Evaluation of Geographic Information Systems-Based Spatial Interpolation Methods Using Ohio Indoor Radon Data

Abstract

This paper evaluates the performance of six different Geographic Information System based interpolation methods: inverse distance weighting (IDW), radial basis function (RBF), global polynomial interpolation, local polyno- mial interpolation, kriging, and cokriging, using the Ohio homes database developed between 1987 and 2011. The best performing interpolation method to be used in the prediction of radon gas concentrations in the unmeasured areas of Ohio, USA was determined by validating the model predictions with operational performance measures. Additionally, this study performed a zip code level-based analysis that provided a complete picture of the radon gas concentration distribution in Ohio. The RBF method was identified to be the best performing method. While the RBF method performed significantly better than the IDW, it was statistically similar to the other interpolation methods. The RBF predicted radon gas concentration results indicated a significant increase in the number of zip codes that exceeded the United States Environmental Protec- tion Agency and the World Health Organization action limits, thereby, indicating the need to mitigate the Ohio radon gas concentrations to safe levels in order to reduce the health effects. The approach demonstrated in this paper can be applied to other radon-affected areas around the world.