Second to Earth’s glaciers, groundwater is the most abundant and accessible fresh water resource available to humans. About 40% of the human population is dependent on groundwater for their daily water needs. In many areas of the world aquifers are undergoing severe overdraft or have extensive contamination from industrial and agricultural practices. Even in developed nations these problems occur in basins that are not fully characterized. Overdraft is the condition where demand outstrips replenishment (or recharge) of the groundwater. Furthermore, contamination of groundwater stems from recharge that carries the contaminant into the subsurface. The important process of groundwater replenishment is addressed in this chapter, which focuses on natural settings and illustrates the lack of convenient (reliable?) analytical approaches for its characterization. We also discuss attempts to intentionally induce recharge to offset groundwater overdraft.

Natural Groundwater Recharge

Figure 1 shows standing water levels (blue) for a well located in the Sacramento Valley of California. We see an annualized trend that oscillates between high and low levels, where the high stands occur in the middle of winter and the low levels in mid- to late-summer. Also shown in black are the individual precipitation events in a nearby weather station. Note that the rainy season begins after the water level in the well has already begun to rise, and when the cumulative precipitation is maximum the water level has already dropped. These observations strongly suggest that the oscillatory trend is caused by seasonal

Figure 1. A well in the Sacramento Valley of California where annualized water levels (blue) oscillate from high in the winter to low in the late summer. Precipitation (black bars) is maximized when standing water levels are high. The oscillatory trend is primarily caused by agricultural pumping for summer irrigation, with little pumping in the winter. Also note the progressive decline in water levels. Data from California Department of Water Resources CASGEM

pumping for agricultural irrigation, which lowers the water levels even when precipitation is high. Remarkably, water levels then begin to rise in late summer, when precipitation is low, clearly because pumping has ceased. However, the data trend leads to some important questions about the groundwater aquifer and how it supports the annual pumping. Firstly, where is the source of the aquifer recharge located? Is it replenished directly from the surface, or laterally conveyed from a distant area? What is the importance of the annual precipitation to this replenishment process? Does a nearby river or stream replenish the groundwater? How fast is the replenishment?

Firstly, to address these questions we need more information about the well location and the surrounding topography and geology. The well is located on the southwest side of the Sacramento Valley, California which is a broad low-lying area surrounded by higher elevations to the west, east, and north. In the center of the valley runs the Sacramento River, which is mostly fed by snowmelt released from numerous surface reservoirs. The well in Figure 1 is located several miles to the west of the Sacramento River and up-gradient. The area land use is dominated by irrigated agriculture that applies about 3.5 feet of standing water equivalent per year for crop growth. Thousands of wells throughout the valley pump groundwater during the growing season. During the wintertime, precipitation generates runoff in the high elevation areas to west and snowpack in the higher elevations to east and north. The high elevations are dominated by either low permeability sandstones and shales or crystalline rock, whereas coarse gravel underlies the valley edge. This coarse material is highly permeable and an area where surface water runoff is partially lost to groundwater seepage. Further downhill and toward the valley center, the alluvial fill becomes more sand and silt dominant. At depth, the geology is all recent unconsolidated fluvial material laid down over the time by bed loads of streams and the river which created approximately 2000 feet of sand and gravel lenses inter-fingered with low permeability silts and clays.

The change in slope that occurs at a valley edge is important to recognize for both the geologic development of the valley floor, but also for streamflow behavior over time. Streamflow under a steep topographic gradient has sufficient energy to cause rock erosion over time. The eroded material supplies the sedimentary bed load carried downstream and ultimately generates the valley fill below. Where the topographic slope undergoes a sharp decrease, the rapid reduction of stream energy loss causes the coarsest part of the bed load to deposit. This is why coarse material can always be found at the boundary

Figure 2. Google Earth showing the steep reduction in topographic slope on the east side of Colorado Rockies, which causes streams to fan out and deposit coarse gravel, while inducing groundwater recharge because of stream energy loss.

between a valley and the bordering steep, high elevation terrane. The finer material in the bedload can be carried further downstream into the valley. In addition, the stream bed capacity to carry water also decreases during the change in topographic gradient. This forces streams to spread out or form multiple channels. This creates greater surface area of flow that now is traversing coarse material deposited on the surface. In many cases this is an area of surface water seepage into the groundwater accommodated by the course surficial sediments. Figure 2 illustrates the Rocky Mountain front-range looking west from the Great Plains. In this exaggerated Google Earth image, the break in slope between the front-range and the lower plains area is clearly seen. Note also how stream flows emerging from the range marked in the green areas spread out as they progress downhill. As a matter of fact, the change in slope from the front range to the valley is about a factor of two and half, which equates to over half of the stream energy being lost with the decrease of topographic slope. As a result, this area is the principal recharge area of both the Ogallala Aquifer in the Great Plains and the underlying Dakota sandstone aquifer. So in the case of our example in Figure 1, recharge to the pumping well is supplied by a similar recharge mechanism. However, recharge directly from the overlying surface is also an important in the Sacramento Valley, as discussed later.

