Spatiotemporal Dynamics of Ecosystem Service Balance in the Beijing-Tianjin-Hebei Region and Its Ecological Security Barrier with Inner Mongolia

Abstract

In the context of the global decline in the capacity of ecosystem services (ESs) to meet increasing human demands, assessing and quantifying ESs is crucial for ecological policy formulation. To address this, our study employs an adjusted land-use matrix method and the patch-generating land-use simulation (PLUS) model for a quantitative analysis of the ES balance in the Beijing–Tianjin–Hebei–Inner Mongolia (JJJM) region from 2000 to 2020, projecting to 2040. Focusing on the JJJM region, a focal area for ecological policy exhibits significant socioeconomic disparities, revealing a synergistic interplay in the ESs balance. Areas with high vegetation cover, like forests and grasslands, demonstrate an elevated ESs balance, with Inner Mongolia having the highest total ESs balance at 71.40. Conversely, highly urbanized areas, such as Beijing and Tianjin, face deficits in the ESs balance, with Tianjin recording the lowest at 17.83. Our results show an upward trend in total ESs balance in the JJJM region (slope: 0.08 year⁻¹). In particular, regulating services are declining (slope: −0.04 year⁻¹), cultural services are increasing (slope: 0.08 year⁻¹), and provisioning services remain relatively stable. Projecting to 2040, our analysis indicates a slight decline in ESs balance, attributed to Inner Mongolia’s urban expansion. This emphasizes the need for industrial transfers and proactive urbanization promotion to enhance ESs balance and support sustainable management and ecological civilization development in the JJJM region.

1. Introduction

Ecosystem services (ESs), which are crucial for human survival, can be defined as the benefits of maintaining the natural environment for human existence. These also include the products and services provided directly or indirectly through ecosystem functions [1,2]. Against the current backdrop of climate change, environmental pollution, and the loss of biodiversity, the synergistic preservation and enhancement of nature’s numerous contributions to the quality of human life are crucial challenges for the achievement of the sustainable development goals [3,4]. Therefore, assessing the temporal and spatial changes in ESs is crucial for fostering socioeconomic development and facilitating the recovery of ecological benefits [5].

Studies on ESs have been conducted across various scales using diverse methods. The concept of ESs was explained clearly for the first time by Daily in “Nature’s Service: Social Dependence on Natural Ecosystems” [6]. Subsequently, Costanza et al. employed the value equivalence method to estimate the economic value of 17 ESs across 16 global biomes [5]. The Millennium Assessment Reports categorized ESs into four major types: provisioning, regulating, supporting, and cultural services [7]. However, some studies indicated that supporting services, being prerequisites and foundations for the other three types of ESs, have no direct connection to human social systems [8,9]. Xie et al. formulated an equivalence table for the functions of six major and nine minor categories of ESs based on expert knowledge [10], and, by relying on the study of crop inputs and outputs conducted by Hu and Leng in 2005 [11], they calculated the economic value of ESs. Later, based on the 2010 Statistical Yearbook data, Xie et al. proposed an improved unit area equivalence table for the basic values of ESs [11], which has been widely used in the quantification research of 4 major and 11 minor categories of ESs within China. Furthermore, Liu et al. employed multiple data sources to assess the spatiotemporal distribution of the current and future contributions of 18 natural factors to human well-being under climate models [1], establishing an indicator system with two dimensions: the potential contribution of nature and the actual local human demand. Lastly, Burkhard et al. developed a matrix associated with land use and land cover to calculate the supply, demand, and balance of ESs [12], a method that has since been applied in multiple regions [13,14,15].

Numerous studies indicate that rapid societal development, driven by climate change and land-use alterations, imposes significant uncertainty on ecosystems. The implementation of a series of policies has accelerated urbanization and industrialization, leading to the loss of natural habitats [16,17]. Due to unsustainable land use, approximately 60% of ESs have experienced a partial or complete reduction [18]. In particular, climate change can adversely impact vegetation growth and alter land-use patterns by affecting soil quality through factors such as reduced precipitation and increased evaporation [19,20]. However, Abel et al., who investigated the sensitivity of vegetation in tropical arid lands to rainfall, discovered that positive changes in developed countries may be attributed to climatic advantages and improved land-use management [21]. Li et al., using the Landsat data, extracted impervious surfaces and explored the varied impacts of urbanization on urban heat island trends [22]. Further, Venter et al. constructed the global human footprint index and analyzed its impact on the conservation of biodiversity using data on infrastructure, land cover, and human entry into natural areas [23]. Therefore, a series of ecosystem services programs have been introduced to protect ecosystems and alleviate the degradation and reduction in ESs [24].

