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 6:971-984 (2007)
DOI: 10.2136/vzj2006.0166
© 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 (1)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Oostrom, M.
Right arrow Articles by Rohay, V. J.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Oostrom, M.
Right arrow Articles by Rohay, V. J.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Oostrom, M.
Right arrow Articles by Rohay, V. J.
Related Collections
Right arrow Flow
Right arrow Multi-Fluid Systems
Right arrow Multiphase Fluid Flow

SPECIAL SECTION: HANFORD SITE

Carbon Tetrachloride Flow and Transport in the Subsurface of the 216-Z-9 Trench at the Hanford Site

M. Oostroma,*, M. L. Rockholda, P. D. Thornea, M. J. Truexa, G. V. Lasta and V. J. Rohayb

a Environmental Technology Division, Pacific Northwest National Lab., P.O. Box 999, MS K9-33, Richland, WA 99352
b Fluor Hanford, Inc., Richland, WA 99352
* Corresponding author (mart.oostrom{at}pnl.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 9 November 2006.



   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 Materials and Methods
 Results and Discussion
 Conclusions and Conceptual Model...
 REFERENCES
 
As a result of past practices, up to 580 m3 carbon tetrachloride (CT) was discharged to waste sites at the 200 West Area of the USDOE's Hanford Site near Richland, WA. Three-dimensional modeling was conducted to enhance the current conceptual model of CT distribution beneath the major disposal site (216-Z-9). The simulations, using the STOMP code, focused on migration of dense nonaqueous phase liquid (DNAPL) consisting of CT and codisposed organics under scenarios with differing sediment properties, sediment distribution, waste properties, and waste disposal history. Simulation results support a conceptual model for CT distribution where CT in the DNAPL phase migrated primarily in a vertical direction below the disposal site and where some CT DNAPL likely migrated across the water table into the regional aquifer. Results also show that the lower permeability Cold Creek unit retained more CT DNAPL within the vadose zone than other hydrologic units during the infiltration and redistribution process. Due to the relatively high vapor pressure of the CT, the resulting vapor plumes are extensive and influenced by density-driven advection. Any continued migration of CT from the vadose zone to the groundwater is likely through interaction of vapor phase CT with the groundwater and not through continued DNAPL migration. Additional simulations assessed the impacts of soil vapor extraction (SVE) as a remediation method. These simulations showed rapid CT removal associated with the assumed local equilibrium of CT between the phases. Additional efforts are needed to enhance the understanding of rate-limited volatilization to improve simulation of the SVE process and to provide a basis for refining the design and operation of SVE systems.

Abbreviations: CT, carbon tetrachloride • DNAPL, dense nonaqueous phase liquid • EV, EarthVision • NAPL, nonaqueous phase liquid • SVE, soil vapor extraction


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
Materials and Methods
 Results and Discussion
 Conclusions and Conceptual Model...
 REFERENCES
 
Plutonium recovery operations within the 200 West Area near the Plutonium Finishing Plant of the USDOE's Hanford Site in Richland, WA, resulted in organic and aqueous wastes that were disposed of at several cribs, tile fields, and French drains. The organic wastes consisted of CT mixed with lard oil, tributyl phosphate, and dibutyl butyl phosphonate. The main disposal areas were the 216-Z-9 trench, the 216-Z-1A tile field, and the 216-Z-18 crib. These three major disposal facilities received a total of about 13,400,000 L of liquid waste containing 363,000 to 580,000 L of CT. The disposal site locations are shown in Fig. 1 , with the 216-Z-9 site situated in the middle of the site-specific model domain and the 216-Z-1A and 216-Z-18 sites located to the southwest of the site-specific domain. The Plutonium Finishing Plant is directly to the west of the 216-Z-9 disposal site. Assuming a maximum aqueous CT solubility of 800 mg L–1 and an organic liquid density of 1.59 g cm–3 (Schwille, 1988), the 13,400,000 L of liquid waste would be able to contain approximately 6700 L of CT in dissolved form, indicating that the majority of the CT entered the subsurface as an organic liquid.


Figure 1
View larger version (69K):
[in this window]
[in a new window]

  		FIG. 1. Outline of regional and 216-Z-9 trench geologic model domains at the Hanford Site near Richland, WA.