Because in fluvial basins the deposits were laid down by the previous courses of streams and rivers, modern day surface flows have intimate contact with permeable layers below the river or stream bed. Where flow passes through an area of topography where both banks have groundwater elevations higher than the stream level, then groundwater will tend to flow into the stream. This is referred to as a gaining stretch of a stream. Alternatively, where the opposite occurs and groundwater levels on either side of the surface flow are lower than the stream surface, then surface water will seep into and recharge the groundwater system. This is referred to as a losing stretch of a stream or river. Both conditions can occur along a single stream, as illustrated in Figure 3. Rivers and streams represent important recharge sources in basins that are being over pumped, but this results in stream flow loss as it replenishes the underlying aquifer. Such conditions are seen in areas such as the Ogallala Aquifer in the Great Plains, the Owens Valley, and the Central Valley in California.

Figure 3. Left hand image illustrates a gaining stream with real contoured groundwater levels adjacent to the Sacramento River in northern California. The right hand image illustrates a losing stream also with the Sacramento River in the Sacramento City area. Data from California Department of Water Resources

Tertiary volcanic terranes are also areas where rapid groundwater recharge can take place. Lava flows and ash flow tuffs develop extensive fracture networks as they cool, which provide conduits for rapid infiltration of precipitation (Figure 4). Furthermore, soil development is poor and no impeding features exist to direct the infiltration into evapotranspiration.

Figure 4. An example of recent volcanic basalt layer with no soil development and source of rapid infiltration.

Karst terranes also facilitate the rapid infiltration of surface water and precipitation. Limestones are soluble and subject to dissolution in freshwater settings, because organic acids generated in soil zones lower the pH of infiltrating water. This water has a tendency to dissolve the limestone over time along preferred paths, commonly forming caves in these terranes. As caves grow, they thin the overlying rock and eventually form collapse structures referred to as sinkholes. Karst topography is easily recognized by numerous sinkholes, each a topographic depression where entering surface waters must flow downward (Figure 5). Thus each sinkhole represents a rapid path for water infiltration into the groundwater aquifer below.

So far we have discussed recharge mechanisms through highly accessible pathways through permeable material. However, in many other areas precipitation will infiltrate directly into developed soil profiles and the fate of the water is not straightforward. Recall in the previous chapter on the Hydrological Cycle that we characterize water balance as P = R + ET ±ΔS, where P is the precipitation, R is the runoff, ET

Figure 5. Example of karst terrane adjacent to the Mississippi River in southwestern Illinois. Rainfall entering these sinks must move downward into the groundwater system, which then flows laterally through fissures and caves, ultimately to emerge along the river bluff, as illustrated by Falling Spring (see map and bottom photo).

is the evapotranspiration, and ΔS is the change in storage in groundwater aquifers and soil zones. ET and ΔS are strongly coupled in areas of well-developed soils. This is due to the fact that plant roots are effective at intercepting infiltrating water in soils. In developed soils parent bedrock has been entirely altered by physical and chemical breakdown processes. Much of this is facilitated by plant root growth coupled with acidic exudates that promote rock dissolution. Over time soil becomes layered with an organic rich upper horizon, then by an underlying, mineral and organic rich layer where most roots occur, and then either by a clay rich or clay deficient zone, depending on the local climate (Figure 6). Generally, below this is the uppermost bedrock in various stages of disintegration. Often the agronomist will refer to soils as either poor or well drained, depending on whether infiltrating water can pass through the root zone efficiently enough to not compromise plant growth. This need have no bearing on groundwater recharge because excess soil drainage can easily be drained laterally, towards streams or other points of lower head.

Figure 6. Different vegetative regimes as a function of precipitation and mean annual temperature. After Whittaker (1962).

