Ecosystem Services Response to the Grain-for-Green Program and Urban Development in a Typical Karstland of Southwest China over a 20-Year Period

Abstract

Southwest China is an ecologically fragile area with a high proportion of karstland and large variations in terrain, and it can be greatly affected by land use change. From 2000 to 2020, with the acceleration of urbanization in the whole country, the macro-scale Grain-for-Green Program (GFGP) has been developed in the karstland of southwest China. This has resulted in the expansion of forest and construction land with a reduction in cultivated land. The response of ecosystem services (ESs) to these changes needs to be investigated. However, there is a lack of in-depth analysis of the karstland of southwest China, and current studies mostly focus on the spatio-temporal variation in individual ESs or how the whole land use change affects ESs. Thus, our study uses an InVEST model and geographically and temporally weighted regression (GTWR) to examine the spatio-temporal variation in land use associated with four ESs, namely water conservancy (WC), soil conservancy (SC), carbon sequestration (CS), and habitat quality (HQ). We found that the GFGP area first increased and then decreased, aggregating to 4.48 × 10⁴ km² from 2000 to 2020. We also observed that from 2000 to 2020, ESs remained stable or gradually increased (despite fluctuations), SC was the most stable, whereas WC, CS, and HQ first decreased and then fluctuated more. This may be related to the destruction of topsoil in the early stages of the GFGP and the relatively weak ES supply capacity of the young trees. Moreover, the response of ecosystem services to the GFGP was spatially heterogeneous, suggesting a strong driving influence of the local environment, mainly caused by the distribution of karstland and terrain, differentiated urbanization levels, and the intensity of the GFGP. Specifically, the relatively significantly positive effects of WC, SC, and CS were found in western and northern cities, and so was the HQ in southeastern cities. On the contrary, the negative impacts of urbanization were found to weaken over time, suggesting the trade-off effect of the GFGP. Our findings would contribute to the development of effective forest management strategies and provide valuable insights for policymakers and stakeholders involved in ecosystem restoration and conservation efforts by exploring the impacts of the GFGP and urbanization on ESs.

1. Introduction

Ecosystem services (ESs) refer to the natural conditions and benefits provided by ecosystems and eco-processes to people. These can include the provision of food and water, shelter from climatic events, and even the provision of important cultural values associated with the local environment [1,2]. Broadly, ESs can be further split into ecosystem products and functions. Ecosystem products provide humans with food, raw materials, and energy for production and life. Ecosystem functions provide a guarantee for the quality of the human living environment by conserving water and soil. For example, water conservancy (WC) functions include regulating runoff, increasing available water resources, and improving water quality [3]; soil conservation (SC) functions are closely related to human food production and health [4]; habitat quality (HQ) affects biodiversity and environmental quality; and carbon sequestration (CS) affects human production and life by influencing soil fertility and carbon pools in both soil and the atmosphere. The practice of ‘valuing’ ESs has been widely addressed and discussed by the scientific community due to its direct impact on human development interests [5,6,7,8].

As a part of the natural and human social system, ESs are constantly in a state of change under the influence of environmental factors and human activities [9,10]. Natural factors such as global climate change (temperature, precipitation), terrain slope, and vegetation cover directly affect it. Mutual constraints and synergies among changing ES elements also play a role [10,11,12]. Furthermore, human activity has complex and profound effects on the provision of ESs. Land use change directly affects terrain and topography, and indirectly affects ESs by changing the local climate [8,13]. Present studies mainly show that human activities such as urbanization and overgrazing are harmful to ESs, while protective measures such as ecological programs and land planning are ecologically conducive to recovery [14,15,16]. Nevertheless, it is controversial how land use variation affects ESs, considering the differences in nature and societal situations, as well as the trade-off relationship between ecological construction and constructive expansion. For example, Wang et al. found that the GFGP has improved ESs and restored ecology in the basin by increasing the area of forests and grasslands [17]. Meanwhile, Qian et al. pointed out that in the absence of landscape pattern planning, simply expanding green areas is not sufficient to reduce the risk of ecological degradation, because a relatively high fragmentation of green space weakened its ecological maintenance function [18]. Through multiscale analysis, some researchers pointed out that due to national environmental protection policies, urbanization of the Beijing–Tianjin–Hebei urban agglomeration is not negative for ESs [19]. Ouyang et al. concluded that different regions have prominent ESs, and the spatial interaction between urbanization and ESs depends on the type of ESs chosen [20]. Considering the spatial heterogeneity of the variation in ESs and its driving factors, a thorough analysis should be conducted based on a specific region.