 
In recent years, two major remediation technologies have been applied to remove CT from the vadose zone and groundwater at Hanford. Beginning in 1991, about 78,000 kg of CT was removed using an SVE system in the vadose zone (Fluor Hanford, 2006). In addition, a pump-and-treat system for the unconfined aquifer removed 9700 kg of CT from groundwater beginning in 1994 (USDOE, 2006).

Several conceptual models and modified versions have been proposed to explain the behavior of CT mixtures in the subsurface of the Hanford 200 West Area (e.g., Rohay et al., 1994; Swanson et al., 1999; USDOE, 2004). Figure 2 shows the conceptual model recently outlined in USDOE (2004), depicting the basic lithology and potential flow and transport pathways. This conceptual model considered DNAPL, aqueous phase liquid, and vapor phase CT from multiple source areas but was developed with only minor input from multifluid flow simulations. Except for two-dimensional simulations by Piepho (1996), a numerical assessment of subsurface CT behavior had not been previously conducted.


Figure 2
View larger version (53K):
[in this window]
[in a new window]

  		FIG. 2. Current conceptual model of subsurface carbon tetrachloride behavior at the Hanford Site (after USDOE, 2004).

 
Because of the relatively low permeability, the Cold Creek unit is thought to cause lateral spreading of both DNAPL and the aqueous phase (Fig. 2). The question marks in the figure denote the uncertainties related to the model, including DNAPL movement to groundwater, the extent of lateral DNAPL movement in the vadose zone, the vertical extent of the dissolved CT plume in the groundwater, and the effects of potential undocumented discharges. The numerical modeling presented in this paper primarily deals with infiltration, redistribution, and removal of DNAPL disposed of at the 216-Z-9 "organic rich" crib and the behavior of CT vapors (Fig. 2, items 2 and 4) in the vadose zone. Carbon tetrachloride vapors may be considerably denser than the ambient soil gas, potentially causing rapid downward and lateral movement through density-driven advection (Mendoza and McAlary, 1990). Simulations addressing undocumented discharges (Fig. 2, item 3) have been reported by Oostrom et al. (2006), who concluded that the likelihood of DNAPL reaching the groundwater from such a source is negligible. Future work will examine the effect of aqueous phase disposal from other sites on CT subsurface behavior, mass transfer of CT to the gas phase from the groundwater, and CT flow and transport in the groundwater (Fig. 2, items 1, 5, and 6, respectively).

The main objective of the current study is to develop an updated conceptual model for CT behavior under the 216-Z-9 site, based on a series of multifluid flow simulations, and compare the updated conceptual model with the existing model described by the USDOE (2004), as shown in Fig. 2. The numerical simulations focus on the 216-Z-9 site because of the three major DNAPL disposal sites, the 216-Z-9 has the smallest footprint and has received the most DNAPL waste. Because of these characteristics, it is expected that DNAPL disposed at the 216-Z-9 trench may have moved deeper into the subsurface compared to the other Hanford CT disposal sites. A series of three-dimensional multifluid flow simulations was conducted using the STOMP code (White and Oostrom, 2006), including a base case simulation and 28 sensitivity analysis simulations, to examine the impact of parameter variation on the simulated migration of CT in the subsurface beneath the 216-Z-9 disposal area over the period from 1954 to 2005. An additional objective was to investigate the impact of SVE, implemented at the site in 1993, on CT distributions and fluxes.


   Materials and Methods
TOP
 ABSTRACT
 INTRODUCTION
 Materials and Methods
Results and Discussion
 Conclusions and Conceptual Model...
 REFERENCES
 
The numerical model was developed using available information regarding the hydrogeology of the modeled area, measured fluid properties for the likely mixtures of disposed aqueous phase and organic liquids, and estimates of hydraulic properties and hydrologic boundary conditions. The hydrogeologic setting was established by assembling a geologic model based on interpretations of borehole geologic information using EarthVision (EV) software (Dynamic Graphics, Inc., Alameda, CA). The EV geologic model enabled porous media properties to be readily mapped to the numerical model grid. Fluid properties were determined in the laboratory as part of the USDOE's Remediation and Closure Science Project (Oostrom et al., 2004). Additional work under this program provided the basis for an updated theory of nonaqueous phase liquid (NAPL) residual saturation formation that was incorporated into the model for selected simulations (Lenhard et al., 2004). Hydraulic properties were obtained from Khaleel and Freeman (1995) and Khaleel et al. (2001). Appropriate ranges for organic liquid and water disposal conditions for the local model were established based on a thorough review of historical information (Anderson, 1976).