Zhang et al., after analyzing the dynamics of vegetation in China’s arid lands, where widespread ecological engineering has been implemented, highlighted that ecological engineering and CO₂ have accelerated vegetation greening [25]. The policy implementation of such ecological engineering plays a crucial role in driving land-use changes; however, the studies carried out thus far have tended to assess the impact of historical land use on ESs. Thus, they lack predictions of changes in ESs under the prospects of social development. Such predictions could provide valuable insights for the formulation of ecological policies and the sustainable development of human societies.

The Beijing–Tianjin–Hebei–Inner Mongolia (JJJM) region, located in the northern part of China, comprises a considerable amount of arid land and has experienced significant socioeconomic development over recent decades. Specifically, the JJJM region has undergone rapid urbanization and modernization of agriculture and animal husbandry, which has prompted the implementation of numerous ecological projects that have resulted in an increasingly complex land-use situation [26]. As part of the “Capital Economic Circle”, the Beijing–Tianjin–Hebei (JJJ) region has experienced rapid economic development, even though an imbalance in the distribution of ESs persists [27]. The Inner Mongolia Autonomous Region is rich in green areas, which serve as a natural ecological barrier for the JJJM region; however, the arid grassland ecosystem in Inner Mongolia is relatively fragile, and excessive human activities have led to habitat fragmentation, which can cause ecological issues such as land fragmentation and dust storms. These issues may have far-reaching impacts on the JJJM region [28,29]. In October 2023, the Chinese government issued new policy guidelines that emphasized Inner Mongolia’s role as a crucial ecological security barrier in the north of China, urging it to coordinate the governance of mountains, rivers, forests, fields, lakes, grasslands, and deserts, along with prioritizing ecological considerations and promoting green development. Previous studies, constrained by administrative divisions, often focused on the JJJ region while overlooking the crucial role played by Inner Mongolia in supporting its ESs.

Therefore, this study made targeted improvements to Burkhard’s land-use matrix model to adapt to the unique environmental conditions and ecological project implementation in the JJJM region. The introduction of ecological factors to the model served as fine adjustments to the supply differences of ESs under the same land-use types. This study quantified the balance of three types of ESs—provisioning, regulating, and cultural—in the region and delved into the spatiotemporal evolution patterns, trade-offs, and synergies among them. Further, by utilizing the patch-generating land-use simulation (PLUS) model and considering multiple driving factors, this study predicted land-use changes in 2040 and quantified the ESs balance. Thus, this study not only reveals the current status and future trends of the ESs balance in the JJJM region but also provides data support, thereby establishing a decision-making foundation for maintaining the balance of regional ecosystems and formulating policies for the management of ESs. Additionally, this study has significant implications for promoting the sustainable development of human societies.

2. Materials and Methods

2.1. Study Area

The JJJM region encompasses Beijing, Tianjin, Hebei, and the Inner Mongolia Autonomous Region in China, with administrative divisions situated between 36°05′ N–53°19′ N and 126°04′ E–97°12′ E. The total area of the region is approximately 1.40 × 10⁶ km² (Figure 1) and is characterized by predominantly arid and semi-arid climates. Moreover, the typical temperate continental monsoon climate in the region causes abundant summer precipitation, with significant interannual variations, while the topography features higher elevations in the northwest and lower elevations in the southeast, with significant variations in slope and terrain undulation. Abundant natural resources also characterize the region, encompassing diverse types of ecosystems, such as agricultural areas, agricultural–pastoral transition zones, forests, rivers, lakes, and deserts. This diversity results in a complex and dynamic land-use pattern. By 2022, the JJJM region, which has a population of approximately 130 million, reached CNY 123,159 trillion in GDP, accounting for 10.2% of China’s total economic output.

2.2. Data Processing

Land-use data, with a resolution of 1 km × 1 km at five-year intervals from 2000 to 2020, served as the foundation for ecosystem services assessment and PLUS model predictions. These data were sourced from the Chinese Academy of Sciences Resource and Environmental Data Cloud Center (http://www.resdc.cn/DOI, accessed on 1 December 2023). To align with the land-use matrix, the data were reclassified into seven types of land use, namely, cropland, forest, grassland, water, wetland, impervious, and barren. Afterward, Python 3.9 was utilized with Cartopy, Numpy, and other tools for mapping [30,31,32], and the Normalized Difference Vegetation Index (NDVI) was employed to adjust the supply of ecosystem services after considering the spatial and temporal resolution, along with regional characteristics. MOD13A3 (https://lpdaac.usgs.gov, accessed on 1 December 2023) was selected, and its maximum and mean values were synthesized into an annual scale. For a more accurate simulation of future land use within the JJJM region, three natural factors (DEM, slope, and surface undulation) and eight anthropogenic factors (GDP, electricity consumption, grazing density, population, CO₂ emissions, distance to roads, distance to water bodies, and nighttime lights) were simultaneously considered, which served as expansion factors in the PLUS model. All data in this study had a 1 km resolution, were processed on an annual scale, and were uniformly projected in WGS 84.

