Modeling future urban expansion and land use transitions in Gorgan Region of Iran using spatial prediction techniques

Abstract

The continuous expansion of urban areas without considering all the influencing factors has led to numerous problems in the field of urbanization. However, urban development modeling serves as a valuable tool for explaining the interrelationships between the man-made and the natural environment. In this context, the current research aims to model the spatial development of human-made areas using Land Change Model (LCM) module and Markov Chain in Gorgan region. For this purpose, LULC maps for the years 1991, 2006, and 2021 were prepared using Landsat images and Support Vector Machine (SVM) model and the trend of changes during the study period was evaluated. Furthermore, potential transition modeling and predictions for the years 2036 and 2051 were performed using Multilayer Perceptron (MLP) and Markov Chain methods, respectively. The results indicate an accuracy of over 75% for the generated LULC maps in the study area. The increasing trend in residential, orchard, and rainfed agricultural LULC classes and the decreasing trend in forest, irrigated agricultural, and rangeland LULC classes during the research period are among the findings of this study. Additionally, urban development intensity during the period from 2006 to 2021 has been greater than the period from 1991 to 2006. The conversion of 1.83 square kilometers of irrigated agricultural land to residential and orchard areas in the Gorgan region area occurred during the time period from 1991 to 2006. In general, the transformation of LULC classes from irrigated agriculture, forest, and rangeland to residential, rainfed agriculture, and orchard classes in Gorgan region predominates. The results and the approach used in this research highlight the efficiency of satellite data and information in the preparation and monitoring of LULC changes and its application in urban development management.