Published in Vadose Zone Journal 4:72-78 (2005)
© 2005 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
ORIGINAL RESEARCH
Chloride Mass Balance
Cautions in Predicting Increased Recharge Rates
G. W. Geea,*,
Z. F. Zhanga,
S. W. Tylerb,
W. H. Albrightb and
M. J. Singletonc
a Pacific Northwest National Laboratory, Richland, WA
b University of Nevada and Desert Research Institute, Reno, NV
c Lawrence Berkeley Laboratory, Berkeley, CA
* Corresponding author (glendon.gee{at}pnl.gov)
Received 21 February 2004.
 |
ABSTRACT
|
|---|
The chloride mass balance (CMB) method has been used extensively to estimate recharge in arid and semiarid environments. Required data include estimates of annual precipitation, total Cl– input (from dry fallout and precipitation), and pore water Cl– concentrations. Typically, CMB has been used to estimate ancient recharge, but recharge from recent land-use change has also been documented. Recharge rates below a few millimeters per year are reliably detected with CMB; however, estimates above a few millimeters per year appear to be less reliable. We tested the CMB method against 26 yr of drainage from a 7.6-m-deep lysimeter at a simulated waste burial ground located on the Department of Energy's Hanford Site in southeastern Washington, USA where removal of vegetation has increased recharge rates. Measured drainage from the lysimeter for the past 26 yr averaged 62 mm yr–1. Precipitation averaged 190 mm yr–1 with an estimated Cl– input of 0.22 mg L–1. Initial pore water Cl– concentration was 88 mg L–1 and decreased to about 6 mg L–1 after 26 yr, while the drainage water Cl– concentration decreased to <1 mg L–1. A recharge estimate made using Cl– concentrations in drain water was within 26% of the measured drainage rate. In contrast, recharge estimates using 1:1 (water/soil) extracts were lower than actual values by factors ranging from 2 to 8 or more. The results suggest that when recharge is above a few millimeters per year, soil water extracts can lead to unreliable estimates of recharge.
Abbreviations: CMB, chloride mass balance • HMS, Hanford Meteorological Station
|
INTRODUCTION
|
|---|
IN THE CMB METHOD, measurements of Cl– in pore water are used to estimate the recharge rate when both precipitation and Cl– inputs are known. The CMB for a soil profile at steady state can be written as
 | [1] |
where P is the average annual precipitation (mm yr–1); Clp is the average Cl– input from all sources, including wet and dry fallout (mg L–1); Cls is the average Cl– concentration of pore water below the root zone (mg L–1); and R is the average annual recharge rate (mm yr–1).
Key assumptions are (i) steady influx of water and Cl–; (ii) steady, vertical efflux of Cl– below the root zone; (iii) no soil sources or sinks for Cl–; and (iv) piston flow of Cl– such that point measurements of solute concentrations can be used to represent a true spatial average of the soil Cl– flux. The CMB has been used to estimate recharge in arid and semiarid regions throughout the world (Allison et al., 1994; Prych 1995; Murphy et al., 1996; Hendrickx and Walker 1997; Scanlon et al., 1997; Scanlon 2000). The CMB method appears to be useful in estimating paleoclimate recharge rates dating back thousands of years (Stone 1984; Murphy et al., 1996; Tyler et al., 1996), but CMB also has been used for estimating modern recharge rates, including those that have increased in response to land-use change, specifically where vegetation had been altered and deep-rooted trees were replaced with shallow-rooted grasses (Jolly et al., 1989; Walker et al., 1991). Because of its simplicity, CMB is an attractive method. With proper assumptions about average annual precipitation and Cl– inputs, the only direct measurement required is the volume-averaged Cl– concentration in the pore water.
The general principles of the CMB method have been known and practiced in irrigation management for years (USDA, 1954). In a manner similar to the CMB method, salt mass balance in irrigation water is often expressed as the leaching requirement (LR), defined as the applied water volume (Q) times the input salt concentration (Ci) divided by the drainage volume (D) times the salt concentration (Cs) of the drainage water; that is,
 | [2] |
For an LR of 1 and considering only Cl–, Eq. [1] is identical to Eq. [2].
