Radon in Ground Water: A Study of the Measurement and Release of Waterborne Radon and Modeling of Radon Variation in Bedrock Wells

Naturally occurring radon gas (222Rn) exists in ground water and drinking water supplies. Research involving radon in ground water requires the ability to accurately measure radon in water. In the absence of a national program, an intercomparison study of laboratories was sanctioned by the State of Maine. The University of Maine research laboratory supplied each laboratory with water samples of various radon concentrations, served as the reference laboratory, and analyzed the results presented here. The external review of the University of Maine laboratory and agreement with some of the participating laboratories verifies its accuracy in measuring radon in water. A study of nine elementary schools in Maine examined the release of waterborne radon during water use. The release of radon into the kitchen air was measured to be greater than the EPA action level of 0.150 Bq L-1 (4 pCi L-1) in all schools but negligible concentrations of radon were found in adjacent classrooms. In two schools over a three-fold spatial radon variation was measured suggesting that multiple detectors are needed to accurately measure waterborne radon in air. During water use, the radon in water concentration was measured periodically and many of the schools showed an increase in the radon concentration by 200 Bq L-1 or more. To explore this effect, nine bedrock wells were studied in detail. Measurements of the ambient and purged radon profiles in the wells showed variations of radon concentration of samples within the well. The rock chips removed during well-drilling were analyzed for radionuclides in the 238U decay series. The 226Ra concentrations in the rock chips do not explain the measured vertical variation of dissolved radon. The vertical flow and fracture locations were previously determined by borehole logging to determine location of ground water inflow. A mathematical model of the ground-water flow into and through the well with radon as a tracer was tested. The model was successfully fit to data obtained from the wells that had a variation in radon concentration.