Published in Vadose Zone Journal 3:560-569 (2004)
© 2004 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
SPECIAL SECTION: UNCERTAINTY IN VADOSE ZONE FLOW AND TRANSPORT PROPERTIES
Application of a Darcian Approach to Estimate Liquid Flux in a Deep Vadose Zone
J. M. Hubbell*,a,
M. J. Nichollb,
J. B. Sissona and
D. L. McElroya
a Idaho National Engineering and Environmental Laboratory, Geosciences Research Department, P.O. Box 1625, MS 2107, Idaho Falls ID 83415
b Geosciences Dep., Univ. of Nevada, Las Vegas, NV 89122
* Corresponding author (jmh{at}inel.gov).
Received 10 July 2003.
 |
ABSTRACT
|
|---|
Approaches for estimating liquid flux in the shallow (0–2 m) vadose zone are hindered by the high degree of spatial and temporal variability present near the land surface. It is hypothesized that high-frequency variations in flux will be damped with depth. This study was conducted to estimate deep liquid flux using the Darcian approach at a waste disposal site in south-central Idaho that is underlain by a complex sequence of unsaturated basalt flows intercalated with thin sedimentary layers. Flux is estimated by combining in situ water potential measurements from sedimentary interbeds located at depths of 34 and 73 m below land surface (bls) with laboratory estimates for the unsaturated hydraulic conductivity. Tensiometer data at seven locations indicated nearly constant conditions for 30 mo, while nine of the other 10 sites showed small gradual trends. Assumption of a unit hydraulic gradient led to flux estimates ranging from 0.2 to 10000 cm yr–1. Estimates in the 34-m interbed ranged across four orders of magnitude while flux estimates for the 73-m interbed ranged three orders of magnitude. While the tensiometer data appear to reflect in situ conditions and are a sensitive indicator of hydrologic conditions in the deep vadose zone, the laboratory-developed hydraulic properties introduce a high degree of uncertainty, potentially affecting predictions by orders of magnitude. There is a need to develop techniques for assessing flux rates for the range of applicable field conditions to improve the confidence in deep flux estimates.
Abbreviations: bls, below land surface • ESRP, Eastern Snake River Plain • INEEL, Idaho National Engineering and Environmental Laboratory • RWMC, Radioactive Waste Management Complex • SDA, Subsurface Disposal Area
|
INTRODUCTION
|
|---|
FORECASTING the transport of waterborne contaminants through the vadose zone requires estimates for liquid flux between the land surface and the water table. While there is an extensive body of literature regarding the estimation of flux at shallow depths, the deeper vadose zone has received much less attention. Instruments used for direct measurement of flux in the near-surface environment (0–2 m) include pan lysimeters (e.g., Jordan, 1968), tension lysimeters (Byre et al., 1999), and vadose zone flux meters (Wagenet, 1986; Gee et al., 2002). There are also Darcian approaches in which shallow measurements of moisture content or water potential (
) are combined with laboratory developed relations for the unsaturated hydraulic conductivity to estimate flux (Stephens and Knowlton, 1986). However, the utility of all such measurements is limited by the inherent spatial and temporal variability of flux in the shallow vadose zone (Wagenet, 1986).
At sites with thick vadose zones, flux estimates obtained at depth may be more representative of mass transfer to the water table than those obtained from the near-surface environment. Also, many of the complicating effects found in the near-surface environment will be damped or eliminated with increasing depth. For these reasons, flux measurement at depth would appear to be an attractive alternative at such sites. However, borehole instruments for direct measurement of deep flux do not exist at this time. Environmental tracers (e.g., Scanlon et al., 1997; Phillips, 2001) may be used to provide information on average historical flux at some sites, but this approach is not conducive to monitoring activities. Conversely, Darcian approaches will have a more widespread applicability and are amenable to monitoring.
Darcian approaches are founded on an assumption of one-dimensional vertical flow. One then needs sufficient information on either the in situ moisture content or water potential to calculate flux from laboratory-derived unsaturated hydraulic conductivity. Previous applications of this approach have used tensiometers (Stephens and Knowlton, 1986), thermocouple psychrometers (Andraski, 1997), or heat dissipation sensors (Montazer et al., 1986) to measure water potential gradients along boreholes at depths of >2, 5, and 200 m, respectively. Tensiometric data have a distinct advantage in that it is a direct measure of water potential, whereas the other two methods are calibration-dependent, indirect measures. Available techniques for in situ measurement of moisture content are not only calibration dependent, but also physically difficult to install at depths beyond a few meters.
