Seasonal Dynamics and Trade-Offs/Synergies of Cultural Ecosystem Services in Urban Parks: A Case Study of Chengdu, China

Abstract

Urban parks provide diverse cultural ecosystem services (CESs), which are crucial for residents’ mental well-being. However, few studies have investigated how urban parks’ CESs and their interactions vary across seasons. In this study, we used the downtown area of Chengdu, China, as a case study, and evaluated urban parks’ CESs based on social media comments and further explored their seasonal dynamics. We then analysed the seasonal trade-offs/synergies of these CESs for service pairs using Pearson correlation and for multiple services using bundle identification. The results show the following: (1) Most CESs except for social interaction had the highest intensities in autumn, and recreational activities and education were the CESs with the highest and lowest intensities among the four seasons, respectively. Education service showed the greatest seasonal variation, while recreational activities and physical and mental recovery were stable among different seasons. (2) Some CES pairs exhibited trade-offs/synergies, but those relationships changed over seasons. Specifically, there were trade-off/synergy relationships between seven CES pairs in spring, three CES pairs in summer and autumn, and four CES pairs in winter. (3) In terms of the trade-offs/synergies among multiple CESs, we identified three types of CES bundles, i.e., physical and mental recovery- and aesthetics-dominated, inspiration- and education-dominated, and social interaction- and recreation-dominated bundles. More than 50% of the urban parks exhibited the physical and mental recovery- and aesthetics-dominated bundle in four seasons, and the seasonal change between this bundle and the social interaction and recreation-dominant bundle was the most obvious among all the bundle changes. This study revealed urban parks’ CES seasonal dynamics and identified the seasonal variations in CES trade-offs/synergies, providing a reference for CES management in urban parks.

1. Introduction

As global urbanisation continues to advance, changing urban environments and urban lifestyles have posed enormous challenges to people’s health. Studies have demonstrated that modern urban life is a main factor for health problems such as anxiety, insomnia, and depression [1]. Cultural ecosystem services (CESs) refer to spiritual and non-material benefits that people obtain from nature, which are crucial for human physical and mental health [2]. As important urban green infrastructure, urban parks provide urban residents with a number of CESs (such as physical and mental recovery, natural recreation, and inspiration) [3,4,5,6,7]. Urban parks are important places that allow people to release psychological pressure and relieve mental tension in their leisure time. In the context of increasingly severe health problems of urban residents, optimising CESs in urban parks has become a key pathway to improving the quality of residents’ lives. The study of urban parks’ CESs is important for improving residents’ well-being and promoting sustainable urban development.

The CES evaluation of urban parks has received significant attention in the field of urban planning. Many researchers analysed CESs using interviews and questionnaire surveys [8,9], while some researchers further explored CES distributions using public participatory GIS [10]. In recent years, social media has gradually become an important data source for CES evaluation. Compared with questionnaires and interviews, CES evaluations based on social media can not only effectively save manpower and material costs but also cover a larger number of urban parks at the same time. For example, Wang et al. collected comments on 50 urban parks in Beijing from a Chinese social media website and created a CES vocabulary, which are used to reveal the spatial pattern of urban parks’ diverse CESs [3]. Other scholars mainly focused on investigating overall public’s perception and satisfaction of urban parks and did not distinguish different CESs [11,12]. Generally, social media data have greatly promoted CES assessment research at the urban and regional scales, and the integration of social media data into CES assessment has become an important current trend.

Exploring the trade-offs/synergies between CESs is a necessary prerequisite for meeting the diverse needs of residents. There may be a trade-off relationship between some CESs, i.e., the increase in one CES at the cost of other CESs, while a synergy can be observed in other CESs, i.e., the simultaneous increase among different CESs [13]. To date, scholars have explored the trade-off/synergy between CESs in two dimensions: between two services (i.e., service pair) and across multiple services. For the CES pair, scholars have often used correlation statistics analysis to reveal the trade-off/synergy between CESs. For example, Ament et al. analysed the CESs perceived by tourists in 19 national parks in South Africa based on the Spearman correlation coefficient [14]. The results revealed that there were trade-off relationships among most CESs, with the strongest trade-off being between natural history and recreation. Moreover, Zhao et al. conducted pairwise correlation analysis between CESs in Wuyishan city and reported that, except for religious value and ecotourism, there were significant trade-offs/synergies among other CESs [15]. In terms of the trade-offs/synergies among multiple CESs, the service bundle was the most frequently used analysis method. An ecosystem service bundle refers to a group of services that occur repeatedly in time or space, which can simultaneously exhibit trade-offs and synergies among multiple services [16]. For example, Cheng et al. identified CES bundles at the urban park scale and proposed that the synergistic relationships among these services can be strengthened by adjusting landscape characteristics [17]. Overall, the elucidation of the trade-offs/synergies between CESs of urban parks requires deeper research on the relationships among ESs at the urban park scale, which not only has important theoretical value but also provides scientific support for park management.