Infiltration refers to the penetrating capacity of precipitation into the soil zone. The capacity to infiltrate depends on the nature of the soil surface, such as the presence of plant litter, the amount of water already in the soil, the rooting structure, and the overall porosity and permeability. Infiltration is high beneath forest canopies but can be low in fire-damaged areas such as burnt natural grasslands. Infiltrating water is driven downward by gravitational force through pore spaces, but in very small pores capillary forces can dominate. This latter process is driven by electrostatic attraction between mineral surfaces and polar water molecules. For larger pores capillary action is not strong enough to overcome the gravitational force. Much of the infiltrating water, particularly in small pores, is intercepted by plant roots forming the ET component of the water balance equation. The more intense the vegetative growth the higher the root water demand. This demand is commonly large in tropical to temperate climates, but much smaller in semi-arid to arid regions (Figure 6). Commonly, soil development is poor in dry areas but advanced in wet regions.

It is very difficult for a student of hydrology to build an intuitive grasp of evapotranspiration in the absence of simple observation. The standard approach involves convoluted model representations of the process and tedious calculations while overlooking essential comparisons. For instance, Hasenmueller and Criss (2013) devised a simple way to envision the role of ET in water mass balance, by computing the ratio of mean annual runoff from a watershed to the mean annual precipitation, using the mass balance equation 1-R/P = (ET±ΔS)/P. Two different climatic regions can be compared this way to illustrate the relationship. Firstly, in Missouri which has a mean annual precipitation of about 38 in/yr, the Meramec River basin has a mean annual runoff measured at the Eureka station of 3261 cubic feet per second, which normalized over the watershed area of 3788 mi2, equate to 11.7 in/yr of precipitation equivalent. One minus the ratio of 11.7 in/yr to 38 in/yr (1-R/P) is 0.69, which is the ET±ΔS normalized to P. Since most of this watershed is karst terrane, any groundwater recharge re-emerges as stream flow from springs down gradient, so that nearly 2/3 of the mean annual precipitation is lost as ET, consistent with the dense forest growth of this watershed.

For comparison, the same calculation is made for a Mojave Desert watershed in the New York Mts where intermittent discharge emerges from the Caruthers Canyon area. Here, mean annual precipitation measured over several years was 13 in/yr for a 0.8 mi2 watershed. Long term mean annual discharge from Caruthers Canyon is computed to be 0.14 cubic feet per second, representing a precipitation equivalent of 2.4 in/yr. Consequently, the computed value for (ET±ΔS)/P is 0.82, which is much higher than that computed for the Meramec River drainage. A quick comparison of the two landscapes reveals a stark difference in vegetative regimes and plant water demand (Figure 7). So how can ET±ΔS be so much higher in the Mojave Desert than in Missouri? The obvious answer is that for the Meramec River watershed, ET dominates over the ±ΔS, whereas in the Mojave watershed the opposite is true. This should not be surprising given the lack of soil development in these type of desert settings and the highly porous and permeable surface material to take up infiltrating water. The other factor in the Mojave case is that most of the annual precipitation falls as snow, which accumulates in one or two events and rapidly melts within days, providing a rapid pulse of infiltrating water that quickly creates zones of saturated infiltration and recharge.

Figure 7. The left hand image illustrates the Meramec River in Missouri with its huge plant water demand and high ET. The right hand image illustrates the Mojave Desert with no soil development and implied rapid precipitation infiltration and low ET.

Now returning back to Figure 1 and addressing the vertical component to recharge we look to the cultivated land use and the method of irrigation. In the southwestern Sacramento Valley soil development is extensive, but heavily ameliorated with added nutrients because of decades of intensive irrigation practices leached large amounts of natural nutrients from the native soil. However, the flood irrigation practice repeatedly inundates the field furrows with standing water to which anhydrous ammonia has been added. On average 3.5 feet of standing water is added in this manner every year. This combined with annual precipitation rate in the area (18 inches) total to about 5 feet of water on the surface per year.

Over the decades the excess nitrogen added to the irrigation water has migrated downward, and now reaches the drinking water wells of nearby municipalities, indicating downward recharge. This downward recharge was further verified and quantified using isotope measurements that showed about one-third of the total water (irrigation + precipitation) applied annually infiltrates deeply and becomes groundwater recharge. This indicates that the ET rate in this area is about 3.3 feet, representing crop demand plus natural evaporation.

