Resumen
Assessment of the impacts of landscape patterns on regional precipitation will help improve ecosystem management and strategies for adaption to global changes. This study aimed to identify the key landscape metrics that affect precipitation across three sub-climatic regions in Inner Mongolia, China, using 266 landscape metrics and daily precipitation data from 38 weather stations for 1995, 2000, 2005, and 2015. Pearson correlation, stepwise linear regression, and Redundancy analysis were used to identify the contributions of landscape patterns to local precipitation in each sub-climatic region. Three-year datasets were used for model development and a one-year data set was used for validation. It was found that the contribution of landscape patterns is higher than that of climatic variations in semi-arid or humid regions. The Core Area Coefficient of Variance (CACoV) of grasslands and Landscape Area (TLA) in non-irrigated croplands have a negative relationship with precipitation in arid regions. Further, the Total Core Area Index (TCAI) of grasslands has a negative correlation with precipitation, while the area proportion (C%LAND) in waters has a significant positive relationship with precipitation in semi-arid regions. Additionally, the Mean Core Area (MCA), Core Area (CA), and Core Area Standard Deviation (CASD) of grasslands and Total Core Area Index (TCAI) of waters are negatively related to precipitation in humid regions. Suitable land use configuration and composition, especially the proportion of grasslands and waters, should be considered in ecosystem management for alleviating the possible harmful effects due to climate change.