2.3. Quantitative Assessment of ESs Balance

Burkhard et al. developed a semi-expert matrix that associates land use with ESs, and it encompasses provisioning, regulating, and cultural services [12]. Unlike most methods of quantifying ecosystem services that focus solely on the supply of these services, this matrix enables the simple and rapid quantification of the supply, demand, and balance of ESs, even in data-scarce situations, while producing reliable results, which has led to its wide application [33,34,35,36]. While some researchers have raised questions about its reliability, existing comparative studies indicate that its results for ecosystem services (ESs) supply exhibit spatial patterns similar to those derived from biomass methods [37,38]. Moreover, this simpler mapping relationship is more suitable for informing decision making [39]. Through the selection of studies conducted in the Chinese region, this study refined the classification of secondary land use into primary categories for the evaluation matrix, thereby adapting it to the scoring matrix [15] (Figure 2). At the same time, after considering the variation in the capacities for supplying ESs within the same land-use unit, as well as the extensive implementation of grassland cover and various ecological projects in the JJJM region, NDVI was introduced to adjust the supply of ESs, thereby refining the spatial distribution of the ESs balance [40,41].

The area of each type of land use within the area under study was multiplied by the corresponding quantitative values of ESs as follows to calculate both the supply and demand of ESs:

$$\begin{array}{c}E{S}_j={\displaystyle \sum _{i=1}^7}{S}_i\times V_{ESi}\left(i=1,2,3\dots 7\right)\end{array}$$

(1)

where $V_{ESi}$ represents the quantitative value of the ecosystem for land-use type $i$, $S_i$ is the area of land-use type $i$, and $E{S}_j$ denotes the ecosystem services value for ESs category $j$.

NDVI was thus employed to adjust the ESs supply within the JJJM region during the research period:

$$\begin{array}{c}{ES}_s=F_{nj}\times A_i\end{array}$$

(2)

$$\begin{array}{c}{ES}_d=F_{nj}\times 1\end{array}$$

(3)

where $E{S}_s$ represents the adjusted value of ESs supply; $E{S}_d$ represents the value of ESs demand; $F_{nj}$ is the initial value of ESs $j$; and $A_i$ is the adjustment coefficient, defined as the ratio of the annual maximum NDVI to the annual average NDVI for each grid. Moreover,

$$\begin{array}{c}E{S}_b={ES}_s-{ES}_d\end{array}$$

(4)

where $F_s$ is the adjusted value of ESs supply, $E{S}_{de}$ is the demand value of ESs, and $E{S}_b$ is the ecosystem services balance adjusted by NDVI.

Simultaneously, we employed a partial correlation analysis to investigate the potential influencing factors of the heterogeneity in ESs balance distribution.

$$\begin{array}{c}{\rho }_{XY\cdot ZW}=\frac{{\rho }_{XY\cdot Z} - {\rho }_{XW\cdot Z}{\rho }_{YW\cdot Z}}{\sqrt{(1 - {\rho }_{XW\cdot Z}^2)(1 - {\rho }_{YW\cdot Z}^2)}}\end{array}$$

(5)

where ${\rho }_{XY\cdot Z}$, ${\rho }_{XW\cdot Z}$, and ${\rho }_{YW\cdot Z}$ are the partial correlation coefficients between $X$ and $Y$, $X$ and $W$, and $Y$ and $W$, controlling for $Z$, respectively.

2.4. Future Land-Use Simulation

The PLUS model, which was employed in the present study to simulate future land-use scenarios in the JJJM region, is an innovative land-use simulation model based on the cellular automata (CA) model developed by Liang et al. can better explore the causal factors of various types of land-use changes, and performs better in the change of patches of natural land-use types [42]. This model facilitates a higher-level exploration of the causes behind various land-use changes and simulates these changes at the patch level. It primarily comprises the land expansion analysis strategy (LEAS) module, which utilizes the random forest algorithm and the CA model based on a variety of random patch seeds.

The LEAS module is a rule-mining method based on a land expansion analysis, and it extracts parts of different types of land-use expansion between two periods, samples these parts, and employs the random forest algorithm to mine the expansion of various land-use scenarios and driving factors. This process yields the development probability of various types of land use and the contribution of a specific driving factor to the expansion of each type of land use during that period [43].

Random forest classification (RFC) is an ensemble classifier based on decision trees, and it can handle high-dimensional data and multiple collinear variables. Ultimately, it outputs the growth probability $P_{i,k}^d\left(x\right)$ for land-use type $k$ on spatial unit $i$; thus,

$$\begin{array}{c}{P}_{i,k}^d\left(x\right)=\frac{{\Sigma }_{n=1}^{m}I\left(h_n\left(x\right)=d\right)}{M}\end{array}$$

(6)

where $d$ takes values of 0 or 1, 1 indicates a transition from other types of land use to land-use type $k$, 0 represents other transition modes, $x$ is a vector composed of multiple driving factors, $I(\ast )$ is the indicator function of the decision tree geometry, $h_n\left(x\right)$ is the predicted type of vector $x$ by the $n$-th decision tree, and $m$ is the total number of decision trees.

