Changes of Ecosystem Service Value in the Water Source Area of the West Route of the South–North Water Diversion Project

Abstract

To ensure water source security and sustainability of the national major strategic project “South-to-North Water Diversion”, this study aims to evaluate the spatio-temporal evolution characteristics of the ecosystem service value (ESV) in its water source area from 2002 to 2022. This study reveals its changing trends and main influencing factors, and thereby provides scientific support for the ecological protection and management of the water source area. Quantitative assessment of the ESV of the region was carried out using the Equivalence Factor Method (EFM), aiming to provide scientific support for ecological protection and resource management decision-making. In the past 20 years, the ESV has shown an upward trend year by year, increasing by 96%. The regions with the highest ESV were Garzê Prefecture and Aba Prefecture, which increased by 130.3% and 60.6%, respectively. The ESV of Xinlong county, Danba county, Rangtang county, and Daofu county increased 4.8 times, 1.5 times, 12.5 times, and 8.9 times, respectively. In the last two decades, arable land has decreased by 91%, while the proportions of bare land and water have decreased by 84% and 91%, respectively. Grassland had the largest proportion. Forests and grasslands, vital for climate regulation, water cycle management, and biodiversity conservation, have expanded by 74% and 43%, respectively. It can be seen from Moran’s I index values that the dataset as a whole showed a slight positive spatial autocorrelation, which increased from −0.041396 to 0.046377. This study reveals the changing trends in ESV and the main influencing factors, and thereby provides scientific support for the ecological protection and management of the water source area.

1. Introduction

China’s South–North Water Diversion Project, a critical hydraulic infrastructure initiative, mitigates regional water disparities and promotes equitable economic development across regions. The unimplemented Western Route aims to resolve western water shortages through inter-basin transfer from the Yangtze to the Yellow River. The ecosystem service value (ESV) can measure regional ecological quality, which is the foundation of ecological protection and an important basis for ecological engineering decision-making [1]. Accurately calculating ESV is crucial for scientifically managing ecosystems [2]. This study focuses on the water source area and explores the impact of water diversion on the ESV of the water source area, providing a research basis for ecological compensation after the implementation of the project.

Ecosystem services are the various goods and services that natural systems provide, directly or indirectly, to human societies through their processes and structures [3]. Spearheaded by the United Nations, the Millennium Ecosystem Assessment (MEA) constituted the inaugural global effort to systematically appraise ecosystem services’ conditions and trajectories. It laid a conceptual foundation for ecosystem service science and significantly advanced the development of valuation methodologies worldwide. Among these, monetary valuation methods have gained prominence for translating ecosystem functions into quantifiable economic indicators, thus offering valuable support for policy-making, natural resource management, and the design of ecological compensation mechanisms. Such monetary assessments not only help reveal the hidden economic contributions of ecosystems to regional development, but also serve as key metrics for evaluating ecological civilization progress and sustainability performance at the regional level [4].

The quantitative assessment of ESV primarily relies on two methods: the functional value method and the equivalent factor method [5]. While the functional method offers greater theoretical accuracy, it requires extensive ecological and socioeconomic data, making it less practical for large-scale or data-scarce areas [6]. Owing to its straightforward methodology and extensive relevance, the equivalent factor approach has seen predominant adoption. Costanza’s study pioneered ESV estimation using unit-area values of land use categories. However, the global averages used limit its direct applicability in ecologically diverse regions such as China [7,8]. To address this, Xie Gaodi et al. [9] developed a localized equivalent factor system tailored to China’s ecological conditions. This approach is now extensively applied across various spatial scales in China, including watersheds [10], provinces [11,12], and cities [13], supporting ecological compensation, land use planning, and sustainability efforts.

