Evaluating Spatial Interpolation Methods and Fuzzy C-Means Clustering for Effective Management Zone Delineation: A Case Study of Soil Health Management in India

Understanding soil variability is vital for implementing sustainable agricultural practices at the village level. The study was conducted at Bondi Madugula village, Kurnool district, Andhra Pradesh, India. A total of 538 geo-referenced soil samples (0 ~ 15 cm depth) were collected from agricultural fields. Soil parameters like pH, electrical conductivity (EC), organic carbon (OC), macronutrients (N, P, K, and S), and micronutrients (B, Mn, Fe, Cu, and Zn) were analyzed. The study found signif-icant variability in soil properties, with pH exhibiting the lowest coefficient of variation and S showing the highest. The logarithmic and box-cox transformations are employed to address data normality. Different methods, like Inverse Distance Weighting (IDW), Radial Basis Function (RBF), Ordinary Kriging (OK), and Empirical Bayesian Kriging (EBK) to map soil properties were Compared. Cross-validations and coefficient of determination results reveal that IDW (Power-1) performed best for pH and Cu, whereas RBF’s Inverse Multiquadratic (IMQ) was better for the remaining soil parameters. The soil maps and variability were analyzed using the best-fitted model. The study results identified deficiencies in N, OC, Zn, and Fe, with percentages of 99, 76, 51, and 32%, respectively. Fuzzy C-means cluster analyses were computed through a management zone analyst to delineate the study area into five zones, revealing heterogeneity identified in soil parameters except Zn, Fe, and Mn. The study results indicate that a single spatial interpolation model may not fit all soil variables with different types of soil datasets. The findings of this study provide valuable insights for local agricultural management strategies, promoting sustainable practices, and enhancing soil health to ensure food security in the region.