Predicting and validating the post-harvest soil nitrogen, phosphorus and potassium in direct seeded rice–wheat cropping sequence of northern India

Initial soil fertility affects fertilizer recommendations derived from soil test crop response. Therefore, predicting post-harvest soil available nitrogen (AN), phosphorus (AP) and potassium (AK) by this technique reduces the dependency on frequent analysis of post-harvest soil analysis. To investigate this, a field experiment was conducted at G.B. Pant University of Agriculture and Technology, Pantnagar from 2016 to 2018. The primary goal was to develop predictive equations for post-harvest soil test values (PHSTVs) for AN, AP and AK for direct-seeded rice–wheat cropping sequence (DSR-W-CS) utilizing an inductive and targeted yield model. The fractional factorial design was employed comprising twenty-four treatments across three fertility gradient strips for this study. The dependent variable was PHSTVs, while the independent variables included initial soil nutrient status, fertilizer nutrient, along with crop grain yield (GY) or total uptake (TU) of nutrient. Throughout the investigation, multiple linear regression (MLR) models’ were implicated to predicting AN, AP, and AK after individual crop and after whole cropping sequence also. Notably, the MLR models demonstrated improved predictive accuracy for individual crop compared to the entire DSR-W-CS. Additionally, these models exhibited satisfactory precision in predicting post-DSR-W-CS AN, AP, and AK. Following the DSR-wheat sequence, apparent balance of organic carbon (OC), AN, AP, and AK were − 3.28%, 8.67%, 13.19%, and 3.24%, respectively. These results highlight MLR-based models’ effectiveness in predicting PHSTVs in DSR-W-CS, thereby offering a practical approach to guide soil test crop response based fertilizer recommendations for individual crop or whole sequence.