ESV encompasses four key categories: cultural, regulating, provisioning, and supporting services [14]. The majority of ESV [15] focuses on agricultural lands, forests, wetlands, and so on. Grassland and forest degradation and biodiversity loss threaten the stability of the regional ecological environment [16,17]. This assessment method can effectively guide the restoration of ecosystems, maintain ecological security, and inform land use planning [18]. The ecological environment in the western region of China is fragile. Frequent alterations to land use patterns have undermined the inherent configuration and equilibrium of regional ecosystems, leading to a degradation of ecosystem service capacity, amplifying ecological vulnerabilities, and ultimately jeopardizing the prospects for sustainable development [19,20]. The change in land use involved multiple factors, which was a complex process [21]. The water source area is sensitive to both climate change and human activities [22]. As critical safeguards for preserving ecosystem integrity, water retention capacity, and environmental sustainability, quantifying the ESV of key eco-functional areas helps in the rational planning of land use and implementation of good environmental protection measures, ultimately achieving the goal of ecological economic coordination, stability, and sustainable development.

With this background, the present study was conducted with three objectives: (i) to develop an ESV assessment model suitable for the Western Route region; (ii) to analyze the spatiotemporal patterns and evolution of ESV from 2002 to 2022 using Moran’s I index; and (iii) to explore the spatial response mechanisms of ESV under land use change.

2. Methods

2.1. Study Area and Region Division

For this study, three provincial-level administrative regions—Sichuan, Yunnan, and Qinghai provinces—were selected, covering 9 prefecture-level administrative divisions and 47 county-level administrative units (Figure 1). The study area is primarily located in the northwestern plateau region of Sichuan Province, characterized by its remote geographical location, underdeveloped economy, and fragile ecological environment.

Three typical years of 2002 (Figure 2a), 2012(Figure 2b), and 2022 (Figure 2c) of land use status data were selected, classifying land use types into six categories based on regional characteristics: cropland, forest, grassland, wetland, bare land, and water bodies. This study merged similar land use types to obtain six types of land. The farmland ecosystem in the research area includes dry land and paddy fields, collectively referred to as farmland; woodland, shrubland, sparse woodland, and other forest types, etc., are integrated into a forest ecosystem; high, medium, and low coverage grasslands, etc., are merged into grassland ecosystems; lakes, rivers, reservoirs, etc., are merged into aquatic ecosystems; bare land and unused land form two additional separate ecosystems. This article focused on the study of natural ecosystems and therefore did not consider construction land, such as urban land, rural residential areas, and tourist attractions. The spatial raster data resolution was 30 m, using the CGCS1984 coordinate system.

2.2. Ecosystem Service Value Assessment

This study quantifies the ESV of the research area by adapting Xie Gaodi’s “equivalent value per unit area” table, which is integrated with the ecological characteristics of regions such as Garzê and Aba in the Western Route Project area. Data sources included land use planning documents, statistical yearbooks, and peer-reviewed literature. In these regions, ecosystems such as farmland, grasslands, and forests have significant contributions to the local social, economic, and ecological environment services; therefore, accurately assessing their ESV has high practical value.

The equivalent factor method quantifies ESV by establishing a benchmark unit’s ESV equivalent factor. This standard serves to gauge relative contributions across ecosystems [9]. The standard equivalent of this study is defined as the economic value of the annual natural grain output of farmland with a national average yield of 1 hectare, and all subsequent studies are based on this parameter; for example, according to the research method of Xie [9], the net profit of food production in agricultural ecosystems is used as a standard equivalent factor for the value of ecosystem services. The calculation formula is as follows:

$$E_{\mathrm{a}}={\displaystyle \frac{1}{7}}{\sum }_{\mathrm{e}=1}^{\mathrm{h}}{\displaystyle \frac{m_{\mathrm{e}}{p}_{\mathrm{e}}{q}_{\mathrm{e}}}{M}}$$

(1)

$$m_{\mathrm{e}}={\displaystyle \frac{1}{\mathrm{h}}}{\sum }_{\mathrm{i}=1}^{\mathrm{h}}{m}_{\mathrm{i}}$$

(2)

$$p_{\mathrm{e}}={\displaystyle \frac{1}{\mathrm{h}}}{\sum }_{\mathrm{i}=1}^{\mathrm{h}}{p}_{\mathrm{i}}$$