Artificial Recharge

Artificial recharge is a term used to describe an intentional modification of a natural system to enhance groundwater recharge to satisfy an immediate or future demand. Water stressed areas that depend on a groundwater supply for daily needs are target areas for artificial recharge, because where ever population density or agricultural demand is high, groundwater withdrawals exceed the natural replenishment rate. Accordingly, where excess water can be obtained, the natural system can be engineered to enhance the groundwater replenishment rate, either through surface percolation or forcible groundwater injection.

A common method of conducting artificial recharge is through enhanced surface water percolation. This is accomplished in geological areas that have a high percolation rate, such as abandoned gravel mining areas or along the sandy bottoms of rivers or creeks. Typically, the surface is modified to maximize contact between the flowing stream and the percolation surfaces; for example, water can be diverted to infiltration ponds lined with sand. Water sources for recharge might represent natural base flow, storm water runoff, or even treated waste water.

A high-profile example of surface percolation for artificial recharge is the Anaheim Forebay of Orange County, California (Figure 8). Active large scale gravel mining occurred in the early to mid-20th century in this area, which now is a dense industrial area along the Santa Ana River that drains the inland empire area of Riverside and San Bernardino counties. The river maintains a base flow of tens of cubic feet per second throughout the year, which is largely made up of tertiary treated waste water discharged from upstream cities. However, during winter storms the runoff can exceed several thousand cubic feet per second. The Orange County Water District owns much of the real estate in the river channel and many of the off-channel basins (Figure 8). The river channel is connected to the basins through subsurface pipes which direct water into the basins for percolation. The basins are periodically scraped of fine-grained deposits that are replaced with sand to enhance percolation (Figure 8). The river bed is treated the same but the channel is lengthened by means of T and L levees to maximize the recharge of streamflow.

Figure 8. The Anaheim Forebay of Orange County California includes large areas groomed for artificial recharge by percolation (blue shading).

Figure 9 provides an example of recharge capacity for intentional percolation at Anaheim Lake in the Orange County Forebay. The data represent two cycles of recharge and basin cleaning and illustrates the rate at which surface water can infiltrate into subsurface aquifers. The amount of water percolated over this four month period in Figure 9 is equivalent to approximately 11,000 acre-ft (3.6 billion gallons), enough to supply around 14,000 southern California households for a year.

Figure 9. Two cycles of percolated recharge and basin cleaning equates to a fast percolation rate.

Together the recharge basins, river channel, and injection wells contribute about 275,000 acre-ft to the groundwater basin per year. Half of that recharge is derived from recycled waste water, already treated in a 100 million a gallon per day, state-of-the-art water recycling plant. This plant is located downhill in south Orange County, but the recycled water is pumped uphill to supplement inflow to the recharge basins.

Another form of artificial recharge, called aquifer storage and recovery (ASR), is to directly inject excess water into an aquifer. . Injection wells are similar to production wells, but care has to be taken on the selection of materials for the well construction. Unprotected metal for casing the well invites biological buildup in the well casing and could lead to clogging. Stainless steel or polyvinyl chloride are preferred for well construction.

An ASR well field is well suited to confined aquifers where horizontal conductivity is tens to hundred times greater than vertical conductivity. If the well field is for temporary storage of groundwater, an aquifer with low horizontal conductivity is preferred as it allows the same groundwater to be extracted later. Alternatively, if an unconfined aquifer is over drafted and the water table is low (>100 feet below the surface), then ASR is feasible.

Another application for groundwater injection is to create a groundwater mound to prevent seawater intrusion. Coastal Los Angeles in southern California has operated a 10 mile-long seawater intrusion barrier for many decades to accomplish this. Similarly, two projects in Orange County to the south inject advanced treated waste water for this purpose. The Orange County basin was over drafted in the 1930s, causing seawater to intrude the basin. The injection barriers have operated since the 1970s and currently the advance of seawater has been halted (Figure 10).

Figure 10. Cross-section, looking south, of a part of Orange County near the Pacific Ocean, where an injection barrier operated since the 1970s attempts to prevent salt water from encroaching into the fresh groundwater basin. After Davisson et al. (1999).