Although the abovementioned studies have provided beneficial explorations for revealing the trade-offs/synergies of CESs in urban parks, they have not considered the potential variations in the relationships between CESs in different seasons. Relevant studies have shown that CESs may exhibit significant seasonal heterogeneity. For example, Guo et al. examined the distributions and levels of multiple CES perceptions in different seasons and found that the distributions and levels were season-dependent [18]. Huang et al. further investigated how the seasonal variation affects residents’ preferences for CESs in parks with different urban thermal environments in Shanghai [19]. The results revealed that people’s perceptions of CESs in spring and autumn were significantly greater than those in summer and winter. Thus, seasonal variation factors need to be integrated into CES research. On the one hand, such analysis can more accurately reveal the trade-offs/synergies between CESs, and on the other hand, it can help refine the management of CESs in urban parks.

To address the deficiencies in current studies, this study has selected 84 urban parks in the downtown area of Chengdu and selected six typical CESs for analysis. This study aims to address three research questions from the perspective of seasonal change: (1) What are the seasonal dynamics of the CES intensity and distribution in urban parks? (2) What trade-off/synergy relationships exist between different CESs (from the dimensions of CES pairs and multiple CESs) in urban parks? (3) What is the seasonal dynamic of the CES trade-off/synergy relationships? This study contributes to the understanding of the seasonal dynamics of CESs and their interactions on the urban park scale and provides a theoretical basis for residents’ well-being-oriented park management and planning.

2. Methodology

2.1. Study Area

Chengdu (102°54′~104°53′ E, 30°05′~31°26′ N) is located in the central part of Sichuan Province, China (Figure 1), and has a mild climate with four distinct seasons. The study area was the downtown area of Chengdu, including 12 administrative districts and 2 economic functional zones, with a total area of approximately 3639 km² (Figure 1). With accelerated urbanisation and population agglomeration in Chengdu, the rapidly changing urban environment has posed a considerable challenge to people’s mental health, and thus, Chengdu urgently needs to explore how to manage urban parks to meet residents’ diverse demands for CESs. Furthermore, the Chinese central government has proposed a novel urban development model named “Park City”, which emphasises the integration of green space systems and park networks into urban planning [20]. Such a model aims to enhance the urban ecological environment and improve residents’ quality of life by managing urban parks and has played an increasingly important role in Chinese urban planning. In 2022, the State Council of China approved Chengdu’s plan to establish the first Park City Demonstration Area in China [21], and thus, Chengdu is a representative study area for studying urban parks’ CESs.

Urban parks in the downtown area of Chengdu were selected as the research objects for this study and were further screened based on the research questions and feasibility. The screening criteria were as follows: (1) This study focused on public parks that are accessible to all residents, so the selected parks have to be open to the public for free. (2) The number of social media comments for the parks was sufficient to support the CES evaluation for multiple seasons; thus, the parks with less than 20 comments for four seasons were excluded. In the end, 84 parks in the downtown area of Chengdu were selected for this research (Figure 1).

2.2. Research Design

This research was conducted in the following steps (Figure 2). In step 1, we selected typical CESs based on CES classification systems and local stakeholders’ experiences. Then we processed data by collecting social media comments, data cleaning, and seasonal counts. After that, we constructed a comment thesaurus reflecting CESs and used the CES intensity proxy to evaluate different CESs. In step 2, Global Moran’s I, Lorenz curve, and Gini coefficient were used to reveal CES distribution dynamics, and seasonal statistics and seasonal concentration degree were used to analyse CES intensity dynamics. In step 3, we analysed CES trade-offs/synergies and their seasonal dynamics for CES pairs and multiple CESs, respectively. Pearson correlation was used to reveal the trade-offs/synergies of CES pairs, and CES bundles were employed to analyse the trade-offs/synergies among multiple CESs. Each analysis technology is detailed in the following subsections.