Conventional tensiometers require a continuous water column that extends from the measurement point to the sensor location at or near the land surface. The vaporization of water in the water column limits the depth of emplacement to about 8 m and has precluded the use of tensiometer data for Darcian estimates of flux below that depth. This problem was recently overcome by development of the advanced tensiometer (Hubbell and Sisson, 1998), which has been successfully deployed to make direct measurements of water potential at depths up to 145 m. The advanced tensiometer has two parts, a permanently installed porous cup assembly with casing that extends to land surface and a removable electronic pressure transducer assembly for installation from land surface. Positioning the sensor close to the measurement point (porous cup) eliminates the need for a water column extending to land surface, thus removing the restriction on depth of emplacement.
Here, we present a first attempt to estimate flux at depth using long-term monitoring data obtained from the deployment of advanced tensiometers. Instruments were installed within two laterally extensive sedimentary interbeds at depths of 34 and 73 m for the purpose of monitoring hydrologic activity within a complex sequence of unsaturated basalt flows that underlie a waste disposal facility in south-central Idaho. At most of the 17 measurement sites, data collected during a 30-mo period (McElroy and Hubbell, 2003b) indicated near steady-state conditions. From Richards' equation for vertical flow, steady-state flow occurs for a unit hydraulic gradient, where flux is given by the in situ unsaturated hydraulic conductivity. Using laboratory-measured moisture retention data and saturated hydraulic conductivity on samples from the sedimentary interbeds, we model the unsaturated hydraulic conductivity as a function of water potential (van Genuchten, 1980), and then estimate flux by applying that relationship to measured water potentials.
We begin this paper by briefly describing the physiography, stratigraphy, and hydrology of the study site. Well locations and instrument installation are presented next. Data analysis begins with estimation of the unsaturated hydraulic properties from core samples. Then we use tensiometric data from a 30-mo period to argue for applicability of a Darcian approach at this site. Resulting estimates of flux vary across a range from 0.2 to 10000 cm yr–1, with the latter value exceeding precipitation at this site by more that two orders of magnitude. We then discuss possible sources of error in our estimates, along with ideas for improving this approach. Finally, we conclude with a short summary of key observations.
|
MATERIALS AND METHODS
|
|---|
Site Overview
Data were collected at the Subsurface Disposal Area (SDA) of the Idaho National Engineering and Environmental Laboratory (INEEL). The SDA is situated in a local topographic depression on the Eastern Snake River Plain (ESRP) of south-central Idaho (Fig. 1)
and is part of the Radioactive Waste Management Complex (RWMC) at the INEEL. Trenches, pits, and soil vaults sited within thin surficial sediments at the SDA have been used to dispose of approximately 2 x 105 m3 of low-level, mixed, and transuranic radioactive waste (DOE, 1998). Beneath the shallow sediments, a thick sequence of fractured volcanic rock hosts the ESRP aquifer, which has been designated as a sole source aquifer by the Environmental Protection Agency (Federal Register, 1991).
View larger version (29K):
[in this window]
[in a new window]
|
Fig. 1. Location map of the Eastern Snake River Plain (ESRP), Idaho National Engineering and Environmental Laboratory (INEEL), and Subsurface Disposal Area (SDA).
|
|
Stratigraphy at the SDA (Fig. 2)
consists of thin (0–8 m thick) fluvial and aeolian sediments resting on a thick sequence (>0.8 km) of basalt flows (Hackett et al., 1986). The basalt profile is intercalated with thin (
0–10 m thick) sedimentary interbeds deposited during quiescent periods of the basalt emplacement. Of particular interest to this study are interbeds located at depths of about 34 and 73 m bls. Following local convention, these are referred to as the 34-m and 73-m interbeds. Based on limited subsurface data, the 34-m interbed appears to cover about 90% of the study area, while the 73-m interbed is believed to be contiguous. Both have textures similar to the surficial sediments. The continuity of these two interbeds and their position between the SDA and the water table at about 180 m bls make them natural targets for monitoring activities. Unless flow is extensively diverted within the basalt layers, recharge to the ESRP aquifer must either pass through or perch on top of these interbeds. Given the relative homogeneity of the sedimentary interbeds, we expect point estimates of flux to be more representative than in the basalt layers, where flow is likely to be episodic and highly localized.
