Assessment of multiple ecosystem services and interrelation-ships in Forest-rich regions: a case study of Guilin, China

Ecosystem services (ES) are fundamental to achieving human well-being. To systematically evaluate ES and their interrelationships in forest-enriched regions, this study takes Guilin, a representative forest-rich area in China, as the study area and develops an integrated coupling model for the comprehensive measurement and interaction analysis of provisioning, regulating, and cultural ES. The InVEST model was used to quantify provisioning and regulating ES. A discrete choice experiment (DCE) was applied to estimate the economic value of cultural ecosystem services (CES), and participatory geographic information systems (PPGIS) were employed to assess its spatial distribution. Based on these results, a spatial analytic geometry model was introduced to calculate the coordination deviation degree (CDD) and coordination transformation degree (CTD), which, respectively, characterize the synergy/trade-off relationships and the comprehensive development level among the three ES types. Finally, a spatial econometric model was used to identify the main influencing factors. The results show that provisioning ES in Guilin are generally at a medium level, with higher values in the southern and central regions and lower levels in the north; regulating ES are relatively high, exhibiting higher levels in the east and west but lower values in the central area; the economic value of CES varies considerably across space, forming a distinct spatial gradient. At both municipal and county levels, the coordination among the three types of ES remains relatively high, while the comprehensive development level is comparatively low. Further analysis reveals that forest coverage and GDP exert significant positive effects on ES, whereas temperature, terrain, population density, and industrial structure have negative effects. This study provides a feasible analytical framework for CES research, effectively captures the interrelationships among provisioning, regulating, and cultural ES, and offers valuable insights for forest ecosystem management.

Ecosystem Services (ES) collectively refer to the products and benefits that humans obtain from ecosystems through direct or indirect interactions, serving as the foundation for human survival and development (Costanza et al., 2017). The United Nations Millennium Ecosystem Assessment (MA) classifies them into provisioning, regulating, supporting, and cultural services (Corvalan et al., 2005). The Common International Classification of Ecosystem Services (CICES) further refined this framework by merging regulating and supporting services into a single cate-gory and classifying ES into three main types—provisioning, regulating, and cultural—effectively avoiding double counting that may arise from the overlap between regulating and supporting services (Haines-Young and Potschin, 2018). Provisioning ES refer to the products and services directly provided to humans through the biological production processes of ecosystems. Regulating ES reflect the ways in which ecosystems regulate and control natural processes, thereby indirectly maintaining the environmental conditions essential for human survival (Guo et al., 2024). Cultural Ecosystem Services (CES) represent the non-material benefits humans obtain from ecosystems through spiritual fulfillment, cognitive development, reflection, recreation, and aesthetic experiences. As intangible bonds connecting humans and nature, CES contribute to enhanced human well-being (Huynh et al., 2022).

Although the assessment of ES has been increasingly applied to mitigate the human–land contradiction, different types of ES have received varying degrees of attention. The assessment of biophysically oriented ES, such as provisioning and regulating services, has become relatively well-established. However, CES, due to their intangible and non-material nature, present theoretical and methodological challenges. Consequently, they are often overlooked by researchers and policymakers and are rarely incorporated into decision-making processes (Guo and Zhang, 2024). However, in the context of international sustainable science development, the utility of CES in revealing the relationship between humans and nature, as well as their close connection to human well-being, has become increasingly evident. CES has gradually emerged as a frontier and hotspot in research across geography, ecology, and related disciplines (Kumar and Kumar, 2008). The focus of ES research is shifting toward CES with its non-material values and the interrelationships between provisioning, regulating, and cultural ES (Meng et al., 2020). Therefore, accurately estimating the economic value of CES and achieving its spatial visualization, while simultaneously conducting a comprehensive assessment together with provisioning and regulating ES, has become a critical issue that policymakers urgently need to address.

Currently, the assessment of provisioning and regulating ES has established a mature research framework (Liu et al., 2024). Studies often operate at specific scales, selecting appropriate indicators to quantify ES expression, identify spatial patterns, analyze synergies and trade-offs, and ultimately determine driving factors to formulate corresponding measures (Liu and Wu, 2024). In terms of methodology, model-based approaches such as the InVEST model and the value equivalence method have gradually replaced traditional abstract textual descriptions, enabling spatially explicit quantification of ES (Guo et al., 2024). Among these, the InVEST model has become a primary choice for calculating provisioning and regulating ES due to its advantages in simplified data requirements and accurate simulation assessments (Bastos et al., 2023).

In CES research, defining CES categories is a prerequisite for value assessment. However, due to the intangible nature of CES, accurately classifying them remains challenging. The delineation of CES types is highly subjective, as individuals may perceive CES differently depending on their background, experience, age, and gender (Scholie et al., 2015). In addition, CES categories may overlap, intersect, or fail to account for local cultural characteristics (Richards and Bige, 2018). Therefore, to avoid controversy, this study combines authoritative frameworks such as CICES with field survey data to classify the CES types considered in this paper.

CES assessment is a fundamental aspect of CES research. Currently, CES assessment methods are mainly divided into non-monetary and monetary approaches (Hirons et al., 2016; Christie et al., 2012). Each approach can be further classified based on Revealed Preference (RP) and Stated Preference (SP) (Christie et al., 2012; Spangenberg and Settele, 2010). In non-monetary approaches, RP indirectly determines human preferences for CES by observing behaviors or analyzing literature and data. Specific methods include literature analysis (Scholie et al., 2015), social media-based approaches (Czembrowski et al., 2016), and observational methods (Unnikrishnan and Nagendra, 2015). SP, on the other hand, evaluates CES by directly eliciting the values and preferences of individuals or groups, using methods such as surveys and interviews (Schmidt et al., 2016), expert judgment (Nahuelhual et al., 2013), Q methodology (Winkler and Nicholas, 2016), and participatory mapping (Brown and Raymond, 2007). In monetary approaches, RP evaluates CES value by observing actual market behaviors related to CES, with common methods including the hedonic pricing method (Spangenberg and Settele, 2010) and the travel cost method (Cheng et al., 2019). SP evaluates value by constructing hypothetical markets, in which respond-ents directly state their willingness to pay, accept, or for go services. This includes methods such as deliberative valuation (Kenter et al., 2016), contingent valuation method (CVM), and Discrete Choice Experiments (DCE) (Xiao et al., 2026). Compared with non-monetary approaches, monetary assessment methods not only provide policymakers with a more intuitive understanding of the economic value of CES, offering strong economic evidence for ecological protection and management (Spangenberg and Settele, 2010), but also enable the comparison of synergies and trade-offs between CES and other ecosystem services, facilitating the integration of CES into overall ES assessments (Cheng et al., 2019). Therefore, this study employs a monetary approach to quantify the eco-nomic value of CES.

Although RP in monetary approaches can monetize CES value, it only captures a portion of CES, such as aesthetics and cultural heritage, and cannot measure its full value (Cheng et al., 2019). SP methods, in contrast, can cover all CES categories (Christie et al., 2012). However, deliberative valuation requires that participants be highly familiar with CES in order to express accurate preferences (Christie et al., 2012), which makes it difficult to apply at larger scales. The CVM determines CES value by asking respondents how much they are willing to pay, but it cannot estimate the marginal value of different CES types and is highly subjective (Xiao et al., 2025). DCE, on the other hand, present respondents with a series of hypothetical yet realistic choice scenarios, allowing them to select among different CES types (attributes). This approach reveals which CES types (attributes) are more important to respondents and how much they are willing to pay for improvements in these CES types (attributes), thereby determining CES value (Nie et al., 2025). Compared with CVM, DCE not only reduces subjective bias but also allows for a clear estimation of respondents’ willingness to pay for each CES type (attribute) and the calculation of marginal values for each CES type (attribute), making it a mainstream method in current environmental and resource valuation studies (Nie et al., 2023). Therefore, in this study, we use DCE to estimate CES value, designing choice scenarios with CES types as attributes to survey respondents’ willingness to pay, thereby calculating the economic value of CES. It should be noted that DCE can only quantify the economic value of CES and cannot provide a spatially explicit representation of this value. Spatializing the economic value of CES not only helps reveal regional differences and inform differentiated management strategies but also allows for the identification of patterns in the relationships between CES and other ecosystem services. Additionally, it provides scientific support for ecological protection and optimized land use. Non-monetary methods such as PPGIS, however, can effectively achieve spatial representation of CES. Therefore, this study combines the monetary method DCE with the non-monetary method PPGIS to leverage the strengths of both approaches, enabling a scientifically rigorous measurement and spatial representation of CES economic value.

