Key Areas of Ecological Restoration in Inner Mongolia Based on Ecosystem Vulnerability and Ecosystem Service

Abstract

Inner Mongolia is located in China’s arid and semi-arid regions, with sensitive and fragile ecosystems at risk of increased desertification, necessitating ecological restoration. However, economic resources for large-scale ecological restoration are often scarce, so it is vital to identify key areas for ecological restoration. Previous desertification research has focused mainly on the condition and changes in soil or vegetation. However, not all changes in soil or vegetation directly impact humans. New perspectives are increasingly needed to bridge the gap between biophysical and human well-being. We construct a framework to identify priority restoration areas based on ecosystem services and ecosystem vulnerability over a long time series. The results show that: (1) soil conservation services in northeast and southwest Inner Mongolia have degraded. Sand fixation services in central and eastern Inner Mongolia have shown a degradation trend. Habitat quality has been generally stable and sporadic in the past 20 years. (2) The areas with higher ecosystem vulnerability are concentrated in the northeast, mainly due to higher climate exposure and climate sensitivity but relatively lower climate resilience in the northeast. (3) Compared with the results of ecological restoration areas identified based on the trends of traditional vegetation indicators (fractional vegetation cover and net primary productivity), we found a greater proportion of land in northeastern Inner Mongolia in need of restoration. Additionally, there was identified a decreased restoration proportion in southwestern Inner Mongolia to ensure the self-restoration and regulation of desert ecosystems, which is conducive to realizing nature-based solutions.

1. Introduction

Under the combined pressure of climate change and human activities, two-thirds of the world’s drylands are moderate to severely degraded, with 12 million hectares of new degraded land added each year [1]. Land degradation’s direct or indirect cost to society is approximately USD 300 billion [2]. Desertification is the land degradation of dryland ecosystems, a continuous decline in the ability of dryland ecosystems to provide services over time [3]. China’s arid and semi-arid regions are vast, with sensitive and fragile ecosystems, and the risk of desertification increases under the pressure of climate change and human activities [4,5,6]. Ecological restoration can increase renewable and nurturing natural capital, thereby maintaining biodiversity and improving the quality of ecosystem goods and services on which our health and well-being depend [7,8]. Therefore, ecological restoration is an effective method to mitigate desertification. Ecological restoration has become increasingly important in global and national treaties, coalitions, and UN conventions [9]. The Biodiversity Strategy 2030 program [10] and Convention on Combatting Desertification [11] provide vital support to call for large-scale ecological restoration. These conventions emphasize ecological restoration as an effective and necessary means of reducing greenhouse gas emissions to achieve land degradation neutrality and carbon neutrality [9]. However, economic resources for large-scale ecological restoration are often scarce [12], so we cannot implement ecological restoration projects in every degraded area. At the same time, the degree of degradation in each ecosystem is different, and there should be priorities in ecological restoration. Therefore, it is necessary to find areas with a high risk of desertification. Identifying key ecological restoration areas will help maximize the supply of multiple ecosystem services and enhance human well-being within limited economic conditions.

The more widely accepted definition of ecological restoration is the process of assisting in the recovery of degraded, damaged, or destroyed ecosystems [13]. This definition facilitates understanding the purpose of ecological restoration, and its primary step is to find the degraded areas. Previous Ecological restoration research mainly focuses on single object restoration or single function restoration [14,15]. In addition, biodiversity conservation and ecosystem vulnerability are also important research contents of ecological restoration [15,16,17]. Ecosystem vulnerability (EV) is the risk of damage to ecosystems due to disturbance [18]. Ecosystems possess different vulnerabilities due to their different natural conditions and show different sensitivity and resilience to human activities and climate change [19]. One of the key points of ecological restoration is to restore areas with high vulnerability. In recent years, there has been a growing view that ecological restoration should be approached from the perspective of human wellbeing [20] in order to meet social needs and restore important ecological and social functions.