2.3. Data Source and Preparation

The data used in this study mainly included the park boundary and Dianping website’s comments. The Dianping website is a leading social media site in China that records various types of life information, such as parks and restaurants, and is one of the earliest independent third-party review websites established in China. In this study, the comments of the park on the Dianping website were used as the basic data for quantifying CESs. The advantage of such a data source is that there are no distracting factors such as advertisements in the relevant reviews, and the website has a large number of comments, which can guarantee the number and representativeness of the samples. First, the Python 3.9 tool was used to crawl the comment text data of 84 urban parks from Dianping.com from 2010 to 2023, for a total of 35,328 comments. We selected such a study period for two reasons: (1) Since 2010, the Chinese government has gradually prioritised the role of parks in sustainable urban development. For example, in 2010, the Chinese government has launched an initiative named “Leisure City” in Chengdu [22], which aims to attract more tourists and improve living quality by leveraging urban parks’ CESs. Furthermore, the most recent data available for urban park comments collected from the Dianping website is up to the year 2023. (2) During this period, Chengdu’s population increased from 1.41 million to 21.40 million [23], sharply driving up the demand for CESs provided by urban parks.

Next, the comment data were cleaned by manually removing comments with the same content and blanks and removing meaningless emojis. All validated comments were then categorised by seasonal quarters: (spring: March to May; summer: June to August; autumn: September to November; winter: December to February). The vector data of the park boundary and distribution is from Gaode map, which was also crawled via Python 3.9.

2.4. CES Evaluation Method

2.4.1. Construction of a Structured Text for Urban Park CESs

In this study, we used two commonly accepted CES classification systems, i.e., the Common International Classification of Ecosystem Services and MEA Ecosystem Service Classification, as main references for CES selection. The selection of CESs for analysis should be based on the specific conditions of urban parks in Chengdu, because urban parks may not provide all the CESs listed in those classification systems. For example, prey for hunting or collecting was not allowed in Chengdu’s parks and, thus, was not selected for urban parks’ CES evaluation. Following preliminary CES selection, we further invited relevant stakeholders to judge whether these CESs are important for living quality and mental health via interviews. The stakeholders included twenty Chengdu residents aged 18 to 65, ten landscape architecture experts with a master’s or higher degree, and six urban park managers. Finally, the focused types of CESs provided by urban parks include cultural education, social interaction, aesthetic appreciation, physical and mental recovery, recreation, and inspiration (

Table 1. CESs and their corresponding vocabulary (the vocabulary was determined based on expert judgement and previous studies [3,4,6,17,24,25,26,27,28,29]).

CES Type	Explanation	Representative Vocabulary
Education	Urban parks provide opportunities for the public to obtain knowledge and education.	Science popularisation; Research; Crafts; Teaching; Commentary
Social interaction	Urban parks provide places for residents to socialise and meet their social needs.	Friends; Family; Dinner; Chatting; Singing; Picnic
Aesthetics	Aesthetic information obtained from viewing the interior scenery of the park.	Taking pictures; Beautiful; Scenery; Colourful; Scenery
Physical and mental recovery	Urban parks provide a series of natural and man-made landscapes to meet people’s needs for rest and mental stress reduction.	Comfortable; Calm; Stress-relieving; Refreshed
Recreational activities	Urban parks provide people with the services and resources for leisure, entertainment, and recreation.	Boating, Playground; Leisure; Garden party; Chess game
Inspiration	Artistic inspiration is inspired by the natural or cultural landscape of the park.	Art exhibitions; Places of interest; Romanticism; Museums; A sense of history

).

We then developed the corresponding comment text thesaurus for each CES and identified the types of CESs that appeared in the park comment. The Python-based Jieba package was used to conduct text segmentation of the park comments. After that, the word2vec model was used to vectorize the text information. Next, based on the word segmentation text in the previous step, we conducted word frequency statistics on the comment texts and extracted words related to CESs of urban parks from the high-frequency words to form the essential vocabulary. Additionally, we further used the word2vec module of Python to expand the range of high-frequency words and finally manually classified the expanded words into six CES categories to form the CES thesaurus.

