Investigating the Spatio-Temporal Evolution of Land Cover and Ecosystem Service Value in the Kuye River Basin

Abstract

Land cover change influences the provision of regional ecosystem services, posing a threat to regional ecological security and sustainable development. The Kuye River Basin, a vital tributary of the Yellow River Basin, has experienced significant land cover changes due to intense human activity. Building on analysing the spatiotemporal evolution of land use cover and ecosystem service values from 1990 to 2022, this study predicted the land cover structure and ecosystem service value with two future scenarios, the NDC and the EPC, to provide insights into guiding sustainable policy interventions. We found the predominant land cover types were greensward and forest land, accounting for 67.22% of the total area. Forest land, greensward, and farmland have increased, while desert, water area, and other land types have decreased from 1990 to 2022. Forest land, greensward, farmland, and water areas are the main contributors to ecosystem service value in the Kuye River Basin. However, water area services have significantly decreased from 1990 to 2022. Under the NDC scenario, land development primarily relies on greensward and farmland, reducing forest and water areas and weakening the ecosystem’s regulatory and supporting functions. In contrast, the EPC scenario enhances ecosystem services by protecting critical ecological regions. Ecological protection measures significantly increase the ecosystem service values of the Kuye River Basin, and well-planned land use can effectively balance economic development with ecological preservation. This study provides scientific evidence to inform policies integrating ecological protection and economic growth, contributing to the sustainable development of the Kuye River Basin.

1. Introduction

Rapid urbanisation and industrialisation have catalysed the change in global land cover significantly through human activities [1]. These unplanned and damaging changes have led to worldwide ecosystem degradation, water scarcity, and food security crises [2]. The advantages that humans receive from ecosystems are known as ecosystems; they support and maintain the environmental conditions necessary for human life [3,4]. Human activities alter land cover and ecosystem services, affecting human welfare. Understanding the mechanism of land cover changes and their impacts on ecosystems is crucial for developing sustainable strategies to protect the environment and enhance human well-being.

The beginnings of the assessment of ecosystem services can be traced back to the 1980s. In 1997, Professor Daily provided the first systematic definition of ecosystem services in his book “Nature’s Services that Society Depends on: Natural Ecosystems” [5,6]. Their purpose was to improve human well-being and contribute to the goals of sustainable development [7]. In 1994, Professor Costanza calculated THE global biosphere’s ecosystem services were worth USD 33 trillion [8]. After the two milestone studies, ecosystem service value has become a popular issue of ecological economics worldwide [9]. Forest land and greensward are among Earth’s most productive natural ecosystems [10]. They provide many ecosystem services to humans, including material provision, climate regulation, flood mitigation, water purification, soil conservation, biodiversity protection, etc. In addition, the provision of ecosystem services is changing, driven by various factors, including demographic, economic, socio-political, technological, physical, or biological [11,12,13]. The primary driver is the configuration of land cover. Research has shown that changes in the structure of land cover directly impact the quality and quantity of ecosystem services [14,15,16]. The efficient use and allocation of land resources are essential to securing and enhancing ecosystem services [17].

Extensive research has been conducted on various valuation methods for ecosystem services [18,19,20]. However, a widely accepted valuation method for ecosystem services has yet to be available. Significant discrepancies in the methodologies used make it challenging to understand ecosystem service functions and their values accurately. Currently, two primary approaches for ecosystem service valuation are price per unit of service function and value equivalent factor per unit area [21,22,23]. The initial approach entails constructing a production equation that establishes a connection between a certain ecosystem service function and the ecological elements present in a given area. This approach quantifies how ecological characteristics in a region contribute to the delivery of ecosystem services—such as quantifying the climate regulation service by calculating the carbon sequestrated by the soil and biomass. However, despite its precision, this approach involves multiple input parameters and a complex computational process and presents challenges in standardising assessment methods and parameter requirements for each service value [24]. These difficulties make it challenging to ensure consistency and comparability across different ecosystems and services. The second approach calculates the ecosystem service value based on quantifiable and consistent criteria after distinguishing ecosystem service functions [25]. This simple, user-friendly method is ideal for valuing regional and global ecosystem services [8,25]. For example, Xie et al. developed a set of equivalent value factors of 11 ecosystem services from 14 ecosystem types in China (CN) based on numerous experts’ opinions [26]. This method was improved in 2015 by integrating experts’ opinions, available quantification methods, and remote sensing data [27]. However, due to the pedoclimatic and landscape variance in CN and the diversity of local human activities, applying the same equivalent coefficients of ecosystem service value (ESV) presents specific challenges [28]. Therefore, it is essential to localise the equivalent value factors by integrating local cultural and socio-economic conditions to ensure a more precise and context-specific valuation of ecosystem functions and services.

When dealing with watersheds, it is essential to take into account the distinct natural and social developmental traits of each administrative region within the watershed. This remains a significant challenge and a key focus of ongoing research. Most current studies have concentrated on examining the effects of past and current alterations in land cover on the values of ecosystem services [29]. Most studies applied CA-Markov, CLUE-S, and FLUS models to simulate future land cover [30,31,32]. Compared to other ones, the PLUS model is better at combining the drivers of various types of land cover change to simulate changes in the number of patches in each category [33]. The PLUS model includes the LEAS and CARS modules. The LEAS module identifies areas where land cover type shifts occurred during two phases of land cover change and analyses the impact of various drivers on land cover expansion. Using the Random Forest algorithm, the LEAS module calculates the development probability for each land cover type [34]. The CARS module combines the development probabilities obtained from LEAS to simulate future land cover using a meta-cellular automata model, which allows fine-grained simulation of regional land cover at the patch level and further prediction of ecosystem services [35]. Therefore, studies need to delve deeper into predicting future land cover changes and their impact on ecosystem service values based on historical and current land use trends.

