Introduction

Ecosystem services (ES) refer to the benefits that humans derive directly or indirectly from ecosystems (Costanza et al. 1997), which are generally categorized as provisioning, regulating, cultural, and supporting services (Mea 2005). The continuous and stable supply of ES lays the foundation for environmental, economic, and social sustainability (Wu 2013, 2021). An in-depth understanding of the value of ES is conducive to increasing the attention and investment in ecological protection and improving human well-being (HWB) (Daily et al. 2009). Provisioning services (e.g., food, water, and raw materials) are the fundamental source for residents to satisfy their basic material needs (Zhang et al. 2007). Regulating services (e.g., gas regulation, water conservation, and environmental purification) and supporting services (e.g., soil conservation and habitat maintenance) safeguard the environmental conditions for human survival and development (Yee et al. 2021). Cultural services (e.g., recreation and eco-tourism) can provide opportunities for residents to interact with ecosystems for spiritual relaxation and satisfaction, education, and aesthetic experience (Willis 2015). Thus, ES and HWB form a strong interconnection (Carpenter et al. 2009; Qiu et al. 2021). However, the Millennium Ecosystem Assessment (MA) found that 63% of ES have been in serious decline between 1950 and 2000 globally, and will continue to decline sharply over the next 50 years with myriad detrimental effects on the well-being of all humankind and sustainable development (Mea 2005). Enhancing HWB is the ultimate goal of sustainable development, and ES are critical to maintaining and improving HWB (Smith et al. 2013; Wu 2013). Therefore, analyzing the relationships between ES and HWB is not only an important basis for better understanding the coupling mechanism of human–environment systems but also a practical need for promoting regional sustainable development (Bennett et al. 2015; Daw et al. 2016; Liao et al. 2020).

In recent years, the relationships between ES and HWB have become the research frontier and hotspot of modern ecology, geography, and sustainability science. The MA project has conducted the first conceptual framework for the ES-HWB relationships since its inception in 2001. The framework not only provides a scientific basis for ecosystem-related decision-making but also triggers the active attention of scholars and the public from different countries around the world (Zhao and Zhang 2006). Existing studies have been conducted mainly in ecologically fragile areas (e.g., loess hilly areas and mountain-oasis-desert areas in arid regions) (Wei et al. 2018; Liu et al. 2019), poverty-stricken areas (e.g., sub-Saharan Africa and the Wuling-Qinba Mountain poverty-stricken area in Chongqing, China) (Stringer et al. 2012; Sandhu and Sandhu 2014; Li et al. 2017), and specific ecosystems such as forests (Kalaba et al. 2013), grasslands (Dai et al. 2014) and wetlands (Pedersen et al. 2019). Recently, several studies have paid more attention to the rapidly urbanizing areas (Huang et al. 2020; Richards et al. 2022; Li et al. 2023). Previous studies have shown that the relationships between ES and HWB are generally influenced by geographic locations and socioeconomic conditions associated with different levels of urbanization (Cumming et al. 2014; Liu et al. 2022, 2023). However, most of these areas are located in arid and semiarid regions with relatively low socioeconomic status. The regions with better ecological, production, and living conditions are usually undergoing rapid urbanization with pronounced human–environment interactions. Therefore, focusing on ES, HWB and their relationships in rapidly urbanizing regions with better ecological, production, and living conditions is not only an extension of current study areas, but also facilitates validating and enriching the existing findings, which can deepen the understanding of the coupling mechanism of human–environment systems in the process of urbanization (Wu 2022; Fang et al. 2023).

Regarding research scales, the relationships between ES and HWB have been investigated at different spatial scales, including national scales (Santos-Martín et al. 2013; Liu et al. 2022), regional scales (Delgado and Marín 2016; Ciftcioglu 2017), and local scales (Pereira et al. 2005; Abunge et al. 2013). Duraiappah (2011) highlighted the importance of scale and indicated that most of the changes in HWB (e.g., some poverty indicators) occurred at smaller spatial scales, such as the community level. The number of studies at finer administrative scales (e.g., town or village scale) has tended to increase in recent years (Wang et al. 2017; Ren and Zhou 2019), contributing to landscape management and planning. However, most of these studies were conducted at a single scale, lacking a comparative and multiscale understanding of the ES-HWB relationships at finer administrative scales (e.g., the “town-village” scale). A multiscale analysis of the relationships between ES and HWB has been carried out at the “province-prefecture-county” scale in China, clearly indicating that the relationships varied with spatial scales (Liu and Wu 2021). Thus, it is necessary to carry out a multiscale analysis at finer administrative scales to more comprehensively capture the local characteristics of ES, HWB, and their relationships and further verify the scale dependence of the relationships.

