Arid grassland ecosystems undergo degradation because of increasing environmental and human pressures. Degraded grasslands show vegetation cover reduction and soil-patch development, leading to grassland fragmentation and changes in spatial heterogeneity. Understanding grassland degradation that involves soil-patch development remains a challenge over large areas with limited accessibility such as the Qinghai-Tibetan Plateau. We hypothesized that vegetation cover, its spatial heterogeneity and changes thereof over time retrieved from satellite data can indicate grassland development and degradation levels. To test the hypothesis, we studied these indicators from 2000 to 2016 and related them to previously described degradation levels on the eastern Qinghai-Tibetan Plateau (QTP) in 2004. We further use these indicators to map the new grassland development and degradation levels in 2016.

We found that lower vegetation cover does not always indicate a more severe degradation; instead, higher spatial heterogeneity is a better correlate of degradation. Combined temporal changes in grassland cover and its spatial heterogeneity are related to the literature-defined degradation levels. We found that grassland areas on the eastern QTP have moved into new degradation stages from 2000 to 2016 using changes in grassland cover and its spatial heterogeneity as indicators. The normalized difference vegetation index (NDVI) as a proxy for grassland cover declined over time in the literature-defined degraded areas but increased in the desert areas from 2000 to 2016. Spatial heterogeneity generally increased across different degradation levels from 2000 to 2016; however, this increase was less pronounced in severely degraded and slightly deserted areas. Our newly defined degradation levels in 2016 included degradation, desertification, and improving levels. Across our study area, 63% of all areas were classified as degraded and 2% were at risk of desertification. The remaining areas (35%) classified as improving and re-growing occurred in higher-elevation or previously severely degraded grassland. Our study demonstrates that a combination of changes in grassland cover and in its spatial heterogeneity can indicate grassland degradation levels and serve as an early-warning signal for desertification threats.