A Spatiotemporal Change Detection and Forest Cover Dynamics in the Fandoghlo Region (NW Iran)

In this research, in order to identify changes, Landsat-7 ETM+ and Landsat-8 OLI/TIRS L1TP images were used for 2003 (August) and 2023 (June), respectively, under supervised classification using Support Vector Machine (SVM) classification algorithm. Five main land use classes including agriculture, forest, rangeland, built-up, and water areas were extracted. The accuracy evaluation results for the classified images showed an overall accuracy of > 92% and a Kappa coefficient of > 0.89 for both years. During the period, agriculture and built-up areas increased, while forests, rangelands, and water bodies decreased. The highest rate of change (62.3 hectares per year) was observed in rangelands. The results of forest cover change maps after classification show a significant decrease of 8.5 percent in forest cover over a period of 20 years. The main reason for this change is the observed replacement of native forest cover with rangelands (16.9 %), built-up areas (4.3 %), and agricultural areas (0.8 %). Most of these land use changes are unplanned and mainly a result of human activities with a wide range of negative effects on natural ecosystems. Therefore, effective policies and land use management are needed to protect the environmental and aesthetic benefits of ecosystems, as well as to manage natural hazards caused by these changes. Apart from the assessment of land use changes, this research provides insight into the dynamics of forest cover, along with its causes and consequences, and in this way shows the hot spots of forest loss.