Evaluation and Driving Determinants of the Coordination between Ecosystem Service Supply and Demand: A Case Study in Shanxi Province

Abstract

Understanding the coordination relationship between ecosystem service (ES) supply and demand and elucidating the impact of driving factors is critical for regional land use planning and ecological sustainability. We use a large watershed area as a case to map and analyze ES supply, demand and the coordination relationship, and identify the associated socio-ecological driving variables. This study assessed the supply and demand of five ESs (crop production, water retention, soil conservation, carbon sequestration, and outdoor recreation) in 2000 and 2020, and evaluated the coordination between them employing the coupling coordination degree model (CCDM). Additionally, we utilized the geo-detector model (GDM) to identify driving determinants and their interactive effects on the spatial pattern of the coupling coordination degree (CCD) between ES supply and demand. The results showed that mountainous regions with abundant forest coverage were high-value areas for ES supply, while the ESs were predominantly required in city center areas within each basin area. From 2000 to 2020, there was a slight decline in ES supply and a significant increase in ES demand. Counties were grouped into four coordination zones in the study area: extreme incoordination, moderate incoordination, reluctant coordination, and moderate coordination. The number of counties with extreme incoordination linked to regions with a mountain ecosystem is increasing, where the ES supply is much greater than the demand. The moderate incoordination counties dominated by a cropland ecosystem exhibited slightly higher levels of ES supply than demand. The moderate and reluctant coordination were linked to counties with distinct ecological characteristics. Construction land played a major role in the characteristics of the CCD, followed by grassland. The interaction between construction land and all other factors significantly increased the influence on the CCD. These findings offered valuable insights for land managers to identify areas characterized by incoordination between ES supply and demand and understand associated factors to develop optimal ES management strategies.

1. Introduction

Ecosystem services (ESs) are defined as the benefits that humans directly or indirectly receive from ecosystems [1]. The Millennium Ecosystem Assessment (MA) [2], first conducted in 2005, established a framework for the global assessment of ecosystems. This framework divides indicators of ESs into four categories: provisioning, regulating, cultural, and supporting services. These categories are based on the connection between ESs and human well-being. Due to regional socio-economic factors and the rapid, high-intensity expansion of land for human use, ecological systems are facing continuous destruction [3]. This has caused a significant reduction in the ecosystem service supply (ESS), while the demand for a better living environment continues to grow [4]. Consequently, this has intensified the incoordination relationship between ESS and ecosystem service demand (ESD), negatively affecting sustainable development and human well-being [5]. Thus, it will be difficult to effectively manage and optimize regional ecosystems and encourage sustainable development via only the supply of ESs, while ignoring the human demand for ESs. Investigating the relationship between ESS and ESD not only addresses the challenges of sustainability arising from increasing human demand, but also establishes a solid theoretical foundation for adept and efficient ES management and utilization practices [6,7]. Therefore, understanding the relationship between ESS and ESD is essential for effective land use planning and decision-making processes that ensure the long-term resilience of ecosystems and social well-being.

As a closer relationship is established between ESs and human well-being, ES demand (ESD) gradually integrates with ESS. ESS is generally defined as the beneficial effect that ecosystems have on society [8]. The structure and function of ecosystems often result in simultaneous positive and negative changes in ESS [9]. However, there is no universally accepted definition of ESD. Currently, it is mainly defined from two perspectives. From a consumption perspective, ESD refers to the services provided by ecosystems that are useful to consumers, reflecting the actual demand for ESs [10]. From a preference perspective, ESD refers to the ecosystem services requested by social groups [11], reflecting not only the actual demand for ESs, but also the potential demand that cannot be met due to certain conditions. ESD includes the requirements, desires, or aspirations of human societies in relation to the benefits and contributions provided by ecosystems [12]. These needs arise from the dependence of human well-being and quality of life on the services and resources provided by natural ecosystems. ESD is driven by various factors, including population growth, economic activities, urbanization, and societal preferences. When the demand for ESs exceeds the natural capacity of ecosystems to provide them, it can lead to the overexploitation, degradation, or loss of those services. Conversely, if the supply of ES surpasses the demand, it can result in the underutilization or inefficient allocation of resources. Therefore, the relationship between ESS and ESD is crucial for creating a dynamic balance through which ecosystem products and services are transferred from ecosystems to social systems [13,14].

In recent years, an increasing number of studies have focused on evaluating the relationship between ESS and ESD, considering the coordination or conflict between regional social systems and ecosystems [15,16]. Thus far, these studies have primarily concentrated on quantifying and comparing the supply and demand of distinct ES indicators, such as water supply [17], flood regulation [18], air purification [19], and erosion control [20], and identifying the degree of mismatch, imbalance or incoordination between ESS and ESD. The majority of these investigations are centered around specific types of ecosystems, including forests [21], croplands [22], and urban regions [23]. These studies were mainly conducted in European contexts; however, in recent years, an increasing number of studies have been conducted in China, driven by ecological regionalization policies at different scales. Most previous studies assessed ESS, mainly focusing on quantifying and identifying the patterns and functions of ESs, as well as the effects of land use and land cover changes on ESs [24]. Compared to ESS, there are two main distinct approaches to ESD assessment: One is to evaluate the demand for each individual ES, and the other involves conducting a comprehensive assessment of the overall demand for ecosystem services. The land development index (LDI) has been widely used to comprehensively assess the ESD [25]. The rationale behind using the LDI to assess demand for ESD lies in its capacity to gauge the intensity of land development and its corresponding impact on the demand for these services [26]. Thus, ESS and ESD can be expressed from the perspective of the interaction between ecosystems and social systems; hence, the coordination relationship between ESS and ESD is a significant reflection of whether regional socio-economic structure and natural ecological backgrounds can develop in a coordinated manner [27].

Previous studies mainly analyzed temporal dynamic variations and spatial imbalances or mismatches in ESS and ESD [28,29], using methods such as modeling, mapping, participatory methods, etc. [30]. The concept of “coupling coordination” provides a framework for quantitating the coordination between ESS and ES [31,32]. The coupling coordination degree (CCD) is a measure used to assess the level of coordination and interdependence between different components or subsystems within a larger system [33]. In the context of ES, the coupling coordination degree model (CCDM) can be used to evaluate the level of coordination between the ecological system that provides services and the social system that demands and utilizes such services. It can also be applied to assess how effectively the ESS and ESD function together and whether their interactions are balanced and mutually beneficial. In recent years, many studies have applied the CCDM to analyze the relationship between ESS and ESD. For example, Guan et al. [34] provided valuable insights into the evolving characteristics of the spatial coupling between ESS and ESD using the CCDM. Li et al. [35] analyzed the dynamic characteristics of the supply and demand coupling of ESs in Lanzhou, China. Yang et al. [36] identified the coupling coordination relationship between sustainable development and ESs in Shanxi Province, China. Therefore, CCD is regarded as an efficient tool for researchers and policymakers to gain insights into the functioning of the coupled ecological and social systems and identify areas that require attention or intervention. Consequently, this helps us to understand the complex interactions between systems as well as design strategies for the sustainable management and conservation of ESs.

