Delineating Ecological Restoration Zoning Integrating Functional and Structural Models in Horqin Sandy Land, China

Abstract

Escalating human–land conflicts have exacerbated ecosystem degradation, threatening regional sustainable development. As the largest sandy land in China, the Horqin Sandy Land (HSL) in eastern Inner Mongolia exhibits high ecological fragility. Delineating ecological restoration zones (ERZ) is critical to transition from localized restoration to system-wide stability, thereby enhancing regional ecological security, which reflects ecosystem health and integrity. Ecological security patterns (ESP), as spatial configurations that support and maintain ecological security, serve as the foundational framework for ERZ planning. Unlike conventional applications of InVEST and MSPA, this study integrates an ecosystem service assessment with morphological spatial pattern analysis (MSPA) under a “Source–Resistance–Corridor–Note” paradigm to develop a novel “ecological network–zoning” approach. This framework transforms ecological connectivity analysis into actionable restoration zoning, bridging theoretical ESP construction with practical management needs. Key findings include the following: (1) In total, 76 vital ecological source regions were mapped, representing about 10,204.38 km² of ecologically significant land, with primary distribution in the northwestern mountainous regions; (2) A total of 169 ecological corridors were extracted, spanning 4071.94 km in length. Ecological pinch points with 239.91 km² and barrier points with 568.85 km² were systematically identified; (3) A “Five Zones, Three Belts, One Core” spatial strategy was proposed, aligning with regional ecological conditions and development goals. This study provides a transferable methodology for ecosystem restoration in global arid and semi-arid regions, bridging theoretical frameworks with actionable zoning practices.

1. Introduction

Under the dual pressures of intensifying climate change and escalating human activities, global ecosystem structures and functions are experiencing widespread degradation, leading to increasingly severe ecological security challenges worldwide. The United Nations Decade on Ecosystem Restoration (2021–2030) reveals that over 2 billion hectares of ecosystems globally face degradation risks, with declining ecosystem service capacities becoming a fundamental constraint on sustainable development [1,2], particularly in arid and semi-arid developing countries, where ecosystem restoration encounters multifaceted challenges that urgently require systematic and scientifically grounded spatial rehabilitation strategies. In response, China has proposed building a “life community integrating mountains, waters, forests, farmlands, lakes, grasslands and deserts” while progressively establishing key regions, such as the “Three Zones and Four Belts”, thereby developing an ecosystem conservation and restoration philosophy that is based on ecological security patterns [3]. Against this background, the construction of ESP based on ecosystem service and structural characteristics has emerged as the primary foundation for ecological restoration zoning [4,5,6].

Based on landscape ecology theory, the “ecological source-resistance surface-ecological corridor” framework has become a canonical paradigm for constructing ESP [7,8,9,10]. Ecological sources refer to patches with a high ecosystem service and landscape connectivity [11]. Currently, the identification methodology for ecological source areas remains a focal point of scholarly attention. While qualitative extraction methods that directly designate well-preserved ecosystems, such as forests, grasslands, wetlands, and nature reserves, as source areas offer objectivity, they often overlook fragmented ecological source patches [12,13]. Recent studies increasingly identify ecological sources through integrated assessments of ecosystem structural integrity and functional performance, utilizing tools like the InVEST and MSPA model to pinpoint areas exhibiting both high ecosystem service capacity and superior landscape connectivity [14,15,16,17,18]. Building upon these foundations, this study advances the InVEST–MSPA integration by identifying ecological sources to support the delineation of ERZ. The ecological resistance surface quantifies the difficulty of species’ movement between landscape patches. Common resistance surface construction methods involve either calibrating land-use type resistance values or a weighted summation of multiple resistance indicators. For instance, Cao et al. [19] employed the normalized difference vegetation index (NDVI), slope, and digital elevation model (DEM) as calibration factors, while Zhang et al. [20] integrated landscape type, DEM, NDVI, slope, distance to urban areas, distance to road networks, and ecological barriers as composite resistance indicators. Currently, no unified indicator framework exists for resistance surface construction, with land-use type calibration remaining more prevalent due to its operational simplicity [21,22]. Ecological corridors serve as critical conduits for material/energy flows and species migration between sources, effectively mitigating ecosystem fragmentation [23,24]. Although the minimum cumulative resistance (MCR) model remains widely used for corridor extraction [25,26], its limitation to identifying single optimal paths often fails to capture ecosystem complexity. Recent advances demonstrate that circuit theory models better simulate species’ stochastic movement patterns, significantly improving corridor identification accuracy [27,28,29].

