Evaluation of ecosystem services from 2000 to 2020 and their trade-offs/synergies in a coalfield: a case study in the Pingshuo mining area of the Loess Plateau, China

Determining the spatiotemporal dynamics in land use and ecosystem service value (ESV) and understanding the trade-offs/synergy relationships between ecosystem services (ESs) are crucial for ecosystem management and achieving sustainable development in mining areas. However, existing research on ESV and ESs has not paid sufficient attention to the special coalfield in arid/semiarid areas. In this study, we investigated the Pingshuo mining area and used the standard equivalent factor to evaluate ESV variations resulting from the spatiotemporal changes of land use based on remote sensing and land use data (2000, 2010, and 2020). Simultaneously, the trade-offs/synergies between ESs were further explored using the ESs trade-offs/synergies degree (ESTD) model. We found that (1) the land use changed considerably in this area, which was mainly reflected in farmland decrease and built-up land increase by 3580.60 hm2 and by 5103.44 hm2, respectively, from 2000 to 2020. (2) ESV in our study area declined by 7116.53 × 104 RMB yuan over the investigated period. High ESV mainly appeared in the north/south of the study area, while the low ESV was concentrated in the middle and northeast/southeast of the mining area. (3) The interactions between paired ESs were mainly the synergies that generally appeared among eight ESs except soil conservation, while trade-offs mainly existed between soil conservation and other paired ESs in the study region. To achieve socio-economic and ecological benefits, the local government should take effective measures to improve the environment of the coalfield and reverse the falling tendency of ESV.


Introduction
Ecosystem services (ESs) are defined as the various products and benefits directly or indirectly obtained from nature and are categorized into provisioning, supporting, regulating, and cultural services (MA 2005;Pan et al. 2021;Wang et al. 2021). ESs provide a conceptual link between the natural ecosystem and human society (Bandana et al. 2021), which are closely associated and interact with each other (Takeuchi 2010;Gao et al. 2021). The natural ecosystem provides the environmental conditions and the basic resources for human well-being (Daily 1997;Nelson et al. 2009;Mohsen and Azam 2020). The development of the human society improves people's ability to transform natural ecosystems and promotes rapid economic development (Mendoza-Gonz´alez et al. 2012;Gao et al. 2021). In recent decades, ESs have become one of the most important issues in sustainable development and utilization of natural resources (Mohsen and Azam 2020).
Since the 1970s, ESs have been a much-discussed topic in ecology. Ecologists have conducted numerous studies concerning the definition, valuation, driving mechanism, and management strategies of ESs (Su et al. 2018;Bandana et al. 2021;Ma and Wen 2021). A milestone in the study of ESs was the evaluation of the theory and method of ESs on the global scale by Costanza et al. (1997) (Schaefer et al. 2015;Fu et al. 2016;Wang et al. 2018;Pan et al. 2021). Based on the assessment results of Costanza et al. (1997Costanza et al. ( , 2014, Xie et al. (2008Xie et al. ( , 2017 presented a more practical and specific method for estimating terrestrial ESs in China through expert consultation. Globally, researchers have investigated ESs at different spatial scales (e.g., global, national, provincial, and city) (Costanza et al. 1997(Costanza et al. , 2014Xie et al. 2008Xie et al. , 2017Goldstein et al. 2012;Bateman et al. 2013;Wang et al. 2018;Ma and Wen 2021) and ecosystem types (e.g., mountains, basins, rangelands, and coasts) (Wang et al. 2017Mohsen and Azam 2020;Bandana et al. 2021;Gao et al. 2021). The most common outcomes have related ESs to ecosystem services value (ESV) assessment and spatiotemporal patterns as well as to trade-offs/synergies and human wellbeing. However, these studies have not elucidated ESV in coalfields, and very few researchers have considered the trade-offs/synergies among the ESs of mining areas (Su et al. 2018;Jiang et al. 2021). Thus, existing theoretical knowledge of mining areas is insufficient for the management of ESs.
Mining areas represent a unique semi-natural and semiartificial ecosystem under the influence of resource exploitation. The long-term development and utilization of resources in these areas result in a series of negative impacts on the natural environment, such as land destruction, soil erosion, water pollution, and vegetation degradation. These greatly alter the structure and function of the natural ecosystem in the mining area and restrict the stable and sustainable development of the regional social economy. Therefore, it is essential to conduct basic ESs research in mining areas, which will help optimize decision-making and relevant policies for the maintenance of ecosystem functions in the process of resource development, minimize the degradation of ecosystems, ensure the availability of resources necessary for economic development, and ultimately realize the harmonious development of the regional economy and environmental protection (Cao et al. 2021).
The Pingshuo mining area is the largest open-pit coal mining area in northern China and is located at the edge of the Loess Plateau and the farming-pastoral ecotone. The mining area represents a microcosm of the ecological changes occurring in coalfields in China. Therefore, the study of ESs in the Pingshuo mining area is representative and demonstrative. Accordingly, the Pingshuo mining area was selected as the focus area in this study, and spatiotemporal changes in ESs and their trade-offs/synergies were examined by introducing remote sensing data, land use data, and mathematical models. The main objectives of this study are as follows: (1) to quantify the land use changes in the Pingshuo mining area during 2000-2020, (2) to analyze the spatial and temporal change process for ESV, and (3) to identify the trade-offs/synergies between pairs of ESs. Our findings will aid in the restoration of the ecological environment and the sustainable development of coal resources not only at a regional scale in the Loess Plateau, but also on the global scale.

