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


     

Published online 20 November 2007
Published in Vadose Zone J 6:935-945 (2007)
DOI: 10.2136/vzj2006.0176
© 2007 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
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 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 Web of Science (2)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Khaleel, R.
Right arrow Articles by Kristofzski, J. G.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Khaleel, R.
Right arrow Articles by Kristofzski, J. G.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Khaleel, R.
Right arrow Articles by Kristofzski, J. G.
Related Collections
Right arrow Effective Parameters
Right arrow Variably Saturated Fluid Flow
Right arrow Infiltration

SPECIAL SECTION: HANFORD SITE

Impact Assessment of Existing Vadose Zone Contamination at the Hanford Site SX Tank Farm

Raziuddin Khaleela,*, Mark D. Whiteb, Martinus Oostromb, Marcus I. Woodc, Frederick M. Mannd and John G. Kristofzskid

a Fluor Government Group, P.O. Box 1050, Richland, WA 99352
b Pacific Northwest National Lab., P.O. Box 999, Richland, WA 99352
c Fluor Hanford, Inc., P.O. Box 1000, Richland, WA 99352
d CH2M HILL Hanford Group, Inc., P.O. Box 1500, Richland, WA 99352
* Corresponding author (raziuddin_khaleel{at}rl.gov).

All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.


Received 11 December 2006.



   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 Modeling Approach
 Numerical Simulation Results
 Conclusions
 REFERENCES
 
The USDOE has initiated an impact assessment of existing vadose zone contamination at the Hanford Site SX tank farm in southeastern Washington State. The assessment followed the Resource Conservation and Recovery Act (RCRA) Corrective Action process to address the impacts of past tank waste releases to the vadose zone at the single-shell tank farm. Numerical models were developed that consider the extent of contamination presently within the vadose zone and predict contaminant movement through the vadose zone to groundwater. The transport of representative mobile (technetium-99) and immobile (cesium-137) constituents was evaluated in modeling. The model considered the accelerated movement of moisture around and beneath single-shell tanks that is attributed to bare, gravel surfaces resulting from the construction of the underground storage tanks. Infiltration, possibly nearing 100 mm yr–1, is further amplified in the tank farm because of the umbrella effect created by percolating moisture being diverted by the impermeable, sloping surface of the large, 24-m-diameter, buried tank domes. For both the base case (no-action alternative) simulation and a simulation that considered placement of an interim surface barrier to minimize infiltration, predicted groundwater concentrations for technetium-99 at the SX tank farm boundary were exceedingly high, on the order of 106 pCi L–1. The predicted concentrations are, however, somewhat conservative because of our use of two-dimensional modeling for a three-dimensional problem. A series of simulations were performed, using recharge rates of 50, 30, and 10 mm yr–1, and compared to the base case (100 mm yr–1) results. As expected, lowering meteoric recharge delayed peak arrival times and reduced peak concentrations at the tank farm boundary.

Abbreviations: bgs, below ground surface • BTC, breakthrough curve • RCRA, Resource Conservation and Recovery Act


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
Modeling Approach
 Numerical Simulation Results
 Conclusions
 REFERENCES
 
The Hanford Site is located on the Columbia River plateau, a semiarid region in south-central Washington State (Fig. 1 ), with an average annual precipitation of about 180 mm. The site served as a plutonium production facility for nuclear weapons from 1944 to the end of the Cold War era in 1989. A total of 177 large underground tanks were constructed in the Hanford vadose zone to store waste fluid streams from the plutonium extraction facilities. Many of the older single-shell tanks have leaked radioactive waste fluids, posing a contamination hazard for the underlying aquifer and ultimately, the Columbia River. This analysis specifically addresses the SX tank farm (Fig. 1), where highly radioactive and chemically hazardous aqueous fluids of high ionic strength have leaked into the vadose zone. The nature and extent of subsurface contamination; the past, present, and future migration of contaminants (Pruess et al., 2002); and issues posed by possible future remedial actions are of interest in assessing the impacts associated with the SX tank farm.


Figure 1
View larger version (31K):
[in this window]
[in a new window]

  		FIG. 1. Location map of single-shell tank Waste Management Area (WMA) S-SX and surrounding facilities in the 200 West Area, Hanford Site, Washington State.

 
The SX tank farm is located within the 200 West Area (Fig. 1). Perturbations of the natural hydrogeologic system at the SX tank farm occurred because the ground was excavated down approximately 15 m below ground surface (bgs), and 15 large cylindrically shaped storage tanks with approximate dimensions of 14-m height and 24-m diameter were emplaced in a regular pattern with about 30-m spacing between tank centers (Fig. 2 ). The excavated material was then backfilled, and an approximate 2-m gravel layer was placed on top. The altered hydrogeologic properties in the backfilled region led to a substantial increase in net infiltration. The moisture migration within the unsaturated zone was further altered by the umbrella effect of the tanks that diverts moisture around the tank perimeters (Kline and Khaleel, 1995; Ward et al., 1997).