View larger version (40K):
[in this window]
[in a new window]
|
Fig. 2. East–west cross section across the study area. Surficial sediment and interbeds comprise 5% of the total thickness of the material to the water table at 180 m. Wells 01, 02, 03, and 07 are projected north–south onto the cross section. (See Fig. 3 for well locations.)
|
|
View larger version (33K):
[in this window]
[in a new window]
|
Fig. 3. Map of well locations in and around the SDA, with measured water potential shown as inserts. For presentation, the 30-mo-long data streams were subsampled at 12-h intervals. Well names indicate the tensiometer depth in meters (i.e., Well I1-69 has a tensiometer located at 69 m bls).
|
|
In addition to the subsurface complexity at this site, the upper boundary condition, surface infiltration, is expected to vary in space and time. The study area receives about 22 cm yr–1 of precipitation (Clawson et al., 1989), with about one-half of that amount occurring as snow from November to February. Infiltration is believed to be concentrated in drainage channels and topographic lows that receive overland flow of snowmelt (Barraclough et al., 1976). Runoff in the SDA is routed primarily into an east–west drainage channel down the center of the site (Fig. 3)
, while flow from the surrounding area is diverted around the SDA and then eastward. Other local sources of recharge include the Big Lost River channel (2 km to the north) and spreading areas (1.3 km to the west and 1.1 km south) used to infiltrate floodwaters from the Big Lost River (Fig. 1). Given the spatial and temporal variability of infiltration, plus the potential for lateral diversion and focused flow within the intervening basalt layers, we expected a range of flux and moisture contents within the sedimentary interbeds.
Well Locations, Installation, and Data Collection
Of the 19 vadose zone wells that have been installed in or around the RWMC (see McElroy and Hubbell, 2003a, 2003b), 16 were completed near the tops of the 34-m or 73-m interbeds (one was completed in both) and are thus relevant to this study. Ten of those wells are inside the SDA, and the remaining six are within 0.5 km of the perimeter (Fig. 3). One well (76-5) was installed in 1995 (details in Hubbell et al., 2002), while the remainder (I and O prefixes) were installed during 1999 and 2000 (details in Dooley and Higgs, 2003). All wells were drilled by air-rotary, using either a downhole hammer bit (25.1-cm diam.) or a wireline core barrel (9.63- or 12.26-cm diam.). Where possible, 6.35- or 8.5-cm-diameter core samples were taken from the sedimentary interbeds using a wireline technique (Christensen Geobarrel, Salt Lake City, UT). All of the wells completed in the 73-m interbed were reamed to 25.1 cm before installing the instruments. In the 34-m interbed, the O-series wells and I5-30 were reamed to 25.1 cm.
In each well, an advanced tensiometer (Hubbell and Sisson, 1998) was installed in an approximately 1.5-m-thick backfill layer spanning the contact between interbed materials and the overlying basalt. To assure hydraulic connection with the interbed materials, backfill consisted of either a silica-flour slurry (I and O prefix wells) or dry native loam material (Well 76-5). The slurry backfill was chosen to speed equilibration (e.g., Sisson et al., 2000). Granular bentonite layers were placed above and below the backfill to isolate the monitoring intervals and seal the remainder of the borehole. Tensiometric data were collected at 1- or 4-h intervals using ±800 cm water pressure (differential) transducers with an accuracy of ±3 cm (Electronic Engineering Innovations, Las Cruces, NM) connected to a data logger (CR23, 21X, or 510X, Campbell Scientific, Inc., Logan, UT). Suction lysimeters installed approximately 0.3 m above the tensiometers were sampled at irregular intervals (I and O prefix wells only).
Estimation of Hydraulic Properties
Arbitrarily selected segments from the cored interbed sediments were used to estimate hydraulic properties at the tensiometer locations. Transparent sampling tubes allowed on-site examination of the core to identify segments exhibiting both minimal disturbance and materials representative of those found at the tensiometer location (i.e., top of the interbed). The selected segments were then cut from the sample tube, sealed, and transported for analysis. Bulk density, effective porosity, and saturated hydraulic conductivity (Ks) were measured on the intact core (Klute, 1986a). Portions of each sample were disaggregated, and the <2-mm fraction classified on the basis of texture (Table 1, Fig. 4)
. The <2-mm fraction of the disaggregated sample was subdivided and repacked for measurement of moisture retention at arbitrarily chosen pressures of approximately –200, –400, –600, –800, –1000, and –10000 cm of water (ASTM, 1994; Klute, 1986b).