The interrelationships among CES, provisioning ES, and regulating ES are also key issues addressed in this study. As direct material products and services provided by eco-systems, provisioning ES form the foundation for the delivery of regulating ES and CES (Daily, 1997). However, this does not imply that more provisioning ES is always better, as excessive provisioning ES may encroach upon the spatial carriers of regulating ES and CES, thereby negatively impacting these two types of services (Raudsepp-Hearne et al., 2010). Regulating ES maintain ecosystem stability and health, ensuring the sustainability of both provisioning ES and CES (De Groot et al., 2002). Yet, an overemphasis on regulating ES may also constrain the delivery of provisioning ES and CES (Rodríguez et al., 2006). CES originate from provisioning and regulating ES, with well-functioning provisioning and regulating ES serving as prerequisites for CES (Bezerra et al., 2025). However, excessive focus on CES may intensify human activities and negatively affect provisioning and regulating ES (Sharma et al., 2025). Therefore, synergies and trade-offs coexist among these three types of ES, necessitating an assessment of their integrated development level.

Forests provide critical ES essential to the Earth’s ecological environment. With the growing understanding of CES, an increasing number of studies emphasize the need to examine the synergies and trade-offs between forest CES and provisioning and regulating ES. Miroslava Hochmalová et al. (Hochmalová et al., 2022) using forest-rich regions in the Czech Republic as a case study, highlighted that forest ecosystem assessments must consider the interactions among provisioning, regulating, and cultural ES to reveal the multifunctional characteristics of forests and their balance relationships. Helseth et al. (Helseth et al., 2022) analyzing the dynamic changes in Norway’s forest ecosystem services from 1950 to 2020, found that forest functions are gradually shifting from provisioning and regulating ES toward CES. This shift reflects changes in societal demand structures and indicates that governments should balance the three types of ES in forest management. Chen and Wu (2024) argued that quantifying the synergies and trade-offs between forest CES and provisioning and regulating ES is crucial for developing scientifically informed forest ecosystem management strategies. Deng et al. (2023) further noted that in karst regions, the interactions among the three types of ES in forest ecosystems are more complex and require in-depth analysis considering regional ecological characteristics. Guilin, China, is not only a typical forest-rich area but also features a unique karst landscape, making its ecosystem service relationships highly representative and valuable for research. Therefore, this study uses Guilin as a case study to systematically analyze the comprehensive relationships among provisioning ES, regulating ES, and CES in forest ecosystems, based on the quantification of forest CES economic value.

In terms of quantitative methods for assessing synergies and trade-offs among ES, current research commonly employs approaches such as correlation analysis and spatial overlay (Jopke et al., 2015). For example, Jaligot et al. (2019). conducted a correlation analysis of provisioning, regulating, and cultural ES in Switzerland. Canedoli et al. (2020) used correlation analysis to evaluate the relationships among multiple ES in forest and grassland habitats in the mountainous regions of Italy. Cademus et al. (2014) applied spatial overlay to analyze the relationships among timber production, carbon sequestration, and water yield in the southeastern United States. Li et al. (2023) conducted a spatial overlay analysis to examine trade-offs and synergies among ES in Hebei Province, China. These methods have produced substantial results. However, correlation analysis and similar methods typically focus on pairwise relationships among ES and cannot quantitatively capture interactions among all three types of ES. While spatial overlay can identify synergies and trade-offs among three types of ES, it struggles to reflect their overall development and lacks an integrated analysis. We argue that a region’s overall ES condition should be considered, rather than merely measuring ES magnitudes or pairwise synergies/trade-offs. For instance, even if all ES are high in a region, poor coordination among them can still reduce overall ES effectiveness. Therefore, it is necessary to integrate ES values with their synergies and trade-offs for a comprehensive assessment. Spatial analytic geometry model can not only calculate the degree of coordination among ES but also link ES values with their synergies and trade-offs (Liu et al., 2023). Using the Coordination Transformation Degree (CTD), these models provide a more reasonable and effective representation of the comprehensive development level of regional ES.

This study, based on a multi-model coupling framework, evaluates provisioning, regulating, and cultural ES in Guilin, China. It comprehensively analyzes the synergies and trade-offs among the three types of ES, as well as their overall development level, and explores the natural and socioeconomic driving factors. The specific objectives are: (1) to quantify the levels of provisioning, regulating, and cultural ES; (2) to reveal the integrated relationships and spatial patterns among the three ES using a spatial analytic geometry model; and (3) to identify the main natural and socioeconomic factors influencing the comprehensive development of ES. This study integrates the monetary method DCE with the non-monetary method PPGIS, addressing the limitations of monetary approaches in spatial representation while overcoming the shortcomings of non-monetary approaches, which cannot reflect CES economic value or easily allow horizontal comparisons with other types of ES. This combination provides a scientifically robust and operationally feasible framework for both quantifying and spatially visualizing CES economic value. Additionally, by introducing a spatial analytic geometry model, this study attempts to calculate the comprehensive development level of the three types of ES, offering methodological support for effectively assessing integrated ES and optimizing ecosystem management.

Study area

Guilin is located in southern China (24°15′–26°24′N, 109°36′–111°29′E), covering a total area of 27,800 km 2 with a population of 4.93 million and a forest coverage rate of 75%, making it a typical forest-rich region (Figure 1). Guilin’s unique natural and cultural eco-logical landscapes make it an ideal area for assessing provisioning, regulating, and cultural ES. First, Guilin is a representative karst landform area characterized by the inter-woven distribution of rivers, forests, grasslands, and peaks. Its overall ecological environment is fragile and sensitive, thus requiring evaluation of provisioning and regulating ES to support ecological improvement and restoration. Second, Guilin’s distinctive landforms provide local residents with a range of cultural services such as recreation, aesthetic enjoyment, and inspiration. Meanwhile, the region preserves numerous historical sites, traditional settlements, and ethnic cultural heritage, reflecting rich cultural diversity and strong social identity. The coexistence of multiple ethnic groups has also given rise to di-verse folk customs and local cultural expressions, making Guilin highly representative in terms of cultural experience and spiritual identity—hence the need to evaluate CES.

