*3.2. Drivers of Change in EEQ*

Unlike other karst areas in the world, where the population density is low, the karst mountains in southwest China are populous and ecologically fragile. There is sparse coordination between peasant–land conflicts and high ecological pressure [42]. Various ecological restoration projects have been promoted since 2000, such as the Grain to Green Project (from 2000) and the Karst Rocky Desertification Restoration Project (from 2008) [43]. According to the mean values of the four indicators presented in Table 4, the overall indexes were relatively stable, and the average RSEI values were also relatively stable.


**Table 4.** Mean indicators of RSEI from 2000 to 2020.

During 2010–2015, the mean value of NDVI increased by 23.92%, and the WET mean value also increased significantly. The effectiveness of the ecological restoration project is also highlighted during this period, having contributed to a significant decrease in the mean value of LST. The ecological restoration project also contributed to an increase in the mean value of RSEI in the study area (Table 4).

From 2015 to 2020, PAR was promoted and completed. During this period, NDISI was reduced significantly, by 17.24%. Relocation to alleviate poverty caused many rural people to move to towns and cities, while their original home base was reclaimed and re-greened. In this case, the impact of human activities on rural areas was significantly reduced, which directly contributed to a significant reduction in the dryness index. The NDVI was increased, along with the WET, which means that the conflict between humans and land was fundamentally relaxed, and the effectiveness of various ecological restoration projects was maintained and further improved. PAR was a key factor in the improvement of RSEI in the study area (Table 4).

#### *3.3. The Contribution of PAR to EEQ Changes*

A multiple linear regression analysis model was used to calculate the residual trends in PAR implementation, whereby the turning point of RSEI change caused by PAR was determined to occur in 2016. The variation in RSEI was influenced by natural factors, including MAT and ATP, as well as human activities, including the Grain to Green Project and the Karst Rocky Desertification Restoration Project. We took these four factors as the independent variables, i.e., the influencing factors, and RSEI was adopted as the dependent variable, i.e., the resulting factor. Standardized coefficients of the regression models were analyzed by the SPSS with 95% confidence (Table A1). The results show that the observed cumulative probability and the predicted cumulative probability were normally distributed in the linear regression analysis model (Figure 5). Furthermore, the standardized residuals were randomly distributed without outliers (Figure 6), and the regression model significance was 0.038 at a significance level of 95%. Therefore, the multiple linear regression equation was verified to be stable. The analysis results show that the RSEI variable could mathematically be represented as Variable (RSEI) = 0.273 × Variable(MAT) + 0.285 × Variable(ATP) − 0.144 × Variable (funds of the Grain to Green Project) + 0.520 × Variable (funds of the Karst Rocky Desertification Restoration Project). This equation predicted the mean RSEI values in 2017, 2018, and 2020 to be 0.5676, 0.5660, and 0.5701, respectively, under the influence of the four factors. The observed RSEI values in 2017, 2018, and 2020 were 0.5942, 0.6374 and 0.6363, respectively. The residual difference between the observed and predicted mean values of RSEI was likely due to the PAR-driven RSEI improvement. The improvements in RSEI caused by PAR were 0.0266, 0.0715, and 0.0662 in 2017, 2018, and 2020, respectively, according to the actual increase of 0.0266, 0.0715, and 0.0662. Compared with the turning point of 2016, the PAR contributed to the RSEI growth contribution rate of 70.56%, 88.38%, and 82.96% in the three years, respectively. Collectively, after 2016, the RSEI values increased significantly, and from there on, the average RSEI reached good levels and remained relatively stable.

**Figure 5.** P-P plot of regression standardized residuals.

**Figure 6.** Scatter diagram.

#### *3.4. Correlation between PAR and EEQ Changes*

The correlation between relocation and EEQ changes was analyzed by taking the village area as the basic unit. A map showing the change in RSEI during 2015–2020 is shown in Figure 7.

**Figure 7.** Change map of RSEI during 2015–2020.

There were 1330 administrative village units in the study area. The number of villages in the area with reduced RSEI was 205, accounting for 15.41% of the total. A total of 1125 villages exhibited improved RSEI. These villages were mostly located in the central and northern regions of Pu'an and Qinglong. The original EEQ of these regions was relatively low, and the effect of the upgrade in terms of RESI was obvious. The relocation of the five counties on the northeast side of the study area of Southwest Guizhou Autonomous Prefecture, including the counties of Ceheng, Qinglong, Wangmo, Zhenfeng, and Pu'an, was the most concentrated, whereby 74,600 households were relocated, accounting for 82.47% of the total (Figure 2).

Figure 8 shows the spatial distribution of the coupling coordination between the relocation population density and RSEI variables in 1222 administrative villages in eight counties. There were 247 villages with extreme coordination, accounting for 25.33% of the area, mainly concentrated in Ceheng, Wangmo, Qinglong, and Pu'an, which are the most densely relocated areas. There were 411 general coordination villages, accounting for 33.90% of the area, most of which were located in the northeast and southeast of the study area. There were 403 general detuning villages, accounting for 29.99% of the area, and 161 extreme detuning villages, accounting for 7.63% of the total villages. These villages were predominantly located in economically developed areas, such as Xinyi. After the spatial analysis of the coupling coordination between relocation population density and RSEI variables, a significant coupling coordination relationship was found between EEQ enhancement and PAR in the study area.

**Figure 8.** Coupling coordination degree of density of relocated population, and change in RSEI.