The river basin is a complex system made up of natural and socioeconomic components, and its biological processes show some degree of independence and systematicity [36]. Its ecological status is brought to light, and possible issues with surface change are identified through an analysis of the land cover dynamics and ecosystem services. Nonetheless, prior research has mostly employed administrative districts as the assessment unit and has concentrated on the study of ecological services [37]. There has been limited systematic research on ecosystem dynamics and future trends across the multiple administrative divisions within the entire basin system. “Ecological Conservation and High-Quality Development of the Yellow River Basin” is a key development strategy in CN [38,39]. It emphasises that the Yellow River Basin (YRB) is of great economic importance and is an essential ecological barrier in the country. Improving ecological protection and management can facilitate the whole basin’s high-quality development and thus contribute more to the country’s revitalisation [40,41]. The Kuye River Basin (KRB) is a significant tributary of the YRB [42] and a vital area for CN’s ecological security barrier in the northern region. The extent of soil erosion in the area is over 95%, as indicated by research [43]. The average annual amount of sediment transported is 111 million tonnes, with a maximum yearly volume of 335 million tonnes. The highest documented sand concentration is 1700 kg/m³ [42]. The primary constituent of the sediment carried by the YRB is predominantly silt. The KRB is distinguished by widespread human activity, substantial economic growth, and distinct ecological circumstances. The Shenfu-Dongsheng (SD) mega-coal field, ranked as the sixth largest globally, is located in the central section of the basin [44]. The mining region comprises 28.51% of the entire basin area [45]. The “14th Five-Year Plan” plays a vital role in advancing ecological protection and attaining high-quality growth in the YRB [46]. Nevertheless, the impact of human activities on land cover change and their influence on ecosystem service supply in the KRB remains unknown.

The YRB is a critical ecological and economic region in CN, and its sustainable development is essential for the country’s broader environmental and financial goals. The overall objective of this study is to assess the land cover dynamics and ecosystem service values in the KRB, a vital tributary of the YR, to support high-quality ecological conservation and development strategies. Specifically, this study aims to (1) identify the land cover dynamics in the region from 1990 to 2022, (2) evaluate the impact of land use changes on ecosystem service values, and (3) predict the land use structure and ecosystem service values in 2030. The significance of this research lies in providing a scientific foundation for ecological governance and sustainable land management policies in the KRB, contributing to the broader efforts of ecological conservation in the YRB.

2. Materials and Methods

2.1. Overview of the Study Area

The Kuye River Basin (KRB) originates in Ordos (ERDOS) City, Inner Mongolia (IM), and flows into the Yellow River (YR) at Hejiachuan (HJC) Town in Yulin (YL) City, Shaanxi (SN) Province (Figure 1). The geographical location is 109°28′–110°45′ E and 38°22′–39°50′ N. The total length of the main stream is 242 km (including 83.9 km in Inner Mongolia Autonomous Region (IMAR) and 158.1 km in Shaanxi (SN) Province), with a total drainage area of 8706 km².The average temperature in the basin for many years is 8.9 °C, and the average annual precipitation is 386 mm. The precipitation varies greatly within the year, mainly concentrated in the summer. On the terrain, it shows a change of being higher in the northwest and lower in the southeast, with an altitude of 800–1300. The topography shows a changing pattern from upstream to downstream, from sandy grassland areas to gully and hilly areas, and then to wet beaches. At the same time, the vegetation coverage in the watershed is sparse, and it is severely eroded by water and wind erosion.

KRB is also home to the Shenfu Dongsheng (SD) mega coalfield, the 8th largest coal field in the world. This coalfield stretches across the central part of the basin and covers 28.51% of the basin’s total area. As a result, the basin has become a focal point for developing the “Jin-Shaan-Mengguang (JSM) Energy and Heavy Chemical Industry Base” in CN for the coming century. However, rapid urbanisation, growing industrial and mining activities, and inadequate governance have led to significant unsustainable development. This has further exacerbated soil erosion, land fragmentation, and degradation.

2.2. Data Sources

This study used remote sensing photos, biological and environmental data, and socioeconomic data to methodically evaluate land cover, ecosystem services, and future development possibilities (

Table 1. Data sources.

Category	Data Details	Year	Original Resolution	Data Resource	Material Source	Data Processing Platform	Equipment Source
Land cover structure	Other land, water area, forest land, desert, greensward, and farmland	1990, 1995, 2000, 2005, 2010, 2015, and 2022	30 m	“http://www.gscloud.cn (accessed on 20 May 2024)”.	Computer Network Information Center (CNIC), Beijing (BJ), China (CN)	ArcGIS (version: 10.8) and ENVI (version: 5.4)	ArcGIS (version: 10.8) source from Environmental Systems Research Institute (ESRI), Redlands (R), State of California (CA), The United State of America (USA) and ENVI (version: 5.4) source from Exelis Visual Information Solutions (EVIS), The United State of America (USA)
Natural data	DEM (Digital Elevation Model)	2000, 2010, and 2020	30 m	“http://www.gscloud.cn (accessed on 20 May 2024)”.	Computer Network Information Center (CNIC), Beijing (BJ), China (CN	ArcGIS (version: 10.8)	Environmental Systems Research Institute (ESRI), Redlands (R), State of California (CA), The United State of America (USA)
	Slope			“http://www.gscloud.cn (accessed on 20 May 2024)”.	Computer Network Information Center (CNIC), Beijing (BJ), China (CN
	Temperature			“https://www.resdc.cn/ (accessed on 20 May 2024)”.	Resources and Environmental Science Data Center (RESDC), Beijing (BJ), China (CN)
	Rainfall			“https://www.resdc.cn/ (accessed on 20 May 2024)”.	Resources and Environmental Science Data Center (RESDC), Beijing (BJ), China (CN)
Social, economic data	Distance from road	2000, 2010, and 2020	30 m	“https://www.resdc.cn/ (accessed on 20 May 2024)”.	Resources and Environmental Science Data Center (RESDC), Beijing (BJ), China (CN)	ArcGIS (version: 10.8)	Environmental Systems Research Institute (ESRI), Redlands (R), State of California (CA), The United State of America (USA)
	Night light
	GDP (Gross Domestic Product)
	Population density
	distance from city

).

