Examination of the Impact of Land use and Land cover Changes on Soil Erosion and Deposition Using the SIMWE Model and the Google Earth Engine Platform on the Archaeological Site of Amathous in Cyprus

Abstract

"The main aim of the present paper is to assess and simulate the spatial pattern of erosion, sediment transport, and deposition from an event-based rainstorm on the archaeological site of Amathous in Limassol district (Cyprus) through the application of the Simulation of Water Erosion (SIMWE) model in combination with open-source GRASS GIS. The SIMWE model implementation requires various spatial datasets, such as meteorological data (rainfall intensity), soil, land cover and geological data, the Digital Elevation Model (DEM) and multispectral satellite images of medium or high spatial resolution. The output is water depth, water discharge, erosion and deposition maps. The estimated soil erosion and deposition at the archaeological site varied from 0.0054 kg/m2s to 0.0075 kg/m2s. The highest amount of both erosion and deposition are expected in areas with high concentrated sediment flow. High sediment flow rates were observed in the middle and northern parts of Amathous and varying flow rates in adjacent areas. The approach employed in this paper is based on using the land use and land cover changes as a dynamic indicator of the erosion process. Land cover changes were determined utilising and comparing two machine learning algorithms applied to Landsat data from the archive from the Google Earth Engine platform. The results show that Random Forest (RF) achieves a better performance in detecting land-use changes than CART. "