In terms of research methods, the estimation of ES mainly includes biophysical analysis (Nelson et al. 2009; Gao et al. 2017) and monetary valuation (Costanza et al. 1997; Xie et al. 2008, 2015b). Both of these have been widely applied at global, national, and regional scales (Xie et al. 2017; Hu et al. 2021). The biophysical analysis of ES can help investigate the mechanism of how one type of ES influences HWB. If different types of ES are considered, the monetary valuation of ES (i.e., ecosystem service values, ESV) is one way to standardize and synthesize them, thus facilitating comparisons among regions. HWB mainly consists of two types: objective well-being (OWB) and subjective well-being (SWB) (Summers et al. 2012). OWB is represented by social or economic well-being (Vemuri and Costanza 2006; Hou et al. 2014), and SWB is based on the subjective perception of stakeholders (Wang et al. 2017; Huang et al. 2020). Compared with the OWB indicators (e.g., the Human Development Index, HDI), the SWB indicators focus more on residents' perceptions of their own situation and the external environment (Summers et al. 2012) and are more instructive for testing the rationality of policies and strategies for sustainable development. This study focuses on ESV, SWB, and their relationships, which can not only reflect the traditional perspective of economic value outputs of ecosystems while taking into account the perceptions of well-being but also provide a scientific basis for the practice of ecological value-added or compensation (Hernández‐Blanco et al. 2022).

The Yangtze River Delta region (YRD) is one of the regions with the most active economic development, the highest degree of openness, and the strongest innovation capability in China. The integrated development of the YRD was elevated to a national strategy in November 2018, and the Yangtze River Delta Ecological Greening Development Demonstration Area was designated to practicing the concept of ecological civilization, enhancing HWB, and providing a reference for the YRD as well as other cross-regional management. The pilot demonstration area is located in the core area of the Yangtze River Delta Ecological Greening Development Demonstration Area. This area is an important water conservation area and is undergoing rapid urbanization and urban–rural integration processes, with intense interaction between humans and the environment (Xia et al. 2024). Therefore, it provides an ideal place to study the relationships between ES and HWB at multi-finer administrative scales for improving landscape sustainability across urban and rural regions.

To this end, this study takes the pilot demonstration area of the YRD in China as the study area and carries out a multiscale analysis of ESV, SWB, and the ESV-SWB relationships at the town and village scales. The study aims to answer the following two research questions: (1) What are the spatial characteristics of ESV and SWB in the study area? and (2) How is SWB related to ESV at the town and village scales?

Study area and data sources

Study area

The pilot demonstration area is located at the junction of three province-level administrative units of China (i.e., Shanghai, Jiangsu, and Zhejiang) (Fig. 1a). The region consists of five towns, including two towns of Zhujiajiao and Jinze in Shanghai, one town of Lili in Jiangsu, and two towns of Xitang and Yaozhuang in Zhejiang, with a total area of about 660 km2 (Fig. 1b). The area has a typical subtropical monsoon climate, with a large number of rivers and lakes, such as Dianshan Lake in the south serving as the main waterway connecting southern Jiangsu Province and downtown Shanghai. It is also the main water source of Shanghai and has various ecological potentials. There are different land use/cover types, with cropland, water, and woodland accounting for 39.2%, 32.2%, and 6.8% of the total area of the region in 2020, respectively (Fig. 1b). According to the statistical yearbooks, the study area has a leading position in socioeconomic development in Chinese towns and villages, with a population density of about 650 people/km2. The GDP per capita of the region reached 16,116 USD, more than 1.5 times the national average in 2020.

Fig. 1
figure 1

Map of the study area: a Location, b land use and land cover pattern in 2020. Sample points indicate the approximate locations of the questionnaire survey, including six communities (i.e., the administrative scale equivalent to the village) and nine villages. ZJJ refers to Zhujiajiao Community (consisting of four communities of Beidaxin, Daxinjie, Xihuxincun, and Dongdamen); XC refers to Xicen Community; LL refers to Lili Community (including two communities of Lixin and Xingli); LX refers to Luxu Community (including two communities of Zhendong and Zhenxi); YZ refers to Yaozhuang Community; CND refers to Chaonandai Community; SC refers to Shuichan Village; XJ refers to Xuejian Village; ZJG refers to Zhoujiagang Village; LH refers to Lianhu Village (incorporating Liansheng Community); JZ refers to Jinze Village; YD refers to Yuandang Village; DL refers to Donglian Village; CN refers to Cuinan Village; and HL refers to Hualian Village. More information about sample points is shown in Table S3

Full size image

In China, the administrative divisions generally consist of five levels: provincial, prefectural, county, township, and village levels. The five levels formed a spatially nested administrative hierarchy as a latter level belongs exclusively to a former level (Ma et al. 2016). Liu and Wu (2021) investigated the ES-HWB relationships in China on three administrative levels (i.e., provincial, prefectural, and county levels). Our study focused on the two finer administrative levels of town and village, with a village being part of its corresponding town. The village level, also known as a basic level of autonomy or a fundamental organizational unit, is divided into urban communities and rural villages according to the urbanization level (Tian 2020; Sun et al. 2024). The boundaries of town and community/village scales in our study area are shown in Fig. 1b.