Physical Methods to Determine Recharge Rates

A popular method to determine infiltration of rain water is to employ an infiltrometer. Technically, infiltration and recharge connote two different meanings. Infiltration is wetting of the soil, whereas recharge is infiltration that passes through the root zone to enter the water table below. An infiltrometer is a cylindrical tube that is forced into the soil. Water is added to the tube by means of Mariotte bottle, which keeps a constant head on the infiltrating water. The rate at which the water infiltrates is a measure of the infiltration rate in a particular area. Sometimes two concentric tubes are used, where the outer one buffers lateral flow to make the infiltration one-dimensional. Typically, infiltration is fast in the beginning, but then slows down and reaches a steady-state as the soil is wetted. A common problem with infiltrometers is that they distort the soil when they are emplaced. Also, air can be trapped by the infiltrating water and preferential flow paths are established. The biggest problem with infiltrometers is their one-dimensional nature. Soils are very heterogeneous and it would take hundreds of infiltrometers to map out the infiltration rate of a real recharge zone. The time and expense required render this methodology rather impractical.

Another method to quantify recharge is to measure stream flow loss. As streams encounter porous media they tend to lose water to the subsurface as illustrated in Figure 2. Alternatively, if the groundwater basin is being pumped, surface water will tend to replace the extracted groundwater. Discharge measurements made both upstream and downstream, and the difference is the loss of surface water into the subsurface.

Another means to estimate recharge rates is to calculate a groundwater basin mass balance. To accomplish this pumping well discharges need to be quantified and static water levels in wells need to be measured. Stream discharges need to be measured for loss of flow into the subsurface. The amount of annual precipitation must be known. Recall the mass balance equation, P = R + ET ±ΔS; the only value not measured is ET. Static water levels in wells will provide the change in storage, the stream measurements will provide the runoff and precipitation measurements will provide the independent variable P. ET can then be estimated within a factor of two, and compared to aerial images of vegetation. Once ET is estimated, change in storage is the extracted groundwater per annum plus the recharge component.

Geochemical Methods of Estimating Recharge

Geochemical measurements are one of the best ways to estimate recharge by direct observation. Geochemical measurements of isotope ratios or water quality parameters can yield important information on the timing and amount of groundwater recharge. For instance, if geochemical ages are determined on groundwater from a number of wells, this information can be combined with their water levels and position to yield a distance/age relationship. Darcy’s law then provides a linear velocity v such that

where K is hydraulic conductivity of the media, n is porosity, and dh/dl is the hydraulic gradient. Any geochemical marker of age can provide valuable information that yields hydraulic properties of the aquifer material. For example, in an agricultural area the depth of recharge of groundwater contaminated with nitrate gives an important marker of time. Nitrate was heavily applied after World War II, so measurements of its concentration illuminate land use impacts and help define the speed of groundwater recharge.

Chlorofluorocarbons (CFCs)

Chlorofluorocarbon (CFC) compounds have been synthesized on an industrial scale since 1931. They have primarily been used as refrigerants and aerosol can propellants, but also as foam blowing agents, solvents, and in insulation. Production maximized during the 1970s and 1980s before it was recognized that CFCs contribute to the Earth’s ozone destruction. Production was subsequently banned in the 1990s as part of a global agreement. Three principal CFC compounds were used during the 20th century: dichlorodifluoromethane, trichlorofluoromethane, and trichlorofluoroethane, whose trade names are CFC-11, CFC-12, and CFC-113, respectively. The CFCs production and release to the atmosphere have been measured and reconstructed back to 1940, as shown in Figure 11 (McCarthy et al., 1977; Gamlen et al., 1986; Wisegarver and Gammon, 1988; Fisher and Midgley, 1993; Fraser et al., 1996).

Figure 11. Atmospheric concentration of three principal CFCs produced since 1940 based on annual measurements from approximately 1980 to present and reconstructed based on release rates prior to 1980. Source of data is University of Utah Noble Gas Lab.

The basis for age-dating with dissolved CFC measurements in groundwater is comparing the measured values to those of the atmospheric concentrations at the time of recharge. This is accomplished by recognizing that the dissolved concentration Ci is

where KH is the Henry’s constant and pi is the partial pressure of the CFC in air. The concentration is related back to atmospheric concentration through pi

where xi is the dry air mole fraction of the CFC, P is the atmospheric pressure and pH2O is the water vapor pressure. Henry constants have been carefully measured for the three CFCs of interest and solubility determined as a function of temperature and salinity. A number of comparative age-dating studies have shown the reliability of the CFC approach (Busenberg and Plummer, 1992; Busenberg and Plummer, 1993; Ekwurzel et al., 1994; Cook and Solomon, 1997).

