| Radon is an important natural tracer for certain aquatic environmental studies. New methods for rapidly determining submarine groundwater discharge (SGD) into surface waters in the coastal zone rely on the accuracy of the air/water partitioning coefficient for doing radon mapping of coastal waters. The original research that determined the radon partitioning coefficient was conducted in 1916 and more recent published analysis of these data only accounted for temperature changes in fresh water.;My research was designed to: 1) determine the partition coefficient by accurately measuring the gas and water phase concentrations of 222 Rn when the radon is in equilibrium between both phases; 2) minimize the error due to multiple analytical system calibrations; and 3) calculate the partition coefficient at different temperatures and salinities.;I focused on 2 different experimental setups, one designed to directly measure the radon concentration of the gas phase and the dissolved concentration in the liquid phase (known as the direct measurement or volume independent method), and the other was hoped to be a rapid measurement system that only required a continuous measurement of the gas phase to determine the partitioning coefficient at different temperatures (the indirect measurement or volume dependent method).;I combined my research results with parallel research experiments being conducted in Leipzig, Germany and together we redefined the air/water partitioning coefficient of radon gas with respect to both salinity and temperature. By analyzing our combined results, we developed a new set parameters a1 -- b3 for use in the Weiss Equation, which could then be used to determine the partitioning coefficient of radon at different salinities and temperatures. Our results showed that if salinity effects are neglected in warm, relatively low salinity waters (e.g., coastal waters from the Gulf of Mexico), there was a slight overestimation of the radon concentration (up to 10%), but in colder, higher salinity waters, this overestimation could exceed 20%. |
