VZJ sign up for etocs
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (5)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by McElroy, D. L.
Right arrow Articles by Hubbell, J. M.
Right arrow Search for Related Content
PubMed
Right arrow Articles by McElroy, D. L.
Right arrow Articles by Hubbell, J. M.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by McElroy, D. L.
Right arrow Articles by Hubbell, J. M.
Related Collections
Right arrow Flow
Right arrow Field-Scale Studies
Right arrow Recharge
Published in Vadose Zone Journal 3:170-182 (2004)
© 2004 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA

SPECIAL SECTION: UNDERSTANDING SUBSURFACE FLOW AND TRANSPORT PROCESSES AT THE IDAHO NATIONAL ENGINEERING & ENVIRONMENTAL LABORATORY (INEEL) SITE

Evaluation of the Conceptual Flow Model for a Deep Vadose Zone System Using Advanced Tensiometers

Deborah L. McElroy*,a and Joel M. Hubbellb

a Applied Geosciences Department, Idaho National Engineering and Environmental Laboratory, Bechtel BWXT Idaho, LLC., Idaho Falls, ID 83415
b Geosciences Research Department, Idaho National Engineering and Environmental Laboratory, Bechtel BWXT Idaho, LLC., Idaho Falls, ID 83415
* Corresponding author (dlm2{at}inel.gov).

Received 9 February 2003.



   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 SITE DESCRIPTION
 CONCEPTUAL FLOW MODEL
 METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Conceptual flow models provide a framework for predictive modeling of contaminant transport. This study tests the assumptions of steady-state flow and a unit hydraulic gradient in a 177-m-thick vadose zone beneath a mixed waste site, using a network of advanced tensiometers. The conceptual flow model at the waste site, located on the Idaho National Engineering Laboratory (INEEL), describes moisture movement through a geologically complex site comprising basalt flows intercalated with sedimentary interbeds. The presence of sedimentary interbeds is expected to dampen and store much of the episodic recharge, resulting in near steady-state conditions and unit gradient flow. Thirty advanced tensiometers in 18 wells provided field water potential data at depths ranging from 6.7 to 73.5 m below land surface (bls), beneath and adjacent to the waste site. Measured water potentials from February 2000 through August 2002 ranged from near saturation (–30 cm of water) to about –400 cm of water. Above 17 m, the observed long-term drying trends were presumed to be a response to the cumulative effect of lower than average annual precipitation for the last 3 yr (2000–2002). Below 17 m, steady-state conditions were observed at more than one-half of the monitored locations. However, long-term drying and wetting trends were also observed at 9 of the 25 monitored locations below 17 m, in contrast to the steady-state flow assumptions in the conceptual model. Long-term water potential changes ranged from about 20 to 200 cm of water. It is hypothesized that these drying trends are related to areas of focused infiltration, such as drainage ditches, and are a response to decreased runoff from three years of less than average precipitation. A unit gradient was indicated by aligning dispersed monitoring locations along a presumed vertical profile.

Abbreviations: bls, below land surface • INEEL, Idaho National Engineering and Environmental Laboratory • RWMC, Radioactive Waste Management Complex • SDA, Subsurface Disposal Area


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
SITE DESCRIPTION
 CONCEPTUAL FLOW MODEL
 METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
CONCEPTUAL FLOW MODELS in an unsaturated zone describe the hydrologic processes of a vadose zone system from infiltration to deep percolation and provide the framework needed to model and predict contaminant transport. The addition of site-specific data with time may augment or change conceptual flow models, and conceptual models will continually evolve. Site-specific characterization activities may include the introduction of tracers to track flow patterns and flow rates; analysis of water and soil for contaminants; the use of isotopes to determine source, age, or recharge rates; analyses of the media to determine hydrologic properties; and instrumentation to monitor water potentials and moisture contents. Flint et al. (2001) described the evolution of the conceptual flow model for Yucca Mountain, an arid site under consideration as a high-level waste repository, during 15 yr. The primary driver behind the evolution of the conceptual flow model was the accumulation of site-specific data from neutron logging of boreholes, bomb-pulse isotopes, perched water analyses, and thermal analyses.

Collecting site-specific temporal data in deep vadose zone systems present unique challenges because of the difficulty and expense of installing and maintaining instrumentation for extended time periods (years). However, the in situ measurement of water or matric potentials provides data needed to characterize flow processes, track infiltration or drainage, and estimate deep percolation.

