Preparing Distribution Maps of Available Phosphorus in Soils Using Various GIS-based Spatial Interpolation Methods

Abstract

The study, was conducted from January, 2022 to December, 2024 at Jawaharlal Nehru Krishi Vishwa Vidhyalaya, Jabalpur, Madhya Pradesh (482004), India. We have collected two thousand two hundred sixteen (2216) Global Positioning System (GPS) based soil samples from depth of 0–15 cm from farmer's field of the Kymore Plateau and Satpura Hills zone of Madhya Pradesh, India. Laboratory analysis showed that the available P content ranged from 1.11 to 117.7 kg ha⁻¹ (mean: 10.64 kg ha⁻¹). Significant correlations were found between pH, electrical conductivity (EC), organic carbon (OC), and P (r = 0.073). The study aimed to identify the most suitable interpolation method for mapping available P in soils, employing three geo-statistical (Ordinary Kriging, Simple Kriging, and Empirical Bayesian Kriging) and three deterministic methods (Radial Basis Function, Local Polynomial Interpolation, and Inverse Distance Weighting), as well as two barrier-based methods (Kernel Smoothing and Diffusion Kernel). Geo-statistical results indicated OK (Box-Cox) spatial interpolation method for estimating available P distribution in soils that followed exponential model with ranges of 3652.22 meters, nugget values of 16.90 and a N/S ratio of 0.47 which showed moderate spatial dependency. Among the methods tested, Empirical Bayesian Kriging (EBK) provided the most accurate estimates of Olsen P distribution, followed by Ordinary Kriging (OK) and Simple Kriging with Box-Cox transformation.