Abstract: Fractional vegetation coverage (F_VC) is an important indicator of vegetation growth in arid and semi-arid regions. Trend analysis was used to study the spatial-temporal change of F_VC in Xilingol of Inner Mongolia from 2000 to 2019, and ordinary least squares (OLS) and geographical weighted regression (GWR) were used, respectively, to explore the dominant factors of F_VC based on 10 km×10 km grid. The results show that the annual growth of F_VC was 0.007 and F_VC owned the characteristics of overall improvement but with some local reduction. Spatially, F_VC decreased in parts of the east, significantly improved in the middle and was relatively stable in the west. The influencing factors of vegetation show that precipitation, atmospheric temperature and population were the dominant factors of F_VC with strong spatial aggregation, and the dominant proportion accounted for ≥ 82.10%. Specifically, precipitation was positively dominant. Temperature was negatively dominant. Population played a dual role, which gradually increased with the passage of time. The influences of elevation, slope gredient, slope aspect, sand, and clay on vegetation were not significant and with broken spatial aggregation. The dominant proportion accounted for only ≤ 17.10%. Mining had various effects on vegetation and vegetation had not been improved significantly around the open mining areas. When exploring the causes of spatial heterogeneity of vegetation in Xilingol, GWR was better than OLS. The study on F_VC has certain theoretical and practical significance for arid and semi-arid areas in Xilingol, and the results are expected to provide scientific supports for the ecological management in local areas of Xilingol.

Key words: fractional vegetation coverage (F_VC), multicollinearity testing, dominant factors, geographically weighted regression, Xilingol