Thermocouple psychrometers, heat-dissipation sensors, and tensiometers measure components of the total energy potential of water. The total energy potential of water is the sum of the contributions of the gravitational, pressure, and osmotic potentials (Hillel, 1980). Thermocouple psychrometers measure water (matric and osmotic) potential in the –2000 to –80000 cm of water range and have been used to monitor temporal variations in water potentials in deep vadose zones located in arid settings (Montazer et al., 1988; Scanlon and Goldsmith, 1997; Scanlon et al., 1999). Heat dissipation sensors measure matric potential between –100 and –14000 cm of water and have been used in deep vadose zone investigations (Montazer et al., 1988) that require measurement in a higher range than provided by psychrometers.

Tensiometers measure the sum of the matric and pneumatic (air pressure) potential (Scanlon et al., 1997). Water potential is used interchangeably with matric potential, where osmotic potential is negligible and the instrument is referenced to the gas pressure at the point of measurement. Tensiometers measure water potentials in the range of 0 to –800 cm of water (Gardner et al., 1922; Richards, 1931; Richards et al., 1937; Cassel and Klute, 1986; Hubbell and Sisson, 1998); thus, they are most useful in water potential ranges that are indicative of higher water fluxes. Conventional tensiometry is limited by depth as a result of the length of the hanging column of water needed to operate the tensiometer. The development of the advanced tensiometer minimized the hanging water column by placing a pressure transducer near the porous cup and allowed deployment of tensiometers at depths >9 m (Hubbell and Sisson, 1998). In addition to the increased depth of placement, the use of advanced tensiometers has minimized thermal noise and sluggish response problems encountered in conventional tensiometers (Sisson et al., 2002). The advanced tensiometers have been proven to be robust and reliable instruments, operating for as long as 2 yr without servicing (Sisson et al., 2002).

Much of the available temporal water potential data from deep vadose zones were collected from arid and semiarid regions of the southwestern United States. These regions tend to exhibit low water potentials and little temporal variation below depths of approximately 5 m. At a semiarid site in Texas (Scanlon and Goldsmith, 1997), temporal water potentials below 0.8 m remained fairly uniform during the monitored period and ranged from about –10000 to –100000 cm of water for a 22-m profile. In the Chihuahaun Desert of Texas (Scanlon et al., 1999), water potentials showed little temporal variation below 5 m, and water potential means ranged from –3000 to –76000 cm of water for a 30-m profile.

Temporal water potential data from thick vadose zones in the high deserts in northwestern United States are limited in number. Water potentials in the northwest appear to be higher than water potentials typical of the southwestern desert vadose zones. Sully et al. (1994) noted higher water potentials (–1000 cm of water) at Hanford, WA than at the Nevada Test Site (–6000 cm of water) in the southwestern United States. In Idaho, temporal water potentials ranged from +100 to –250 cm of water at the 2- to 30-m depths (Hubbell et al., 2002).

Temporal water potentials at greater depths and with larger spatial coverage are needed to test conceptual models of water flow in deep vadose zones at high desert sites in the northwestern United States. This investigation uses temporal water potential data collected from advanced tensiometers to evaluate assumptions used in the conceptual model of flow in the deep vadose zone beneath the Subsurface Disposal Area (SDA), a waste disposal site at the INEEL. The appropriateness of steady-state flow and unit gradient assumptions were examined in the upper 74 m of the 177-m-thick vadose zone, for an area of approximately 1 km2, within a complex geologic setting of thick fractured basalt layers intercalated with sedimentary interbeds.


   SITE DESCRIPTION
TOP
 ABSTRACT
 INTRODUCTION
 SITE DESCRIPTION
CONCEPTUAL FLOW MODEL
 METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The SDA is a 39-ha portion of the Radioactive Waste Management Complex (RWMC), located in the southwest portion of the INEEL, in southeastern Idaho (Fig. 1) . Low-level, transuranic, and mixed wastes were buried in shallow pits and trenches in the SDA from 1952 until 1970, when the burial of the transuranic portion of the waste ceased. Since 1982 only low-level waste has been buried at the SDA. Contaminants of concern that have been detected in the unsaturated zone beneath the SDA include nitrates, carbon tetrachloride, 14C, 99Tc, and U isotopes (Holdren et al., 2002).



View larger version (31K):
[in this window]
[in a new window]
  		Fig. 1. Location of the Subsurface Disposal Area (SDA) relative to the Idaho National Engineering and Environmental Laboratory, the Big Lost River, and the spreading areas.

