How effective soil depth modulates imagery vegetation dynamics in a semi-arid West African Savanna
Plinthosols; Rooting-zone limitation; Normalized Difference Vegetation Index; Soil-plant-water relations; Remote Sensing
In a semi-arid landscape, soil properties control water availability and plant productivity. Soils with plinthic or petroplinthic horizons impose a severe physical restriction on root growth, limiting the exploitable soil volume. However, the functional impact of this limitation and how it interacts with different land-use systems, which alter vegetation physiology, remains poorly quantified. This study aimed to analyze how the depth of the plinthic contact influences the seasonal and interannual vegetation dynamics across different land cover types (croplands, grasslands, and shrublands) in a West African landscape, Mali. We integrated field-measured effective soil depth with a five-year (2019–2023) Normalized Difference Vegetation Index (NDVI) time series derived from Sentinel-2 imagery, processed on the Google Earth Engine platform. A General Linear Model (GLM) was used to assess the main effects and interactions between soil depth, land cover, and season on vegetation vigor. Results revealed a significant interaction (p < 0.01) between soil effective depth and land cover. The positive relationship between increasing soil depth and NDVI was strong and clear in perennial systems (shrublands and grasslands), particularly during the dry season. In contrast, NDVI in croplands showed no significant relationship with soil depth, suggesting that annual management practices override the influence of this intrinsic soil property. The overall model explained 57% (R² adjusted = 0.569) of the NDVI variability. We conclude that the functional limitation imposed by a shallow plinthic contact is not uniform but is fundamentally dependent on the vegetation system. These findings provide a quantitative basis for site-specific land-use planning and demonstrate the potential of remote sensing to map functional edaphic constraints in data-scarce regions.