns. The cultural, social, historical, and natural laws contained in the space can only be expressed by text, and the two form a basic element of space. In this study, various public space types and components in coastal villages were dissected and reassembled to discern the essential iconographic language emblematic of coastal villages in western Guangdong. They were arranged according to specific cultural significance, with each component bearing a particular spatial relevance [33]. The process of extracting spatial genes is shown in Figure 3.

The spatial genes of Naozhou Island’s coastal villages were stratified into five concentric layers, starting from the outermost external environment layer and moving inward to the boundary form, street organization, and public space layers, culminating in the traditional architectural layer [34] (Figure 4). The five basic layers are then deconstructed and categorized based on the need to continue extracting the constituent elements; the extracted contents are elements, patterns, structures, and text [35]. Element extraction analyzes the composition of landscape features one by one. Pattern extraction refers to finding patterns with distinctive characteristics and iconic cultural features. Structure extraction analyzes the spatial morphological features of traditional villages to summarize their implied landscape intentions. Finally, text extraction is used to describe cultural features with special meanings but without material shapes, such as the sacrificial rituals of the fishermen in the temple plaza before going out to sea. The public space gene strata were deconstructed, classified, and distilled into discrete spatial gene components, as shown in

Table 1. Classification of spatial gene extraction methods for coastal villages.

Public Space Hierarchy	Identifying Factor	Genome Component Indicators	Extracted Contents
Plot genetics	Parcel subdivision type	Squares, courtyards, beaches, streets, village entrances	Pattern
	Plot function	Places of faith Open space Tourist places	Text
	Number of public spaces	Rich, common, crowded	Text
Borderline gene	Boundary composition type	Sea areas, beaches, windbreaks, fences, roads	Pattern and text
	Boundary layout form	Vertical, parallel, separate	Pattern
	Boundary space forms	rectangles, irregular	Pattern
Material gene	Building footprint form	L-shaped, rectangular	Pattern
	Ancillary structure	Theater, cultural pagoda, village gate	Pattern and text
	Location of buildings in the space	In space, enclosing space, embedding space	Pattern and text
	Building function	Faith-based buildings, residential buildings, public facilities	Pattern and text
	Vegetative	Form of planting, number, species, symbolism	Pattern and text

.

Spatial genome components denote schemas of individual elements, whereas spatial types are more complex schemata amalgamated from these components. Owing to computational constraints and the processing capabilities of the simulation software, modeling all coastal villages on the island was unfeasible. Consequently, the analysis was confined to certain public spaces within specific villages. In alignment with the villagers’ utilization patterns of public spaces and the prevalence of congregational activities, three public spaces were selected for their congruence with the aforementioned spatial gene classification and embodiment of western Guangdong’s coastal village characteristics. The spaces selected for analysis were the temple plaza in Gangtou Village, a public courtyard in Dedou Village, and an open beach in Cunliang Village. The composite schema of the public space of the seafront village is in

Table 2. Identification of three typical public spaces in coastal villages.

Location	Space Type	Space Genome Components	Graphical Language
Gangtou Village	Temple annex plaza	Temple buildings in the center
		Rectangular boundary
		Vertical sea
		Subsidiary theater
Dedou Village	Public courtyard	Residential building enclosure
		Rectangular boundary
		Subsidiary theater
		Isolated sea area
Cunliang Village	Open beach	Residential building embedded
		Banded rectangular boundary
		Parallel sea area

. The extracted graphic language is a reference for modeling.

2.3. Typical Public Spaces in Coastal Villages

Temple square (Figure 5a): situated in the northern sector of Naozhou Island, Gangtou Village’s temple, known as Zhen Tian Shai Fu, stands at the village forefront, facing the sea. In close proximity, the theater serves as the temple’s cultural complement, hosting operatic events and other activities integral to the villagers’ daily lives. Flanked by buildings and vegetation, this space is steeped in the islanders’ longstanding tradition of deity worship. The square adjacent to the temple is a repository of rich cultural faith that serves as the primary venue for villagers to practice their beliefs and congregate for various activities, especially during major festivals, sea blessings, and theatrical performances. The dashed red line in Figure 5 shows the modeling scope.

Public courtyard (Figure 5b): nestled in the south-central region of Naozhou Island, Dedou Village, though not directly coastal, is significantly influenced by the marine climate. As a semi-open space framed by cultural buildings, the village’s centrally located public courtyard is predominantly used by residents and features a theater stage and concrete-paved lower cushion, facilitating daily life and leisure activities, thereby making it a quintessential public space for local inhabitants where the community engages in routine and recreational pursuits.

Open beach (Figure 5c): Cunliang Village, located in the southern expanse of Naozhou Island, features Baiping Beach, a scenic locale that attracts numerous visitors and is a renowned tourist destination on the island. The expansive spatial scale and unobstructed surroundings of the open beach, coupled with a coastal road and a rich array of landscape elements, create an inviting atmosphere. Strategically placed cultural buildings serve a dual purpose: they enhance the visual landscape by seamlessly blending into the environment and cater to commercial needs, supporting the area’s tourism industry. The beach subsurface is a composite of concrete and sand, which significantly influences the thermal comfort experienced by tourists and residents. The spatial elements in the typical spatial aerial photographs obtained in this study were organized to serve as a model basis for the thermal environment simulation.