2.3. Research Methodology

2.3.1. Research Framework

This article focuses on the KRB, a vital tributary of the YRB, as the research subject, covering the period from 1990 to 2022. The study includes a current situation analysis, dynamic evolution analysis, and future development trend prediction to provide a practical and well-founded basis for ecological environment planning and land cover structure management in the KRB. The research process consists of four main parts (Figure 2).

First, we collected Landsat TM 4–5 and Landsat 8 OLI_TRIS remote sensing (RS) images of the KRB. The images were processed using ENVI software, and six land use categories in the KRB were identified through supervised classification. We analysed the land cover structure, geographical and temporal evolution, and land cover transfer status of the KRB in 1990, 1995, 2000, 2005, 2010, 2015, and 2022 using ArcGIS and ENVI software. Secondly, based on the land cover data, we used ArcGIS software to visually analyse the spatial and temporal differentiation characteristics and evolutionary elements of land cover structure in the KRB from 1990 to 2022. Thirdly, we revised the equivalent factors of ecological service value per unit area in CN (

Table 2. Equivalent factors of ecological service value per unit area in CN.

ESF	TES	FL		F				GL			WL	D		W
		DL	PF	AL2	MN	BL	S	G	SG	M	W	D	BS	WS	GSC
PS	FP	0.12	0.19	0.03	0.04	0.04	0.03	0.01	0.05	0.03	0.07	0.00	0.00	0.11	0.00
	PM	0.06	0.01	0.07	0.10	0.09	0.06	0.02	0.08	0.05	0.07	0.00	0.00	0.03	0.00
	SWR	0.00	0.36	0.04	0.05	0.05	0.03	0.01	0.04	0.02	0.36	0.00	0.00	1.15	0.30
RS	GC	0.09	0.15	0.24	0.33	0.30	0.20	0.07	0.27	0.16	0.26	0.02	0.00	0.11	0.02
	CC	0.05	0.08	0.70	0.97	0.90	0.59	0.19	0.72	0.42	0.50	0.01	0.00	0.32	0.07
	CUO	0.01	0.02	0.21	0.28	0.27	0.18	0.06	0.24	0.14	0.50	0.04	0.01	0.77	0.02
	HR	0.04	0.38	0.46	0.49	0.66	0.46	0.14	0.53	0.31	3.36	0.03	0.00	14.16	0.99
SS	SC	0.14	0.00	0.29	0.40	0.37	0.24	0.09	0.33	0.19	0.32	0.02	0.00	0.13	0.00
	NCM	0.02	0.03	0.02	0.03	0.03	0.02	0.01	0.02	0.02	0.02	0.00	0.00	0.01	0.00
	B	0.02	0.03	0.26	0.36	0.33	0.22	0.08	0.30	0.18	1.09	0.02	0.00	0.35	0.00
CS	AL1	0.01	0.01	0.11	0.16	0.15	0.10	0.03	0.13	0.08	0.66	0.01	0.00	0.26	0.01

) by incorporating the cultural and socio-economic conditions of the watershed to derive the coefficient of ecosystem service value per unit area in the KRB (

Table 3. Coefficient of ecosystem service value per unit area in KRB (dollar·km⁻²·a⁻¹).

ESF	TES	Land Cover Type
		OL	WA	FL	D	G	F
PS	FP	0.0036	0.2881	0.0900	0.0036	0.0828	0.3061
	PM	0.0108	0.0828	20.8842	0.0108	0.1224	0.1440
	SWR	0.0072	2.9850	0.1080	0.0072	0.0684	0.0072
RS	GC	0.0396	0.2772	0.6877	0.0396	0.4357	0.2413
	CC	0.0360	0.8246	2.0560	0.0360	1.1486	0.1296
	CUO	0.1116	1.9984	0.6013	0.1116	0.3781	0.0360
	HR	0.0756	36.8138	1.3431	0.0756	0.8426	0.0972
SS	SC	0.0468	0.3349	0.8354	0.0468	0.5293	0.3709
	NCM	0.0036	0.0252	0.0648	0.0036	0.0396	0.0432
	B	0.0432	0.9182	0.3349	0.0180	0.2124	0.0216
CS	AL1	0.0180	0.6805	0.33349	0.0180	0.2124	0.0216

). We calculated the ecosystem service value of the KRB from 1990 to 2022 (Formula (7)). Finally, the land cover structure and ecosystem service value of the KRB in 2030 were predicted based on the PLUS model.

Initially, this study gathered remote sensing images of the KRB, specifically Landsat images (including TM 4–5, 8 OLI_TRIS). The ENVI programme was utilised to process the images, and a supervised classification technique was employed to identify six distinct land cover types. Then, using ArcGIS 10.8 and ENVI 5.4 software, we conducted an analysis of the land cover structure, spatial and temporal changes, and land cover transfer status in the years 1990, 1995, 2000, 2005, 2010, 2015, and 2022. Furthermore, utilising ArcGIS 10.8 software, we conducted a visual analysis of the regional and temporal differentiation characteristics and evolutionary aspects of the land cover structure in the KRB between 1990 and 2022, using the available land cover data. In addition, modified the corresponding components of ecological service value per unit area in CN (Table 2) by considering the cultural and socio-economic circumstances of the study area. This allowed us to calculate the coefficient of ecosystem service value per unit area in the KRB (Table 3). We computed the monetary worth of the ecological services from 1990 to 2022 using Formula (7). The PLUS model was used to anticipate the land cover structure and ecosystem service value of the KRB in 2030.