View this table:
[in this window]
[in a new window]
|
Table 1. Textural classification and hydraulic properties of sediment from the 34-m and 73-m interbeds. The calculated properties are from sediment samples collected from the interbed and well, except where specifically noted. Wells identified with an O are located outside the SDA; others are located inside. Gravel refers to the sample fraction (by mass) exceeding 2 mm.
|
|
View larger version (27K):
[in this window]
[in a new window]
|
Fig. 4. Soil textural triangle. Grain size analyses of the <2-mm fraction are used to place sediment samples into textural categories ranging from silty clay loam to loamy sand.
|
|
Laboratory measurements and soil textural classifications were used to estimate the unsaturated hydraulic conductivity, K(), of each sample as a function of water potential () following the model presented by van Genuchten (1980):
 | [1] |
where adjustable fitting parameters 
, m, l, and n are dependent on pore structure (pore size distribution and connectivity) of the media. Further, we assumed that and m are related to moisture retention as described by Mualem (1976); n was approximated as m = 1 – 1/n, and the value of l was set to 0.5 (e.g., van Genuchten et al., 1991). The aforementioned assumptions were implemented numerically using the model RETC (van Genuchten et al., 1991), with the further constraint that the estimated parameter 
s (saturated water content) must be consistent with the observed textural classifications as defined in the RETC model. After initial trials, water retention data at –10000 cm was noted to place an undue control on model outcome. As all of the in situ tension measurements fell between –21 and –388 cm, we discarded the –10000 cm data and used water retention measurements from –200 to –1000 cm to estimate the van Genuchten parameters (Table 1). At three locations within the SDA, drilling operations failed to obtain representative samples from the 34-m interbed. For those locations we employed van Genuchten parameters previously estimated for the 34-m interbed by Magnuson and McElroy (1993).
|
RESULTS
|
|---|
Tensiometric Data and Hydraulic Gradients
Water potential measurements obtained from advanced tensiometers targeting the 34-m and 73-m interbeds are shown in Fig. 3 for February 2000 to August 2002 (further details in McElroy and Hubbell, 2003a, 2003b). Measurements ranged between –21 and –388 cm of water (Fig. 3, Table 2), with most locations showing little variability. Note that some of the apparent variability seen in Fig. 3 may be attributed to measurement artifacts. The numerous downward spikes in the data coincide with sampling from nearby suction lysimeters, while the exponential declines in water potential at early time (Wells 05-32 and I5-30) are believed to result from equilibration of the backfill to ambient conditions. Data from the other wells were collected after equilibration. A final artifact occurs when a tensiometer slowly loses water and must be refilled. For this condition, the data show a gentle increase in water potential followed by a sharp decline (Wells O2-33 and I3-70). All of the tensiometers are refilled on an irregular basis.
View this table:
[in this window]
[in a new window]
|
Table 2. Range, mean, trend, and slope of trend for water potential from tensiometers, February 2000 to September 2002. Wells identified with an O are located outside the SDA; others are located inside.
|
|
Ignoring the measurement artifacts, and with minor exception, water potential data taken from the 34-m and 73-m interbeds were either stable or gradually trending toward a stable condition. Water potentials in the 34-m interbed were near steady at five locations; three locations suggested gradual drying, and three wetting (Table 2). Two instruments monitoring the 73-m interbed yielded steady readings, while three showed gradual drying, and one wetting. On the basis of this apparent long-term stability, we concluded that it was appropriate to assume steady-state vertical flow across the interbeds and employ a Darcian approach to estimate flux. From Richards' equation for vertical flow, true steady-state behavior can only occur when water potential gradients are negligible, and flow is driven solely by the gravitational component of the hydraulic gradient, which is unity. Long-range horizontal gradients within the interbeds will be much smaller than gravity and are ignored. We also assume that the short-range horizontal gradients within the interbeds are negligible, although they may be of local importance.
In addition to the apparent steady-state behavior at most of our measurement sites, supporting evidence for a near unit vertical gradient within the interbeds comes from evaluation of tensiometric data at this site as a whole, and from measurements at a nearby analog site. McElroy and Hubbell (2003b) evaluated water potential measurements from 30 locations around the SDA, including the 17 considered here and additional sites within the basalt. They concluded that the correlation between measured water potential and elevation of the measurement suggested a unit hydraulic gradient in the upper 73 m of the unsaturated zone. More detailed vertical measurements of water potential were collected at an analog site 11 km northwest of the SDA in the context of a transient infiltration experiment. At the analog site, five wells, each with multilevel tensiometers at depths of 20 to 40 m, penetrate a stratigraphy similar to that at the SDA. Before the ponded infiltration experiment, measured gradients across the interbeds ranged between 0.94 and 1.04, providing additional support for our assumption of a unit gradient under ambient conditions.