2.4.2. CES Evaluation Index

Based on the constructed thesaurus and park comments, the lexical matching method is used to analyse each comment sample. If the content of a comment matches a word in a CES thesaurus, it is recorded that park users perceive that corresponding type of CES in the park; if there is no matching vocabulary, it indicates that park users do not perceive any CES in the park. This study used the following calculation formula to evaluate the CES intensity for each park from the users’ perception perspective [3]:

$$Pi=Si/Ci$$

(1)

where $Pi$ is the intensity of CES $i$, $Si$ is the number of times a CES is perceived in the park, and $Ci$ represents the total number of comments for the i-th park. $Pi$ was calculated in four different seasons individually.

2.5. Seasonal CES Dynamics

2.5.1. CES Distribution Characteristics

We used the global Moran’s I and the Gini coefficient to analyse the spatial distribution characteristics of different CESs.

(1)

Global Moran’s I

The global Moran’s I was used to analyse the spatial agglomeration characteristics of CESs in urban parks. The global Moran’s I can test the agglomeration effect of attributes in space. The value of the global Moran’s I is between −1 and 1. The closer the index is to 1, the more the CESs are aggregated in space; −1 indicates more scattered CESs in space, and a value equal to 0 indicates that CESs are randomly distributed [30]. Relevant operations were performed via the ArcGIS 10.8 spatial analysis tool.

(2)

Lorenz curve and Gini coefficient

In this study, the Gini coefficient and the Lorenz curve were used to analyse the spatial unevenness of the CES intensity in urban parks. The Lorenz curve can visually represent the cumulative distributions of the area proportions and the CES intensities of different parks. The X-axis shows the proportion of the cumulative park numbers to the total park number, sorted from lowest to highest, and the Y-axis shows the proportion of the cumulative intensity of the CESs of the parks to the total intensity of all the parks. The Gini coefficient is the area ratio between the Lorenz curve and the absolute fairness line [31]. The Gini coefficient is between 0 and 1, with a higher value indicating a more uneven spatial distribution of the perceived intensity of CESs in a park. The calculation method is shown in Equation (2).

$$G=1-\sum _{k=1}^n{(P}_k-P_{k-1}\left)\right(R_k-R_{k-1})$$

(2)

where $n$ is the total number of parks. $P_k$ is the cumulative proportion of the CES intensity for the urban park $k$, and $R_k$ is the cumulative number of the urban park for $k$.

2.5.2. CES Intensity Characteristics

(1)

CES intensity statistics in different seasons

First, we calculated the mean value of each type of CES in all parks in each season. Using spring as the baseline, the ratios of CESs in other seasons to CESs in spring were calculated. The four-season variation map of the CES was drawn to reflect the seasonal variation trends of different CESs.

(2)

Seasonal concentration degree index across the four seasons

We used the seasonal concentration ratio as an important indicator to reflect the degree of time concentration of the samples.

$$S=\sqrt{{\displaystyle \frac{\sum _{i=1}^n{\left(X_i-25\right)}^2}{4}}}$$

(3)

where $S$ is the CES seasonal concentration ratio, and $X_i$ represents the ratio of the CES intensity of urban parks in each season to that of the whole year. In the equation, the value of 25 stands for an assumed ratio where four seasons have a similar CES intensity ratio, and the value of 4 represents the season number. A larger $S$ value reflects that the CES intensity in urban parks is more concentrated in a certain season; in contrast, a smaller $S$ value indicates more uniform CES intensity in different seasons [32].

2.6. Trade-Offs/Synergies of CESs

We used Pearson analysis and ecosystem service bundle analysis to reveal the trade-offs/synergies between CESs in urban parks from the aspects of CES pairs and multiple services.

2.6.1. Pearson Analysis

For this study, we used the Pearson correlation coefficient method to analyse the trade-off synergy of CES pairs. The Pearson correlation coefficient is widely used to measure the degree of correlation between two variables, and it has a wide range of applications and high reliability; thus, it is widely used in trade-off/synergy analysis for any two ecosystem services (service pair).

2.6.2. Ecosystem Service Bundle Analysis

We used different parks in each season as independent samples and used the six items related to the perceived intensity of CESs in the parks as attributes to obtain a total of 336 samples. First, z-score standardisation was performed on the six CESs. The dimensionality of the six indicators was subsequently reduced via the principal component analysis (PCA) method in Python [5]. Two principal components with the highest explanatory effects were ultimately selected as the main bundle factors, and k-means clustering was further used to identify the CES bundle types. The bundle number was determined by the elbow method [3].

