Investigating surface water–well interaction using stable isotope ratios of water

Because surface water can be a source of undesirable water quality in a drinking water well, an understanding of the amount of surface water and its travel time to the well is needed to assess a well’s vulnerability. Stable isotope ratios of oxygen in river water at the City of La Crosse, Wisconsin, show peak-to-peak seasonal variation greater than 4‰ in 2001 and 2002. This seasonal signal was identified in 7 of 13 city municipal wells, indicating that these 7 wells have appreciable surface water contributions and are potentially vulnerable to contaminants in the surface water. When looking at wells with more than 6 sampling events, a larger variation in dO compositions correlated with a larger fraction of surface water, suggesting that samples collected for oxygen isotopic composition over time may be useful for identifying the vulnerability to surface water influence even if a local meteoric water line is not available. A time series of dO from one of the municipal wells and from a piezometer located between the river and the municipal well showed that the travel time of flood water to the municipal well was approximately 2 months; non-flood arrival times were on the order of 9 months. Four independent methods were also used to assess time of travel. Three methods (groundwater temperature arrival times at the intermediate piezometer, virus-culture results, and particle tracking using a numerical groundwater-flow model) yielded flood and non-flood travel times of less than 1 year for this site. Age dating of one groundwater sample using H–He methods estimated an age longer than 1 year, but was likely confounded by deviations from piston flow as noted by others. Chlorofluorocarbons and SF6 analyses were not useful at this site due to degradation and contamination, respectively. This work illustrates the utility of stable hydrogen and oxygen isotope ratios of water to determine the contribution and travel time of surface water in groundwater, and demonstrates the importance of using multiple methods to improve estimates for time of travel of 1 year or less. q 2004 Elsevier B.V. All rights reserved.