Estimated Flux
Given that most of our tensiometric data indicated gravity-dominated flow under near steady-state conditions, we employed the Darcian approach to estimate unsaturated flux. In the absence of water potential gradients, flow through homogeneous unsaturated media is driven solely by gravity, and the Richards equation reduces to
 | [2] |
where j represents the downwards flux, and K() is the unsaturated hydraulic conductivity at a given water content (e.g., Charbeneau, 2000). Therefore, for unit gradient conditions, knowledge of and K() is sufficient to determine flux. Lacking data on and K(), we combine tensiometric data with the K() relations from Table 1 to estimate flux at the minimum, maximum, and mean water potentials recorded during the 30-mo data collection period (Fig. 5)
.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 5. Flux estimates from the (a) 34-m and (b) 73-m interbeds. The horizontal bars represent the range of water potentials measured at a given site, with the solid dot placed at the mean. The vertical bars represent the range of K() estimated from those values. The dashed lines represent the generic curves developed by Magnuson and McElroy (1993).
|
|
Although water potential values varied by about an order of magnitude across the site, flux estimates ranged across five orders of magnitude (Fig. 5). With respect to the 34-m interbed, flux estimates for the 73-m interbed showed a smaller range, and were generally less sensitive to small changes in . Both observations are consistent with the generally finer grained sediments (Fig. 4) and lower water potentials (Table 2) in the 73-m interbed. For the 34-m interbed, flux estimates were significantly higher within the SDA than outside (Table 3). Conversely, flux estimates in the 73-m interbed were higher outside of the SDA than within; however, this is based on a limited sample size.
View this table:
[in this window]
[in a new window]
|
Table 3. Range of water potentials with calculated flux over period from February 2000 to September 2002. Wells identified with an O are located outside the SDA, others are located inside.
|
|
Closely spaced pairs of wells along the centerline of the SDA (Fig. 2 and 3) provide the opportunity to compare vertical flux estimates between the two interbeds. For the westernmost pair of wells (I1-31 and I1-69), estimated flux decreased by a factor of about 2 from the shallower interbed to the deeper one (Table 1). The difference became almost an order of magnitude for the next pair (I2-29 and I2-68). Continuing eastwards (I3–28 and I3-70), the situation reversed, as estimated flux in the 73-m interbed was about 2.5 orders of magnitude larger than in the 34-m interbed, and reversed again for the most eastern pair (I4-30 and I4-69), where estimated flux in the 34-m interbed was about five orders of magnitude larger than in the 73-m interbed. Given that these pairs of wells are associated with a drainage channel that flows from west to east, we could be observing the results of locally focused flow. However, the degree of variability and magnitude in estimated flux seems extreme, especially considering the apparent lack of lateral migration (perched water) along the tops of the interbeds. We do note that data from other wells indicated localized perched water in dense basalts above the interbeds (Fig. 2).
At three locations within the 34-m interbed (I5-30, 765-31, O4-33) fluxes were estimated using generic properties developed by Magnuson and McElroy (1993) for the 34-m interbed. Flux predicted using the generic properties fell into the middle of the data set, but was much more sensitive to than most of the other data points. In general, the K() relations developed for this investigation led to flux estimates (Fig. 5) substantially different than if the generic relations (34-m and 73-m interbeds) were utilized. Most of our data points plotted above the generic curves, but the two wettest measurements plotted substantially below the generic curves (Fig. 5a). For these two wet points, use of the generic properties would lead to a three- to four-order of magnitude increase in estimated flux. At the dry end of our data, use of the generic curves would lead to a decline in estimated flux of about two orders of magnitude.
The sensitivity of the K() relationship is illustrated by comparing temporal changes in estimated flux at Wells O4-33, 04-69, and I1-69 (Fig. 6)
. The water potential at O4-33 varied over about 130 cm, indicative of an apparent wetting event in September 2000, followed by drainage (Fig. 6a). The estimated flux rate varied by nearly three orders of magnitude, with about 70% of the flow during the initial 5 mo. The Well O4-69 tensiometric reading steadily increased from –360 to about –300 cm, with flux increasing by a factor of two. Figure 6b shows data from I1-69, where the water potential and calculated flux suggest several wetting events. The water potential varied 100 cm from –375 to –275 cm with the calculated flux rate change of a factor of two. Recognizing that our assumption of a unit hydraulic gradient is least tenable for a dynamic , the apparent sensitivity of estimated flux to small changes in suggests that long-term water potential data are a sensitive indicator of hydrologic changes in the deep vadose zone.