Pore water Cl– is typically obtained from soil samples taken at depths well below the root zone and analyzed first for water content (gravimetrically) and then for Cl– with Cl– determined typically from a 1:1, 2:1, or sometimes 3:1 (solution/soil) extract (Murphy et al., 1996; Scanlon and Goldsmith, 1997; Tyler et al., 1999; Scanlon, 2000). The Cl– concentration, Cls (mg Cl L–1 soil solution), is subsequently computed by dividing the measured extract concentration by the gravimetric water content (g water/g oven-dry soil) and multiplying by the solution/soil ratio such that
 | [3] |
where Clext is Cl– concentration in extract water, w is gravimetric water content, and n is solution/soil ratio (mass of water added/mass of oven dry soil).
For example, with a measured Cl– concentration, [Clext] of 0.025 mg L–1 (25 ppb), a gravimetric water content, w, of 0.05, and a soil solution ratio, n, of 2, the pore water concentration is calculated to be 1 mg L–1, indicating, in this case, that the extract water is 40 times more dilute than the actual pore water. It is this dilution that is of concern particularly at low pore water Cl– concentrations.
The use of pore water extracts works best at high Cl– concentrations and low recharge rates because, as the Cl– concentration approaches the detection limit of the analytical method used, analytical errors significantly increase the uncertainty of the recharge estimate (Tyler et al., 1999). Also, at low concentrations, systematic contamination of sample and extracts can become an important source of error. The better analytical methods have an operational resolution for Cl– of around 0.02 mg L–1 (20 ppb), so it should be apparent that for estimates of soil pore water Cl–, uncertainties mount rapidly as the extract concentrations drop below 1 mg L–1 and analytical errors become more prominent. In addition, treatment of a soil low in Cl– with deionized water can potentially release mineral Cl– (Murphy et al., 1996). Another consideration for soils with significant clay content is the possibility of anion exclusion, which could adversely affect the recharge estimate using the CMB method (Gee and Hillel 1988). Due to anion exclusion, pore water samples may not represent the true area-averaged Cl– value. The impact is greatest when the water content is lowest. For clay-rich soils, pore water samples obtained from direct extraction (e.g., centrifugation) would be impacted the most.
There is ample evidence that a range of conditions including unstable (e.g., finger) flow, funnel flow, and preferential flow through worm holes or root channels, may cause highly nonuniform flow of Cl– in and below the root zone. Allison et al. (1994) provided examples of extreme heterogeneities in field soils giving rise to highly variable Cl– concentrations. McCord et al. (1997) showed that textural variations can lead to zones of slowly leached soil under steady-flux conditions. Roth (1995) also showed that flow channeling can occur in mildly heterogeneous soils, which can significantly increase the distance over which solute concentrations stabilize and reflect reliable estimates of solute mass balance. The successful application of CMB at a given site is therefore dependent on site conditions that minimize preferential flow of solutes.
Chloride mass balance has been applied for years to estimate recharge, but in situations where there has been land-use change and recharge has increased, there have been few attempts to compare CMB with any independent or direct estimates of recharge. In this study, we tested the CMB method for predicting recharge in a semiarid climate setting, where land use has changed (i.e., soil disturbed and vegetation removed) and where the Cl– flux could be quantitatively checked with lysimetry. Lysimetry is a method that can be used to directly measure the percolation of water through soils and determine both the flux rate and soluble constituents removed in the drainage. (Gee and Hillel 1988; Gee et al., 2003).
|
MATERIALS AND METHODS
|
|---|
The study was conducted at the U.S. Department of Energy's Hanford Site, near Richland, Washington, USA. The Hanford Site has a northern, steppe (semiarid) climate typified by dry, hot summers and cool, wet winters (Hoitink et al., 2003). For the past 25 yr, the precipitation at the Hanford Meteorological Station (HMS) has averaged 180 mm yr–1, about two-thirds of this amount coming in winter months (Nov.–Mar). Potential evaporation, controlled by the aridity of the climate, is about 1600 mm yr–1, nearly nine times the annual precipitation (Gee et al., 1989). Actual evaporation (AE) is significantly different than potential at Hanford. For undisturbed sites, with shrub-steppe vegetation, AE is approximately equal to annual precipitation, so we would expect little drainage. Chloride measurements made at the Hanford Site in areas of undisturbed shrub-steppe vegetation growing on coarse soils have shown significant bulges of high Cl– (>100 mg L–1) at shallow depths, with corresponding recharge rates estimated to be much less than 1 mm yr–1 (Prych, 1995; Murphy et al., 1996; Fayer et al., 1999). In contrast, for disturbed sites with little or no vegetation and coarse soils, AE can be less than two-thirds of the annual precipitation (Gee et al., 1992), resulting in drainage rates that exceed 50 mm yr–1. The corresponding Cl– distributions are expected to be low at the disturbed sites containing coarse soils.