Figure 2
View larger version (134K):
[in this window]
[in a new window]

  		FIG. 2. 241-SX tank farm under construction.

 
As part of site characterization, a number of studies have been conducted at the S and SX tank farms (Fig. 1). These include characterization of subsurface geology and subsurface conditions (Price and Fecht, 1976; Myers et al., 1998; Johnson et al., 1999; Myers, 2001), vadose zone contamination and contaminant inventories (Raymond and Shdo, 1966; Goodman, 2000; Jones et al., 2000), laboratory investigations (McKinley et al., 2004; Liu et al., 2003; Steefel et al., 2003; Zachara et al., 2002; Serne et al., 2001a, b), and modeling (Ward et al., 1997; White et al., 2001; Pruess et al., 2002; Steefel et al., 2003; Lichtner et al., 2004). A compendium of knowledge on subsurface conditions, hydrogeochemical processes, and nature and extent of vadose zone contamination for the S-SX tank farms has been provided in a recently published field investigation report (Knepp, 2002). Because infiltration is a major driver for transport of long-lived mobile radionuclides (e.g., technetium-99), the report considered several management options for controlling infiltration. One of the management options is emplacement of interim surface barriers to reduce infiltration of moisture and thereby reduce long-term risks from groundwater contamination at the SX tank farm. The objective of our investigation was to evaluate the effectiveness of interim barriers to the infiltration (recharge) of meteoric water (from precipitation and snowmelt) on the migration of contaminants from previous leak sources. The reference suite of simulations (base case or no-action alternative) considers the migration of contaminants from field estimates of concentration distributions through the vadose zone to the tank farm fenceline boundary with no interim barriers but a closure barrier by the year 2040. The base case results were compared to results for a scenario with an interim barrier placed in 2010, followed by a closure barrier in 2040. The two-dimensional cross-sectional simulations were run for 1000 yr. Two radionuclides, cesium-137 (short-lived with a half-life of 30.14 yr and relatively immobile) and technetium-99 (long-lived with a half-life of 2.03 x 105 years and mobile) were evaluated. The simulations predicted that cesium-137 does not migrate to the fenceline boundary and that technetium-99 does; therefore, results are presented only for technetium-99.

Geology and Vadose Zone Contamination Data
The generalized stratigraphy beneath the Hanford Site 200 West Area consists of, in ascending order, the Columbia River Basalt Group, the Ringold Formation, the Cold Creek unit, and the Hanford formation. The geologic cross-section used in two-dimensional cross-sectional modeling for the tank row consisting of tanks SX-107, SX-108, and SX-109 is shown in Fig. 3 . Unlike other strata that are comprised of sandy or gravelly sand/sandy gravel, the Cold Creek unit consists primarily of fine sand and silt and has higher moisture-holding capacity compared with other units. Another noticeable feature of the cross-section is the southwesterly dip of the coarse-grained unit. The fenceline boundary is located on the lower right corner of the cross-section (Fig. 3).


Figure 3
View larger version (22K):
[in this window]
[in a new window]

  		FIG. 3. Geologic cross-section through tank row SX-107, SX-108, and SX-109. Depths and distances are shown in feet (ft) because of its usage at the Hanford Site. (bgs, below ground surface.)

 
Table 1 provides the estimated contaminant inventory in the vadose zone at the S and SX tank farms. For the tank farm cross-section considered in this study, the vadose zone contamination in the region is dominated by the waste leaked from tank SX-108 between 1962 and 1967. Smaller-volume leaks occurred from tanks SX-107 and SX-109. While other tanks are listed in Table 1, the focus of our study is contamination that is primarily derived from tank SX-108 leak.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Tank leak volume estimates and estimated contaminant inventory in the vadose zone around the S and SX tank farms.

 
In 1998 borehole 41-09-39 (W23-234 in Fig. 4 ), located near tank SX-108, was sampled in two episodes following its initial completion. First, the borehole was extended via drilling from about 39.6 m (130 ft) bgs to groundwater, about 65.5 m (215 ft) bgs, with near continuous sampling obtained during the extension. Then the upper portion was sampled via coring of borehole sidewall from 39.6 to 7.6 m (130–25 ft) bgs as the borehole was decommissioned; the sidewall core samples were of 2.5 x 2.5 x 28 cm.


Figure 4
View larger version (58K):
[in this window]
[in a new window]

  		FIG. 4. Technetium-99/(nitrate + nitrite) ratios for Waste Management Area (WMA) S-SX network wells (the ratios highlighted in green and red). The technetium-99/(nitrate + nitrite) ratios for tanks SX-108 and SX-115 are 0.43 and 0.42 pCi µg–1, respectively. All well numbers are prefixed by 299-. RCRA = Resource Conservation and Recovery Act.

