Multi-Scale Geographically Weighted Regression Modeling of Urban and Rural Construction Land Fragmentation-A Case Study of the Yangtze River Delta Region

Landscape fragmentation, which can be natural or man-made, leads to decreasing species diversity and environmental degradation. The analysis of landscape fragmentation helps people understand the relationship between land and ecosystems in order to achieve sustainable development. Using the data from 306 counties in the Yangtze River Delta region, we calculated the landscape pattern index and constructed an evaluation index for urban and rural construction land fragmentation. We used a multi-scale geographically weighted regression model to analyze the determinants of land fragmentation and their spatial heterogeneity. We found that land fragmentation has significant spatial autocorrelation. The spatial distribution of the fragmentation is "ladder-like" and higher in the southwest and lower in the northeast. The determinants' spatial scales from largest to smallest are average yearly precipitation, population (PS), temperature (TEM), gross domestic product (GDP) per capita (GPC), GDP per unit area (GPUA), proportion of secondary production (PSP), slope (SLO), and elevation (ELE). The mean contributions of the regression coefficient from highest to lowest were PER, ELE, TEM, location, GPC, SLO, GPUA, PSP, and PS. The evaluation index could help policymakers make urbanization plans and implement effective urban and rural spatial structures.