The CARS module is an improved CA model that, by combining random seed generation and a threshold decrease mechanism, builds upon the development probability generated by the LESA module. By achieving a multi-type random patch-seeding mechanism based on a descending threshold, it dynamically simulates the self-generation of patches in both space and time. When the neighborhood impact of land-use type $k$ is equal to 0, this mechanism uses the Monte Carlo method in the following manner to generate seeds of change on each land-use type growth probability surface $P_{i,k}^{d=1}$ [42]:

$$\begin{array}{c}O{P}_{i,k}^{d=1,t}=\left\{\begin{array}{cc}{P}_{i,k}^{d=1}\times \left(r\times {\mu }_k\right)\times D_k^t& if {\mathsf{\Omega }}_{i,k}^t=0 \mathrm{and} r<P_{i,k}^{d=1}\\ P_{i,k}^{d=1}\times {\mathsf{\Omega }}_{i,k}^t\times D_k^t& all others\end{array}\right.\end{array}$$

(7)

where $r$ is a random value ranging from 0 to 1, and ${\mu }_k$ is the threshold for generating new patches of land-use type $k$. Moreover,

$$\begin{array}{c}If {\Sigma }_{k=1}^N\left|G_c^{t-1}\right|-{\Sigma }_{k=1}^N\left|G_c^t\right|<Step Then,l=l+1\end{array}$$

(8)

$$\begin{array}{c}\left\{\begin{array}{c}Change P_{i,c}^{d=1}>\tau \mathrm{and} T{M}_{k,c}=1\\ No change P_{i,c}^{d=1}\le \tau \mathrm{or} T{M}_{k,c}=0\end{array}\right.\tau ={\delta }^l\times rl\end{array}$$

(9)

where $Step$ represents the step size of the PLUS model, which is used to approach the future land-use target demand; $\delta$ is the decay factor of decreasing threshold $\tau$, ranging from 0 to 1 and set according to different scenario requirements; rl is a normally distributed random value with an average of 1, ranging from 0 to 2; and $l$ is the decay order. Furthermore, $T{M}_{k,c}$ is the transformation matrix that defines whether land-use type $k$ is allowed to convert to type $c$.

As

Table 1. Transition matrix and neighborhood weight of different types of land use.

	Cropland	Forest	Grassland	Water	Wetland	Impervious	Barren
Cropland	0.2077	Y	Y	N	N	Y	Y
Forest	N	0.1740	Y	N	N	N	Y
Grassland	Y	Y	0.3454	N	N	Y	Y
Water	Y	Y	Y	0.0261	Y	N	Y
Wetland	Y	Y	Y	Y	0.0470	N	Y
Impervious	N	N	Y	N	N	0.0940	N
Barren	Y	Y	Y	N	N	Y	0.1058

shows, the land-use conversion matrix for the study area was based on existing research [44,45]. Combined with random seed generation and the threshold decrease mechanism, the PLUS model can simulate the self-generation of patches under the constraint of development probability.

3. Results

3.1. Historical, Temporal, and Spatial Analyses of the ESs Balance in the JJJM Region

Applying the land use matrix method, we adjusted the supply of ES by incorporating NDVI. This approach facilitated the quantification of both the supply and demand dynamics for three key ESs categories–provisioning, regulating, and cultural–along with the overall ESs, spanning the period from 2000 to 2020. Cartopy in Python 3.9 was used for mapping and assessing the spatiotemporal distribution of four types of ESs balance. During the study period, the ESs balance in the JJJM region and Inner Mongolia showed an initial rise, followed by a decline. The JJJ region in particular exhibited varying levels of decline, with that of Tianjin being the most severe (Figure 3). Beijing, Hebei, and Inner Mongolia exhibited higher provisioning services balance, with mean values of 18.16, 17.92, and 20.24, respectively. In contrast, Tianjin demonstrated a significantly lower provisioning services balance, with a mean value of 7.25. Further, the mean values of regulating services balance for Beijing, Hebei, Inner Mongolia, and Tianjin were 26.15, 27.93, 27.78, 26.68, and 27.85, respectively. It should be noted that Tianjin’s regulating services balance has been consistently negative since 2000, having only decreased year by year. Additionally, the cultural services balance for each province was relatively high, ranging from 16.69 to 17.67. As for the total ESs balance in the JJJM region, the mean values were 64.71, 68.96, 68.45, 65.92, and 68.15, respectively, indicating a decline in 2010 that was followed by an overall upward trend.