 
A slant (~30° off vertical) borehole (W23-64 in Fig. 4) was also drilled near tank SX-108. Sixteen core samples were recovered from the slant borehole at depths from 16.8 to 43.9 m (55–144 ft) bgs. The sediment characterization data were augmented with spectral gamma logging data from 98 drywells (not shown) surrounding the SX tanks. The spectral gamma logs show that the vast majority of the cesium-137 is in the upper regions of the vadose zone between 0 and 12.2 m (40 ft) below the tank bottoms, that is, 16 to 28.3 m (53–93 ft) bgs, and within the coarse-grained unit that dips to the southwest (Fig. 3). Some cesium-137, however, is present as deep as 30.5 m (100 ft) below the tank bottoms in several places.

The sediment characterization data from the recent boreholes show that the bulk of the cesium-137 under tank SX-108 is found between 0 and 12.2 m (40 ft) below the tank bottom, that is, 16 to 28.3 m (53–93 ft) bgs. At borehole 41-09-39, the bulk of the cesium-137 is found between 1.8 and 16.8 m (6–55 ft) below tank bottoms, that is, 18 to 33 m (59–108 ft) bgs. Figure 5 shows the vadose zone contaminant profiles for cesium-137, technetium-99, and sodium, based on borehole 41-09-39 sampling for tank SX-108 leak. As described below, Fig. 5 is primarily used to develop initial conditions for vadose zone contaminant distribution.


Figure 5
View larger version (56K):
[in this window]
[in a new window]

  		FIG. 5. Vadose zone contamination for cesium-137, technetium-99, and sodium near tank SX-108 (after Serne et al., 2001a). A composite sample is obtained by combining two adjacent sleeves of the borehole liner. Depths are shown in feet (ft) because of its usage at the Hanford Site.

 
Conceptual Model for Migration of Tank Leak
Historically, it was assumed that cesium-137 sorbs readily on to Hanford sediments, drops out of solution quickly, and hence is not available for transport through the vadose zone. However, in the SX tank farm, cesium-137 was found farther into the vadose zone than the preceding hypothesis would suggest. An extensive effort (Serne et al., 2001a, b) to understand this observation was undertaken, and it was found that for tank leak SX-108, supernate chemistry controlled the local chemical environment near the leak. Historical knowledge of tank supernate chemistry and laboratory sorption work (McKinley et al., 2004; Liu et al., 2003; Zachara et al., 2002), as well as numerical calculations (Steefel et al., 2003; Lichtner et al., 2004), indicated that optimum conditions existed at the time of tank leak to cause higher-than-postulated cesium-137 mobility. The primary enabling conditions were a combination of very high sodium and cesium-137 concentrations in tank fluid. Under these conditions, sodium was more strongly sorbed, causing cesium-137 to remain in solution.

From these observations, a simplified two-step model for how the SX-108 leak migrated into the vadose zone was postulated. In the first step, a rather rapid release of the leaking fluid occurred, providing a hydraulic driver for the fluid to move through the vadose zone. Stratigraphic variability in the vadose zone geology exerted sufficient influence to induce movement initially in the lateral direction to the southwest (Fig. 3). The supernate chemistry controlled the local geochemical conditions to enhance the mobility of cesium-137 so that the cesium was poorly retarded and migrated essentially with the leaking tank fluid. As discussed in Serne et al. (2001a), the ratio of cesium-137 to technetium-99 activity in the shallow sediments from 18.3 to 25 m (60–82 ft) bgs was greater than 10,000. Such high ratios suggest that cesium-137 and technetium-99 may be traveling at the same rate through these shallow sediments below the tank bottom. However, the potential exists that the presence of some of the cesium-137 in deep vadose zone sediments is due to drag-down (push-down) along the borehole.

In the second step, the hydraulic driver for the leak event eventually relaxed, and the moisture distribution within the far-field (away from the leak source and at depths in excess of 10 m below tank bottom) vadose zone equilibrated with natural infiltration. Far-field conditions reverted to more natural soil-water conditions (the soil essentially buffered the high sodium tank chemistry), thus again causing cesium-137 to sorb strongly to sediments. As indicated in Fig. 5, subsequent movement of cesium-137 was held to a minimum. Between the depths of 27.4 and 40.9 m (90 and 134 ft) bgs, the cesium-137 to technetium-99 ratio ranges from less than 1 to about 100 (Serne et al., 2001a), which suggests that cesium-137 eventually is removed from the fluid. However, technetium-99 as well as other nonsorbing constituents (e.g., nitrate, nitrite, and sodium) remained mobile. Figure 5 shows a reduction of sodium concentration and a distinct chromatographic separation of cesium-137 and sodium with depth. Once cesium-137 and sodium were sufficiently separated, cesium-137 sorption became favorable, fixing it in place. Site characterization data (Fig. 5) also suggest that the bulk of the radionuclide inventory from the SX-108 leak is still within the vadose zone and, as discussed below, only a small fraction of contamination from the leak reached the unconfined aquifer.


