The Impact of (DEM) Accuracy and (LC/LU) Resolution on the Watersheds Areas

Abstract

Land Cover/Land Use (LC/LU) and Digital Elevation Model (DEM) are the main inputs for watershed modelling. Recently, (DEM) and (LC/LU) are freely available online open-source products in varied accuracies, spatial and spectral resolutions as a result of several remote sensing platforms. Therefore, it is very imported to determine which one is the optimum for modeling the watersheds in the selected study area, which is represented by five valley watersheds of diverse characteristics, located on east and west sides of the Mosul dam reservoir, Nineveh province, Iraq.In this research, the different accuracy of 30m resolution DEMs with Satellites (Copernicus (GLO-30), ASTER, and SRTM), besides to supervised classification (Support vector machines (SVM) classifier) results (LC/LU) main layers (green land, bare soil, urban areas, and water) of different spatial and spectral resolutions images with Satellites (10m Sentinel-2, 30m Landsat-8, and resampled 15m from 30m Landsat-8) are examined by using the techniques of Remote Sensing (RS) and Geographic Information System (GIS). Analysis of the results led to find that Copernicus DEM (GLO-30) 30 m spatial resolution is the most accurate and optimum DEM in the research study area with 1.1615 m vertical accuracy and 2.276 m at 95% confidence level. The optimum most accurate image for (LC/LU) thematic map production in the selected area is (Sentinel-2 l0 m) satellite at overall Classification accuracy (97%) and the overall kappa statistics (95%). The optimum remote sensing data (sentinel-2 image and Copernicus DEM) are maped, 3D simulated and analyzed by GIS to calculate the areas of (LC/LU) main layers and to find that the considerable part of the selected watersheds is the area of green lands 404.32676 km2 was about 54.396% of the total area in the study, and these areas are vary depending on many factors such as remote sensing data, precipitation, cultivation, season, and human activities.