Analisis Spasial PH dan Konduktivitas Listrik (EC) di Lahan Pertanian dengan Menggunakan Perbedaan Metode Interpolasi Kriging pada Kecamatan Sumbersari

Abstract

Sumbersari District is one of the rice production buffer areas in Jember Regency, where productivity could potentially be increased through nutrient management based on local soil chemical conditions. This study aims to analyze the spatial distribution of soil pH and electrical conductivity (EC) values, alongside comparing the accuracy of Ordinary Kriging (OK) and Universal Kriging (UK) methods in mapping both parameters. A total of 33 sample points scattered across six sub-districts were analyzed in the laboratory, then interpolated using three semivariogram models, namely exponential, spherical, and gaussian, with accuracy evaluated through the Root Mean Square Error (RMSE) values from cross-validation. Laboratory results showed pH values ranging from 5.7 to 7.9, with 57.6 percent of the points categorized as neutral, while EC values ranged from 0.217 to 0.662 dS/m, classifying all points as non-saline. The spatial pattern of pH displayed a gradient from higher values in the northern part to lower values in the southern part, which relates to differences in fertilization patterns and crop rotation between regions. The UK gaussian model provided the best accuracy for pH interpolation with an RMSE of 0.463014, while the gaussian model in both OK and UK provided equivalent accuracy for EC with an RMSE of 0.10655, leading to OK being selected as the best model for EC on the grounds that the data lacks a spatial trend. The resulting distribution maps can serve as an initial reference for farmers and agricultural extension workers in determining priority zones for liming and fertilizer dosage adjustments in Sumbersari District.