   Modeling Approach
TOP
 ABSTRACT
 INTRODUCTION
 Modeling Approach
Numerical Simulation Results
 Conclusions
 REFERENCES
 
The simulations in this study did not model the impact of a tank leak itself but modeled the impact from past tank leaks, as quantified by existing contamination within the vadose zone (Fig. 5). All simulations were composed of steady-flow and transient components, where flow fields developed from the steady-flow component were used to initialize the transient simulation. From the starting conditions, transient simulations of solute transport were conducted for a 1000-yr period (i.e., year 2000 CE to 3000 CE) that involved changes in the flow fields in response to placement of surface barriers.

Fluid flow within the vadose zone was described by Richards' equation, whereas the contaminant transport was described by the conventional advective–dispersive transport equation with an equilibrium linear sorption coefficient (Kd) formulation. Data on laboratory measurements for moisture retention, particle-size distribution, saturated and unsaturated hydraulic conductivity, and bulk density for individual stratum were based on samples from 200 East and 200 West Areas (Khaleel et al., 2000; Khaleel and Freeman, 1995). Samples that contained measurements of soil moisture retention as well as unsaturated hydraulic conductivity were used. This was primarily to avoid extrapolating unsaturated conductivities (van Genuchten, 1980; Mualem, 1976) to the dry end, based on moisture retention and saturated conductivity estimate (Khaleel et al., 1995). Also, to reflect field conditions, the laboratory data were corrected for the presence of any gravel fraction in the sediment samples (Khaleel and Relyea, 1997). As with flow modeling, each stratum was modeled with a range of transport parameters (i.e., bulk density, diffusivity, and macrodispersivity).

Initial Conditions for Flow
Steady-state initial conditions were developed by simulating from a prescribed unit hydraulic gradient condition to a steady-state condition, dictated by the initial meteoric recharge at the surface, variation in soil hydraulic properties, location of impermeable tanks, and no flux vertical boundaries above the water table. The water table boundary was prescribed by the water table elevation (65.5 m [215 ft] bgs) and an unconfined aquifer hydraulic gradient (0.0012 m m–1) from west to east. No-flux boundaries were used for the lower boundary (Fig. 3). No solute transport was considered during the steady-state simulation.

Model Setup and Boundary Conditions
A two-dimensional (west–east) vertical (x–z) slice of the flow domain was used for modeling flow and transport. The flow domain includes the tank farm fenceline boundaries and extends vertically from the ground surface to about 6 m below the water table. The geologic strata are assumed continuous but not of constant thickness. For flow modeling, Neumann boundary conditions are prescribed at the surface with the flux equal to the recharge rate estimate. For transport modeling, a zero flux boundary was prescribed at the surface for cesium-137 and technetium-99. Above the water table, the western and eastern boundaries were assigned no-flux boundaries for both flow and transport.

Recharge Estimates, Flow and Transport Parameters, and Inventory Estimates
This section provides the recharge estimates as well as the effective (upscaled) values of flow and transport parameters for the vadose zone. Specific flow parameters include moisture retention and saturated and unsaturated hydraulic conductivity. Transport parameters include bulk density, diffusivity, sorption coefficient for cesium-137, and macrodispersivity. Details on deriving the effective (upscaled) parameters are addressed in Khaleel et al. (2000). The section concludes with how the contaminant inventory was assigned as a function of elevation within the vadose zone.

Recharge Estimates
The tank farm surfaces are covered with gravel to prevent vegetation growth and provide radiation shielding for site workers. Bare gravel surfaces, however, enhance net infiltration of meteoric water compared with undisturbed naturally vegetated surfaces (Gee et al., 1992). Infiltration is further enhanced in the tank farms by the effect of percolating water being diverted by the impermeable, sloping surface of the tank domes. The basis for recharge estimates (Table 2) is presented in Khaleel et al. (2000).


View this table:
[in this window]
[in a new window]

  		TABLE 2. Estimated recharge for interim and closure barriers.

 
Effective Flow Parameters
Table 3 lists the composite, fitted van Genuchten–Mualem parameters (van Genuchten, 1980) for various strata at the tank farm. As stated above, to obtain a better fit for unsaturated conductivity for the moisture regime of interest, the parameters are based on a simultaneous fit of both moisture retention and unsaturated conductivity data so that the saturated hydraulic conductivity, Ks (Table 3) is a fitted parameter, not the laboratory-measured value. Also, because of a lack of information, sandy gravel data were used to represent properties for backfill sediments.