3. Results

3.1. CESs and Their Seasonal Variations

3.1.1. Seasonal CES Intensity Dynamic

As shown in Figure 3a, the intensity rank of various CESs was relatively stable during different seasons, with recreation activities and education services being the highest and the lowest among the four seasons, respectively. With the intensity of CESs in spring as the baseline (value = 1), the change trend of intensity of CESs during four seasons was calculated (Figure 3b). Education, inspiration, and physical and mental recovery showed an “increase–increase–decrease” trend from spring to winter. Social interaction showed a “decrease–increase–increase” trend from spring to winter, and recreation activities and aesthetics showed a “decrease–increase–decrease” trend from spring to winter. Autumn was the season when all the CESs (except for social interaction) reached their highest intensities among the four seasons.

Seasonal concentration ratios of CESs are shown in

Table 2. Seasonal concentration index of CES intensities.

Recreational Activities	Education	Physical and Mental Recovery	Aesthetics	Inspiration	Social Interaction
0.39	4.20	0.64	1.21	1.72	0.76

. The seasonal concentration ratios of education, aesthetics, and inspiration were greater than 1, indicating that these three CESs had significant seasonal differences, with the seasonal difference in the intensity of education service being the most significant (seasonal concentration = 4.20). The seasonal distributions of recreational activities and physical and mental recovery among the four seasons were relatively more even.

3.1.2. Spatial Variation Characteristics of CESs

Spatially, the spatial autocorrelation results in

Table 3. Spatial autocorrelation of the CES perceived intensity based on Moran’s I (four seasons).

	Recreational Activities	Education	Physical and Mental Recovery	Aesthetics	Inspiration	Social Interaction
Spring	0.036520	−0.005314	0.062788	0.027328	0.074008 *	0.050884
Summer	0.008186	−0.033007	−0.019778	0.034717	0.058878	0.017375
Fall	−0.037574	−0.031924	−0.007281	−0.004967	0.067429 *	0.022975
Winter	0.000493	0.024646	−0.00282	−0.001377	0.13378 *	−0.002811

* represents statistics significance.

reveal that the spatial distributions of most CESs in urban parks were random, and only the inspiration service showed obvious spatial aggregation. The Moran’s I values for inspiration service were statistically significant (p < 0.05) and positive (>0) across spring, autumn, and winter, indicating that this CES had a significant positive spatial correlation in these seasons; that is, parks with high inspiration tended to spatially aggregate.

As shown in Figure 4, the Lorenz curve and the Gini coefficient revealed the differences in the spatial distributions of the six CESs across the four seasons. The Gini coefficients of recreation activities and physical and mental recovery were the smallest in the four seasons, between 0.075 and 0.121, indicating that these two CESs are evenly distributed in space. In contrast, inspiration and education had higher Gini coefficients in the four seasons (between 0.516 and 0.665), indicating that the spatial distributions of these two CESs were highly uneven. Additionally, the Gini coefficients of social interaction in summer and winter were significantly higher than those in spring and autumn, indicating that the spatial unevenness of this CES was exacerbated in summer and winter.

3.2. CES Trade-Offs/Synergies and Their Seasonal Variation

3.2.1. CES Trade-Offs/Synergies for the Service Pair

According to the correlation analysis of CES pairs (Figure 5), there were significant correlations between seven pairs of CESs in spring, three pairs of CESs in summer, three pairs of CESs in autumn, and four pairs of CESs in winter. In terms of the synergetic relationship, aesthetics and physical and mental recovery consistently showed synergy in all four seasons; inspiration and education had synergy in spring, summer, and autumn; and inspiration and education had synergy in summer and autumn. In terms of the trade-off relationship, education and physical and mental recovery exhibited a significant trade-off in spring, summer, and winter, and inspiration and recreation exhibited a significant trade-off in spring, summer, and winter.

3.2.2. CES Trade-Offs/Synergies for Multiple Services

Three typical CES bundles were identified to indicate CES trade-offs/synergies for multiple services. The spatial distribution and CES composition of each bundle across the four seasons are shown in Figure 6. Each bundle was named based on the dominant CES type.

Type 1 is the physical and mental recovery- and aesthetics-dominated bundle. This type of CES bundle provided high levels of physical and mental recovery and aesthetic services. The representative parks with this bundle included Baihuatan and Donghu Parks (four seasons). These types of urban parks often have good ecological quality, which can increase visitors’ physical and mental recovery and aesthetic appreciation. Among the four seasons, the number of parks with this bundle was the largest, accounting for more than 50% of the total number of parks in the four seasons.