Data sources

The land use data were sourced from the China Land Cover Dataset (CLCD) developed by Wuhan University, comprising 2023 remote sensing monitoring data with 30 × 30 m grid resolution. The natural environment data comprise evapotranspiration, precipitation, soil properties, and NDVI datasets sourced from the Resources and Environment Science Data Center of the Chinese Academy of Sciences, with an original resolution of 1 km × 1 km. DEM data were obtained from the Geospatial Data Cloud at 30 × 30 m resolution, with slope data derived therefrom. All raster datasets were uniformly processed using Albers Equal Area Conic projection and resampled to a consistent 30 m × 30 m spatial resolution. The survey data were obtained from a questionnaire survey conducted in Guilin in 2023. A total of 618 respondents were randomly interviewed, including 306 urban residents and 312 county-town/rural residents. In total, 607 questionnaires were returned, while 11 respondents did not submit their questionnaires. After further excluding 7 invalid questionnaires with incomplete responses or highly consistent answers, 600 valid questionnaires were obtained, with a valid response rate of 97.09%. Among them, 297 valid samples were collected from urban areas and 303 from county-town/rural areas. The proportions of the two groups were relatively close, indicating relatively balanced spatial coverage across the main survey areas (See the Supplementary Materials for the detailed table). Statistically, the respondents’ ages ranged widely from 24 to 68 years, with an average age of approximately 38 years. Since the survey required respondents to have a certain degree of familiarity with local culture, most respondents were local residents. Non-local respondents were required to have lived in Guilin for a certain period of time. Respondents were randomly selected and asked to express their behavioral choices. The questionnaire consisted of three parts. The first part focused on respondents’ spatial perceptions of CES. First, respondents were shown a map of Guilin indicating current land cover conditions and major scenic spots. They were then asked to mark, as comprehensively as possible, the locations that they considered important and capable of providing specific CES, and the coordinates of these locations were recorded. Finally, respondents were asked to assume that 100 CNY of conservation funds was available for each CES type and to allocate the funds among the locations they had just marked. The second part was the discrete choice experiment, in which respondents selected the option they perceived as maximizing their benefits. The third part collected respondents’ basic demographic information.

Research methods

The classification of provisioning ES, regulating ES, and CES

Based on the CICES framework, this study classifies ES in Guilin into three categories: provisioning, regulating, and cultural services. Guilin is rich in forest resources and is also characterized by typical karst landforms. Regional ecological processes are influenced not only by forest vegetation coverage, ecological regulation functions, and landscape connectivity, but also by karst geomorphological conditions, such as high rock exposure, shallow soil layers, strong permeability, and weak soil water retention capacity. Therefore, the ES indicators selected in this study should effectively reflect the ecological functional characteristics shaped by the combined effects of abundant forest resources and karst landforms. Provisioning ES include water yield and grain supply. Water yield can characterize hydrological supply capacity under the combined influence of forest vegetation, precipitation, actual evapotranspiration, soil water retention capacity, and karst leakage processes. In Guilin, forest vegetation affects water resource formation by regulating precipitation interception, infiltration, and runoff processes (Yan et al., 2021), while the strong leakage and weak water retention characteristics of karst areas further highlight the importance of water yield (Liu et al., 2024). Therefore, water yield is an important indicator for evaluating provisioning services in this region. Grain supply mainly reflects the productive capacity of limited cultivated land resources under a forest-dominated landscape, as well as the trade-off between agricultural production and ecological conservation. Because mountains, hills, and karst landforms are widely distributed in Guilin, cultivated land resources are relatively limited, and agricultural production is easily constrained by fragmented terrain, shallow soil layers, and weak soil water and nutrient retention capacity. Thus, assessing grain supply is important for characterizing regional provisioning ES (Cao et al., 2020; Cao et al., 2021). Regulating ES include soil conservation, carbon stock, and habitat quality. Soil conservation reflects the protective role of forest vegetation on fragile karst surfaces through processes such as rainfall interception, root-induced soil stabilization, and the reduction of runoff scouring. It is an important indicator for evaluating regional soil erosion and rocky desertification risks (Zhu et al., 2021; Li et al., 2024). Carbon stock reflects the important functions of forest ecosystems in vegetation carbon storage, soil carbon accumulation, ecological restoration, and rocky desertification control, and plays an irreplaceable role in regional material cycling, energy flow, and climate regulation (Li et al., 2020; Zhu et al., 2025). Habitat quality characterizes the relationships among forest patch integrity, landscape connectivity, human disturbance intensity, and biodiversity maintenance capacity, and provides an important basis for assessing the ecological integrity of forest-dominated landscapes and formulating biodiversity conservation strategies (Wang et al., 2021; Wu et al., 2021).

CES can comprehensively reflect the landscape aesthetic, recreational, ecological education, and local cultural values formed through the interactions among forest vegetation, karst landforms, mountain–water landscapes, and human activities in Guilin. They are important indicators for characterizing the comprehensive value of Guilin’s forest–karst composite landscape. Unlike provisioning ES and regulating ES, CES rely more strongly on human perception, experience, and cultural connections. The aesthetic enjoyment, relaxation, cultural identity, and spiritual pleasure that individuals obtain from ecosystems constitute the fundamental elements of CES (Zhu et al., 2024). This study classified CES types through a two-step process: First, building upon the CICES framework while incorporating local cultural characteristics of Guilin, we initially categorized CES into 15 types after eliminating overlapping categories (see Supplementary Materials). Through questionnaire surveys, we examined respondents’ understanding of landscape types, historical culture, and local environment, documenting their annual participation in cultural activities and identifying which natural or cultural elements they associated with specific CES types. Second, we applied Principal Component Analysis (PCA) for dimensionality reduction (detailed in Supplementary Materials) (Jiang et al., 2024) to clarify distinctions between different cultural types. Consequently, we consolidated CES into four categories: Aesthetic Information, Recreation and Well-being, Natural and Cultural Heritage, and Biodiversity Value (Table 1).

Calculation of provisioning ES

The grain supply is determined based on total grain yield and the Normalized Difference Vegetation Index (NDVI) (Li T. et al., 2024), while water yield is calculated using the InVEST model (Liu et al., 2024). The specific formulas are presented in Table 2, through which we can quantify these two key components of provisioning ES.

Calculation of regulating ES

Soil conservation, carbon stock, and habitat quality were quantified using the Sediment Delivery Ratio (SDR) module (Li et al., 2024), Carbon module (Zhu et al., 2024), and Habitat Quality module (Wu et al., 2021) of the InVEST model, respectively. The specific formulas are presented in Table 3, through which the three key components of regulating ES can be systematically assessed.

CES value calculation

Given the unique characteristics of CES, this study employs the DCE to estimate their non-market economic value. As a stated preference research method, the DCE is grounded in the random utility theory proposed by Lancaster. This theory posits that the utility de-rived by respondents is determined by the attributes of the product or service in question (Nie et al., 2023a). The DCE method primarily involves designing choice scenarios based on the attributes of a product or service and their varying hypothetical levels (a monetary attribute is included when economic valuation is required), thereby constructing choice sets. Respondents are then asked to select their most preferred option from these sets through questionnaires or surveys, which allows researchers to elicit their preferences for different attributes and their willingness to pay for changes in attribute levels (Wang et al., 2025). Conducted in a hypothetical market setting, the DCE estimates the value of product or service attributes by analyzing respondents’ WTP. It has been extensively applied in environmental and eco-logical valuation studies and is increasingly recognized as a mature and reliable method for preference assessment (Nie et al., 2023b).

The design of DCE involves four fundamental steps: defining attributes, determining attribute levels, generating choice alternatives, and constructing choice sets. In this study, the four reclassified types of CES were selected as attributes, to which a monetary attribute was subsequently added. The levels of each attribute were determined based on existing literature and field investigations (Nie et al., 2025). The levels for Aesthetic Information were defined in relation to vegetation coverage and water clarity; Recreation and Well-being was associated with the number of ecological trails and parks; Natural and Cultural Heritage was linked to the quantity of natural landscapes and historical sites; Biodiversity Value was characterized using the number of species listed in the IUCN Red List; and the levels of the monetary attribute were based on local residents’ income derived from tourism. Detailed attributes and their corresponding levels are presented in Table 4.

This study incorporated five attributes with level counts of 3, 3, 3, 3, and 4, respectively. The full factorial combination of these attributes would theoretically yield 243 (3 × 3 × 3 × 3 × 4) candidate alternatives. However, including all alternatives is practically in-feasible, as the cognitive load and limited rationality of respondents can lead to decreased decision quality when faced with excessive options. To ensure respondent comprehension and survey efficiency, an orthogonal experimental design was implemented using SAS 9.4 software. After eliminating unrealistic alternatives, 18 choice scenarios were retained and organized into 6 choice sets. Each set consisted of three improved alternatives and one “opt-out” option (indicating no change from the status quo). The D-efficiency value of 92.53% indicates a highly effective orthogonal design. An example of a choice set is illus-trated in Figure 2. Following data collection on respondents’ choices and spatial perceptions, we applied random utility theory to quantify cultural service values (see Supplementary Materials for equations), with spatial visualization achieved through PPGIS (detailed methodology in Supplementary Materials).

