Abstract: Spatial patterning of rural settlements can reflect how land resources are spatially segregated in agrarian societies. In this study, we conducted a systematic sampling at a national scale and used highresolution remotelysensed images to extract information on the rural settlements in eastern mainland China. We used spatial analyses and multivariate regression to quantitatively analyze structural attributes of rural settlements in relation to a set of environmental and socioeconomic factors at a large spatial scale. Results showed that the macroenvironmental factors including climate and topography can well explain the geographic patterns of settlement density, betweensettlement nearest neighbor distance and distribution of Thiessen polygons generated based on settlement points. Settlement density is mainly constrained by topography. In areas with higher waterenergy flux and complex topographic relief, the betweensettlement spatial segregation of land resources tends to present smaller spatial units, while the heterogeneity of such segregation tends to be elevated. In general, spatial structure of rural settlements is co-influenced by land resource supply and topographic constraint, reflecting a macroscale humannature relationship. Our work can help to better understand spatial characteristics of rural settlements at broad scales, and have useful implications to the current practice of ‘New Countryside Constructions’ in China.

Key words: biomass conversion and expansion factors., Cubist model, random forest, boosted regression trees, national forest inventory