However, current research predominantly focuses on ecological security pattern construction, with ecological restoration zoning often relying on administrative divisions [14,15], ecosystem service supply–demand relationships [30], ecological vulnerability evaluations [31], or ecosystem health assessments [32], combined with natural breaks classification. These approaches lack systematic pathways for translating “ecological networks” into “restoration zoning”, failing to support regionally differentiated conservation and restoration strategies. Several recent studies have attempted to link ecological networks with restoration zoning, providing valuable methodological insights [33]. However, most of these efforts have been concentrated in humid or mountainous regions and have not fully addressed the challenges of translating ecological network structures into restoration-oriented spatial strategies in arid and semi-arid sandy ecosystems. Moreover, ecological restoration strategies remain ambiguous for high human-activity areas like croplands, particularly in sandy lands and ecotones where ecosystem degradation manifests not as visible bare land expansion, but as “ecological function weakening”, demanding precise identification and zonal restoration approaches. Therefore, the HSL was selected as a typical case area for ecological restoration to supplement existing research. Historically, the Horqin Grassland was rich in water and vegetation. However, with the intensification of drought, caused by global climate change and overgrazing activities, the region has experienced grassland degradation, land desertification, and groundwater decline. These processes have led to the weakening of key ecosystem services, such as soil and water conservation, wind prevention and sand fixation, carbon sequestration, and biodiversity maintenance, and the grassland has gradually evolved into the largest sandy land in China. Conducting ecological restoration research in the HSL is of great importance for enhancing the supporting capacity of the Xiliao River Basin, strengthening the dominant function of the northern sand-prevention belt, and consolidating the ecological security barrier of northern China, which also contributes to global ecological governance.

This paper contributes to addressing these research gaps through a comprehensive study of the HSL. As shown in Figure 1, under the “Source–Resistance–Corridor–Note” paradigm, an ecological security pattern was constructed to reveal the spatial organization of ecological processes, which served as the foundation for establishing a restoration zoning scheme. Specifically, this study integrated an ecosystem service assessment and MSPA to identify ecological sources, and applied circuit theory to extract ecological corridors and key ecological nodes, such as pinch points and barriers. Building upon these findings, we propose a “Five Zones, Three Belts, One Core” spatial framework for ecological conservation and restoration, aiming to provide both scientific evidence and innovative paradigms for ecosystem rehabilitation in arid and semi-arid regions. The study specifically addresses three key scientific questions: (1) Within the ecosystem structure–function coupling framework, how can multi-source data be synthesized to identify critical ecological sources and characterize the distribution patterns and functional attributes of core ecosystem units? (2) How should we construct an ecological security pattern incorporating ecosystem services (ES), landscape connectivity, and resistance mechanisms to delineate ecological corridors, nodes, and barrier areas, thereby revealing spatial pathways of ecological processes? (3) How can we translate ecological networks into functional restoration zoning, proposing spatially differentiated management strategies “conservation–restoration–transition” to establish a science-informed spatial governance framework?

2. Materials and Methods

2.1. Study Area

The HSL is a vast sandy belt extending approximately 400 km in northeastern China, characterized by dynamically shifting boundaries without a clearly defined unified demarcation. As the Horqin Grassland Ecological Function Zone represents a nationally designated key ecological functional area located in the heart of this sandy land system, this study selects it as the focal research area.

The Horqin Grassland Ecological Function Zone is situated in a transitional ecotone between the Songnen Plain and Inner Mongolian Plateau, with geographic coordinates ranging from 121°42′–123°02′ E, and 43°22′–43°58′ N (Figure 2). Located within the northern shelterbelt region of China, this region suffers from severe desertification and soil erosion, demonstrating high ecological sensitivity. Covering an area of approximately 111,103 km² across eleven counties [34], the zone features a distinctive saddle-shaped topography, with higher elevations in the western and southern sections and relatively flat eastern plains. The area experiences a temperate continental monsoon climate, characterized by hot, rainy summers and cold, dry winters, with the mean annual precipitation ranging between 350 and 450 mm (National Bureau of Statistics of Chin). Grassland and cropland constitute the dominant land-use types, while severe grassland degradation has led to increased bare land areas and accelerated landscape fragmentation.

From the perspective of ESP, the study area exhibits three distinctive spatial characteristics. First, forest and grassland ecosystems demonstrate a pronounced northwest–southeast bimodal distribution pattern, where the western hilly forest–grassland complexes and southeastern plain wetland remnants collectively form critical ecological core areas. Second, the central zone experiences intensive anthropogenic disturbances that disrupt ecological connectivity, creating a typical ecological barrier zone with significant restoration challenges. Third, the region shows strong interdependencies among ecological processes, with highly coupled ES, such as water conservation and windbreak-sand fixation functions, making it a priority area for conservation–restoration zoning. As such, the HSL serves not only as a nationally significant ecological functional zone, but also as a key component of China’s northern ecological security barrier and desertification control strategy belt. The area’s complex topography, diverse ecosystem structures, and dynamic land-use transitions provide an exemplary case for developing ESP and restoration zoning based on structure–function coupling mechanisms, with broad applicability across similar ecosystems.