2.3.2. Land Cover Analysis

(1)

Land cover dynamic degree

Land cover dynamics, which primarily reflect regional variations in the extent and severity of changes with land cover. Characterises the quantitative shift of land cover types over a given period of time. Its two types are individual and comprehensive land cover dynamics [47]. The particular calculating approach makes use of a formula where the dynamic degree of each type of land cover represents the rate at which that specific land cover type changes within the region over a certain period of time [48]. The dynamic degree of land cover is used to analyse the overall status of land cover transfer between land cover categories during the study period [49,50]. It is capable of displaying the degree of regional land cover changes and conducting a thorough comparison of land cover changes between local and total areas. Below is a description of the precise calculation process.

$${LUT}^{now-before}=\frac{{\mathit{LUT}}^{\mathit{now}}-{\mathit{LUT}}^{\mathit{before}}}{{\mathit{LUT}}^{\mathit{before}}\times \mathit{Time}}\times 100\%$$

(1)

$$LUT \mathit{synthesize}=\left[{\displaystyle \frac{{\sum }_{i=1}^{ n}∆{\mathit{LUT}}^{x-y}}{2\sum _{i=1}^{ n}{\mathit{LUT}}^{ \mathrm{x}}}}\right]\times {\displaystyle \frac{1}{Time}}\times 100\%$$

(2)

The land cover type’s area at the beginning and conclusion of the study period is reflected in the LUT^now−before. LUT indicates its level of dynamic variation. The following variables are represented: Time (study duration), LUT^synthesize (land cover dynamic degree), LUT^x (number of beginning land types), ∆LUT^x–y (area of x-type land transformed to other land types) are all represented.

(2)

Land cover transfer matrix

The land cover transfer matrix reflects the dynamic changes between land cover types over a specified period [51]. It includes static data and detailed information on area transfers between categories from the beginning to the end of this study (Formula (3)).

$${\mathit{SQ}}_{\mathit{ij}}=\left[\begin{array}{ccccc}{\mathit{SQ}}_{11}& {\mathit{SQ}}_{12}& {\mathit{SQ}}_{13}& \dots & {\mathit{SQ}}_{1n}\\ {\mathit{SQ}}_{21}& {\mathit{SQ}}_{22}& {\mathit{SQ}}_{23}& \dots & {\mathit{SQ}}_{2n}\\ {\mathit{SQ}}_{31}& {\mathit{SQ}}_{32}& {\mathit{SQ}}_{33}& \dots & {\mathit{SQ}}_{3n}\\ \dots & \dots & \dots & \dots & \dots \\ {\mathit{SQ}}_{n1}& {\mathit{SQ}}_{n2}& {\mathit{SQ}}_{n3}& \dots & {\mathit{SQ}}_{\mathit{nn}}\end{array}\right]$$

(3)

where i and j (i, j = 1, 2,…, n) are the land cover type area; S is the area; n is the number of land cover types both before and after the transfer; and SQ_ij is the area of the i land class both before and after the transfer.

(3)

Land use intensity

Land use intensity reflects the depth, extent, and environmental and human impacts on the land [52]. This study follows the comprehensive method of Liu Jiyuan et al. [53] to analyse the land use intensity from an ecological point of view. The intensity of land cover in the KRB is divided into three levels: other land and desert as the first level, forest land, greensward, and water area as the second, and farmland as the third level. The land use intensity increases from level 1 to level 3. Formulas (4) and (5) are used to determine the overall land use intensity index and its change in the basin.

$$L=\sum _{i=1}^n(A_i\times C_i)$$

(4)

$$C_i={\mathit{CC}}_i/\mathit{HJ}$$

(5)

where n is the land cover degree; CC_i is the total area of the level i land cover type; HJ is the total area; and L is the comprehensive land cover intensity of the study area. A_i is the land cover intensity classification index of the level i land cover type. C_i is the proportion of the level i land area to the total area.

2.3.3. Estimation of Ecosystem Service Value

(1)

Equivalent factors of ecological service value per unit area in China

Ecosystem services provide individuals with direct or indirect advantages. Benefits include providing energy and materials to the economy and society, accepting and transforming wastes created by these systems, and providing services directly to society [54]. The Millennium Ecosystem Assessment (MEA) classifies ecosystem services as supplying, regulating, sustaining, and cultural. Carpenter, Stephen R. et al. [55] establish the framework and research emphasis for ecosystem service valuation. Local researchers tried it in several regions in CN for related research and found that the method had some flaws, and the ecosystem service value calculation was inaccurate.

After two revisions, Xie et al. [28] created the equivalent factor table of ecological service value per unit area in CN (Table 2) using a questionnaire poll of over 200 professionals and academics. The equivalent value of ecosystem services per unit area is the entire economic value of ecosystem services in a specific area, calculated by assessing their economic and social contributions. A more advanced method uses unit area value equivalent variables to assess 11 ecosystem services and 14 ecosystem categories nationwide (Table 2). In the table, ESF stands for ecosystem service function. PS stands for provision services. RS stands for regulating services. SS stands for supporting services. CS stands for cultural services. TES stands for types of ecosystem services. FP stands for food production. PM stands for production of material. SWR stands for supply of water resources. GC stands for gas conditioning. CC stands for climate control. CUO stands for clean-up operation. HR stands for hydrological regulation. SC stands for soil conservation. NCM stands for nutrients cycle maintenance. B stands for biodiversity. AL1 stands for aesthetic landscape. FL stands for farmland. F stands for forest. GL stands for grassland. WL stands for wetland. D stands for desert. W stands for water area. DL stands for dry land. PF stands for paddy field. AL2 stands for acerose leaf. MN stands for mixed needle. BL stands for broad leaf. S stands for shrub. G stands for grassland. SG stands for shrub grass. M stands for meadow. W stands for wetland. BS stands for bare soil. WS stands for water system. GSC stands for glacier and snow cover.

(2)

According to the MEA, ecosystem services were classified into provisioning, regulating, supporting, and cultural ecosystem services [56]. This research, following Sun et al. [57] (Formula (6)), modified CN’s ecosystem service value per unit area [28] to adapt it to the study area’s ecosystem (Table 3). This modification was made based on the basin’s natural geography and human and social conditions.

$$X=E_a\times \left({\displaystyle \frac{y^{\prime }}{y}}\right)$$

(6)

where X represents the economic value of a single ecosystem service in the KRB, measured in dollars per square kilometre. E_a represents the monetary value of a single ecosystem service equivalent factor in CN, also measured in dollars per square kilometre. y′ represents the average grain yield per unit area of farmland in the KRB from 1990 to 2022, measured in kilogrammes per square kilometre. y represents the average grain yield per unit area of farmland in CN from 1990 to 2022, also measured in kilogrammes per square kilometre.

$$ESV={\displaystyle \sum _{i=1}^m{\displaystyle \sum _{j=1}^n(A_i}}V{C}_{kij})$$

(7)

This equation represents the calculation of the overall value of ecosystem services (ESV). It takes into account the area of each land cover type (A_i) and the corrected coefficient of the value of each ecosystem service (VCk_ij) for each land cover type. The values for VCk_ij may be found in Table 3.