Measurement of ES

This study first standardized all data and assigned equal weights to each specific type of provisioning ES and regulating ES, then calculated their composite scores through weighted summation. Due to the inherent complexity in determining appropriate weights for different CES types, we introduced a multidimensional model that quantifies CES through polygonal area calculations (Jiang et al., 2020). Specifically, let a, b, c, and d represent the standardised values of four cultural function types for the i-th unit. The angle α between any two adjacent dimensions is set at 90° (α = 360°/4). The polygonal area S is calculated using the formula: S = (ab + bc + cd + de) × sin α/2. Since the calculated area varies depending on the ordering sequence of CES types, we used the average of all possible polygonal configurations as the final CES value. To facilitate subsequent calculation of Guilin’s ES coordination deviation degree and coordination transformation degree, we also standardized the CES values (Figure 3).

Coordination deviation degree and coordination transformation degree of ES

This study employs a spatial analytic geometry model to assess the coordination and integrated development among provisioning, regulating, and cultural ES in Guilin. In this model, the three types of ES are mapped into a geometric space, where each ES is represented as a vector force in a distinct direction. The combined effects of these directional vector forces are then used to identify the coordination deviation degree (CDD) and the coordination transformation degree (CTD) among the ES (Yang et al., 2018). As shown in Figure 4, the vector forces representing provisioning, regulating, and cultural ES form a three-dimensional space, within which there must exist a point A that represents the integrated state of ES interactions. The distance between point O and point A (OA) denotes the total coupling degree (TCD), reflecting the level of interconnection among the three ES, which is closely related to their respective assessment values. Although TCD can indicate how tightly the three ES are connected, it does not reveal whether these connections represent synergies or trade-offs. Therefore, it is necessary to further calculate the synergistic and trade-off characteristics among the three ES to provide a more comprehensive under-standing of their interactions. Assuming that all ES share the same origin point (O), a point located on the spatial diagonal OA₁ represents a state of synchronous development and perfect coordination among the three ES. If point A does not fall on OA₁, it indicates the presence of trade-offs among the ES, meaning that full coordination has not been achieved. The perpendicular distance d from point A to the diagonal OA₁ represents the deviation between the actual relationships among the ES and the ideal coordinated state. This deviation reflects the difference between the actual coupling of ES and their ideally coordinated development. Since it arises from trade-offs among the ES, the distance d can be regarded as the CDD, which characterizes the extent of synergy or trade-off among the ES. Finally, the CTD is calculated by subtracting the CDD from the TCD to reflect the over-all development level of ES (see Supplementary Materials for the specific formula). By excluding the trade-off component from the interrelationships among the three ES, we consider that the remaining part represents the genuine development status of the ES. Within this framework, TCD represents the overall coupling condition of ES, CDD characterizes the synergistic or trade-off relationships among them, and CTD integrates these two aspects to describe the comprehensive development level of ES. CDD ranges from 0 to 1, with higher values indicating poorer coordination, while CTD also ranges from 0 to 1, with higher values representing better overall effects.

Factors influencing CTD of ES

This study selects independent variables from two dimensions—natural and socio-economic—to explain the CTD. Based on existing studies (Zhu et al., 2025) and the availability of relevant data, eight explanatory variables were identified, as summarised in Table 5.

This paper integrates spatial information into conventional regression frameworks and employs spatial econometric techniques to examine the factors influencing CTD (Fu and Yang, 2024). Spatial econometric models primarily comprise two forms: the spatial lag model (SLM) and the spatial error model (SEM). The appropriate model specification is selected based on spatial dependence diagnostics (Assaf and Assaad, 2024). The specific vector form of the model is presented in Table 6.

ES spatial patterns

Provisioning ES spatial patterns

Based on the calculation results of grain supply and water yield, as well as the weighted summation method, this study analyzed the spatial distribution characteristics of provisioning ES and their specific types in Guilin (Figure 5). Overall, grain supply, water yield, and the integrated function of provisioning ES all showed clear spatial differences. However, the distributions of their high- and low-value areas were not entirely consistent, indicating that provisioning ES in Guilin are jointly influenced by cultivated land resources, topography and geomorphology, vegetation coverage, and other factors.

In terms of grain supply, its spatial pattern showed a relatively distinct belt-like clustering feature. High-value areas were mainly concentrated in Quanzhou County, Lingui District, the urban districts, Yangshuo County, and Pingle County, and the county-level mean values of grain supply showed similar results. These areas are mostly located in river valley plains, basins, or relatively flat terrain, where cultivated land is more contiguous and the quality of cultivated land and agricultural infrastructure conditions are relatively favorable. Therefore, they have strong grain supply capacity. In contrast, both the spatial distribution and county-level mean values indicate that Ziyuan County and Longsheng Various Nationalities Autonomous County are low-value areas for grain supply. These two counties are dominated by mountains and hills, with well-developed karst landforms, fragmented and limited cultivated land resources, and constrained agricultural production conditions, resulting in relatively low levels of grain supply. In terms of water yield, Guilin generally had a high water yield level, but with clear spatial differentiation, showing an overall pattern of “higher in the south and lower in the north.” Specifically, the urban districts, Lingui District, Yongfu County, and Pingle County in the central and southern parts of Guilin had relatively high water yield. This was mainly because these areas have higher proportions of construction land and cultivated land, relatively lower vegetation coverage, and lower actual evapotranspiration. As a result, under similar precipitation conditions, they showed stronger water yield capacity. By contrast, Ziyuan County and Longsheng Various Nationalities Autonomous County in the north had relatively low water yield. These areas are dominated by mountainous forest ecosystems, with high forest coverage, strong vegetation evapotranspiration, and strong water conservation functions, leading to relatively low water yield as represented by surface water available for supply. Meanwhile, these counties have higher elevations and greater topographic relief, and water yield capacity also tended to decrease with increasing elevation. This indicates that in forest-rich mountainous areas, forest vegetation can maintain the stability of hydrological processes through water conservation and runoff regulation, while also affecting the spatial distribution of water yield due to strong evapotranspiration.

After integrating grain supply and water yield, the overall level of provisioning ES in Guilin was moderate, with relatively small differences among counties. Spatially, provisioning ES showed a pattern of “higher in the southern and central regions and lower in the northern region.” High-value areas were mainly concentrated in the urban districts, Lingui District, Yangshuo County, Lipu City, and Pingle County. These areas have both favorable agricultural production conditions and relatively high water yield and water supply capacity, resulting in a higher integrated level of provisioning ES. Low-value areas were mainly distributed in Ziyuan County and Longsheng Various Nationalities Autonomous County in the north, where both grain supply and water yield were relatively low, leading to weaker integrated provisioning ES functions.

Regulating ES spatial patterns

Based on the calculation results of soil conservation, carbon stock, and habitat quality, as well as the weighted summation method, this study analyzed the spatial distribution characteristics of regulating ES and their specific types in Guilin (Figure 6). Overall, the three types of regulating ES showed clear spatial heterogeneity and were closely related to topographic conditions, vegetation coverage, human activity intensity, and land-use structure. Among them, soil conservation and habitat quality exhibited a high degree of spatial similarity, both performing better in mountainous and hilly areas, whereas carbon stock showed a relatively high overall level and a more continuous spatial distribution pattern.