Areas with a higher provisioning services balance are mainly distributed in the central grasslands and northern forests of Inner Mongolia, along with the northern part of the JJJ region (Figure 4). Cities with a high provisioning services balance include Hulun Buir, Hinggan, Tongliao, Xilingol, and Ulanqab, located in the northern part of the JJJM region. After calculating the slope of the provisioning services balance for the JJJ region in each grid, the areas with a decreasing provisioning services balance were primarily concentrated in the central and southern parts, the northern and western parts of Inner Mongolia, and the city of Ordos in the west. The decreasing trend was observed to be more pronounced in Beijing, Tianjin, and Hebei, with Tianjin experiencing the most drastic decline, reaching a slope of −0.60 and decreasing to 2.92 in 2020, which is only 30% of the level observed in 2000.

Regarding the regulating services balance, apart from Inner Mongolia, which exhibited an upward trend, other provinces were found to have experienced a decline (Figure 5). Inner Mongolia consistently maintained the highest level of regulating services balance, with an average of 29.73. Although Beijing maintained a relatively high average of 25.02 for regulating services balance, it had a pronounced downward trend with a slope of −0.43. Conversely, Tianjin reached a deficit in regulating services balance, and its steep decline with a slope of −0.33 is noteworthy. The central and southern parts of the JJJ region generally exhibited a lower regulating services balance, while the sandy areas in the western part of Inner Mongolia maintained higher levels than the JJJ region. Specifically, cities with a high regulating services balance included Hulun Buir, Hinggan, Tongliao, and Xilingol, while cities with a low regulating services balance were mainly located in the central and southern parts of the JJJ region. These areas in Inner Mongolia, including Hulun Buir, Ordos, and Bayan Nur, demonstrated a more pronounced declining trend in regulating services balance located in the western part of the JJJM region.

The cultural services balance in the JJJM region was generally high, showing a rising trend overall (Figure 6). Between 2000 and 2015, the cultural services balance among the provinces and cities exhibited minimal variation, with the mean values staying around 17; however, after 2015, the differences in cultural services balance increased. Though the number of areas with an upward trend in cultural services balance in the JJJM region increased, the overall mean values showed a downward trend, particularly in Beijing and Tianjin, both of which experienced a rapid decline with a slope of −0.24. Additionally, Tianjin’s cultural services balance dropped to 12.91. Generally, the cultural services balance in the JJJM region remained relatively high, with lower values being observed mainly in the central and southern parts of the JJJ region and in Alxa near the western part of Inner Mongolia.

Figure 7 illustrates the temporal evolution of the total ESs balance in the JJJM region. The balance initially rose, followed by a slight decline, maintaining an overall increase with a slope of 0.08 year⁻¹. Inner Mongolia consistently had the highest total ESs balance among the four provinces, which ultimately reached 69.43. Beijing also had a relatively high total ESs balance, with a mean value of 63.87, which was second only to that of Inner Mongolia. Further, Hebei’s mean value was found to be 57.69, and Tianjin was found to have the lowest mean value at 26.17. Spatially, there were noticeable differences in the total ESs balance in the JJJM region. Specifically, the regions with the highest total ESs balance were in the northeastern part of Inner Mongolia. The relatively higher areas were primarily distributed in the northern part of JJJM, as well as in the central and eastern parts of Inner Mongolia, whereas the lower areas were distributed in the western sandy areas of Inner Mongolia, as well as in the central and southern parts of JJJM. The deficit areas in the JJJM region demonstrated an expanding trend, with deficit areas appearing in the central districts of Beijing, Tianjin, Baoding, Shijiazhuang, and Hohhot, located in the south central part of the JJJM region. Moreover, many deficit areas were also found in the coastal regions of Cangzhou and Tangshan in the eastern part of the JJJ region. The total ESs balance in the three provinces of Beijing, Tianjin, and Hebei demonstrated a declining trend, with Beijing and Tianjin being the most severely affected, with slopes of −2.23 and −2.26, respectively. Although Hebei’s declining trend was relatively weaker, it still reached −1.05. Inner Mongolia’s total ESs balance demonstrated a weak increasing trend, with a slope of 0.08.

A correlation analysis was conducted on the balance of the three types of ESs, and it revealed a positive correlation among them throughout the study period, indicating a significant and increasingly strengthened synergistic relationship. In particular, the synergy between the balance of provisioning services and regulating services was found to be the most robust, with the correlation coefficient increasing from 0.88 in 2000 to 0.95 in 2020. Conversely, the synergistic relationship between cultural services and regulating services was found to be relatively weaker, with the correlation coefficient rising from 0.84 in 2000 to 0.87 in 2020, albeit with a slightly lower increase. In 2020, the balances of various services were found to be more concentrated than in 2000, with a reduction in both high and low values for regulating and cultural services balances, whereas the balance of provisioning services showed a relative decline. Figure 8 illustrates this trend.