Data sources

Land use/cover data with a spatial resolution of 30 m × 30 m in 2020 for the study area and the administrative boundary data were obtained from the Resource and Environmental Science and Data Center of the Chinese Academy of Sciences (http://www.resdc.cn/). The land use/cover data in 2020 were generated by artificial visual interpretation based on Landsat remote sensing imagery (Xu et al. 2018). High-resolution Google Earth images and sampling surveys in fields were used to verify the classification accuracy of land use/cover data in our study area. The overall accuracy of the land use data was 87.14%, with an average Kappa coefficient of 0.81. Both user and producer accuracy were above 80%. Thus, the land use data in our study area had relatively high accuracy and can be used for further research. The study area contained five primary land use/cover categories and twelve secondary land use/cover categories: (1) cropland, including paddy fields and dry land; (2) woodland, including forest, shrub, and others; (3) grassland, including dense grass; (4) water body, including rivers, lakes, reservoir and ponds, and others; and (5) built-up land, including urban land, rural settlements, and other built-up land.

Socioeconomic data on grain production, grain output value, and sown area were obtained from statistical yearbooks of the corresponding districts, counties, and towns in the study area. Currency exchange rate data were obtained from the China Foreign Exchange Trade System (https://iftp.chinamoney.com.cn/english/). We collected SWB data, including individual socioeconomic characteristics of respondents and their satisfaction scores for different SWB indicators, by face-to-face questionnaires conducted from July to August 2019.

Methods

Quantification of ESV

Estimation of the ESV

The ESV estimation can provide information on the relative scarcity of ecosystems, reflect the social costs of environmental degradation, and form the basis of decision-making for ecological compensation (Howarth and Farber 2002). Costanza et al. (1997) proposed a method for assessing ESV globally for the first time. Based on this methodology, Xie et al. (2008) took into account the ecological and socioeconomic characteristics of China and carried out an expert knowledge-based assessment of ESV (named the equivalent factor method) with about 700 Chinese ecologists involved in structured questionnaires. This method was more suitable for estimating the ESV in China and has been widely applied in different regions of China (Liu et al. 2014; Hu et al. 2021; Yang et al. 2021).

According to the existing studies (Xie et al. 2008, 2017; Liu et al. 2014), we distinguished four primary ES categories (including provisioning, regulating, supporting, and cultural ES) and nine secondary ES categories (including food supply, raw material supply, gas regulation, climate regulation, environment purification, hydrologic regulation, maintenance of soil fertility, biodiversity, and landscape aesthetics). This study used the equivalent factor method to calculate their corresponding ESV with the equivalent value per unit area (i.e., the equivalent coefficient) of each type of ES for different land use/cover types shown in Table S1. Referring to the equivalent coefficients used in China (Xie et al. 2017) and in the YRD (Liu et al. 2014; Zhu and Zhong 2019), we adjusted the equivalent coefficients for different land-use types according to the actual situation of the study area. The economic value of the standard equivalent factor was equal to 1/7 of the market value of the average grain price per unit area in the current year. Based on the statistical data of grain production, grain output value, and sown area, we calculated the standard equivalent factor for each town of the study area (Table S2), and then estimated the ESV at the town and village scales using the following formula:

$${ESV}_{i}=\sum_{j=1}^{n}{EC}_{ij}\times {SF}_{i}\times {N}_{ij}\times PS$$
(1)
$$ESV=\sum_{i=1}^{m}{ESV}_{i}$$
(2)

where \(ESV\) is the total value of ES in the study area; \({ESV}_{i}\) is the ESV of i town/village; \({EC}_{ij}\) is the equivalent coefficient for j land use/cover type of i town/village; \({SF}_{i}\) is the standard equivalent factor of i town/village; \({N}_{ij}\) is the number of pixels for j land use/cover type of i town/village; \(PS\) is pixel size; m and n are the number of towns (i.e., 5) and secondary land use/cover types considered in this study, respectively. The value for each ES primary indicator equals the sum of its corresponding secondary indicators.