The coordination between ESS and ESD is influenced by the rapid development of regional socio-economic factors and the ecological factors, especially when socio-economic factors interact with natural and ecological factors [37]. Natural factors, land use/land cover, and socio-economic factors have been identified as the primary determinants of the relationship between ESS and ESD. For instance, Sun et al. [38] conducted an empirical study examining the correlations among 12 natural and socio-economic variables related to both ESS and ESD. Their study sheds light on the disparity between these two aspects within the United States. Wu et al. [16] analyzed the relationships between ES supply and demand and identified the effect of forest area on ESS, as well as the effects of per capita GDP, energy consumption per unit of GDP, and permanent population on ES demand in China. Peng et al. [39] systematically analyzed the impact of urbanization on ESs in metropolitan areas. Previous studies identified multiple influencing factors and encompassed different ESs. However, the majority of studies primarily focus on identifying the individual influencing factors of ESS or ESD. Alternatively, some studies have solely examined the influencing factors of ESS and ESD as substitutes for the actual relationship between the two [40]. As a result, there is a lack of studies investigating factors that directly impact the relationship between ESS and ESD. Research on the relative importance of socio-ecological drivers of coordination between ESS and ESD remains limited, and little attention has been given to the relationships between various drivers and the coordination between ESS and ESD, as well as associated spatial influences [41,42].

Shanxi Province is a typical resource-based area heavily reliant on coal and other resources for rapid economic growth. However, this excessive consumption of resources has created an exceedingly fragile ecosystem across the province. This increase in ecosystem degradation is having a negative effect on the economy and society, posing significant challenges to the sustainable development of Shanxi Province [43]. For example, intensive agriculture and mining activities are contributing to soil degradation in Shanxi. Erosion, the loss of topsoil, and contamination from pollutants can have serious consequences for agricultural productivity and ecosystem health. Unsustainable mining practices, particularly in coal mining, are causing land subsidence issues in Shanxi [44,45]. This phenomenon can lead to infrastructure damage, waterlogging, and the disruption of ecosystems [46]. In addition, some parts of Shanxi are vulnerable to desertification due to factors such as soil erosion, overgrazing, and unsustainable land use practices. This threatens agricultural productivity and ecosystem stability [47]. Since 2000, large-scale ecological restoration projects, such as the Natural Forest Protection Projects (NFPP) and the Grain for Green Program (GFGP), have significantly bolstered vegetation restoration in Shanxi Province. Efforts to address these ecosystem problems likely involve a combination of policy interventions, stricter environmental regulations, technological innovations, public awareness campaigns, and sustainable development practices.

Based on this context, this study aims to analyze the coupling coordination relationship between ESS and ESD and identify the associated socio-ecological driving variables in Shanxi Province. The objectives of this study are to (i) quantify and map the spatial distribution of the ES supply and demand, respectively; (ii) analyze the spatial–temporal characteristics of CCD between ESS and ESD based on CCDM; and (iii) determine the decisive influencing factors and the effects of interactions between factors using GDM. The results are anticipated to provide valuable information for achieving a harmonious balance between economic development and ecological restoration on a national scale within the provincial administrative units of China.

2. Materials and Methods

2.1. Study Area

Shanxi Province, located in the northern part of China (110°14′–114°33′ E, 34°34′–40°44′ N), covers an area of 156,700 km², accounting for 1.6% of the country’s territory (Figure 1a). It consists of 107 counties and is characterized by a typical mountain plateau terrain. Its topography is complex and diverse, including mountains, hills, plateaus, basins, and platforms. Mountains and hills make up 80% of its area, with altitudes ranging from 208 m to 2988 m above sea level (Figure 1b). The study area falls within the temperate continental monsoon climate region, the annual average temperature is between 4 °C and 14 °C, and the average annual rainfall is 468 mm. The dominant land use/land cover types include cropland, grassland, forest, and construction land (Figure 1c). Shanxi Province straddles the Yellow River basin and the Haihe River basin, and the river system is a self-generated outflow. The total population of the study area accounts for 2.48% of the national population and 1.71% of China’s total GDP [48].

Shanxi Province is one of the most important provinces with coal and mineral resources in China. As a region highly dependent on coal and mineral resources, its economic and social development is closely linked to its ecological environment [49]. However, excessive coal and mineral resource exploitation, the overexploitation of groundwater, rapid urbanization, and accelerated water and land resource exploitation have made the ecosystem in the study area extremely fragile. In 2019, the Chinese Government put forth the objectives of “ecological protection and high-quality development” for the Yellow River basin. As part of the Yellow River basin, in recent decades, Shanxi Province has faced daunting challenges in coordinating population, resources, ecosystems, and economic development.

2.2. Data Collection

In this study, we selected 2000 and 2020 as representative years to collect and analyze data. The data encompass both spatial and statistical information. The spatial data were processed at a grid cell resolution of 1 km × 1 km, while statistical data were aggregated at the county level.

Table 1. Datasets used in the study.

ES Variable	Data Type	Spatial Resolution	Data Source
Land use/land cover	Raster	30 m	National Geomatics Center of China (http://www.globallandcover.com/GLC30Download/index.aspx, accessed on 11 December 2020)
NDVI	Raster	250 m	National Aeronautics and Space Administration and United States Geological Survey (http://e4ftl01.cr.usgs.gov/MOLT/MOD13Q1.006/, accessed on 7 December 2020)
DEM	Raster	90 m	Geospatial Data Cloud (https://www.gscloud.cn/#page 1, accessed on 20 December 2020)
Meteorological data	Numeric	Sites	China Meteorological Data Sharing Service System (http://www.escience.gov.cn/metdata/page/index.html, accessed on 5 November 2020)
Soil database	Raster	30 arc-second	Harmonized World Soil Database (http://www.fao.org/soils-portal/soil-survey/soil-maps-and-databases/harmonied-world-soil-datebse-v12/en/, accessed on 1 December 2020)
Administrative map	Vector	County	National Geomatics Center of China (http://ngcc.sbsm.gov.cn/ngcc/, accessed on 11 December 2020)
Crop yield, GDP and population	Numeric	County	Statistical Yearbook

provides a comprehensive list of the primary data required to calculate the ES indicators. To ensure consistency, all spatial data were transformed to a common spatial reference system, specifically the WGS84 coordinate system and Albers equal-area conic projection. The flowchart depicted in Figure 2 illustrates the methodology that we employed to achieve our study objectives.