2.4. Numerical Simulation Methods

Numerical simulations facilitate the recreation of the field environment through computer modeling, enabling the calculation of pertinent data and scientific validation of theoretical constructs. The ENVI-met5.1.1 software was employed for modeling and simulation [36,37], capable of replicating the wind, heat, humidity, and insolation conditions within small- to medium-sized microclimate settings. This tool has been instrumental in investigating thermal comfort in rural areas [18] and is widely recognized for its efficacy in analyzing the climatic impact of subsurface materials, building components, and vegetation [38,39]. Given the complex interplay between meteorological elements, outdoor underlays, air, and vegetation, ENVI-met5.1.1 was used to simulate thermal comfort in the public spaces of Naozhou Island’s seafront villages.

2.4.1. Thermal Environment Modeling Setup

The location of the site is chosen as Naozhou Island, the number of modeling grids is 150 m × 160 m × 20 m, and the unit grid size is 2 m × 2 m × 1 m; considering that there is a height difference of 2 m between the public space and the seashore, the grid size in the Z-direction takes the initial value of 4 m and increases by 20% in equal proportions.

The initial meteorological data conditions are the measured data of the public space of the three villages on 18 July 2022, including air temperature, relative humidity, wind speed, and direction. Meanwhile, the plant species corresponding to the site of Yingming Village are found in the ENVI-met5.1.1 plant database, and the subsurface settings of sand and clay, concrete, granite, grass, and pervious tiles are assigned based on field research. The parameters are set to the default software values, and the detailed parameters are shown in

Table 3. ENVI-met5.1.1 initial model parameters.

Name	Parameter Selection
Simulation location	Naozhou Island (20.91° N, 110.59° E)
Number of grids (x, y, z)	80, 80, 20
Size of x, y, and z grid cells	2 m, 2 m, 1 m
Simulation time	18 July 2022, 8:00–20:00, 12 h
Temperature range	28–36 °C
Humidity range	70–90%
Wind speed and direction	Data recorded by KESTREL 5500 anemometer (Nielsen-Kellerman Co., Boothwyn, PA, USA), wind direction 135° (southeast)
Surface roughness	0.01 m (ENVI-met5.1.1 default)
Cloudiness	0
Vegetation	Trees 8 m high and 10 m wide, LAD = 4.0 m2/m3; trees 10 m high and 12 m wide, LAD = 5.0 m2/m3; trees 8 m high and 6 m wide, LAD = 3.5 m2/m3
Buildings	Thermal transmittance of walls and roofs: 0.5 W m−2 K−1 (default value) Albedo of walls and roofs: 0.5 (default value)
Underlayment	Sand and clay (albedo: 0), concrete (albedo: 0.5), granite (albedo: 0.5), grass (albedo: 0.2), pervious tiles (albedo: 0.2)

.

In the model setup, it is necessary to correspond the on-site measurement point locations to the corresponding locations in the model grid so that the data corresponding to the actual measurement point locations can be extracted from the model for comparison.

2.4.2. Software Feasibility Verification

In order to verify the applicability of the ENVI-met5.1.1 software on Naozhou Island, five points in the public space of Yingming Village on Naozhou Island were selected for verification, and this public space belongs to the open beach space in the public space layer of the genetic public space of the seafront village space mentioned in the previous section. The verification results are shown in Figure 6. From the overall results, the numerical simulation results of measurement points 1~5 are more consistent with the trend of the measured values. Measurement points 1, 4, and 5 are the wind source entrances to the public space of Yingming Village, mainly measuring wind speed, temperature, and relative humidity; measurement points 2 and 3 were chosen to measure the intensity of solar radiation on the same day, based on the measurement of wind speed, temperature, and relative humidity, using the presence of shade as a control, and placing a solar radiometer at the same time. The actual measurement process is inevitably affected by other external factors, resulting in some fluctuations in the measured values. For example, the measured values of relative humidity at each measurement point are slightly larger than the simulated values due to the difference in the wind direction and size of the sea breeze, which carries away the heat in the space, so there is a certain error between the simulated values and the measured values.

In order to further determine the reliability of the simulation software, this paper introduces the Root Mean Square Error (RMSE) and the Mean Absolute Percentage Error (MAPE) to compare and analyze the values of air temperature and relative humidity at each measurement point obtained from the simulation with the measured values, and the formulas of the two are as follows:

$$RMSE=\sqrt{{\displaystyle \frac{1}{n}}\sum _{\mathit{t}=1}^n{\left(A_t-F_t\right)}^2}$$

(1)

$$MAPE={\displaystyle \frac{1}{n}}\sum _{\mathit{t}=1}^n\left|{\displaystyle \frac{A_t-F_t}{F_t}}\right|$$

(2)

In both equations, A_t is the simulated value, F_t is the measured value, and n is the number of measurements.

From the validation results of air temperature and relative humidity at each measurement point in

Table 4. Quantitative evaluation between simulated and measured values at each measurement point.

Meteorological Factors	Validation Metrics	Measurement Point
		1	2	3	4	5
Air temperature	RMSE/°C	0.58	0.71	0.52	0.70	0.56
	MAPE/%	1.63	1.71	1.49	1.96	1.55
Relative humidity	RMSE/°C	4.72	4.71	4.87	4.66	4.78
	MAPE/%	4.95	4.76	4.80	4.98	4.87

, the RMSE variance for air temperature is in the range of 0.52–0.71 °C, and the error margin of the MAPE value is in the range of 1.49–1.71%, which is a relatively small range of variation. Furthermore, the RMS variance of relative humidity is in the range of 4.66–4.87%, and the MAPE value is in the range of 4.76–4.98%. According to the relevant studies, the software calibration results are within the error margins [40]. Therefore, the results of this experiment show that the ENVI-met5.1.1 software has a good fit between the simulated values and the measured results when using the measured meteorological data as the initial boundary conditions. Considering potential human and instrumental errors, the simulation software can be used as a numerical simulation tool to assess thermal comfort in the public spaces of the seafront villages on the island of Naozhou in this study.

