Assessment of Groundwater Discharge to Lake Barco via Radon Tracing

Groundwater-surface water interactions in lakes have been gaining attention in recent years as scientists have recognized the potential significance that groundwater has as a source of nutrients and contaminants to aquatic ecosystems. Such interactions need to be understood in order for us to protect important ecosystems and quantify nutrient loading into lakes. This project set out to test the idea that good estimates could be made of groundwater inputs into Florida lakes using a simple geochemical tracer technique. We hypothesize that a relatively small number of measurements would be sufficient to provide a reasonably good (a factor of 2) estimate of groundwater discharge. Naturally occurring 222Rn makes an ideal tracer because it exists in enriched concentrations in groundwater relative to surface water. We used Lake Barco, a small seepage lake in the Katherine Ordway Preserve, for a detailed pilot study in order to test this hypothesis. The preserve is located about 34 km east of Gainesville, Florida and makes an excellent research site because it is maintained in a pristine state as it is closed off to public access. Two intensive samplings were completed, one in the "dry" season (March 2002) and one in the "wet" season (August 2003). In addition, periodic sampling trips were made to the lake every few weeks to evaluate the spatial and temporal patterns of radon distribution. A 222Rn mass balance was constructed for the lake to evaluate radon fluxes. Once the 222Rn fluxes were determined, groundwater inputs were estimated by dividing these fluxes by the concentration of radon in the water seeping into the lake. The radon concentration is estimated by sediment equilibration experiments using grab samples of sediment and radon activity measurements from monitor wells. We also constructed a traditional hydrologic water budget for Lake Barco. All hydrological variables were either measured or estimated and substituted into a water balance equation, which was then solved for the net groundwater flow term. These estimates were then compared to those from the 222Rn model. Our 222Rn measurements show little discernible spatial variation of radon inventories in the lake on any given day of sampling. Inventories measured at five different stations during the two different intensive sampling trips were all within ±13% of the mean value for each sampling. This amount of variation is considered insignificant, as the estimated analytical variation for the radon measurements is ±10%. We also found that we can estimate the lake inventory at any one station by collecting triplicate samples approximately one meter over the bottom at a single central sampling station. However, our results did display significant temporal variability, responding to rainfall trends that can influence the recharge rates. So, in a lake the size of Lake Barco (12 ha), reasonable groundwater discharge estimates at any given time can be made by triplicate analysis of radon in near-bottom waters and collection and processing of a few sediment samples. The estimations of groundwater discharge to the lake over the period of our study were, on average, 13,000 m3/month and 18,000 m3/month for the radon model and hydrologic budget, respectively. The water balance based estimate might be somewhat low due to differences in hydrogeologic movement under such dry conditions or the large error associated with budget estimations. Considering the large uncertainties in estimating groundwater flow via a hydrologic budget, agreement within a factor of two is considered satisfactory.