Monitoring urban growth in Google Earth Engine from 1991 to 2021 and predicting in 2041 using CA-MARKOV and geometry: case study—Tehran

Urban growth leads to land use/land cover (LULC) changes, especially in developing countries. Knowledge of land transition, especially agricultural land, into the built-up region can help urban planners design and plan for better urban growth management. The purpose of this study is to monitor urban growth using Landsat time series images in Google Earth Engine (GEE). Using two algorithms, random forest (RF) and support vector machine (SVM), images were classified to generate the LULC maps. Land change modeler (LCM) analysis was performed to produce FROM-TO maps and check the direction to monitor the changes in Tehran in the last 30 years. Also, prediction maps were produced for 2027 and 2041 using the cellular automata-Markov (CA-Markov) model to predict the urban growth trends. By comparing the two algorithms, the RF algorithm had the best performance and the highest accuracy, which was chosen for this research. According to the classification results, it shows the overall accuracy and kappa coefficient of 0.94 and 0.90 for monitoring urban growth. Urban LULC has increased from 55,133 ha in 1991 to 69,923 ha in 2021, which shows that rapid urban growth has occurred in Tehran in the last 30 years. The results of LCM analysis showed that the major changes were from the bare land to built-up and the direction of development and expansion of Tehran was to the west and northwest. It is also predicted that 0.5% of non-urban areas will become urban by 2027 and 2% by 2041.