Examining the temporal and spatial change of current land cover types in Demre District using machine learning

The study, conducted to analyze the temporal and spatial changes in current land use types in Demre district of Antalya Province and to establish a foundation for future land management studies, was supported by multispectral satellite images obtained from the Landsat 5- Thematic Mapper (TM) and Landsat 8- Operational Land Imager (OLI) remote sensing satellites. Land use maps showing the spatial distribution of land cover changes were prepared using composite images compiled for the years 2004, 2014, and 2024. The Support Vector Machines (SVM) algorithm was used as the classifier model. High classification performance was achieved for all three images (kappa = 0.90, 0.89, 0.87, overall accuracy = 90.7%, 90.3%, 87.9%). Following the classification process, land use change maps were created for each decade, and statistical analyses related to land cover change were conducted. Over the past 20 years, according to the land use types in Demre district, quarry and mining areas (↑227.49%) and settlement and greenhouse areas (↑72.88%) have increased significantly, while scrubland (↓41.21%), agricultural land (↓9.45%), and forest areas (↓8.89%) have decreased significantly. In addition, sparse maquis areas, dune areas, and water surfaces areas have declined, indicating a situation that is detrimental to the region's natural areas. In conclusion, the study provides an overview of how land use types have changed in the region and reveals the current state of land use preferences in the district.