Geostatistical analysis of arsenic contamination in soil and comparison of interpolation techniques in Nadia district of Bengal, India

Abstract

The contamination of soil and water with arsenic directly or indirectly affects millions of people, particularly in Southeast Asia. Efficiently managing contaminated sites cost-effectively requires an understanding of the spatial distribution of contamination in soil. In this study, different interpolation methods, including Ordinary Kriging (OK), Inverse Distance Weighted (IDW), Radial Basis Function (RBF), and Empirical Bayesian Kriging (EBK), were evaluated in the Bengal region to determine their effectiveness in predicting the Olsen extractable As content in the soil. The study found that the mean Olsen extractable content in soil was 1.45 mg kg-1 , with a range of 0.48 to 3.57 mg kg-1 . Geostatistical analysis showed that the northern side of Nadia had relatively high contamination, while the southern side had relatively lower contamination. The Root Mean Square Error (RMSE) values of the different interpolation methods ranged from 0.52 to 0.54, with corresponding mean cross-validation (CV) values ranging from -0.005 to 0.008. The predicted minimum and maximum values of as-in soil were in close agreement with the measured values for IDW interpolation, followed by OK, RBF, and EBK. The study found that IDW consistently provided the most precise predictions of pollution in the soil throughout space. These findings have significant implications for managing contamination in the Nadia West Bengal and other regions facing similar challenges.