Distribution of Groundwater Ages at Public-Supply Wells: Comparison of Results from Lumped Parameter and Numerical Inverse Models with Multiple Environmental Tracers

Estimates of groundwater age distributions at public-supply wells can provide insight into the vulnerability of these wells to contamination. Such estimates can be used to explore past and future water-quality trends and contaminant peak concentrations when combined with information on contaminant input at the water table. Information on groundwater age distributions, however, is not routinely applied to water quality issues at public-supply wells. This may be due, in part, to the difficulty in obtaining such estimates from poorly characterized aquifers with limited environmental tracer data. To this end, we compared distributions of groundwater ages in discharge from public-supply wells estimated from age tracer data (SF6, CFCs, 3H, 3He) using two different inverse modeling approaches: relatively simple lumped parameter models and more complex distributed-parameter numerical flow models with particle tracking. These comparisons were made in four contrasting hydrogeologic settings across the United States: unconsolidated alluvial fan sediments, layered confined unconsolidated sediments, unconsolidated valley-fill sediments, and carbonate rocks. In all instances, multiple age tracer measurements for the public-supply well of interest were available. We compared the following quantities, which were derived from simulated breakthrough curves that were generated using the various estimated age distributions for the selected wells and assuming the same hypothetical contaminant input: time lag to peak concentration, dilution at peak concentration, and contaminant arrival and flush times. Apparent tracer-based ages and mean and median simulated ages also were compared. For each setting, both types of models yielded similar age distributions and concentration trends, when based on similar conceptual models of local hydrogeology and calibrated to the same tracer measurements. Results indicate carefully chosen and calibrated simple lumped parameter age distribution models can give predictions of contaminant breakthrough at individual public-supply wells that are similar to predictions from more labor-intensive numerical groundwater flow simulation models for contaminants with spatially uniform but time-varying inputs. The simulated contaminant trends illustrate important potential errors or uncertainties associated with well vulnerability assessments based on either mean ages or discrete age estimates (apparent tracer ages), and highlight the advantage of having relatively simple tools that can be used to gain some insight into the groundwater age distribution at a public-supply well.