2.4.3. Thermal Comfort Evaluation Criteria

PET, a prevalent metric for evaluating outdoor thermal comfort [41], assesses the effect of meteorological and physiological factors on human comfort within the village milieu [42]. The evaluation criterion amalgamates diverse elements such as meteorological conditions, geographic location, attire, and individual characteristics. Owing to the extensive application of PET across various climatic zones, an adjusted PET assessment range was adopted in this study and tailored for the hot and humid Lingnan region, as shown in

Table 5. Scope of assessment of summer PET in the hot and humid Lingnan region.

PET Value (°C)	Thermal Sensation
≤11.3	Cool
11.3–19.2	Slightly cool
19.2–28.7	Neutral
28.7–36.0	Slightly warm
36.0–43.2	Warm
43.2–53.6	Hot
≥53.6	Very hot

[43].

2.5. Remodeling

This study aimed to ascertain the extent to which landscape elements affect thermal comfort in Lingnan villages and devise an experimental optimization scheme for their collaborative function. The field’s original landscape element amalgamation did not fully meet the research requirements, prompting the reconfiguration of the thermal comfort model for public spaces using an orthogonal experimental method.

The orthogonal experimental method, which is a multifaceted and multilevel design technique, leverages orthogonality to select representative points from comprehensive tests. These points are characterized by their “uniform dispersion and comparability”, offering a rapid, efficient, and convenient means of conducting multifactorial tests. This approach aimed to illustrate the entire spectrum of potential landscape element combinations. Using this method, a typical public space model was organized into one three-factor, three-level, and two four-factor, three-level orthogonal experimental schemes. PET served as the experimental target to ultimately quantify the impact of each landscape element on the thermal comfort of public spaces.

2.5.1. Establishment of Test Factors and Levels

Mitigating thermal discomfort through landscape elements has been rigorously examined in previous studies. The typical spatial types of coastal villages in Naozhou Island exhibit both commonalities and diversity in their landscape elements. After a thorough collection and organization of public space data from coastal villages, the landscape elements that most significantly influence thermal comfort were identified as follows: tree species (T), plant configuration (C), green space ratio (G), planting distance (D), planting method (P), activity site location (A), structure (S), and underlay type (U). These eight categories were analyzed for their interactions with each typical space.

Tree species (T): Tall trees effectively block solar radiation and only columnar planting is permissible in temple plazas because of their sacred nature. Common tree species in temple-affiliated plazas include Ficus microcarpa, Ficus olivaceus, and Hibiscus tiliaceus. The average values of these dimensions were calculated from field research data. The tree species of the temple-affiliated plaza in Gangtou Village was categorized into three levels: Ficus microcarpa (T1), whose height, crown, and height below branches are 8, 8, and 2.5 m, respectively; elecampane (T2), whose height, crown, and height below branches are 7, 7, and 2 m, respectively; and Hibiscus sabdariffa (T3), whose height, crown, and height below branches are 7.5, 6, and 2.5 m, respectively.

Plant configuration (C): Strategically matched tree groupings significantly enhance outdoor thermal comfort [44]. The plant structure in the temple plaza is limited, whereas the beach, oriented towards tourists, typically features a three-dimensional plant structure. The planting configuration of the public courtyard varies with villagers’ financial resources. Thus, the planting configuration in the public courtyards was categorized into three levels: trees + grass (C1), shrubs + grass (C2), and trees + shrubs + grass (C3).

Green space ratio (G): The public courtyard in Dedou Village, which serves as a communal square, must accommodate periodic group activities, limiting its green space ratio. Following Chinese urban green space standards [45], green coverage was divided into three levels: less than 35 (G1), 35–65 (G2), and >65% (G3). Three green space areas were set for Dedou Village’s public courtyard.

Planting distance (D) and planting method (P): In hot and humid climates, random tree planting can squander the cooling effects of trees, and excessive irrational planting fails to improve the thermal environment [22]. On Naozhou Island, the temple plaza planting method is restricted to columnar planting, and the planting distance significantly affects thermal comfort. In Gangtou Village’s temple plaza, canopy spacing for columnar planting was set at 3 (D1), 6 (D2), and 9 m (D3). The public courtyard and open beach, with their diverse planting methods, had three planting levels: columnar (P1), clump (P2), and a combination of both (P3), established for the public courtyard in Dedou Village and the open beach in Cunliang Village.

Activity space location (A): For an open beach frequented by tourists and residents, the location of the main activity site is crucial because it often serves as the core of the entire space and determines the primary area for human activity. Thus, identifying these areas is vital for the thermal comfort model of open beach spaces. Three levels were designated for beach location: the western part near residential areas (A1), the central part (A2), and the eastern part close to the coast (A3).

Structure (S): Beaches require artificial structures to satisfy the increased thermal comfort needs of tourists and villagers [46]. For this case study, three prevalent structures—the pavilion (S1), promenade (S2), and tensile membrane (S3)—were selected as variables to establish three levels with surrounding tree planting to mitigate the influence of other factors.

Underlay type (U): In landscape design, various underlay types differentiate activity areas and influence the thermal comfort of public spaces based on properties such as reflectivity and water permeability [47]. Consequently, three underlay combinations were designated for Gangtou Village’s temple plaza and Cunliang Village’s open beach: concrete + grass (U1), permeable brick + grass (U2), and granite + grass (U3).