2.2. Data Sources

The dataset in this study includes land-use data, topographic data, NDVI data, meteorological data and soil data (

Table 1. Data sources and introductions.

Data	Type	Resolution	Year	Sources
Land use	Raster	30 m	2020	https://zenodo.org/records/5816591 (accessed on 15 August 2024)
NDVI	Raster	1000 m	2020	https://www.resdc.cn (accessed on 15 August 2024)
Soil	Raster	1000 m	-	https://www.fao.org/soils-portal/data-hub/soil-maps-and-databases/harmonized-world-soil-database-v12/en/ (accessed on 17 August 2024)
DEM	Raster	30 m	2013	https://lpdaac.usgs.gov/ (accessed on 15 August 2024)
Slope	Raster	30 m	2013	Calculated from DEM
Snow depth	Raster	0.25°	2020	https://data.tpdc.ac.cn/ (accessed on 25 August 2024)
Precipitation, potential evapotranspiration	Raster	1000 m	2010 2015, 2020	https://data.tpdc.ac.cn/ (accessed on 25 August 2024)
Meteorological data (weed direction, speed, etc.)	Raster	1000 m	2020	https://data.cma.cn/ (accessed on 10 September 2024)
Plant available water content	Raster	%	-	https://data.isric.org/ (accessed on 17 September 2024)
Root restricting layer depth	Raster	250 m	-	https://data.isric.org/ (accessed on 17 September 2024)

). The same projection coordinate system (Albers_Conic_Equal_Area) and spatial resolution (30 m × 30 m) were utilized for all data and all grid data, respectively.

2.3. Methodology

2.3.1. Identifying Ecological Sources

Ecological sources are patches that play a critical role in providing fundamental ES. This study employs a structure–function coupling model to identify ecological sources. First, five key ES were selected, including water yield, soil conservation, carbon storage, habitat quality, and windbreak and sand fixation. The selection was based on the study area’s natural characteristics, consistency with prior studies conducted in arid regions [35,36,37] and alignment with the targets of Sustainable Development Goal 15 (life on land), proposed by the United Nations in 2015. After consulting the relevant literature and expert advice [19,38,39], these services were equally weighted to generate a comprehensive ecosystem service index through overlay analysis [40]. Using the natural breaks method, the ESs were classified into five levels [11,23], and areas identified as “extremely important” and “important” ecological sources were extracted as preliminary ecological sources. Second, MSPA was applied to compute the ecosystem structure and identify core areas. Finally, the overlapping regions between the core areas and the preliminary ecological source areas were delineated as the final ecological source areas.

(1)

ES

This study employed the revised wind erosion equation (RWEQ) model to quantify the windbreak and sand fixation service, while the other four ES were evaluated using the InVEST 3.16.1 software.

• Soil conservation.

The InVEST model’s sediment delivery ratio (SDR) module was employed to estimate soil conservation, encompassing both potential retention and potential soil loss. The computation follows the equations derived from references [35,36]:

$$SD=RKLS-USLE$$

(1)

$$RKLS=R\times K\times LS$$

(2)

$$USLE=R\times K\times LS\times P\times C$$

(3)

where $SD$ denotes soil retention (t/hm²·a), $RKLS$ represents potential soil erosion (t/hm²·a), and $USLE$ refers to potential soil loss (t/hm²·a). $R,K$ and $LS$ orrespond to rainfall erosivity (MJ·mm/m²·h·a), soil erodibility factor (t·hm²·h/hm²·MJ·mm), and the topographic factor, respectively, while $C$ and $P$ account for the factor of vegetation cover and crop management, respectively: a factor of soil and water conservation measures without dimension

• Windbreak and sand fixation.

The function of windbreak and sand fixation was assessed through the revised wind erosion equation (RWEQ) model, which quantifies potential and actual soil loss driven by aeolian processes. The relevant equations are as follows [37]:

$$Q_{r max}=109.8\times \left[WF\times EF\times SCF\times K^{\prime }\right]$$

(4)

$$S_{\mathrm{r}}=150.71\times \left[WF\times EF\times SCF\times K^{\prime }\right]$$

(5)

$$S_{\mathrm{Lr}}={\displaystyle \frac{2z}{S_r^2}}{Q}_{r max}\times e^{-{\left({\displaystyle \frac{z}{S_r}}\right)}^2}$$

(6)

$$Q_{max}=109.8\times \left[WF\times EF\times SCF\times K^{\prime }\times C\right]$$

(7)

$$S=109.8\times {\left[WF\times EF\times SCF\times K^{\prime }\times C\right]}^{-0.3711}$$