In terms of soil conservation, its spatial distribution generally showed an opposite pattern to that of grain supply, with certain differences among counties. High-value areas were mainly scattered in mountainous and hilly areas along county boundaries, such as Ziyuan County, Longsheng Various Nationalities Autonomous County, Guanyang County, and Gongcheng Yao Autonomous County. These areas have favorable ecological baseline conditions, high vegetation coverage, and relatively low human development intensity. Ecological land such as forests and grasslands can effectively reduce slope runoff and soil loss, resulting in strong soil conservation capacity. In contrast, low-value areas were mainly concentrated in river valley plains, such as the urban districts and Lingui District. These areas have concentrated construction land, cultivated land, and population activities, with relatively high land development intensity and obvious surface disturbance, leading to weaker soil conservation functions. In terms of carbon stock, its spatial pattern was somewhat similar to that of soil conservation, but its overall level was relatively higher, with the mean carbon stock values of all counties generally remaining at a high level. High-value areas were mainly distributed in mountainous and hilly areas with high forest coverage, indicating that forest ecosystems are important carriers of carbon storage functions in Guilin. Medium-value areas mostly overlapped with agricultural production areas and some river valley zones, suggesting that relatively high vegetation coverage can still maintain regional carbon stock functions to some extent. By contrast, low-value areas were mainly concentrated in the urban districts and areas with dense urban construction. Urban expansion has led to vegetation reduction, intensified soil disturbance, and increased impervious surfaces, thereby weakening regional carbon storage capacity. In terms of habitat quality, there were certain differences among counties in Guilin. High-value areas were mainly distributed in mountainous and hilly areas with favorable natural baseline conditions, high forest coverage, and weak human disturbance. Their spatial distribution was relatively consistent with the high-value areas of soil conservation. These areas have relatively continuous ecological spaces and low construction and development intensity, providing stable spatial conditions for species habitats and the maintenance of ecological processes. In contrast, low-value areas were mainly concentrated in the urban districts, Lingui District, and some plain and river valley areas. Affected by urban construction, transportation corridors, and agricultural production activities, these areas have high land-use intensity and strong ecological patch fragmentation, resulting in relatively low habitat quality.

After integrating soil conservation, carbon stock, and habitat quality, regulating ES showed a relatively high overall level, with comparatively small differences among counties. However, spatially, they still exhibited a distribution pattern of “higher in the eastern and western regions and lower in the central region.” High-value areas were mainly distributed in Ziyuan County, Longsheng Various Nationalities Autonomous County, and some mountainous and hilly areas, whereas low-value areas were concentrated in the urban districts and surrounding areas with high development intensity. Overall, the spatial differentiation of regulating ES in Guilin reflects the important role of mountainous forest ecosystems in soil conservation, carbon storage, and habitat maintenance. It also indicates that areas with intensive human activities should be regarded as key areas for future improvement of regulating services and ecological restoration.

CES spatial patterns

By integrating the RPL model results with PPGIS, we obtained the spatial values of each CES type and subsequently calculated CES using a multidimensional model. Each CES type exhibits a spatial value gradient while maintaining broadly similar distribution patterns (Figure 7). Among them, Aesthetic information holds the highest value, reaching a maximum of 978 yuan. Simultaneously, aesthetic information demonstrates the second most extensive spatial distribution after recreation and well-being, indicating that respondents could perceive the beauty of Guilin’s landscapes across most locations and were willing to pay substantial amounts for this aesthetic appreciation. Biodiversity value shows comparable characteristics to natural and cultural heritage value. Although these two types have relatively limited spatial distributions, respondents demonstrated relatively higher willingness-to-pay for them. Biodiversity value primarily concentrates in nature reserves, which constrains its areal extent. Similarly, natural and cultural heritage involves traditional culture, religion, and spiritual fulfillment—elements that can only be experienced in specific locations, thus similarly limiting their distribution. Recreation and well-being displays the lowest monetary value but the most widespread distribution. This aligns with reality, as numerous locations can support outdoor activities while also facing substantial substitution effects—people can achieve recreation and well-being through alternative means, resulting in lower willingness-to-pay. When synthesizing all four types, CES exhibit spatial patterns characterized by coexisting concentration and dispersion, synergy and differentiation, along with persistent value gradients. Influenced by natural features (land cover types, species distributions) and socioeconomic factors (local culture, economic development), high-value areas concentrate in urban districts and Yangshuo County, demonstrating multi-type synergies. In contrast, low-value areas scatter across other counties with varying dominant CES types depending on local characteristics.

ES CDD and CTD

This study standardized the raster scales of provisioning ES, regulating ES, and CES in Guilin through resampling using ArcGIS 10.8, then measured their CDD and CTD using a spatial analytic geometry model. At the municipal level, Guilin exhibits a CDD of 0.236 and CTD of 0.279, indicating relatively good ES coordination but suboptimal integrated development that requires strengthened optimization of linkages among the three ES types. As shown in Figure 8, areas with low/lower CDD values are primarily concentrated in the urban districts, Yangshuo County, and Quanzhou County. The urban core and Yangshuo demonstrate medium regulating ES but high provisioning ES and CES, resulting in good coordination, while Quanzhou County shows low-level coordination due to uniformly low performance across all three ES types. Medium CDD zones are pre-dominantly distributed across plains and valleys, where high provisioning ES combines with moderate regulating ES and CES to produce intermediate coordination. High/higher CDD areas mainly include Guanyang, Ziyuan, Gongcheng Yao Autonomous County, and Longsheng Multi-Ethnic County, where high regulating ES coexists with poor provisioning ES and CES, leading to weak coordination.

The CTD exhibits a spatial pattern with higher values in the south compared to the north. Areas with low and relatively low CTD values are primarily located in Quanzhou, Xing’an, Lingchuan, Lingui, Ziyuan, and Longsheng counties. In these regions, either all three types of ES are at low levels or there exists a significant deficiency in one particular ES category, both of which substantially constrain the improvement of CTD. The medi-um-value zones represent extensions of the low and relatively low-value areas, where all three ES types demonstrate moderate performance, resulting in intermediate CTD levels. Regions with relatively high CTD values mainly include Guanyang, Gongcheng Yao Au-tonomous County, and Yongfu County, where the presence of at least one comparatively strong ES type has contributed to some degree of CTD enhancement. The high-value zones are concentrated in parts of Yangshuo County and certain urban districts, where the robust performance of both provisioning ES and CES serves as a crucial foundation for achieving high CTD levels.

Analysis of factors influencing the CDD of ES

Before accounting for spatial relationships, it is essential to examine whether the de-pendent variable exhibits spatial correlation (Basu et al., 2024). The dependent variable in this study is the CTD of Guilin City, which shows a Moran’s I index of 0.2178, indicating the presence of spatial correlation. This justifies the use of spatial econometric models to analyse the influencing factors. The results of spatial OLS estimation indicate that the LM-lag test is significant at the 0.1 level, thus supporting the use of a spatial lag model.

As shown in Table 7, temperature, topography, population density, and industrial structure exert significant negative effects on the CTD, while forest coverage and GDP show positive effects. Precipitation and per capita retail sales of consumer goods have no significant influence. Specifically, Guilin, located in a subtropical monsoon climate zone, frequently experiences high temperatures, which negatively affect various types of ES. High temperatures intensify evapotranspiration, reducing crop yields and water supply capacity; decrease vegetation cover, weakening soil conservation and habitat quality; accelerate organic matter decomposition, lowering carbon storage; and suppress outdoor cultural and recreational activities, thereby diminishing CES values. The karst topography also has a negative effect on ES. Its fragmented terrain and shallow soils reduce agricultural productivity and water yield, while severe soil erosion undermines soil retention and carbon storage capacities. Ecosystem fragmentation and the expansion of rocky desertification further lead to biodiversity loss. Moreover, the complex terrain and ecological fragility constrain cultural spaces and weaken cultural accessibility and continuity. High population density imposes additional pressure through land occupation, resource consumption, and environmental degradation, resulting in reduced cropland, weakened water yield, deteriorated soil and habitat quality, and diminished carbon sequestration. The compression of natural and cultural spaces also weakens aesthetic enjoyment and cultural identity. An industry structure dominated by tourism further intensifies ES pressures. Tourism development occupies farmland and increases surface hardening, impairing water regulation and soil retention functions. Infrastructure expansion fragments habitats and reduces biodiversity; energy use and construction activities lower carbon storage; and excessive commercialization and tourist concentration lead to landscape degradation and loss of cultural authenticity. Overall, while tourism-driven economic growth enhances local income, it exerts systemic negative impacts on regional ES.