Formula (7) is used to calculate the ESV for 11 ecosystem services (food production, production of material, supply of water resources, gas conditioning, climate control, clean-up operation, hydrological regulation, soil conservation, nutrient cycle maintenance, biodiversity, and aesthetic landscape) of 6 land cover types (greensward, farmland, desert, forest land, water area, and other land) in the KRB. This calculation was performed for 7 time points, namely 1990, 1995, 2000, 2005, 2010, 2015, and 2022. In the table, ESF: ecosystem service function; PS: provision services; RS: regulating services; SS: supporting services; CS: cultural services; TES: types of ecosystem services; FP: food production; PM: production of material; SWR: supply of water resources; GC: gas conditioning; CC: climate control; CUO: clean-up operation; HR: hydrological regulation; SC: soil conservation; NCM: nutrient cycle maintenance; B: biodiversity; AL1: aesthetic landscape; WA: water area; FL: forest land; D: desert; G: greensward; F: farmland; OL: other land (revised from Xie et al. [28]). For the six land cover types and eleven ecosystem services in KRB, Formula (7) is used to calculate the ESV. Seven time points—1990, 1995, 2000, 2005, 2010, 2015, and 2022—were used in this computation. In Table 3, including ecosystem service function (ESF), provision services (PS), regulating services (RS), supporting services (SS), and cultural services (CS) are listed in the table. FP: food production; PM: material production; TES: types of ecosystem services; water resource supply (SWR); gas conditioning (GC); climate control (CC); clean-up operation (CUO); hydrological regulation (HR); soil conservation (SC); maintenance of the nutrient cycle (NCM); biodiversity (B); attractive landscape (AL1); water area (WA); desert (D); greensward (G); farmland (F); and other land (OL) (revised from [28] Xie et al.).

2.3.4. Prediction of Land Cover and Ecosystem Services

This study utilised the PLUS model to anticipate land cover and ecosystem service developments in the KRB by 2030 under scenarios of natural development and ecological protection based on the land cover structure and ecosystem services (from 1990 to 2022). The predictions contribute to improving ecological quality, enhancing ecosystem services, and preventing conflicts related to land use.

(1)

Simulation parameters and neighbourhood weight settings for the CARS module

The neighbourhood weight parameter represents each land category’s expansion intensity, influenced by both socioeconomic and natural forces, with values ranging from 0 to 1. A higher value (closer to 1) indicates greater expansion capacity and a reduced likelihood of being converted into other land categories [33]. This parameter helps model the growth potential of land types, but due to the complexity of the interactions between influencing factors, direct calculation of expansion intensity is not feasible. Instead, historical land use changes are analysed to calculate the expansion capacity of each land category, providing a practical approach to understanding land dynamics over time.

$$X^{\ast }={\displaystyle \frac{X-X_{\min }}{X_{\max }-X_{\min }}}$$

(8)

where X* represents the standardised deviation value. X refers to the change in area of each land category between different years of land cover data. X_max is the maximum change in location among all land categories, while X_min is the minimum change.

This paper calculates the expansion intensity of each land use type in the KRB by combining land cover data with natural factors such as rainfall, DEM, slope, temperature, and factors in socio-economics like proximity to nightlight, highways, provincial roads, GDP, railways, population density, and cities (Figure 3). The resulting neighbourhood weights, which represent the expansion potential of each land type, are shown in Figure 4.

(2)

The setting of land conversion cost matrix parameters

In this paper, the conversion rules between land cover types are represented by a cost conversion matrix, which indicates whether land cover types in the watershed can be converted into each other [58]. A matrix value of 1 means a land class can be converted to another, while a value of 0 indicates it cannot [58].

(3)

Future development scenarios

The 2030 land cover was simulated using 2000, 2010, and 2020 data under two scenarios. This considered the KRB’s land cover structure, ecosystem service evolution, and territorial development. Based on Table 3 and

Table 4. Land conversion cost matrix settings for natural development scenarios and ecological protection scenarios.

Land Cover Type	Natural Development Scenarios						Ecological Protection Scenarios
	Other Land	Water Area	Forest Land	Desert	Greensward	Farmland	Other Land	Water Area	Forest Land	Desert	Greensward	Farmland
Farmland	1	1	1	1	1	1	0	1	1	1	1	1
Forest land	1	0	1	1	1	1	1	0	1	1	1	1
Greensward	1	1	1	1	1	1	1	1	1	1	1	1
Water area	0	1	0	1	1	0	0	1	0	1	0	0
Desert	1	0	1	1	1	1	0	0	0	1	0	0
Other land	1	1	1	1	1	1	1	0	1	1	0	1

, the future ecosystem service trends for 2030 were calculated.

The natural development scenario (NDC) serves as the foundation for the other scenarios. This is a theoretical development scenario that predicts the spatial pattern of land cover will only develop linearly in accordance with its historical evolutionary trajectory, or linearly in accordance with its evolutionary past. It is predicated on how the KRB’s land cover changed between 2000 and 2020, as well as natural and socioeconomic factors. The limitations imposed by different laws and rules, plans for land cover and urban expansion, etc., are disregarded in this scenario. As a parameter for land demand in the PLUS model, the Markov chain projects the demand for each land type in 2030.