The HMS is located about 15 km northwest of the simulated waste site. Continuous records of precipitation have not been kept at the lysimeter test site because of a hiatus in project funding over the years. However, precipitation at the lysimeter test site was previously found to be about 6% more than at the HMS (Gee 1987), so we have estimated the precipitation to be 190 mm yr–1 for the past 26 yr. Chloride input from wet and dry fallout has been studied extensively by Murphy et al. (1996) and found to range from 0.22 to 0.23 mg L–1. For the purposes of this study we selected 0.225 mg L–1 for the Cl– input at the lysimeter site. We recognize that the resolution of the Cl– input is uncertain and later show how increases in this value can affect the CMB results.
Soil samples were taken at a simulated waste burial ground about 5 km north of Richland from two sand-filled, 7.6-m-deep lysimeters that were kept vegetation free for the past 26 yr. The sand is screened material containing only 1% gravel (material between 2 and 10 mm) taken from the lysimeter excavation and is Hanford formation material that is largely granitic in origin (Baker et al., 1991). The lysimeter soil is classified as a sand with 1% gravel, 95% sand, 3% silt, and 1% clay. The hydraulic conductivity is 2.0 x 10–3 cm s–1 at saturation and is 1.0 x 10–7 cm s–1 at –5 kPa. The volumetric water content at –5 kPa is 0.10 m3 m–3. Chloride analysis performed on samples of lysimeter soils taken at the time of construction showed the initial Cl– concentration of the pore water in the lysimeter soil was 88 mg L–1. The initial soil water content in the lysimeter was 0.06 m3 m–3. The lysimeters were installed in 1978 and began draining in 1981 (Gee et al., 1989; Tyler et al., 1999). It should be noted that 3 yr between installation and the inception of drainage were required for the soil to wet up to the point of drainage (
0.10 m3 m–3). Figure 1
shows the cross section of the lysimeters along with an instrument caisson that provided entry for collecting drainage water and for soil sampling. Most of the sampling was done in the south lysimeter (Fig. 1) by coring through the side ports (1996, 1998) or by direct coring using a split-spoon sampler (2002). Lysimeter drainage was measured in two ways. From the inception of drainage in 1981 until April 2000, drainage was measured by periodically collecting water in tared containers, returning the containers to the laboratory where they were manually weighed on an electronic balance to a resolution of 1 g (precision of <0.001 mm). In April 2000, the drainage collection was switched to an automatic sampling scheme. A tipping spoon (Pronamic Ltd., Silkeborg, Denmark) rain gauge with a resolution of 4.7 mL per tip was installed and used thereafter. While the precision of the drainage collected by the tipping spoon is considerably less than that obtained by manual weighing, the ability to constantly monitor the drainage and electronically display the cumulative drainage volume has proved to be a distinct advantage. Periodic checks of the tipping spoon showed that the calibration has not changed in time and the precision of the measured drainage is well within 10% of the true value. During the past 20 yr, the lysimeter drainage from the bare soil has averaged 34% of the total precipitation, or about 55% of the winter precipitation.
View larger version (50K):
[in this window]
[in a new window]
|
Fig. 1. Schematic of the cross section of two deep lysimeters at a simulated waste burial ground at the USDOE Hanford Site near Richland, WA.
|
|
In December 1996, soil cores were taken from the south lysimeter at 10 depths, from the surface to a depth of 7.1 m. Each soil sample was analyzed for water content, and then a 1:1 (solution/solid) extract was prepared and analyzed for Cl–. Similarly, samples were taken in March 1998 and then again in September of 2002. Samples taken in 1996 and 1998 were analyzed by the Desert Research Institute Hydrology Group Laboratory (Reno, NV). Samples taken in 2002 were analyzed by the Earth Sciences Group at Lawrence Berkeley Laboratory (Berkeley, CA). All samples were analyzed using similar equipment (Dionex chromatograph, Sunnyvale, CA) and best available laboratory practices. Estimated precision for the 1998 samples was 0.02 and 0.05 mg L–1 for the 1996 and 2002 analyses. Samples taken in 1998 were not oven dried but brought to 1:1 dilution using known water contents obtained from subsamples. The other two sampling sets (1996 and 2002) were oven dried and analyzed in a manner similar to that reported by Murphy et al. (1996).