In contrast, a high forest coverage rate significantly enhances various ES. Forests regulate microclimates, strengthen water retention and soil conservation, and promote agricultural productivity and water yield. They also improve habitat quality and carbon sequestration while enriching aesthetic enjoyment and cultural identity, achieving synergistic improvement and sustainable development of ES. GDP growth, when accompanied by effective ecological regulation, can also promote ES enhancement. Economic development supports agricultural infrastructure and technological upgrading, improves water management and ecological restoration, advances green energy and afforestation initiatives, and strengthens public services and cultural industries. Consequently, it fosters a win–win relationship between economic growth and ES improvement.

Spatial pattern characteristics of ES in Guilin

The results of this study show that provisioning ES in Guilin are mainly concentrated in plain and river valley areas, providing the material production foundation for the ecosystem. Regulating ES are mainly distributed in mountainous and hilly areas, forming an ecological security barrier and ensuring ecosystem stability. CES are concentrated in the central urban area and surrounding regions with abundant landscape resources and frequent tourism activities, such as the Li River Basin, Yangshuo, and the Longji Rice Terraces, showing a spatial clustering effect of human activities and cultural perceptions. The formation of this spatial pattern is not the result of a single factor, but rather the combined influence of natural geographical conditions, land-use structure, human activity intensity, and regional functional division. Specifically, plain and river valley areas have relatively flat terrain, concentrated cultivated land resources, and favorable agricultural production conditions, making them more likely to develop strong grain supply and water supply capacities. Mountainous and hilly areas have high forest coverage and relatively low human development intensity, and therefore show clear advantages in regulating services such as soil conservation, carbon storage, and habitat maintenance. In contrast, the central urban area, Yangshuo, and areas along the Li River are characterized by concentrated landscape resources, good transportation accessibility, and well-developed tourism facilities, making them more likely to form high CES perception and value agglomeration.

From the perspective of spatial distribution patterns, provisioning ES and regulating ES show a spatial mismatch but complementary relationship. CES and regulating ES also exhibit a mismatch in their main distribution areas, while showing marginal overlap. In contrast, CES and provisioning ES show a relatively high degree of spatial overlap. These spatial complementary and overlapping relationships help maintain a relative balance of regional ES as a whole, which also explains the relatively good coordination among the three types of ES. However, the partial spatial mismatch among the three ES types leads to a relatively low CTD in Guilin, indicating that the linkage mechanisms among provisioning ES, regulating ES, and CES need to be further strengthened in the future. Specifically, the spatial mismatch between provisioning ES and regulating ES reflects the ecological trade-off within natural systems: areas with higher resource use intensity are often accompanied by a decline in regulating functions. This is consistent with the findings of Maes et al. (2012) and Costanza et al. (2014). Therefore, future ecosystem management should reduce the trade-off between provisioning ES and regulating ES and promote their coordination. Although regulating ES and CES are spatially mismatched, their marginal overlap indicates that the formation of CES often depends on a favorable ecological environment, while intensive development and utilization of CES may also disturb regulating ES. Therefore, in strengthening ES linkages, the relationship between CES and regulating ES should be coordinated to avoid a situation in which CES are enhanced at the expense of regulating ES degradation. In addition, high-value CES areas largely overlap with high-value provisioning ES areas, indicating that the formation and manifestation of CES depend on a strong provisioning foundation. Areas with high provisioning ES usually possess abundant natural and cultural resources, such as agricultural landscapes, traditional settlements, and agricultural heritage. These resources not only have high aesthetic and cultural symbolic value, but also provide key places for tourism, recreation, festivals, and spiritual experiences. Therefore, in strengthening ES linkages, the positive synergy between CES and provisioning ES should be further optimized to promote both the realization of ecological product value and the enhancement of cultural experience functions.

From the perspective of climate justice, the spatial pattern of ES in Guilin further reflects regional inequalities in ecological conservation responsibilities, climate risk burdens, and access to ecological benefits. The northern mountainous and hilly areas, as well as key ecological conservation areas, undertake substantial provisioning and regulating ES functions, such as soil conservation, carbon storage, habitat maintenance, and water conservation. However, constrained by topographic conditions, transportation accessibility, and industrial foundations, these areas obtain relatively limited benefits from provisioning ES and CES, which may lead to an imbalanced situation characterized by “high ecological contribution but insufficient development benefits.” In contrast, the central urban area, Yangshuo, and areas along the Li River can obtain more benefits from cultural tourism, the value transformation of ecological products, and infrastructure improvement. However, their higher intensity of human activities may also increase ecosystem pressure and weaken some provisioning and regulating ES functions. Therefore, climate justice or climate injustice may affect the spatial pattern of ES through the uneven distribution of ecological responsibilities and development benefits. Future ES management in Guilin should not only focus on increasing the overall supply and strengthening the coordination of provisioning ES, regulating ES, and CES, but also pay attention to regional differences in ecological conservation responsibilities, climate risk exposure, ecological service benefits, and adaptive capacity. For mountainous and hilly areas that undertake important provisioning and regulating ES functions, ecological compensation, green industry support, and public service improvement should be adopted to enhance the development capacity and benefit-sharing level of ecological conservation areas. For areas where CES and tourism benefits are highly concentrated, ecological capacity control, low-carbon tourism guidance, and ecological disturbance monitoring should be strengthened to prevent the development of cultural services from further intensifying the degradation of provisioning and regulating ES. In this way, trade-offs among ES can be mitigated, while regional equity between ecological conservation and development benefits can also be improved from the perspective of climate justice.

CES assessment of contributions to ES research

Taking Guilin, a typical forest-rich region in China, as the study area, this paper systematically evaluates provisioning, regulating, and cultural ES. Compared with the relatively mature research on provisioning and regulating ES, CES—due to their intangible, non-material, and invisible nature—are often overlooked or insufficiently understood by policymakers (Cheng et al., 2019). At present, CES assessment mainly follows two methodological pathways: non-monetary and monetary approaches. Non-monetary methods can capture human perceptions and emotional connections and have strong spatial explicitness, yet they fail to quantify the economic value of CES, providing limited decision-making support for policy and making cross-comparison with other types of ES difficult (Martin and Mazzotta, 2018). In contrast, monetary methods can translate cultural preferences into monetary indicators, offering a quantitative foundation for policymaking and for analyzing synergies and trade-offs among ES, but they often lack spatial explicitness (D'amato et al., 2016). To address these limitations, this study integrates a monetary method (DCE) with a non-monetary method (PPGIS), providing an approach that combines economic comparability with spatial visualization for CES assessment. The DCE can reveal trade-off choices among different CES attributes in a simulated market environment, thereby quantifying the public’s marginal willingness to pay (MWTP) for various CES. It is considered one of the most promising valuation tools in environmental economics (Rakotonarivo et al., 2016). The PPGIS, on the other hand, collects spatial information on cultural perceptions through public participation, assigning geographic coordinates and spatial weights to CES. Plieninger et al. (2013) applied PPGIS to identify CES and revealed their spatial clustering characteristics at the landscape scale; however, the results mostly remained at the level of perception frequency, failing to reflect their economic value. In this study, the willingness-to-pay (WTP) data obtained from the DCE were overlaid with the spatial coordinates collected through PPGIS, and the kernel density estimation method was applied to spatially express the economic value of CES. This approach not only achieves a quantitative coupling between cultural perceptions and ecological space but also establishes a three-dimensional “psychological-economic-spatial” integrated framework for CES assessment in forest-rich regions.