Subsequently, an orthogonal test factor level table was established to quantify the influence of each landscape element on the thermal comfort of public spaces. The impact of thermal comfort in Gangtou Village’s temple plaza was based on tree species, planting distance, and underlay type; that of Dedou Village’s public courtyard considered the green space ratio, plant structure, planting method, and underlay material; and that of Cunliang Village’s open beach focused on activity site location, structure, planting method, and underlay type, as listed in

Table 6. Factor levels for the TS, PC, and OB orthogonal test for public space modeling.

Public Space Model	Factor		Level
		1	2	3
Temple square (TS)	T	Ficus (T1)	Olive tree (T2)	Hibiscus (T3)
	D	3 m (D1)	6 m (D2)	9 m (D3)
	U	Concrete + grass (U1)	Permeable brick + grass (U2)	Granite + grass (U3)
Public courtyard (PC)	G	35% (G1)	50% (G2)	65% (G3)
	C	Arbor + grass (C1)	Shrub + grass (C2)	Tree + shrub + grass (C3)
	P	Column planting (P1)	Cluster planting (P2)	Column + cluster (P3)
	U	Concrete + grass (U1)	Permeable brick + grass (U2)	Granite + grass (U3)
Open beach (OB)	A	Space west (A1)	Center of space (A2)	Space east (A3)
	S	Pavilion (S1)	Promenade (S2)	Tensioned membrane (S3)
	P	Column planting (P1)	Cluster planting (P2)	Column + cluster (P3)
	U	Concrete + grass (U1)	Permeable brick + grass (U2)	Granite + grass (U3)

.

2.5.2. Test Modeling

In this study, three typical public space abstract models were established using 1 × 1 m grids in ENVI-t, and the data of typical summer weather days on Naozhou Island were used as the initial boundary conditions to simulate the working conditions. By analyzing the simulation results of different influencing factors in the public space at 1.5 m above ground, the simulation time was chosen as the hottest time of the day and time when the residents were more concentrated, that is, 8:00 (morning), 14:00 (afternoon), and 18:00 (evening).

Considering the effects of the characteristics of the landscape elements and their interactions on thermal comfort, we used the standardized table L27 (3¹³) to develop the test protocols for the three typical public space models, as detailed in Appendix A 

Table A1. L27(3¹³) test protocol table of public space TS.

Exp	1	2	3	4	5	6	7	8	9	10	11	12	13
	T	D	T × D	T × D	U	T × U	T × U	D × U	Error Column	Error	D × U	Error	Error
1	1	1	1	1	1	1	1	1	1	1	1	1	1
2	1	1	1	1	2	2	2	2	2	2	2	2	2
3	1	1	1	1	3	3	3	3	3	3	3	3	3
4	1	2	2	2	1	1	1	2	2	2	3	3	3
5	1	2	2	2	2	2	2	3	3	3	1	1	1
6	1	2	2	2	3	3	3	1	1	1	2	2	2
7	1	3	3	3	1	1	1	3	3	3	2	2	2
8	1	3	3	3	2	2	2	1	1	1	3	3	3
9	1	3	3	3	3	3	3	2	2	2	1	1	1
10	2	1	2	3	1	2	3	1	2	3	1	2	3
11	2	1	2	3	2	3	1	2	3	1	2	3	1
12	2	1	2	3	3	1	2	3	1	2	3	1	2
13	2	2	3	1	1	2	3	2	3	1	3	1	2
14	2	2	3	1	2	3	1	3	1	2	1	2	3
15	2	2	3	1	3	1	2	1	2	3	2	3	1
16	2	3	1	2	1	2	3	3	1	2	2	3	1
17	2	3	1	2	2	3	1	1	2	3	3	1	2
18	2	3	1	2	3	1	2	2	3	1	1	2	3
19	3	1	3	2	1	3	2	1	3	2	1	3	2
20	3	1	3	2	2	1	3	2	1	3	2	1	3
21	3	1	3	2	3	2	1	3	2	1	3	2	1
22	3	2	1	3	1	3	2	2	1	3	3	2	1
23	3	2	1	3	2	1	3	3	2	1	1	3	2
24	3	2	1	3	3	2	1	1	3	2	2	1	3
25	3	3	2	1	1	3	2	3	2	1	2	1	3
26	3	3	2	1	2	1	3	1	3	2	3	2	1
27	3	3	2	1	3	2	1	2	1	3	1	3	2

,

Table A2. L27(3¹³) test protocol table of public space PC.