The ecological restoration under desertification needs to identify areas with a high risk of desertification. Previous desertification studies have mainly focused on the changes in soil or vegetation [21,22]. However, not all changes in soil or vegetation directly impact humans. Ecosystem services are the bridge connecting humans and nature [23], and it is the manifestation of human beings benefiting from nature. A decline in ecosystem services represents an impairment of human wellbeing. However, as defined by a series of global greening studies, vegetation restoration may mask the degradation of ecosystem services, and new perspectives are increasingly needed to bridge the gap between biophysical and ecosystem services. Some researchers have conducted ecological restoration studies based on ecosystem services in recent years. However, these ecological restoration studies are mainly based on ecological security patterns [24,25], the current status of ecosystem services [26], and ecosystem services trade-offs [27,28] to identify ecological restoration areas. Previous studies have not included the dynamic changes of ecosystem services over time in ecological restoration studies.

Overall, we can identify priority areas for ecological restoration from ecosystem services and ecological vulnerability. However, areas with degraded ecosystem services and high ecological vulnerability do not entirely overlap due to meteorological factors and geographical conditions. Therefore, combining long time-series ecosystem service analysis with ecological vulnerability is beneficial for identifying areas with continuously declining ecosystem services and high ecological vulnerability and determining priority areas for ecological restoration.

Inner Mongolia is located in the arid and semi-arid region of China, with typical desertification characteristics, and it is essential to identify priority areas for ecological restoration. We assessed the trends of soil conservation services, sand fixation services, and habitat quality in Inner Mongolia from 2000 to 2020. Then, we classified Inner Mongolia into five degradation categories according to the trends and significance of ecosystem services. At the same time, we calculated the vulnerability of Inner Mongolia to climate change. By combining the degradation trends and ecosystem vulnerability of Inner Mongolia to identify priority areas for desertification restoration, we provide a reference for ecological restoration in the region. Our objectives are: (1) to construct a framework of priority restoration areas based on ecosystem services and ecosystem vulnerability over a long time series, (2) to classify restoration areas (3) to propose ecological restoration recommendations. Identifying ecological restoration priority areas is beneficial to maximize the restoration, which is vital to carbon sequestration [29] and reducing desertification in Inner Mongolia.

2. Materials and Methods

2.1. Study Area

Inner Mongolia is located between 97°12′–126°04′E and 37°24′–53°23′N, in the arid and semi-arid region of China (Figure 1a). The average elevation of Inner Mongolia is 1000 m, and the primary landform type is the plateau. The Daxinganling, Yinshan, and Helen Mountains form the natural ecological barrier between northeast, north, and northwest China. Forests, grasslands, deserts, and wetlands are widely distributed on the northwest side of the mountain line, while arable land and cities are concentrated on the southeast side of the mountain line (Figure 1b,c). The temperate continental monsoon climate dominates the climate of Inner Mongolia, and the wind direction changes significantly with the seasons. Ecological restoration in Inner Mongolia will benefit residents and play an important role in reducing the sand in the interior of China.

2.2. Dataset

We used the below data to calculate soil conservation service, sand fixation service, habitat quality, and ecosystem vulnerability (

Table 1. Data list, types, sources, datasets, and original resolution. International Soil Reference and Information Centre (ISRIC); National Aeronautics and Space Administration (NASA); Cesium (Cs); Land Processes Distributed Active Archive Center (LP DAAC); Shuttle Radar Topography Mission (SRTM); Moderate-Resolution Imaging Spectroradiometer (MODIS); Climate Change Initiative (CCI).