Typical karstland in China has a fragile and sensitive ecological environment due to limestone bedrock, large undulating terrain, and thin soil layers [21]. This landscape predominates across southwest China, where inappropriate land use and resource exploitation have caused ecological problems like soil erosion and rock desertification. To protect and restore the environment, China launched the Grain-for-Green Program (GFGP) in 1999 [22]. The GFGP is the world’s largest ecological restoration project in terms of investment and scale. As an important measure to enhance land utilization structure, restore the ecological environment, and achieve balanced progress, the GFGP aimed to increase the green belt area and reduce soil erosion by converting cultivated land into forest and grassland [23]. Among all regions, the southwest karstland has been the focus of China’s 20-year GFGP, with a prominent phase achieved by 2020. The afforestation rate of the area is the highest in China, covering a remarkable 64.92% of forest and 1.01% of grassland. This has resulted in significant ecological advantages for the local area through the preservation of natural forests (conifers, beech, Myrtaceae, etc.), the implementation of artificial afforestation (pines, cypresses, moso bamboo, etc.), and the establishment of economic forests (citrus, prickly ash, bamboo, camellia, pomegranate, etc.). These efforts contribute to the enhancement of ESs and sustainable land management in the region. Wang et al. discovered that implementation of the GFGP increased soil conservation and reduced water production in karstland [24]. Similarly, Hu et al. found that the GFGP has significant positive contributions to HQ, CS, and SC, while obviously reducing water yield [25]. This indicates that the southwest karstland presents a unique combination of ecological and economic development contradictions, which makes it an intriguing case study for understanding the impacts of the GFGP and urbanization on ESs. Therefore, studying the impacts of the GFGP and urbanization in this context has both academic and practical significance, providing insights that can inform sustainable development strategies in similar karst areas worldwide.

At the same time, with rapid economic development, there has been an expansion of construction land within the study area, becoming another critical factor that affects local ESs. According to early research findings, urbanization destroys natural surfaces, resulting in reduced vegetation and soil erosion, and hardened surfaces reduce water and carbon storage, which may consequently disturb ecology systems. Presumably, the damage caused by increasing development has been mitigated to some extent by the influence of the GFGP on ESs. However, despite the substantial land use change across the karst landscape of southwestern China [26], there remains a lack of research on how ecological engineering and urbanization affect ESs in this area. Most of the existing research focuses on the isolated response of a single ES function to the GFGP [27,28], thereby lacking a comprehensive understanding of the integrated effects of the GFGP on ESs within the context of urbanization.

Due to the contradictions of ecological–economic development and its distinctive natural characteristics, which highlight the vulnerability of the region to land use changes and collectively contribute to the ESs, we selected a typical karstland in China as the study area. Based on land use data, meteorological data, soil property data, and others, we analyzed the spatio-temporal change in areas protected by the GFGP and areas that have been developed during urbanization. Then, the response of ESs to the GFGP and urbanization was quantified using the InVEST model and geographically and temporally weighted regression (GTWR) method. In addition, the objectives of this study were to (1) quantitatively analyze the spatiotemporal changes in the GFGP and urbanization in a representative karstland in Southwest China using the Markov model and land use dynamics model; (2) investigate the variations in ESs, specifically soil conservation, water conservation, carbon storage, and habitat quality via InVEST; and (3) explore the impacts of the GFGP and urbanization on ESs. Ultimately, the findings of this study will contribute to the development of effective forest management strategies and provide valuable insights for policymakers and stakeholders involved in ecosystem restoration and conservation efforts.

2. Study Area

The Xijiang River Basin (21°36′ N−27° N, 102°16′ E−113°23′ E) and the Wujiang River Basin (22°07′ N−30°22′ N, 104°18′ E−109°22′ E) were selected in this study as the karstlands representative of southwestern China (Figure 1). The Xijiang River is the longest river in the Pearl River system. It has a total length of 2000 km and a basin area of about 350,000 km², accounting for about 66% of the Pearl River basin. It is located in the subtropical to tropical monsoon climate zone. The average annual temperature falls within the range of 14–22 °C, while the annual rainfall varies between 1000 and 2200 mm. Precipitation is unevenly distributed throughout the year. Eighty percent of the precipitation falls between April and October [29]. The Wujiang River watershed is the largest tributary of the southern headwaters of the Yangtze River. It has a humid subtropical monsoon climate. The average annual rainfall is 850–1600 mm. Across the basin, rainfall decreases first from the east to the west and (within the longitudinal decrease) from the south to the north [30]. The karstland coverage of the two watersheds is 43.5% and 77.2%, respectively. Elevation decreases from west to east. The terrain is very hilly. The underlying karstland includes the Yunnan–Guizhou Plateau and the Guangxi Hills. It is located near the Tropic of Cancer and has a typical humid and rainy subtropical monsoon climate zone. Therefore, the natural vegetation types include coniferous forests (pine, fir, spruce, etc.), broad-leaved forests (nanmu, teak, oak, Dalbergia, etc.), and shrub forests (Camellia, masson pine, Osmanthus, etc.) [31,32].