Tritium-Helium-3 Age Dating

Attempts have been made in the past to date groundwater with the radioactive (unstable) hydrogen-3 isotope tritium (3H; see Mazor, 1991 and references therein). Because of its short radioactive half-life of 12.43 years, it is ideally a good chronometer for young (≤40 years) groundwater flow. Unfortunately, from a dating standpoint, 3H concentrations in precipitation have varied considerably over the past 50 years due to 3H production from surface and atmospheric testing of thermonuclear weapons (Figure 12).

Figure 12. Changes in the 3H concentration in precipitation have varied over an order of magnitude due to fallout of thermonuclear-produced tritium from surface testing. IAEA/WMO (2001). Global Network of Isotopes in Precipitation. The GNIP Database. Accessible at:

Tritium measurements in groundwater 40 years ago were useful from the standpoint of tracing the "bomb-pulse" 3H that had recharged into groundwater in the early 1960s, and calculating the groundwater travel time based on the observed depth of the "bomb pulse" water. Today, however, "bomb-pulse" groundwater has become less distinctive because of 3H decay and groundwater dispersion. Tritium measurements alone cannot be used for dating groundwater reliably because of the uncertainty in what the original 3H concentration was at the time of recharge, but it does help define relatively young groundwater whenever it is observed.

In more recent years with the development of high-precision noble gas mass spectrometry, the radioactive decay product of 3H, helium-3 (3He), can be measured. The advantage to this lies in the dating equation, where

3H is the concentration of the tritium today, and 3Ho is the original tritium concentration at the time of recharge. Since the 3Ho has a large uncertainty due to the spatially and temporally variable "bomb pulse" tritium, the resulting age calculation will have large uncertainties. By simultaneously measuring the 3He produced by tritium decay (known as the tritiogenic 3He or 3Hetrit) we can reconstruct the 3Ho by adding together the measured tritiogenic 3Hetrit and the 3H which leads to

Dissolved 3He measured in a groundwater is actually derived from several sources that include:

where 3Hemeas is the total 3He analytically measured, 3Heequil is the amount of 3He dissolved in a non-turbulent surface water in equilibrium with the atmosphere and is temperature dependent, 3Heexcess is the amount of 3He dissolved in water exceeding the equilibrium amount (a common phenomenon in groundwater due to excess dissolved air), and 3Herad is the amount of 3He produced from radioactive decay of isotopes other than tritium. The latter species is very minor and totals only about 0.2% of the total 3He. Separating these different components of the 3He requires additional measurements of the 4He abundance which comprise:

where the subscripts are the same as those for 3He. In the case of 4Herad, a product of uranium-thorium decay, the abundance can be significant where older waters are involved (e.g. >1000 years old) and has been used numerous times as an independent groundwater age measurement due to its steady state accumulation. This comparison is important because the 3He/4He ratio relative to equilibrium will be less than 1.0 if appreciable radiogenic 4He exist. This is due to the accumulation of 4He from uranium-thorium decay, where there are indications of radiogenic 4He we can correct for it in the age calculations.>

Noble Gas Abundance

The noble gases of helium, neon, argon, krypton, and xenon naturally occur at trace abundance in the atmosphere. They also dissolve in groundwater during recharge. Their concentration in groundwater is controlled by 1) equilibrium solubility and 2) incorporation of excess air. The solubility of the noble gases in non-turbulent, free-standing water is temperature dependent, with increasing solubility with decreasing temperature. This temperature dependency is most pronounced in the argon, krypton, and xenon concentration (Fig. 13).

Figure 13. Solubility of noble gases in water at various temperatures can be used to calculate groundwater recharge temperatures. Figure from Mazor (1991); see for examples and further discussion.

The curves in Figure 13 provide a means to calibrate measured dissolved noble gas abundances in groundwater against its recharge temperature. During most groundwater recharge, the mean soil temperature dictates the equilibrium noble gas concentrations dissolved in the recharging water, which in most regions is around 2°C greater than the mean annual air temperature.