Study area
The Pingshuo mining area is located in the north of Shanxi Province (112°05′E-112°29′E, 39°18′N-39°37′N); it has a total area of 380 km 2 and consists of three open-cut (Antaibao, Anjialing, and East Surface Mines) and three underground coal mines (Well Mine I, II, and III) (Fig. 1). The landscape in this region is gently undulating, and the elevation increases from south to north, ranging from 1095 to 1700 m (Xue et al. 2013). The region is characterized by a semiarid continental monsoon climate, with dry winter, windy spring, warm summer, and cool autumn. The average annual air temperature varies from 5.0 to 7.5 °C, the annual mean rainfall ranges from 410 to 450 mm (mainly from June to September), the annual average evaporation is between 1750 and 2550 mm, and the mean annual wind speed is 5.0 ms −1 (Xue et al. 2021). The main soil types in the study area are Calci-Ustic Isohumosols and Haply-Udic Cambosols (CRGCST 2001), with low organic matter content and weak anti-erodibility, making it susceptible to severe water and wind erosion (Xue et al. 2013). Consequently, this region has become an ecologically fragile area in the Loess Plateau.
The Pingshuo mining area represents the overlap zone between the eastern Loess Plateau and the farming-pastoral ecotone of northern China. The Pingshuo coalfield has a proven geological reserve of approximately 12.60 billion tons, making it the largest and most modernized surface and underground combined coal mine in China (https:// pings huo. china coal. com/). Human economic activities are complex and prominent in this region. The mining area is an agricultural-industrial-mining ecotone area characterized by the comprehensive effects of agriculture, mining, and urbanization (Long et al. 2021). Therefore, the Pingshuo mining area is a typical and representative area to study the change of ESs caused by coal mining subsidence.

Data collection
Spatial and statistical data covering the study area from multiple sources were used in this study. Meteorological station data of temperature and precipitation were obtained from the national Meteorological Scientific Data Sharing Platform (http:// data. cma. cn). The land use datasets were based on Landsat TM/ETM/OLI images of the Pingshuo mining area from 2000, 2010, and 2020, which had a resolution of 30 m and were downloaded from the Resources and Environment Data Cloud Platform of the Chinese Academy of Sciences (http:// www. resdc. cn/). The images were processed following the methodology used by Ma and Wen (2021). The images were classified into six land use types, namely, farmland, forest land, grassland, water areas, built-up land, and unused land. The overall accuracy and the Kappa coefficients were above 88.60% and 0.91, respectively. The digital elevation model (DEM) was obtained from the ASTER Global Digital Elevation Model (ASTER GDE; http:// gdem. ersdac. jspac esyst ems. or. jp/). Socioeconomic statistical data (e.g., grain yield, sown area, and agricultural products market prices) in 2000-2020 were obtained from Shuozhou Statistical Yearbooks by Shuozhou Municipal Bureau of Statistics (http:// sztj. shuoz hou. gov. cn/) and Compilation of National Agricultural Product Cost-benefit Data by National Bureau of Statistics of China. The coefficients of ESV were primarily extracted from the research of Xie et al. (2017), and ESV coefficients of built-up land were proposed by Long et al. (2021).

