Detection of dryland degradation using time series segmentation and residual trend analysis(tss-restrend)

Grassland ecosystem is widely distributed and is an important ecological barrier in China, so it is of great significance for monitoring grassland degradation. Remote sensing data can provide long-term and large-scale records of grassland ecosystem changes. Therefore, remote sensing is an effective way to monitor land degradation in grassland areas. Xilingol grassland is a combination of typical grassland, meadow grassland and desert grassland in arid and semi-arid areas of northern China. We Used NOAA GIMMS NDVI3g data and rainfall time series data from 1983 to 2008 in this article. Meanwhile, time series segmentation and residual trend analysis (TSS-RESTREND) was used to monitor and analyze land degradation in Xilingol grassland. This method can remove the influence of climate factors on NDVI, so as to analyze the change of grassland caused by human activities. The results showed that during 1983-2008, 86.8% of the pixels had no obvious change, 2.5% of the pixels had degradation, 4.7% of the pixels had greenness increase, and 6.0% of the pixels were uncertain. In terms of spatial distribution, land degradation occurred in the eastern part of Xilingol, while land greenness increased in the western part.