Developing a methodology for modelling land use change in space and time

Abstract

The complex, dynamic and non-linear characteristics of land use change (LUC) modelling necessitate the development of advanced models and methods. In this research, a novel and stepwise approach is developed for modelling integrated LUC at the regional level based on land suitability, land demand and land use conversion rules. The aim of this model is to use different models (i.e. geographic information systems, remote sensing, fuzzy logic, cellular automata and multi-criteria decision analysis) in a logical way to improve model development, model evaluation and prediction of LUC. The proposed model was implemented in ‘Borkhar and Meymeh’ township, Iran, for two periods (1986–1998 and 1998–2005). Results for producer accuracy, user accuracy, overall accuracy and figure of merit were 0.345, 0.357, 0.351, 0.234 respectively in the first period (1986–1998) and 0.302, 0.287, 0.0.294, 0.202 respectively in the second period (1998–2005). Our results suggest that the developed model is accurate, when compared with the results of existing models.