Type 2 is the inspiration- and education-dominated bundle. This bundle presented the highest level of inspiration and education and the lowest level of physical and mental recovery and recreational activities. Among the four seasons, the number of parks corresponding to this bundle was relatively stable, ranging from 7 to 9, accounting for approximately 10% of the total number of parks. The corresponding parks are the parks with specific visiting themes, such as Tea Culture and Chengfei (i.e., the abbreviation of aviation in Chengdu) Parks, which were mainly distributed in the city centre area and the northern part of Qingbaijiang District.

Type 3 is the social interaction- and recreation-dominated bundle. This bundle presented the highest level of social interaction and recreation. The numbers of parks with this bundle representing this CES bundle were 32 (spring), 29 (summer), 25 (autumn), and 35 (winter). This type of park includes comprehensive or recreation theme parks, such as People’s Park and Chengdu Open-Air Music Park and was widely distributed in the study area.

Figure 7 illustrates how the park’s CES bundle changes with seasons. Urban parks may exhibit different CES bundle types in different seasons. The transformation between the physical and mental recovery- and aesthetics-dominated bundle and the social interaction- and recreation-dominated bundle was the most significant. For example, from spring to summer, the CES bundle type of nine parks switched from the social interaction- and recreation-dominated bundle to the physical and mental recovery- and aesthetics-dominant bundle. The parks with the inspiration- and education-orientated bundle were relatively stable, indicating that the dominant CES types of the parks with this bundle type were less affected by seasonal variations. The CES bundle change times for each park were recorded due to the season variation (Figure 7). More than half of the urban parks showed no change in their bundle type (transition frequency = 0), indicating high seasonal stability in the CES composition of many parks. The parks with a CES bundle change frequency of 1 were distributed in the core area of the central urban area, Southern Pidu District, Southern Shuangliu District, Northern Qingbaijiang District, and Central Tianfu New District. The parks with a CES bundle change frequency of 2 were mainly distributed in the core area of the downtown area and west of Wenjiang District, south of Xindu District, and south of Longquanyi District. The only park that exhibited CES bundle change frequency of 3 was the Seaside Park, which was newly built in the central area.

4. Discussion

4.1. CES Changes Among Different Seasons

Our study showed that the average intensity of recreational activities in the four seasons was approximately 2.97, which was far higher than that of other CESs. This result is in agreement with the findings in Shanghai and Beijing, China [3,26], and Rotterdam, the Netherlands [33], indicating the dominance of recreational activities among different CESs provided by urban parks. The average intensity of the education service was the lowest, which was consistent with the study results of Guo et al. and Koh et al. [18,34], reflecting that the education service is commonly less perceived by park users. Seasonal analysis revealed that the seasonal concentration index of the education service was 4.2, indicating that this CES has a significant seasonal effect. Moreover, the education service is mainly concentrated in the autumn. This can be explained by the cultural activities that the parks in China normally hold to celebrate the festivals with cultural and educational significance (such as the Mid-Autumn Festival and the Double Ninth Festival) in autumn. In terms of space, the Gini coefficients of inspiration and education services were greater than 0.5 in all four seasons, indicating that the spatial distribution of these two CESs was severely uneven and may be provided by a small number of parks. The uneven distribution of CESs may lead to inequalities in CESs among residents in the study area.

4.2. Seasonal Variation in the CES Trade-Off

4.2.1. CES Pair Relationship

The study results revealed that the CES pairs that showed significant synergetic relationships mainly included the education–inspiration pair and the physical and mental recovery–aesthetics pair, which was consistent with previous findings [35]. The education–inspiration pair relationship was synergetic in all seasons except winter. This may be because the natural conditions, suitable climate, and rich landscapes during spring to autumn provide a favourable external environment for the synergy of education and inspiration, making people more willing to participate in educational activities in parks, thus stimulating inspiration in the outdoor learning environment, whereas unfavourable climatic conditions in winter weakened this synergetic relationship. For the physical and mental recovery–aesthetics pair, previous studies have shown that natural beauty, such as plant landscapes, can elicit a physiological relaxation response, such as reducing heart rate and blood pressure and achieving emotional relaxation [36]. These aesthetic experiences can further promote physical and mental recovery and help people visually and emotionally relax. The relaxed state can enhance people’s perceptual acuity and enable them to experience aesthetics more deeply. The synergy between the physical and mental recovery and aesthetic services was observed in all seasons except summer. The hot weather in summer causes tourists to obtain aesthetic services through quick visits; however, a short stay in urban parks may not provide tourists with the physical and mental recovery service.

