Spatio-temporal land-use dynamics and landscape ecological risk assessment in an artificial oasis, Northwestern China

Understanding long-term land-use changes and their ecological consequences is essential for managing fragile artificial oasis systems in arid regions. This study analyzes annual land-use/land-cover (LULC) dynamics in the Alar Reclamation Area (northwestern China) from 1990 to 2019 using multi-temporal Landsat imagery, maximum NDVI composites, and a supervised SVM classifier. We produced annual LULC maps, quantified area changes and transition matrices, computed landscape pattern metrics (e.g., patch density, edge density), detected abrupt change points, and evaluated ecological risk using a landscape disturbance-vulnerability framework. Socioeconomic and climatic drivers (population, agricultural production value, cotton price, temperature and precipitation) were integrated to explain observed transformations and used in CA-Markov scenario simulations. Key findings: (1) cultivated land, orchards and construction land expanded substantially (net increases of 1147.2 km², 674.2 km² and 36.5 km², respectively), largely at the expense of unused land and natural vegetation; (2) a structural turning point occurred around 2005, associated with policy and market drivers; (3) landscape fragmentation increased, and ecological risk concentrated in reclamation belts adjacent to the Tarim River; (4) scenario simulations show that high-intensity development would markedly raise ecological risk, whereas conservation-oriented management can mitigate risk. The study identifies trade-offs between agricultural development and ecosystem stability, highlights salt-crust degradation and increased erosion as key ecological concerns, and provides spatially explicit evidence to inform land-use planning. Limitations include reliance on medium-resolution imagery and limited in-situ measurements; we therefore recommend future integration of higher-resolution imagery and process-based erosion monitoring.