Inverse Distance Weighting as an alternative interpolation method to create radiometric maps of natural radionuclide concentrations using QGIS

Abstract. A previous study by the authors illustrated the distribution of naturally occurring radionuclides, i.e., potassium (K⁴⁰), thorium (Th²³²), and uranium (U²³⁸), in the sediment of the Berg River estuary. The study also described the Delta Underwater Gamma detection System (DUGS) and its usage to measure the concentrations of these natural radionuclides. It also proposed a novel radiometric mapping technique with QGIS and, more importantly, highlighted the geospatial process through Kernel Density Estimation (KDE) to create radiometric maps. The present study used the same data and a similar design but proposed the use of the Inverse Distance Weighted (IDW) interpolation method in QGIS to display natural radionuclide concentrations. The radiometric maps created in the previous study using the KDE technique created smooth and visually attractive maps. However, the IDW method is an exact interpolation method that predicts a value at a sampling location that is identical to the observed value. This is a requirement for the analysis of natural radionuclide concentrations in sediments. However, the effectiveness of the interpolation methods was evaluated using SPSS statistics software. First, probability-probability (P-P) plots were produced for each interpolation method. Secondly, descriptive regression statistics, including 'goodness-of-fit', Analysis-of-Variance (ANOVA), coefficients, and residuals were evaluated for both the IDW and KDE interpolation methods. This was done to assess which method was more effective for calculating the radionuclide concentrations (actual vs predicted values) in the Berg River area. The results showed that both methods experienced problems to predict unknown values. However, IDW consistently performed better than KDE across most of the interpolation tests. Natural radionuclides are useful predictors to track sedimentation and the results of this paper can serve as a benchmark for future work in tidal and non-tidal coastal environments.