Dissolved noble gas abundances in groundwater other than helium that exceed an equilibrium amount are due to dissolution of excess air. Incorporation of excess air into recharged groundwater is thought to occur when air in the vadose zone is trapped by a plug of recharge water and is transported to sufficient depths that it is dissolved. Groundwater recharged through a vadose zone likely has excess dissolved air. In almost all cases the composition of the excess air is the same as the atmosphere (Heaton et al., 1981). Therefore, the amount of noble gases dissolved in groundwater above the equilibrium amount is a simple arithmetic addition of each noble gas from the atmosphere. Therefore, the amount of each dissolved noble gas relative to each other within a single sample should reflect a single equilibrium solubility temperature at the time of groundwater recharge. The amount of excess air dissolved in a groundwater can also provide qualitative information about the type of groundwater recharge. For instance, high excess air content may suggest recharge by a periodic "piston" flow under vadose zone conditions. Little excess air may suggest recharge with a limited vadose zone such as in river or lake infiltration.

The remaining noble gas effect that requires some consideration is the build-up of radiogenic 4He. There is a constant flux toward the ground surface of 4He derived from radioactive decay of uranium and thorium in the Earth’s crust that, given enough time, can accumulate in groundwater. Typically, groundwater that is thousands of years old will have an appreciable amount of radiogenic 4He, while young groundwater (less than 100 years old) has little or none except in special conditions such as close proximity to large-scale active faults.

To test for the presence of radiogenic 4He, the other noble gas abundances must be measured and calibrated to a recharge temperature. With this recharge temperature, the 4He content can be predicted based on equilibrium solubility. Any 4He that is above this predicted amount can be attributed to radiogenic 4He, and subsequently subtracted (Fig. 14). This will provide a revised 3He/4He ratio that can be used for calculating the groundwater age.

Figure 14. A ratio of measured 4He to measured 20Ne compared to measured 40Ar to 20Ne in an example California groundwater aquifer shows excess amount of 4He buildup for 5 of the 8 measurements above what predicted for equilibrium solubility (dashed line) relative to the other measured noble gases. After Davisson and Rose (2014).

Stable Isotopes

The stable isotope measurements of oxygen-18/oxygen-16 (18O/16O) and deuterium/hydrogen (D/H; deuterium is hydrogen-2) ratios in water are used to delineate different water populations in recharged groundwater. The measured 18O/16O and D/H ratios are normalized to a recognized standard and the converted results are reported in δ notation (pronounced "del"), where

The 18O/16Ostd and D/Hstd are the isotopic ratios of "Standard Mean Ocean Water" (SMOW). A δ value is a per mil (or parts per thousand) deviation from the standard.

Natural processes can preferentially favor or exclude particular isotopes, because differences in isotopic mass give rise to differences in the thermodynamic stability of the water molecules that incorporate them. Significant effects arise during phase transitions (i.e., vapor, water, ice), where the isotopes with highest mass tend to be favored in the most-condensed phase For example, at 25°C the measured δ18O value of water is 9.3 per mil higher than coexisting equilibrium vapor. This small difference is actually large when compared to the typical measurement precision of 0.1 per mil. As a result of this and other processes, a given body of water has a particular isotopic composition, or “fingerprint”, that in many cases is sufficiently distinctive to allow it to be identified and traced.

The D/H and 18O/16O ratios of ocean water are remarkably uniform worldwide, owing to global circulation patterns. However, since all continental precipitation originates from the ocean, isotopic partitioning occurs between water phases, and because continental storm fronts are isolated from the ocean and behave as closed systems, the isotopic ratios of measured precipitation varies systematically. This variation is almost exclusively driven by elevation difference and distance inland from the ocean. An example of δ18O variations in precipitation across British Columbia are illustrated below in Figure 15a. Figure 15b shows how shallow groundwater collected on the western slope of the Sierra Nevada record this systematic δ18O variation in its recharge.

Figure 15. Left hand image shows systematic variation of δ18O values in precipitation across British Columbia (from Yonge et al., 1989). Right hand image shows how shallow groundwater records this systematic variation on the western slope of the Sierra Nevada (from Rose et al., 1996).

The method for comparing the isotopic character of different waters lies in the use of a δD-δ18O plot of the isotope ratios. A plot of δD vs. δ18O values provides a graphical means to distinguish various populations of data relating to different water masses of different origins (Fig. 16).

Figure 16. General δD-δ18O plot showing the Meteoric Water Line (MWL) and the effects of evaporation on natural waters. The slope of the evaporation line can vary between 2 and 6 and depends on the ambient temperature and humidity. The MWL has a constant slope of 8 for global precipitation.

