Multidimensional landscape changes contribute significantly to vegetation distribution: The evidences from the Luo River Watershed in the eastern part of the Qinling Mountains, China.

Assessment of vegetation characteristics plays an important role in monitoring ecosystem health and restoration status, and it is necessary to consider the response characteristics of vegetation assessment indicator (VAI) to multidimensional landscape structure. Therefore, this research constructed a framework for analyzing landscape structure from two dimensions of landscape element and landscape pattern, selected fractional vegetation cover (FVC) and net primary productivity (NPP) as VAI, determined the response degree and spatial correlation of VAI to landscape structure by using geographical detector and bivariate Moran's index, and finally obtained the change threshold of VAI based on the constraints of landscape pattern by building a cubic polynomial. FVC and NPP were significantly higher from 2001 to 2021, with the expansion of the high-value areas. Cropland and woodland were the dominant landscape elements. The spatial explanatory power of landscape element for VAI was better than that of landscape pattern characteristics, while the combined factors in landscape structure all explained VAI better than univariate. The response of FVC to landscape structure was more significant compared to NPP. The high value areas of FVC tended to occur in natural landscape elements. FVC showed spatial correlation and clustering characteristics with both PD and LPI. One or more thresholds existed in the response of FVC to changes in landscape pattern. The research results have important reference value in the optimization of landscape structure and vegetation restoration, and can provide practical guidance for the sustainable development of the ecosystem.