Exp	1	2	3	4	5	6	7	8	9	10	11	12	13
	G	C	G × C	G × C	P	G × P	G × P	C × P	U	Error	C × P	Error	Error
1	1	1	1	1	1	1	1	1	1	1	1	1	1
2	1	1	1	1	2	2	2	2	2	2	2	2	2
3	1	1	1	1	3	3	3	3	3	3	3	3	3
4	1	2	2	2	1	1	1	2	2	2	3	3	3
5	1	2	2	2	2	2	2	3	3	3	1	1	1
6	1	2	2	2	3	3	3	1	1	1	2	2	2
7	1	3	3	3	1	1	1	3	3	3	2	2	2
8	1	3	3	3	2	2	2	1	1	1	3	3	3
9	1	3	3	3	3	3	3	2	2	2	1	1	1
10	2	1	2	3	1	2	3	1	2	3	1	2	3
11	2	1	2	3	2	3	1	2	3	1	2	3	1
12	2	1	2	3	3	1	2	3	1	2	3	1	2
13	2	2	3	1	1	2	3	2	3	1	3	1	2
14	2	2	3	1	2	3	1	3	1	2	1	2	3
15	2	2	3	1	3	1	2	1	2	3	2	3	1
16	2	3	1	2	1	2	3	3	1	2	2	3	1
17	2	3	1	2	2	3	1	1	2	3	3	1	2
18	2	3	1	2	3	1	2	2	3	1	1	2	3
19	3	1	3	2	1	3	2	1	3	2	1	3	2
20	3	1	3	2	2	1	3	2	1	3	2	1	3
21	3	1	3	2	3	2	1	3	2	1	3	2	1
22	3	2	1	3	1	3	2	2	1	3	3	2	1
23	3	2	1	3	2	1	3	3	2	1	1	3	2
24	3	2	1	3	3	2	1	1	3	2	2	1	3
25	3	3	2	1	1	3	2	3	2	1	2	1	3
26	3	3	2	1	2	1	3	1	3	2	3	2	1
27	3	3	2	1	3	2	1	2	1	3	1	3	2

and

Table A3. L27(3¹³) test protocol table of public space OB.

Exp	1	2	3	4	5	6	7	8	9	10	11	12	13
	A	S	A × S	A × S	P	A × P	A × P	P × U	U	Error	P × U	Error	Error
1	1	1	1	1	1	1	1	1	1	1	1	1	1
2	1	1	1	1	2	2	2	2	2	2	2	2	2
3	1	1	1	1	3	3	3	3	3	3	3	3	3
4	1	2	2	2	1	1	1	2	2	2	3	3	3
5	1	2	2	2	2	2	2	3	3	3	1	1	1
6	1	2	2	2	3	3	3	1	1	1	2	2	2
7	1	3	3	3	1	1	1	3	3	3	2	2	2
8	1	3	3	3	2	2	2	1	1	1	3	3	3
9	1	3	3	3	3	3	3	2	2	2	1	1	1
10	2	1	2	3	1	2	3	1	2	3	1	2	3
11	2	1	2	3	2	3	1	2	3	1	2	3	1
12	2	1	2	3	3	1	2	3	1	2	3	1	2
13	2	2	3	1	1	2	3	2	3	1	3	1	2
14	2	2	3	1	2	3	1	3	1	2	1	2	3
15	2	2	3	1	3	1	2	1	2	3	2	3	1
16	2	3	1	2	1	2	3	3	1	2	2	3	1
17	2	3	1	2	2	3	1	1	2	3	3	1	2
18	2	3	1	2	3	1	2	2	3	1	1	2	3
19	3	1	3	2	1	3	2	1	3	2	1	3	2
20	3	1	3	2	2	1	3	2	1	3	2	1	3
21	3	1	3	2	3	2	1	3	2	1	3	2	1
22	3	2	1	3	1	3	2	2	1	3	3	2	1
23	3	2	1	3	2	1	3	3	2	1	1	3	2
24	3	2	1	3	3	2	1	1	3	2	2	1	3
25	3	3	2	1	1	3	2	3	2	1	2	1	3
26	3	3	2	1	2	1	3	1	3	2	3	2	1
27	3	3	2	1	3	2	1	2	1	3	1	3	2

. The experiments considered the interactions between the experimental factors; the numbers 1, 2, and 3 indicate the level of each factor, and the blank columns represent the error columns of the analysis of variance (ANOVA). Each typical public space had 81 test scenarios, of which 27 represented the working conditions. By using the orthogonal test method to filter the scenarios, each typical public space could be simulated only for the 27 corresponding working conditions; the model scenarios are shown in Figure 7.

3. Results

By simulating the 81 transformation schemes, the k value of the orthogonal test results can be compared to obtain a superior ranking of the thermal comfort impact of the working conditions of the landscape elements of the typical public space in coastal villages. Polar analysis and ANOVA were used as the analysis methods for the orthogonal test results. The polar analysis has the advantage of simplicity and intuition and is sufficient for screening experiments that do not require high accuracy of analysis; however, the polar analysis cannot estimate the size of the error and cannot accurately estimate the degree of importance of the influence of each factor on the results, especially when the number of levels is greater than or equal to 3 and interactions must be considered. When the polar analysis is not sufficient, multi-factor ANOVA can be chosen to analyze the results. In this study, ANOVA was performed using SPSS Statistics 25 software to verify the effect of landscape elements on thermal comfort in public spaces.

In analyzing the significance results, a larger F value represents a larger fluctuation of the experimental indicators caused by the change in the factor and a more significant effect. The Sig value is used to test the significance indicators, and when the significance level is 0.05, a Sig value of less than 0.05 indicates that the factor has a significant effect on the experimental indicators; a value greater than 0.05 indicates that the effect is not significant.

While organizing the data, not all of the data conformed to a normal distribution; thus, the results were validated by introducing the contribution rate, calculated by taking the sum of the squares of the column deviations of the factors or interactions as a percentage of the sum of the squares of the total deviations. The contribution margin is calculated as follows:

$${\rho }_j=S_j/S$$

(3)

where ρ_j, S, S_j, and j represent the contribution rate, the sum of the squared total deviations, the sum of the squared column deviations, and the column number, respectively.

