The shallow aquifers discussed above typically have residence times of decades to hundreds of years. In contrast, deeper and less permeable aquifers that extend for many kilometers can have through-flow times of thousands of years. If the flow regime is simple and mixing minimal, such aquifers can serve as archives of information about environmental conditions at the time of recharge. After a long interval with only minor progress, there has been a revival of interest in this approach to paleoclimatic reconstruction [ Fontes et al., 1993] and some results of major significance have begun to emerge.
The major tracers of interest have been the stable isotopes of hydrogen,
oxygen, and carbon, and the concentrations of noble gases. Oxygen-18 and
H in the water molecule mainly reflect the temperature of precipitation
and post-infiltration evaporation, but other climatic factors also play a
role. Understanding the processes responsible for the isotopic composition of
groundwater recharge relies mainly on studies of the isotopic content of
modern precipitation. As the length of time such data have been collected
increases, our understanding of the processes does also, as reflected in
studies such as that of Ronzanski et al. [1992].
One of the best examples of the new prominence of groundwater tracers in
paleoclimatic studies comes from Devils Hole in southern Nevada [ Winograd et
al., 1992; Coplen et al., 1994]. Although these studies did not measure
parameters directly on groundwater, they did so indirectly by sampling a
calcite vein that had recorded the characteristics of the groundwater from
which the calcite precipitated. The results provide a well-dated 500,000-year
record of changes in the 
O of meteoric water (i.e., of
precipitation from the atmosphere). The age obtained for the termination of
the penultimate glaciation is older than previous results, which the authors
argue casts serious doubt on the predictions of the generally-accepted
Milankovich approach, which ascribes initiations and terminations of
glaciations to astronomically-induced variations in solar radiation.
The data on stable isotope compositions of paleowater that have been obtained from Devils Hole are only one example of the expanding amount of information available on the changes in isotopic composition of meteoric water due to glacial/interglacial cycles. In general, the data have indicated that glacial-age water was isotopically lighter than modern groundwater, which is what would be expected under colder glacial climates. However, recent results reported by Plummer [1993] seem to support the existance of an anomalous region of isotopically heavier glacial-age groundwater in the southeastern United States. The origins of this anomaly are at present speculative, but undoubtedly it is an important clue to the nature of glacial-period atmospheric circulation.
Another example of the revival of interest in aquifers as archives of paleoclimate is the application of dissolved noble gas concentrations in groundwater as a paleothermometer. The principles have been described by Stute and Sonntag [1991] and Stute and Schlosser [1993]. A large part of the present interest is stimulated by the desire to resolve a discrepancy between estimates of glacial period temperatures at low latitudes based on the distributions of planktonic foraminifera fossils (which generally show little temperature reduction) and those based on snow-line reconstructions (which indicate large temperature reductions). The temperature dependence of the solubilities of the noble gases provides a simple, physically-based method of testing these paleotemperature estimates. Results from studies of the Carizzo aquifer in south Texas [ Stute et al., 1992] indicate a 5 C to 6 C reduction in temperature, consistent with the snow-line estimates. A similar study on the Ojo Alamo aquifer at higher elevation in New Mexico [ Stute et al., 1993] gave equivalent results, indicating that the snow-line and planktonic estimates could not be reconciled by postulating a reduced temperature lapse rate with elevation during the glacial period.
Paleohydrology studies, relying in large part on isotopic methods, have increasingly led to the realization that modern groundwater flow systems have been strongly influenced by changes in boundary conditions that have resulted from long-term climate change. Examples include the Midwestern studies by Siegel [1990 and 1991] that traced the direct influence of the continental ice sheets on groundwater recharge and that by Smith et al. [1992] which proposed that much of the groundwater in the desert areas of southeastern California was recharged under more humid glacial-period conditions. Love et al. [1994] have demonstrated the effects of glacial/interglacial climate change on recharge in South Australia and Fontes et al. [1991] have used environmental tracers to show how shifts in the position of the Niger River since the end of the last glaciation have affected the distribution of groundwater recharge.
The utility of isotopic tracers in applied research on water quality problems is beginning to be widely appreciated. Joseph et al. [1992] have applied stable isotopes to infer the relationship between atmospheric sources of water vapor and the resultant groundwater recharge (or lack of it) in the Sahel. Simpson and Herczeg [1991] have used stable isotope enrichment as a method of investigating the sources of the salinization that is afflicting the Murray River Basin in Australia. Vengosh and Rosenthal [1994] have included stable isotope methods in the techniques they apply to analyzing the role of saline groundwater in the water crisis in Israel.
Finally, more examples of the use of isotopic tracers in localized, applied, groundwater problems are beginning to crop up. Carrillo-Rivera et al. [1992] applied a suite of isotopic tracers to analyzing the water-supply problems of a small basin in Mexico. McCarthy et al. [1992] used stable isotopes to help quantify groundwater/surface water interactions along the Columbia River. Simpkins and Bradbury [1992] applied tritium to estimating flow velocities in thick till in Wisconsin.
The most common approach to radiometrically determining residence times of
groundwater in large aquifers has been 
C. There have been several advances in this method
over the past three years. One has been the development
of new numerical approaches to correcting for carbon reactions in the
subsurface, especially the NETPATH model of Plummer et al. [1991]. A quite
different approach has been taken by other researchers who, following the lead
of Murphy et al. [1989], have attempted to measure 
C activity in the
organic, rather than the usual inorganic, component of the dissolved carbon.
Wassenaar et al. [1991] have examined the carbon isotope composition in a
shallow aquifer in Ontario, while Purdy et al. [1992] compared organic and
inorganic 
C results in the large Aquia aquifer in Maryland. Murphy et
al. [1992] presented an unusually detailed analysis of the interrelationship
between carbon isotopes, carbon species chemistry, and microbial activity on
the dissolved organic and inorganic components in the Middendorf aquifer.