Towards Modeling Underground Soil Radon Gas Emanation

After cigarette smoking, Radon (222Rn) inhalation has been regarded the second most significant cause of lung cancer in the US and the world by the U.S. Environmental Protection Agency (EPA) and the World Health Organization (WHO). Radon is an inert noble radioactive gas produced by the Uranium (238U) decay series, specifically by the decay of Radium (226Ra) into Radon. Radon emanates naturally from the soil, particularly in regions rich in granite and slate. Thus far, there have been several controlled laboratory radioactive studies on the dependence of Radon gas emanation on atmospheric physical parameters such as humidity/moisture and temperature. However, there is no clear understanding of the exact emanation and transport process of Radon gas from the soil, and no model defines the dynamics. This paper presents our experience, learning, and modeling results from our extensive research on Radon emanation from soil. We consider a data-centric approach to modeling Radon emanation from the soil where we report and analyze data from a real-world underground Radon measurement testbed that we have installed on our site. In addition to presenting the testbed information, we report the results from evaluating ten of the most used and state-of-the-art supervised machine/deep learning time-series analysis models and unsupervised clustering analysis. We conclude this paper by summarizing our inferences and learning about our modeling results, the Radon modeling process, and the need for controlled lab experiments to validate and confirm on-field data-based analysis results. We plan to open the Radon testbed data used in this work to the public unon publication.