3.1. Ranking of the Thermal Comfort Impact Level of Landscape Elements

3.1.1. PET Trends of Typical Public Spaces

The orthogonal test results were analyzed via intuitive analysis to calculate the average PET of each level of each factor in three typical public spaces in different time periods. A trend graph of the relationship between each factor and PET was obtained according to the statistical results, as shown in Figure 8. The PET values of each factor in typical public spaces in the evening are in a slightly warmer range, and the landscape elements have little influence on PET. The PET values in the morning are in a warmer range, the PET values in the middle of the day are in a hotter range, and the landscape elements have a greater influence on PET. Among the typical public space types, the PET values of public space PC are generally higher than the others throughout the day because they are not close to the seashore. Thermal comfort is worse than that of the other two public space types, and there is a greater need to improve thermal mitigation capacity.

3.1.2. Factor Ranking in Order of Preference

Figure 9 shows the k values for each landscape element level in three typical public spaces, with the aim of reducing PET to improve thermal comfort. In this study, the level indicators and k values of each factor were analyzed within the public space model TS across different time periods. The analysis revealed that Ficus microcarpa was the most effective in enhancing thermal comfort during the morning and afternoon hours, followed by Elemi, with Hibiscus tiliaceus being the least effective. Evening thermal comfort was better, with negligible differences among the tree species. Given that Naozhou Island’s most acute thermal discomfort occurs in the afternoon, the overall effectiveness of tree species on thermal comfort was ranked as T1 × T2 × T3. Planting distance demonstrated a negative correlation with PET, indicating that an increased distance led to reduced temperatures. Regarding the underlay types, granite + grass proved to be optimal in the morning, whereas concrete + grass was less effective. The afternoon conditions showed slight differences among factors, with permeable bricks performing the least effectively, whereas the evening conditions showed no significant changes across underlay types. Overall, the preferred underlay types were ranked as U3 × U2 × U1.

For the public space model PC, the level indicator and k value analysis indicated that PET decreased with an increase in green space ratio, particularly in the morning when the impact on thermal comfort was more pronounced, ranking as G3 × G2 × G1. However, no significant changes were observed in the PET values during the afternoon and evening. The plant structure had the most substantial effect on thermal comfort in the morning and afternoon, with C3 being more effective than C2 and C1. In the evening, the plant structure levels did not significantly affect thermal comfort. The effectiveness of the planting methods in improving thermal comfort varied across different times of the day. Overall, P3 exhibited the best thermal mitigation ability, followed by P2; however, during the hottest afternoon hours, P2 was less effective than P1. This result may be attributed to clumped planting, which increases the wind blockage, leading to stuffy and hot conditions. Thus, the ranking of the planting methods was P3 × P2 × P1, with special attention needed for wind direction guidance. Regarding the underlay types, concrete heated faster in the morning than permeable tiles and granite, impacting thermal comfort more significantly, and the ranking was U3 × U2 × U1. Underlay type had a minimal impact on PET during the midday and evening hours.

In the OP model, the level indicator and k value analysis for each factor across different time periods showed that the location of the activity site had the best effect in the morning and afternoon because of the beneficial influence of the sea breeze near the seashore. In the evening, there were no significant changes in PET due to the location of the activity site. Among the structures, S3 was the most effective in regulating thermal comfort in the morning and afternoon, followed by S1, with S2 being the least effective. This result is related to the size of the structures and their shaded areas. In the evening, there were no significant differences among the structures in terms of improving thermal comfort owing to the reduced solar radiation. For tree-planting methods, there were no significant differences in improving thermal comfort between the morning and afternoon periods for each method, whereas, in the evening, tree planting could induce and block wind, improving thermal comfort with the ranking being P3 × P2 × P1. For the underlay types, the order of effectiveness in improving thermal comfort in the morning, afternoon, and evening periods was U3 × U2 × U1.

3.2. Factor Effect Level and Significance Analysis

Figure 10 presents the significance analysis results for each influencing factor for the three typical public spaces, conducted using SPSS STATISTICS 25 software. To facilitate observation, the F value axis is represented by a curve, with intervals of 0.025 on the left and 0.2 on the right; the Sig value threshold was set at 0.05. Thus, landscape elements significantly influenced thermal comfort when the Sig value was less than 0.05 [48].

The significance of each influencing factor within the public space model TS was analyzed using SPSS STATISTICS 25. The results indicate that the Sig values for tree species, planting distance, and their interactions were all below 0.05, indicating a substantial effect on public space PET. Conversely, the underlay type did not have a significant impact on PET.

The S_j and ρ_j for factors and their interactions were computed in SPSS STATISTICS 25 using the contribution formula (Equation (3)), with results shown in Figure 11. Planting distance emerged as a significant factor influencing PET across all time periods, particularly in the evening, with a contribution of 76.84%. The impact of tree species on PET varied with species change, reaching a maximum contribution of 29.66% in the morning. The effect of the underlay type on PET was less pronounced than the other factors, becoming more noticeable only during daytime hours. Among the interactions affecting PET, the combination of tree species and planting distance was notably significant in the morning, contributing 3.55%, whereas the interaction between tree species, planting distance, and underlay type was relatively minor, with a contribution of ~1% across different time periods.

In the temple annex plaza, tree species and planting distance are pivotal in determining thermal comfort. During daytime, the influence of these factors on the thermal comfort of public spaces varied, with the impact of the underlay being relatively limited and less pronounced. The ANOVA and contribution rate analysis indicated that the factors affecting public space TS PET were ranked as follows: D × T × D × T × U.