 
The SDA is located on the Eastern Snake River Plain, an arcuate depression extending from Yellowstone Park and northwestern Wyoming westward across Idaho to the Oregon border. In the vicinity of the INEEL, the Eastern Snake River Plain has an approximate elevation of 1500 m and is bounded by mountains and high plateaus on the north, east, and south. The INEEL receives approximately 22 cm of precipitation annually, based on a 38-yr record (Clawson et al., 1989), and the majority of the November to April precipitation falls as snow. The region is classified as arid to semiarid.

The RWMC lies within a natural topographic depression. The subsurface beneath and adjacent to the SDA comprises a thin (0–7 m) cover of loess underlain by thick sequences of fractured basalt intercalated with thin sedimentary interbeds (Fig. 2) . The upper 70 m of the subsurface is composed of three primary basalt flow groups, called the A, B, and C basalts. Two laterally extensive sedimentary interbeds occur at 34 and 73 m bls, and are named after the basalt flow groups they lie between, the BC (at approximately 34 m) and the CD (at approximately 73 m) sedimentary interbeds. A less extensive, discontinuous sedimentary interbed is located at approximately 9 m bls, and is referred to as the AB interbed. The thickness of the sedimentary interbeds varies from 0 to about 10 m. The Snake River Plain Aquifer underlies the SDA at a depth of approximately 177 m.



View larger version (40K):
[in this window]
[in a new window]
  		Fig. 2. West to east trending cross section through the Subsurface Disposal Area.

 

   CONCEPTUAL FLOW MODEL
TOP
 ABSTRACT
 INTRODUCTION
 SITE DESCRIPTION
 CONCEPTUAL FLOW MODEL
METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
In our conceptual flow model of the deep vadose zone at the RWMC, long periods of constant flux result in stable moisture contents and water potentials and a unit hydraulic gradient. Under a unit gradient, there is little-to-no change in water potential with respect to depth or elevation, and moisture flow is driven by gravity rather than by differences in water potentials.

These periods of stability are occasionally interrupted by episodic recharge events. Major sources of water at the surface are direct precipitation and run-off from snowmelt, which concentrates water in topographically low areas. Rapid snowmelt, combined with heavy rain, caused flooding in pits and trenches within the SDA in 1962, 1969, and 1982 (Barraclough et al., 1976; Bargelt et al., 1992). Local run-off from late winter and early spring snowmelt has the greatest potential for extensive infiltration because these events occur when evapotranspiration rates are low. Hubbell et al. (2002) monitored a nested series of advanced tensiometers from 6.7 to 31.4 m bls, in a single borehole at the SDA. Based on 4 yr of water potential data, they found water potentials at this location were generally constant, but were interrupted by episodic recharge that was tracked to the 17-m depth, during the 1999 spring snowmelt.

In general, infiltration within the SDA is thought to be higher than outside the SDA. McElroy (1990) found higher (comparatively wetter) water potentials in surficial sediments within the SDA compared to surficial sediments outside the SDA. Cecil et al. (1992) estimated recharge rates of 0.36 to 1.1 cm yr–1 outside the SDA, at a site located near the northern boundary of the SDA. In contrast, McElroy (1993) and Bishop (1998) estimated recharge that ranged from 0.1 to 49.4 cm yr–1 during spring recharge from snowmelt inside the SDA. Laney et al. (1988), McElroy (1990), and Bishop (1998) suggested surface infiltration within the SDA is highly nonuniform and is concentrated in surface depressions such as drainage ditches or in pits or trenches that flooded in the past.

During localized recharge events, wetting fronts move through the surficial sediments and into the underlying basalts. Downward movement through the basalts is assumed to occur primarily through open or sediment-filled fractures or joints, rather than the basalt matrix (Holdren et al., 2002). Hubbell et al. (2002) monitored the advance of a wetting front in the SDA that moved through basalt at 0.1 m d–1. With a continuously ponded water source, such as during the Large-Scale Infiltration Test performed near the RWMC (Wood and Norrell, 1996), the advance of the wetting front through the basalt was on the order of 5 m d–1.

It is assumed the downward pulse of the wetting front is eventually slowed as the moisture is stored in sediment and basalts or diverted laterally by geologic media with contrasting hydraulic conductivities, such as dense unfractured basalt layers or sedimentary interbeds. Two major sedimentary interbeds are located at approximately 34 m (the BC sedimentary interbed) and 73 m (the CD sedimentary interbed). In the Large-Scale Infiltration Test (Dunnivant et al., 1998) water movement was predominantly vertical through fractured basalt to the BC interbed at a depth of 55 m, but perching and lateral movement was reported above the BC sedimentary interbed.

