Abstract: A rational assessment of landscape ecological risk (LER) in the ecological restoration region of loess hills in northern Shaanxi is crucial for optimizing regional landscape pattern and maintaining ecosystem function. Landscape pattern index and GIS spatial analysis were used to demonstrate the level and spatial-temporal evolution pattern of LER using data from land use of Yan’an from 2000 to 2020. The results showed that: (1) The landscape ecological risk index (ERI) was in the range of 0-0.449, with low, medium-low, and medium risk areas dominating risk grades that were patchily dispersed in the southwest, along the Yellow River, and northwest. The high-risk area was centered in the Baota district. In 2020, the area of high and medium-high risk regions increased by 1.27×10⁴ hm² and 25.25×10⁴ hm² compared to 2000, respectively, with a notable expanding tendency of LER. (2) From 2000 to 2020, Moran’s I of LER was positive, with a significant geographic correlation. The main spatial aggregation modes of LER were “high-high” and “low-low”, but the spatial aggregation initially strengthened and then declined, and the local difference gradually became prominent. (3) The LER hotspots were located in the Baota region, where built-up construction land and farmland predominated, with substantial system disturbance and fractured and split landscape, indicating a high-value agglomeration pattern. The LER cold spots were primarily located in Huanglong, Huangling, and Fuxian counties, where the high regional altitude and large patch area of forest and grassland were responsible for the low intensity of human disturbance and the low degree of landscape separation and fragmentation, resulting in a low-value agglomeration pattern. In the past 20 years, the area of the cold-hot spot regions changed dramatically, and its distribution pattern shifted frequently. Consequently, LER in Yan’an was relatively low, with variations of risk levels, spatial agglomeration patterns, and cold-hot zones. Land use structure should be planned to promote the security and stability of landscape.

Key words: landscape pattern, ecological risk, spatial autocorrelation, local hotspot analysis, Yan’an.