(3)

$$q_{\mathrm{e}}={\displaystyle \frac{1}{\mathrm{h}}}{\sum }_{\mathrm{i}=1}^{\mathrm{h}}{q}_{\mathrm{i}}$$

(4)

$$M={\displaystyle \frac{1}{\mathrm{h}}}{\sum }_{\mathrm{i}=1}^{\mathrm{h}}{M}_{\mathrm{i}}$$

(5)

where E_a is the unit equivalent factor economic value, CNY/ha; p_e is the average price of grain, CNY/kg; q_e is the yield per unit area of grain planted, kg/ha; m_e is the area for planting crops, ha; M is the total planting area of major crops, ha.$m_{\mathrm{i}}$, $p_{\mathrm{i}}$, $q_{\mathrm{i}}$, and $M_{\mathrm{i}}$ are the area planted with e crop in year i (hm²), the national average price (CNY/t), the production per unit area (t/hm²), and the area planted with all crops in year i (hm²), respectively.

$$ESV={\sum }_{n=1}^i{A}_i{S}_i{E}_a$$

(6)

where A_i refers to the coefficient for land type; S_i represents the land area; and ESV represents the ecosystem services value.

2.3. Data Sources

The statistical data for 2002, 2012, and 2022 (

Table 1. Statistical data sources from 2002 to 2022.

Data	Data Sources
Land use data	Institute of Geographic Sciences and Resources, Chinese Academy of Sciences
	Resource Environmental Science and Data Center
	China City Yearbook
Crop-related data	China Statistical Yearbook (2002–2022)
National average price of food crops	Compilation of National Data on Costs and Returns of Agricultural Products (2002–2022)

) were sourced from the national compilation of agricultural product cost–benefit data and China statistical yearbook and city yearbooks. The specific value of the standard equivalent can be calculated and applied to quantify and compare different ecosystem services.

2.4. Moran’s I Index

Inter-county ESV spatial distributions within the source basin were analyzed through Moran’s I autocorrelation testing [23]. Moran’s I index measures the degree to which similar values tend to cluster together in spatial data, and is used to reflect the overall spatial autocorrelation strength throughout the study area with a single value. ${\left(x_{\mathrm{e}}-\overline{x}\right)}^2$: The deviation between the values of different regions multiplied by spatial deviation and the average value. $\left(x_{\mathrm{e}}-\overline{x}\right)\left(x_{\mathrm{j}}-\overline{x}\right)$: Multiplying the deviations of two regions (i and j), the positive or negative value of this product represents the degree of similarity between the two regions.$W_{\mathrm{e}\mathrm{j}}$: representing the spatial relationship between region e and region j, and weighting the “trend similarity” between all adjacent regions. Moran’s I is −1~1. Moran’s I > 0 signifies spatial clustering, I < 0 denotes dispersion, and values approaching 0 suggest random spatial distribution. The calculation formula is as follows:

$$I={\displaystyle \frac{n}{{\sum }_{e=1}^n{\sum }_{j=1}^n{W}_{\mathrm{e}\mathrm{j}}}}\times {\displaystyle \frac{{\sum }_{e=1}^n{\sum }_{j=1}^n{W}_{\mathrm{e}\mathrm{j}}\left(x_{\mathrm{e}}-\overline{x}\right)\left(x_{\mathrm{j}}-\overline{x}\right)}{{\sum }_{e=1}^n{\left(x_{\mathrm{e}}-\overline{x}\right)}^2}}$$

(7)

where I is Moran’s I index; n is the number of spatial units; x_e and x_j are the attribute values of the location where evaluation units e and j are located, respectively; and W_ej is the spatial weight adjacency matrix.

3. Results and Discussion

3.1. Changes in Land Use Types

The changes in land use types identified by remote sensing images from 2002 to 2022 are shown in

Table 2. Changes in land use types from 2002 to 2022.