3. Method and Material

Our study data include land use (LUCC), digital elevation model (DEM), soil property, and precipitation (

Table 1. Data sources of the study.

Data Type	Data	Source	Resolution
LUCC	Annual land cover product of China (CLCD)	Earth System Science Data [33]	30 m
DEM	ASTER GDEM	Geospatial Data Cloud site [34]	30 m
Soil property	Harmonized World Soil Database v 1.2	FAO SOILS PORTAL [35]	30 arcsecond
Precipitation	China ground meteorological data	National Earth System Science Data Center [36]	1000 m

Note: All data above are open access.

). The source of parameters used in the remaining models is shown in the Supplementary Materials.

3.1. Ecosystem Service Evaluation

With the development of studies on the correlation between land use change and ESs, new methods of analysis have been developed and widely applied, including the InVEST model [37], which is a widely used model for assessing ES dynamics [38,39]. The model takes regional natural (such as land use, elevation, climate, and soil data, etc.) and socioeconomic data under current or future scenarios as input, and outputs the distribution status and evolution trend of ES functions under this scenario [40]. There are many indicators used to measure regional ESs (i.e., supply and support services) in the InVEST model, and the selection of indicators in existing studies varies, mainly influenced by the natural environmental characteristics of the region and the implementation of major projects [9,14]. Considering the natural features of our study area, such as large undulations, a thin soil layer, and fragile structure, as well as the prominent ecological issues such as soil erosion and rocky desertification, we selected four aspects, namely water conservation (WC), soil conservation (SC), carbon storage (CS), and habitat quality (HQ), to measure local ESs comprehensively. In order to evaluate WC, SC, CS, and HQ services in the study area, we used the established InVEST modules of water yield (WY), sediment delivery ratio (SDR), carbon storage and sequestration (CSS), and habitat quality (HQ). The following calculation processes all refer to the InVEST 3.13.0 User’s Guide [41].

3.1.1. Water Conservation

Water conservation is the regulation of water by the ecosystem, which can be improved through afforestation and the construction of water conservation reserves. The InVEST Water Yield model estimates the relative contributions of water from different parts of a landscape, offering insight into how changes in land use patterns influence annual surface water yield [42,43,44,45] (Tables S1 and S2). The calculation formulas for the WC of the InVEST WY model are as follows:

$$Y_x=\left(1-\frac{{AET}_x}{P_x}\right)\times P_x,$$

(1)

Then, the annual water yield ($Y_x$) is corrected using the water conservation calculation formula [46] to obtain the water conservation amount.

$$\mathrm{W}\mathrm{R}=\mathrm{M}\mathrm{i}\mathrm{n}\left(1,\frac{249}{Velocity}\right)\times \mathrm{M}\mathrm{i}\mathrm{n}\left(1,\frac{0.9\times TI}{3}\right)\times \mathrm{M}\mathrm{i}\mathrm{n}\left(1,\frac{K_{sat}}{300}\right)\times Y_x,$$

(2)

where $Y_x$ is annual water yield of grid unit x in the study area (mm); ${AET}_x$ is annual actual evapotranspiration of grid unit x (mm); $P_x$ is annual precipitation of grid unit $x$(mm);$\mathrm{W}\mathrm{R}$ is water conservation (mm); $Velocity$ is velocity factor; $TI$ is topographic index; and $K_{sat}$ is soil saturated hydraulic conductivity.

3.1.2. Soil Conservation

The InVEST Sediment Delivery Ratio (SDR) model is a spatially explicit model working at the spatial resolution of the input digital elevation model (DEM) raster. Based on DEM, it calculates soil erosion with precipitation erosion factors and soil erosion factors, slope length and slope, and cover and management factors [47,48] (Tables S3 and S4). Finally, the difference between potential in the absence of vegetation and actual erosion with land cover and management (RKLSCP) indicates the role of local factors in avoiding erosion. The calculation formulas for the SC of the InVEST SDR model are as follows:

$${SEDRET}_x={RKLS}_x-{USLE}_x,$$

(3)

$${RKLS}_x=R_x\times K_x\times {LS}_x,$$

(4)

$${USLE}_x=R_x\times K_x\times {LS}_x\times C_x\times P_x,$$

(5)

where ${SEDRET}_x$ is soil retention of grid unit x in the study area (ton, t); ${RKLS}_x \mathrm{a}\mathrm{n}\mathrm{d} {USLE}_x$ represent the potential soil erosion amount (t) and actual soil erosion amount (t) of grid unit $x$, respectively; $R_x$ is the rainfall erosion factor; $K_x$ is the soil erodibility factor; ${LS}_x$ is the slope and slope length factor; $C_x$ is the vegetation coverage factor; and $P_x$ is the soil and water conservation factor.