2.3. Quantifying ES Supply

2.3.1. Selection of ES Indicators

The selection of appropriate ES indicators is a critical step in assessing ecosystem services. In this study, we followed certain criteria for selecting ES indicators: (i) aligning with the classification of ES according to the Millennium Ecosystem Assessment [2] to ensure comparability with other studies; (ii) considering existing case studies conducted in Shanxi Province and selecting ES indicators that are closely related to the natural, ecological, social, and economic conditions of the study area; and (iii) the availability of the primary data required for evaluating ES indicators. Based on these criteria, our study focused on five key ES indicators relevant to the study area. These included one provisioning service (crop production), three regulating services (water retention, soil conservation, and carbon sequestration), and one cultural service (outdoor recreation).

2.3.2. Calculation of the ES Indicators

To quantify the selected ES indicators, we employed existing and widely used assessment models originally developed for this purpose. Specifically, the assessment of crop production was based on annual crop yield data [50]. The water balance equation served as a proxy for measuring water retention. The Universal Soil Loss Equation (USLE) model was used to calculate soil conservation [51]. Net primary production (NPP) was used as a proxy for carbon sequestration [52,53] and was assessed using the Carnegie–Ames–Stanford Approach (CASA) model, a widely adopted approach for NPP estimation [54,55]. The spatial distribution of individual ES indicators was visualized via mapping in ArcGIS.

To analyze the relationships between the five ES indicators, we employed ArcGIS 10.2, which serves as a common spatial unit, to aggregate all ES indicators at a national level. According to Raudsepp-Hearne et al. [56], administrative boundaries are suitable for identifying socio-ecological systems in a landscape, as management decisions at this level influence the provision and consumption of ES. The specific models and processes used for assessing the ES indicators are summarized in

Table 2. ES indicators from the MA categories and their quantitative methods, units, and ES variable requirements.

ES Indicator	Code	Model or Proxy	Unit	Assessment Process
Crop production	Cro	Crop yield per square kilometer	ton/km2	Crop production was calculated by dividing the crop yield of each county by its territory to illustrate per-unit provision service.
Water retention	Wret	Water balance equation	m3/km2	$TQ={\displaystyle \sum _{i=1}^j}(P_i-R_i-E{T}_i)$ TQ is water conservation, Pi is the annual average rainfall (mm), Ri is the annual average surface runoff (mm), and ETi is the annual evapotranspiration (mm).
Soil conservation	Scon	USLE (Universal Soil Loss Equation)	t/(hm2·a)	$\Delta A=R\times K\times L\times S(1-C\times P)$ ΔA = soil conservation (t/ (hm2·a)), R = rainfall erosivity index (MJ·mm/(hm2·h·a)), K = soil erodibility factor (t·hm2·h/ (MJ·mm·hm2)), L S = slope length and steepness factor (unitless), C = cover and management factor (unitless), P = conservation practice factor (unitless). The parameters R were from Wischmeier and Smith [57], K from Williams [58], L S from McCool et al. [59] and Liu et al. [60], C from Cai et al. [61], and P from Kumar et al. [62].
Carbon sequestration	Cseq	CASA (Carnegie–Ames–Stanford approach)	kg C/km2	$NPP=APAR\times \xi$ NPP = net primary productivity (g C/m2), APAR = absorbed photosynthetic active radiation (MJ/m), ξ = the utilization rate of light energy (g C/MJ).
Outdoor recreation	Rec	Tourists per square kilometer	persons/km2	Outdoor recreation was calculated via the area of forest land in each county.

.

2.3.3. Assessment of ES Supply Index

Since each ES has its own measurement unit, we individually standardized the ES values and then summarized them within each county to mitigate the influence of magnitude and variability. We employed min–max normalization to standardize the values of the five ES indicators [63,64]. This normalization method removes the units of the input data and scales them to a common range. After standardization, the standardized values were accumulated to obtain the ES supply index (ESSI), which represents the total ecosystem service supply. The calculation equation for ESSI is as follows:

$${\mathrm{E}\mathrm{S}\mathrm{S}}_{\mathrm{i}\mathrm{j}}=\frac{{\mathrm{E}\mathrm{S}}_{\mathrm{i}\mathrm{j}}-{\mathrm{m}\mathrm{i}\mathrm{n}\mathrm{E}\mathrm{S}}_{\mathrm{j}}}{{\mathrm{m}\mathrm{a}\mathrm{x}\mathrm{E}\mathrm{S}}_{\mathrm{j}}-{\mathrm{m}\mathrm{i}\mathrm{n}\mathrm{E}\mathrm{S}}_{\mathrm{j}}}$$

(1)

$${\mathrm{E}\mathrm{S}\mathrm{S}\mathrm{I}}_{\mathrm{i}}={\sum }_{\mathrm{j}=1}^{\mathrm{n}}{\mathrm{E}\mathrm{S}\mathrm{S}}_{\mathrm{i}\mathrm{j}}$$

(2)

where ESSI_i is the ES supply index of county i; ESS_ij is the standardized value for ES j of county i; ES_ij is the initial value for ES j of county i; max ES_j is the maximum value of ES over 107 counties, and min ES_j denotes the minimum value of ES_j; n = 5.

2.4. Quantifying ES Demand

The ES demand represents the human demand and preference for ecosystem products and services within a specific time period. In this study, we used a research method [26,65] to quantify ESD by considering land development intensity, population density, and gross domestic product (GDP) per area. This helped us to understand the coordination between the development and preservation of essential natural processes that sustain human well-being and environmental quality [29,66,67]. Specifically, land development intensity was measured as the percentage of construction land in the total land area. It reflects the intensity of the human consumption of ES. A higher percentage of construction land indicates a greater intensity of human land development in a given area, and consequently, a higher demand for ES. Population density serves as an indicator of the amount of ES demand. A higher population density corresponds to a greater ESD. GDP per area reflects the economic development of the region and indirectly indicates how much humans wish to consume or utilize ESs. Logarithmic methods were employed to remove fluctuations in the data. The ESD index is calculated using the following formula:

$${\mathrm{E}\mathrm{S}\mathrm{D}\mathrm{I}}_{\mathrm{i}}={\mathrm{D}}_{\mathrm{i}}\times \mathrm{l}\mathrm{g}\left({\mathrm{P}}_{\mathrm{i}}\right)\times \mathrm{l}\mathrm{g}\left({\mathrm{G}}_{\mathrm{i}}\right)$$

(3)

where ESDI_i is the ES demand index of each county i; Di, Pi, and Gi are the land development intensity (%), population density (person/km²), and GDP per area (yuan/km²) of county i, respectively.