Land Use Types	2002		2012		2022
	Area (ha)	Proportion (%)	Area (ha)	Proportion (%)	Area (ha)	Proportion (%)
Cultivated land	390,800	13.63	56,400	1.86	33,800	0.95
Forests	621,700	21.68	641,900	21.20	1,084,300	30.51
Grasslands	1,691,800	58.98	2,306,600	76.19	2,415,500	67.96
Wetland	25	0.00	17	0.00	15	0.00
Bare land	86,500	3.01	10,000	0.33	13,700	0.39
Water bodies	77,400	2.70	12,300	0.41	6700	0.19

. In 2002, the majority of land use types were cultivated land, forests, and grasslands, accounting for 390,800 ha, 621,700 ha, and 1,691,800 ha, respectively, with proportions of 13%, 21%, and 59%. In 2012, the majority of land use types were forests and grasslands, accounting for 641,900 hectares and 2.3066 million hectares, respectively, with proportions of 21% and 76%. In 2022, the majority of land use types were forests and grasslands, accounting for 1.0843 million hectares and 2.4155 million hectares, respectively, with proportions of 30% and 68%. Cultivable acreage has plummeted 91% over the last two decades, and the proportions of bare land and water have decreased by 84% and 91%, respectively. Forests and grasslands showed an increasing trend, increasing by 74% and 43%, respectively. National ecological conservation policies and land use reconfigurations partially explain observed land use transitions. Notably, sustained implementation of initiatives like “Grain for Green” (GFG) has accelerated the transformation of cropland into forest and grassland ecosystems [24,25]. This pattern is consistent with findings from other ecologically vulnerable areas in western China, where policy interventions and ecological restoration efforts have jointly contributed to the gradual recovery of natural ecosystems [26].

Meanwhile, the substantial reduction in barren land and water bodies may be influenced by a combination of factors, including climatic variability, hydrological regulation, and potential limitations in remote sensing classification accuracy. To better understand the underlying drivers of these changes, future research should consider integrating ground-based observations with meteorological and hydrological datasets to enhance the explanatory power of land use change analyses.

3.2. Ecosystem Service Value of Districts in the Water Source Area

Figure 3a–e illustrate the 2002 ESV distribution across the water source area, totaling CNY 108.45 billion. Garzê Prefecture’s ESV exceeded Aba Prefecture’s during this period. The lowest among them was Jinchuan county; the highest was Daofu county. Xiaojin county and Luhuo county were both relatively high, at CNY 10.97 billion and CNY 7.51 billion, respectively. Service value composition: provisioning 7%; regulating 68%; supporting 21%; cultural 4%. Regulating and supporting services collectively accounted for 89% of total ESV. The largest proportion of regulating services was hydrological regulation and climate regulation, while the largest proportion of supporting services was biodiversity. The largest proportions of provisioning services were water resource supply and raw material production. The provisioning service was CNY 7.248 billion, with Jinchuan county having the lowest provisioning value and Daofu county having the highest value of CNY 679 million. Xiaojin county, Luhuo county, Luding county, and Xinlong county have higher values of CNY 670 million, CNY 422 million, CNY 321 million, and CNY 318 million, respectively. The value of regulatory services is CNY 68.589 billion, with Jinchuan county having the lowest regulatory value, Daofu county having the highest, and Xiaojin county, Luhuo county, Xinlong county, Luding county, and Hongyuan county having higher regulatory values of CNY 6.816 billion, CNY 4.75 billion, CNY 3.569 billion, CNY 3.147 billion, and CNY 2.78 billion, respectively. The values of supporting services and cultural services were CNY 21.092 billion and CNY 4.274 billion, respectively. The distribution of the highest supporting services and cultural services was consistent across the counties. Daofu county, Xiaojin county, Luhuo county, Xinlong county, and Aba county had higher supporting services, at CNY 2.731 billion, CNY 2.336 billion, CNY 1.593 billion, CNY 1.041 billion, and CNY 964 billion, respectively.