Sensitivity analysis

Considering the uncertainty of equivalent coefficients for different land use/cover types, this study used the coefficient of sensitivity (CS) to determine the percentage change in ESV for a given percentage change in an equivalent coefficient. Accordingly, the extent to which the ESV depends on the change in an equivalent coefficient was quantified with the following formula:

$${CS}_{j}=\left|\frac{({ESV}_{aj}-{ESV}_{oj})/{ESV}_{oj}}{({EC}_{aj}-{EC}_{oj})/{EC}_{oj}}\right|$$
(3)

where \({CS}_{j}\) is the coefficient of sensitivity of j land use/cover type; \({ESV}_{aj}\) represents the ESV of j land use/cover type calculated using the adjusted equivalent coefficient (\({EC}_{a}\)); \({ESV}_{oj}\) is the original ESV of j land use/cover type calculated using the original equivalent coefficient (\({EC}_{oj}\)). If CS is greater than 1, the ESV can be regarded as elastic with low credibility, and it is recommended to adjust the equivalent coefficient. If CS is lower than 1, the ESV is considered to be inelastic and the result of the ESV is credible (Kreuter et al. 2001; Hu et al. 2021).

The CS was calculated after adjusting the equivalent coefficients upward and downward by 50%, respectively. The values of CS were all less than 1, indicating that the ESV was inelastic to the equivalent coefficient and our results were reliable (Table 1).

Table 1 Changes in ESV after adjusting the equivalent coefficients (CS)
Full size table

SWB assessment

Selection of the SWB indicators

The conceptual framework of the contribution of ES to HWB was first introduced in the MA. The framework of MA has been widely used in previous studies on HWB assessments (Smith et al. 2013; Wang et al. 2017) and the ES-HWB relationships (Ciftcioglu 2017; Wei et al. 2018; Huang et al. 2020).

Based on the five dimensions of HWB adopted by MA, this study identified five primary SWB indicators, including the basic materials for a good life, health, security, good social relations, and freedom of choice and action. The selection of secondary SWB indicators mainly took into account the socioeconomic situation of the study area and the United Nations Sustainable Development Goals (SDGs) closely related to SWB (e.g., SDG 3—Good health and well-being and SDG 11—Sustainable cities and communities). Based on the indicators commonly used in existing studies on the quantification of SWB (Ciftcioglu 2017; Wang et al. 2017) and data availability, we selected a total of 11 secondary SWB indicators to establish the SWB assessment framework in this study (Table 2).

Table 2 The set of 11 items related to the five dimensions of SWB used in this study
Full size table

Questionnaire survey

We selected 15 villages/communities (i.e., basic level autonomies at the village level in China) of the study area to conduct face-to-face interviews (Fig. 1). The content of the questionnaire involved: (1) basic social-demographic information of the respondents, including gender, household registration, age, and income; and (2) satisfaction ratings of the respondents on 11 secondary SWB indicators corresponding to the five dimensions. Satisfaction scores were quantified using a 5-point Likert scale (Likert 1932), with 1 representing the lowest score (very dissatisfied) and 5 representing the highest score (very satisfied) (Bryce et al. 2016). The Cronbach’s α of the questionnaire was 0.895, indicating a relatively high internal consistency of the questionnaire (Creswell 2002). 413 valid questionnaires were finally obtained, with a 92.6% validity. The basic information of the participants was shown in Table 3.

Table 3 Characteristics of the respondents in the study area
Full size table

Analysis of SWB

Respondents gave a satisfaction score (ranging from 1 to 5) for each SWB secondary indicator during the questionnaire survey. The satisfaction score for each SWB secondary indicator in a village or town was obtained by averaging satisfaction scores by its respondents, and the satisfaction score for each SWB primary indicator was equivalent to the average score of its corresponding secondary indicators. Previous studies usually determined the weights for different SWB indicators by their importance scores recorded by respondents in the interviews (Wang et al. 2017; Huang et al. 2020). In our questionnaire survey, most respondents gave a similar score on the importance of the 11 secondary SWB indicators, and thus we adopted an equal weight method to calculate the score of the composite SWB. Meanwhile, considering that communities were dominated by the urban population while villages were dominated by the rural population, this study quantified the SWB level of communities and villages separately and made a comparison. For consistency, this study also compared the ESV between communities and villages.