View this table:
[in this window]
[in a new window]

  		TABLE 3. Average bulk density and composite van Genuchten–Mualem parameters for various strata.{dagger}

 
The modeling accounts for moisture- or tension-dependent anisotropy (Yeh et al., 1985) for Hanford sediments. The heterogeneous nature of Hanford sediments is effective in smearing out the effects of large natural or manmade water applications. This was illustrated by the moisture content profiles at a controlled field injection experiment site in the 200 East Area (Sisson and Lu, 1984; Ward et al., 2000; Ye et al., 2005). The field data for the injection experiment also suggests moisture- or tension-dependent anisotropy (Yeh et al., 2005; Ward et al., 2006). Variable, tension-dependent anisotropy provides a framework for upscaling laboratory-scale measurements to the effective (upscaled) properties for the large-scale tank farm vadose zone.

The Polmann (1990) model was used to describe tension-dependent anisotropy for sediments at the tank farm; details are in Khaleel et al. (2000, Appendix C). To account for moisture-dependent anisotropy, the Gardner (1958) relationship was used to describe unsaturated hydraulic conductivity (K) as a function of Ks and tension, h. The Gardner model is referred to as the log-linear model, because lnK is linearly related to h through the Gardner slope ß, the pore-size distribution parameter. However, such a constant slope is often inadequate in describing lnK(h) over ranges of tension of practical interest for field applications. Therefore, the slope ß was approximated locally by straight lines over a fixed range of tension. The lnKs in Gardner equation was then derived by extrapolating the local slopes back to zero tension.

Using a linear correlation relationship between the log-conductivity zero-tension intercept and ß, the Polmann (1990) generalized model accounts for the cross-correlation of the local soil property (i.e., lnKs and ß) residual fluctuations. Compared to the uncorrelated lnKs and ß model (Yeh et al., 1985), a partial correlation of the properties has a significant impact on the magnitude of the effective parameters derived from the stochastic theory.

Effective Transport Parameters
Similar to flow parameters, because of natural variability, the transport parameters are spatially variable. The purpose was again, similar to the flow parameters, to include the effect of such variability on the large-scale transport process.

The average bulk density estimates (Table 3) were based on data in Khaleel et al. (2000) for the five strata. A sorption value of 500 mL g–1 was used for cesium-137 (Kaplan and Serne, 1999) as being representative of undisturbed sediments. The sorption coefficient for technetium-99 was estimated to be zero. The effective, large-scale diffusion coefficients for tank farm soils were assumed to be a function of volumetric moisture content, {theta}, and are expressed using an empirical relation (Millington and Quirk, 1961).

Macrodispersivities were defined in a manner similar to saturated media estimates (Gelhar and Axness, 1983). Macrodispersivity estimates are needed for both reactive (i.e., cesium-137) and nonreactive (i.e., technetium-99) species. The estimates are different for cesium-137 than those for technetium-99 because, unlike nonsorbing technetium-99, the sorption for cesium-137 varies as the soil properties experience spatial variability (Gelhar, 1993; Talbott and Gelhar, 1994). Such variation affects the contaminant velocity, which, in turn, enhances the spreading of the plume. For reactive species (cesium-137), the enhanced spreading is defined by a macrodispersivity enhancement, relative to spreading for nonreactive (technetium-99) species.

An extended review is provided in Khaleel et al. (2000) on the rationale for macrodispersivity estimates; Table 4 gives the nonreactive estimates for technetium-99. The transverse values were estimated to be 10% of the longitudinal macrodispersivities. For cesium-137, the macrodispersivity enhancement ranged from about 1.07 for backfill sediments to about 2.35 for Cold Creek unit sediments. Stochastic theory (Gelhar, 1993) indicates that the macrodispersivity enhancement only occurs in the longitudinal direction.


View this table:
[in this window]
[in a new window]

  		TABLE 4. Nonreactive longitudinal (AL) and transverse (AT) macrodispersivity estimates for various strata for technetium-99.

 
Inventory Assignment by Depth
Figure 5 was used to assign inventory within the vadose zone profile. First, the estimated inventory (Table 1) in the vadose zone due to contribution from tanks SX-108 and SX-109 were added; the total inventory under both tanks was then distributed in the lateral direction across the two tanks (including the space between the tanks) (Fig. 6a and 7a ). For tank SX-107, the inventory is distributed across the diameter of that tank only (Fig. 7a).


Figure 6
View larger version (42K):
[in this window]
[in a new window]

  		FIG. 6. Progression of saturation profile at year (a) 2000, (b) 2050, (c) 2540, and (d) 3000. Depths and distances are shown in feet (ft) because of its usage at the Hanford Site. (bgs, below ground surface.)