Therefore, the contribution of this study lies in establishing a closed-loop CES assessment pathway through the integration of DCE and PPGIS methods—linking the entire cognitive chain of CES from the elicitation of residents’ preferences (psychological dimension), to the monetary quantification of cultural preferences (economic dimension), and finally to their spatial representation (geographical dimension). At the same time, this approach balances subjective perception and objective measurement: it quantifies the intensity of cultural preferences through DCE and identifies their spatial heterogeneity using PPGIS, thereby achieving an organic integration of the two. More importantly, the method establishes a replicable regional assessment framework that does not rely on any specific data source and can flexibly adjust CES attributes and levels according to different region-al cultural characteristics, demonstrating strong cross-regional applicability and scalability.

Regarding evaluation results, scholars have noted that non-specialists are more sensitive to landscape aesthetics, identifying aesthetic landscapes as a primary type of CES (Li L. et al., 2024). Our study confirms this finding: respondents exhibited the highest willingness to pay for aesthetic information, and the spatial distribution of aesthetic information is ex-tensive. This also indirectly validates the accuracy of our CES assessment results.

Analysis of the relationship between ES

Provisioning, regulating, and cultural ES together form a tightly interconnected network characterized by interdependence, synergy, and potential trade-offs. Among them, regulating ES serves as the foundation, as its healthy functioning provides the necessary environmental conditions for the stable production of provisioning ES and the realization of CES. Provisioning ES directly supports human survival and development, yet its sustainability highly depends on the maintenance of regulating ES. CES acts as a key link among the three, profoundly shaping human use and management decisions of provisioning and regulating ES by influencing perceptions, emotions, and values toward nature, thereby serving as an intrinsic driver for the sustainable conservation of ecosystems (Martín-López et al., 2014). Quantifying synergies and trade-offs among the three types of ES has become a critical issue in policymaking and ecosystem management. It is important to emphasize that a scientific assessment of regional ES should consider not only ES levels but also the integrated performance of their internal synergies and trade-offs (Liu et al., 2023). To this end, this study introduces a Spatial Analytic Geometry model that not only quantifies the synergies and trade-offs among provisioning ES, regulating ES, and CES but also measures their overall development. This approach overcomes the limitations of previous studies that analyzed ES and their interactions separately and provides a feasible methodology for comprehensive ES assessment in forest-rich regions.

To further elucidate the relationships among provisioning ES, regulating ES, and CES, this study employed Pearson’s correlation coefficient method to analyze pairwise ES interactions. This approach was selected because Pearson’s correlation coefficient is particularly effective at identifying synergy/trade-off relationships among ES in complex regions (Yang et al., 2022). The results indicate that the correlation coefficient between provisioning ES and CES is approximately 0.15, showing a weak positive relationship, while the correlation coefficient between regulating ES and CES is about −0.13, indicating a weak negative relationship. This suggests that the connections between CES and both provisioning and regulating ES in Guilin are relatively weak, which is consistent with the findings of Zhu et al. (2024). The underlying mechanisms can be explained as follows: While provisioning ES emphasize material outputs and CES focus on spiritual experiences—sharing common origins but exhibiting fundamental differences—they can coexist (e.g., Grain supply and pastoral landscapes), yet intensive agricultural practices may compromise aesthetic values, and vice versa. This explains their positive but weak correlation. Furthermore, areas with high CES values typically experience significant human influence, and such intensive human activities often negatively impact regulating ES, resulting in the observed weak negative correlation between them.

Moreover, these divergent results may be related to our chosen research scale and methodology. Chen et al. combined the improved equivalent factor method with Pearson’s correlation coefficient to analyze the synergy/trade-off relationships of forest ES in China’s Yangtze River Delta, finding positive correlations between both provisioning/regulating ES and CES (Chen and Wu, 2024). The improved equivalent factor method is more suitable for large-scale ES valuation, with its CES calculations based primarily on land use types. In contrast, our CES assessment method, being resident preference-oriented, is better suited for small-scale studies, which explains the differences in research outcomes. To enhance the accuracy of our findings, we plan to conduct future analyses at dual scales—both county-level and grid-based—to systematically examine the relationships among provisioning ES, regulating ES, and CES.

Policy implications

To optimize the ES management in Guilin and promote the synergistic development of provisioning ES, regulating ES, and CES, this study proposes the following policy implications:

(1) Provisioning ES: In the short term, grain supply capacity should be improved; in the long term, a stable grain security system should be established.

The short-term objective is to strengthen grain production capacity in core agricultural production areas. Guilin should focus on consolidating the functions of the two major agricultural production areas in the north and south, namely Pingle County and Quanzhou County. Within the red line for permanent basic farmland protection, the construction of high-standard farmland and the integrated restoration of cultivated land in terms of “quantity–quality–ecology” should be promoted to enhance the production potential of cultivated land and the resilience of grain supply. Meanwhile, in priority development areas such as the municipal districts and Lingui District, cultivated land protection mechanisms should be strengthened during urban expansion. Policy tools such as transfer of development rights (TDR) should be used to maintain the dynamic balance of total cultivated land area and the stability of cultivated land quality, thereby preventing urban construction from continuously squeezing the space available for grain supply. The long-term objective is to establish a provisioning service system that integrates grain security, ecological protection, and industrial integration. Guilin could designate permanent basic farmland protection zones in key ecotourism counties such as Yangshuo to safeguard the foundation of grain production. On this basis, agritourism integration projects, such as ecological agricultural sightseeing, farming experiences, and agricultural cultural exhibitions, could be moderately developed to promote a positive interaction among cultivated land protection, farmers’ income growth, and regional economic development. In addition, to optimize the north–south differences in water yield, Guilin could gradually promote the construction of water storage projects and the cross-regional allocation of water resources in the northern region, thereby improving water resource support for agricultural production and providing support for the long-term stability of the regional grain security system.

(2) Regulating ES: In the short term, soil conservation capacity in key areas should be improved; in the long term, forest ecological networks and ecological compensation mechanisms should be strengthened.

The short-term objective is to rapidly improve soil conservation and soil erosion prevention capacity in river valleys, plains, and areas with intense human activities. In these areas, Guilin can enhance rainwater retention, regulation, and infiltration through sponge city construction, thereby weakening the scouring effect of surface runoff on soil and improving soil erosion resistance. Wetland ecological corridors can be restored to effectively reduce soil loss by making use of the soil-fixing and sediment-retention functions of wetland vegetation roots. Urban green spaces can also be expanded to increase surface vegetation coverage and strengthen soil conservation functions in urban areas. Meanwhile, strip-shaped spaces such as marginal land, abandoned land, barren slopes, and transportation corridors should be fully utilized to develop ecological shelterbelts and landscape forests, thereby constructing a forest ecological network with a more complete structure and stronger connectivity, and reducing the intensity of surface runoff and the risk of soil erosion. The long-term objective is to build a stable, connected, and sustainable forest ecological regulation system. In mountainous and hilly areas, particularly within the ecological conservation red line, Guilin should continuously promote the improvement of low-efficiency forests, the restoration of degraded forests, and the revegetation of logged areas. These measures can increase forest coverage and ecosystem stability, thereby enhancing slope stability, water conservation capacity, and soil-fixing and water-retention functions. At the same time, a cross-regional horizontal ecological compensation mechanism should be established to provide fiscal transfer payments and green industry support for areas undertaking important ecological function maintenance tasks. This would strengthen the endogenous development capacity of ecological conservation areas and prevent the degradation of ecological baseline functions caused by socioeconomic development pressures.

