Two Decades of Desertification Dynamics in Northeastern Algeria: Insights from ANN-MLP Modeling

Aures region is one of the most affected area by the phenomenon of desertification. The main aim of the present study is to assess the desertification degree during three periods: 2003, 2013, and 2023 across the Aures region by integrating remote sensing and GIS techniques and the Artificial Neural Network—Multilayer Perceptron (ANN–MLP) approach. Five factors were used as indicators of desertification degree namely: Modified Soil Adjusted Vegetation Index (MSAVI), Normalized Difference Moisture Index (NDMI), Bare soil index (BSI), Topsoil Grain Size Index (TGSI), and Land use land cover (LULC). Results show that desertification degree changes over space and time. From 2003 to 2023; very low degree class increased from 642.4 to 1003.8 km2, while the class of low degree decreased from 1305.2 to 1037.4 km2, then medium degree class increased from 335.7 to 354.5 km2after high degree class decreased from 2830.1 to 2661.7 km2, finally very high degree class increased from 99.7 to 155.6 km2. Kappa coefficient and Root Mean Square (RMS) values exceed 0.60 and 0.20, respectively indicating a good performance for the current model. The results obtained constitute a valuable support in the decision-making process, encouraging the application of protection strategies and combating desertification.