Abstract: Loess Plateau is featured with fragile ecological environment. Affected by monsoon climate, it has obvious distinction among four seasons and apparent vegetation changes, which offer a good testing ground for researching the consistency of different remote sensing data. According to the Normalized Difference Vegetation Index (NDVI) derived from GIMMS in 2001-2006 and from MODIS in 2001-2014, we analyzed the variations of vegetation in the loess plateau. And furthermore we investigated the differences between the two data in aspects of spatial distribution, seasonal and annual variations of vegetation in the loess plateau of Shaanxi Province. Results suggest that data obtained by the two methods both reflect the scarce coverage in the northwest and good coverage in the southeast of the loess plateau. What’s more, MODIS data is sensitive to the variation of vegetation differences. In terms of the trend, vegetation index increased year by year, and the most rapid increment was in autumn, indicating that the vegetation restoration work in recent years has achieved remarkable result. Compared with GIMMS, MODIS is more suitable for reflecting the spatial distribution of vegetation cover in the loess plateau.

Key words: NDVI, loess plateau, GIMMS, MODIS, coefficient of correlation, comparative analysis of time series data