Ecological Protection Scenario (EPC): This scenario focuses on spatial control measures for land cover to support watershed ecosystem protection and promote ecological civilisation amidst rapid socio-economic development. Because the study area is rich in mineral resources and a key energy hub in CN but faces severe soil erosion with over 100 million tonnes of annual sand transport. Population density, GDP, urbanisation, and infrastructure (Figure 3) are all increasing. To prevent ecosystem degradation during this growth, future land cover changes must prioritise enhanced protection of ecological areas (see Table 4 for details).

(4)

Simulation accuracy test

To ensure the scientific validity and effectiveness of the prediction results, it is necessary to verify the accuracy of the land cover prediction results for the KRB in 2030.

The accuracy of land cover modelling is evaluated using the Kappa and FoM coefficients. Excellent model performance is indicated by a Kappa coefficient greater than 0.8 [59]. Higher values of the FoM coefficient, which gauges the accuracy of meta-cellular simulations, correspond to higher accuracy. It usually ranges from 0.01 to 0.25 [60]. The FoM coefficient can be computed using the following formula:

$$FoM={\displaystyle \frac{B}{A+B+C+D}}$$

(9)

where A denotes the error resulting from an actual change in the regional area’s land use status but being predicted to remain unchanged; B is the accurately predicted area; C is the error resulting from prediction errors; and D is the error resulting from an actual change in the region’s land cover but variations in the predicted outcomes.

3. Results

3.1. Land Cover Structure

The land cover structure of the KRB over 32 years shows forest land and greensward comprised the most significant proportion of the KRB land, averaging 5260.126 km², or 60.42%, covering over half the total area (Figure 5). Spatially, this ecosystem spans the northern, central, and southern parts. Farmland followed, with areas ranging from 1257.8817 km² to 1492.0524 km², representing 14.45% to 17.14%. Farmland is evenly distributed across the basin, serving as the primary food source for local communities. Desert areas decreased significantly, from 1307.006 km² (15.01%) in 1990 to 382.1620 km² (4.39%) in 2022, with a sharp decline starting in 1995.

Figure 6 shows that in 1990, large desert areas were spread across the north, centre, and south of the basin, but by 2022, deserts are primarily concentrated in the northwest and scattered in central and southeastern regions. The continuous decrease in desert areas reflects growing awareness of ecological and environmental protection alongside socio-economic development. This issue has been increasingly addressed, leading to a gradual reduction in desertification.

In 2022, the distribution of other land types, like construction land, saline land, wasteland, and unused land, covered the following areas: 1206.58 km² (13.86%), 896.94 km² (10.30%), 902.80 km² (10.37%), 571.61 km² (6.57%), 416.76 km² (4.79%), 541.68 km² (6.22%), and 747.33 km² (8.58%). These land types exhibited a mix of both increasing and decreasing trends over time. The northern portion of the basin contains the majority of these areas, a densely populated region that includes the Shenfu-Dongsheng (SD) mega coal field [44]. Additionally, water resources, an essential part of the ecological environment and production, have declined. By 2022, the water area had shrunk to 232.538 km², just 10.84% of the basin’s land, and affecting the branches.

3.2. Characteristics of Land Cover Change and Transfer

Land cover change generally follows a trend of three increases and three decreases: forest land, greensward, and farmland have increased, while desert, water area, and other land types have decreased (Figure 7). From 1990 to 1995, desert areas were widespread at the start of the study period, the land cover structure was inefficient, and ecosystem quality was low. However, by the early 21st century, the watershed’s land cover had increased significantly, with increases in forest land, greensward, and farmland and a sharp decline in desert areas. Forest land, greensward, and farmland grew at 1.61%, 0.92%, and 0.60%, respectively, adding 1.6576 km², 11.9159 km², and 2.3417 km² by 2022. In contrast, desert areas decreased by 9.2484 km², reducing the desert percentage to 5.82% by 2022.

Land cover transfers led to a significant increase in forest and other land and a substantial decrease in the area of desert (Figure 8). While some greensward and farmland were converted to forest, 49.19 km² and 12.96 km² remained unchanged, with an additional 3.66 km² of greensward and 1.69 km² of farmland added. Together, these areas accounted for 77.53% of the watershed. Overall, desert areas showed a trend of “weak transfer in and strong transfer out,” with desert land being converted to greensward, residential, transport, and other land types.

3.3. Land Use Intensity

The composite index of land use extent measures the impact of socio-economic development, including human activities and urbanisation, on land use changes over time. The composite index values for 1990, 1995, 2000, 2005, 2010, 2015, and 2022, which are 1.8556, 1.9025, 1.9080, 1.9847, 2.0542, 2.0535, and 2.0417, respectively, illustrate an overall increase in land use intensity (Figure 9). From 1990 to 2022, the land use intensity for other land and desert areas remained low, at 0.09 and 0.11, respectively. In contrast, greensward, forest land, and water areas exhibited higher land cover, with greensward increasing by 0.27 and forest land by 0.03. However, the decline in water area use intensity by 0.05 is concerning. Farmland showed a moderate land use intensity of 0.50.

3.4. Characteristics of Changes in Ecosystem Services

3.4.1. Changes in Ecosystem Values across Varied Service Functions

The overall ecosystem service value fluctuated between 1990 and 2022, initially decreasing and then increasing, as seen in Figure 10. The entire worth was USD 497.2469 million in 1990. Then, the values decreased to USD 470.210270 million in 2000. From the third phase (2000–2005), ecosystem service values rise again, reversing the previous trend. This upward trajectory continues until the seventh phase (2015–2022), when they reach a peak of USD 501.7703 million.

Over 32 years, the value of provisioning services (food, materials, and water), regulatory services (gas conditioning, climate control, purification, and water regulation), supporting services (soil conservation, nutrient cycling, biodiversity), and cultural services steadily increased, reaching USD 501.7855 million by 2022. The provisioning service grew from around USD 110.80 million in the first decade (1990–2000) to over USD 117.725 million in the second (2000–2010), continuously increasing since 2010. The regulating service followed a “two increments and two decrements” pattern, with gas conditioning and climate control rising by USD 31.44 million and USD 76.04 million, while clean-up operation and hydrological regulation declined. This trend is linked to changes in land cover and the influence of natural and socio-economic factors in the watershed.