(3) CES: In the short term, the accessibility and synergy of cultural ecosystem services should be improved; in the long term, a coordination mechanism between ecological restoration and cultural inheritance should be established.

The short-term objective is to activate the mobility of CES and enhance the spatial accessibility and ecological synergy of cultural resources. Guilin can rely on digital platforms to develop maps of intangible cultural heritage and design ecological study and cultural experience routes, thereby promoting the sharing and interconnection of cultural resources across regions, breaking the value gradient of CES, and alleviating spatial equity issues caused by “green gentrification” in tourism hotspots. Meanwhile, for areas with high CES value, such as Yangshuo, the areas along the Li River, and the Longji Rice Terraces, refined management should be implemented. Measures such as optimizing the ecological design of infrastructure, using permeable materials and reversible structures to construct walking trails and viewing platforms, and controlling tourist flows and activity ranges can reduce the disturbance of tourism activities to the ecological environment and landscape integrity, thereby promoting the coordinated development of ecological conservation and cultural utilization. The long-term objective is to promote the deep integration of cultural value enhancement and ecological restoration. For areas with low CES value, a comprehensive strategy combining long-term protection and gradual improvement should be implemented, and the linkage mechanism between ecological restoration and cultural empowerment should be strengthened. A high-quality ecological environment should serve as the foundation for the generation of cultural value. Measures such as restoring forest connectivity, rehabilitating wetlands, reconstructing vegetation, and optimizing landscape aesthetics can improve the accessibility and perceptibility of the natural environment, enabling the public to form emotional connections and cultural experiences in authentic and aesthetically appealing ecological spaces. At the same time, community co-management and public participation mechanisms should be established, and activities such as local knowledge sharing, traditional festivals, ecological education, and nature-based experiences should be incorporated into ecological governance. This would encourage residents to jointly shape, maintain, and inherit regional cultural values through participation, thereby achieving the coordinated enhancement of ecological conservation and cultural revitalization.

(4) ES synergy: In the short term, positive linkages among ES should be strengthened; in the long term, mechanisms for mitigating ES trade-offs should be established.

Guilin needs to maintain ES synergies, mitigate ES trade-offs, and strengthen the linkages among multiple ES to improve CTD. The short-term objective is to enhance the positive linkage between provisioning ES and CES. Guilin can rely on agricultural cultural heritage and traditional agricultural models, such as the Longji Rice Terraces and rice–fish co-culture systems, to promote agricultural heritage conservation and ecological agriculture demonstration projects. This would help establish an integrated development mechanism of “agricultural heritage + ecology + tourism” and create comprehensive demonstration areas with ecological, economic, and cultural values. Meanwhile, ecological and cultural tourism activities, such as rice culture festivals and farming experiences, should be promoted to enhance the added value of ecological products and cultural attractiveness. The long-term objective is to establish a mechanism for mitigating ES trade-offs. Guilin should gradually improve ecological capacity monitoring and early-warning systems in tourism areas, dynamically monitoring tourist numbers, activity intensity, ecological disturbance, and landscape carrying status. This would guide CES-related activities toward greener, low-carbon, and low-disturbance development, thereby reducing the negative impacts on regulating ES during the realization of CES values. At the same time, the relationship between provisioning ES and regulating ES should be coordinated over the long term. In core ecological source areas, high-intensity grain production activities should be strictly restricted, and ecological restoration and natural recovery within ecological red lines should be promoted to safeguard regulating ES. In ecological transition zones, eco-friendly planting systems, such as returning sloping cultivated land to forest and grassland, intercropping, and organic agriculture, should be promoted to achieve a balance between provisioning ES and regulating ES. In key agricultural production areas, efficient agricultural water use, reduced inputs of chemical fertilizers and pesticides, ecological irrigation, and circular agriculture technologies should be promoted to enhance regulating ES while maintaining the output of provisioning ES.

Limitations and prospects

First, although this study classified CES types based on an internationally authoritative CES framework and the practical survey context of Guilin, the relevant variables can still be further expanded due to the influence of landscape diversity, human subjective agency, and other factors. Meanwhile, respondents’ cultural backgrounds, individual experiences, length of residence, and familiarity with local landscapes may all affect their perceptions of CES and value judgments. Therefore, future research should further enrich CES-related variables and develop a more refined and context-specific CES indicator system by incorporating the characteristics of natural landscapes, historical cultures, and social activities in different regions. This would help improve the explanatory power of CES assessment and reduce the influence of subjectivity. Second, the survey period of this study was relatively short, and samples were collected only within approximately one week during spring and summer. However, different seasons may lead to changes in respondents’ CES preferences. For example, under the influence of natural factors such as low temperatures in winter, residents’ preferences may shift from natural landscapes to cultural landscapes. Therefore, in future research, we plan to conduct surveys during selected periods across all four seasons and gradually expand the survey scope to obtain a more comprehensive understanding of residents’ perceptions of CES. Third, the survey respondents in this study were predominantly local residents. Future research should include both residents and tourists in the survey sample to reveal the perceptions of CES among different social groups from a multi-stakeholder perspective. Finally, in terms of the selection of provisioning ES and regulating ES indicators, this study only selected several representative ES indicators. Future research could expand the range of provisioning ES and regulating ES indicators and further consider the impacts of climate change on these services, thereby improving the accuracy of the research.

Guilin, as a typical forest-rich region in China, serves not only as an important natural ecological barrier but also as a repository of rich cultural and ecological values, making ES research particularly significant. This study measured provisioning and regulating ES, estimated the economic value of CES, and realized the spatial representation of all three ES types. In addition, a comprehensive development assessment of the three ES was conducted. The main findings are as follows: Overall, provisioning ES in Guilin is at a medium level, with a spatial pattern showing higher values in the south and central regions than in the north. High-value areas are concentrated in the municipal districts, Lingui, Yangshuo, Lipu, and Pingle, while low-value areas are mainly found in resource counties and the Longsheng Autonomous Counties. Regulating ES exhibit generally good levels, following a spatial pattern of high values in the east and west and lower values in the central region, with high-value areas mainly in resource counties and the Longsheng Autonomous Counties, and low-value areas concentrated in the municipal districts. CES shows significant spatial differences, with high-value areas in the municipal districts and Yangshuo County, while low-value areas are scattered across other counties. Overall, the spatial patterns of CDD and CTD are relatively consistent. The low CDD values indicate a good level of ES coordination, but the low CTD values suggest that the overall develop-ment level of ES remains relatively weak. Factor analysis shows that Guilin’s CTD is influenced by multiple factors: temperature and topography have significant negative effects, population density and industrial structure exert weaker negative effects, and forest coverage and regional GDP have positive effects, although their influence is relatively limited.

The “psychological-economic-spatial” integrated framework applied in this study not only addresses the limitations of CES monetary valuation, but also overcomes the difficulty that non-monetary valuation methods cannot provide economic value references for policy-making. It therefore offers a transferable approach for studies in other forest-rich regions. During the extension and transfer of this framework, considering that the greatest challenge currently facing CES assessment is regional heterogeneity, a two-layer structure of “core framework + local modules” can be established. First, a core CES classification framework should be developed to identify common CES types, based on which standardized DCE attributes, attribute levels, and PPGIS place-based elements can be constructed. Second, given the significant regional heterogeneity and local cultural embeddedness of CES, different regions should adaptively adjust CES attributes and place types according to their local natural landscapes, historical cultures, ethnic traditions, and residents’ lifestyles when applying this framework, thereby preserving local cultural values and regional characteristics. In addition, to improve the efficiency of framework extension, online questionnaire surveys, mobile-based map annotation, and open-source PPGIS platforms can be further integrated to expand sample coverage, reduce the cost of field surveys, and enhance the practical operability of CES assessment in different forest-rich regions.