|
RESULTS AND DISCUSSION
|
|---|
Table 1 shows the Cl– concentrations for the soil samples and drainage water collected at three dates. Recharge estimates from soil cores are presented in Table 2 along with recharge estimates based on Cl– in the drainage water. These values are compared with the measured drainage rate. It is apparent that the soil-core estimates underestimate the recharge by a factor of 5 or more for samples taken in 1996 and 2002 and by a factor of 2 or more for samples taken in 1998.
View this table:
[in this window]
[in a new window]
|
Table 1. Chloride concentrations found in soils and drain water taken from the south lysimeter at selected times.
|
|
View this table:
[in this window]
[in a new window]
|
Table 2. Comparison of measured and estimated drainage rate from Cl– mass balance recharge estimates using drainage water and 1:1 extract (pore water) data sets. Assumes 190 mm yr–1 average precipitation with 0.225 mg L–1 Cl– input or 0.275 mg L–1 Cl– input.
|
|
There are a number of uncertainties associated with the CMB method. One of which is the total Cl– input. Chloride input was not measured directly but was derived from estimates made by Murphy et al. (1996). Table 2 shows that if the Cl– input is increased 22%, from 0.225 mg L–1 to 0.275 mg L–1, agreement is achieved between the 26-yr average lysimeter drainage rate and that estimated from the CMB method using the drainage-water Cl– as an estimate of the true pore water Cl–. While this modification can explain the difference between the CMB estimate using the drain water, it does not explain the large underestimation of recharge using the soil water extracts, because (i) it would take an unreasonably high Cl– input value of above 1 mg L–1 and (ii) if such a high Cl– input value was used it would make the drain water correspondingly too high (above 3 mg L–1).
We looked at possible impacts of interaction of the concrete floor as a possible sink for Cl–. Samples of concrete were collected from the floor in the access caisson, and Cl– water at 1 mg L–1 was equilibrated with the concrete. The analysis of the Cl– water after contact with the concrete showed no material loss of Cl– so we determined that the concrete was not creating a sink for Cl– and that the drainage water reflected the true pore water concentration.
In an attempt to explain the underestimation of the recharge using the soil water extracts, we compared the soil pore water estimates with numerical simulations. The simulations were done for a period of 26 yr (1978–2003). Using the available soil hydraulic properties from the lysimeter and assuming a constant recharge flux of 62.5 mm yr–1, numerical simulations of the leaching of the lysimeter were conducted using STOMP (White and Oostrom, 2004), which is designed to solve a variety of nonlinear, multiple-phase, multidimensional flow and transport problems for unsaturated porous media. The transport was modeled as a one-dimensional process, and we assumed transport to be controlled by advection and dispersion alone. The simulation domain was discretized into 750 nodes with node spacing of 0.01 m. The initial Cl– concentration in the lysimeter in 1978 was taken to be 88 mg L–1 based on Cl– analysis of archived soil from the lysimeter, and the Cl– input was assumed to be steady at 0.684 mg L–1, on the basis of the assumptions of 190 mm of annual precipitation, 0.225 mg L–1 input, and a drainage rate of 62.5 mm yr–1. After drainage started in 1981, the water storage was nearly constant, there was no runoff, and the average evaporation rate was 127.5 mm yr–1. The simulations were run in an isothermal mode. The lower boundary condition was assumed to be unit gradient, based on the observed unit gradient conditions measured with tensiometers (Gee, 1987; Sisson et al., 2002). We ran the simulations two ways. Case 1 assumed the drainage was steady at 62.5 mm yr–1. Case 2 assumed the net infiltration varied with precipitation input, with the net infiltration mimicking the variability observed in the annual drainage rates. We varied the water storage before 1981 and the drainage thereafter by assuming that the net infiltration was a function of the winter precipitation. Gee (1987), Gee et al. (1989), and Tyler et al. (1999) reported annual drainage rates from the test lysimeter that vary from <50 mm yr–1 to more than 100 mm yr–1 depending on extreme variations in winter precipitation. Variations of drainage since 2000 can be seen at http://vadose.pnl.gov/300n/drainage2.asp (as of July 2004; Pacific Northwest National Laboratory, 2004). Figure 2