Quantification of the ESV-SWB relationships

Both the Spearman correlation coefficient and the Pearson correlation coefficient have been widely used in analyzing the relationships between ES and SWB (Ren and Zhou 2019; Yang et al. 2019). The Pearson correlation coefficient focuses more on the linear correlation between variables, while the Spearman correlation coefficient utilizes the rank order of two variables for correlation analysis and does not require the distribution of the original variables (Liu et al. 2023). Therefore, this study chose the Spearman correlation coefficient to quantify the correlation between different ESV and SWB indicators at the town and village scales, and then compared their strengths and directions between scales. Specifically, the ESV per unit area and the average satisfaction score of SWB were first calculated for each primary and secondary indicator at the town and village scales in the study area, respectively. Then the correlations between different indicators of ESV and SWB were calculated by Spearman’s coefficient at two scales. The correlation analysis and mapping were done using the corrplot package in R.

Results

ESV at the town and village scales

The total ESV in the study area was 1.00 billion USD, and the ESV per unit area was 15,202.90 USD/ha. There was an obvious difference in the total ESV for different ES types (Fig. 2j), with the highest value for regulating services at 0.86 billion USD, followed by supporting services (79.57 million USD), cultural services (35.22 million USD), and provisioning services (29.42 million USD) (Table S4). In terms of secondary ES indicators, hydrologic regulation had the highest value of 0.72 billion USD, accounting for 72% of the total ESV in the study area. The remaining secondary indicators had the ESV ranging from 5.80 million USD to 57.54 million USD, accounting for about 1 to 6% of the total ESV (Fig. 2j).

Fig. 2
figure 2

Spatial distribution of the ESV for 11 secondary ES indicators (a–i) and their percentage of the total ESV (j) in the study area

Full size image

The discrepancies in the total ESV and the ESV per unit area were also pronounced at the town scale (Fig. 3a). The Lili town had the highest total ESV (0.38 billion USD) and the Jinze town had the highest ESV per unit area (25,289.86 USD/ha), which were nearly 9 times and 5 times the corresponding values in the Xitang town, respectively. Regarding the secondary ES indicators, the Lili town had the highest values for all the secondary ES indicators among the five towns (Table S6), and the Xitang town had the highest value of food production per unit area at 365.54 USD/ha.

Fig. 3
figure 3

The total ESV and the ESV per unit area at the town scale (a) and their comparisons between communities and villages at the village scale (b). The dashed line (b) indicates the mean values of the total ESV and the ESV per unit area for communities and villages. Please see the annotations in Fig. 1 for the abbreviations on the x-axis

Full size image

At the village scale, the ESV varied between communities and villages. Both the averages of the total ESV and the ESV per unit area for communities (i.e., 2.79 million USD and 5362.32 USD/ha) were lower than those for villages (i.e., 5.67 million USD and 10,724.64 USD/ha) (Fig. 3b, Table S5). The total ESV and the ESV per unit area in the YD village were the highest, which were 21.16 million USD and 24,318.84 USD/ha, respectively. For the secondary ES indicators, the total value of environment purification and hydrologic regulation as well as their ESV per unit area in the YD village were the highest as well, with the lowest values found in the CDD community (i.e., 33,420.29 USD and 37.93 USD/ha).

SWB at the town and village scales

The satisfaction score of the composite SWB in the study area was 4.12. Among the five dimensions of well-being, good social relations had the highest satisfaction score of 4.46, while freedom of choice and action had the lowest score of 3.77 (Table 4). For the secondary SWB indicators, the higher scores corresponded to family relationships (4.53), neighborhood relationships (4.39), and law and order situation (4.36), while the relatively low scores were found for income, leisure and recreation, and employment and work situation, all of which scored less than 4 (Table 4).

Table 4 The satisfaction score of SWB in the study area
Full size table

There were differences in the level of SWB among the five towns (Fig. 4a). The Lili town had the highest score of the composite SWB (4.31), while the towns of Zhujiajiao and Yaozhuang had a relatively low score of 3.97. Among the secondary SWB indicators, family relations were the highest-scoring indicator for all five towns. Food and water supply was rated the number two with a score of 4.27 for the Jinze town. Residents in the Yaozhuang town rated public security higher, with a score of 4.39. However, the satisfaction of income received the lowest scores for all the five towns (Table S6).

Fig. 4
figure 4

Satisfaction score of each primary SWB indicator at the town scale (a) and that at the village scale (b) as well as the scores of the composite SWB for different communities and villages (c). The abbreviations at the village scale are shown in Fig. 1

Full size image

At the village scale, differences also existed in the level of SWB between communities and villages. The satisfaction score for each primary SWB indicator was higher for communities than for villages (Fig. 4b). Whereas among the secondary SWB indicators, environmental quality was the only indicator with the satisfaction score for villages higher than that for communities. The highest score of the composite SWB was found in the DL village (Fig. 4c), especially in the satisfaction of two secondary SWB indicators (i.e., family relations and public security) with their scores reaching 4.94 and 4.88, respectively. In comparison, the ZJG village had the lowest score of 3.60, with the satisfaction scores of income and mental health downward to 2.83 and 3.33, respectively.

