Integration of logistic regression and markov chain analysis in modelling and predicting future urban growth in Kaduna,Nigeria

Abstract

Land use changes in developing countries generally occur in the form of new building construction in the peripheral areas of cities mostly for residential purposes. This fast pace of urban development has thrust to the fore a number of research initiatives that make use of new applications for modeling such phenomenon. In this study, urban expansion is examined in Kaduna, Nigeria, using a hybrid model that combines Cellular Automata (CA) techniques and Markov chain analysis with logistic regression. Land use change between 1990 and 2001 was analyzed and the results showed huge conversion of green areas and open lands to built-up areas. Environmental variables dealing with urban expansion were evaluated to determine their influence on urban expansion and operationalized to develop transition potential sub-models. A transition potential map is then derived from the transition potential sub-model which was employed to simulate future urban growth. The model was subsequently evaluated and validated using ROC statistics and the results were found to be acceptable.