In terms of the trade-offs, the recreation service–inspiration pair exhibited a trade-off relationship in summer, spring, and winter. Such a trade-off was also observed in the previous study conducted by Cheng et al. [17]. This is mainly because excessive recreational activities may destroy the quiet environment and interfere with residents’ meditation in the park. However, autumn is the season when different CESs can be perceived and enjoyed by park users, thus weakening the trade-off relationship between these two CESs. Additionally, spring is the season with the largest number of the CES pairs exhibiting trade-off relationships (including physical and mental recovery–education pair, aesthetics–education pair, inspiration–recreation pair, and inspiration–social pair), indicating that urban managers should develop adaptive strategies that specifically target CES trade-offs in spring.

4.2.2. Relationships Between Multiple CESs

In this study, the ES bundles were used to indicate the relationships between multiple CESs and found that the inspiration- and education-dominated bundle was the most stable among different seasons. This is mainly because, within the study area, the inspiration and education in urban parks were closely related to historical relics and man-made cultural infrastructure, which were less affected by seasonal variation. Additionally, many parks with the physical and mental recovery- and aesthetics-dominated bundle maintained the same CES bundle in all four seasons, such as Tazishan Park and Donghu Park. These parks usually have unique ecological advantages, good ecological quality, and an aesthetic atmosphere. Natural elements such as green plants, lakes, and mountain views are not severely affected by seasonal changes. Specifically, in Chengdu, where the vegetation is evergreen, parks are covered by green space, and landscape layout and infrastructure (such as walkways, seats, sculptures, etc.) are often available year-round. As a result, physical and mental recovery and aesthetic appreciation are stably provided in the long term. For these parks, even though seasonal changes may affect the growth cycle of some plants and the flowering time of some flowers, tourists can still enjoy the atmosphere and visual beauty of being close to nature in different seasons. However, some parks have also switched from the bundle dominated by physical and mental recovery and aesthetics to the bundle dominated by recreational activities and social interaction, such as Shuangliu Central Park and New Jinniu Park. These parks usually have diverse functional areas, such as leisure areas, sports fields, and social spaces. In a certain season, the recreational activities and social interaction may be enhanced and become the dominant CESs.

4.3. Management Implications for Urban Parks

4.3.1. Focus on Maintaining Recreation Activities and Improving Education Service

In terms of the CES type, the service of recreational activities is the basic CES provided by parks, and the intensity of tourists’ perception of recreational activities is relatively high in the four seasons. This indicates that this type of CES is the main CES obtained from parks; thus, urban parks need to continue to provide fine maintenance and management of recreational activities. The intensity of education is low, and thus the education service can be a key improvement target in subsequent park renovations to meet the diverse needs of residents for CESs. Previous studies exhibited that the support provided by facilities, structures, and site paving for natural education cannot be ignored. Thus, creating an effective educational environment in parks requires the following: (1) Developing comprehensive science popularisation facilities, educational media, and interpretation systems in urban parks. (2) Integrating tourism route planning to guide the public in spontaneously acquiring knowledge during visits. In terms of the seasonality of CESs, the CES intensity in other seasons is significantly lower than that in autumn. Thus, seasonally adaptive activities and facilities should be developed according to climatic and botanical features. In spring, managers can organise floral festivals (e.g., spring flower festival) and picnics to enhance park engagement. In summer, the shaded areas with large-canopy trees should be established to attract visitors seeking cool respite. In winter, park managers can provide sun-exposed venues with tea-making facilities for residents to enjoy warmth and leisure.

4.3.2. Promotion of CES Synergy and Alleviation of the CES Trade-Off

This study revealed that there were trade-offs/synergies between CESs in urban parks. For example, aesthetics and physical and mental recovery had significant synergistic effects in all four seasons. This may be because the aesthetics and physical and mental recovery complement each other in terms of the environment, activities, and needs and have an interactive reinforcement effect. However, some ES pairs, such as inspiration–recreational activities, have significant trade-offs in spring, summer, and winter. This may be due to the interference of recreational activities on tourists’ acquisition of inspiration. Therefore, multiple strategies should be implemented to enhance the CES synergy. For instance, creating open rest areas near aesthetically valuable natural landscapes enables tourists to achieve physical and mental recovery through scenic engagement, thus synergistically improving landscape aesthetics and restorative benefits. At the same time, relevant cultural activities, such as art and meditation workshops, can be held, in which tourists can meditate and relax while appreciating works of art (obtaining the education service), thus enhancing the comprehensive experience of the activities.