Figure 4a–e depict the 2012 ESV distribution across the water source area, totaling CNY 158.45 billion, with Garzê Prefecture exhibiting higher ESV than Aba Prefecture. The lowest was Seda county; the highest was Xinlong county. Danba county and Banma county were both relatively high, at CNY 24.91 billion and CNY 23.346 billion, respectively. The proportion of provisioning services, regulating services, supporting services, and cultural services was 8%, 64%, 23%, and 5%, respectively, with regulating services and supporting services accounting for 87% of the value of ecological services. Hydrological and climate regulation dominated regulating services, while biodiversity constituted the primary supporting service. The largest proportions of provisioning services were water resource supply and raw material production. The provisioning services value was CNY 12.747 billion, with Songpan county having the lowest provisioning services value and Xinlong county having the highest value of CNY 2.592 billion. Banma county, Danba county, Daofu county, and Rangtang county have higher values: CNY 1.91 billion, CNY 1.889 billion, CNY 1.85 billion, and CNY 1.588 billion, respectively. The value of regulatory services was CNY 101.596 billion, with Songpan county having the lowest regulatory value and Xinlong county having the highest value, at CNY 20.508 billion. Danba county, Daofu county, Bama county, and Rangtang county have higher regulatory values of CNY 16.109 billion, CNY 14.783 billion, CNY 14.782 billion, and CNY 12.28 billion, respectively. The value of supporting services and cultural services was CNY 36.704 billion and CNY 7.407 billion, respectively. The distribution of the highest supporting services and cultural services was consistent among the counties, with Xinlong county, Danba county, Bama county, Daofu county, and Rangtang county having higher supporting services, at CNY 7.57 billion, CNY 5.754 billion, CNY 5.536 billion, CNY 5.454 billion, and CNY 4.601 billion, respectively.

The ESV of the water source area in 2022 was also investigated (Figure 5a–e). The ESV in 2022 was CNY 198.518 billion, and the ESV of Garzê Prefecture was greater than that of Aba Prefecture. Lixian county had the lowest ESV; Xinlong county had the highest. Daofu county and Danba county were both relatively high, at CNY 31.547 billion and CNY 25.203 billion, respectively. Ecosystem service values comprised provisioning (6%), regulating (65%), supporting (24%), and cultural (5%) categories, where regulating and supporting services collectively constituted 89% of total ESV. Hydrological/climate regulation dominated regulating services, while biodiversity represented the primary supporting service component. The largest proportions of provisioning services were water resource supply and raw material production. The supplying service value was CNY 12.126 billion, with Songpan county having the lowest value and Xinlong county having the highest at CNY 1.988 billion. Daofu county, Danba county, Rangtang county, and Banma county have higher values of CNY 1.912 billion, CNY 1.48 billion, CNY 1.476 billion, and CNY 1.414 billion, respectively. The value of regulating services was CNY 129.871 billion, with Songpan county having the lowest regulating value and Xinlong county having the highest at CNY 207.44 billion. Daofu county, Danba county, Rangtang county, and Banma county have higher regulatory values of CNY 20.605 billion, CNY 16.583 billion, CNY 15.374 billion, and CNY 14.364 billion, respectively. The values of supporting services and cultural services were CNY 47.039 billion and CNY 9.482 billion, respectively. The distribution of the highest supporting services and cultural services was consistent among the counties, with Xinlong county, Daofu county, Danba county, Rangtang county, and Banma county having higher supporting services, at CNY 7.671 billion, CNY 7.516 billion, CNY 5.944 billion, CNY 5.696 billion, and CNY 5.386 billion, respectively.

The ecosystem service values of different cities in the water source area of the South-to-North Water Diversion Project were calculated for the years 2002, 2012, and 2022 (

Table 3. Changes in ESV from 2002 to 2022 at city (prefecture) level.