Additional sources of water to the subsurface beneath the SDA may be derived through lateral underflow from the Big Lost River and spreading areas (Fig. 1). The spreading areas, located west and east of the SDA, are infiltration basins that hold water diverted from the Big Lost River during times of high flow. Rightmire and Lewis (1987) and Hubbell (1990) presented evidence that suggests the spreading areas are a source for perched water located in basalt above the CD interbed. Through the use of a tracer, Nimmo et al. (2002) concluded some of the perched water above the CD interbed beneath the SDA was derived from the spreading areas. The investigators hypothesize that water from the spreading areas moves primarily downward, but a portion is diverted laterally by perching above low hydraulic conductivity layers.


   METHODS
TOP
 ABSTRACT
 INTRODUCTION
 SITE DESCRIPTION
 CONCEPTUAL FLOW MODEL
 METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The advanced tensiometers used in this study are composed of a porous cup (0.1 MPa [1 bar], standard flow, ceramic, 50–60 mL volume) installed at a specified depth with an attached polyvinyl chloride (PVC) pipe (2.54 and 3.81 cm, class 200) that extends to land surface (Fig. 3) . A volume of water is placed in the PVC pipe to fill the porous cup. A pressure transducer is placed inside the PVC pipe and seated just above the porous cup by means of a rubber stopper, sealing the water chamber in the porous cup from the water in the PVC pipe. The water in the porous cup will move into or out of the formation until the partial vacuum in the cup is equal to the subatmospheric water pressure in the surrounding soil. The pressure transducer measurement of this partial vacuum is then considered equivalent to the water potential. The pressure transducers (±800 cm of water, differential pressure relative to atmospheric; Electronic Engineering Innovations, Las Cruces, NM) were connected to Model 510X, 10X, or 23X Campbell data loggers (Campbell Scientific, Inc., Logan, UT) or a Tumut Gadara data logger (Electronic Engineering Innovations, Inc., Las Cruces, NM). This system collects continuous water potential measurements at each instrumented depth. Data were collected at least every 4 h and data loggers were generally downloaded on a monthly basis. The pressure transducers were calibrated before installation, field checked periodically, recalibrated after field placement if measurements were questionable, and replaced as needed.



View larger version (57K):
[in this window]
[in a new window]
  		Fig. 3. Schematic of advanced tensiometer, showing porous cup, transducer, and outer PVC guide pipe.

 
Figure 4 shows the locations of the wells containing the tensiometers. Tensiometer depths and adjacent lithology are listed in Table 1. In the following discussion we will use well number and the depth (rounded off), such as I1S-31 for the tensiometer at the 31.4 m depth in I-1S, to identify tensiometers. Well names beginning with "I" (for inside the SDA) and "O" (for outside the SDA) were drilled and instrumented in 1999 and 2000 as part of the hydrologic characterization activities (Dooley and Higgs, 2003) at the RWMC. The three remaining wells, 76-5, 77-2, and 78-1, were drilled and instrumented before 1999. A nested series of advanced tensiometers were installed in Well 76-5 in June 1996 (Sisson and Hubbell, 1999). Wells 77-2 and 78-1 were instrumented with advanced tensiometers in December 1995, with the addition of a portable tensiometer (Hubbell and Sisson, 1996) at the bottom of 77-2 in December 1999. Previous tensiometer data from Wells 76-5, 77-2, and 78-1 are presented in Hubbell et al. (2002) and McElroy and Hubbell (2003).



View larger version (33K):
[in this window]
[in a new window]
  		Fig. 4. Locations of wells with tensiometers. Blue highlighted squares indicate steady-state conditions and red highlighted circles indicate long-term drying or wetting trends below depths of 17 m (except nested Well 76-5, represented by tensiometers below 18 m). Transient (red) wells tend to be located along the main east–west road through the center of the SDA.

 

View this table:
[in this window]
[in a new window]
  		Table 1. Well names, instrument depths below land surface (bls), and lithology adjacent to each instrument.

 
Tensiometers in the I and O series wells were installed in a silica flour slurry that was placed around the porous cup to obtain a hydraulic connection between the cup and the geologic formation (Dooley and Higgs, 2003). The height of the silica flour slurry interval ranged from 1 to 4 m. Borehole diameters at depth of instrumentation were 9.6 cm in the I series wells (with the exception of I-5S, which was 25.1 cm) and 25.1 cm in the O series wells. Granular bentonite was used to seal the remainder of the borehole between the instrumented depths. At two of the wells (I-3D and I-4S), a silica sand interval (12.2 and 7.6 m, respectively) was placed above the silica flour interval. The BC and CD sedimentary interbeds (at approximately 34 and 73 m) were generally targeted for tensiometer placement, although some instruments were placed adjacent to basalt.

