Abstract:

Vegetation constitutes a critical nexus among the atmosphere， soil， water， and biosphere. Investigating its spatiotemporal dynamics and driving mechanisms is critical for understanding ecosystem evolution. This study leverages 22-year MOD13Q1 NDVI remote sensing data （2000-2022） and integrates natural factors （digital elevation models， soil types， precipitation， temperature， evapotranspiration） and socioeconomic parameters （land use types， population， GDP）. Analytical techniques including pixel binning models， trend analysis， and geographic detectors were applied to systematically evaluate vegetation coverage patterns and their drivers in the Kubuqi Desert. Results demonstrate a sustained upward trend in vegetation coverage， with an annual growth rate of 0.387% （R²≈0.832）. Seasonal variations reveal peak coverage during June-August and minimal levels in November-January. Notably， 73.62% of the study area exhibited vegetation improvement： the desert core transitioned from extremely low to low coverage grades， the southwest shifted from low-medium to medium-high grades， and the southeast achieved widespread high-grade coverage. Single-factor analysis identified land use types （0.493）， precipitation （0.461）， and population density （0.443） as dominant individual drivers. Multi-factor interaction detection highlights precipitation-elevation synergy （interaction q=0.731） as the most significant combined influence on vegetation dynamics.

Key words: Kubuqi Desert, vegetation coverage, temporal and spatial variation, geographical detector