shows the measured cumulative drainage from the lysimeter compared with that simulated for Case 1 (steady) and Case 2 (variable drainage rates). Note that the measured drainage exhibit periods when drainage was temporarily stopped (because of irregular sampling before 2000). Less noticeable but also present are short periods after 2000 when air locks in the drainage line temporarily stopped drainage. The periods of hiatus in drainage are artifacts of the method of collection and do not reflect the actual lysimeter drainage rate. During periods when drainage stopped, water was simply stored in the gravel base until the collection system was operational again and drainage commenced. Figure 2 illustrates that the variable case (Case 2) more closely matches the measured values. Figure 3
shows the simulated Cl– concentrations in the drainage water as a function of time, and Fig. 4
shows the Cl– concentration profiles at two selected times. The variable net infiltration case (Case 2) yielded the lowest Cl– concentrations and approached a steady-state low value more quickly than when the net infiltration was assumed to be steady. However, by 1998 and subsequently as late as September 2002, the drainage water concentrations simulated for both cases ranged from 0.8 to 0.7 mg L–1. The simulations for Cases 1 and 2 show that after 1998, the drainage water was within 20% of the assumed net Cl– input (0.7 mg L–1). The simulated values were close to the measured Cl– values in the drainage water, which for the same time period ranged from 0.8 to 1.0 mg L–1. After drainage started in 1981, the water storage was nearly constant and there was no runoff from the lysimeter.
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 2. Cumulative drainage measured from the 7.6-m-deep south lysimeter, compared with simulated drainage under steady (Case 1) and variable drainage (Case 2) conditions. Measured values are shown as the solid line, Case 1 as crosses, and Case 2 as open circles.
|
|
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 3. Time course of the simulated Cl– concentration at the south lysimeter. Case 1 (steady drainage) open box symbol, Case 2 (variable drainage) cross symbol.
|
|
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 4. Simulated Cl– concentration profiles in March 1998 and September 2002 under steady (Case 1) and variable drainage (Case 2) conditions.
|
|
To further investigate the discrepancy between the soil Cl– concentrations and the observed drainage flux, the transport was modeled using a mobile–immobile water transport approach (van Genuchten and Wagenet, 1989). Simulations were conducted with CXTFIT (Toride et al., 1995) under conditions of steady drainage (65 mm yr–1) over the first 16 yr of drainage. Previous laboratory data (Gee and Campbell, 1980) suggested that the mobile water/total water content ratio was approximately 0.65. Several simulations using this ratio failed to produce any similarities to the observed drainage and soil sample concentrations. The first simulation consisted of assuming a very slow exchange coefficient (
= 1 x 10–5 yr–1), and a ratio of mobile water to total water content of 0.65 was able to match the drainage water concentration, but soil water Cl– (resident) concentrations were much higher than observed. The exchange coefficient was slowly decreased in subsequent cases; however, this produced higher concentrations of Cl– in the drainage water than observed, and also resulted in Cl– concentrations varying down the length of the lysimeter at the end of 16-yr simulation period. Additionally, the Cl– concentrations in the drainage water changed significantly with time, as opposed to the observed concentrations that are essentially unchanging in time between 1996 and 2002. Decreasing the mobile/total water content ratio only produced much higher soil water Cl– concentrations and suggested that preferential flow though a small portion of the lysimeter was not a viable hypothesis.
Only when we increased the ratio of mobile to total water content to 0.9 and chose an extremely slow exchange between the mobile and immobile phase ( = 1 x 10–5 yr–1) could simulated soil water and drainage water concentrations be well matched to those observed in the 1:1 dilutions and drainage water. Under this simulation, the Cl– in the immobile phase is essentially unavailable for transport. These simulation results produced a fairly uniform pore water Cl– concentration that changed only very slowly with time and a Cl– concentration in the drainage water that also did not vary significantly with time, after the initial flush of Cl–.
The parameters used above suggest that a small portion of the Cl– in pore water is inaccessible to the recharge waters. The lack of significant temporal change in either the soil water Cl– or drainage water, combined with the fairly uniform vertical distribution of soil water Cl–, suggests that the majority of pore water is mobile yet disconnected from a small mass of soil water Cl–.
