Modeling the Impact of Soil Health and Fertility on Crop Production

Abstract

Enhancing crop yields per land unit to satisfy future food and fiber demands accelerates soil nutrient depletion, highlighting the critical need for effective nutrient management strategies to restore soil fertility. Significant advancements in optimizing nutrient-use efficiency in agriculture require more precise assessments of plant-available nutrients within the root zone, increased crop responsiveness to applied nutrients, along with reduced off-site nutrient transfer. It involves data of agricultural yields on major crops taken through several growing seasons, using machine learning techniques for the analysis of soil fertility. Various algorithms were utilized, including Logistic Regression (LR), K-Neighbors Classifier (KNC), Random Forest Classifier (RFC), Boost Classifier, Decision Tree Classifier (DTC), and Gaussian Naive Bayes (GNB). The results show a significant correlation between improved soil health and rising agricultural yield, particularly in nutrient-deficient soil. According to the modelling, improving soil fertility can result in a production increase of up to 30%, emphasizing the need for environmentally-sustainable soil management practices. Of all the algorithms that were used for testing, the RFC and Boost Classifier both achieved the highest accuracy at 100%. These results highlight the importance of prioritizing soil health practices tailored to crop type for improving yields and sustainability in agriculture.