View larger version (29K):
[in this window]
[in a new window]
|
Fig. 6. Temporal changes in water potential and estimated flux for Wells (a) O4-33, O-69, and (b) I1-69.
|
|
|
DISCUSSION
|
|---|
Application of the Darcian approach to our data leads to vertical flux estimates that range from approximately 0.2 to 10000 cm yr–1. The latter value is about 500 times the local precipitation (22 cm yr–1), which raises concerns about the accuracy of our estimates. However, we cannot arbitrarily dismiss flux estimates that exceed precipitation. The data points were preferentially located along drainage features that periodically receive significant quantities of snowmelt; thus, recharge may be focused, even at depth. Also, it has been suggested (Hubbell, 1990; Nimmo et al., 2002) that water infiltrated at spreading areas to the south and west (Fig. 1) may migrate laterally toward the SDA in perched zones within the basalt sequence that separates the 34-m and 73-m interbeds. However, there were no releases to the spreading areas during our monitoring period that would provide data relevant to that supposition. Since the conceptual model of the site cannot be used to eliminate the possibility of locally enhanced recharge at depth, the potential sources of uncertainty within our flux estimates must be examined.
To consider the uncertainty associated with our flux estimates, we look at the validity of our primary assumptions, which are accurate measurements of water potential, steady one-dimensional flow, and a well understood relationship for K(). Provided that the backfill is at equilibrium with the surrounding sediments and neglecting barometric induced fluctuations, the tensiometric data should be accurate to within instrument error (±3 cm). Effects of this uncertainty on predicted flux will vary with the K() relationship, but will be much less than an order of magnitude, except possibly under much wetter conditions than those encountered here. Likewise, the small temporal changes in suggest that uncertainty associated with the assumption of a unit hydraulic gradient will also be much less than an order of magnitude. Conversely, there may be substantial uncertainty in the K() relationship.
To estimate K(), we measured Ks, effective porosity, and moisture retention (–200, –400, –600, –800, and –1000 cm of water). Even if our samples were truly undisturbed (highly unlikely) and representative (somewhat likely), we must still contend with the fact that we did not directly measure K() under in situ conditions, nor did we consider the effects of hysteresis. Although our only physical measurement of K() was at saturation, we extrapolated to estimate values of K() that are one to six orders of magnitude less than Ks. Moreover, that extrapolation was based on water retention measurements taken on a disturbed sample. The range of moisture retention measurements do not provide substantial information in the wet range, where the K() relation will be particularly sensitive to small changes in . As an example of the potential for error in estimating K(), we look at Well I4-30, which produced our highest flux estimate (6690–10000 cm yr–1). The sample taken from this location led to an estimate for the van Genuchten of 0.0041 cm–1 (Table 1), which is substantially lower than those of other samples from the 34-m interbed, particularly given the textural description of loamy gravel. If one were to raise that value of by an order of magnitude to more closely reflect the other samples, the estimated flux would drop by two orders of magnitude.
The sensitivity of our flux estimates to the K() relation suggests several areas where estimates could be improved, and areas for further research. It would be best to avoid the need to estimate flux from water potential measurements, and instead directly measure flux in situ. However, current technology does not allow for direct measurement of flux at depth. The next best option would be direct measurement of K() on undisturbed samples and across the range of expected field water potentials for both wetting and drying conditions. Recent advances, such as the geocentrifuge or the laboratory technique presented by Butters and Duchateau (2002) may make such data a possibility in the near term. There are also ways to reduce the uncertainty associated with the conventional approaches used in this investigation. Measurement of water retention on undisturbed samples and/or for a narrower range of water potentials that closely reflects expected field conditions would improve our estimate. Finally, where possible, one might also obtain one or more laboratory measurements of K() at values of close to that expected in the field. While doing so is much more difficult than measuring moisture retention, even a single data point would help to constrain the estimate for K().
|
CONCLUSIONS
|
|---|
A Darcian approach was used to estimate flux in a thick vadose zone that underlies a waste disposal facility in south-central Idaho. Given a complex geology of intercalated basalt flows, two laterally extensive sedimentary interbeds were targeted for monitoring. Wells were drilled to collect sediment core and install advanced tensiometers at the 34-m and 73-m interbeds. Laboratory-measured hydraulic properties on the core samples were used to estimate the unsaturated hydraulic conductivity as a function of water potential. Tensiometric data collected during a 30-mo period were considerably more stable than expected for the near surface environment, suggesting that a Darcian approach would be appropriate. Seven of the 17 sites showed nearly steady readings for the 30-mo time period. Four sites showed an increasing water potential trend, while six sites showed a downward trend; with one exception, all of the trends were gradual and smooth. Assuming steady flow (unit gradient), we then estimated flux on the basis of in situ water potential measurements. Flux estimates ranged from 0.2 to about 10000 cm yr–1 at a site with 22 cm yr–1 precipitation.