The small mass of immobile water is consistent with an incomplete flushing of the vadose zone by the infiltration and is consistent with that seen by Jolly et al. (1989) following a change (increase) in recharge. The nature of the incomplete flushing could be either microscopic, with remnants of Cl– remaining in dead end pores that are ubiquitous, or as large unflushed areas resulting from large-scale preferential flow. As will be discussed below, large-scale preferential flow is very unlikely to be occurring in the lysimeter. Rather, the findings of elevated soil water Cl–, as compared with the drainage water, is consistent with widely dispersed storage of small amounts of residual Cl– originally present when the soils were emplaced in the lysimeters.
We ruled out preferential flow as an explanation of the high Cl– concentrations found in the soil samples. The Cl– concentrations of side port samples taken in 1996 were nearly the same as those found in borehole samples taken from the center of the lysimeter in 2002. If there were wall effects or other preferential flow occurring, it should have shown up as marked differences in the two tests.
The reported 1:1 soil Cl– concentrations are always too high to correctly predict the drainage. To match the measured recharge rate, the 1:1 extract Cl– concentrations would have to be 0.041 mg L–1 or lower, compared with measured values, which were as high as 0.5 mg L–1 or more. Table 3 shows what the pore water and 1:1 concentrations would be for various recharge rates. It is clear that as the recharge rate increases, the pore water Cl– becomes more dilute. At 100 mm yr–1, 1:1 extracted Cl– is at the detection limit for the Dionex ion chromatograph.
View this table:
[in this window]
[in a new window]
|
Table 3. Recharge rates and estimated Cl– concentrations expected for 1:1 extracts under Hanford climate and estimated Cl– input.
|
|
We have considered mineral dissolution as a possible mechanism for some of the high Cl– numbers. Fayer et al. (1999) reported Cl– concentrations in Hanford formation minerals ranging from 100 to 230 ppm. If 0.1% of the Cl– leached from the mineral, this would account for a soil solution concentration ranging from 0.10 to 0.23 mg L–1. Such levels of Cl– could possibly have leached from the samples, particularly after they had been dried and subsequently wetted with deionized water. The fact that the samples in 1998 were not oven dried and gave the best results suggests that oven drying may release Cl– from the sediments tested. While we have not exhaustively tested this hypothesis, we suspect that mineral dissolution may contribute to the high Cl– values obtained with the 1:1 extracts. Additional Cl– contamination from outside sources at the very low levels of Cl– is also possible.
As a result of these observations, we have concerns about using soil pore water sampling in the CMB method for recharge rates much above a few millimeters per year at the Hanford Site and possibly at other locations where slow mineral dissolution or other sources of Cl– release can confound the low concentration values of the soil Cl–. It is further recommended that minimum soil water dilutions be used when measuring Cl– concentrations in soils to reduce the impacts of analytical errors and possible dissolution. Dilution errors may be reduced by using less than 1:1 extract ratios, by centrifuging field-most samples (when the sample is wet enough), or by obtaining pore water directly in the field using solution sampling or wick lysimetry (Gee et al., 2003). It is clear for our study that at soil pore water concentrations below a few milligrams per liter, Cl– sources other than from precipitation and fallout may contribute to errors in estimating recharge using the CMB method.
In summary, Cl– concentrations found in drainage waters from our test lysimeter were in good agreement with those required in CMB estimates to accurately predict lysimeter drainage rates. In contrast, the soil pore water Cl– was always elevated with respect to the drainage water. Mass balance predictions using the soil pore water Cl– underpredicted the measured drainage (recharge) by factors of 2 to 8 or more. Modeling of drainage, assuming no immobile Cl–, yielded Cl– drainage-water concentrations consistent with observations. However, for the same simulations, modeled soil Cl– profiles were in poor agreement with observed pore water Cl–. When mobile Cl– was reduced to 0.9 and a low exchange-coefficient was simulated, soil Cl– profiles were more consistent with observed values of the pore water Cl–. Slow dissolution of mineral Cl– is a possible explanation for the observed elevated soil pore water Cl– and consistent with the 0.9 mobile (0.1 immobile) Cl– transport analysis. In addition, dilution of pore waters low in Cl– using standard 1:1 or 2:1 (solution/solid) methods can cause the measured Cl– to approach the analytical detection limits and give rise to additional uncertainties in reported Cl– concentrations.