Installation methods for tensiometers in Wells 76-5, 77-2, and 78-1 differed from the methods used for the I and O series wells. At Well 76-5, the porous cups of the tensiometers were placed in a 0.3- to 1-m layer of silt loam (15.2-cm-diam. borehole) to hydraulically connect the porous cup to the fractured basalt (Hubbell et al., 2002). Granular bentonite (about 0.3 m) layers were placed above and beneath the loam-filled monitoring depths to isolate the monitoring intervals. Coarse sand (2.4–3.3 mm) filled the remaining portions of the borehole between the tensiometer monitoring depths, with thin layers of bentonite placed about every 2 m to inhibit moisture flow through the borehole. Tensiometers in Wells 77-2 and 78-1 were placed in dry native fill (loam), and bentonite layers were placed between instrumented depths to isolate each instrumented depth. The porous cups of these tensiometers were placed in a 1.8-m dry loam layer (Well 77-2, 12.4-cm-diam. borehole) and in a 0.6- to 0.7-m dry loam layer (Well 78-1, 7.62-cm-diam. borehole).


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 SITE DESCRIPTION
 CONCEPTUAL FLOW MODEL
 METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Water potentials from February 2000 through August 2002 are shown in Fig. 5 for above the 17-m depth and in Fig. 6 and 7 for below the 17-m depth. Water potentials ranged from a near-saturated –30 cm of water in the BC sedimentary interbed at I1S (Fig. 6b) to about –400 cm of water in the CD sedimentary interbed at O4 (Fig. 7b). Data gaps in water potential profiles were due to equipment malfunctions (such as battery or transducer failure) or loss of water from the tensiometer cup.



View larger version (19K):
[in this window]
[in a new window]
  		Fig. 5. Long-term drying trends at advanced tensiometer locations above depths of 17 m.

 


View larger version (19K):
[in this window]
[in a new window]
  		Fig. 6. Stable moisture contents indicating steady-state conditions in the (a) basalt at or below depth of 17 m, (b) BC sedimentary interbed, and (c) CD sedimentary interbed.

 


View larger version (22K):
[in this window]
[in a new window]
  		Fig. 7. Long-term wetting and drying trends below 17 m in the (a) BC sedimentary interbed and basalt and the (b) CD sedimentary interbed.

 
The advanced tensiometer data may be affected by equipment drift, barometric pressure fluctuations, and changes in the length of the hanging water column (<15 cm of water) from air entry into the water chamber. The type of backfill used to complete the well may also have influenced data. The use of fine-grained backfill for an interval of about 1.5 m provides a composite measurement that may dampen the instruments response. The influence of the dampened response is lessened by the small temporal changes evidenced in the ambient water potential.

Tensiometric data presented in Fig. 5 through 7 are believed to represent equilibrated ambient conditions, with the possible exception of Well 05-32. Two-month equilibration times were observed in well completions using dry backfill at Wells 76-5, 77-2, and 78-1 (Hubbell et al., 2002; McElroy and Hubbell, 2003). Sisson et al. (2000) found that wet backfill equilibrated much more rapidly (days) than dry backfill. Wet backfill was used at the I and O series wells to reduce equilibration times, and data collection did not begin for 1 to 6 mo after the boreholes were backfilled.

Temporal Trends
Water potentials at monitored locations from February 2000 through August 2002 were evaluated for temporal trends. Water potentials that do not change with time suggest steady-state conditions, implying a constant water flux. Steady state is relative to the monitored period; some long-term changes in water potentials may not be large enough to discern on the 2.5-yr time period. These long-term changes may become more evident as the monitoring continues. Transient conditions imply changes in moisture with time, presumably from changes in infiltration at land surface. Specific recharge events in short time periods, such as from snowmelt and run-off, are easily identified because of the large changes in water potentials. Small changes in water potentials with long time frames are more difficult to discern. These long-term transient trends are arbitrarily defined in this report by the presence of a long-term shift in water potentials of >15 cm of water, from February 2000 through August 2002.

