Abstract: Vegetation is a key link connecting the atmosphere, pedosphere, hydrosphere and biosphere. The relationship between the spatiotemporal variation of vegetation and its driving factors is of great significance in the research on regional environmental changes. Based on the datasets of vegetation NDVI, climate, vegetation type, geomorphology, soil type, and DEM, we analyzed the spatiotemporal variation of vegetation NDVI and driving factors in Qilian Mountain National Park from 2000 to 2019 using a geographic detector model. The results showed that: (1) Vegetation in Qilian Mountain National Park had gradual improvement from 2000 to 2019; (2) The area with low and medium-low grade NDVI accounted for more than 55% of total area, the area of middle and high grade accounted for about 35%, and the proportion of high-grade area was extremely low. Meanwhile, the vegetation NDVI showed a downward trend from east to west; (3) Precipitation, altitude, soil type, and temperature were the main factors affecting the variation of vegetation NDVI; (4) There were interactions among the evaluation indicators on vegetation NDVI, which were doublefactor enhancement and nonlinear enhancement effects; (5) The most suitable scope or categories for vegetation growth in Qilian Mountain National Park were as follow: mean annual rainfall of 320.7-385.8 mm, moisture index of 2.9-21.3, mean annual temperature of 0.2-5.2 ℃, vegetation types of coniferous and broad-leaved forest, soil types of black and gray forest soil, DEM of 2875-3365 m, ≥10 ℃ accumulated temperature of 346.2-720.5 ℃, geomorphic type of medium undulating mountain, slope of 17.1°-22.8°, slope direction of 0°-31.55°.

Key words: vegetation NDVI, Qilian Mountain National Park, geographical detector model, spatiotemporal variation.