The Sig values for plant structure and its interaction with the green space ratio during the morning and afternoon hours in the public space model PC were less than 0.05, indicating a significant effect on PET. In the evening, the Sig values for green space ratio, plant structure, planting method, and their interactions were also less than 0.05, significantly influencing PET.

In terms of the specific influence of each landscape element and their interactions on PET in the public space PC, as depicted in Figure 9, plant structure had the most substantial influence, with contribution rates of 81.21% in the morning, 74.56% in the afternoon, and 68.63% in the evening. The impact of the green space ratio on PET increased over time, with contributions of 8.18%, 8.25%, and 13.08%, respectively. Planting methods contributed less than 2% in both morning and afternoon hours, with an increased contribution of 8.91% in the evening. Underlay type contributed 3.55% in the morning and over 8% in both afternoon and evening; despite its ANOVA insignificance, it should be considered based on its contribution rate. Comparing the different interactions shows that the combination of plant structure and green space ratio significantly affected PET, contributing 4.64% in the morning, 5.27% in the afternoon, and negligibly in the evening, whereas other interactions had low or negligible contributions.

Thus, in public courtyards, plant structure is the key factor regulating thermal comfort, and the green space ratio and underlay type also have a notable effect to some extent, although not decisively. As time progresses, the effect of the interaction between plant structure and green space ratio diminishes, yet plant structure remains dominant. The combined analysis ranked the influences on public space PET as follows: C × C × G × G × P × U.

For public space OB, the Sig values for activity site location and structures were less than 0.05 in the morning, with the significance order being S × A. In the afternoon, the interaction between the activity site location, structures, and their combination had Sig values of less than 0.05, with the significance order being S × A × S × A according to the F value size. In the evening, the Sig values for the tree-planting method, activity site location, and their interaction were less than 0.05, with the significance order being P × A × P × A.

The Sj and ρj for factors and interactions in the OB public space model were calculated, as shown in Figure 9. The results indicate that structures had the most significant influence on PET in public spaces during morning and afternoon hours, with contributions of 68.09% in the morning, 62.53% in the afternoon, and negligible in the evening. This result is attributed to the structures’ effectiveness in shading solar radiation, which diminishes in the evening, thus weakening their impact on PET. The activity site location had a more pronounced effect during daytime and a reduced impact on PET in the evening, contributing 23.15%, 29.42%, and 9.01%, respectively. The planting method had a minimal effect on PET in the morning and afternoon, with negligible contributions, but contributed over 71% to PET in the evening as the sea breeze intensified. The influence of subsurface type on PET diminished over time, with contributions of 4.87%, 2.09%, and 0.08%, respectively, which, despite being minimal, should not be overlooked. When examining the interplay of various factors, the interaction between the location of activity sites and structures significantly affected PET, contributing 1.57%, 3.94%, and 0.16% during the morning, afternoon, and evening, respectively. The synergy between planting methods and site location contributed 1.70% in the morning, 0.64% in the afternoon, and over 9% in the evening, whereas other interactions were negligible.

In open beach settings, structures are pivotal in regulating thermal comfort during the morning and afternoon, whereas the location of activity sites plays a significant but not decisive role. In the evening, the importance of simple shading decreased, and well-ventilated arboricultural planting emerged as the most influential factor for thermal comfort. A comprehensive analysis ranked the impact of landscape elements on open beach PET as follows: structure (S) × activity site location (A) × planting method (P) × interaction between activity site and structure (A × S) × interaction of activity site and planting method (A × P) × substrate type (U).

3.3. Optimal Combination Analysis Considering Factor Interactions

When factor interactions are present and exert a more substantial effect than individual factors, considering the hierarchy of influence under these interactions and re-evaluating the levels with significant interplay is essential. The most notable interaction within the public courtyard (PC) model was between green space ratio and plant structure, denoted as G and C, respectively. This interaction was not as impactful in the temple space (TS) or open beach (OB) models and was hence not considered in these contexts.

Table 7. Thermal comfort values for C and G interactions in public space PC.

		Morning			Afternoon			Night
Space PC		G1	G2	G3	G1	G2	G3	G1	G2	G3
	C1	39.87	39.85	39.92	48.79	49.72	48.12	33.64	33.34	33.67
	C2	40.02	40.03	40.62	49.52	48.19	48.37	33.30	33.67	33.85
	C3	39.45	39.87	40.02	48.21	48.35	49.54	33.85	33.88	33.27

presents the mean PET values corresponding to each factor–level interaction across the three periods. The results indicate that for PC, the combination of C3G1 offered the best thermal comfort in the morning, and the thermal comfort of CiG1 (where i = 1, 2, or 3) surpassed the others. In the afternoon, C1G3 was optimal, and in the evening, the PET value for C3G3 was lower than the rest. Overall, C3G1 provided the most favorable thermal comfort throughout the day.

Consequently, the optimal landscape element configuration for thermal comfort during the full time period in the corresponding public space model is as follows: temple space (TS): T1D1U3; public courtyard (PC): C3G1P3U3; and open beach (OB): S3A3P3U3. These findings are shown in Figure 12.

4. Discussion

The results of this study demonstrated that altering the configuration of landscape elements considerably affects the thermal comfort of public spaces in coastal villages. Specifically, in spaces associated with religious faith, planting spacing was the most influential factor, contributing an average of 65.4% of thermal comfort throughout the day. In public courtyards, the planting structure consistently played a dominant role, with an average contribution rate of 74.8%. For open beach areas, the planting method in the evening contributed up to 71.27% of thermal comfort. However, not all synergistic effects between the types of landscape elements were significant. Only interactions between tree species and planting distance in temple annex spaces, between green space ratio and plant structure in public courtyards, between the location of activity sites and structures, and between the location of activity sites and planting mode in open beaches were significant, with Sig values of less than 0.05.

