Tracking seasonal coastal dynamics of Ağyatan wetland using object-based Classification and Regression Tree

Authors : Mustafa Hayri Kesikoğlu, Tolga Kaynak

Pages : 308-317

Doi:10.28948/ngumuh.1559034

View : 65 | Download : 112

Publication Date : 2025-01-15

Article Type : Research Paper

Abstract :In recent years, the management of water and wetlands has become increasingly important. Ağyatan Wetland is a natural lake formed by the Ceyhan River, located west of the river. The area around the wetland is home to various sea and land creatures, as well as endemic plants and endangered birds. This study aimed to analyze the seasonal changes occurring in the Ağyatan wetland in 2016. Landsat 8 images from four seasons of 2016 were utilized to identify these changes. The study areas were classified into six categories: lake, sea, cultivated agricultural land, barren land, building area, and water channel. Thematic maps for the four seasons were generated using the object-based Classification and Regression Tree (CART) method. Additionally, change detection analyses were conducted using the post-classification comparison method. The study revealed a decrease in the lake area from winter to spring (0.8919%), followed by an increase from spring to summer (0.3627%) and summer to autumn (0.1953%) in 2016. This fluctuation was attributed to groundwater, river water, and melting snow.
Keywords : Değişim belirleme, nesne tabanlı sınıflandırma, CART, sulak alan, Landsat 8, uydu görüntüsü