	Variables	Data Type	Source	Dataset	Resolution
Soil conservation service	Daily rainfall data	Station-based observation data	China Meteorological Administration (2000–2020)	China surface climate data daily data set (V3.0)	1 km, simple kriging interpolation
	Soil organic carbon	Raster data	ISRIC	SoilGrids	250 m
	Sand content	Raster data	ISRIC	SoilGrids	250 m
	Silty content	Raster data	ISRIC	SoilGrids	250 m
	Clay content	Raster data	ISRIC	SoilGrids	250 m
	Digital elevation model (DEM)	Raster data	NASA	SRTM Digital Elevation Database v4.1	90 m
	Normalized difference vegetation index (NDVI)	Raster data	LP DAAC (2000–2020)	MODIS (MOD13A1 product)	1 km
Sand fixation service	Daily wind speed data	Station-based observation data	China Meteorological Administration (2000–2020)	China surface climate data daily data set (V3.0)	1 km, simple kriging interpolation
	Daily temperature data	Station-based observation data	China Meteorological Administration (2000–2020)	China surface climate data daily data set (V3.0)	1 km, simple kriging interpolation
	Solar radiation	Station-based observation data	China Meteorological Administration (2000–2020)	China surface climate data daily data set (V3.0)	1 km, simple kriging interpolation
	Snow depth	Raster data	National Qinghai-Tibet Plateau Scientific Data Center	Long-term series of daily snow depth datasets in China (1979–2020)	25 km
Habitat quality	Land use	Raster data	European Space Agency (ESA)	CCI land cover V2	1 km, simple kriging interpolation
Vulnerability	Leaf area index (LAI)	Raster data	NASA	Global Inventory Modelling and Mapping Studies (GIMMS) LAI3g dataset	1 km
	Daily atmospheric pressure data	Station-based observation data	China Meteorological Administration (2000–2020)	China surface climate data daily data set (V3.0)	1 km, simple kriging interpolation
	Daily relative humidity data	Station-based observation data	China Meteorological Administration (2000–2020)	China surface climate data daily data set (V3.0)	1 km, simple kriging interpolation
Validation	Sediment transport	Book	National Library of China	hydrologic data yearbook	-
	137Cs	Point	[30], Shapotou Desert Experimental Research Station of the Chinese Academy of Sciences	137Cs	-
	Above-ground biomass carbon	Raster data	[31]	Passive microwave-based global above-ground biomass carbon dataset (1993–2012) version 1.0	25 km
	Tree cover	Raster data	[32]	Global Forest Change dataset	30 m

). Station-based observation data are converted into raster data by simple Kriging interpolation. All data periods are 2000–2020, with a resolution resampling of 1 km.

2.3. Key Restoration Areas Framework

We construct a framework for identifying priority restoration areas based on ecosystem services and ecosystem vulnerability over a long time series (Figure 2). In our study, priority is a level of urgency, and ecological restoration funds should be spent where it is most urgent. In this framework, priority ecological restoration in Inner Mongolia is the areas with both ecosystem services degradation and high ecosystem vulnerability. We used the Revised Universal Soil Loss Equation (RUSLE) model to calculate the soil conservation service, the Revised Wind Erosion Equation (RWEQ) model to calculate the sand fixation service, and the Integrated Valuation of Ecosystem Services and Trade-offs (InVEST) model to calculate the habitat quality. Then, we classified each ecosystem service into five degradation categories according to the Theil–Sen procedure and a Mann–Kendall significance test: significant increase, slight increase, insignificant change, slight decrease, and significant decrease. We superimposed the three ecosystem service change trends, which can spatially form the ecosystem service degradation map. In addition, ecosystem vulnerability is classified using the Jenks optimization method to identify extreme vulnerable and vulnerable areas. Finally, the areas with degraded ecosystem services and high vulnerability were superimposed to obtain the key areas for ecological restoration in Inner Mongolia.

2.4. Soil Conservation Service

Soil conservation service (SCS) is an ecosystem’s capacity to conserve soil and control erosion [33]. We used the RUSLE model to calculate soil conservation services. The calculation formula was as follows (Equation (1)):

$$Q_{sr}=R\times K\times L\times S\times \left(1-C\right)\times P$$

(1)

$Q_{sr}$ is the plot sediment retention (Mg ha⁻¹ yr⁻¹); R is the rainfall erosivity factor (MJ mm ha⁻¹ h⁻¹ yr⁻¹), calculated from the Xie equation [34]; K is the soil erodibility factor (Mg ha h ha⁻¹ MJ⁻¹ mm⁻¹), calculated using the EPIC model [35]; L is the slope length factor (dimensionless) [36]; S is the slope steepness factor (dimensionless) [37]; C is the vegetation cover and crop management factor [36]; and P is the soil and water conservation measures factor [36].