(8)

$$S_{\mathrm{L}}={\displaystyle \frac{2z}{S^2}}{Q}_{max}\times e^{-{\left({\displaystyle \frac{z}{S}}\right)}^2}$$

(9)

$$G=S_{\mathrm{Lr}}-S_{\mathrm{L}}$$

(10)

where $G$ is the amount of sand fixation (kg/m²); $S_{\mathrm{Lr}}$ and $S_{\mathrm{L}}$ are potential soil wind erosion (kg/m²) and actual soil wind erosion (kg/m²), respectively; $Q_{r max}$ and $Q_{max}$ are the potential and actual maximum transfer volume of blown sand (kg/m), respectively; $S_{\mathrm{r}}$ and $S$ are potential block length (m) and critical block length (m), respectively; $z$ is the maximum downwind erosion occurrence distance (m); $WF$ is the meteorological factor (kg/m); $EF$ is the soil erodibility factor (no dimension); $SCF$ is soil crust factor (dimensionless); $K^{\prime }$ is the surface roughness factor (dimensionless); and $C$ is vegetation cover factor (dimensionless).

• Carbon Storage.

Based on the carbon density method, the carbon storage and sequestration module in the InVEST model is used to calculate carbon storage. The calculation formula is as follows [41]:

$$C={\displaystyle \sum }_{m=1}^n{A}_m\times \left(C_{m·a}+C_{m·b}+C_{m·c}+C_{m·d}\right)$$

(11)

where $C$ denotes the total carbon storage of the study area, $A_m$ is the area of land-use type $m$ (hm²), and $C_m$ is the carbon density value of the four carbon pools corresponding to land-use type $m$: namely, above-ground carbon density ($C_{m·a}$, t/hm²), underground carbon density ($C_{m·b}$, t/hm²), soil carbon density ($C_{m·c}$, t/hm²), and carbon density of dead organic matter $C_{m·d}$, t/hm²).

With reference to the relevant literature [42,43], the temperature and precipitation were used to correct the carbon density value. We revised the national land carbon density data and the Loess Plateau carbon density data when the land types were not in the national land carbon density data. Finally, the carbon density value of the study area was shown in

Table 2. The carbon density for each land-use type in the study area (mg C/hm²).

Land-Use Type	Aboveground Biomass	Belowground Biomass	Soil Organic Matter	Dead Organic Matter
Cropland	2.58	0.39	79.11	0
Forest	36.88	8.6	99.58	1.85
Shrubland	2.69	1.74	78.89	0.78
Grassland	5.81	5.59	79.81	0.08
Water	3	0	0	0
Bare land	0.83	4.46	57.8	0
Impervious surface	2.5	0.08	68.93	0
Wetland	3.09	7.47	143.59	0

.

The formula is as follows:

$$K_{BP}={\displaystyle \frac{0.03\times P^{\prime }+14.4}{0.03\times P^{″}+14.4}}, K_{BT}={\displaystyle \frac{-0.4\times T^{\prime }+43.0}{-0.4\times T^{″}+43.0}},K_B=Average\left(K_{BP, }{K}_{BT}\right)$$

(12)

$$K_{SP}={\displaystyle \frac{0.07\times P^{\prime }+79.1}{0.07\times P^{″}+79.1}}, K_{ST}={\displaystyle \frac{-3.4\times T^{\prime }+157.7}{-3.4\times T^{″}+157.7}},K_S=Average\left(K_{SP, }{K}_{ST}\right)$$

(13)

$$K_{DP}={\displaystyle \frac{0.001\times P^{\prime }+0.58}{0.001\times P^{″}+0.58}}, K_{DT}={\displaystyle \frac{-0.03\times T^{\prime }+2.03}{-0.03\times T^{″}+2.03}},K_D=Average\left(K_{DP, }{K}_{DT}\right)$$

(14)

where $K_{BP}$ and $K_{BT}$ are the biomass coefficients for the precipitation and temperature, respectively; $K_B$ refers to the biomass correction coefficient; $K_{SP}$ and $K_{ST}$ correspond to the soil coefficients for the precipitation and temperature, respectively; KS denotes the soil organic correction coefficient; $K_{DP}$ and $K_{DT}$ represent the dead coefficients for the precipitation and temperature, respectively; $K_D$ represents the final dead organic correction coefficient; P′ and P″ represent precipitation in the study area and on a national scale or the Loess Plateau (mm); T′ and T″ represent temperature in the study area and on a national scale or the Loess Plateau (°C).

• Water Yield.