 

Figure 7
View larger version (44K):
[in this window]
[in a new window]

  		FIG. 7. Progression of technetium-99 distribution at year (a) 2000, (b) 2050, (c) 2540, and (d) 3000. Depths and distances are shown in feet (ft) because of its usage at the Hanford Site. (bgs, below ground surface.)

 
Next, both contaminant concentration (Fig. 5) and mass (Table 1) were constrained in distributing the inventory in the lateral direction for each depth within the vadose zone. In assigning the contaminant inventory by depth, it was assumed that higher inventories are associated with higher concentrations. Thus, the depth distribution consisted of a series of concentric disks of varying concentration and diameter with depth (Fig. 7a); the disk diameter was sized to maintain the total solute mass for a particular depth. Note that Fig. 5 is based on borehole 41-09-39 data; contaminant concentrations were slightly higher at borehole 41-09-39, compared to those at the slant borehole.


   Numerical Simulation Results
TOP
 ABSTRACT
 INTRODUCTION
 Modeling Approach
 Numerical Simulation Results
Conclusions
 REFERENCES
 
All simulations reported herein were performed using the STOMP simulator (White and Oostrom, 2000a, 2000b). A detailed discussion on the numerical implementation for STOMP and simulation results is presented in White et al. (2001). Selected results are summarized in Fig. 6 through 12 and are discussed below.

Base Case: No Action Alternative
The base case simulation investigated solute transport considering natural surface infiltration, no interim surface barriers, but with a closure barrier by the year 2040. The closure barrier limits infiltration to 0.1 mm yr–1 for the first 500 yr. The simulations were initialized using a steady-state flow solution defined by a surface recharge rate of 100 mm yr–1 and a hydraulic gradient in the unconfined aquifer. Simulated results are presented for relative saturation ({theta}/{theta}s where {theta} = moisture content and {theta}s = saturated moisture content) distribution as well as for the solute concentration distribution for years 2000, 2050, 2540, and 3000. The moisture content field for these simulations remains unchanged from the initial steady-flow field until the year 2040, when the closure barrier becomes effective.

The saturation field is dependent on the surface recharge, impermeable structures (e.g., single-shell tanks), various strata, and their hydraulic parameters. The steady-flow saturation field with 100 mm yr–1 of meteoric recharge is shown in Fig. 6a. This field shows the impact of the tanks on the relative saturation, where higher than ambient saturations occur above and between the tanks and lower than ambient saturations occur just below the tanks.

By 2040 a closure barrier was assumed to be active, which lowered the meteoric recharge from 100 mm yr–1 to 0.1 mm yr–1. The saturation field dried in response to the change in surface recharge, as shown in Fig. 6b for the year 2050. A comparison of Fig. 6a and 6b shows that, at a lower value of surface recharge, the impact of the impermeable tanks on saturation is reduced. The closure barrier was assumed to remain effective for 500 yr, at which point it degrades, allowing meteoric recharge to increase to 3.5 mm yr–1. In the 500-yr period between 2040 and 2540, the saturation field continued to slowly dry reaching a minimum average level at 2540, as shown in Fig. 6c. In the 460-yr period between 2540 and 3000, the saturation field wetted, in response to the increased meteoric recharge of 3.5 mm yr–1 for the degraded surface barrier. The saturation field at year 3000 is shown in Fig. 6d. The closure barrier was assumed to decrease the meteoric recharge from 100 mm yr–1 to 0.1 mm yr–1 for 500 yr and then degrade to 3.5 mm yr–1 for the next 460 yr. The variations in surface infiltration had the greatest impact on saturation in the region between tanks within the backfill material and the region immediately below the bottom of the tanks. As evident from Fig. 6a to 6d, the Cold Creek unit showed the least change in saturation with change in the surface recharge. The regions directly beneath the tanks additionally showed lower variability in relative saturation.

Color-scaled images for technetium-99 are shown as a series of time sequences in Fig. 7a to 7d. The solute concentrations are color valued using exponential scaling from the maximum concentration limit of 900 pCi L–1 for technetium-99 to the maximum initial inventory concentration. Note the differences in time sequences (Fig. 7a–7d); the initial inventory was spread across the region between tanks SX-108 and SX-109 and centered beneath tank SX-107. Although differences are noticeable in the plume located between tanks SX-109 and SX-108 with the plume beneath tank SX-107, the overall rate of migration toward the groundwater is nearly identical. The initial inventory beneath tank SX-107 shows a slight delay in arrival times. Some variations in solute migration direction are noted due to the sloped gravelly sand strata, but none substantially alter the breakthrough concentrations at the SX fenceline boundary, located on the lower right-hand corner of the domain (Fig. 7a–7d).

