Abstract: Based on data of GIMMS (Global Inventory Modelling and Mapping Studies) NDVI (Normalized Difference Vegetation Index), temperature, and precipitation, we analyzed the spatiotemporal variation of vegetation cover and its seasonal response to climate change in the Loess Plateau from 1982 to 2015, using the Ensemble Empirical Mode Decomposition (EEMD), linear regression analysis, and partial correlation analysis. In the interannual variation trend from 1982 to 2015, NDVI on the Loess Plateau increased significantly in the growing season, spring, summer, and autumn. The increasing rate of NDVI in each season also increased during 1982 and 2015, with the most obvious change in summer. Spatially, NDVI gradually increased from northwest to southeast in the growing season, spring, summer and autumn, which was statistically significant in most areas. As shown by the linear regression analysis, temperature had a significant increase trend in the growing season, spring, summer, and autumn, while precipitation increased in the growing season and autumn and decreased in spring and summer. The results of EEMD analysis further showed that in the growing season, spring, summer, and autumn, temperature first increased and then decreased, and precipitation first decreased and then increased. In terms of spatial variation, temperature had a significant increase trend in most areas during the growing season, spring, summer, and autumn, while precipitation had a significant increase trend only in some areas during autumn. There was a significant positive correlation between NDVI and temperature in the northeast and southwest of the Loess Plateau, and a significant positive correlation between NDVI and precipitation in the north and northwest of the Loess Plateau.

Key words: abandoned cropland, soil physical property, Ordos Plateau., shrub encroachment