The water yield module of the InVEST model is used to calculate the water yield. The specific formula is as follows [36,37]:

$$Y_{xj}=(1-{\displaystyle \frac{AE{T}_{xj}}{P_x}})\times P_x$$

(15)

$${\displaystyle \frac{AE{T}_{xj}}{P_x}}={\displaystyle \frac{1+w_x+R_{xj}}{1+w_x+R_{xj}+1/R_{xj}}}$$

(16)

where $Y_{xj}$ is the annual water quantity (mm) of the land-use type $j$ in cell $x$; $AE{T}_{xj}$ is the actual evapotranspiration (mm) of the land-use type $j$ in cell $x$. $P_x$ is the annual precipitation of cell $x$ (mm). $R_{xj}$ is the Budyko dimensionless index of the land-use type $j$ in cell $x$. $w_x$ is the ratio of annual vegetation water requirements to precipitation.

• Habitat Quality.

Habitat quality was assessed using the InVEST habitat quality module, which evaluates the relationship between habitat suitability and degradation. The formulation is given in Equation [44]:

$$Q_{xj}=H_j\left[1-\left({\displaystyle \frac{D_{xj}^z}{D_{xj}^z+k^z}}\right)\right]$$

(17)

where $Q_{xj}$ refers to the habitat quality index, $H_j$ represents the habitat sensitivity of the land-use type $j$, $D_{xj}^z$ denotes the habitat degradation degree of the land-use type $j$ in grid $x$, and $k^z$ accounts for the half-saturation constant; and z corresponds to the normalization constant, typically taken as 2.5 [45,46].

(2)

MSPA

The MSPA model is developed to identify areas with better connectivity. Land-use types are divided into foreground and background. The foreground is the ecological space including forest, grassland, shrub, water, and wetland, and their value is 2. Bare land, impervious surfaces, and cropland are regarded as the background, and their value is 1. Based on the Guidos Toolbox 3.3 software, the MSPA calculated core areas, islands, bridges, edges, voids, loops, and branches. The core area, with less disturbance and most species, has been widely regarded as an ecological source [32,47].

2.3.2. Revising Ecological Resistance Surface

Ecological resistance refers to the impedance encountered during the exchange of material flow, energy flow, information flow, and other ecological processes between different ecological sources. With reference to existing studies [19,22], the construction of the ecological resistance surface in this study involved three main steps. First, initial resistance values were assigned to different land-use types. Based on research by Wang et al. [48] and Cao et al. [19], and considering the regional characteristics of the study area, the initial resistance values for forest, grassland/shrubland, water/wetland, cropland, bare land, and impervious surfaces were set as 1, 10, 30, 50, 70, and 100, respectively. Second, the digital elevation model (DEM), normalized difference vegetation index (NDVI), and slope were selected to modify the resistance surface [19]. These three factors were normalized and assigned equal weights. Third, the initial resistance values of land-use types were adjusted using the modified resistance factors. The correction formula is as follows [19]:

$$R_a^{\prime }=C_f\times R_a$$

(18)

$$C_f={\displaystyle \frac{1}{3}}DEM+{\displaystyle \frac{1}{3}}NDVI+{\displaystyle \frac{1}{3}}slope$$

(19)

where $R_a^{\prime }$ is the final resistance surface, $C_f$ is the modified resistance surface, and $R_a$ is the initial resistance value of land-use type $a$.

2.3.3. Extracting Ecological Corridors and Ecological Nodes

Ecological corridors establish connectivity between ecological sources, facilitating the exchange of energy, information, and ecosystem functions among species, and serve as vital channels for ecosystem restoration. Ecological barrier points, on the other hand, denote locations where such ecological connectivity is weakened or interrupted, hindering species dispersal and energy transmission. The circuit theory model treats the ecological resistance surface as a resistor network, simulating the migration process of species between ecological sources. Using the Linkage Mapper Toolkit in ArcGIS (https://www.arcgis.com/home/index.html), ecological corridors and nodes were extracted [49,50].

The Build Network and Map Linkages function was applied using both cost-weighted and Euclidean distance adjacency analyses to delineate optimal corridors under varying cost thresholds. The pinch point Mapper module, adopting an all-to-one analytical framework, was used to compute the cumulative current density, with high-current regions representing critical bottlenecks in the ecological network. Finally, the Barrier Mapper tool, employing a 1000 m search radius, was used to locate areas with elevated barrier scores that signified potential ecological obstacles.

2.3.4. Dividing ERZ

Building upon the established literature and considering the fundamental characteristics of ecological sources, corridors, and nodes, we propose the following zoning criteria (

Table 3. Ecological zoning classification criteria.

Landscape Feature	Zoning Category
Ecological sources	Conservation zone
High-resistance areas	Restoration zone
Pinch points	Priority protection area
Barrier points	Critical restoration area
Corridor connectivity	Core area

and

Table 4. A classification of ecological restoration.

