Limitations to and Cautions About the Application of Environmental Tracers for Dating


Limitations to the use of tracers for groundwater-dating purposes have been documented (for example: Busenberg and Plummer, 2000; Solomon and Cook, 2000; International Atomic Energy Agency, 2006). Most of the uncertainties are environmental in nature; analytical uncertainty generally is only a minor source of uncertainty.


Limitations include the following for CFCs: mixing (dispersion, diffusion, mixing at sampling points), degradation, sorption, anthropogenic CFC contamination in the aquifer, CFC contamination during sampling, local atmospheric anomalies in CFC concentrations, lack of equilibrium between atmospheric gases and unsaturated-zone gases, effects associated with recharge from sources other than diffuse recharge (for example, recharge of river water that was not at equilibrium with the atmosphere), flattening and overturning of the CFC input curves in the 1990s and early 2000s, and uncertainties in estimation of recharge altitude, recharge temperature, and excess air. Many of these effects are evaluated quantitatively in Busenberg and Plummer (1992) and International Atomic Energy Agency (2006).


Limitations for SF₆ include: mixing, natural subsurface production, degassing in the aquifer or during sampling (potentially problematic because of the low solubility of SF₆), anthropogenic SF₆ contamination in the aquifer, SF₆ contamination during sampling, spatially variable atmospheric SF₆ concentrations, lack of equilibrium between atmospheric gases and unsaturated-zone gases, effects associated with recharge from sources other than diffuse recharge, and uncertainties in estimation of recharge altitude, recharge temperature, and excess air. Many of these effects are evaluated quantitatively in Busenberg and Plummer (2000) and International Atomic Energy Agency (2006).


Limitations for ³H/³He include: mixing, fractionation of He isotopes and He mass-balance problems associated with He degassing, uncertainties in corrections for ³He derived from terrigenic sources, uncertainties in age interpretation for samples with low (< 1 TU) tritium content, and uncertainties in estimation of recharge altitude, recharge temperature, and excess air. Quantitative evaluation of several of these effects is discussed in Schlosser and others (1989) and Solomon and Cook (2000).


A major limitation common to all three of these dating methods is that tracer-based piston-flow ages are based on the simplifying assumption that tracer transport is by advective flow without hydrodynamic dispersion and mixing in wells (Bethke and Johnson, 2002; Weissmann and others, 2002). Thus, piston-flow concepts represent an end-member condition that may be approached in some simple flow systems, but a condition that is never actually attained. Mixing can be problematic particularly in supply wells—the typical well type in most MAS networks. Supply wells (including domestic supply wells) often are characterized by long screened intervals, which promote mixing in well bores during sampling. High pumping rates in supply wells can promote mixing owing to the hydraulic stresses that pumping can place on aquifers (Busenberg and Plummer, 1992). Supply wells typically are installed at greater depths than are monitoring wells and this can lead to increased mixing in the former due to increased mixing along the typically longer flow paths. A limitation also related to mixing but specific to tracers such as CFCs, SF₆, and ³H/³He is that these tracers provide information primarily on recharge from recent decades, and generally have provided little information about the distribution of times-of-travel of the older components (if present) in groundwater mixtures. (A groundwater sample may be composed of a multitude of different water parcels, each with a distinct time-of-travel.)


Mixing generally should be less extensive in samples from LUS wells (the dominant network in this compilation) than in samples from MAS wells. The use of short-screened monitoring wells in most LUS networks (85 percent of the LUS wells in this report were monitoring wells) and the typical placement of these wells close to the water table in recharge areas may result in less mixing in well bores and less hydrodynamic dispersion along flow paths (International Atomic Energy Agency, 2006). Low pumping rates may give an additional advantage to monitoring wells because these rates normally cause only minimal disturbance to natural groundwater flow conditions, therefore minimizing mixing in the aquifer. However, all groundwater samples are mixtures of multiple components.


Finally, most of the datasets in this report consist of sites selected by a stratified random design. That is, wells were randomly located in areas that were representative of a particular hydrogeologic setting (MASs) or land-use setting (LUSs). Data from sites that are selected in this manner, rather than from sites located along a targeted groundwater flow path, can sometimes be difficult to interpret because the understanding that can be derived from analysis of position within a flow system may be lacking. Many of the MAS network wells could fall into this category. However, LUS wells typically are positioned close to the water table, which provides flow-system context.


Because the piston-flow model rests on many simplifying assumptions, it has been argued that tracers would be used more appropriately to calibrate transport models than to generate tracer-based piston-flow ages (for example, Bethke and Johnson, 2008). Although such uses may prove to be superior, it is worth noting that tracers do represent point measurements that have the potential to provide understanding that may be lacking in transport models that inherently must simplify the representation of groundwater systems. This understanding might not be generated from a transport model in which geologic heterogeneity is simplified relative to the scale of measurement points (wells). Despite simplifying assumptions and numerous limitations, tracer-based piston-flow ages derived from the commonly used tracers of young groundwater (CFCs, SF₆, and ³H/³He) can be robust under favorable environmental and sampling conditions. For example, CFC-based piston-flow ages from diverse agricultural areas have been shown to be consistent with pesticide application histories (Burow and others, 2007; Tesoriero and others, 2007), distribution patterns of CFC-based piston-flow ages in central Oklahoma have been linked to periods of increased precipitation and recharge (Busenberg and Plummer, 1992), and tracer-based piston-flow ages based on CFCs, SF₆, and ³H/³He have been shown to agree well with those derived from other tracers (for example, Ekwurzel and others, 1994; Busenberg and Plummer, 2008).


The tracer-based piston-flow ages in this report were derived using a consistent approach and are reported with the limitations described above. If the limitations associated with these tracer-based piston-flow ages are ignored, incorrect conclusions may be drawn. However, these tracer data and interpretations contain useful information that can be used in the context of other lines of evidence to gain an improved understanding of flow and chemistry in groundwater systems.


First posted January 27, 2011