What can environmental tracer concentrations tell us about groundwater residence time?

Quantification of groundwater age distributions through the use on lumped parameter models and environmental tracer concentrations is an ongoing practice. These models are based on simple geometries and sampling conditions and may not be applicable outside of ideal conditions, limiting the results that can be obtained. We have developed a method that requires no underlying assumptions about the structure of the groundwater age distribution. This method assumes that the convolution relationship describes the relationship between temporal atmospheric concentrations and measured groundwater sample concentrations. We looked at synthetically generated groundwater age distributions and a number of levels of tracer data including a long time series of tritium, the use of CFCs and SF6 sampled at various times and the use of CFCs, SF6, 3H and 14C sampled at a single time. In general, we found that environmental tracers are not very informative of groundwater age distributions resulting in estimates being highly non-unique. The exception was the use of a time series of tritium data. Tritium is more informative due to the temporal variations of atmospheric concentrations and a short half-life reducing the correlation of sequential measurements. The ability of CFC compounds to estimate entire distributions was limited due to the lack of variation in temporal atmospheric concentrations and correlation between sampled concentrations. Despite varying ability of the method to predict whole distributions, some indicators such as the mean age and some arrival time statistics were estimated well by the method. The implications of this show that care should be taken in the choice of lumped parameter model used in conjunction with environmental tracer techniques to constrain entire groundwater age distributions. Interestingly, the ability to correct for mixing bias of mean ages despite poor estimates of entire groundwater age distributions shows some promise.