The relationships between ESV and SWB at the town and village scales

In terms of the primary indicators of ESV and SWB, there was no significant correlation between the two at both the town and village scales (Tables S7, S8). Therefore, this section focused on the relationships between the secondary indicators of ESV and SWB.

At the town scale, there were a total of 19 groups between ESV and SWB being positively correlated (p < 0.05). For example, the satisfaction scores of food and water supply, mental health, and leisure and recreation were positively correlated with the values of five ESV indicators (i.e. raw material supply, environment purification, hydrologic regulation, biodiversity, and landscape aesthetics) at the 0.05 significance level. In addition, the satisfaction score of environmental quality was positively correlated with the values of two ESV indicators (i.e., food production and gas regulation) at the 0.05 significance level (Fig. 5a, Table S9).

Fig. 5
figure 5

Spearman correlation coefficients between ESV and SWB at the town scale (a) and at the village scale (b). The red box indicates a 0.05 level of significance with a thin line and a 0.01 level of significance with a thick line

Full size image

At the village scale, the strength of the correlations weakened, with only 4 groups between ESV and SWB showing significant correlations (p < 0.05), and the direction of the correlations shifted from positive at the town scale to negative at the village scale. The correlation coefficients between the satisfaction scores of two SWB indicators (i.e., employment and work environment as well as leisure and recreation) and the values of two ESV indicators (i.e., raw material production and climate regulation) were all about − 0.5 (Fig. 5b, Table S10).

Discussion

What are the regional characteristics of ESV in the study area?

We found that the study area had a relatively higher ESV per unit area but a relatively lower ESV per capita, in contrast to other regions of China. The ESV per unit area in the study area was considerable at about 15,202.90 USD/ha in 2020, which was nearly three times the Chinese average (5753.62 USD/ha) (Xie et al. 2015a) and even four times the ESV per unit area (3985.51 USD/ha) in the Beijing-Tianjin-Hebei region (i.e., one of the three largest national-level urban agglomerations in China) (Li et al. 2022). The ESV per unit area in the study area was still higher than that in the YRD which the study area belongs to. The ESV per unit area in the YRD (including four province-level administrative units of Jiangsu, Zhejiang, Anhui, and Shanghai) in 2015 was about 1753.62 USD/ha, less than 1/9 of the value of the study area in 2020 (Zhu and Zhong 2019). The relatively higher ESV per unit area is mainly because the study area has favorable climate and hydrological conditions with flat land, fertile soil, and numerous rivers and lakes (Ma et al. 2022). The subtropical monsoon climate not only provides abundant water for rivers and lakes but also brings appropriate heat for vegetation growth, which accelerates food production as well as the supply of other types of ES. For example, the study area had a relatively high value of food production, with an average grain price of 0.41 USD per kilogram and a standard unit equivalent factor of 454.06 USD/ha in 2019. In comparison, the standard unit equivalent factor was 352.99 USD/ha in the Beijing-Tianjin-Hebei region in 2019 (Li et al. 2022) and was only 251.91 USD/ha in the YRD in 2010 (Liu et al. 2014).

However, the bottleneck in the study area was the relatively lower ESV per capita (322.11 USD per person in 2020), which was less than 1/9 of the average of 2898.55 USD per person in China (Xie et al. 2017) and even slightly lower than the ESV per capita in other metropolitan areas (e.g., 512.07 USD per person for a village in the Xi’an metropolitan area, China) (Ren and Zhou 2019). The relatively lower ESV per capita is mainly due to the dense population in the study area. The population of the YRD increased by 6.40 million from 2015 to 2019, with a population growth rate of 2.90% and a population density of 657 people/km2 in 2019, which was far higher than the national level of 147 people/km2. The imbalance between supply and demand of ES has been found to increase in the YRD caused by rapid urban expansion and a large loss of cultivated lands (Tao et al. 2018). This is an important issue to be resolved for regional sustainability (Baró et al. 2015; Shou et al. 2020).

To better realize regional sustainable development, Tao et al. (2022) proposed that urban encroachment onto cultivated and ecological lands needs to be strictly limited. The Territorial Spatial Master Plan for the Yangtze River Delta Eco-Green Integrated Development Demonstration Zone (2019–2035) explicitly states that cultivated lands in our study area should be retained at a minimum of 165 square kilometers, and the construction lands can not exceed the existing total scale. Moreover, increasing green infrastructure (i.e., the interconnected network of greenspaces) and determining proper population size, growth, and distribution might also make potential benefits to help enhance the ESV per capita (Tang et al. 2016; Fang et al. 2023).