2.5. Assessing Coordination between ES Supply and Demand

In this study, we employed the CCDM to investigate the interactive coordination relationship between ESS and ESD. The CCDM highlights the interdependence of ecological and social systems and aims to understand how they interact and mutually influence each other. By adopting this model, we can develop a holistic understanding of the relationship between ecosystems and human societies.

The CCDM recognizes the interconnectedness of ecological and social systems and emphasizes the importance of studying them together. This enables us to analyze the developmental pattern of these systems or indicators, progressing from disorder to order [68]. This representation provides insights into the overall effectiveness and synergistic impact between systems [69].

Mathematically, the CCDM is expressed as follows:

$${\mathrm{C}\mathrm{C}\mathrm{D}}_{\mathrm{i}}=\sqrt{{\mathrm{C}}_{\mathrm{i}}·{\mathrm{T}}_{\mathrm{i}}}$$

(4)

$${\mathrm{C}}_{\mathrm{i}}={\left\{\left({\mathrm{S}}_{{\mathrm{E}\mathrm{S}\mathrm{S}\mathrm{C}\mathrm{I}}_{\mathrm{i}}}·{\mathrm{D}}_{{\mathrm{E}\mathrm{S}\mathrm{S}\mathrm{D}\mathrm{I}}_{\mathrm{i}}}\right)/{\left[\left({\mathrm{S}}_{{\mathrm{E}\mathrm{S}\mathrm{S}\mathrm{C}\mathrm{I}}_{\mathrm{i}}}+{\mathrm{D}}_{{\mathrm{E}\mathrm{S}\mathrm{S}\mathrm{D}\mathrm{I}}_{\mathrm{i}}}\right)/2\right]}^2\right\}}^{1/2}$$

(5)

$${\mathrm{T}}_{\mathrm{i}}=\mathsf{\alpha }·{\mathrm{S}}_{{\mathrm{E}\mathrm{S}\mathrm{S}\mathrm{C}\mathrm{I}}_{\mathrm{i}}}+\mathsf{\beta }·{\mathrm{D}}_{{\mathrm{E}\mathrm{S}\mathrm{S}\mathrm{D}\mathrm{I}}_{\mathrm{i}}}$$

(6)

where CCD_i represents the coupling coordination degree of county i (0 ≤ CCD_i ≤ 1) between ESS and the ESSD; C_i refers to the coupling degree between ESS and the ESSD; T_i is the comprehensive development index of ESS and the ESSD; and S_ESSIi and D_ESDIi are the values of standardized ESSI and ESDI (0 ≤ SESSI_i ≤ 1, 0 ≤ DESDI_i ≤ 1). α and β are the weights to be determined; due to the equal importance of ESS and the ESSD in the coordination, α and β are given the same weight, that is, α = β = 0.5. Referring to previous research [32], we divide the CCD into five levels: When 0 ≤ CCD_i ≤ 0.20, the ESS and ESSD are in extreme incoordination; when 0.20 < CCD_i ≤ 0.35, they are in moderate incoordination; when 0.35 < CCD_i ≤ 0.55, they are in reluctant coordination; when 0.55 < CCD_i ≤ 0.70, they are in moderate coordination; and when 0.70 < CCD_i ≤ 1, they are in superior coordination.

2.6. Driving Variables of ES Coordination

2.6.1. Critical Driving Variables

In this study, we selected socio-ecological variables to explain the spatiotemporal differences between ESS and ESD, based on relevant research [70,71]. Potential explanatory variables were chosen from three sources: (1) the variables used to quantify ESS or ESD in our study, (2) variables identified in the literature as directly or indirectly driving individual ESs and/or their associations [72], and (3) variables for which quantitative data were available. After considering these factors, we ultimately selected thirteen potential socio-ecological variables, including natural variables such as elevation (DEM), slope (SLOPE), average annual precipitation (PRE), and average annual temperature (TEM); ecological variables such as NDVI, percentage of crop land (CROP), percentage of forestland (FOREST), and percentage of grassland (GRASS); and socio-economic variables such as percentage of construction land (CON), total population (POP), GDP, proportion of urban population (URBAN) and distance to the nearest county center (COUNTY) (

Table 3. Details of the driving variables for coupling coordination between ESSI and ESDI in this study.

Variable	Code	Description	Unit
Elevation	DEM	Derived from the SRTM3 global digital elevation model	Meter
Slope	SLOPE	Derived from the SRTM3 global digital elevation model	Degree
Precipitation	PRE	Annual trends of precipitation for the period 1956–2017	mm
Temperature	TEM	Annual trends of temperature for the period 1956–2017	°C
Normalized Difference Vegetation Index	NDVI	Vegetation cover	%
Cropland	CROP	County land area that is occupied by area that is classified as cropland	%
Forestland	FOREST	County land area that is occupied by area that is classified as forest	%
Grassland	GRASS	County land area that is occupied by area that is classified as grassland	%
Construction land	CON	County land area that is occupied by area that is classified as construction land	%
Population	POP	Annual total population	person
Economic level	GDP	Gross domestic product	yuan
Urbanization rate	URBAN	Urban population proportion	%
Distance to the nearest county	COUNTY	Distance to the nearest county center	km

).

2.6.2. Effects of Driving Variables on Coordination via Geo-Detector Model

The geo-detector model (GDM) can help identify the most influential factors or variables that contribute to specific spatial patterns, and reveal how different factors interact in a spatial context. It offers insights into the potential effects of human activities on ecosystems, water resources, or air quality. In this study, we employed the GDM to assess the spatial correlation between the explanatory variables and the dependent variables through spatial variance analysis (SVA) [73,74]. The GDM is a valuable analytical tool to identify and quantify the spatial associations between driving factors and specific outcomes. This insight can guide decision making by highlighting where interventions or ecological resource allocation should be focused for maximum impact. By employing statistical and spatial analytical techniques, the GDM enables researchers to identify dominant driving factors and their interactive effects, as well as explore spatial patterns and trends in complex geographical processes [75]. The fundamental assumption of the GDM is as follows: if an independent variable X significantly affects a dependent variable Y, then the spatial distributions of X and Y should exhibit similarity. SVA is used to compare the spatial consistency between the dependent variable and independent variables. Based on this comparison, the interpretation of independent variables in relation to the dependent variable can be quantified.