Advanced tensiometers in basalts and interbed sediments above monitored depths of 17 m (772-10, 781-11, and 765-7, 765-9, and 765-12) showed decreasing water potentials for the 2.5-yr monitoring period, indicating long-term drying trends at each of these locations (Fig. 5). These decreases ranged from approximately 40 cm of water at 765-9 to 130 cm of water at 765-7. At Wells 772-10, 765-7, 765-9, and 765-12, initial drainage was in response to episodic recharge that occurred in spring of 1999 and was previously recorded at these locations (Hubbell et al., 2002; McElroy and Hubbell, 2003). However, water potentials at each of these locations, except Well 765-12, have decreased to less than the steady-state values found before 1999 (McElroy and Hubbell, 2003).

Low precipitation over the last few years may have been a factor in the decreased water potentials at these shallower (<17 m) depths. Figure 8 shows the annual precipitation (October–September) since 1991 at the Central Facilities Area at the INEEL (NOAA, 2002). Annual precipitation has decreased to <14 cm for the past 3 yr (2000, 2001, and 2002), which is less than the average annual precipitation (22 cm yr–1) (Clawson et al., 1989). The shallower (<17 m) sediments and basalts are likely responding to decreased surface infiltration.



View larger version (23K):
[in this window]
[in a new window]
  		Fig. 8. Precipitation at the nearby Central Facilities Area, approximately 8 km northeast of the Radioactive Waste Management Complex, for each water year (October–September) since 1993. The historical average annual precipitation is 22.1 cm (Clawson et al., 1989).

 
In the deeper basalts and sediments, at depths of 17 m or greater, temporal trends were mixed. The majority (15 of 25) of these deeper locations indicated steady-state conditions (Fig. 6), but long-term water potential changes were observed at nine of the 25 locations. One location, I5S-30, was indeterminate. Water potentials were stable in basalt at 772-17, 772-27, 765-24, 765-30, I2D-54, O3-27, and O7-74 (Fig. 6a). Two data spikes do occur in 772-27, in March 2001 and December 2002, but these spikes may be artificially induced. This depth (772-27) is monitored by a portable tensiometer in an open standpipe, and there is a possibility condensation formed in the borehole, resulting in a temporary rise that dissipated quickly. In the BC sedimentary interbed (Fig. 6b), water potentials were stable at 765-31, I1S-31, I2S-29, O2-33, O5-32, and O7-37. In the CD sedimentary interbed (Fig. 6c), water potentials were stable at I2D-68 and O3-67. The slow 12 cm increase in water potential values at O2-33 from May 2000 until September 2002 is indicative of air entry into the tensiometer. Refilling the water chamber with water returned the measurements to near the initial readings.

Suction lysimeter sampling caused the steep drops in water potentials at some locations (such as in I2S-29, Fig. 6b). The suction lysimeters are located within 1 m of these tensiometers, within the same silica flour backfill. Steep drops in water potential measurements correspond directly with the timing of vacuums applied to nearby suction lysimeters. Smaller upward spikes may follow these steep declines and correspond with the release of vacuum and application of positive pressures to the lysimeters. This pressure response dissipated within 2 to 4 d.

Although the majority (15 of 25) of the tensiometer locations below 17 m exhibited steady-state water potentials, long-term changes in water potentials were also observed (Fig. 7). Water potentials gradually decreased in the basalt at 765-17 (Fig. 7a), in BC interbed sediments at I3S-28, I4S-30, and O4-34 (Fig. 7a), and in the CD interbed sediments at I1D-69, I3D-70, and I4D-69 (Fig. 7b). The decreasing water potentials at 765-17 and O4-34 follow recharge events that peaked in September 2000 and October 2000, respectively, and the drying trends represent subsequent drainage. Water potential decreases in the sedimentary interbeds ranged from a slow decrease of approximately 16 cm at I4S-30 to a more substantial decrease of 58 cm at I1D-69 for the 2.5-yr study. The cyclic rise and decline in water potentials (high frequency oscillations) that is most prominent in I3S-28 and I1D-69, but also occurs in I4S-30, appears to be related to barometric fluctuations.

Several different mechanisms might result in the observed drying trends at depth. First, it is possible the drying trends are the result of decreased surface recharge. The wells are shown in Fig. 4, with red circles, for comparison with steady-state locations (blue squares). These wells that reflect drying trends are, with the exception of 76-5 and O-4, located along the main east–west drainage ditch that runs through the center of the SDA. For sampling purposes, these wells were sited to increase the chances of encountering perched water. These ditches collect surface water from surrounding areas and focus run-off and infiltration along the road. Areas near drainage ditches receive more surface infiltration during years of higher precipitation and run-off than areas farther from these ditches (Bishop, 1998). Sedimentary interbeds beneath those areas of focused infiltration may be responding to decreased run-off from 3 yr of less than average precipitation (2000–2002). Development of this hypothesis is limited by the lack of wells sited away from drainage ditches in the SDA, and by the need for nested tensiometers near the ditches to track surface recharge.