Types	Classification Criteria
Naturally dominated recovery	Relatively low resistance values
Engineering restoration dominant	Extremely high resistance values

): (1) ecological sources are designated as conservation zones; (2) areas with high resistance values are classified as restoration zones; (3) ecological pinch points are identified as priority protection areas; (4) ecological barrier points are demarcated as critical restoration areas; and (5) the centroid of the ecological source with the highest corridor connectivity is designated as the primary core area, while sources with relatively high connectivity are classified as secondary core areas. Through systematic integration of these elements with the regional ecological baseline, we developed a comprehensive ecological protection and restoration zoning framework.

3. Results

3.1. Ecological Sources

3.1.1. Evaluation of ES

Ecosystem service functions in the Horqin Grassland Ecological Function Zone have significant spatial heterogeneity (Figure 3). Overall, habitat quality (Figure 3b), soil conservation (Figure 3e), and carbon storage (Figure 3d) services have high value in the western mountain area, whereas the windbreak and sand fixation (Figure 3a) and water yield (Figure 3c) services’ high value was predominately located in the eastern plain area. Specifically, the Kailu County, Horqin left wing back banner, Hure Banner, Naiman Banner, and Ongniud Banner showed dominant high-value areas of windbreak, sand fixation, and water yield services, in which the main land-use types are cultivated land and bare land. Habitat quality and carbon storage services’ high-value areas are distributed in the northwest of the Arukorqin Banner, Baarin Right Banner, Jarud Banner and Horqin Right Wing Middle Banner, where the land-use types are mainly grassland and woodland. Soil conservation’s spatial distribution trend was similar to the habitat quality and carbon storage services, but it was more widespread and distributed in the south of the Hure Banner, Naiman Banner, and Ongniud Banner.

The importance levels of ESs in the study area were obtained by the weighted superposition of five ESs and a natural breakpoint method (Figure 4). The northwest mountainous and the southeast plain areas showed extremely important and important ESs. The reason for this is that the areas have a high proportion of forest and grassland. The importance level of ES in the central region is relatively low, which distributed cropland, bare land, and impervious surfaces. Extremely important and important ecosystem service areas were identified as primary ecological sources, with an area of about 19,261.45 km², accounting for 17.3%.

3.1.2. The Consequence of MSPA

The landscape based on MSPA shows that the core area is mainly distributed in the northwest and southeast regions, with the largest area accounting for 39.8% (Figure 5). The core area is mainly distributed in the ecological conservation redline area, such as the Greater Khingan Mountains water conservation and biodiversity maintenance function ecological conservation redline area in the northwest, and the windbreak and sand fixation function ecological conservation redline area in the southeast sandy grassland. These places have a good ecological environment and landscape connectivity, which make them suitable for choosing as ecological sources. The bridge, loop, and branch areas play the role of landscape connection, accounting for 5.2%, 2.6%, and 2.9%, respectively. Isolated islands, known as “stepping stones”, are of great significance in maintaining ecological security, accounting for 2% of the total area. The edge and perforation area are distributed in the inner and outer edges of the core area, which have a certain effect on protecting the core area. In the study area, the edge and perforation areas account for 7.7% and 3%, which are relatively large areas and could protect the core area well. In general, the core areas are concentrated and contiguous, with good protection and good landscape connectivity, forming a healthy and stable ecosystem.

3.1.3. Ecological Sources Identication

From a structure–function perspective, this study integrated landscape pattern analysis with ecosystem service assessment to identify 76 critical ecological source areas, primarily distributed in the western mountainous regions and southeastern plains, including the Jarud Banner, Arukorqin Banner, Hure Banner, Horqin Left Wing Middle Banner, and Baarin Right Banner (Figure 6). These sources cover a total area of 10,204.38 km², accounting for 9.18% of the study area. Dominated by forest and grassland ecosystems with minimal anthropogenic disturbances, these zones play a vital role in maintaining both the structural integrity and service functions of the ecosystem. The largest ecological source patch has an area of 1270.83 km² and is located in the northwestern mountainous region, while the smallest patch, with an area of 50.12 km², is situated adjacent to the largest patch. Overall, aside from the largest patch, the remaining ecological source patches exhibit relatively small differences in size, with a mean area of 134.27 km².

3.2. Resistance Surfaces

Through the revision of land-use types’ resistance values, the spatial distribution of each resistance factor and comprehensive resistance surface is shown in Figure 7. The high resistance values were found in places with high urbanization levels and serious ecological degradation, such as the central part of Kailu County and the southern part of the Kulun Banner and the Naiman Banner, where there are concentrated bare lands and impervious surfaces. Meanwhile, the spatial distribution trend of low resistance value was similar to ecological source areas, where there was concentrated high vegetation coverage and less human disturbance.