3.1.3. Carbon Storage

The InVEST Carbon Storage and Sequestration (CSS) model uses maps of land use along with stocks in four carbon pools (aboveground biomass, belowground biomass, soil, and dead organic matter) to estimate the amount of carbon currently stored in a landscape or the amount of carbon sequestered over time [49] (Table S5). The calculation formula for the CS of the InVEST CSS model is as follows.

$$C_{tot}=C_{above}+C_{below}+C_{soil}+C_{dead},$$

(6)

$\mathrm{w}\mathrm{h}\mathrm{e}\mathrm{r}\mathrm{e} C_{tot}$ is total C sequestration (t); $C_{above}$ is aboveground C sequestration (t); $C_{below}$ is underground C sequestration (t); $C_{soil}$ is soil C sequestration (t); and $C_{dead}$ is litter C sequestration (t).

3.1.4. Habitat Quality

Biodiversity is closely related to the production of ESs. Patterns in biodiversity are inherently spatial, and therefore can be estimated by analyzing land use and threats to species’ habitats [50]. The InVEST Habitat Quality (HQ) model estimates HQ by integrating the suitability of land types for habitats, evaluating the impact of distance and stress factors, and considering the sensitivity of each habitat type to these stress factors [51,52] (Tables S6–S8). The calculation formula for HQ of the InVEST HQ model is as follows.

$$Q_{xj}=H_j\times \left[1-\left(\frac{D_{xj}^z}{D_{xj}^2+k^z}\right)\right],$$

(7)

where $Q_{xj}$ is an index of habitat quality index referencing grid unit $x$ in land use type $j$. Values range between 0 and 1, with higher values close to 1 representing a better habitat quality and a more stable ecosystem. $H_j$ refers to the ecological suitability of land use types; $D_{xj}$ refers to the degree of habitat degradation in grid unit $x$ in land use type $j$; $k$ is the semi-saturation constant, which is half of max degradation degree; and $z$ is the normalization constant, the default value of the model.

3.2. Response of ESs to GFGP and Urbanization

The availability and demand of ESs are affected to varying degrees by natural factors and human activities. To better analyze the impact of the GFGP and urbanization on ESs, we used a scenario simulation [53] estimating the change in ESs after a single land use change to quantify the influence of human activities. Only the impacts of the GFGP and urbanization on ESs were considered. To exclude the effect of conversion of other land types, based on the broad definition of the GFGP, we extracted the cumulative change area of cultivated and unused land converted to forest and grassland as the independent variable of the GFGP. We extracted the cumulative change area of other land types converted to construction land as the independent variable of urbanization. The results for the GFGP, urbanization, and ESs were standardized to reduce the error caused by numerical differences. Finally, the spatial differences of the GFGP and urbanization and the response of ESs to these two types of land use changes in each city were investigated using the GTWR method [54] (Figure 2).

3.2.1. Scenario Simulation

Observations of water conservation (WC) provisioning ESs should be the same as those under the simulation scenario if the land use in the theoretical scenario did not change in the past. Thus, the difference between the WC in the simulated scenario and that in the actual scenario can be considered a quantitative estimate of the effect of land use/land cover change [53]. In this study, the simulated WC was calculated for the years 2010 and 2020. The actual WC was subtracted from the simulated WC.

$$∆W_l=W_O-W_s,$$

(8)

where $∆W_L$ is the impact of land use; $W_s$ is the simulated WC; and $W_O$ is the actual WC.

3.2.2. GTWR

As a local regression model, GWR can not only effectively estimate data with spatial autocorrelation, but can also reflect the spatial heterogeneity of parameters in different regions. The GTWR extends the traditional GWR by adding a temporal dimension, which can obviously make the regression result more accurate. According to existing research [54], it is calculated as follows.

$$Y_i={\beta }_0\left(u_i+v_i\right)+{\sum }_j{\beta }_j\left(u_i+v_i\right)x_{ij}+{\epsilon }_i,$$

(9)

where $Y_i$ is the dependent variable for geographical unit i; $u_i+v_i$ is the geographical mark of the ith regional unit; j is the number of independent variables; i is the number of regional units; ${\beta }_0(u_i+v_i)$ is the intercept ${\beta }_j\left(u_i+v_i\right)$ as a continuous function, the value of ${\beta }_j\left(u_i+v_i\right)$ in the unit i; $x_{ij}$ is the value of the j^th independent variable in the i^th geographic unit; and ${\epsilon }_i$ is the random error term.

