Long-term vegetation change and its driving factors in a typical catchment of the Loess Plateau

Abstract

Vegetation information is critical for understanding terrestrial water and carbon dynamics, and is a key factor for regional environment management under climate change. This study selected a typical arid and semi-arid catchment, in the Loess Plateau, the Zuli River Basin (ZRB) to investigate the long-term vegetation cover change and its driving factors. Different from available Normalized Difference Vegetation Index (NDVI) products that were produced at coarse spatial resolutions and with short temporal ranges, high-resolution (30m) and long-term (31 years) NDVI data was generated in this study by using 30m resolution Landsat dataset analyzed within the Google Earth Engine (GEE), a cloud platform with planetary-scale analysis capabilities. We evaluated the relationships between NDVI and precipitation, runoff coefficient and sediment concentration using cross wavelet, and explored the impact of precipitation on vegetation and the impact of vegetation cover on runoff and sediment before and after the implementation of Grain for Green Project (GGP, conservation program designed to mitigate and prevent flooding and soil erosion in China) in different parts of ZRB in 1956-2016. Our results show that there was an increasing trend in average maximum annual NDVI across the ZRB during 1987-2016 and a significantly (p < 0.05) increased trend in most areas. The vegetation cover increases in upstream Chankou and upstream Huining (two representative catchments within the ZRB) was higher than that in the middle stream and downstream, but with a smaller increase rate. Widespread increase in vegetation mostly occurred after GPP because the area with significant increase from 1987-2016was higher than the increased area in 1987-1999. The precipitation had significant correlation (p < 0.05) with NDVI series across the ZRB except for the upstream Chankou where human activities played a major role. The relationship between NDVI and the runoff coefficient, sediment concentration was significantly negative (p < 0.05), which indicates that the vegetation cover is an important reason for reduction of runoff coefficient and sediment concentration. Meanwhile, the human activities also played a positive role on both restoration of vegetation and reduction of runoff coefficient and sediment concentration across the ZRB.