Determination of ESV
Based on Costanza et al. (1997Costanza et al. ( , 2014, Xie et al. (2008Xie et al. ( , 2017 summarized an equivalent factor of ESV per unit area suitable for the evaluation of ESV in China by conducting a questionnaire investigation of 500 Chinese scholars. Based on the data in MA (MA 2005) and the study results of Costanza et al. (1997), ESs in this study were categorized into four types and nine sub-types (food production-FP, material production (MP), air regulation (AR), climate regulation (CR), hydrologic adjustment (HA), purified environment (PE), soil conservation (SC), biodiversity protection (BP), and aesthetic landscape (AL)) ( Table 1). The economic value of the standard equivalent factor of ESV is equal to 1/7 of the regional mean market value of grain produced by a hectare of farmland (Xie et al. 2003). Corn is the main food crop cultivated in the Pingshuo mining area (Yang et al. 2018). The average yield of corn was 5425.50 kg.hm −2 , and average price was 1.20 Yuan.kg −1 from 2000 to 2020. Based on these, the equivalent factor of ESV was 930.09 Yuan.hm −2 (Xue et al. 2021). Subsequently, the ESV equivalents per unit area of different land use types in the Pingshuo mining area were determined based on this calculation (Table 1). The total value of ESs was the sum of the four functional services provided by the natural ecosystem and was calculated as follows (Gao et al. 2021;Pan et al. 2021): where ESV is the total ESV (Yuan.y −1 ); A i is the area of the i-th land use type (hm 2 ); VC ij is the value coefficient (VC) of i-th land use type with the j-th ESs (Yuan.hm −2 .y −1 ); i is the number of land use types; j is the number of ESs types; m = 6; n = 9.

Contribution rate of ESV
The contribution rate (CR) of ESV refers to the influence of the ESV changes of each land use type on the total ESV changes in the monitoring period. It can reveal the main contributing and sensitive factors affecting the total ESV changes in the study area. The CR is calculated as follows  where CR it is the CR of i-th land use type in t-th time period and ∆ESV it is the ESV changes of i-th ESs land use type in t-th time period (Yuan.y −1 ); m = 6.

Sensitivity analysis of ESV
Considering VC' uncertainties, the VC for each land use type was adjusted by 50%. The sensitivity analysis model was introduced to estimate ESV results (Kreuter et al. 2001;Bryan et al. 2018), which is expressed as: where CS represents sensitivity coefficient; ESV and ESV' are initial and adjusted total estimated ESV (Yuan.y −1 ), respectively; and VC i and VC i ' refers to initial and adjusted VC for i-th ESs land use type (Yuan.hm −2 .y −1 ), respectively. If CS is more than 1, it implies that ESV is elastic to VC; if CS is equal to 1, indicates that ESV is unit elastic to VC; if CS is less than 1, it indicates that ESV is inelastic to VC. Furthermore, the closer the absolute value of CS is to 0, the higher is its accuracy.

Trade-offs/synergies degree of ESs
The model of ecosystem services trade-offs/synergies degree (ESTD) reflects the direction and degree of action between two ESs, which is used to evaluate the overall interaction of ESs changes. The equation of the model is as follows : where ESTD ij represents the ESTD of i-th and j-th of ESs; ESC ib and ESC jb are the changes of i-th and j-th of ESs in the b period, respectively; ESC ia and ESC ja are the changes Provisioning services, food production (FP) and material production (MP); regulating services, air regulation (AR), climate regulation (CR), hydrologic adjustment (HA), and purify environment (PE); supporting services, soil conservation (SC) and biodiversity protection (BP); and cultural service, aesthetic landscape (AL) of i-th and j-th of ESs in the 'a' period, respectively. ESTD less than 0 indicates the trade-offs of i-th and j-th of ESs. ESTD more than 0 represents the synergies of i-th and j-th of ESs. Simultaneously, the absolute value of ESTD implies the change degree of i-th of ESs compared with that of the j-th.