Also on this plot lies what is referred to as the Global Meteoric Water Line (MWL), a linear regression through the values of various unevaporated precipitation collected world-wide, which results in an empirical equation of δD = 8δ18O + 10. The slope of this line originates from the fact that isotopic partitioning of deuterium between water vapor and liquid is approximately 8 times greater than for 18O. Since global precipitation forms a slope of 8 indicates that cloud water establishes isotopic equilibrium between vapor and liquid.

However, when liquid water evaporates from the surface of water body, a non-equilibrium partitioning develops between the relative deuterium and 18O abundances, causing isotopic enrichment of the remaining liquid water. On a δD-δ18O plot, progressive evaporation causes a shift of the remaining liquid to the right of the MWL along a straight line (see Fig. 16). The slope of this evaporation line depends on temperature and humidity of the surrounding air. The proximity of an evaporated isotopic value relative to the MWL is proportional to the extent of evaporation or isotopic enrichment.

In short, isotopic and geochemical data have found great application in identifying water sources, tracing their movement and mixing, and determining water age. These methodologies are commonly more precise, and less expensive, than methods that involve drilling and testing.


Busenberg E and LN Plummer, 1992, Use of chlorofluorocarbons (CCl3F and CCl2F2) as hydrologic tracers and age-dating tools: The alluvium and terrace system of Central Oklahoma. Water Resour. Res. 28, 2257-2283.

Cook PG and DK Solomon, 1997, Recent advances in dating young groundwater: chlorofluorocarbons, 3H/3He, and 85Kr. J. Hydrol. 191, 245-265.

Criss, R.E., 1999, Principles of Stable Isotope Distribution. Oxford University Press: New York, 264 pgs.

Davisson, M.L., Hudson, G.B., Esser, B.K., Ekwurzel, B., Herndon, R.L., 1999, Tracing and age-dating recycled waste water recharged for potable reuse in a seawater injection barrier, southern California, USA. International Atomic Energy Agency Symposium on Isotope Hydrology in Water Resources Management, May 1999, Vienna.

Davisson, M.L., T.P. Rose, 2014, Recharge and flow in the Medicine Lake Volcano – Fall River Springs groundwater basin, California. J. Environ. Forensics, 15, 66-77

Ekwurzel B, P Schlosser, WM Smethie Jr, LN Plummer, E Busenberg, RL Michel, R Weppernig, and M Stute, 1994, Dating of shallow groundwater: Comparison of the transient tracers 3H/3He, chlorofluorocarbons and 85Kr. Water Resour. Res. 30, 1693-1708.

Fisher DA and PM Midgley, 1993, The production and release to the atmosphere of CFCs 113, 114, 115. Atmos. Environ., 27A, 271-276.

Fontes, J.Ch., 1980, Environmental isotopes in groundwater hydrology. In Fritz, P., Fontes, J.Ch. (eds.) Handbook of Environmental Isotope Geochemistry. Vol. 1 Elsevier: New York, 75-140.

Fraser P, D Cunnold, F Alyea, R Weiss, R Prinn, P Simmonds, B Miller, R Langenfelds, 1996, Lifetime and emission estimates of 1,1,2-trichlorotrifluorethane (CFC-113) from daily global background observations June 1982-June 1994. J. Geophys. Res. Atmos., 101, 12585-12599.

Gamlen PH, BC Lane, PM Midgley, 1986, The production and release to the atmosphere of CCl3F and CCl2F2 (chlorofluorocarbons CFC 11 AND CFC 12). Atmos. Environ., 20, 1077-1085.

Hasenmueller, E. A. and Criss, R. E. (2013) Water balance estimates of evapotranspiration rates in areas with varying land use. in Evapotranspiration-An Overview, S.G. Alexandris, ed., ISBN 978-953-51-1115-3, Ch. 1, p. 1-21.

Heaton T.H.E. and Vogel J.C., 1981, "Excess air" in groundwater. J. Hydrol., 50, 210-216.

Mazor, E., 1991, Applied Chemical and Isotopic Groundwater Hydrology. Halsted Press: New York, 274 pgs.

McCarthy RL, FA Bower, JP Jesson, 1977, The fluorocarbon-ozone theory – I. Production and release – world production and release of CCl3F and CCl2F2 (fluorocarbons 11 and 12) through 1975. Atmos. Environ., 11, 491-497.

Wisegarver DP and RH Gammon, 1988, A new transient tracer: measured vertical distribution of CCl2FCClF2 (F-113) in the North Pacific subarctic gyre. Geophys. Res. Lett., 15, 188-191.