A spatial logistic regression model for simulating land use patterns: A case study of the shiraz metropolitan area of Iran

Abstract

Changes in global land cover and land use are occurring at a rate, magnitude, and spatial extent unprecedented in human history. Land use and land cover change is one of the most important components and major causes to global environmental changes. This paper presents an urban expansion model which utilizes Logistic Regression (LR) as a means to simulate and predict urban expansion pattern. LR is used as the core algorithm in this model and remote sensed image with arbitrary time interval and environmental variables have been employed in geospatial information system. Socio-economic and environmental variables were used as inputs while urban and non-urban areas were considered as outputs for LR model. Evaluation of the model was performed with Relative Operating Characteristic (ROC). A case study of urban expansion in Shiraz Metropolitan Area (SMA) is presented to simulate land use change. This paper presents a version of LR model which is parameterized for SMA and explores how factors such as road, building area, service centre, green space, elevation and slope can influence urban expansion. The area under the operating characteristic curve produced an accuracy of 68%. Having model parameters with specific time interval and assuming the existence of the same rate of urban expansion, we used the LR model presented here to generate locations of future urban expansion. © 2010 Springer Science+Business Media B.V.