What are the regional characteristics of SWB in the study area?

The satisfaction score of the composite SWB in the study area reached 4.12 on a 5-point scale, which was higher than that in other regions of China (e.g., Huailai Mountain basin or Baiyangdian watershed located in the Beijing-Tianjin-Hebei region) (Wang et al. 2017; Huang et al. 2020). This suggests that respondents in the study area were generally satisfied with the current living conditions. The three primary SWB indicators (i.e., health, safety, and good social relations) got relatively high scores at 4.20 or more, indicating that respondents were more satisfied with individual physical and mental health as well as public security and environmental safety. However, other regions had relatively low subjective satisfaction with health and safety, with scores of less than 3.5 for health and less than 3.7 for safety in the Huailai mountain basin (Wang et al. 2017) and the Baiyangdian watershed (Huang et al. 2020). The positive feedback on SWB in the study area may be attributed to its high-level socio-economic development and the relatively sound and complete social security system. According to the local statistical yearbooks, the per capita disposable income in the study area reached 8242.03 USD in 2019, which was nearly two times the national average of 4665.07 USD in China; and the number of health technicians per 1000 population was 15.21, which was also higher than the national average of 13.81. By contrast, respondents were less satisfied with freedom of choice and action with a score of only 3.77. During the questionnaire survey, respondents showed a stronger preference for traveling, indicating the growing demand for a better life and the need for higher levels of spirituality and recreation.

How is SWB related to ESV at the town and village scales?

Our study showed that the strength of the ESV-SWB relationships decreased from the town to village scales, and the positive correlations at the town scale changed to non-existent relationships at the village scale (Fig. 5a and b). Previous studies have also shown that the ES-HWB relationships can be positive, negative, or non-existent (Delgado and Marín 2016; Wei et al. 2018; Liu and Wu 2021). At the town scale, we found that significant positive correlations existed between the values of seven secondary ESV indicators (e.g., food production and hydrologic regulation) and four secondary SWB indicators (e.g., the satisfaction of food and water supply and that of leisure and recreation) in our study area. The significant positive correlations at the town scale may be related to the government’s promotion of agricultural green production and the collaborative management of cross-boundary water bodies.

Cultivated lands account for about 40% of the total area of our study area (Fig. 1b). The scattered agricultural land has been gradually integrated into large-scale and high-standard farmlands under the guidance of town governments (Liu et al. 2018). Therefore, it is more efficient to enhance the land utilization ratio and excavate land use potentials (Long and Qu 2018), which may improve food production and then meet the basic material needs of residents at the town scale. Meanwhile, the study area is rich in water resources with numerous lakes and rivers accounting for more than 30% of the total area (Fig. 1b). The collaborative management of cross-boundary water bodies (e.g., Dianshan Lake and Yuandang Lake) is conducive to the improvement of the water quality and has made the two lakes reach the water quality targets for the year 2025 in advance, which helps enhance the environmental safety and well-being of residents. Except for safeguarding water quality, the study area has built different forms of tourism landscapes based on the resource endowment of rivers and lakes, which may not only improve the aesthetic value of landscapes to a certain extent but also increase the satisfaction of residents for leisure, recreation, and tourism.

In comparison, the ESV-SWB relationships at the village scale turned negative or uncorrelated. This may be due to the differences in the implementation of town planning for improving ES or HWB at the village scale and the impacts of different gradients of urbanization (Wang et al. 2019). Additionally, local non-ES (e.g., technology and innovation, convenient transportation, educational resources) and remote ES through trading or biophysical flow can also weaken the dependence of HWB on local ES (Liu and Wu 2021; Yang et al. 2023). Specifically, significant correlations existed only between the values of two secondary ESV indicators (i.e., raw material supply and gas regulation) and two secondary SWB indicators (i.e., employment and work environment as well as leisure and recreation), and the direction of the correlations changed from positive to negative (Fig. 5a and b). It suggests that residents' employment and recreational well-being did not increase with the values of provisioning and regulating services.