|
CONCLUSIONS
|
|---|
The CMB method was tested using a deep lysimeter at a site where vegetation removal increased recharge from a very low rate (less than a few millimeters per year) to a much higher rate (>60 mm yr–1). Using the Cl– data from the lysimeter drainage as the true estimate of the pore water, recharge estimates agreed within 26% of the measured drainage rate. In contrast, soil samples subjected to leaching (1:1) yielded pore water estimates that always underestimated the drainage by a factor of two (i.e., 200%) or more. The data show that the CMB method can be used where land use has caused increased recharge rates over relatively short time periods (decades) but that collecting pore water samples from drainage is preferred over conventional soil water extraction procedures. The apparent error in pore water Cl– concentrations is attributed to dilution effects of the extract water coupled with mineral dissolution and possibly other external sources of Cl– in sampling low concentration Cl–. When recharge rates are much above a few millimeters per year, it appears that dilution of the pore water Cl– resulting from the 1:1 extraction technique, and possible Cl– contamination from mineral dissolution or others sources, can cause pore water extracts to give unreliably low estimates of recharge.
|
ACKNOWLEDGMENTS
|
|---|
We thank Ray Clayton and Karen Waters-Husted of Pacific Northwest National Laboratory for technical support in collecting and pretreating the soil samples. The study was conducted as part of an overall recharge evaluation effort for the Hanford Site, and the work was supported by the Remediation Decision Support Task of the Groundwater Remediation Program and the Remediation and Closure Project sponsored by the U.S. Department of Energy under contract DE-AC06-76RL01830.
|
REFERENCES
|
|---|
- Allison, G.B., G.W. Gee, and S.W. Tyler. 1994. Vadose zone techniques for estimating groundwater recharge in arid and semiarid regions. Soil Sci. Soc. Am. J. 58:6–14.[Abstract/Free Full Text]
- Baker, V.R., B.N. Bjornstad, A.J. Busacca, K.R. Fecht, E.P. Kiver, U.L. Moody, J.G. Rigby, D.F. Stradling, and A.M. Tallman. 1991. Quaternary geology of the Columbia Plateau. p. 215–250. In R.B. Morrison (ed.) Quaternary nonglacial geology: Conterminous U.S. The geology of North America. v. K-2. Geological Society of America, Boulder, CO.
- Fayer, M.J., E.M. Murphy, J.L. Downs, F.O. Khan, C.W. Lindenmeier, and B.N. Bjornstad. 1999. Recharge data package for the immobilized low-activity waste 2001 performance assessment. PNNL-13033. Pacific Northwest Natl. Lab., Richland, WA.
- Gee, G.W. 1987. Recharge at the Hanford Site: Status report. PNL-6403. Pacific Northwest Natl. Lab., Richland, WA.
- Gee, G.W., and A.C. Campbell. 1980. Monitoring and physical characterization of unsaturated zone transport—Laboratory analysis. PNL-3304. Pacific Northwest Natl. Lab., Richland, WA.
- Gee, G.W., M.J. Fayer, M.L. Rockhold, and M.D. Campbell. 1992. Variation in recharge at the Hanford Site. Northwest Sci. 66:237–250.
- Gee, G.W., and D. Hillel. 1988. Groundwater recharge in arid regions: Review and critique of estimation methods. J. Hydrol. Proc. 2:255–266.
- Gee, G.W., M.L. Rockhold, and J.L. Downs. 1989. Status of FY 1988 soil-water balance studies on the Hanford Site. PNL-6750. Pacific Northwest Natl. Lab., Richland, WA.
- Gee, G.W., Z.F. Zhang, and A.L. Ward Ward. 2003. Modified water fluxmeter with solution collection capability. Available at www.vadosezonejournal.org. Vadose Zone J. 2:627–632.[Abstract/Free Full Text]
- Hendrickx, J.M.H., and G.R. Walker. 1997. Recharge from precipitation. p. 19–111. In I. Simmers (ed.) Recharge of phreatic aquifers in (semi-) arid areas. A.A. Balkema, Rotterdam, the Netherlands.
- Hoitink, D.J., K.W. Burk, J.V. Ramsdell, and W.J. Shaw. 2003. Hanford Site climatological summary 2002 with historical data. PNNL-14242. Pacific Northwest Natl. Lab., Richland, WA.
- Jolly, I.D., P.G. Cook, G.B. Allison, and M.W. Hughes. 1989. Simultaneous water and solute movement through an unsaturated soil following an increase in recharge. J. Hydrol. (Amsterdam) 111:391–396.
