Answers are grounded in this paper and cite their sources.
Demonstrates ecological risk patterns in the Kaidu-Kongque River Basin, indicating effective land management strategies.
Landscape ecological risk assessment (LERA) is of key importance for maintaining the stability of natural systems and guiding land spatial optimization. However, quantifying landscape ecological risk (LER) in arid river basins remains challenging because of fragile ecosystems, rapid land-use transitions, and increasing human disturbances. Taking the Kaidu-Kongque River Basin as a typical region, this study integrated land-use transition analysis, PLUS simulations, LERA, and XGBoost-SHAP to systematically examine land-use evolution, spatiotemporal patterns of LER, and associated predictors. Based on validation using 2020 land-use data, future LER distributions were further estimated under multiple development pathways for 2030. Results showed that (1) During 2005–2024, the regional landscape was overwhelmingly characterized by barren land and grassland, together covering over 90% of the basin. Cropland increased by 2046.60 km 2 , while barren land decreased by 1353.55 km 2 . Grassland experienced a total outflow of 3512.74 km 2 , mainly converted to cropland and barren land. (2) LER exhibited an overall downward tendency, with the high- and highest-risk areas decreasing by 3387.79 km 2 . Global Moran's I remained significantly positive from 2005 to 2024, ranging from 0.776 to 0.803, indicating persistent spatial clustering of risk. (3) The XGBoost-SHAP framework demonstrated relatively good predictive performance under five-fold spatial block cross-validation. Among the selected predictors, natural and river-proximity variables exhibited higher model-based predictive contributions to LER variation than the socioeconomic variables. (4) Scenario simulations for 2030 indicated that the LER distribution would largely maintain the historical pattern, but with varying risk levels and class structures. Relative to 2024, the natural development (ND) scenario reduced the combined area of high- and highest-risk classes by 5936.92 km 2 . Relative to the 2030 ND scenario, the ecological protection (EP) and economic development (ED) scenarios further reduced this area by 5773.84 km 2 and 2568.52 km 2 , respectively. These changes can be interpreted through four coupled land-process pathways: oasis agricultural redistribution, grassland-buffer degradation, aquatic-patch fragmentation, and desert-margin disturbance. These findings provide landscape-index-based evidence for spatially targeted land management and pattern-based ecological risk control in arid basins.
No takes yet. Share an insight, caveat, or question.
Luo et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: