Variation in fractional vegetation cover and its attribution analysis of different regions of Beijing-Tianjin Sand Source Region, China

Abstract

Based on the MODIS NDVI data from 2000 to 2018, we estimated the fractional vegetation cover (FVC) using the dimidiate pixel model and analyzed the spatiotemporal characteristics of FVC in the Beijing-Tianjin sand source region (BTSSR). The geographical detector model was used to estimate the impacts of natural and human factors on FVC spatial distribution at the regional scale. The results showed that the FVC of the BBTSR showed an increasing trend from 2000 to 2018, with an annual growth rate of 0.013·(10 a)-1 and a vegetation increase rate of 8.2%. The area with high FVC was concentrated in the Yanshan Mountain water source protection area, followed by the pastoral transitional zone desertified land control area and the Otindag sandy land area. The area with poor FVC was concentrated in the northern arid grassland area. The explanatory power of driving factors to FVC varied across different regions. Among the natural factors, annual precipitation was the main driving factor for the spatial distribution of FVC in the northern arid grassland area, the Otindag sandy land area and the Yanshan Mountain water source protection area. Slope was the main driving factor for the spatial distribution of FVC in the pastoral transitional zone desertified land control area. Among different human activities, the number of large livestock at the year-end was the main driving factor controlling the spatial distribution of FVC in the northern arid grassland area and the pastoral transitional zone desertified land control area, while population density was the main driving factor controlling the spatial distribution of FVC in the Otindag sandy land area and the Yanshan Mountain water source protection area. There were regional differences in the influen-ce of other factors on FVC spatial distribution. The results of the interaction detector showed that the two-factor interactions were mainly the double-synergy and nonlinear synergy. The interaction of human activities with annual precipitation and slope could more fully explain the spatial variations of FVC. The range of suitable vegetation growth identified by the risk detector was the area with annual precipitation of 316.4-486.0 mm, average relative humidity of 48.4%-57.6%, and average annual temperature of 2.5-7.9 ℃, while other driving factors were different in different zones.