Tracer-Based Piston-Flow Ages and Related Information 


This section documents previously published tracer-age analyses for 618 sites, and newly interpreted tracer data for 781 sites evaluated as part of this project. Additionally, some insights about recharge temperatures, derived from analysis of major dissolved-gas data, are discussed briefly.


Interpretations of tracer data are given as tracer-based piston-flow ages and as tracer-based piston-flow recharge dates. As an example, a groundwater sample collected during calendar year 2000 and attributed with a tracer-based piston-flow age of 10 years would have a tracer-based piston-flow recharge date of (calendar year) 1990. Tracer-based piston-flow ages and tracer-based piston-flow recharge dates often are censored with a “>” (greater than), “<” (less than), “≥” (greater than or equal to), or “≤” (less than or equal to). For example, if CFC degradation appeared to be present in a sample, the tracer-based piston-flow age and tracer-based piston-flow recharge date could have an old bias. The tracer-based piston-flow recharge date could be accompanied by a “>” to indicate a greater (more recent) date, and the tracer-based piston-flow age by a “<” to indicate a smaller (younger) age. Sites with age-dating results that have been censored with >, <, ≥, or ≤ do not have discrete tracer-based piston-flow ages or discrete tracer-based piston-flow recharge dates, and these sites are not included in statistical characterizations of tracer-based piston-flow ages.


Previously Interpreted and Published Environmental-Tracer Data


Tracer data from 618 sites in 22 Study Units have been interpreted and reported (as tracer-based piston-flow ages) in 30 publications. Tracer concentrations also have been reported in many of these publications. The level of interpretation of tracer data in these publications is variable. For example, an urban LUS [LUSRC1, table A1 (appendix A)] in the Great Salt Lake Basins (GRSL) Study Unit was sampled for ³H/³He in 1999. The ³H/³He-based piston-flow ages were used to reconstruct initial ³H concentrations that were expected to be present in recharge water at the time of recharge. These reconstructed initial ³H concentrations matched historical ³H inputs (discussed and referenced in appendix F), demonstrating consistency between age interpretations and an independent data set. However, many of the CFC-based piston-flow ages from the Kanawha-New River Basin (KANA) Study Unit, although thoroughly documented in a report that focused on age dating (appendix F), may have an old bias as a consequence of the extensive occurrence of highly reducing conditions. (In table F1, potentially affected sites from the KANA Study Unit have comments in the remark fields about this potential problem.)


The original interpretations for these 618 sites have been compiled, without re-interpretation, in table F1. Table F1 lists sampling date, tracer, tracer-based piston-flow age, tracer-based piston-flow recharge date, and citable reference for each site, organized by network and Study Unit. Appendix F also contains a short summary for each group of age results listed in table F1, including information describing the extent to which age interpretation was addressed. These summaries are provided, in part, to facilitate the process of matching these results, or appropriate subsets of these results, to the intended uses.


Newly Interpreted Environmental-Tracer Data


Tracer data from 781 sites in 26 Study Units that previously had not been interpreted and published were interpreted here as tracer-based piston-flow ages. The data are organized by Study Unit and network. Some networks were combined for the purposes of analysis and reporting. Networks were combined where multiple networks were composed of similar land use, or where REF wells and LUS wells were co-located. Data from the Apalachicola-Chattahoochee-Flint Basin (ACFB) and Georgia-Florida Coastal Plain Drainages (GAFL) Study Units also were combined because these two (adjacent) Study Units shared a continuous, cross-Study-Unit LUS. 


The measured tracer-concentration data, interpreted tracer-based piston-flow ages, and ancillary data for each network are reported in appendix G. The derived tracer-based piston-flow ages are examined for consistency with available tritium data and local age gradients in appendix G. A summary of the tracer-based piston-flow ages from appendix G is given in table 1 (at back of report).


Where tracer data from more than one tracer type (CFCs, SF₆, ³H/³He) were available for a given site, interpretations from both tracers are provided in appendix G. Multiple tracers were compiled for 20 sites: 2 sites in the Connecticut, Housatonic, and Thames River Basins (CONN); 5 in the South-Central Texas (SCTX); and 13 in the Upper Illinois River Basin (UIRB) Study Units. Some of these cases of multiple tracers occurred on different sampling dates, and some occurred on the same sampling dates. At each of 3 sites (1 in the CONN and 2 in the SCTX Study Units), tracer data from 2 sampling dates were available, and these separate results are included in table 1. At each of 17 sites (1 in the CONN, 3 in the SCTX, and 13 in the UIRB Study Units), the multiple tracers were collected on the same date, but only the most reliable tracer-based piston-flow age is reported in table 1. The multiple tracers were ³H/³He in combination with either CFCs or SF₆. Generally, ³H/³He was considered the more reliable tracer because the parent compound (³H) and the decay product (³He) are measured, and because this tracer pair is not subject to microbial degradation. 


