Monitoring Soil Degradation Using Sentinel-2 Imagery and Statistical Analysis of Spectral Indices in a Semi-Arid Watershed of the Moroccan High AtlasMonitoring Soil Degradation in Semi-Arid Contexts Using Sentinel-2 Imagery and Statistical Analysis of Spectral Indices

Oussama Nait-Taleb^1*Ismail Karaoui¹Kamal Abdelrahman²Mustapha NAMOUS^1,3Samira Krimissa¹Maryem ISMAILI^1*Mohammed S Fnais²sana elomari¹Jaouad El Atiq⁴Insaf Ouchkir¹Abdenbi Elaloui¹

• ¹Data Science for Sustainable Earth Laboratory (Data4Eart), benimellal, Morocco
• ²Department of Geology, College of Science, King Saud University, Riyadh, Saudi Arabia
• ³Faculté des Arts et des Sciences (FAFS), Université de Saint-Boniface, 200, Avenue de la Cathédrale, Winnipeg, MB R2H 0H7, Canada, Winnipeg, Canada
• ⁴Geomatics, Georesources and Environment Laboratory, Faculty of Sciences and Techniques, Sultan Moulay Slimane University,x, Beni Mellal, Morocco

The existence of serious water erosion problems in different parts of a watershed is often evidenced by the presence of high levels of suspended sediment in watercourses. The indirect assessment of erosion through the measurement of suspended sediments transported to catchment outlets serves as a robust indicator of the environmental impact of agricultural practices. The aim of this study is to propose a model for assessing the risk of soil degradation in the upstream Tassaoute watershed (in the Moroccan High Atlas). The methodology is based on the statistical analysis of spectral indices derived from Sentinel-2A satellite images acquired during the year 2021, including four vegetation indices and nine soil indices. These indices are aggregated to form a composite image (the independent variable), which is then subjected to regression analysis against the individual indices (the dependent variable) to determine correlation coefficients and coefficients of determination. Principal Component Analysis (PCA) is then used to condense the information from all the spectral indices, providing factorial coordinates and facilitating the identification of positive and negative correlations. The principal component captures soil-related information, while the secondary component focuses on vegetation characteristics. The final predictive model is developed by assigning weights to each index based on its coefficient of determination and the coordinates of the factors. This approach produces a quantitative map delineating four categories of soil potentially at risk of degradation. The results show that incorporating the spectral bands of Sentinel-2A's C-MSI sensor into the calculation considerably improves accuracy and provides an accurate representation of ground reality. The existence of severe water erosion issues in different parts of a watershed often evidenced by the presence of elevated levels of suspended sediments in watercourses. The indirect evaluation of erosion Formatted: Not Highlight