3.2. Future Land-Use Trends of the JJJM Region in 2040

By leveraging the PLUS model, we employed the land-use data from 2010 and 2020 to extract patterns of land-use expansion. For this, we selected 12 categories of driving factors and utilized the LEAS module to integrate these factors and generate a map depicting the development potential of land use. Afterward, this potential map was input into the CARS module, wherein a Markov chain-based approach facilitated the estimation of the quantity of each type of land use for 2020 and 2040. The simulation results for 2020 were achieved through parameter tuning, which yielded a satisfactory Kappa coefficient of 0.656, indicating high simulation accuracy. The chosen driving factors and parameter settings met the simulation requirements. Further, the FoM index that was used to measure the error stood at 0.016, which, being below 0.2, fell within the acceptable range. Subsequently, the previously adjusted PLUS model was applied to simulate the specific quantities and spatial distributions of the types of land use in the JJJM region for 2040 (Figure 9).

In 2020, grasslands accounted for 41.44% of the total land use in the JJJM region and were mainly concentrated in the central areas of Inner Mongolia, whereas croplands accounted for 15.33% of the land use and were primarily distributed in the central and southern parts of the JJJ region, falling within the agricultural regions in North China. Arable land was found to be present in the agro-pastoral transition zones in the central and southern parts of Inner Mongolia, and forests were predominantly located in the northeastern part of Inner Mongolia, specifically in the Greater Khingan Mountains area and in the central-northern part of the JJJ region, where various ecological projects have been implemented. Additionally, water bodies and wetlands accounted for 0.81% and 2.17% of the land use, respectively. Furthermore, impervious surfaces were mainly distributed in the JJJ region and city centers. Notably, in the central and southern parts of the JJJ region, there is a widespread interspersion with arable land, which accounted for 2.71% of the land use. Barren land was predominantly found in the western part of Inner Mongolia, specifically in the desert area of Alashan, which accounted for 2.09% of the land use. The simulation results for 2040 revealed that the proportion of impervious surfaces will reach 3.01%, which is primarily attributed to the increased urban areas in Inner Mongolia and the increase in cropland in the agricultural regions of the JJJ region. The proportion of grassland is expected to increase to 41.44%, whereas the area of arable land is expected to slightly decrease by 0.4%.

3.3. Quantification of the Future ESs Balance of the JJJM Region

Based on the simulated land-use scenarios in 2040, the land-use matrix model was used to quantify the balance of provisioning, regulating, cultural, and total ESs, which was adjusted by the mean adjustment factors from 2000 to 2020 (Figure 10). The ESs balance in the JJJM region continued to exhibit a pattern of high vegetation cover and high balance. In the JJJ region, all balances of ESs demonstrated an upward trend along with a downward trend in deficit areas. Mean values for Beijing, Tianjin, Hebei, and Inner Mongolia were 64.02, 27.14, 56.23, and 66.12, respectively. In comparison to 2020, the regional averages of Beijing, Tianjin, and Hebei increased by 11.69, 9.31, and 3.31, respectively. Moreover, for Beijing, Tianjin, and Hebei, the provisioning services balance increased by 3.41, 0.55, and 0.61; the regulating services balance increased by 1.72, 1.52, and 0.50; and the cultural services balances increased by 1.42, 1.53, and 0.67, respectively. In Inner Mongolia, all balances of ESs demonstrated a downward trend. The reduction in the provisioning services balance occurred in the central and northwestern regions, whereas the reduction in the regulating services balance occurred in the central-northern part of Inner Mongolia. The decrease in the cultural services balance was more substantial, with the entire northern region of Inner Mongolia experiencing a downward trend. Inner Mongolia also experienced significant urban expansion, particularly in the northeastern and central regions, expanding into some deficit areas, which included Hulun Buir, Xilingol, and Ordos. The total ESs balance in Inner Mongolia decreased by 2.29. Specifically, its provisioning services balance decreased by 0.71, the regulating services balance decreased by 0.95, and the cultural services balance decreased by 0.61.