In this study, we utilized the “factor detector” module of GDM to identify the driving factor(s) that determine the distribution of CCD. This module identifies the extent to which the driving variables explain the spatial differentiation of CCD. The calculation results of the factor detector include the q-statistic and p-value. The q-statistic represents the influencing coefficient of the driving variable on CCD, with larger values indicating a stronger impact of the driving variable on CCD. The p-value indicates the significance level of the explanation, and a significance level of 0.1 (p-value < 0.1) is considered statistically significant. The formula for the factor detector is as follows:

$$q=1-\frac{{\sum }_{h=1}^l{N}_h{\sigma }_h^2}{N{\sigma }^2}$$

(7)

where q signifies the influencing coefficients of the driving variables for the ES (q -statistic), the values of which range from 0 to 1, where 0 corresponds to no correlation between the two and 1 to CCD’s complete dependence on a driving variable. σ² is the variance of the CCD, and N is the size of CCD. The superposition of the driving variables and CCD forms L layers in CCD, which are indexed by h = 1, 2… l, and N_h and ${\mathsf{\sigma }}_{\mathrm{h}}^2$ represent the scale and variance of layer h, respectively.

The “interaction detector” module of GDM was used to examine whether two factors have a stronger or weaker effect on ESs than they do independently. The types of interactions between two variables are as follows:

Enhance: if q (D₁ ∩ D₂) > q (D₁) or q (D₂)

Enhance, bivariate: if q (D₁ ∩ D₂) > q (D₁) and q (D₂)

Enhance, nonlinear: if q (D₁ ∩ D₂) > q (D₁) + q (D₂)

Weaken: if q (D₁ ∩ D₂) < q (D₁) + q (D₂)

Weaken, univariate: q (D₁ ∩ D₂) < q (D₁) or q (D₂)

Weaken, nonlinear: if q (D₁ ∩ D₂) < q (D₁) and q (D₂)

Independent: if q (D₁ ∩ D₂) = q (D₁) + q (D₂)

where the symbol “∩” denotes the intersection between the layers D₁ and D₂. The attributes of layer (D₁ ∩ D₂) are determined by the combination of the attributes of layer D₁ and D₂ using a spatial overlay to form a new layer. q (D₁), q (D₂), and q (D₁ ∩ D₂) were calculated using Equation (1). By comparing the sum (q (D₁) + q (D₂)) of the factors’ contribution to two individual attributes (q (D₁), q (D₂)) with the contribution of the two attributes when combined (q (D₁ ∩ D₂)), the interactive effects of the two factors can be defined using the above seven types.

3. Results

3.1. Spatial–Temporal Patterns of ES Supply and Demand

Using a quantitative method, we calculated ESSI and ESDI data for the 107 counties of Shanxi Province for the reference years 2000 and 2020. We normalized the two indices to a range of 0–1 and mapped them to facilitate comparisons (Figure 3 and Figure 4).

The unique geographical location of Shanxi Province resulted in significant variations in natural conditions, leading to considerable heterogeneity in the counties’ ability to provide ESs. The ESSI exhibited substantial variation across Shanxi Province in both 2000 and 2020, which was similar to the distribution characteristics of outdoor recreation (Figure 3). Areas with high ESS were dispersed across the mountainous regions of the study area, surrounded by counties with higher ESSI. Specifically, Mount Taiyue, Mount Zhongtiao, and Mount Wangwu, located in the southern parts of the province, have a particularly abundant supply of ESs. The areas with low supply were concentrated in the northwest edge of Shanxi and the western part of Mount Luliang. From 2000 to 2020, the spatial pattern of ESSI remained largely unchanged in most counties. However, there was a slight overall decrease in the level of ESS, and the ESSI of Yuanping and Xinzhou, located in the northwestern part of Shanxi Province, experienced significant increases. There was a clear upward trend in the number of counties with the lowest supply, increasing from 20.6% for the 107 counties in 2000 to 32.7% in 2020. However, the number of counties with the highest supply remained static.

The spatial distribution hierarchy of ESDI was weaker compared to ESSI (Figure 4). In both 2000 and 2020, ESDI displayed spatial distribution characteristics with higher values in the central areas and lower values in the outer areas of the study area, which closely resemble the spatial distribution of construction land. The areas with the highest and higher ESS grades were primarily concentrated in central Shanxi, southeastern areas, and the northern plains, indicating a relatively concentrated distribution. The areas with the lowest demand were contiguous and distributed in mountainous regions such as Mount Luliang in the west and Mount Taihang in the east. From 2000 to 2020, due to population growth and economic development in the study area, there was a clear increase in ESDI. The number of counties with higher grades of ESS increased, while the counties with the lowest demand decreased. Some areas with medium demand in 2000 shifted to higher demand categories by 2020, indicating a transformation from relatively lower to higher grades. Overall, the spatial distribution of ESDI displayed noticeable differences between the outskirts and the middle regions of the study area.

3.2. Coupling Coordination Characteristics of ES Supply and Demand

Using the CCDM, we measured and mapped the CCD of ESSI and ESDI for the 107 counties of Shanxi Province in 2000 and 2020 (Figure 5). The CCD values ranged from 0 to 0.57 in 2000 and from 0 to 0.43 in 2020 (

Table 4. The statistical values of standardized values of ES supply, ES demand, and coupling coordination degree in 2000 and 2020.

	ESSI		ESDI		CCD
Year	2000	2020	2000	2020	2000	2020
Minimum value	0.528	0.512	0.649	0.638	0.002	0.059
Maximum value	0.859	0.870	0.979	0.980	0.566	0.427
Mean value	0.670	0.649	0.044	0.067	0.212	0.196

), indicating a relatively low level of coupling coordination.

