Remote Sensing for Within-Plot Soil Variability Assessment Using NDVI Dispersion Metrics

While within-plot soil variability strongly influences crop development and agronomic management efficiency, practical and transferable methods for characterizing this variability remain limited. This study proposes a remote sensing framework to assess soil-driven crop heterogeneity using Normalized Difference Vegetation Index (NVDI)-derived parameters and phenological analysis. Crop dynamics were monitored using Sentinel-2 imagery to identify key phenological stages (start of season, maximum development, and senescence), while high-resolution PlanetScope imagery was used to analyze within-plot spatial variability. NDVI dispersion metrics were used as proxies for spatial variability in crop development across soil units and under homogeneous management conditions to isolate soil effects. Results showed that NDVI standard deviation effectively captured spatial heterogeneity in crop development, revealing greater variability in Glacis Slope soils compared with the more homogeneous Platform unit. Crop senescence emerged as the most informative stage for detecting soil-driven variability, while emergence patterns were influenced by sowing dates and early establishment conditions. Analyses under homogeneous management confirmed that soil properties controlling soil moisture dynamics govern spatial variability in crop response. The proposed methodology provides a robust, low-cost, and transferable approach for identifying within-plot variability and supports precision agriculture by enabling site-specific management decisions based on satellite-derived indicators.