Gis-based simulation of land use change

This paper describes an application of ArcGIS to simulate land use changes in Lagos, Nigeria, using both Ordinary Least Squares (OLS) regression and Geographically Weighted Regression (GWR), over three epochs - 1963-1978, 1978-1984 and 1984-2000. Twelve salient, causal factors thought to be related to urban land use change in Lagos were used, such as distance to water, distance to residential structures, income potential and population potential. OLS was used to establish the regression coefficients, gauge the significance of each explanatory variable and estimate conformity with linear regression criteria. GWR was then used to simulate the urban form based on the results of the OLS model. For the three epochs the respective Kappa statistics for the simulated maps were 0.8858, 0.8366 and 0.8812, indicating an almost perfect agreement with the data of 1978, 1984 and 2000.