Second, it was considered that the long-term drying trends detected at some locations in the BC and CD interbed sediments could be the result of drainage and re-equilibration after flooding of pits and trenches (Bargelt et al., 1992). This flooding, caused by rapid snowmelt combined with heavy rains, recharged volumes of water to the subsurface, estimated at 37000, 24700, and 10200 m3 during 1962, 1969, and 1982, respectively (Keck, 1995). Although open pits and trenches that were flooded were identified, the lateral extent of the flooding can only be estimated. An estimate of the lateral coverage during the flooding events included most of the SDA, with the exception of the area north of the main east–west road, near I-4S and I-4D (Holdren et al., 2002). Wells I-4S and I-4D are not located within the estimated flooded areas, yet monitored depths within those wells exhibit long-term drying trends. Other wells, such as I-2S, I-2D, 76-5, and 77-2 are located within the estimated flooded areas; however, steady-state conditions, rather than long-term drying trends, are observed in the BC and CD sediments at these wells. These observations suggest the drainage is not related to the historical flooding.

A third possible mechanism for the drying trends observed in the BC and CD interbed sediments is drainage following lateral underflow from the spreading areas, west and south of the SDA (Fig. 1), into the sedimentary interbeds beneath the SDA. The last discharge to the spreading areas occurred in 1999, a year before the start of the advanced tensiometer monitoring. Monitoring during periods of discharge to the spreading areas, with nested advanced tensiometers that monitor a vertical profile, would be needed to fully evaluate the influence of lateral flow from the spreading areas.

Two tensiometers, O1-30 and O4-69, showed a gradual rise in water potentials over the monitoring period. The 20 cm of water rise in water potentials at O1-30 indicated long-term wetting. At O4-69, in the CD sedimentary interbed (Fig. 7b), a 75-cm rise in water potentials occurred from May 2000, when monitoring began, to March 2002. This increase in water potentials occurred in the same time frame as the drying of sediments in the shallower BC interbed at the same well, O4-34. Possible causes for the rise in moisture include downward drainage from overlying basalt or BC sedimentary interbed or delayed inflow from the spreading areas. Monitoring is needed during and after periods of inflow to the spreading areas to determine if moisture changes in these tensiometers are related to inflow from the spreading areas.

The long-term transient trends observed at nine of the advanced tensiometers, from 17 to 70 m bls, are in contrast to the near steady-state assumption proposed in the conceptual model. These long-term water potential trends suggest portions of the deep vadose zone may be responding to recent (the last 3 yr) changes in surface infiltration, and the response occurs at variable depths and locations.

Unit Gradient
Vertical flow is driven by gravity rather than differences in water potential under a unit hydraulic gradient. Intermittent recharge becomes dampened with depth, resulting in water movement by the force of gravity. This results in constant percolation rates at depth, which allow assignment of constant infiltration at the surface for environmental assessment modeling. The assumption of a unit gradient also simplifies calculation of deep percolation rates. Water potential values from tensiometers can be used to calculate percolation, assuming a unit gradient and hydraulic conductivities are known.

A unit gradient exists when there is little to no change in water potential with respect to depth or elevation, resulting in a hydraulic gradient of one. Assuming one-dimensional, vertical flow, Darcy's Law states

[1]
where q is the flux, K({Psi}) is the hydraulic conductivity as a function of water potential, and dh/dz is the hydraulic gradient. The hydraulic gradient, dh/dz, is composed of both water and elevation potentials such that

[2]

If there is no change in water potential with depth, then the change in total head is only due to changes in elevation and the hydraulic gradient is one, or unity. Under these conditions, the flux, or deep percolation rate, equals the hydraulic conductivity at the specified water potential.

Measured water potentials from 22 Aug. 2002, 0800 h at each advanced tensiometer location were used to calculate hydraulic head (the sum of the average water potential, in meters, and the elevation head) and are shown vs. elevation head in Fig. 9 . The vertical variation in soil water potentials (–30 to –400 cm of water) was small compared with differences in elevation. The hydraulic gradient, or change in hydraulic head vs. elevation is close to one, indicating a unit gradient condition exists in the upper 73 m of the unsaturated zone at the SDA.