4. Results

4.1. GFG and Urbanization

In this study, land use was divided into six categories, namely grassland, cultivated land, construction land, forest, water, and unused land. Our study area is dominated by forest, cultivated land, and grassland (Figure 3). In 2020, the area of these three categories was 4.54 × 10⁸ km², which accounts for more than 90% of the total area. In particular, the predominant form of land use change was the conversion of cultivated land to forest and grassland, which underscores the significant role of the GFGP.

The land use dynamics continuously increased by 20.74%, with the GFGP dynamics exhibiting intensity (

Table 2. Changes in the proportion of GFG, urbanization, and land use dynamics.

Year	Proportion of GFGP (%)	Proportion of Urbanization (%)	Comprehensive Dynamics (%)	Single Dynamic (Shrink Gradually)
2000–2005	1.67	0.46	16.39	Construction (+), unused, cultivated (+), grassland, forest, water (+)
2006–2010	3.82	0.56	19.71	Unused (+), construction (+), grassland, water (+), cultivated, forest (+)
2011–2015	2.56	0.71	19.76	Construction (+), grassland, unused (+), water (+), cultivated (+), forest
2016–2020	2.42	0.90	19.79	Unused (+), construction (+), grassland, water, cultivated, forest (+)

+: increase in dynamism.

). This suggests an elevated influence of human activities on land use changes. Simultaneously, in this process, the rate of construction land changes also drastically reflects the coexistence of the GFGP and urbanization (Figure S2). In the past 20 years, the total area of the GFGP first increased significantly and then decreased, reaching 4.48 × 10⁴ km². Specifically, the expansion of the GFGP was most prominent between 2006 and 2010. After 2011, it tended to flatten. This is because many regions coordinated the balance between the GFGP and food production so that some reforestation actions were temporarily suspended at that time [22] (Figure S1). Furthermore, the dynamics of construction have seen a sustained and rapid increase from 0.22 to 0.27. This trend also reflects the expansion of construction, which limited the proportion of the GFGP.

4.2. Spatial–Temporal Change in ESs

In terms of time, the degree of change in ESs showed an obvious difference. In general, while SC has relatively low variability, WC, CS, and HQ have relatively high variability. From 2000 to 2010, ESs demonstrated a relatively consistent trend of decline followed by a rise. With 2011 as the node, each module showed a different trend. CS and HQ continued to fluctuate and rise, WC began to fluctuate and decline, while SC remained stable (Figure 4).

From 2000 to 2005, several urban areas, such as Enshi and Yunfu, showed a slight increasing trend, but the relative provisioning of each ES was mainly decreasing. From 2005 to 2010, ESs exhibited a trend of rising in the east and west, and declining in the north and south of the central region. SC and WY showed a similarly upward trend in the remaining 36 cities and the most significant increase was in Liupanshui and Chongzuo. CS and HQ showed a downward trend, which was more significant in six cities including Liupanshui and Liuzhou. From 2011 to 2015, the spatial variation pattern of each ES had changed again, with an increasing proportion of urban areas exhibiting a downward trend.

Nanning, Laibin, Guiyang, and other urban areas showed a slight increase, intermittently distributed in the southern and northwestern regions. From 2016 to 2020, WY and SC showed a pattern of increasing in the west. Among them, the WY growth in Liupanshui was more significant. CS and HQ showed an almost identical pattern, for instance HQ showed the increase was significant in 10 urban areas such as Chongqing and Zunyi in the west (Figure 5).

4.3. Response of ESs to GFGP

The GFGP not only has a single positive effect on the karstland of southwest China, but also shows a significant negative effect in some areas, which changed over time (Figure 6). For all ESs except soil conservation, the spatial pattern of the GFGP’s impact on each plate is basically similar. It shows a relatively stable distribution trend of higher values in the northwest and southeast regions, while lower values were observed in the central region. For soil conservation, there is a spatial distribution trend of decreasing from the northwest to the southeast. The implementation of the GFGP had a significant improvement effect on SC and WC, CS, and HQ in the early western cities with dense karst landforms, represented by Qujing City and Liupanshui City. In southeastern Yunfu City and Zhaoqing City, the GFGP conversion had obvious effects on conserving water, but it did not increase the provision of soil conserving services. In southern cities such as Nanning and Laibin City, implementing the GFGP practices could effectively increase carbon storage and improve habitat quality. However, it had no obvious effect on soil conservation and was not conducive to water conservation.

