Assessment of variability and soil–plant nutrient interactions for crop yield prediction in semi-arid Alfisols using regression analysis

Precision nutrient management is essential for improving crop productivity in semi-arid Alfisols, where soil variability strongly influences yield. This study evaluated the relationships among soil properties, plant nutrient uptake namely nitrogen (N), phosphorus (P), potassium (K) and crop yields of finger millet, maize, field bean, tomato and brinjal under improved practices (IP) and farmers’ practices (FP) across the North and South Transects of Bengaluru. Multivariate regression models were developed to identify the key predictors of yield under both management systems. Finger millet in the North Transect showed strong associations with plant N, P and K, as well as with soil N and pH. The IP models significantly predicted finger millet yield using plant N, P and K, whereas FP models were primarily driven by plant K and soil K. In the South Transect, maize yield was mainly influenced by plant N and soil N, with IP models significant for P and K uptake in the North and N uptake in the South; FP models were significant only for N uptake in the South. Tomato yield under IP was significantly associated with P uptake, while field bean showed significant FP-based predictions for N uptake. Brinjal showed no significant relationships across nutrients and transects. Micronutrient-based models provided additional predictive power for field bean (North) and maize (South), whereas macronutrient-based soil models remained largely non-significant. Overall, the regression analysis highlighted nutrient-yield linkages and underscored the role of plant nutrient uptake as a reliable predictor for site-specific recommendations in semi-arid Alfisols.