Spatiotemporal dynamic changes of land use
Analyzing dynamic changes in land use can help reveal the influence of land use changes on ESs (Gao et al. 2021). From 2000 to 2020, changes in land use were particularly observable in the Pingshuo mining area (  Figure 2 illustrates the land use spatial patterns of the Pingshuo mining area from 2000 to 2020. Farmland was always the main land use type, and its spatial distribution had changed little from 2000 to 2010. However, farmland decreased significantly from 2010 to 2020 and was mainly transformed into built-up and unused lands in Antaibao, East, and Anjialing Surface Mines. Forest and grass lands were mainly distributed in Well Mines I and III. However, they were converted into farmland in Well Mine I, and builtup and unused lands in Well Mine III. The distribution of built-up and unused lands increased dramatically in the last 20 years, as they spread gradually from the central to northeast/southeast regions. In 2000, built-up and unused lands were mainly identified in the middle and south of the study area. In 2010, their expansion in the northeast region was more than that in the other regions; built-up and unused lands in 2020 increased sharply, especially in the central and northeast regions, along Antaibao and East Surface Mines.

Spatiotemporal variation of ESV
From 2000 to 2020, the total ESV in our study area showed a shrinking trend (Fig. 3), and decreased from 18,381.93 × 10 4 yuan in 2000 to 11,265.40 × 10 4 yuan in 2020, with a decrease rate of 38.72%. Simultaneously, the ESV of farm, forest, grass, and built-up lands underwent a significant reduction, with a total decrease of 1469.04 × 10 4 , 776.64 × 10 4 , 2677.76 × 10 4 , and 4066.29 × 10 4 yuan, respectively. In the primary classification of ESs, the growth rates of provisioning, regulating, supporting, and cultural services were − 13.63%, − 98.75%, 20.28%, and − 25.92%, respectively (Table 3). Among them, the proportion of supporting and regulating services are the largest and smallest, respectively. In the secondary classification of ESs, only the ESV of soil conservation increased, with the growth rate of 33.34%. All the other eight types of ESV exhibited a negative growth period, among which hydrologic adjustment (− 406.07%) and food production (− 12.64%) decreased the fastest and slowest, respectively. The growth rates of material production, air regulation, climate regulation, purified environment, biodiversity protection, and aesthetic landscape were − 15.46%, − 35.17%, − 26.96%, − 193.08%, − 27 .22%, and − 25.22%, respectively. Overall, the total and individual ESV in the Pingshuo mining area generally demonstrated a decreasing trend, and their variations were obvious.
Based on Fig. 3 and the CreateFishnet module in Arc-GIS 10.2 (http:// www. esri. com/), we obtained the spatial distribution map of ESV at the grid scale (Fig. 4). It became evident that: in 2000 and 2010, extremely high ESV (1607.21-9814.77 yuan) and high value ESV (748.21-1607.20 yuan) in the study region were mainly distributed in Well Mine III and Well Mine I, respectively. It was mainly concentrated in Antaibao, East, and Anjialing Surface Mines. Therefore, the spatial distribution of ESV was nearly consistent with that of the land use types.

Contribution rate
The results of CR in the study period are listed in Table 4. The CR of built-up land, grassland, and farmland were found to be smaller, i.e., − 42.49%, − 26.75%, and − 15.32%, respectively. The ESV in the Pingshuo mining area decreased by 7116.53 × 10 4 yuan (38.71%) in the last 20 years. This decrease was attributed to the diminution of ESV in built-up land, grassland, and farmland. Therefore, these land cover types were recognized as the main contributing and sensitive

Sensitivity analysis
Sensitivity analysis is an efficient approach for examining the dependence of ESV on land use (Gao et al. 2021). The sensitivity coefficients (CS) of ESV and VC in the Pingshuo mining area are shown in Table 5. It is evident that CS values for different land cover types were less than 1 and changed little in different periods, which implies that the total ESV estimated in our study region was inelastic relative to the VC. Additionally, these results revealed that the change in the corresponding VC had a minor impact on ESV in the Pingshuo mining area. The adjusted equivalent factor can reasonably evaluate the fluctuation of ESV in this study. The CS values of farmland, grassland, and built-up land were higher than those for the other land use types, indicating that these 3 land types greatly impacted ESV.