From a temporal perspective, the spatial effect exerted by the influence of the GFGP on each plate changes little. There was no significant change in the spatial distribution of the impact of the GFGP on each plate from 2000 to 2010. Since 2011, there has been a large-scale decrease, and by 2020, there has been a significant increase in the proportion of negative impact areas. Over time, the ability of the GFGP to improve CS and HQ decreased the fastest and the most, followed by SC. WC provisioning was relatively stable, and did not change across the landscape, though there was a small-scale dip in WC provision in the center of the study area. The areas of high ecological benefit on the remaining three plates showed different degrees and directions of shrinkage. Combined with the distribution map of karst landforms (Figure 1), we found that regions that initially exhibited negative effects or later developed negative effects were typically plain areas with less-pronounced karst landforms. In contrast, in the western typical karst region characterized by significant topographical differences, land restoration through afforestation shows a more apparent positive effect on ESs. This observation suggests that the effectiveness of the GFGP is better demonstrated in ecologically vulnerable areas. However, even so, the overall impact of the GFGP on ESs tends to become negative after 2011, particularly concerning CS and HQ. Provision shrank from north and east to west, and the negative influence of the GFGP on carbon storage and habitat quality expanded from the central and eastern regions, eventually developing in contiguous areas. The positive benefits of the GFGP were predominantly concentrated in the western and southern regions of the study area.

4.4. Response of ESs to Urbanization

Similarly, urbanization has not only shown a negative effect on ESs in the karstland of southwest China, but has also shown a significant positive effect in some areas. Both the positive and negative effects of urbanization have changed over time (Figure 7).

The provision of WC and SC services was higher in the east and lower in the west, and similarly CS and HQ provisioning services were abundant in the north and rarer in the south. In addition to habitat quality, the other three plates maintained a relatively stable spatial distribution pattern in the past 20 years. The extreme value distribution of positive and negative effects between different plates is quite different. Urbanization has a stable and obvious water conservation effect in the eastern cities, represented by Qiandongnan and Guilin, while it is not conducive to soil and water conservation in the western cities such as Qujing. Early urbanization (pre-2005) in Chongqing was conducive to soil conservation and CS and HQ improvement, and urbanization in Kunming and Yuxi had the effect of improving HQ. The urbanization in the remaining cities has had a negative effect on the ESs of each section. On the whole, although there are some areas where urbanization is conducive to ecological benefits, the urbanization of major cities in the karstland hinders the potential of ESs.

From the perspective of temporal progression, the spatial pattern of the impact of urbanization on each plate remained relatively stable, the ecological weakening effect has maintained a steady downward trend, and the effect difference between cities has narrowed. From 2000 to 2005, the negative effect of urbanization on ESs was obvious. Since then, the deteriorating effect of urbanization on ecology has gradually weakened over time. The low-value area of soil conservation and CS shrinks to the south, and the high-value area expands in the middle. The deterioration of habitat quality by urbanization has remained stable since weakening immediately after 2005, and the overall effect is still obvious. However, urbanization has had little effect on WC, with no significant change in the past 20 years. On the whole, the early urbanization (pre-2005) had a significant weakening effect on the plates except for water conservation, but it has been improved in the later period. As of 2020, the negative effect of urbanization on soil and water conservation is weak, but it still shows significant adverse effects on CS and HQ.

5. Discussion

5.1. Impact of GFGP on ESs

Reflecting the coexistence of ecological construction and economic development, ESs in the study area are affected by the interaction of the GFGP and urbanization. During the period of the GFGP, the decline in Ess caused by urbanization was buffered. This is consistent with the results of existing research that the GFGP has mitigated the loss of Ess [55]. Nevertheless, there are many other aspects of the response of Ess to the GFGP and urbanization that are still worthy of discussion [26,56,57]. For example, there is a lack of research on the simultaneous discussion of the GFGP and urbanization, although this response varies among different regions, time periods, and Ess. Therefore, this paper mainly reflects the relationship and negative relationship of this process, and analyzes the regression coefficients of the GFGP and urbanization and various Ess from the model results.

