Exploring crop yield variability under different land management practices with spectral vegetation indices in the Ethiopian Blue Nile basin

An accurate crop yield forecast at a reasonable cost is required for increasing production and securing food. Therefore, spectral reflectance was measured with a spectroradiometer (λ 350–2500 nm) to predict grain yield (GY) and aboveground biomass yield (ABY) of teff and finger millet in Ethiopia. We calculated six spectral vegetation indices (SVIs): enhanced vegetation index (EVI), normalized difference VI (NDVI), soil-adjusted VI (SAVI), green (GNDVI), green chlorophyll VI (GCVI), and simple ratio (SR). Linear regression models fitted validation data with coefficient of determination (R²) of 0.73, and root mean square error (RMSE) of 0.11 for teff GY, R² of 0.45 and RMSE of 0.52 for teff ABY, R² of 0.92 and RMSE of 0.22 for finger millet GY, and R² of 0.90 and RMSE of 0.24 for finger millet ABY. The study demonstrates the potential of spectroradiometer-derived SVIs to predict finger millet and teff production in the catchment.