4. Discussion

4.1. Spatial Dynamics of ESs Balance and Natural Factors

Analyzing the flow of ESs from nature to societies provides crucial insights for region-al environmental management, facilitating economic and social development [46]. Significant spatial variations were observed in provisioning, regulating, cultural, and overall ESs balances across the JJJM region from 2000 to 2020.. Areas with a high provisioning and regulating services balance were found to be primarily concentrated in the eastern part of the Inner Mongolia Autonomous Region and the central-southern part of the JJJ region, where abundant forests and grasslands are distributed. Most areas exhibited a relatively high cultural services balance, except for the western sandy areas of Inner Mongolia and the central-southern part of the JJJ region. Further, the total balance of ESs continued to exhibit noticeable spatial variations, in a manner similar to the distribution of the balance of regulating and provisioning services. Additionally, in the northeastern part of Inner Mongolia, Hulun Buir, Hinggan, and Tongliao exhibited a higher total balance of ESs. By conducting a cold–hot spot analysis using ArcGIS, it could be confirmed that these three cities were hot spots at the 99% confidence level. Conversely, the entire JJJ region fell within the category of cold spots, primarily due to extensive urban construction, which resulted in a lower balance of ESs [47]. Past studies indicated that topography, vegetation biomass, and climatic factors influence the supply capacity of ESs [24,48,49]. We conducted a partial correlation analysis on the total ESs balance and temperature, precipitation, and DEM using the pcorr function from the pandas library [50]. The results revealed a positive correlation between precipitation and elevation with the total ESs balance, with partial correlation coefficients of 0.50 and 0.66, respectively. In contrast, temperature exhibited a negative correlation, with a partial correlation coefficient of −0.67 (Figure 11). This indicates that the total balance of ESs is largely influenced by hydrothermal conditions. While an increase in precipitation causes more vigorous vegetation growth, elevated temperatures often negatively impact plant growth and development [51]. Low-altitude (<966 m) areas of the JJJM region predominantly correspond to the semi-humid southeastern urban areas and northeastern areas of Inner Mongolia (45°–50° N), whereas high-altitude (966–3249 m) areas are primarily concentrated in the arid western regions of Inner Mongolia (37°–42° N). These complex natural conditions also influence the current distribution pattern of the total balance of ESs in the JJJM region. Therefore, in the process of urbanization, industrial transfer, and the formulation and implementation of ecological projects, it is crucial to fully consider the natural conditions of the JJJM region, adopt a tailored approach, and avoid encroachment on land with a high potential for ESs supply.

4.2. Shaping ESs Balances through Land-Use Changes

The overall ESs balance during the research period exhibited a steady upward trend, with minor fluctuations observed between 2005 and 2010. After 2012, the Chinese government emphasized the importance of prioritizing ecological civilization by proposing key focuses on optimizing land spatial patterns, promoting resource conservation, intensifying natural ecosystem protection, and strengthening the development of ecological civilization systems [52]. The positive impact of these steps are evident in our results.

At the provincial level, Inner Mongolia Autonomous Region, Hebei, and Beijing exhibited relatively high mean values for the balance of provisioning, regulating, cultural services, and total ESs with minimal differences. However, the JJJ region demonstrated a downward trend in all balances of ESs, with Beijing and Tianjin experiencing the most severe decline. The mean level in Beijing in 2000 exceeded that of Inner Mongolia, possibly due to the significant development of the Three-North Shelter Forest Project initiated in 1987. This project, which received an investment of CNY 57.68 billion, has gained international recognition and acclaim [53]. But the level subsequently declined, with a slope of −2.23 for the overall ESs balance, and the decline was more pronounced in the balance of provisioning and regulating services, with slopes of −0.24 and −0.43, respectively. This trend may be attributed to Beijing’s status as China’s political and economic center, as it features robust urban planning and abundant green spaces in the northern region. However, rapid economic development, dense population influx, and accelerated urbanization have led to a rapid increase in the demand for ESs, which has resulted in a reduction in the balance of ESs at the urban level [47]. The results from the PLUS model simulation indicate that the process of urbanization will continue, leading to a total impervious surface area of 26,365 km² in the JJJM region, which will pose a significant challenge to the balance of ESs. Even so, the downward trend in cultural services balance is less pronounced, with a slope of −0.24. This aligns with the findings of previous research, suggesting that urban green spaces and water bodies contribute to certain cultural services.

Moreover, even with Beijing experiencing rapid urbanization, the city sustains a con-sistent proportion of urban green spaces and water bodies. This phenomenon may eluci-date its comparatively slower decline in cultural services [54]. As the largest northern port city in China, Tianjin boasts a water and wetland area of 1274 km², while the northern urban forest and grassland cover only 511 km². The impervious surface area of Tianjin increased from 1213 km² in 2000 to 1999 km² in 2020, which accounted for 26.42% (Figure 12). This urban pattern has led to significant reduction in the balances of provisioning and regulating services in Tianjin compared to in other cities. The balance of regulating services has even reached a deficit, although the balance of cultural services remains relatively favorable [55]. The current ecosystem services projects in Tianjin, such as The Sandification Control Program for Areas in the Vicinity of Beijing and Tianjin, are primarily implemented in areas surrounding Tianjin but are inadequate in addressing the deficit within the city [54]. We suggest that future efforts should concentrate more on protecting and developing Tianjin’s high-yielding ecosystem services farmlands and urban green spaces. Compared to Beijing and Tianjin, the decline in various ESs balances in Hebei is less pronounced, which can be attributed to the establishment of development systems such as urbanization and integration in the JJJ region [56]. Hebei possesses a significant amount of cropland and, thus, shoulders the responsibility of relocating non-core capital functions, as well as undertaking related tasks for industrial upgrading. Therefore, it is imperative to pay attention to the ESs balance in Hebei to expand environmental capacity and ecological space.