2.5. Sand Fixation Service

Sand fixation services suppress and fix wind and sand by ecosystem vegetation [38]. We used the RWEQ model to calculate sand fixation services. Sand fixation service is the difference between potential wind erosion and actual wind erosion. Potential wind erosion is soil erosion under bare soil without vegetation, and actual wind erosion is soil erosion under surface vegetation cover [30]. The calculation formula was as follows (Equations (2)–(8)):

Sand fixation capacity:

$$SR=S_{L\mathrm{p}}-S_L$$

(2)

Potential wind erosion:

$$S_{L\mathrm{p}}=\frac{2\cdot z}{S_{\mathrm{p}}{}^2}{Q}_{MAX\mathrm{p}}\cdot e^{-{(z/s\mathrm{p})}^2}$$

(3)

$$Q_{MAX\mathrm{p}}=109.8\cdot \left[WF\times EF\times SCF\times K^{\prime }\right]$$

(4)

$$S_{\mathrm{p}}=150.71\cdot {(WF\times EF\times SCF\times K^{\prime })}^{-0.3711}$$

(5)

Actual wind erosion:

$$S_L=\frac{2\cdot z}{S^2}{Q}_{MAX}\cdot e^{-{(z/s)}^2}$$

(6)

$$S=150.71\cdot {(WF\times EF\times SCF\times K^{\prime }\times COG)}^{-0.3711}$$

(7)

$$Q_{MAX}=109.8\cdot \left[WF\times EF\times SCF\times K^{\prime }\times COG\right]$$

(8)

$SR$ is the sand fixation modulus (kg m⁻² yr⁻¹); $S_{L\mathrm{p}}$ is the potential wind erosion modulus (kg m⁻² yr⁻¹); $Q_{MAX\mathrm{p}}$ is the potential maximum amount of wind-transported sediment (kg m⁻²); $S_{\mathrm{p}}$ is the potential critical plot length (m); $S_L$ is the actual wind erosion modulus (kg m⁻²); $Q_{MAX}$ is actual maximum amount of wind-transported sediment (kg m⁻²); $S$ is the critical plot length (m); $z$ is the leeward distance of the plot (m), a constant parameter at 100 [39]; WF is the weather factor (kg m⁻¹); $EF$ is the percentage of soil erodibility (%); $SCF$ is the soil crust factor (dimensionless); $K^{\prime }$ is the soil roughness factor (dimensionless); and $COG$ is the combined crop factor (dimensionless).

2.6. Habitat Quality

Habitat refers to the potential of an ecosystem to provide the conditions necessary for species to survive and reproduce and is an important ecosystem service; habitat quality determines the suitability of habitats and plays an important role in biodiversity conservation [40]. InVEST (Integrated Valuation of Ecosystem Services and Trade-off) model calculates habitat quality. The calculation formula was as follows (Equations (9) and (10)):

$$Q_{xj}=H_j\left(1-\left(\frac{D_{xj}^z}{D_{xj}^z+K^z}\right)\right)$$

(9)

$$D_{xj}={\sum }_{r=1}^R{\sum }_{y=1}^{Y_r}(\frac{W_r}{{\sum }_{r=1}^R{W}_r})r_y{i}_{rxy}{\beta }_x{S}_{jr}$$

(10)

$Q_{xj}$ is the habitat quality index of x grids in landscape j; $H_j$ is the habitat suitability of landscape j, and the value range is [0, 1]; K is the half-saturation constant, and its size is 1/2 of the raster size, so the paper is set to 500; z is the scale parameter, generally set to 2.5. $D_{xj}$ is the habitat degradation index; R is the number of threatened sources; $W_r$ is the weight; $Y_r$ is the number of threat cells; $r_y$ is the number of threat factors on the raster cell; $i_{rxy}$ is the influence range of the threat factor; ${\beta }_x$ is the accessibility of the threat source to the raster x; $S_{jr}$ is the sensitivity of landscape j to threat factor r.

2.7. Trend Analysis

We used the Theil–Sen procedure and a Mann–Kendall significance test to analyze the trend of ecosystem services from 2000 to 2020. This method has been widely used in annual vegetation growth analysis [41,42]. Then, we classified Inner Mongolia into five degradation categories according to the Theil–Sen procedure and a Mann–Kendall significance test of ecosystem services: significant increase, slight increase, insignificant change, slight decrease, and significant decrease. The calculation formula was as follows (Equations (11)–(15)):

$$S_D=Median\left(\frac{D_j-D_i}{j-i}\right)$$

(11)

$$S={\sum }_{i=i}^{n-1}{\sum }_{j=i+1 }^{n}sgn\left(D_i-D_j\right)$$

(12)

$$sgn\left(D_j-D_i\right)=\left\{\begin{array}{c}1, D_i-D_j>0\\ 0, D_i-D_j=0\\ -1, D_i-D_j<0\end{array}\right\}$$

