A Comparison of Geospatial Interpolation Techniques for Gold Anomalies in Wassa - Amenfi District, Southwest Ghana

Abstract

In spatial interpolation, the Inverse-Distance Weight (IDW) and Kriging methods are two of the most widely used deterministic and geostatistical models. In deterministic techniques, interpolation is performed using mathematical functions, whereas geostatistics uses statistical and mathematical methods to create surfaces and assess the uncertainty of predictions. This paper aims to compare the Kriging and IDW spatial interpolation methods in the Wassa-Amenfi District using gold concentration data collected by the Ghana Geological and Survey Authority. The prediction accuracy of the spatial interpolation models was improved by reducing the complexity of the gold concentration dataset using the logarithm and Box-Cox data transformation method. The performance of these methods is measured using the Root Mean Square Error (RME) and coefficient of determination (. Kriging model of 0.92 and 0.81 in both Box-Cox and logarithm transformed Au datasets were greater than IDW  squares of 0.78 and 0.87. As a result, the kriging method outperformed the IDW and was the best interpolation technique for mapping and predicting gold concentration anomalies in the Wassa-Amenfi District, located in Southwest Ghana.