There are four main types of coupling coordination relationships between ESSI and ESDI: extreme incoordination, moderate incoordination, reluctant coordination, and moderate coordination (Figure 5). Most counties belong to the extreme incoordination and moderate incoordination patterns, mainly located in the western and eastern parts of Shanxi. In 2000, the areas with reluctant coordination were primarily found in the Taiyuan Basin in central Shanxi, Linfen Basin, and the Yuncheng Basin in the southeastern parts. However, in 2020, reluctant coordination areas were sparsely distributed in only a few municipalities. Overall, there was a clear incoordination relationship between ES supply and demand in Shanxi.

Between 2000 and 2020, the number of counties with extreme incoordination significantly increased from 29.0% to 36.5%. The number of reluctant coordination areas decreased from 34.6% to 24.3%, with fourteen counties transitioning from reluctant coordination to moderate incoordination. This indicated a substantial decline in the coupling coordination between ESSI and ESDI in Shanxi over the 20-year period. The main change characteristics were the negative transitions from relatively high coordination grades to incoordination grades. The most significant changes were observed in the shifts from moderate incoordination to extreme incoordination and from reluctant coordination to moderate incoordination, accounting for 20.31% and 16.28% of the counties, respectively.

In general, the analysis of coordination between ES supply and demand revealed coexisting states of coordination and incoordination, with incoordination being predominant in most counties. The difference in the spatial polarization of CCD in 2020 was more significant than in 2000. Overall, the coordination relationship between ES supply and demand in Shanxi Province deteriorated between 2000 and 2020.

3.3. Determining Drivers for the Coupling Coordination Degree between ES Supply and Demand

The GDM was utilized to identify the most influential socio-ecological drivers for the coupling coordination between ES supply and demand. The factor detection results for thirteen socio-ecological variables yielded the influencing coefficients (q-statistic values) and significance levels (p-values) (Figure 6).

In 2000 and 2020, the variables SLOPE, PRE, and COUNTY did not pass the significance tests. Meanwhile, the variable TEM was a statistically significant driving variable (p-value < 0.001) in 2000, but its significance level was greater than 0.1 in 2020. Conversely, the variable URBAN was not significant in 2000 (p-value > 0.1) but became a significant driving variable in 2020 (p-value < 0.05). Figure 6 presents the sorting results for the q-statistic values for significant variables (p-value ≤ 0.05), revealing their influencing coefficients on CCD in 2000 and 2020. The influencing coefficients of CON, POP, and GDP were greater than 0.5 and higher than the other factors in both 2000 and 2020. This demonstrates that these variables were the main drivers of the spatial pattern of CCD during both time periods. GRASS, CROP, and DEM were considered sub-high determinate variables based on their q-statistic values. Notably, the variables TEM, NDVI, and FOREST had weak effects on CCD in 2000, whereas in 2020, the variables URBAN, FOREST, and NDVI had a weak effect. This suggests that the impact of forest coverage on the coordination relationship was relatively small in the study area.

Furthermore, the interaction detector module of GDM assessed the influencing coefficients of any two socio-ecological variables (Figure 7) and compared them with their separate influencing coefficients (q-statistic values). The results reveal two interaction modes of socio-ecological variables on CCD: nonlinear enhancement and mutual enhancement. This indicates that the explanatory power of the interaction between any two variables for CCD is greater than that of any single variable. The interactive effects on CCD between CON and the other variables were the strongest, with q-statistic values exceeding 0.85. After interacting with POP and GDP in both 2000 and 2020, the q-statistic values of all variables were above 0.65, which was higher than the values for the separate effects of POP and GDP on CCD. It is important to note that SLOPE, PRE, and COUNTY did not exhibit significant effects on CCD in the single-factor detection results (Figure 6). However, after interacting with CON, POP, and GDP, the influencing coefficients of SLOPE, PRE, and COUNTY were 0.88, 0.67, and 0.68 in 2000, respectively, and similar results were observed in 2020. This suggests that even if individual socio-economic variables do not have a significant effect on the spatial distribution of CCD, they may play a key role via interaction with variables that have high influencing coefficients. Thus, SLOPE, PRE, and COUNTY were identified as important external driving factors for CCD between ESSI and ESDI. Additionally, this highlights the importance of considering the interactions between variables in understanding the spatial distribution of CCD.

4. Discussion

4.1. Spatial–Temporal Characteristics of ESSI and ESDI

Our framework of analysis was used to make sense of relationships between ES supply (ESS) and ES demand (ESD) in a complex social–ecological system. Utilizing the ES supply index (ESSI), we have derived the spatiotemporal variation characteristics of ESS (Figure 3). These characteristics provide valuable insights into assessing the level of ES supply, as well as tracking its enhancement or decline in every county within Shanxi Province.

We found that high values of ESSI were predominantly observed in mountainous areas with dense forest coverage. In contrast, low values were mainly found in the Fen River basin, which is dominated by cropland and stretches from northeast to southwest. The Fen River basin is the most active area in terms of social and economic activities in Shanxi Province. The extensive human disturbance in this basin, characterized by high population density and economic growth, has led to the rapid expansion of construction land, causing a reduction in ecological land. As a result, the supply of ESs in the counties within this watershed is relatively low. Similarly, the northwestern parts of the study area, dominated by cropland and grassland, also exhibited relatively low ESSI values. This aligns with findings from previous studies [76]. This region represents a transition zone from scrub steppe to typical steppe, and factors such as increasing population, livestock, and desertification posed significant threats to the area for an extended period. Despite the implementation of the GFGP in these northwestern counties of Shanxi Province since 2000, the supply of local ESs continues to deteriorate [77]. Hence, the grass-planting-based GFGP in this region is required to improve efficiency and achieve ecological protection. Additionally, the western area of Mount Luliang exhibited relatively low ESSI values. This can be attributed to Mount Luliang’s location in the loess hilly and gully region, which is characterized by severe soil and water loss [78]. Moreover, these mountainous areas are contiguous to regions with concentrated mineral resources and frequent human activities. As of 2021, there are 91 coal mines in Luliang, accounting for 13.62% of the total number of coal mines in Shanxi. Extensive mining methods and inadequate management measures have imposed a burden on ecosystems, leading to a diminished supply level of regional ESs.