3.3. Ecological Corridors and Ecological Nodes

The analysis involved varying the corridor width from 5 km to 50 km to evaluate how the ecological connectivity and the identification of key ecological corridors change with different thresholds. The results indicate that the 20 km threshold represents a reasonable balance between ecological connectivity and practical feasibility in the HSL. Using circuit theory modeling, 169 ecological corridors have been identified, totaling 4071.94 km in length, exhibiting a distinct west-dense, east-sparse spatial distribution pattern (Figure 8).

Analysis of current density results identified critical ecological pinch points in areas with higher current density values, which constrain species movement and energy flows. As shown in Figure 8, these pinch points cover 239.91 km² and exhibit clustered distributions in: (1) the southeastern Ar Horqin Banner and Bairin Left Banner, and (2) the northeastern Ongniud Banner. These zones serve as vital connectivity enhancers between northwestern and southwestern ecological source areas, warranting prioritized conservation efforts.

Ecological barrier points represent critical zones that disrupt species migration and energy flows between ecological sources, thereby compromising the ecosystem equilibrium. The analysis reveals that these barriers exhibit a fragmented spatial distribution, covering 568.85 km², primarily along corridors connecting the northwestern and southeastern ecological sources. Dominated by bare land, impervious surfaces, and croplands, these areas demonstrate significantly reduced ecosystem service capacity, and thus constitute priority targets for restoration interventions. The classification criteria are determined by comprehensively evaluating each barrier point’s landscape resistance value and its land-use type.

3.4. Ecological Conservation and Restoration Zones

Based on comprehensive analyses of ecological sources, resistance surfaces, corridors, pinch points, and barrier points, following the proposed zoning criteria, this study established a “Five Zones, Three Belts, One Core” spatial framework for ecological conservation and restoration (Figure 9). The “Five Zones” are the northwest core protection zone, the southwest ecological protection zone, the central ecological restoration transition zone, the southeast wetland ecological restoration zone, and the northeast ecological restoration zone. The “Three Belts” are the wind-prevention and sand-fixation belt, the ecological protection belt of the Xiliao River Basin, and the ecological restoration belt. The “One Center” is the key ecological protection center.

4. Discussion

4.1. Analysis for Improvement Countermeasures

The ecological restoration zoning of the Horqin Sandy Land fully considers multiple factors such as the importance of ecological functions, the connectivity of ecological structure, and the vulnerability of ecosystems, aiming to restore and enhance the stability and sustainability of the ecosystem. Similarly to the approach taken in the Maiji District [35], where divisions were based on “types of outstanding ecological problems,” our zoning strategy responds to the core drivers of desertification.

The “Five Zones” primarily focus on the ecological conservation and restoration through the systematic management of mountains, waters, forests, farmlands, lakes, grasslands, and sandy areas in the Horqin grasslands. The Western Ecological Conservation Zone and the Core Protection Zone, located in mountainous areas such as the foothills of the Greater Khingan Mountains, with dense ecological sources covering 76.27% of the total ecological source area, should continue to strengthen the protection of natural forest resources, actively promote pilot construction of natural parks like the Argun Dulun Grassland Park and the Elun Grassland Park, and rationally adjust the boundaries of protected areas. Initiatives such as precision forest quality improvement projects should be implemented to comprehensively advance the conservation of natural forests, the management of public welfare forests, and the restoration of degraded forests, thereby enhancing forest functions like water conservation, and windbreak and sand fixation. The Central Ecological Restoration Transition Zone and the Northern Ecological Restoration Zone coincide with high resistance, where the average resistance value is greater than 36 and the maximum resistance value is 54.99, and major barrier areas are identified by the circuit theory. These areas are predominantly characterized by three severe sandy belts: the southern bank of the Laoha River, the banks of the Jiaolai River, the Agula Sandy Belt, and the Taminchagan Sandy Belt, consisting mainly of mobile and semi-mobile sandy lands. It is recommended to combine the establishment of shelterbelt forests with the construction of plant-generated sand barriers to achieve mechanical sand fixation. Through the natural development of plants, sand barrier communities can form, accelerating the restoration and reestablishment of vegetation in sandy areas and enhancing the stability of sand-fixing plants. For fixed sandy lands with sparse vegetation, measures such as creating windbreak and sand-fixation forest networks and artificially reseeding grasses should be adopted to increase forest and grassland coverage, reduce the proportion of moderately and severely desertified land, and decrease soil wind erosion in desertified areas. The Southeastern Wetland Ecological Restoration Zone, identified as a pinch point area with a high water yield (Figure 3c), plays a vital role in maintaining regional hydrological balance. Restoration here should prioritize the protection of national wetland parks such as the Hulistai Naor, restore degraded wetlands, enhance wetland ecological functions, and strengthen the conservation of important river sources, upstream wetlands, and critical water sources in key rivers like the West Liaohe River, Xar Moron River, Laoha River, Jiaolai River, Xinkai River, and Ulgai Moron River. This will help to build an ecological protection system and improve water conservation functions. It is noteworthy that greater efforts should be made to protect farmland in sandy areas [3], vigorously promote water-saving technologies, such as shallow-buried drip irrigation, especially in southern areas with a high water yield (Figure 3c), and adjust and optimize the crop cultivation structure in sandy farmland to prevent land degradation and desertification caused by reclamation practices.