The average contribution of green space ratio and plant structure to public courtyards was 74%. Plants are one of the most important elements in thermal comfort research [49], which suggests that the interaction between green space ratio and planting structure in courtyards has a more substantial effect on thermal comfort than individual factors. There is a consensus among researchers on this point, both for urban [50,51] and rural areas [18]. However, while they draw more on the synergistic effects of vegetation with other landscape elements, this paper emphasizes the interaction between the configuration of the plants themselves and the percentage of area occupied by them, highlighting that single-factor studies are insufficient for comprehensively understanding thermal comfort in coastal village public spaces. Therefore, the combined effects of multiple factors warrant further investigation.

The findings of this study revealed that the thermal mitigation capacity of individual landscape elements fluctuates throughout the day. Despite natural cooling in the evenings in hot and humid climates, residents still experience heat discomfort during summer evenings, with PET remaining in a slightly warmer phase. Consequently, the thermal mitigation capacity of landscape elements in the evening influenced the overall ranking in terms of thermal comfort. In the case of open beaches, structures that significantly affected the range of shade during the day were the most crucial factors, with an average contribution of 65.31%, which is the same result as empty public spaces in the city [52]. At night, the influence of structures diminishes, making planting methods that enhance ventilation more critical, contributing over 71% to thermal comfort after being negligible during the day. This also verifies that various thermal environmental factors behave differently during the day and night [53].

The thermal environment of coastal villages at night is only marginally affected by sea–land breezes and does not exhibit significant nocturnal variation, as is commonly reported in most studies [54]. This may be due to insufficient thermal drive and large-scale background wind field suppression. For thermal drives, this is possibly attributed to the fact that most research on the thermal environment of sea–land breezes has been conducted in urban settings rather than in villages [55]. The rise and fall in temperatures in cities due to the heat island effect are relatively significant [56], where the heat island effect is less pronounced and its impact is correspondingly reduced. In addition, the research area is an island near the mainland with a relatively small land area. The land temperature is not enough to produce significant land and sea winds and is suppressed by the large-scale background wind field. The horizontal scale of land wind circulation is limited, and it is easily covered by large-scale flow fields [57]. In contrast, study islands where significant land and sea breezes occur are larger in size [19]. The coastal nighttime of Naozhou Island is often controlled by the southwest monsoon or background easterly winds (e.g., the extension of the South China Sea monsoon trough in summer), which may mask the local land wind signal if the background wind speed is >3 m/s. Figure 13 shows the monthly average nighttime wind rose meteorological data for the two months of our research on the island. The data were obtained from the meteorological station in Zhanjiang, China. The figure illustrates that from 1 July to 31 August, the island experienced winds predominantly from the southeast for a longer period of time, and the wind speed was more than 3 m/s, which weakened the sea and land winds to a greater extent. This also suggests that in the island’s seafront villages, there is little temperature difference between day and night. These insights could provide valuable references for designers considering these aspects in their work.

This paper concludes with the optimal landscape configuration in the sense of thermal comfort, which may not be feasible in actual projects due to economic constraints and maintenance costs. Village space designers need to consider a full-cycle cost–benefit analysis in practice to quantify the cost of maintenance [58] and then integrate these findings with the combinations of the landscape elements in this study for better spatial planning.

5. Conclusions

In this study, spatial genes were extracted from the public spaces of coastal villages and utilized as a foundational model to reconstruct and simulate the thermal environment of typical coastal village public spaces. The impact of landscape configuration on the thermal environment was investigated, and the thermal mitigation capabilities of each landscape element were analyzed to identify differences. The optimal synergistic working methods were then summarized. The main findings of this study can be summarized as follows.

(i) In-depth research on Naozhou Island, coupled with a spatial genetics methodology, facilitated the extraction of elements, structures, patterns, and texts from the island’s spatial types within the public space hierarchy. This approach led to the derivation of three quintessential public spaces on Naozhou Island replete with intricate public space elements: the temple-affiliated plaza in Gangtou Village, a public courtyard in Dedou Village, and an open beach in Cunliang Village.

(ii) The impact of landscape elements on thermal comfort was found to fluctuate over time. The experimental simulation results yielded a hierarchy of the effects of landscape elements on thermal comfort in public spaces in the coastal villages of Naozhou Island.

(iii) Based on the variances in the thermal mitigation abilities of the landscape elements, optimal combinations for the three public space types were deduced through a comprehensive assessment. For the temple-affiliated space, the best landscape element combination was identified as “Ficus spicata”, “3 m”, and “granite + grass”. The public courtyard space type’s ideal combination is “trees + shrubs + grass”, “35% green space ratio”, “columnar + clump planting”, and “granite + grass”. For the open beach space type, the prime combination is “space east (near the sea)”, “tensioned membrane”, “columnar planting + clump planting”, and “granite + grass”. Consequently, this paper proposes thermal comfort strategies for various public spaces in coastal villages.

A suitable range of landscape configurations for coastal villages was established using an orthogonal experimental method for multivariate analysis, achieved through detailed regional research aligned with the functional needs of public spaces and by prioritizing factors that affect the thermal environment in strategic proposals. As shown in Figure 12, the optimized placement of landscape elements can be directly informed by these data after calculating construction and maintenance costs. The insights gained from this study offer valuable references for village development in climatic contexts.