This is similar to the case in Zhejiang Province where farmers experienced a decline in well-being after land acquisition (Li et al. 2015). One possible reason is the time and cost of farmers' job switching. Middle-aged and elderly farmers have limited competitiveness in the job market and take longer to switch jobs. The other reason may be that most of middle-aged and elderly farmers are rooted in rural life and farming practices, have a deep affection for agricultural land, and thus hold a more conservative or rejecting attitude towards agricultural land acquisition or transfer (Raudsepp-Hearne et al. 2010). In addition, respondents claimed that tourism development had some negative impacts on local transportation and sanitary conditions. Therefore, it is necessary to effectively integrate tourism-related livelihoods into the traditional livelihoods of residents to promote the convergence of culture and tourism when improving landscape-specific ES. At the same time, ecological compensation and employment assistance programs (with a particular focus on vulnerable groups such as the elderly) should be carried out to enhance social well-being (Dong et al. 2015).

Therefore, the spatial scale dependence not only exist at the province-prefecture-county scale (Liu and Wu 2021), but also make sense at finer administrative scales such as the town-village scale in our study, implying the necessity of researching the relationships at multiple and broader scales. Moreover, we found that the ESV-SWB relationships varied not only with spatial scale, but also with thematic scales (e.g., primary and secondary categories). ESV and SWB tended to be more significantly correlated based on secondary categories than primary categories. For example, there was no significant correlation between the values of provisioning services and the satisfaction scores of freedom of choice and action at the town scale (Table S7), but when moving to the secondary indicators, there was a positive correlation between the value of raw material supply and the satisfaction score of leisure and recreation at the 0.05 significance level (Fig. 5a). Therefore, it is necessary to be cautious about the use of integrative indicators when carrying out the analysis of the ESV-SWB relationships to avoid unintended results due to the inappropriate selection of thematic scales.

Limitations and future directions

This study used the equivalent factor method to estimate the monetary value of ES. This method accumulates standardized data of different types of ES for inter-regional comparison. The values of ES can be further compared with biophysical quantities and the subjective perception of ES for better understanding ES characteristics from different perspectives. Meanwhile, future studies can supplement objective well-being data, such as social and economic well-being (e.g., HDI), and combine them with subjective well-being data to comprehensively assess the level of HWB. We chose the indicators of ES and HWB commonly used in recent studies (Liu et al. 2014; Xie et al. 2015a; Ciftcioglu 2017; Wang et al. 2017), which facilitates a comparative analysis with existing findings. However, previous studies also showed that the relationships between ES and HWB may vary with indicators or variables (King et al. 2014; Liu and Wu 2021). Thus, it is necessary to select different indicators to test our findings in future research. Meanwhile, we only explored the correlations between ES and HWB, and it remains to be further verified whether there are causal relationships between ES and HWB. In terms of data, the calculations of the ESV, SWB, and their relationships at the village scale were based on the administrative boundary data, but this data suffered from poor timeliness, which may not fully reflect the current situation. The data should be further updated in the future by combining it with field surveys. In addition, this study only explored the relationships for a single year, and the temporal dynamics of the relationships have been not well understood. Furthermore, different urbanization levels may influence the ES-HWB relationship (Xia et al. 2024). Therefore, there is a need to quantify the long-term dynamics of ES and HWB and their relationships under different levels of urbanization to deepen the understanding of the coupling mechanism of human–environment systems.

This study focuses on the ES-HWB relationship, which is a prevalent theme for biodiversity conservation and sustainable development (Sandifer et al. 2015; Naeem et al. 2016), and one of the core questions in landscape sustainability science (Wu 2021). The study area serves as a typical urban–rural integration system (Zhao and Jiang 2022), and thus provides an empirical case for an in-depth exploration of landscape sustainability in urban and rural areas. Our findings demonstrate that the spatial and thematic scale dependence of the ES-HWB relationship still exists at the town and village scales. The findings are helpful for better understanding the nature-society relationship in changing landscapes, which is a research goal of landscape sustainability science (Wu 2013). Future studies should trace the tradeoffs of ES with human outcomes in the coupled human–environment system (Turner 2010) and explore how the relationships can be integrated into decision-making to improve landscape sustainability.

Conclusions

The relationship between ES and HWB is complex and a core topic in landscape sustainability science. This study systematically examined the relationships between 9 ES and 11 HWB measures at the town and village scales (i.e., the fourth and fifth levels in the administrative hierarchy of China). The results show that the ES-HWB relationships varied with spatial and thematic scales. The secondary indicators of ES and HWB tended to be more significantly correlated than their primary indicators. Based on the secondary indicators, most ESV were significantly and positively correlated with subjective HWB at the town scale, but this correlation shifted to uncorrelated and weakly negative correlations at the village scale. Our study suggests that ES-HWB relationships may also vary unpredictably at relatively fine administrative scales. A better understanding of landscape sustainability science and improvement of the ES-HWB relationships requires investigating at multiple and broader scales and choosing appropriate thematic scales.