Concerning the ES demand index (ESDI), we carefully curated three indicators representing land use, population dynamics, and economic factors (Figure 4). These indicators collectively offer a comprehensive representation of ESD, encompassing preferences and requirements. We found that high values of ESDI were primarily distributed in the central Taiyuan Basin, northern Datong Basin, southern Linfen Basin and Yuncheng Basin, and southeastern Changzhi Basin, while low values were observed in the mountainous areas on the east and west sides. These basin regions, with Taiyuan, Datong, Linfen, Yuncheng, and Changzhi as their central cities, have experienced significant population and industrial concentration, resulting in high demand for ESs. Notably, the spatial pattern of ESDI closely resembles that of the degree of land use development (Figure 3). This finding is consistent with existing studies [79] that indicate a strong correlation between ESDI distribution and land use development degree, population density, and per capita GDP. It is a common phenomenon in many parts of China to expand the area of construction land in order to meet social and economic needs, particularly in areas with extensive human activity [16]. From 2000 to 2020, the average ESDI values significantly increased in the Datong Basin, decreased in the Taiyuan Basin and Yuncheng Basin, and the disparity in ESD values among the counties within these basins significantly decreased. Additionally, ESDI levels in Taiyuan city and Changzhi city district, which had the highest ESDI values in the whole study area, remained relatively stable between 2000 and 2020. This stability can be attributed to the slower economic and population growth in Taiyuan and Changzhi, reaching a stable socio-economic agglomeration state within Shanxi Province [16]. Overall, Shanxi Province exhibits clear spatial mismatch characteristics between ESS and ESD.

4.2. Spatial–Temporal Characteristics of CCD between ESSI and ESDI

The coordination mechanism between ESS and ESD primarily revolves around the harmonization of the ecosystem and social system [80]. It involves a feedback loop, where societal choices, such as land use and economic activities, impact the provision of ESs. In turn, the condition and health of the ecosystem shape the quality and quantity of services that can be supplied to meet social needs. Thus, it is rare to gain a clear spatial relationship between the biophysical supply of ESs and their demand; it is a rare achievement. This rarity primarily stems from the utilization of distinct measurement units for assessing supply and demand [78]. Furthermore, this situation underscores a current challenge within this field of study. In this study, we introduced the CCD model to address this challenge, and obtained the spatiotemporal variation characteristics of the coordination relationship between ESS and ESD in various counties of the Shanxi Province.

The majority of counties exhibited a state of incoordination in terms of CCD (Figure 5). Extreme incoordination was predominantly observed in Mount Taihang in the east, Mount Luliang in the west, and Mount Taiyue in the south of Shanxi Province. This can be attributed to the higher supply of the five key ESs compared to the demand in these counties (Figure 3 and Figure 4). Most of these counties are situated in mountainous areas characterized by forest and grassland with high vegetation coverage [36]. Due to the implementation of ecological protection projects, there is minimal interference from human activities in these counties, resulting in little change in the type and quantity of land use. Furthermore, NFPP and GFGP contribute to increased vegetation coverage in these mountainous areas, enhancing the types and capacities of ESs. As a result, a high level of “lock-in effect” of regional ESS has been achieved [35,76], and ESS in these mountains is significantly weaker compared to the plain and basin areas. On the other hand, moderate incoordination was primarily observed in the Datong Basin, Xinding Basin, Taiyuan Basin, Linfen Basin, Yuncheng Basin, and Changzhi Basin, spanning from the north to south of the study area. This can be attributed to the extensive human activity in these basin regions and indicates the need for additional efforts to improve the coordination between ESS and ESD in these areas.

The coordination relationship between ESSI and ESDI exhibited a decrease in incoordination between 2000 and 2020, as evidenced by an increase in the number of extreme incoordination counties and a decrease in reluctant coordination and moderate coordination counties. For example, the relationship between ESSI and ESDI in Taiyuan, Linfen, Yuncheng, and Changzhi cities shifted from moderate coordination in 2000 to reluctant coordination in 2020. No other counties changed to a moderate coordination, resulting in the absence of moderate coordination counties/districts in the study area in 2020. This shift can be attributed to the slower growth rate of ESS compared to ESD in these cities. The increase in population, expansion of construction land, and economic growth have led to higher ESD in these districts [81]. Despite the increase in green spaces within the cities between 2000 and 2020 due to China’s ecological civilization construction projects, it falls short of meeting the substantial demand for ES in urban production and daily life [82].

Overall, the distribution of CCD in 2020 exhibited noticeable spatial differences between the basin region and the mountain region compared to 2000. The coordination relationship between ESSI and ESDI demonstrates incoordination and spatially varies across Shanxi Province. The degree of incoordination intensified in nearly half of the counties from 2000 to 2020. In 2020, the CCD highlighted the spatial disparities between urban areas, agricultural areas, and forest–grassland areas from the perspective of land use, as well as the differences between valleys, basins, and mountains from the perspective of terrain.

4.3. Associations between CCD and Driving Covariates

In our study, we employed a geo-detector model (GDM) to capture the spatially response characteristics of the dependent variable in relation to the independent variable. This approach allowed us to analyze how different factors contribute to the observed spatial patterns of the coordination between ESS and ESD across the study area (Figure 6 and Figure 7). We found that socio-economic factors had a greater impact on the coordination relationship between ESS and ESD than natural and ecological factors, emphasizing the significance of socio-economic factors in shaping the spatial pattern of this relationship. This finding aligns with previous research that highlighted the role of socio-economic variables as determinants of ES supply and demand distribution [43,83]. The dominant factors influencing the spatial pattern of CCD were identified as construction land, followed by population and GDP. The expansion of construction land, driven by extensive human activity, has significantly influenced land use patterns in terms of magnitude, type, and distribution [84]. The complex nature of human activity further complicates these relationships. However, these findings contrast with a study by Yang et al. [82] in China’s Loess Plateau, which found that vegetation cover had the greatest positive effect on the relationship between ES supply and demand. This discrepancy could be attributed to the dominant influence of vegetation coverage on both ESS and ESD in the Loess Plateau.

Interestingly, the effect of grassland on CCD was second only to socio-economic factors, ranking below elevation and cropland. Grassland emerged as the vegetation cover factor with the strongest impact on CCD, indicating its significant role in shaping the spatial pattern of the coordination relationship between ESS and ESD in Shanxi Province. This finding is supported by previous studies that emphasized the influence of land use changes caused by socio-economic factors on ESs [85]. Notably, the effect of forestland and NDVI on CCD was much smaller compared to grassland, suggesting that the increase in grassland area through GFGP mainly influenced the relationship between ES supply and demand in Shanxi Province. According to data from the 2020 “Shanxi Province Third Land Survey Main Data Bulletin”, grassland covers an area of 3.11 million hectares in Shanxi Province, and is primarily distributed in Datong, Xinzhou, Luliang, Jinzhong, and Linfen cities, accounting for 73% of the province’s grassland [86]. However, the grassland ecosystem in the study area remains fragile, with 70% of the grassland experiencing varying degrees of degradation and facing challenges such as insufficient protection and restoration, low utilization and management efficiency, and a lack of effective technological support. Therefore, the coordination degree between ESS and ESD could be improved by implementing quantitative and spatial adjustments to grassland planting policies.