The delineation of the “Three Belts” holds significant importance for establishing ecological barriers and maintaining regional ecological security. The Windbreak and Sand-Fixation Belt, located along forested corridors in the western foothills of the Greater Khingan Mountains, aligns with high corridor connectivity and functions as a landscape-scale windbreak to block sand transport from high-resistance areas. The West Liaohe River Ecological Protection Belt, a typical pinch-point region, faces corridor interruption due to vegetation degradation and land desertification [19]; therefore, measures such as riparian buffer restoration, wetland rehabilitation, and water resource management are necessary to maintain the integrity of ecological corridors [51]. The Ecological Restoration Belt, dominated by barrier points in cropland and bare land in the north of the region, requires reseeding, windbreak establishment, and conversion of marginal farmland to reduce resistance and enhance connectivity. Overall, compared to previous studies, the “Three Belts” identified in this research provide a more in-depth analysis of ecological restoration in the study area from the perspectives of overall ecosystem stability and optimization of ecological conservation strategies, offering greater practical significance. Moving forward, efforts should focus on promoting agricultural water conservation in the West Liaohe River Basin, establishing riparian and lakeside buffer zones, actively advancing ecological restoration and protection of the water environment, and implementing both in situ and ex situ conservation measures for rare and endangered aquatic species, as well as important aquatic genetic resources. These actions will help safeguard water conservation ecological spaces and maintain biodiversity.

The “Core Zone” represents the area with the highest corridor connectivity, linking 11 corridors and serving as the central hub of the ecological network [52]. Located between the Huolin River and the Kunduleng River, this area includes the Inner Mongolia Hanshan National Nature Reserve and is home to important ecological public welfare forests, playing a crucial role in maintaining ecological balance and ecosystem connectivity. Therefore, it should focus on strengthening ecological protection, restoring fragmented patches, and maintaining connectivity among key corridors to ensure regional ecological security and ecosystem stability.

4.2. Limitations and Prospects

This study considers the structure and function of the ecosystem, uses the methods of ecosystem service and morphological spatial patterns, and combines circuit theory to construct an ecological protection and restoration pattern, providing scientific guidance for ecological restoration in arid areas. Nevertheless, there are still some limitations in this study. First, it constructs ecological protection and restoration zones only from an overall and macroscopic perspective, ignoring scale integration and cross-scale connections, which may weaken the effectiveness of ecological protection and restoration [53,54]. Second, it should be noted that some of the datasets used in this study have relatively coarse spatial resolution and temporal inconsistencies, and field validation was limited. Future research could improve accuracy by using higher-resolution and temporally consistent data, combined with more extensive ground verification. In addition, the study area largely corresponds to regions affected by historical land degradation and desertification trends (1990–2020). Integrating multi-decadal desertification trend data could further refine prioritization in future studies, and reinforce the urgency of restoration actions. Future research should focus on spatio-temporal scale issues to improve and refine the ecological protection and restoration pattern.

5. Conclusions

This study takes the HSL as a case area and integrates ecosystem structure and function to support ecological planning in an agropastoral landscape. The research addresses the following:

(1)

Identification of critical ecological sources: By coupling multiple datasets on ecosystem structure and function, key ecological source areas were identified, and their spatial distribution and functional attributes were characterized, highlighting regions that were critical for maintaining landscape connectivity and ecological integrity.

(2)

Construction of ESP: Using circuit theory, ecological corridors, nodes, and barrier points were delineated by integrating ES, landscape connectivity, and resistance mechanisms. This revealed the spatial pathways of ecological processes and identified areas where the ecological function is constrained by human activities, such as agriculture and settlements.

(3)

Translation into functional restoration zoning: The ecological network was transformed into a spatially differentiated protection and restoration pattern (“Five Zones, Three Belts, and One Center”), providing management strategies along a conservation–restoration–transition gradient. Policy recommendations were proposed to balance trade-offs between agriculture, settlements, and ecological protection, including passive restoration in bare lands, establishment of shelterbelts along cropland edges, and guidance for settlement expansion, away from critical ecological nodes.

Overall, this study demonstrates the value of combining multi-source data, MSPA, and circuit theory with a zoning-based framework to generate science-informed spatial governance strategies, reduce reliance on administrative boundaries, and provide practical guidance for sustainable ecosystem management in agropastoral sandy landscapes.