Desertification Assessment in the Hodna Basin, Algeria using Desertification Degree Index and Geospatial Techniques

Abstract

This study assesses desertification in the Hodna Basin, Algeria, using remote sensing and Geographic Information Systems (GIS) through the development of a Desertification Degree Index (DDI) based on feature-space analysis. Landsat 9 imagery was used to derive four spectral indices: the Normalized Difference Vegetation Index (NDVI), surface albedo, the Modified Soil-Adjusted Vegetation Index (MSAVI), and the Topsoil Grain Size Index (TGSI). Five feature-space combinations were evaluated, with the Albedo-NDVI and TGSI-NDVI models exhibiting the strongest statistically significant negative correlations. These two models were subsequently used for DDI generation and comparison. The Albedo-NDVI model achieved the highest performance, with an Overall Accuracy of 91.9% and a Kappa coefficient of 0.89, substantially outperforming the TGSI-NDVI model. The resulting desertification map indicates that approximately 65% of the Hodna Basin is affected by high to severe desertification, particularly in the central and southern regions, where prolonged aridity and intensive human activities have accelerated land degradation. In contrast, the northern part of the basin exhibits relatively low desertification levels owing to denser vegetation cover and more favorable environmental conditions. These findings demonstrate the robustness of the Albedo-NDVI feature-space model for desertification assessment in arid and semi-arid environments and highlight the value of integrating spectral indices with statistical feature-space analysis. The proposed approach provides a reliable and objective framework for large-scale desertification mapping and offers valuable decision support for soil conservation, ecosystem restoration, and sustainable land management in vulnerable dryland regions.