(13)

$$s\left(S\right)=\frac{n\left(n-1\right)\left(2n+5\right)}{18}$$

(14)

$$Z=\left\{\begin{array}{c}\frac{S-1}{\sqrt{s\left(S\right)}}, S>0\\ 0, S=0\\ \frac{S+1}{\sqrt{s\left(S\right)}}, S<0\end{array}\right\}$$

(15)

$S_D$ is the trend of ecosystem services from 2000 to 2020, $S_D$ > 0 represents an increase in ecosystem services from 2000 to 2020, and $S_D$ < 0 represents a decrease in ecosystem services from 2000 to 2020; n is the length of the period; $D_i$ and $D_j$ are the observations at the time i and j; |Z| > 1.64 is significant [43].

2.8. Ecosystem Vulnerability

Ecosystem vulnerability is the degree to which an ecosystem is vulnerable or unable to cope with the adverse effects of climate change, including exposure, sensitivity, and resilience [44]. The sensitivity is an ecosystem to remain stable during a disturbance [45]. Resilience is an ecosystem that returns to its original state after a disturbance [19]. Exposure is the capacity of an ecosystem to keep up with climate change [46]. The calculation formula was as follows (Equation (16)):

$$VI=\sqrt{\frac{EI\ast SI}{1+RI}}$$

(16)

VI is the vulnerability index, EI is the exposure index, SI is the sensitivity index, and RI is the resilience index [19].

2.9. Ecological Restoration Zoning Standards

We overlayed the trend of ecosystem service changes with the ecosystem vulnerability to obtain different ecological restoration zoning. The zoning criteria are listed in the table below (

Table 2. Ecological restoration zoning standards.

Class	Class Standard
Zero degradation	No degradation
Reserve	Areas with high ecosystem vulnerability but do not have the degradation
Ecological restoration areas	Areas with a slight degradation of ecosystem services
Priority areas of ecological restoration	Areas with degraded ecosystem services and high ecosystem vulnerability

).

2.10. Model Validation

To evaluate the accuracy of the RUSLE, RWEQ, and InVEST models, we regressed the model results on the observed data. We used sediment transport modulus as observations of soil conservation service [36]. Nuclear explosions produce ¹³⁷Cs. Wind erosion is one of the main factors contributing to the transport and redistribution of this element in the soil. ¹³⁷Cs is a mature quantitative assessment method in wind erosion, with the most application cases. We used ¹³⁷Cs as observations of sand fixation service [30] and forest cover and above-ground biomass carbon as observations of habitat quality [47].

3. Results

3.1. Model Validation

We regressed the model results on the observed data, and the results show that the model results are reliable and can be further analyzed (Figure 3). The results showed that RWEQ model (sand fixation service) (R² = 0.74, Figure 3c) was more reliable than the RUSLE model (soil conservation service) (R² = 0.37, Figure 3d) and InVEST model (habitat quality) (R² = 0.28, Figure 3a; R² = 0.29, Figure 3b).

3.2. The Changing Trend of Soil Conservation Services, Sand Fixation Services, and Habitat Quality

The results show that the soil conservation service in northeastern and southwestern Inner Mongolia had a sporadic degradation trend from 2000 to 2020 (Figure 4a). The sand fixation service in eastern Inner Mongolia had a significant degradation trend from 2000 to 2020 (Figure 4b). Habitat quality in Inner Mongolia had an overall stable trend from 2000 to 2020 (Figure 4c).

Among these nine types of degradation (Figure 4d), most areas (75.84%) have no ecosystem service degradation. Besides, class I (12.42%) and III (11.35%) accounted for the largest proportion, representing one type of slight degradation of ecosystem services and one type of severe degradation of ecosystem services in most areas.

3.3. Ecosystem Vulnerability Pattern Analysis in Inner Mongolia

The results showed that the ecosystem vulnerability in northeast Inner Mongolia is high (Figure 5a). In contrast, the exposure was high in the northeast Inner Mongolia area (Figure 5b). The resilience was low in the eastern Inner Mongolia area (Figure 5c), while the sensitivity was high in the northeast part (Figure 5d). Northeastern Inner Mongolia has the highest ecosystem vulnerability, mainly due to the high exposure, sensitivity of these ecosystems.