3.4.2. Ecosystem Values Changes across the Varied Land Cover Structure

Based on the findings of the ecological services response to land cover (Figure 11), the ecosystem service values generated by greensward, water area, and forest land are relatively high at USD 1475.025, USD 970.608, and USD 795.2676 million. The value of ecosystem service generated by farmland, desert, and other land is USD 146.1175, USD 26.038, and USD 21.052 million, respectively. Greensward’s high service value is due to its extensive coverage (55.72% of the watershed). At the same time, though smaller in size, forest land and water areas have high per-unit ecosystem service values of USD 27.7688/km²/year and USD 45.2287/km²/year.

Over time, the ecosystem service value of greensward, forest land, and farmland steadily increased from USD 180,991.1075 million in 1990 to USD 232,701.052 million in 2022. In contrast, water area, desert, and other land decreased from USD 208.9965 million to USD 109.5535 million. This is because the water area has decreased by 207.4244 km², i.e., from 5.05% to 2.67%, within 32 years. Compared to the water area, the ecosystem service values of greensward and forest land areas increased by USD 51.799 and USD 45.982 million, respectively. The ecosystem service value of greensward fluctuates, rising from 1990 to 2010 and decreasing. From 1990 to 1995, the increase in ecosystem service value was USD 4.2935 million, followed by a rise of USD 1.662 million from 1995 to 2000 and USD 28.9465 million from 2000 to 2005. However, from 2010 to 2015, the ecosystem service value decreased by USD 40.165 million, with a continued decline until 2022. Despite this, the overall trend in greensward ecosystem service value was upward.

The shifts in the values of ecosystem services like farmland, desert, and other land from 1990 to 2022 were minimal. Farmland saw a USD 3.324 million increase, while desert and other land decreased by USD 3.601 million and USD 1.939 million, respectively. Desert ecosystem service value has steadily declined since 1995. Farmland, water regions, green spaces, and forests are the primary sources of ecosystem service value. However, water area services have significantly decreased from 1990 to 2022, impacting provisioning, regulating, supporting, and cultural services. Although the general trajectory of ecosystem service functionality and values has been upward, the marked decrease in water areas and the substantial expansion of the built-up regions will likely hurt future ecosystem services and values.

3.5. Future Development for Land Cover and Ecosystem Services under Two Scenarios

The land cover structure in the EPC differs significantly from that in the NDC (Figure 12,

Table 5. Prediction of ecosystem services in the KRB in 2030 (unit: USD million).

Ecosystem Service		Natural Development Scenario
		Other Land	Water Area	Forestland	Desert	Greensward	Farmland
Provisioning service	FP	0.0004	0.0061	0.0042	0.0001	0.0416	0.0457
	PM	0.0013	0.0017	0.9681	0.003	0.0616	0.0215
	SWR	0.0008	0.0628	0.0050	0.0002	0.0344	0.0011
Regulating service	GC	0.0046	0.0058	0.0319	0.0012	0.2191	0.0360
	CC	0.0042	0.0173	0.0953	0.001	0.5775	0.0193
	CUO	0.0131	0.0420	0.0279	0.0034	0.1901	0.0054
	HR	0.0089	0.7743	0.0623	0.0023	0.4236	0.0145
Supporting service	SC	0.0055	0.0070	0.0387	0.0014	0.2661	0.0553
	NCM	0.0004	0.0005	0.0030	0.0001	0.0199	0.0064
	B	0.0051	0.0193	0.0354	0.0013	0.2426	0.0070
Cultural service	AL1	0.0021	0.0143	0.0155	0.0006	0.1068	0.0032
Ecosystem Service		Ecological Protection Scenario
		Other Land	Water Area	Forestland	Desert	Greensward	Farmland
Provisioning service	FP	0.0022	0.1822	0.0973	0.0006	0.1957	0.4407
	PM	0.0066	0.0524	22.5833	0.0017	0.2893	0.2074
	SWR	0.0044	1.8884	0.1168	0.0011	0.1617	0.0104
Regulating service	GC	0.0243	0.1754	0.7437	0.0062	1.0295	0.3474
	CC	0.0221	0.5216	2.2233	0.0056	2.7141	0.1867
	CUO	0.0685	1.2642	0.6502	0.0173	0.8933	0.0519
	HR	0.0464	23.2894	1.4523	0.0118	1.9909	0.14
Supporting service	SC	0.0287	0.2118	0.9033	0.0073	1.2507	0.5341
	NCM	0.0022	0.0159	0.0701	0.0006	0.0936	0.0622
	B	0.0265	0.5809	0.3621	0.0028	0.502	0.0311
Cultural service	AL1	0.011	0.4305	0.3621	0.0028	0.502	0.0311

). Under the NDC, greensward is projected to occupy the largest area at 5046.19 km², with forest land and water areas showing significantly smaller areas. In contrast, under the EPC, farmland, forest land, and water areas see substantial increases, with greensward and other land showing a decrease in area. This reflects a shift towards prioritising ecological resources like farmland, forest land, and water bodies in the protection scenario, while the NDC allows for a larger expansion of greensward but at the cost of forest and water areas.

Comparing the two scenarios, both the NDC and the EPC demonstrate the importance of key land types such as farmland, greensward, and water areas in contributing to ecosystem services. However, the EPC scenario greatly enhances the overall value of these services, particularly in provisioning and regulating services. For example, in the NDC scenario, farmland contributes USD 0.9681 million to material production (PM), while in the EPC scenario, this rises dramatically to USD 22.5833 million. Similarly, hydrological regulation (HR) from water areas in the NDC scenario is USD 0.7743 million, but under the EPC scenario, it soars to USD 23.2894 million, reflecting a much stronger focus on conservation.