Although preferential recharge and lateral migration of perched water could act to locally enhance flux at this site, the high degree of uncertainty associated with mapping water potential to the unsaturated hydraulic conductivity limits our ability to reach such conclusions from this data. Small changes in the fitting parameters can induce order-of-magnitude changes in the flux estimates, particularly at near-zero water potentials. This sensitivity suggests a need for improved techniques to assess flux over the range of applicable field conditions. Despite the uncertainty in the K() relationship, long-term water potential data appear to be a sensitive indicator of hydrologic changes in the deep vadose zone, and thus are important for understanding the temporal dynamics of flow.
|
ACKNOWLEDGMENTS
|
|---|
Information contained in this article was developed during the course of work under contract number DE-AC07-94D13223 with the Department of Energy and the Idaho National Engineering and Environmental Laboratory. M.J. Nicholl acknowledges support from the U.S. Department of Energy through the Basic Energy Sciences Geoscience Research Program under contract number DE-FG03-01ER15122 and the Environmental Management Science Program under contract DE-FG07-02ER63499. Mention of a trademark or a propriety product is for the benefit of the reader and does not constitute an endorsement for the product by the Department of Energy to the exclusion of other products that may also be suitable. We would also like to thank the two anonymous reviewers for their constructive comments that greatly improved this manuscript.
|
REFERENCES
|
|---|
- Andraski, B.J. 1997. Soil-water movement under natural-site and waste-site conditions: A multiple-year field study in the Mojave Desert, Nevada. Water Resour. Res. 33:1901–1916.
- American Society for Testing and Materials. 1994. Standard method for capillary-moisture relationships for course and medium textured soils by porous-plate apparatus. D2325-68 (94). In Soil and rock. V. 04–08. ASTM, West Conshohocken. PA.
- Barraclough, J.T., J.B. Robertson, V.J. Janzer, and L.G. Saindon. 1976. Hydrology of the solid waste burial ground, as related to the potential migration of radionuclides, Idaho National Engineering Laboratory. Open-File Rep. 76-471. USGS, Reston, VA.
- Butters, G.L., and P. Duchateau. 2002. Continuous flow method for rapid measurement of soil hydraulic properties: Experimental considerations. Available at www.vadosezonejournal.org. Vadose Zone J. 1:239–251.[Abstract/Free Full Text]
- Byre, K.R., J.M. Norman, L.G. Bundy, and S.T. Gower. 1999. Equilibrium tension lysimeter for measuring drainage through soil. Soil Sci. Soc. Am. J. 63:536–543.[Abstract/Free Full Text]
- Charbeneau, R.J. 2000. Groundwater hydraulics and pollutant transport. Prentice Hall, Englewood Cliffs, NJ.
- Clawson, K.L., G.E. Start, and N.R. Ricks. 1989. Climatology of the Idaho National Engineering Laboratory. 2nd ed. DOE/ID-12118. USDOE Idaho Operations, Idaho Falls, ID.
- Dooley, K.J., and B.D. Higgs. 2003. End of well reports for the OU 7-13/14 fiscal year 2000 well drilling project at the Radioactive Waste Management Complex. INEEL/EXT-2000-00400. Bechtel BWXT Idaho, LLC, Idaho Falls, ID.
- DOE. 1998. Radioactive waste information for 1997 and record-to-date. DOE/ID-10054(97). Idaho National Engineering and Environmental Laboratory. DOE. Bechtel BWXT, Idaho Falls, ID.
- Federal Register. 1991. Sole source designation of the Eastern Snake River Plain Aquifer, Southern Idaho. Vol. 56. Number 194. p. 50634.
- Gee, G.W., A.L. Ward, T.G. Caldwell, and J.C. Ritter. 2002. A vadose zone water fluxmeter with divergence control. Water Resour. Res. 38(8). DOI: 10.1029/2001WR000816.
- Hackett, B., J. Pelton, and C. Brockway. 1986. Geohydrologic story of the Eastern Snake River Plain and the Idaho National Engineering Laboratory. USDOE BP-455-1186-2.5M-A Idaho Operations Office, INEEL, Idaho Falls, ID.
- Hubbell, J. M. 1990. Perched groundwater at the Radioactive Waste Management Complex of the Idaho National Engineering Laboratory. EGG-ER-8779. EG&G Idaho, Idaho Falls, ID.