3.4. Key Ecological Restoration Areas in Inner Mongolia

We divided Inner Mongolia into four restoration areas: zero degradation, reserve, ecological restoration, and priority ecological restoration areas. Figure 6a is the final result of the restoration areas. The results show that the protected areas (1931 km²), ecological restoration areas (2211 km²), and priority ecological restoration areas (2212 km²) is basically the same area (Figure 6a). In addition, we found that the restoration pattern is consistent with the degradation distribution of the fractional vegetation cover (FVC) (Figure 6c). We found a less proportion of southwestern Inner Mongolia in need of restoration compared with the degradation distribution of net primary productivity (NPP) (Figure 6b). Large-scale ecological projects (e.g., afforestation) in desert areas ignore the background conditions of desert areas. This may increase the water shortage and aggravate desertification. Reducing the proportion of restoration in desert areas can ensure the self-restoration and regulation of desert ecosystems, conducive to the realization of nature-based solutions.

4. Discussion

4.1. Ecosystem Services Are Influenced by Climate and Human Activities

By analyzing the trends of ecosystem services, we can find that soil conservation services in Inner Mongolia have increased in most areas year by year. Although the frequency and intensity of precipitation in Inner Mongolia have increased in recent decades [48,49], most of the vegetation cover in Inner Mongolia has increased year by year due to ecosystem restoration and afforestation and climate change [50,51,52,53]. The increase in vegetation cover can offset the pressure brought by rainfall on the soil to some extent, reduce the area of soil erosion, and enhance the ability of soil conservation services [53,54,55,56]. Mechanistically, the vegetation cover increases the interception of rainfall by the vegetation canopy [57]. At the same time, the dead leaves of vegetation have a protective effect on the soil, alleviating the pressure on the soil from increased rainfall [58]. Vegetation roots can also promote water infiltration, and soil microorganisms enhance the cohesiveness of soil aggregates [59]. In addition, vegetation enhances evapotranspiration and photosynthesis, reducing the water available for sediment transport [55,60,61,62].

The widespread distribution of barren lands and farmlands in eastern Inner Mongolia has led to a decline in sand fixation services. The land-use patterns in southeastern Inner Mongolia are barren lands and cropland. The wind erosion modulus and intensity of barren lands and croplands have been shown to exceed those of grassland [63,64,65]. Mechanistically, barren lands have no vegetation cover, resulting in direct soil exposure to wind, while croplands are heavily disturbed by anthropogenic and productive activities. Therefore, barren lands and croplands are vulnerable to wind, leading to severe soil and nutrient losses [66,67,68], with reduced sand fixation services capacity.

The changes in vegetation and climate change (precipitation changes in soil conservation services and wind changes in sand fixation services) are coupled processes. For example, in sand fixation services, the wind speed increases for a long time. The vegetation cover remains unchanged for a long time, indicating that this region’s sand fixation capacity is increasing and has not yet reached its peak. Once the wind speed increases to a high level, it will cause a decrease in vegetation cover, and the sand fixation capacity will also decrease. Simultaneously, the decrease in wind speed is generally due to the increase in vegetation. According to the calculation formula of the sand fixation service, the increase of vegetation leads to the increase of 1-C factor and the increase of the sand fixation service.

4.2. High Climate Variability in Northeastern Inner Mongolia and Poor Resilience in Eastern Inner Mongolia

Northeastern Inner Mongolia has the highest ecosystem vulnerability, mainly due to the high climate variability. Ecosystem exposure and sensitivity are mainly quantified through temporal and spatial variability of temperature and precipitation [46]. Precipitation and temperature variability are high in northeastern Inner Mongolia. Therefore, changes in precipitation and temperature can influence the spatial pattern of ecosystem vulnerability through exposure and sensitivity [19]. Resilience is quantified as the abnormally high lag-1 autocorrelation in vegetation dynamics [69]. Eastern Inner Mongolia is an agro-pastoral ecotone. Consistent with previous studies, farmland ecosystems have lower resilience [19]. In addition, farmland is subject to more anthropogenic management, leading to lower resilience of farmland ecosystems [70,71,72]. Drought, overgrazing, and unsustainable land use have reduced the resilience of the eastern Inner Mongolia agro-pastoral area.

