Estimasi Spasial Kandungan Sulfur Menggunakan Metode Ordinary Kriging: Studi Simulasi di Wilayah Indonesia

Abstract

Sulfur content in coal is a critical parameter that influences combustion quality and environmental impact, particularly due to sulfur dioxide emissions that contribute to acid rain. Therefore, spatial estimation of sulfur concentration is a strategic step in the management of coal resources. This study aims to evaluate the effectiveness of the Ordinary Kriging method in mapping sulfur content in coal based on simulated data representing the Indonesian region. A total of 100 data points were used to construct an experimental semivariogram, which was then compared with three theoretical models: Spherical, Exponential, and Gaussian. The analysis revealed that the Exponential model performed best, with the lowest Root Mean Square Error (RMSE) of 0.01982. The estimation was applied to 10,000 prediction points, resulting in sulfur content values ranging from 0.3143% to 1.3649%, with an average of 0.8306%. Model validation using cross-validation yielded an RMSE of 0.1806 and a Mean Absolute Error (MAE) of 0.143. Moreover, the Moran’s I value of 0.41 (p-value < 0.01) indicates a significant positive spatial autocorrelation. These findings demonstrate that Ordinary Kriging not only provides accurate estimates but also effectively captures the spatial structure of sulfur content, making it a valuable tool for supporting more efficient and sustainable exploration decision-making.