In this study, the slope and precipitation factors were found to have no significant direct effect on CCD. However, after interacting with construction land, population, and GDP, slope and precipitation played an important role in shaping the spatial pattern of CCD. This indicates that socio-economic factors enhance the influence of natural factors on CCD. Generally, slope and precipitation influence the supply capacity and demand preferences for ESs by controlling the spatial distribution of human activity and landscapes [76], and previous studies demonstrated that precipitation facilitates coordination between ESS and ESD in arid and semi-arid areas [79]. The changes in terrain and precipitation affect both social and economic processes, leading to changes in coordination relationships. Previous research provided guidance on identifying factors that contribute to ES supply and demand mismatch [87], ESS, and ESD, including terrain ruggedness for supply and population density for demand [88].

The spatial distribution of the coordination relationship between ESS and ESD is primarily supported by the expansion of construction land, population growth, and economic benefits. However, this puts immense pressure on ES supply. Although forest and grassland coverage significantly increased between 2000 and 2020, the expansion of construction land has outpaced these gains. The growth rates of forestland and grassland in the study area were 4.2% and 4.3%, respectively, while the growth rate of construction land was as high as 105.4%. Since 2000, Shanxi Province has been the subject of economic system reforms, and thus rapid economic development, but the lack of awareness regarding ecological protection has resulted in the degradation of ESs. Consequently, the coordination relationship between ES supply and demand deteriorated somewhat between 2000 and 2020, indicating that the expansion of construction land and the concentration of population and industry have threatened the coordination between ESS and ESD in the counties of Shanxi. To address this issue, policymakers in Shanxi Province must make significant progress in promoting the coordination between ESS and ESD, implementing measures to ensure a dynamic balance between ES supply and demand.

4.4. Limitations

There are several limitations and uncertainties in this study. Firstly, the evaluation of five types of ES was limited in terms of reflecting the overall ESS level due to data availability and quality constraints [89]. In addition, the equal-weight superposition calculation method used to calculate total ESS may have overlooked significance of various ES types in Shanxi Province. Future studies should strive to include a more comprehensive range of ES indicators and consider their relative importance. Secondly, the selection of indicators to represent ESD focused on land development intensity, population density, and GDP per area, assuming that all types of ESs have the same demand. This oversimplification may not accurately reflect spatiotemporal changes in ESD. Further research is needed to refine the measurement of ESD and capture its dynamics. Thirdly, this study compared the changes in ESS and ESD between 2000 and 2020, overlooking the temporal volatility of ES. Incorporating temporal dynamics would enhance our understanding of the coordination relationship between ES supply and demand.

The use of county-level administrative boundaries was advantageous [90], providing official statistical data; however, this limits our understanding of causality and spatial heterogeneity in coordination relationships [91]. Shanxi Province has diverse topography, vegetation and natural geographical features, and there is a significant spatial mismatch between ESS and demand. Thus, it is difficult to reveal the variation of ESS at a local scale in the spatial units of county. In addition, spatial heterogeneity and scale effects impact the relationship between ESS and ESD [92]. Focusing only on a single scale tends to miss information about the correlation between scales, and the influence mechanism is inevitably one-sided. Future research should use a multi-scale analysis to capture the correlation between scales and comprehensively investigate influence mechanisms.

Additionally, our study assumed that ESD in a county is solely provided by the local ecosystem without considering the flow of ecosystem services across county boundaries. For example, water resources can originate from upstream regions, and food shortages in a city can be mitigated via food trade [93]. Future studies should account for the cross-boundary flow of ES, as services from neighboring ecosystems can also contribute to ESD. Furthermore, while GDM helped identify the strength of influence of socio-ecological factors on the spatial pattern of CCD, it was unable to capture whether this effect was positive or negative. Exploring the positive or negative nature of these effects would provide a more comprehensive understanding of the coordination relationship between ESS and ESD.

5. Conclusions

Shanxi Province has experienced a stage of rapid expansion of construction land; there has been a rapid transformation in the intensity, type, and pattern of land use, which has created an urgent need to optimize ESs, social progress, and economic development. Under this background, this study analyzed spatiotemporal changes in ES supply, demand, and their coordination relationship, and identified the relevant socio-ecological driving factors across 107 counties in Shanxi from 2000 to 2020.

The results reveal that the spatial pattern of ESS was closely linked to forest coverage, while ESD was closely related to the degree of land use development. The changes in ESS and ESD exhibited spatial heterogeneity. Over the study period, Shanxi Province experienced a slight decrease in ESS and an increase in ESD. The evaluation using the CCDM demonstrated significant incoordination between ESS and ESD in Shanxi, which worsened between 2000 and 2020. Based on the CCD of ESS and ESD, Shanxi Province was divided into four zones: extreme incoordination, moderate incoordination, reluctant coordination, and moderate coordination. The extreme incoordination zone was mainly located in mountainous regions, where ecosystems dominated by the eastern Taihang Mountain and western Luliang Mountain provided high levels of ESs but low ESD. The moderate incoordination zone was primarily found in basins with intensive human activity, where ecosystems dominated by cropland exhibited slightly higher levels of ESs compared to ESD. The reluctant coordination and moderate coordination zones were mainly located in central cities within basins, where the CCD between ESS and ESD significantly decreased. The spatial distribution of CCD was primarily influenced by construction land, population, and GDP, with grassland playing a secondary role, largely driven by the GFGP policy.

Accordingly, in order to promote the sustainable development of the counties, we propose the following recommendations: Firstly, the government should enhance ecological compensation policies for residents in mountainous regions, with a special focus on areas like Taihang Mountain and Luliang Mountain. This entails raising compensation for ecosystem service providers. Secondly, county and municipal district administrations should adopt advanced strategies for balanced ecological, social, and economic development. This will bolster land use efficiency and curb haphazard expansion of construction projects. Thirdly, policymakers and governments need to comprehensively assess the distinct impacts of various drivers on the interplay between ESS and ESD at both the local and county scales. This is particularly pertinent when formulating strategies for regional management approaches. These policy insights are applicable not just to Shanxi Province, but also to other regions endowed with coal and mineral resources. In future research, it is imperative to prioritize effective land management decisions that account for the interrelation between socio-ecological factors and the supply and demand of ecosystem services.