NDVI and NDMI: what satellite indices tell a farmer

Full article available in Uzbek and Russian.

Satellite indices promise a field-wide "health map" from a single image, but the details matter. NDVI = (NIR − Red)/(NIR + Red), derived from Sentinel-2 bands B8 (NIR) and B4 (Red) at 10m resolution, captures chlorophyll density and vegetation vigor — roughly 0.1-0.2 for bare soil, 0.2-0.4 for sparse vegetation, 0.6-0.9 for dense healthy canopy. NDMI = (NIR − SWIR)/(NIR + SWIR), using B11 (SWIR, 20m), captures canopy water content instead, and often flags moisture stress before NDVI does.

Sentinel-2's ~5-day nominal revisit at the equator sounds frequent, but cloud cover regularly stretches real gaps to 2-3 weeks. Other limits: soil background signal distorts NDVI during early growth or sparse canopy; NDVI saturates around 0.8-0.9 in high-biomass fields, losing sensitivity to further density changes; and — critically — neither index says anything about root-zone soil moisture, only surface/canopy conditions.

The practical answer isn't satellite-vs-ground-sensors, it's both: satellites deliver spatial coverage across an entire field or region to flag problem zones, while ground IoT sensors deliver frequent, root-zone-depth truth at fixed points. Combined, they cover space, time, and depth together — something neither source achieves alone.

Figures are generalized and illustrative; verify against official Sentinel-2/Copernicus documentation for project-specific decisions.