Figure 8 illustrates the simulated water flux at the water table as a function of time, whereas Fig. 9 shows the solute breakthrough curve (BTC) for technetium-99 at the tank farm fenceline. No cesium-137 was transported to the fenceline boundary for the simulated cross-section. This is true for the base case as well as for the interim barrier case.


Figure 8
View larger version (15K):
[in this window]
[in a new window]

  		FIG. 8. Simulated flux versus time at the water table for the base case with no interim barrier.

 

Figure 9
View larger version (20K):
[in this window]
[in a new window]

  		FIG. 9. Simulated breakthrough curve for technetium-99 at the tank farm fenceline for the base case with no interim barrier.

 
Interim Barrier Case
The interim barrier case investigated solute transport considering natural surface infiltration, an interim surface barrier placed in the year 2010, and a closure barrier in the year 2040. Unlike the closure barrier, which limits the infiltration rate to 0.1 mm yr–1 for the first 500 yr, the interim barrier limits the infiltration rate to 0.5 mm yr–1. Again, the simulations were initialized using a steady-state flow solution defined by the surface recharge rate of 100 mm yr–1 and a hydraulic gradient in the unconfined aquifer. Inventories for the two contaminant species and their distribution within the vadose zone were identical to those for the base case. The moisture content field for these simulations remains unchanged from the initial steady-flow. Figure 10 illustrates the simulated water flux at the water table as a function of time, whereas Fig. 11 shows the solute BTC for technetium-99 at the tank farm fenceline.


Figure 10
View larger version (15K):
[in this window]
[in a new window]

  		FIG. 10. Simulated flux versus time at the water table for the interim barrier case.

 

Figure 11
View larger version (19K):
[in this window]
[in a new window]

  		FIG. 11. Simulated breakthrough curve for technetium-99 at the tank farm fenceline for the interim barrier case.

 
Breakthrough times and peak technetium-99 concentrations at the tank farm fenceline boundary are shown in Table 5 for cases with and without interim barrier. For comparison, maximum initial concentrations are also included in Table 5. Results indicate that, compared with the base case, the interim surface barrier reduces solute concentration at the tank farm fenceline but has negligible impact on the technetium-99 peak concentration arrival time. Although results are not shown here, the impact of the interim surface barrier on lowering peak concentrations is most affected by the initial vadose zone contaminant inventory distribution. Inventory distributions having concentrations of solute mass nearer the water table were less affected by the interim barrier compared with the inventory mass located higher up in the vadose zone. Concentrations at the fenceline boundary at year 3000 were always higher for the interim barrier simulation compared with the base case simulation, which indicates that a major benefit of the interim barrier is to dampen the technetium-99 flux into the aquifer, broaden the breakthrough curve at the fenceline, and reduce peak concentrations.


View this table:
[in this window]
[in a new window]

  		TABLE 5. Simulated peak concentrations and arrival times for technetium-99 at the tank farm fenceline.

 
Impact of Recharge Rates
To investigate the impact of meteoric recharge on contaminant transport, a series of simulations were performed, using recharge rates of 50, 30, and 10 mm yr–1, and compared to the base case (100 mm yr–1) results. In all cases, the closure barrier was assumed to control recharge beyond the first 40 yr. The initial flow field for these simulations was the steady-state condition for the prescribed meteoric recharge rate and fixed groundwater conditions (i.e., water table conditions and gradient). As expected, lowering meteoric recharge delayed peak arrival times and reduced peak concentrations at the fenceline boundary (Fig. 12 ). The exception was the 10 mm yr–1 recharge case (Fig. 12), where arrival times were delayed beyond 1000 yr. However, because of the late arrival, the peak concentrations for the low-recharge case continued to increase for the 1000-yr simulation. Nonetheless, the concentration at the boundary, after 1000-yr simulation, was still well below those for other recharge rates.


Figure 12
View larger version (24K):
[in this window]
[in a new window]

  		FIG. 12. Simulated breakthrough curves for technetium-99 at the tank farm fenceline for various recharge estimates.

 
All preceding simulations were performed using an isothermal model. Such an assumption is supported by a comparison of simulation results for nonisothermal and isothermal runs (White et al., 2002). Nonisothermal model simulations indicate that during periods of high heat loads in the 1950s and 1960s, the thermal load from the boiling waste tanks altered flow patterns and caused large-scale redistribution of moisture. As a result, fluid and vapor flow near the high-heat tanks was dominated by vapor–liquid counterflow. To understand the historical behavior, it is therefore important to consider the strong coupling between the thermal and hydrologic environments. For impact assessment, however, long-term simulation results for technetium-99 migration for isothermal and nonisothermal conditions were not significantly different (White et al., 2002), despite the known thermal effects of high-heat tanks.