The ecosystems in Inner Mongolia are relatively fragile, yet the balance of various ESs has been consistently increasing, with a slope of 0.31 for the total ESs balance. The primary contributors to this increase are the balances of cultural services and provisioning services, with respective slopes of 0.04 and 0.03. This can be attributed to the extensive implementation of a series of nationwide ecological projects by the Chinese government from 2000 to 2020, which have been widely executed in Inner Mongolia [57]. These projects include the National Forest Protection Project, the Beijing–Tianjin Sand-storm Source Control Project, the 3-North Shelter Forest Program, the Grain for Green Project, and the Returning Grazing Land to Grassland Project [58]. In regions such as Xilingo, Ulanqab, Bayan Nur, and Baotou, the implementation of multiple ecological projects has led to noticeable ecological improvement. Despite rapid urbanization and industrial development, Inner Mongolia’s extensive green areas have largely met ES demand, except in declining areas like Ordos and Bayan Nur. PLUS simulations show that the increase in impervious surfaces is focused in Inner Mongolia’s urban centers, projecting an increase to 111.52 km² by 2040, up by 3232 km² since 2000. Additionally, the total balance of ESs in Inner Mongolia is expected to decrease by 2.29, a shift that may result in heightened future demand for ESs, underscoring the ongoing importance of prioritizing ESs in Inner Mongolia. Meanwhile, under this simulated spatial pattern shift, the balance of ESs in the JJJ region increased, which indicates that policies involving urban development and industrial relocation in Inner Mongolia, which may shift impervious surface areas, can lead to a significant improvement in the balance of ESs in the JJJ region, albeit at the expense of a relatively smaller loss in the balance of ESs in Inner Mongolia. This finding underscores the continued importance of prioritizing the balance of ESs in Inner Mongolia during the development process to prevent environmental degradation and reduce social inequality.

4.3. Insights, Limitations, and Future Directions

This study utilized an improved land-use matrix and adjusted ESs supply by introducing NDVI. It quantified the balances of provisioning, regulating, cultural, and total ESs in the JJJM region from 2000 to 2020. This analysis involved an examination of their correlations with complex thermal-hydro conditions and the association with land-use change patterns. Further, we used the PLUS model to simulate the land-use situation in 2040 and the estimated future ESs balance. However, it is important to acknowledge the limitations of this study. First, concerning provisioning, regulating, and cultural services, some sub-services may exhibit a negative correlation with vegetation growth, such as freshwater services, highlighting the need for nuanced analysis [59]. Second, although the analysis of the spatial distribution and trends of the ESs balance in the JJJM region used a land-use matrix model that was successfully applied in the Chinese context, the model primarily relies on local expert knowledge; therefore, future improvements could involve incorporating the perspectives of more stakeholders to optimize the assessment matrix. Finally, while the PLUS model can simulate future land use based on historical data, the selection of driving factors in this study is still constrained by data availability, such as specific industrial and mining sites, tourism data, and calculations of grazing density, leaving considerable room for further research. Additionally, the Markov chain cannot predict the quantity of future land use under various development scenarios; therefore, coupling Shared Socioeconomic Pathways (SSPs) with the PLUS model for multi-scenario predictions could be considered for more accurate estimations of the ecosystem services balance under different scenarios.

5. Conclusions

This study incorporated the ecological index NDVI to adjust the ESs matrix, providing a reference approach for addressing the heterogeneity issue in quantifying ecosystem services based on the equivalence factor method. Our study revealed notable spatial disparities in the ESs balance in the JJJM region, with a synergistic effect among the provisioning, regulating, and cultural services balances. Areas with extensive forests and grasslands, such as Hulun Buir, Hinggan, Tongliao, and Xilingo, were found to generally exhibit higher balances of ESs, whereas urbanized areas such as Beijing and Tianjin showed relatively lower balances of ESs (except for cultural services balance), with some areas experiencing services deficits. This distribution pattern is influenced by regional natural conditions, where increased altitude and precipitation positively impact the balance of ESs, with partial correlation coefficients of 0.66 and 0.50, respectively, and temperature showing a negative correlation coefficient of −0.19. This indicates that areas with suitable thermal-hydro conditions generally have a higher ESs balance. Despite fluctuations and declines in the ESs balance in the JJJM region from 2005 to 2010, an overall upward trend was observed during the entire study period, with a slope of 0.08. This was particularly conspicuous in the case of Inner Mongolia, with a slope of 0.31. However, Beijing, Tianjin, and Hebei exhibited downward trends. Tianjin was found to not only have the lowest ESs balance but also experience the fastest decline. The PLUS simulation results indicate an augmentation in the impervious surface area, with the percentage projected to reach 3.01% by 2040. Further, the variations in the total ESs balance from 2020 to 2040 in Beijing, Tianjin, Hebei, and Inner Mongolia were found to be 11.69, 9.31, 3.31, and −2.21, respectively. This suggests that implementing industrial transfer and urbanization in Inner Mongolia can improve the Beijing–Tianjin–Hebei region substantially without excessively compromising its balance of ESs. Thus, based on these findings, it is recommended that the development strategy for the JJJM region should focus on the impact of urbanization.