Before 2005, the Ess (WY, CS, HQ) exhibited a substantial decline, also indicating a mainly negative impact of the GFGP in the early stage. This is because early plantation vegetation has a relatively weak carbon sequestration capacity due to its shallow root depth and low canopy density, while some disturbances such as mechanical soil preparation, pit digging, and tree planting exacerbated the evaporation losses of conserved water and carbon emissions in the early stages of afforestation [58]. After 2005, HQ and CS services increased significantly, indicating that both recovered quickly after afforestation, while water conservation fluctuated and tended downward, because afforestation worsened evapotranspiration. Additionally, since the assessment of the InVEST SC was only affected by land use change, which accounts for only 3.4%, the SC remained basically unchanged.

The GFGP and urbanization have heterogeneous influences on ES functions, considering factors such as tree growth stages and variations in the implementation. According to ‘Guangxi Zhuang Autonomous Region Land Space Ecological Restoration Plan’ (2021–2035) [59], the highest tree age is only about 20 years, the root depth of vegetation is shallow, the capacity for soil fixation is weak, and the forest structure is mainly artificial forest, so the quality of ESs is not high. Therefore, different stages of reforestation in different regions will have different impacts on ESs.

Firstly, in terms of WY, according to numerous existing studies [58,60], it is scientifically observed that under certain circumstances, large-scale afforestation can have a negative effect due to increased evapotranspiration and a subsequent reduction in regional water use. In the early stages of the GFGP, there were issues including high planting density, extensive coverage, and the use of non-native tree species, which resulted in a reduction in watershed runoff [3]. This is consistent with the pronounced spatial pattern of worsening WY provision—which is most threatened in water-poor karst landscapes. Our results underscore that future management in these systems should consider that delicate water balance. In terms of soil conservation, the positive effect of the GFGP showed a trend of decreasing from west to east and decreasing from east to west. This is consistent with the distribution of topography in the study area. To a certain extent, this suggests that the effect of the GFGP on SC is more significant in the hillier areas where the terrain is more heterogeneous. Therefore, the topology should be a consideration in the planning and implementation of the GFGP [61]. However, it is one-sided to analyze the direct relationship between the GFGP and soil conservation because this study did not investigate the associated indirect influences of the GFGP on the local climate. In future research, we will consider natural factors like topology, climate, etc., to enhance our understanding of afforestation’s impact on ESs and identify locally adaptable approaches. Additionally, establishing long-term monitoring programs will enable us to assess the sustainability and stability of the afforestation efforts, while identifying potential challenges and limitations.

For CS, some studies suggest that the impact of the GFGP is mainly positive in the northern region and has limited influence in southern China [62]. The impact of GFGP on CS in the southern region, which is characterized by favorable water and thermal conditions, intact vegetation types, and fast carbon cycling rates, may not be significant in the short term. However, it is believed that GFGP can still contribute to long-term CS through natural processes that span several decades or even centuries [63]. This will be achieved not only through the carbon storage capacity of trees but also by providing abundant litter accumulation, which facilitates soil carbon sequestration. Afforestation can not only use trees to achieve CS, but also facilitate the conversion to organic matter and soil under long-term litter accumulation, thus promoting soil carbon sequestration [64]. This will be a long process due to the amount of time necessary for forest growth and soil formation. Therefore, the results of CS are not significant in the early stage of the GFGP, and may even be detrimental to the effect of ecological restoration due to the increased interference of human activities [62]. In terms of habitat quality, the model mainly combines land use and threat sources and sensitivity coefficients, reflecting the availability of living resources, and urbanization presents a high threat to habitat quality [65]. Our study confirms that the regression coefficient shows significant spatial heterogeneity. This leads us to consider the ecological and economic needs in the zoning management of habitats. In addition, this study also found that the GFGP accounts for the total study area that decreased after 2010, which to some extent reflects the balance of ecological and economic development in different regions. For example, the scale of the GFGP in Guizhou Province is large, and the positive effect on habitat quality is more obvious. The urbanization development pattern of Guangxi Province, especially Nanning, focuses more on economic benefit growth, and the ESs of the GFGP are mainly negative effects.

5.2. Impact of Urbanization on ESs

From the perspective of the impact of urbanization on ESs, in terms of water conservation, urbanization has a more obvious positive effect on the eastern region. This is because the terrain in the eastern region is flat, and the land preparation after urbanization is conducive to preventing water loss, ultimately improving water conservation. It is consistent with some studies suggesting that urban policies and social and economic factors are generally conducive to water conservation [66]. For soil conservation, we observed that urbanization had a significant negative effect on soil conservation before 2005, though that negative effect did weaken from 2006 to 2020. This showed that moderate urban development was conducive to reducing the risk of soil erosion and loss [67]. Slope protection, land preparation projects, and the implementation of land protection policies can effectively reduce soil erosion.