To reduce the trade-offs between CESs, park managers should first implement rational zoning management according to the main functions of different locations [37]. The zoning management can reduce the mutual interference of different services. For instance, within a given park, zones delivering inspiration services should maintain tranquil environments, while recreational activity zones (e.g., amusement facilities) require strategic separation from other zones to prevent noise interference with adjacent CESs. Secondly, implement time-specific zoning in parks. For instance, designate distinct operating hours for inspiration-focused zones versus recreational activity zones. This staggered scheduling minimises CES trade-offs—designating mornings/evenings as peak periods for inspiration activities while restricting visitor flow in recreational zones during these times.

4.3.3. Classified Management of the Urban Parks with Different CES Bundles

For parks with different CES bundle types, a classified management strategy should be adopted based on the CES characteristics. For the parks with the physical and mental recovery- and aesthetics-dominated bundle, the natural landscape in the park should be maintained while providing a quiet and soothing environment to help tourists relax and enhance their aesthetic experience. For the parks with the inspiration- and education-dominated bundle, the emphasis should be on providing tourists with an inspiring and soothing environment. Activities and facilities about educational experience, knowledge exploration, and inspiration should be encouraged. For the parks with social interaction- and recreation activities-dominated bundle, the focus should be on providing spaces that facilitate socialisation, entertainment, and interaction. These areas must accommodate activities for families, friends, and group gatherings, while prioritising the maintenance of amusement facilities, sports fields, and communal spaces. At the same time, the study results reveal that the dominant CESs in some parks may change in different seasons. For example, from autumn to winter, the transformation from bundle type 1 (dominated by physical and mental recovery and aesthetics) to bundle type 3 (dominated by recreational activities and social interaction) in parks was significant. This transformation reflects that the dominant CES types of parks change seasonally. This suggests that park managers may need to take different measures in a particular park during different seasons.

4.4. Limitations and Future Studies

This study still has several limitations, which should be considered in future research. First, CES evaluation in this study relies on social media comments, which may have data representation bias. Social media users are predominantly young people, while usage is significantly lower among older adults and children. Therefore, the CES results may only reflect the perception of young people, and ignore other population groups that do not use social media. Therefore, we suggest that in future studies researchers can validate social media comment-based CES evaluation with in-person interview surveys. Second, the CES types can vary in different study areas. For example, in some parks, collecting foods can be an important CES that should be considered. Thus, CES selection should carefully consider local stakeholders’ perception. Lastly, the seasonal dynamic of CESs and their interactions observed in this study may not apply to some regions. For example, the four seasons are not clearly defined in the tropics, which may reduce CES seasonal dynamics. Therefore, more case studies especially in the regions within different climate zones are needed.

5. Conclusions

From the season change perspective, this study investigated the seasonal dynamics of CESs and their trade-offs/synergies. The study conclusions include the following: (1) Recreational activities and education were the services with the highest and lowest intensities among the four seasons, respectively. Autumn was the season with the highest intensity for most CESs, and the inspiration and education services exhibited significant spatial non-uniformity. (2) In terms of synergy relationships, inspiration and education exhibited significant synergies in spring, summer, and autumn, and aesthetics and physical and mental recovery exhibited significant synergetic relationships in all four seasons. In terms of trade-off relationships, education and physical and mental recovery had significant trade-offs in spring, summer, and winter, and inspiration and recreational activities had significant trade-offs in spring, summer, and winter. (3) The types of CES bundles in the study area included physical and mental recovery- and aesthetics-dominated bundle, inspiration- and education-dominated bundle, and social interaction- and recreation-dominated bundles. Additionally, parks with the inspiration- and education-dominated bundle were more stable and less susceptible to seasonal changes, while the change from the physical and mental recovery- and aesthetics-dominant bundle to the recreational activities- and social interaction-dominant bundle was notable. The methodology for analysing seasonal dynamics of CESs and their interactions in this study can also be applied to other regions, which can help support urban park management in different parts of the world.