- Hubbell, J.M., E.D. Mattson, J.B. Sisson, and D.L. McElroy. 2002. Water potential in fractured basalt from infiltration events. p. 38–56. In M.N. Sarah and L.G. Everett (ed.) Evaluation and remediation for low permeability and dual porosity environments. ASTM Spec. Tech. Publ. 1415. ASTM, West Conshohocken. PA.
- Hubbell, J.M., and J.B. Sisson. 1998. Advanced tensiometer for shallow or deep soil-water pressure measurements. Soil Sci. 163:271–277.
- Jordan, C. 1968. A simple, tension-free lysimeter. Soil Sci. 105:81–86.
- Klute, A. (ed.) 1986a. Methods of soil analysis. Part 1. 2nd ed. Agron. Monogr. 9. ASA and SSSA, Madison, WI.
- Klute, A. 1986b. Water retention: Laboratory methods. p. 635–660. In A. Klute (ed.) Methods of soil analysis. Part 1. 2nd ed. Agron. Monogr. 9. ASA and SSSA, Madison, WI.
- Magnuson, S.O., and D.L. McElroy. 1993. Estimation of infiltration from in situ moisture contents and representative moisture characteristic curves for the 30', 110' and 240' interbeds. Engineering Design File RWM-93-001.1. EG&G Idaho, Idaho Falls, ID.
- McElroy, D.L., and J.M. Hubbell. 2003a. Advanced tensiometer monitoring results from the deep vadose zone at the Radioactive Waste Management Complex. 2003. INEEL/EXT-02-01276. Bechtel, BWXT Idaho, LLC, Idaho Falls, ID.
- McElroy, D.L., and J.M. Hubbell. 2003b. Evaluation of the conceptual flow model for a deep vadose zone system using advanced tensiometers. Vadose Zone J. 3:170–182.
- Montazer, P., E.P. Weeks, F. Thamir, S.N. Yard, and P.B. Hofrichter. 1986. Monitoring the vadose zone in fractured tuff, Yucca Mountain, Nevada. p. 439–469. In Proceedings of the NWWA Conference on Characterization and Monitoring of the Vadose Zone. 19–21 Nov. 1985. National Water Well Assoc., Westerville, OH.
- Mualem, Y. 1976. A new model for predicting the hydraulic conductivity of unsaturated porous media. Water Resour. Res. 12:593–622.
- Nimmo, J.R., K.S. Perkins, P.E. Rose, J.P. Rousseau, B.R. Orr, B.V. Twining, and S.R. Anderson. 2002. Kilometer-scale rapid transport of naphthalene sulfonate tracer in the unsaturated zone at the Idaho National Engineering and Environmental Laboratory. Available at www.vadosezonejournal.org. Vadose Zone J. 1:89–101.[Abstract/Free Full Text]
- Phillips, F.M. 2001. Investigating flow and transport in the fractured vadose zone using environmental tracers. p. 271–294. In Conceptual models of flow and transport in the fractured vadose zone. National Academy Press, Washington, DC.
- 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., A.L. Schafer, and J.M. Hubbell. 2000. Vadose zone monitoring system for site characterization and transport modeling. p. 161–166. In R.W. Smith and D.W. Shoesmith (ed.) Scientific Basis for Nuclear Waste Management XXIII. Symposium Proceedings. Volume 608. Material Research Society, Warrendale. PA.
- Stephens, D.B., and R. Knowlton. 1986. Soil water movement and recharge through sand at a semiarid site in New Mexico. Water Resour. Res. 22:881–889.
- van Genuchten, M.Th. 1980. A closed-form equation for predicting the hydraulic conductivity of unsaturated soils. Soil Sci. Soc. Am. J. 44:892–898.[Abstract/Free Full Text]
- van Genuchten, M.Th., F.J. Leij, and S.R. Yates. 1991. The RETC code of quantifying the hydraulic functions of unsaturated soils. EPA/600/2-91/065. Robert S. Kerr Environmental Research Laboratory, Ada, OK.
- Wagenet, R.J. 1986. Water and soil flux. p. 1055–1088. In A. Klute (ed.) Methods of soil analysis. Part 1. 2nd ed. Agron. Monogr. 9. ASA and SSSA, Madison, WI.
This article has been cited by other articles:
|
|
|
|
 
R. M. Holt and M. J. Nicholl
Uncertainty in Vadose Zone Flow and Transport Prediction
Vadose Zone J.,
May 1, 2004;
3(2):
480 - 484.
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