Comparison of Base Case Model Results with Other Studies
Ward et al. (1997) investigated technetium-99 transport due to tank leaks for the same cross-section considered in this study. Their modeling approach, however, was different from ours. Unlike ours, they modeled tank leaks, whereas we used the existing vadose zone contamination as our initial condition for technetium-99 migration through the vadose zone and on to groundwater. Ward et al. (1997) found that, in addition to leak volume estimates and fluid density estimates of the leaked waste, recharge rates had a significant impact on technetium-99 migration. Under high recharge rates, technetium-99 migrated rapidly to the water table, arriving within 30 yr of the initial leak. For a leak volume of 500 kL (132,000 gal), depending on recharge rates and specific gravity of leaked fluid (1.0 to 1.4), their simulated technetium-99 peak porewater concentrations at the water table were in the range of 2 x 106 to 12 x 106 pCi L–1. Because of groundwater mixing, our simulated peak concentration at the tank farm fenceline is lower (Table 5) but within the same order of magnitude of Ward et al. (1997) results for the no-action alternative.

With respect to field data, RCRA monitoring wells (Fig. 4) in the vicinity of waste management area (WMA) S-SX show evidence of groundwater contamination from tank waste sources. Downgradient of the tanks to the south and east, technetium-99/(nitrate + nitrite) ratios are elevated with respect to those samples analyzed in upgradient wells to the west and north. With relatively greater technetium-99 inventories present in tank leak fluids compared with neighboring cribs and trenches, the difference in ratios is attributed to contribution from tank waste sources. However, current contributions are limited and have depleted little of the vadose zone source term. For example, the field evidence indicates that the bulk of technetium-99 is relatively high up in the vadose zone (e.g., technetium-99 distribution in borehole 41-09-39 [Fig. 5]).

We hypothesize that current contamination in the unconfined aquifer is a small fraction of the initial inventory and has reached the aquifer because of even more severe, but short-term and more localized, recharge events than those assumed in the modeling. The technetium-99/(nitrate + nitrite) ratios for tanks SX-108 and SX-115 are 0.43 and 0.42 pCi µg–1, respectively. Maximum technetium-99 concentrations and technetium-99/(nitrate + nitrite) ratios (e.g., 188,000 pCi L–1 and 0.11 pCi µg–1) in well 299-W23-19 just southwest of tank SX-115 (Fig. 4) are anomalous. These data are assumed to be a small fraction of total inventory in a known leak from tank SX-115 that was driven to the unconfined aquifer by a subsequent undetected leak in a water line located at the southwestern edge of the SX tank farm. Waterline leaks were not part of modeling in this study. While the magnitude and duration of the waterline leak cannot be quantified, local surface vegetation that persisted for several years, as well as uncharacteristically high moisture content in vadose zone sediments sampled from well 299-W23-19 (Fig. 4), suggests several years of leakage.

Similar conditions are not indicated near tank SX-108, where comparable maximum technetium-99 concentrations and technetium-99/(nitrate + nitrite) ratios are lower (e.g., about 8000 pCi L–1 and 0.006 pCi µg–1) at nearby downgradient well 299-W22-39 (Fig. 4). There is little evidence to suggest that the bulk of the contamination from the SX-108 tank leak has yet reached groundwater.


   Conclusions
TOP
 ABSTRACT
 INTRODUCTION
 Modeling Approach
 Numerical Simulation Results
 Conclusions
REFERENCES
 
The major observations and remarks, based on the numerical modeling, are as follows.


   ACKNOWLEDGMENTS
 
The work reported was performed for the U.S. Department of Energy Office of River Protection under Contract DE-AC06-99RL14047 with CH2M HILL Hanford Group, Inc. Suggestions provided by Glendon Gee (Pacific Northwest National Laboratory), Todd Rasmussen (University of Georgia), and an anonymous reviewer are greatly appreciated. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof or its contractors or subcontractors. The views and opinions of the authors do not necessarily state or reflect those of the United States Government or any agency thereof.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 Modeling Approach
 Numerical Simulation Results
 Conclusions
 REFERENCES
 




This article has been cited by other articles:


Home page
 Vadose Zone JHome page
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]

Home page
 Vadose Zone JHome page
J. M. Zachara, J. Serne, M. Freshley, F. Mann, F. Anderson, M. Wood, T. Jones, and D. Myers
Geochemical Processes Controlling Migration of Tank Wastes in Hanford's Vadose Zone
Vadose Zone J., November 20, 2007; 6(4): 985 - 1003.
[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 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 Web of Science (2)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Khaleel, R.
Right arrow Articles by Kristofzski, J. G.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Khaleel, R.
Right arrow Articles by Kristofzski, J. G.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Khaleel, R.
Right arrow Articles by Kristofzski, J. G.
Related Collections
Right arrow Effective Parameters
Right arrow Variably Saturated Fluid Flow
Right arrow Infiltration

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