4.3. Innovation and Limitations

Our study aims to construct a framework to identify the key areas for ecological restoration in Inner Mongolia. We use a suite of models and data to calculate long-term trends in ecosystem services and ecosystem vulnerability and to develop a classification system with four classes of restoration zones: zero degradation; reserve; ecological restoration areas; and priority areas of ecological restoration. A critical difference between our study and other publications assessing ecosystem services and vulnerability at landscape and regional scales is that we consider longer-term trends to a greater extent. In addition, we also combined long time-series ecosystem service analysis with vulnerability to identify the key areas of ecological restoration in Inner Mongolia.

Although our study identified areas of long-term degradation of ecosystem services, the selection of ecosystem service indicators is not enough, which may affect the accuracy of the analysis results. At the same time, the results are also limited by the model’s accuracy. The RUSLE model is a good compromise between soil conservation service estimates [73,74,75]. In calculating the soil conservation service, we used Xie’s model to calculate the rainfall erosivity, which is better simulated [34]. We used NDVI to calculate the C factor; however, NDVI only reflects the green vegetation information and ignores the information on the soil surface. Therefore, NDVI may overestimate the C factor. The P factor is rarely considered in calculating soil conservation services because it is difficult to estimate [76,77]. We set the P factor to 1 in the calculation of the soil conservation service, so our results do not accurately quantify the amount of soil conservation services. The RWEQ model is excellent for calculating sand fixation service [30,78]. However, we lacked hourly wind speed data during the calculation. Finally, we used the daily average wind speed to obtain hourly wind speed data [78]. To calculate the habitat quality for the long time series, we used the CCI land cover V2 data, but the CCI land cover V2 data could not accurately portray the land use information in China. In addition, we resampled all resolutions to 1 km, which leads to information loss and increases the limitations and uncertainty of the results. Although there are many shortcomings, the results of our model are credible. Moreover, in our study, priority is a level of urgency, and ecological restoration funds should be spent where it is most urgent. However, we did not consider the cost and return ratio. In addition, we also did not consider the ecosystem services supply–demand ratio. These can be further explored in future studies.

4.4. Implications for Land Management

We have to use different measures for different ecological restoration zones. We have to focus on natural restoration for protected areas and follow a nature-based solution for management. For restoration areas and priority restoration areas, we can incorporate artificial interventions to implement ecological restoration by restoring native forests or ecological restoration projects (afforestation). However, ecological restoration projects also have certain limitations and shortcomings. Large-scale afforestation increased vegetation evapotranspiration, groundwater depletion, and water shortage [79], bringing other environmental problems. Meanwhile, the monoculture of afforestation plant species may undermine the benefits brought by biodiversity. In addition, restoration of native forests usually has more significant environmental benefits than afforestation in terms of essential ecosystem services such as biodiversity conservation, carbon sequestration, soil and water conservation, and water supply [80]. To avoid creating new ecological and environmental problems, we suggest that nature-based solutions be adopted, and some sustainable ecological restoration projects should be carried out. For example, when we selected priority ecological restoration sites, long-term trends in ecosystem services should be considered. In choosing ecological restoration methods, we should prioritize native forest restoration with assisted natural regeneration of native species [81,82]. In spring, wind erosion is severe in eastern Inner Mongolia. We should use easily implemented physical sand barriers such as straw checkerboard grids, crop residue barriers, and gravel mulching to slow wind and sand transport and ultimately inhibit desertification [83,84,85,86].

5. Conclusions

This study constructs a framework of priority restoration areas based on ecosystem services and ecosystem vulnerability over a long time series and classifies restoration areas. By combining the degradation trends of ecosystem services and ecosystem vulnerability of Inner Mongolia, we found that: (1) soil conservation services in Inner Mongolia have improved overall in the past 20 years; in the northeast and southwest Inner Mongolia, have degraded; sand fixation services in central and eastern Inner Mongolia have shown a degradation trend in the past 20 years; habitat quality has been generally stable in the past 20 years, showing a sporadic degradation trend. (2) The areas with higher ecosystem vulnerability are concentrated in the northeast, mainly due to higher climate exposure and climate sensitivity but relatively lower climate resilience in the northeast. (3) Compared with the results of ecological restoration areas identified based on the trends of traditional vegetation indicators (fractional vegetation cover and net primary productivity), our results have identified the reduced restoration proportion in the southwestern part of Inner Mongolia and the increased restoration proportion in the northeastern part of Inner Mongolia.