Trade-offs/synergies among ESs
We obtained the trade-offs/synergies between ESs in the study area from 2000 to 2020 (Fig. 5A) based on the model of ESs trade-offs/synergies degree (ESTD) constructed in "Methods" and the ESTD calculation method described in Eq. (4). Of the 81 pairs of ESs based on nine secondary classifications, 16 and 65 pairs had negative and positive correlations, respectively. The synergies correlation accounted for 80.25%, indicating that it was the dominant relationship among ESs in the Pingshuo mining area. In the primary classification of ESs, these synergistic relations mainly existed between provisioning and regulating services, and cultural and regulating services. In the secondary classification of ESs, the degree of synergy between hydrologic adjustment and aesthetic landscape, hydrologic adjustment and material production, and hydrologic adjustment and food production were larger, at 28.89, 20.66, and 13.63, respectively. The trade-off degree between soil conservation and aesthetic landscape, soil conservation and material production, and soil conservation and food production were larger, at − 9.87, − 7.06, and − 4.66, respectively. The same tradeoffs/synergies correlations between paired ESs also appeared in 2000-2010 and 2010-2020 ( Fig. 5B and C), except for ESTD, which showed a slight difference among ESs. As shown in Figs. 5, the value of ESTD between paired ESs increased and decreased from 2000 to 2020. For example, among the regulating services, the synergistic relationships between gas regulation and other paired ESs first increased and then decreased. Similarly, the synergies between climate regulation and other paired ESs first decreased and then increased. The synergistic relationships Table 3 The changes of individual ESV in the Pingshuo mining area during 1990-2020 Definitions of acronyms given in Table 1 Types between HA, PE, and other paired ESs did not change substantially. In supporting services, the trade-off relationships between soil conservation and other paired ESs first increased and then decreased. The synergies between BP and other paired ESs first decreased and then increased. Therefore, the trade-offs/synergies relationships among ESs in the Pingshuo mining area were rather complicated.

Changes in land use and ESV
Analysis of land use showed that the land use dynamics of the Pingshuo mining area in the study period changed considerably, which is mainly reflected in the substantial decrease in farmland/grassland and the growth of built-up land. This is consistent with the findings of other studies conducted in arid/semiarid regions (Abulizi et al. 2017;Wei 2017;Pan et al. 2021). Farmland decreased by 3580.6 hm 2 from 2000 to 2020 (Table 2), with a decrease rate of 11.46%. This reduction was mainly driven by the large-scale development of coal resources and the rapid population growth (Xue et al. 2013(Xue et al. , 2021Gao et al. 2021). In the study area, three surface mines were successively put into production (Antaibao Surface Mine in 1987, Anjialing Surface Mine in 2005, and East Surface Mine in 2016, and the raw coal production increased by 68.50 million t (from 12.50 million t to 81.00 million t) in 2000-2020 (2020 Shuozhou statistical yearbooks), with an increase rate of 5.48 times. According to the population data of Shuozhou statistical yearbook, the population of the Pingshuo mining area increased by 6.85 × 10 4 ,  with a growth rate of 9.13 folds, from 2000 to 2020. With the rapid increase in coal development and local population, the local government has had to expand the built-up land to construct factory, transport, and living facilities to meet the needs of the coal industry and livelihood of the people. This has resulted in resource exploitation in arid/semiarid areas in China (Abulizi et al. 2017;Pan et al. 2021). The total area of land in the Pingshuo mining area remained unchanged  Table 1 during the study period. Thus, the increase in built-up land is mainly due to the conversion of other land cover types. The artificial nature of land use in the study area has become apparent over the last 20 years, indicating that human activities, particularly coal exploitation, have disturbed the land cover.
As shown in Fig. 3 and Table 3, ESV in the Pingshuo mining area had experienced a decreasing trend, and land use and value coefficient of ESs were the main factors influencing ESV variations (Yang et al. 2018;Gao et al. 2021). Farmland and grassland were mainly transformed into built-up land, leading to a significant decline in the ESV of farmland and grassland. In addition to not providing ESV, built-up land requires considerable resources to meet the needs of the people, thereby leading to a decrease in ESV. Therefore, ESV of mining areas in arid/semiarid regions are extremely sensitive to changes in land use (Xue et al. 2013;Rao et al. 2018;Pan et al. 2021). Furthermore, the value coefficient of ESs was another influencing factor that restricted ESV. The decreasing equivalent factor was mainly due to the change in land use and the considerable reduction of farmland area; in particular, the planting area of corn was reduced. Furthermore, the downturn of the corn market and increased cost of corn production decreased the net profit per unit corn. Therefore, the land use change, net income of corn, and change in the corn market were the most important factors that contributed to the change of ESV in the Pingshuo mining area.
Over the past 20 years, the local government realized the negative effects of coal development on the ecological environment, and initiated a strategy to implement a land rehabilitation project with the vegetation area improved by 4000 hm 2 . However, the disturbed environment in the coalfield could not be restored to its original natural state in the short term (Xue et al. 2013;Yang et al. 2018); consequently, ESV showed a downward trend during 2000-2020. To achieve economic and ecological benefits, the local government should take various measures (e.g., protective exploitation, artificial restoration, and ecological compensation) to improve the environment of the coalfield and reverse the decreasing tendency of ESV.