While greensward plays a larger role in the NDC scenario—contributing USD 0.5775 million to climate control (CC)—the overall service values are lower due to a reduced emphasis on ecological protection. In contrast, the EPC scenario highlights much higher contributions from water areas, forest land, and farmland, with water resources supply (SWR) increasing from USD 0.0628 million in the NDC scenario to USD 1.8884 million in the EPC scenario. These numbers show that while both scenarios utilise important land types, the EPC scenario results in much stronger ecosystem functionality and service provision, striking a more sustainable balance between human development and ecological conservation.

4. Discussion

The KRB, a key part of CN’s northern ecological security barrier, is crucial to the YRB. Assessing land cover and ecosystem services helps protect the environment, mitigate hazards, and guide governance. While ecosystem service values have steadily increased from 1990 to 2022, population growth and coal industry expansion have driven urbanisation and reduced water availability, causing imbalances in land cover. Environmental planning has lagged behind these changes.

4.1. Analysis of Land Cover Structure and Ecosystem Services

The intricate relationship between environmental factors and human activity significantly influences the functioning of ecosystem services, largely driven by land use patterns [61]. While stressing the inherent value of nature, evaluating the role and worth of ecosystem services in the context of changing land cover helps to clarify the connection between ecological civilisation and sustainable development in the watershed. Predicting the geographical and temporal changes in ecosystem service values helps characterise these shifts. It critically supports integrating socio-economic growth with ecological conservation, facilitating science-based management decisions [62]. Most current studies on ecosystem services’ spatial and temporal evolution focus on analysing past changes using available land cover data. However, fewer studies emphasise predicting future trends, particularly in watersheds spanning multiple administrative regions [63,64]. This gap exists because such studies must account for each specific area’s unique ecological and socio-economic characteristics. As a result, this study’s ability to predict the spatial and temporal development of ecosystem services in the KRB—a significant YR tributary and CN’s long-standing primary energy hub—is both realistic and promising [65].

The PLUS model is a raster-based tool for simulating changes in land cover at the patch scale, in contrast to previous models. It combines multi-type random seed mechanisms in CA models with rule-mining techniques based on land expansion analysis to anticipate the evolution of land use landscapes at the patch-scale and explore determinants of land expansion. This makes it particularly useful for complex geographical environments like the KRB, where it can offer a rationale based on science for decisions made on regional sustainable development by simulating land use changes under various policy scenarios. Accordingly, this study predicts the land cover structure of the KRB in 2030 under NDC and EPC (Figure 12). Based on these predictions, ecosystem service values for both development scenarios in 2030 were calculated (Table 5).

The results indicate that water areas, greensward, farmland, and forest land are critical ecosystems within the KRB. A reduction in these areas will result in a decline in the value of ecosystem services. Under the NDC for 2030, if no proactive measures are taken to capitalise on the basin’s abundant resources [42], human activities, including coal mining, will continue expanding industrial and mining land, with more people involved in mining operations. This will lead to a sharp increase in construction, industrial, and mining land in the basin. However, this expansion will cost crucial ecosystems, such as forest land, greensward, and farmland. Therefore, to promote balanced environmental and socio-economic growth in the KRB, prioritising the conservation of greensward, farmland, and water resources is essential. Enhancing the protection of forests, greenswards, and water bodies should also be a key focus. Currently, urbanisation is predominantly concentrated in the northern part. To ensure effective watershed management in the future, planning must consider the growth of natural features including mountains, water bodies, forest land, lakes, grasses, and sands, as well as socio-economic factors. Initial efforts will be concentrated in the densely populated northern region, with a gradual expansion of management to the southern part of the basin. Moreover, the impact of microbiological and geological hazards on land cover and ecosystems has not been factored into these projections due to challenges in obtaining data and visualising these risks. Consequently, future studies will aim to integrate ecological and socio-economic factors more comprehensively into these projections.

4.2. Optimisation Suggestions for Future Regulatory Measures and Policy Formulation

Based on the background of the NDC and EPC, as well as the land cover and ecosystem services values, it is clear that policy intervention is necessary for sustainable development in the study area.

In the NDC scenario, which assumes unregulated development, ecosystem services decrease as construction and industrial land expand at the expense of key ecological areas. For instance, water areas contribute only USD 0.7743 million to hydrological regulation, and material production from farmland is limited to USD 0.9681 million. This reflects the environmental degradation that occurs when development proceeds without constraints. On the other hand, the EPC scenario, which emphasises ecological protection, significantly enhances ecosystem service values. Hydrological regulation from water areas is increasing, while farm material production is rising. This scenario shows that prioritising ecological preservation leads to stronger ecosystem functionality and long-term sustainability. The stark difference between the two scenarios underscores the necessity of intervening to prevent ecosystem degradation and maintain the balance between development and ecological preservation.

Thus, in order to analyse land cover and ecosystem services in the KRB between 1990 and 2022, this research combined historical data with modelling of potential future development scenarios. The following suggestions are meant to direct socioeconomic growth in the future, settle disputes between people and land, maximise the distribution of land cover, and provide guidance for ecological and environmental policies and initiatives in this basin.

5. Conclusions

Over the past 32 years, significant land cover changes in the KRB have altered its physical characteristics, tightly linking land cover with ecosystem services. Rational land use is crucial for enhancing ecosystem services and maintaining ecological stability. Projections indicate increasing ecological risks from unchecked urbanisation and industrialisation, threatening key resources like forests, greenswards, and water areas, reducing the basin’s ecosystem service value. Nearly half of the basin requires urgent risk management interventions, particularly in the north. Sustainable development hinges on protecting critical areas and regulating growth. Adaptive management through innovative policies, planning, and technology is essential for ecosystem restoration, with the findings providing crucial evidence for decision-making and strategies applicable to other regions with similar socio-economic and environmental challenges.

In addition, due to the difficulty in obtaining data, it is necessary to conduct field investigations on each small watershed after model partitioning to verify the results. However, this work requires a significant amount of time and personnel involvement, so this paper did not divide the KRB into smaller sub-basins, nor did it investigate the land cover structure, ecosystem service characteristics, and future development scenarios between sub-basins. Hence, in the upcoming research, we will carry out this work to make the study more refined and valuable.