Temporal and spatial patterns of NDVI and their relationship to precipitation in the Loess Plateau of China

The relationship between vegetation and precipitation has been studied extensively, but little is known about its mechanistic linkage to agro-ecosystem sustainability. In this study, we used 250 m MODIS NDVI 16-day composite data and precipitation data for the period 2001-2005 to evaluate the connection between vegetation and precipitation in the Jinghe River watershed of the central Loess Plateau, China. A principal component analysis (PCA) was employed to determine the primary feature of vegetation cover. The results indicated that the first two principal components (PC1 and PC2) explained around 89% of total NDVI variances. The spatial distribution of PC1 followed an increasing trend from northwest to southeast, a similar dynamics to the spatial pattern of NDVI, which was determined by the variation of mean annual precipitation in the Jinghe River watershed. Furthermore, PC2 was associated with the spring/summer modes of the annual cycle while their intra-annual variations were largely modified by the regular phonological cycle of the winter wheat-summer fallow cropping practices, a dominant farming practice in severely degraded regions of the Loess Plateau. Winter wheat-summer fallow practice was a dominant factor affecting the precipitation-vegetation relationship. A sharp decrease in vegetation cover after wheat harvesting during the rainy seasons from July to September may be a major cause of extensive soil erosion in this region. Our findings provide new insight into the relationship between vegetation and precipitation which is critical in the restoration of degraded ecosystems and developing appropriate crop rotation systems to achieve a certain amount of agricultural sustainability.