Automatic extraction of highly risky coastal retreat zones using Google earth engine (GEE)

Coasts have been used for settlement, agriculture, transportation, recreation, defense and trade as areas that have attracted human beings for thousands of years. The importance of coasts for humanity is more than many other natural resources. For this reason, it has been an important research topic to know how these areas, which are of ecological, economic and anthropological importance, have changed with appropriate methods. Therefore, approximately 1435 km of coastline changes stretching between the provinces of Artvin and Kirklareli, which is the Black Sea coast of Turkiye, were investigated on the Google earth engine (GEE) platform using temporal Landsat multispectral satellite data. In this study, an approach that can automatically detect coastline changes using the GEE platform is proposed. With the proposed approach, coastline changes for different temporal periods between 1985 and 2021 were analyzed using automatically generated transects. All automatic operations such as image selection, masking, enhancement, coastline extraction and transect creation were coded in GEE cloud environment using Python script language. The results obtained revealed ten hotspots of maximum coastal retreat which are three erosion and seven accretion areas, in the southern Black Sea coastal zone. While the maximum amount of change in the erosion zones is about - 634 m, it has reached up to 1204 m for accretion zones. This study is an automatic GEE-based coastline extraction procedure which uses novel techniques for fast evaluation of coastal retreat, and it is expected to contribute decision-makers as a rapid coastline change evaluation tool for accurate and reliable decisions on coastal management and planning. This work will also serve as a basis for the future activities to be carried out in this field.