- McCord, J.T., C.A. Gotway, and S.H. Conrad. 1997. Impact of geologic heterogeneity on recharge estimation using environmental tracers: Numerical modeling investigation. Water Resour. Res. 33:1229–1240.
- Murphy, E.M., T.R. Ginn, and J.L. Phillips. 1996. Geochemical estimates of recharge in the Pasco basin: Evaluation of the chloride mass balance technique. Water Resour. Res. 32:2853–2868.
- Pacific Northwest National Laboratory. 2004. Vadose zone transport studies. Drainage: VZFS 300N: All data. Available at http://vadose.pnl.gov/300n/drainage2.asp (accessed July 2004; verified 20 Dec. 2004). PNNL, Richland, WA.
- Prych, E.A. 1995. Using chloride and chorine-36 as soil-water tracers to estimate deep percolation at selected locations on the U.S. Department of Energy Hanford Site, Washington. USGS Open-File Rep. 94-514. USGS, Tacoma, WA.
- Roth, K. 1995. Steady state flow in an unsaturated two dimensional, macroscopic homogeneous Miller-similar medium. Water Resour. Res. 31:2127–2140.
- Scanlon, B.R. 2000. Uncertainties in estimating water fluxes and residence times using environmental tracers in an arid unsaturated zone. Water Resour. Res. 36:395–410.
- Scanlon, B.R., and R.S. Goldsmith. 1997. Field study of spatial variability in unsaturated flow beneath and adjacent to playa lakes. Water Resour. Res. 33:2239–2252.
- Scanlon, B.R., S.W. Tyler, and P.J. Wierenga. 1997. Hydrologic issues in arid, unsaturated systems and implications for contaminant transport. Rev. Geophys. 35:461–490.
- Sisson, J.B., G.W. Gee, J.M. Hubbell, W.L. Bratton, J.C. Ritter, A.L. Ward, and T.G. Caldwell. 2002. Advances in tensiometry for long-term monitoring of soil water pressures. Vadose Zone J. 1:310–315.[Abstract/Free Full Text]
- Stone, W.J. 1984. Recharge in the Salt Lake coal field based on chloride in the unsaturated zone. Open File Rep. 214. N.M. Bureau of Mines and Miner. Resour., Socorro, NM.
- Toride, N., F.J. Leij, and M.Th. van Genuchten. 1995. The CXTFIT code for estimating transport parameters from laboratory or field tracer experiments. Res. Rep. 137. U.S. Salinity Lab., Riverside, CA.
- Tyler, S.W., J.B. Chapman, S. Conrad, D.P. Hammermeister, D.O. Blout, J.J. Miller, M.J. Sully, and J. Ginanni. 1996. Soil water flux on the Nevada Test Site: Temporal and spatial variations over the last 120,000 years. Water Resour. Res. 32:1481–1499.
- Tyler, S.W., B.R. Scanlon, G.W. Gee, and G.B. Allison. 1999. Water and solute transport in arid vadose zones: Innovations in measurement and analysis. In J.W. Hopmans and M.B. Parlange (ed.) Vadose zone hydrology: Cutting across disciplines. Oxford University Press, New York.
- USDA. 1954. Diagnosis and improvement of saline and alkali soils. Agric. Handb. 60. U.S. Salinity Laboratory Staff. U.S. Gov. Print. Office, Washington, DC.
- van Genuchten, M.Th., and R.J. Wagenet. 1989. Two-site/two region models for pesticide transport and degradation: Theoretical development and analytic solutions. Soil Sci. Soc. Am. J. 53:1303–1310.
- Walker, G.R., I.D. Jolly, and P.G. Cook. 1991. A new chloride leaching approach to the estimation of diffuse recharge following a change in land use. J. Hydrol. (Amsterdam) 128:49–67.
- White, M.D., and M. Oostrom. 2004. STOMP—Subsurface Transport Over Multiple Phases. Version 3.1. User's guide. PNNL-14478. Pacific Northwest Natl. Lab., Richland, WA.
This article has been cited by other articles:
|
|
|
|
 
G. W. Gee, M. Oostrom, M. D. Freshley, M. L. Rockhold, and J. M. Zachara
Hanford Site Vadose Zone Studies: An Overview
Vadose Zone J.,
November 20, 2007;
6(4):
899 - 905.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