Urbanization has a negative effect on carbon sequestration services, but similar to soil conservation, we observed that the negative effect of urbanization on carbon sequestration gradually weakens over time [68]. This may be because when natural land is converted to artificial land, the land changes to hardened ground, natural vegetation changes to artificial buildings, and carbon losses increases. The implementation of ecological restoration measures, such as urban green space construction [69], is beneficial to the carbon cycle, thus increasing strong CS, which is consistent with the research that urbanization is conducive to the soil carbon cycle [70] and urban ecosystems will begin to compensate for carbon losses after a decline in carbon stocks [71]. Urbanization also showed a negative effect on HQ, though the magnitude of this negative influence gradually weakened over time in the study. And in Chongqing, Yuxi, Qiandongnan, Yunfu, and Zhaoqing, there has been a positive impact observed across multiple years during the period from 1999 to 2020. During urbanization, complex natural ecosystems are replaced by constructed land that is extremely sensitive to ecological perturbations [72]. However, according to research holding the same view [73], it is believed that the inclusion of natural corridors that cut across different ecosystems can promote an increase in habitat diversity. This can weaken the negative effect of urbanization on habitat quality. In summary, the effect of the GFGP on ESs depends on regional abiotic and biotic characteristics, types of GFG, and the methods.

Although this study separately examines the impacts of the GFGP and urbanization on ESs, it is also essential to explore their combined effects. In this study, it was observed that WY demonstrates the most prominent disparity between the effects of the GFGP and urbanization on ESs, closely followed by SC and HQ. Despite the existence of conflicting impacts, a comprehensive understanding of the effects from the GFGP and urbanization is still lacking. For example, there is a study [74] on the impact of urbanization and ecological construction on ESs that considered the trade-offs involved. Additionally, the selection of the GFGP and urbanization as the main themes for studying ESs is not enough to fully capture the complex mechanisms underlying the influence on ESs. Therefore, a more comprehensive and specific approach would be to conduct a quantitative assessment considering various elements [75]. Overall, consistent with many existing studies, this study believes that the GFGP has a promoting effect on services [26,55,76,77,78], and urbanization has an inhibiting effect on services [19,27]. Due to the complexity of geographical factors in karstland, the effects of the GFGP on SC, WY, CS, and HQ have significant spatial heterogeneity. Through measures such as green space construction planning and policy protection in the process of urbanization, the negative effects of urbanization on ESs can be reduced to some extent [79]. Therefore, a comprehensive consideration of local needs for individualized forest restoration measures is an effective measure to realize ecological benefits.

6. Conclusions

In general, we found that the area of the GFGP first increased and then decreased, reaching a maximum of 3.82% from 2006 to 2010. The dynamic degree of land use increased from 16.39% to 19.79%. Reflecting a remarkable triumph of GFGP, the mutual conversion between forest and cropland is the main form of land use transformation. On the other hand, ESs have changed with land use change. The undulating karstland terrain and fragile topsoil are easily affected by disturbances such as digging and overturning of the land in the early stage of the GFGP, causing certain damage. At the same time, the ecological protection function of young trees is relatively weak, which results in improvements in ESs in the GFGP clearly lagging behind. Although the provision of ESs in SC has not changed in the past 20 years, the provision of ESs in the karstland of southwest China showed a fluctuating upward trend. From 2000 to 2012, WC, WS, and HQ showed obvious fluctuation. After 2013, they remained relatively stable or slightly increased. The large-scale excavation carried out during the early stage of the GFGP in fragile soil layers with steep slopes has exacerbated soil erosion to some extent. Additionally, the GFGP has also contributed, to some extent, to increased carbon emissions. However, over time, the rooting systems of these planted forests have improved, increasing the ability of forests to conserve water and soil, suggesting that ES provision had recovered by 2010. The spatial heterogeneity of the impact of the GFGP on the ES function of each section was significant. The negative effects of WC, SC, and CS are relatively small, and the positive effects are more obvious in the western and northern cities, where the distribution density of karst landforms is relatively large, indicating that ESs under karst landforms are more sensitive to the GFGP. The negative effects of HQ are smaller, and the positive effects are more obvious in the southeastern cities. In addition to having a negative effect on ESs in the SW karstland, urbanization has a significant positive effect in some areas, which changes over time. The negative and positive effects of SC, CS, and HQ were smaller and more obvious in the cities of Northeast China. The negative and positive effects of WC were smaller and more obvious in the cities of Northwest China. According to the initial assessment of this study, as time passes, the negative impact of urbanization on ecology is gradually weakening and the difference between regions will grow smaller, which may reflect that the influence of urbanization is gradually mitigated by green initiatives.