How effective soil depth modulatesimagery vegetation dynamicsin a semi-arid west african savanna

Abstract

Soil properties govern water availability and plant productivity, particularly in semi-arid environments. Soils with petroplinthic horizons can impose severe physical constraints on root growth, limiting the exploitable soil volume and influencing water drainage. However, the functional impact of this constraint and its interaction with different land-use systems remains poorly quantified. This study evaluated how effective soil depth shapes the seasonal and interannual dynamics of vegetation, using the NDVI as an indicator across land-cover types (cropland, grassland, and shrubland) in Bandiagara II, a plinthic landscape typical of the semi-arid Sikasso region in southern Mali. Field measurements of soil depth were integrated with a five-year (2019–2023) NDVI time series derived from Sentinel-2 imagery processed on the GEE platform. A GLM revealed a significant interaction (p < 0.01) between soil depth and land cover. In grasslands, the relationship was positive and significant (slope = +0.0032) and more pronounced in the dry season, whereas the trend in shrublands did not reach statistical significance. In contrast, NDVI in croplands showed no significant relationship with soil depth, suggesting that annual management practices can override the influence of this intrinsic edaphic property. The overall model explained 57% of NDVI variability (adjusted R2 = 0.569). The functional limitation imposed by a shallow petroplinthic contact is not uniform but depends fundamentally on vegetation type. These results provide a quantitative basis for local land-use planning and demonstrate the potential of DSM approaches integrating remote sensing to map functional edaphic constraints in data-scarce regions.