Relation Between Tracer-Based Piston-Flow Ages and Network Type


To what degree have LUS sites achieved the network design goal (Gilliom and others, 1995) of accessing groundwater younger than 10 years of age? Because all groundwater samples are mixtures and represent an age distribution, both the LUS network design goal and this question oversimplify the problem. Ignoring the fact that the groundwater samples are mixtures, 530 of the 859 LUS sites (62 percent) have an uncensored tracer-based piston-flow age (table 1; table F1). For these 530 sites (which probably are not representative of all NAWQA LUS sites), the median tracer-based piston-flow age is 11 years, and 39 percent of the sites had tracer-based piston-flow ages of less than 10 years. (Another 6 percent of the datable LUS sites had a tracer-based piston-flow age equal to 10 years.) Therefore, without consideration of the age distributions in the groundwater samples, the tracer-based piston-flow ages could be construed as indicating that the goal of accessing groundwater younger than 10 years of age was achieved at nearly one-half of these sites. However, each sample actually represents an age distribution, and many of these samples probably contain fractions of groundwater with times-of-travel less than 10 years and fractions with times-of-travel greater than 10 years. So a more meaningful, albeit more challenging, LUS network design goal could have been framed in terms of age distributions.


As might be expected based on the design of LUS and MAS networks, typical tracer-based piston-flow ages of age-dated groundwater samples from MAS networks are somewhat greater than those from LUS networks (fig. 2). The median tracer-based piston-flow age of datable MAS sites was 17 years, compared to 11 years for datable LUS sites (fig. 2). These patterns of generally greater time-of-travel for MAS sites than for LUS sites likely reflect generally greater well depths in MAS wells than in LUS wells. This also may reflect the fact that these LUS and MAS sites were from different populations of wells, from different aquifers with different boundary conditions (for example, recharge rates), geologic composition, geometry, and size. Few data in this compilation are available for direct comparisons between nested LUS and MAS networks (LUS and MAS networks co-located in the same Study Units and principal aquifers), but data from nested LUS and MAS networks in the White, Great, and Little Miami River Basins (WHMI) Study Unit (newly interpreted data, table 1) and the Potomac River Basin and Delmarva Peninsula (PODL) Study Unit (previously interpreted data, table F1) demonstrate results similar to those for the more generalized analysis discussed above. (These two sets of nested networks were the only nested networks from this report that consisted of more than 10 uncensored, datable sites in the LUS and the MAS categories). Median tracer-based piston-flow ages were 2 years (33 WHMI LUS wells) and 9 years (25 WHMI MAS wells), and 8 years (29 PODL LUS wells) and 10 years (29 PODL MAS wells). Differences in tracer-based piston-flow ages between nested LUS and MAS wells (2 years and 7 years for two direct comparisons) were of similar magnitude to the differences observed for the more generalized comparison between the various LUS and MAS networks (6 years). 


Major-Dissolved-Gas-Based and Climate-Based Recharge Temperatures


The compilation of tracers in this report is accompanied by a compilation of major dissolved-gas data that were used for, among other things, estimation of recharge temperatures. In the absence of major or noble dissolved-gas data, recharge temperature commonly is assumed to be close to either the MAAT (Andrews, 1992) or the MAAT+1°C (Stute and Schlosser, 2000). (Where major dissolved-gas data were not available in this report, recharge temperatures were assumed to be equal to the MAAT+1°C.) The dataset of major dissolved gases presented herein for aquifers across the United States provides an opportunity to compare a large and diverse group of N₂/Ar-inferred recharge temperatures to those based on climate data. Considering sites in aquifers composed of sediments (most of the sites analyzed in this report), differences between N₂/Ar-inferred recharge temperatures and recharge temperatures based on MAAT+1°C were, on average, about 0.1°C (n = 277) (fig. 3). The mean difference was +0.1°C (N₂/Ar-based estimates slightly greater than the climate-based estimates), and the median difference was -0.1°C (N₂/Ar-based estimates slightly less than the climate-based estimates). However, the standard deviation of these differences was 3.1°C. Thus, although this comparison could provide support for the use of climate data for estimation of average recharge temperatures, recharge temperatures vary greatly around this central tendency and characterization of site-specific recharge temperatures benefits from site-specific data. Factors contributing to differences between climate data and N₂/ Ar-inferred recharge temperatures are not well understood, but include environmental processes, such as type of recharge (for example, focused compared with diffuse recharge), variable intensity or seasonality of recharge, geothermal warming in thick unsaturated zones, and warming or cooling from land surface where unsaturated zones are thin, and also include uncertainties associated with resolving excess-air components (Kipfer and others, 2002; Plummer and others, 2004).


First posted January 27, 2011