Impacts of Land-Use Change on Past and Future Carbon Stocks in the Tianshan North Slope Economic Belt

To study the connection between land use and regional ecological carbon stocks, predicting the future evolution of land use and understanding the trends of changes in carbon stocks are essential to environmental protection and sustainable development. However, the changes in carbon stocks within ecosystems driven by land-use and land-cover change (LUCC) are characterized by large uncertainties. This study took the Tianshan North Slope Economic Belt in Xinjiang Autonomous Region, China, as an example to investigate the spatiotemporal changes of carbon stocks and their relationship with LUCC, and used 11 variables in a Geo Detector model to analyze the drivers of spatial differentiation in carbon stocks. We trained six variables to predict the changes in carbon stocks under a natural development scenario (NDS) and ecological protection scenario (EPS) in 2030. The following results were obtained: (1) the land use was dominated by unused land in the 20-year study period (2000–2020). Grassland showed a continuous decrease; unused land decreased and then increased, while others continued to increase. The most drastic change was for cropland, which was 7785 km² (an increase of 39.76%), while grassland was reduced by 9402 km² (a decrease of 9.05%). (2) Carbon stocks showed an increasing and then a decreasing trend, with an overall increase of 2.05 × 10⁶ t. The spatial distribution was more centralized in the southwest, showing a continuous distribution with slice-like bands, while the higher values took an irregular form in the northeastern portion of the region. (3) Carbon stocks under NDS reached 1427.50 × 10⁵ t, an increase of 6.26 × 10⁶ t compared with 2020; while under EPS, they reached 1427.79 × 10⁵ t, an increase of 6.29 × 10⁶ t, mainly due to grassland and unused land conversion. Therefore, the protection and restoration of grassland should continue to be strengthened in the future. (4) NDVI and soil erosion had strong explanatory power for the spatial variability of carbon stocks. There was two-factor and nonlinear enhancement interaction in different factors, indicating that natural and human factors enhance the explanation of the spatial variation, the results of which can be applied to the ecosystems of the region.