Evaluating the impact of spatial dependence on land use change analysis in New Castle

Abstract

Sustainability concerns have aroused great interest among policy makers in the interrelated land use change systems. Considerable attention has been devoted to study the methods in which land use change affects spatial patterns. Less specific attention has been paid to analyze the effects of spatial dependence inherent in land use change. Learning how the spatial dependence impacts the accuracy of land use change analysis is thus of our great interesting. This paper compares several widely used spatial statistic regressive methods to study the impact of spatial dependence on land use change. Specifically, logit model and the autologit model coupled with GIS are used for analyzing the impacts of spatial dependent factors such as population density, slope and surrounding land use, etc. on land use change. In addition, the impact of spatial neighbor effects with logit model for the influence of location and change is also evaluated. Case studies of land use change patterns in New Castle County, Delaware has been implemented, and the results demonstrates that spatial neighbor effects provide an illustrative description of spatial dependencies between data attributes and land use change patterns, and better support explaining the complex processes of urban development and change. © 2014 Taylor & Francis Group.