View larger version (12K):
[in this window]
[in a new window]
  		Fig. 9. Slope of elevation head vs. hydraulic head is near one, indicating a unit gradient over the upper 73 m of the unsaturated zone in the subsurface in and around the Subsurface Disposal Area.

 
It is recognized that there are limitations in applying the unit gradient method to the tensiometer data. The tensiometers are widely spaced across an area of approximately 1 km2 and do not represent a vertical profile at one location. Local hydraulic gradients within the interbeds cannot be determined from this tensiometer array. The unsaturated subsurface at the RWMC is heterogeneous and is not an ideal, one-dimensional, steady-state flow system. Episodic infiltration events occur at the surface that may extend to the depth of the BC interbed. Long-term drying and wetting trends were observed at some of the deeper tensiometers. There is a potential for subsurface lateral flow from the spreading areas to the SDA, lateral flow along sloping surfaces and local discontinuities, and water may take preferential paths through basalt fractures.

However, application of the unit gradient method does have some validity. The lack of water in the spreading areas and lower than average precipitation during this time period (spring 2000 through September 2002) decreased the potential impact of infiltration events at the surface or lateral flow from the spreading areas. No infiltration events were observed for at least 1 yr before the 22 Aug. 2002 date used for gradient calculations. Steady-state conditions were observed at the majority of the deeper (17 m or deeper) tensiometers, and the transient conditions that were observed occurred slowly, in the 2.5-yr time period. A previous investigation (Hubbell et al., 2002) indicates near unit gradients existed along a 31-m vertical profile in the SDA, using the nested tensiometers in 76-5.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 SITE DESCRIPTION
 CONCEPTUAL FLOW MODEL
 METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Water potential data from 30 deep advanced tensiometers were evaluated to determine the appropriateness of the steady-state flow and unit gradient assumptions used to describe water flow in a geologically complex vadose zone. Temporal water potential trends were evaluated in basalts and sediments from 6.7 to 73.5 m bls, beneath and adjacent to the SDA from February 2000 through August 2002. Measured water potentials ranged from near saturation (–30 cm of water) to about –400 cm of water.

In the near surface basalts and sediments (above 17 m), long-term drying trends were observed, presumably in response to the cumulative effect of less than average annual precipitation for the last 3 yr (2000–2002). Below 17 m, steady-state conditions were observed at 15 of 25 monitored locations. Nine sites below 17 m showed long-term drying and wetting trends, in contrast to the steady-state assumption in the conceptual model. Long-term drying, observed at seven of the nine trending sites, occurred at locations near the east–west drainage ditches. It was hypothesized that the subsurface beneath areas of focused infiltration (such as in drainage ditches) was also responding to decreased run-off from the last 3 yr (2000–2002) of less than average precipitation. The appropriateness of the unit gradient assumption in the conceptual flow model was evaluated by aligning dispersed monitoring locations along a presumed vertical profile. The resulting hydraulic gradient was close to one, suggesting a gravity gradient dominated vertical flow from 7 to 73 m.


   ACKNOWLEDGMENTS
 
Work supported by the U.S. Department of Energy, Assistant Secretary for Environmental Management, under DOE Idaho Operations Office Contract DE-AC07-99ID13727. Mention of trademark propriety products are for the benefit of the readers and do not constitute an endorsement for the products by the Department of Energy to the exclusion of other products that may also be suitable. The authors thank Indrek Porro, Swen Magnuson, Earl Mattson, and two anonymous reviewers for their insightful comments on this report.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 SITE DESCRIPTION
 CONCEPTUAL FLOW MODEL
 METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 Vadose Zone JHome page
J. C. Seaman, B. B. Looney, and M. K. Harris
Research in Support of Remediation Activities at the Savannah River Site
Vadose Zone J., May 17, 2007; 6(2): 316 - 326.
[Abstract] [Full Text] [PDF]

Home page
 Vadose Zone JHome page
E. D. Mattson, S. O. Magnuson, and S. L. Ansley
Interpreting INEEL Vadose Zone Water Movement on the Basis of Large-Scale Field Tests and Long-Term Vadose Zone Monitoring Results
Vadose Zone J., February 1, 2004; 3(1): 35 - 46.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (5)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by McElroy, D. L.
Right arrow Articles by Hubbell, J. M.
Right arrow Search for Related Content
PubMed
Right arrow Articles by McElroy, D. L.
Right arrow Articles by Hubbell, J. M.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by McElroy, D. L.
Right arrow Articles by Hubbell, J. M.
Related Collections
Right arrow Flow
Right arrow Field-Scale Studies
Right arrow Recharge

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome