Landscape ecological risk assessment of Yulin Region in Shaanxi Province of China

Landscape ecological risk assessment is a basic content of studying the relationship between landscape and ecosystem. It can reveal the spatial-temporal heterogeneity and scale effects of ecological risks, and provide the comprehensive spatial characterization and visualization of multisource risks. Yulin is a typical agropastoral zone in the north of China. Land use and land cover (LULC) changes are frequent and intense, and the landscape ecological environment is very fragile. This study analyzed the change characteristics of land use and land cover of Yulin by ArcGIS, assessed the landscape ecological risks using 4 landscape metrics, and discussed the spatial-temporal trends and the influencing factors for the landscape ecological risk. The results show that the absolutely dominant land use and land cover types in Yulin are grassland and cultivated land. The changes mainly happened in grassland, artificial surface, cultivated land, and bare land, with the variation of - 2.83%, 1.85%, 1.35%, and - 0.23%, respectively. High landscape ecological risks are mainly distributed in Yuyang and part of Jingbian. Most part of Yulin region belongs to the risk grade of Low/Medium and Medium. No statistically significant temporal trends of landscape ecological risk is observed. Variation of grassland, cultivated land, and bare land has more impacts on the landscape fragmentation than other land use and land cover. Risk grades of High and Medium/High mainly cover the lower normalized difference vegetation index (NDVI) region (0.1-0.4) at the edge of the Mu Us desert. The relationship between the landscape ecological risk and the NDVI reveals that the extent, continuity, pattern, and the type of vegetation cover are influencing factors for the landscape ecological risk from the view of the local habitats. The study results may be helpful for the vegetation restoration, land use planning, and sustainable development of the Yulin and other similar regions.