City	ESV (CNY 100 Million)			ΔESV (%)
	2002	2012	2022	2012–2002	2022–2012	2022–2002
Garzê Prefecture	431.64	858.50	993.91	0.99	0.16	1.30
Aba Prefecture	248.60	283.64	399.26	0.14	0.41	0.61
Liangshan	144.23	98.38	192.51	−0.32	0.96	0.33
Leshan	84.90	23.44	24.77	−0.72	0.06	−0.71
Panzhihua	35.28	2.42	2.99	−0.93	0.24	−0.92
Yaan	25.73	79.59	145.03	2.09	0.82	4.64
Yushu Prefecture	4.17	118.22	50.28	27.35	−0.57	11.06
Guoluo Prefecture	20.56	236.05	225.02	10.48	−0.05	9.94
Lijiang	16.90	1.47	1.35	−0.91	−0.08	−0.92

). The ESVs of water source areas in 2002, 2012, and 2022 were CNY 101.202 billion, CNY 158.453 billion, and CNY 198.518 billion, respectively. In the past 20 years, the ESV has shown an upward trend year by year, increasing by 96%. The regions with the highest ESV were Garzê Prefecture and Aba Prefecture, which increased by 130.3% and 60.6%, respectively, over the past 20 years. The value of ecological services in Xinlong, Danba, Litang, and Daofu counties has increased significantly, by 4.8, 1.5, 12.5, and 8.9 times, respectively, over the past 20 years, suggesting that localized ecological restoration and land management policies have achieved positive results. In terms of land use types, farmland has decreased by 91%, while water bodies have decreased by 98%; forests and grasslands have increased by 1.6 times and 40 times, respectively. Forests and grasslands play a crucial role in climate regulation, hydrological regulation, and maintaining biodiversity. This structural shift has had a profound impact on regional ecosystem functioning. Forests and grasslands play key roles in climate regulation, hydrological cycle, and biodiversity maintenance, especially in high-altitude areas where ecosystems are more fragile but have higher potential for ecological services [27]. The results of this study are consistent with previous research on ecosystem evolution in mountainous areas of western China, which concluded that the area covered by natural ecosystems has continued to expand under the combined effects of policy intervention and ecological restoration [28]. Therefore, accurate quantification of regional ESV is important for improving public ecological awareness, promoting the construction of ecological compensation mechanisms, and supporting land use policy formulation [29].

3.3. Moran’s I Index of ESV

Moran’s I indexes of ESVs from 2002 to 2022 are shown in

Table 4. Moran’s I index of ESV from 2002 to 2022.

Year	Moran’s I Index	z	p
2002	−0.041396	−0.31705	0.751202
2012	−0.004779	−0.209876	0.833788
2022	0.046377	0.886732	0.375223

. The p-value was used to determine the significance of spatial autocorrelation. The z-value was used to evaluate the degree of deviation between Moran’s I index and the expected random distribution, helping to explain whether the data have a tendency toward clustering or dispersion in space. Moran’s I index itself was used to quantify the strength of spatial autocorrelation, helping to determine whether data were positively or negatively correlated. It can be seen from Moran’s I index values that the dataset as a whole showed a slight positive spatial autocorrelation, and this spatial autocorrelation gradually increased from −0.041396 to 0.046377 over the past 20 years.

4. Conclusions

In this study, based on the equivalent factor method, combined with remote sensing and geographic information system (GIS) techniques, we quantitatively evaluated the spatial and temporal evolution characteristics of ESV in water source areas in typical years (2002, 2012, and 2022), and revealed its spatial agglomeration and distribution pattern through the Moran index. The results show that the growth of ESV has significant spatial heterogeneity, and some counties, such as Xinlong, Litang, and Daofu, have significant growth in ecological value, indicating that the effectiveness of local ecological management is beginning to be realized. Collectively, this research elucidates alpine ecosystem service dynamics while informing regional ecological compensation, territorial spatial optimization, and ecological restoration policy design. Advancing ecosystem service applications for sustainable development requires synthesizing multi-source climate, socioeconomic, and field data to decode human–nature interdependencies.