Changes in trade-offs/synergies among ESs
Under the influence of natural and anthropogenic factors, especially coal mining activities, the landforms and land cover in the coalfield underwent a considerable and irreversible change (Doley and Audet 2013;Xue et al. 2013;Yang et al. 2018). In the Pingshuo mining area, where both coal surface mining and well mining are practiced, the surface landscape is severely disturbed and reshaped as a result of mining subsidence, and is characterized by land degradation, water pollution, biodiversity loss, buildings destruction, and mining pits and gangue. This will inevitably intensify the land use conflict, damage ecosystem functions, and affect their respective ESs (Lechner et al. 2016;Xiao et al. 2018;Li et al. 2021). ESs in the Pingshuo mining area interacted with and restricted each other, forming the association relationships among ESs with unique characteristics. In this study, analysis of ESs interactions showed that synergies were dominant among ESs, while trade-offs mainly existed between soil conservation and other paired ESs. These findings were consistent with those of previous studies (Chisholm 2010;Li et al. 2021).

Research deficiencies and prospects
Although this study analyzed the spatiotemporal changes of land cover and ESV and trade-offs/synergies among ESs according to the true state of the Pingshuo mining area, there were some limitations and uncertainties. First, the value coefficient of ESs was used to estimate ESV in this area, and its accuracy was evaluated by the sensitivity analysis model. However, the determination of the equivalent coefficient of ESV is subjective, which may have led to an inaccurate ESV assessment result (Troy and Wilson 2006;Liu et al. 2020;Pan et al. 2021). Therefore, it is necessary to construct a more comprehensive evaluation method of the equivalent coefficient of ESV to eliminate these impacts in the future. Second, this study only discussed the trade-offs/synergies among ESs in our study area and did not further analyze their driving factors (e.g., natural, human, and policy factors). Therefore, to further guide ecosystem management in the coalfield, future research is essential to obtain appropriate models and data to perfect the internal processes and driving mechanisms of trade-offs and synergies among ESs.

Conclusions
In this study, we analyzed the spatiotemporal dynamics of land use and ESV in the Pingshuo mining area and the tradeoffs/synergies relationships between ESs from 2000-2020. The research can be summarized as follows: the land use types changed considerably in this area, especially farmland and built-up land. The transformation of farmland in the middle and northeast/southeast of the study area was the main contributor to the expansion of built-up land. The total ESV of this area always exhibited a trend of decline over the study period. This change is attributed to the decrease in farmland and the increase in built-up land. The interactions among ESs were mainly synergistic, and generally existed between provisioning and regulating services. The present study promotes a clear understanding of the characteristics of ESV and interactions among ESs in coalfield in arid/ semiarid area and provides a basis for the formulation of ecosystem management policies and restoration programs. To reverse the decreasing trend of ESV, the local government must take measures to improve the environment and promote human welfare in the coalfield while achieving sustainable regional development.