Geologic Model
Development of a geologic model for the subsurface of the 216-Z-9 trench was completed in two stages. First, a regional-scale geologic model was developed, followed by a detailed local-scale geologic model (Fig. 1). The regional geologic model domain included important liquid disposal areas such as the 216-U-14 ditch, located to the east of the 216-Z-9, the 216-U-10 pond to the south, the 200-ZP-1 injection wells to the west, and the old 216-T-4 pond to the north. Geologic data considered information from all boreholes (groundwater wells, dry wells, and soil borings). Initially, 215 boreholes were identified within the regional model, while 21 boreholes in the immediate vicinity of the 216-Z-9 trench were incorporated into the site-specific geologic model. The base geologic framework was simplified into a layered sequence of five main stratigraphic units, listed here from oldest to youngest:

  1. Saddle Mountains Formation. The Saddle Mountains Formation of the Columbia River Basalt Group constitutes the bedrock beneath the site. With the assumption that no fractures are present, this formation essentially acts as a no-flow boundary at the base of the unconfined aquifer.
  2. Ringold Formation. The basalt bedrock is overlain by the Ringold Formation, a sedimentary sequence of fluvial-lacustrine clay, silt, sand, and granule to cobble gravel deposited by the ancestral Columbia River and its tributaries. Locally, the formation has been subdivided into three units: Ringold A, Lower Mud, and Ringold E.
  3. Cold Creek. Overlying the Ringold Formation is the Cold Creek unit. Locally, this unit is differentiated into Cold Creek unit carbonate and the Cold Creek unit silt. The Cold Creek unit carbonate is a calcium-carbonate cemented paleosol that developed on top of the Ringold Formation.
  4. Hanford. The Hanford formation, consisting of Pleistocene flood deposits, has been locally subdivided into five main units. From oldest to youngest, these are lower fine unit (Hanford 4), lower coarse unit (Hanford 3), middle fine unit (Hanford 2), upper coarse unit (Hanford 1), and upper fine unit (Hanford 1a).
  5. Backfill. The 216-Z-9 trench was excavated, and stockpiles of these sediments have been used as backfill in pipeline trenches and other excavations in the vicinity of the 216-Z-9 trench.

To enhance the level of detail, the 5 main hydrostratigraphic units were subsequently divided into 13 units for the site-specific model domain (Fig. 1). The units are, from top to bottom, Backfill, Hanford 1a, Hanford 1, Hanford 2, Lower Gravel (Hanford 3), Lower Sand (Hanford 4), Cold Creek unit silt, Cold Creek unit carbonate, Upper Ringold, Ringold E, Ringold Lower Mud, Ringold A, and Basalt. The EV software was used to create a three-dimensional model of the geologic units. The resulting three-dimensional geologic model is shown in Fig. 3 . The figure shows that the EV interpretation yields a layered system with minor undulations. Details of the geologic model development can be found in Oostrom et al. (2004, 2006).


Figure 3
View larger version (38K):
[in this window]
[in a new window]

  		FIG. 3. Three-dimensional geologic domain with a cut-out below the 216-Z-9 trench. The stratigraphic units in the legend denote Backfill, Hanford 1a (h1a), Hanford 1 (h1), Hanford 2 (h2), Lower Gravel (low_grvl), Lower Sand (low_sand), Cold Creek unit silt (pplz), Cold Creek unit carbonate (pplc), Upper Ringold (up_ring), Ringold E (ring_e), Ringold Lower Mud (low_rmud), Ringold A (ring_a), and Basalt.

 
Numerical Model
The water–oil–air operational mode (STOMP-WOA) of the STOMP simulator (White and Oostrom, 2006) was used to simulate multifluid flow and transport beneath the 216-Z-9 trench. The fully implicit, integrated finite difference code has been used to simulate a variety of multifluid systems (e.g., Oostrom and Lenhard, 2003; White et al., 2004; Oostrom et al., 2005). The applicable governing equations are the component mass-conservation equations for water, organic compounds, and air. The governing partial differential equations are discretized following the integrated-volume finite difference method by integrating over a control volume. Using Euler backward time differencing, yielding a fully implicit scheme, a series of nonlinear algebraic expressions is derived. The algebraic forms of the nonlinear governing equations are solved with a multivariable, residual-based Newton–Raphson iterative technique, where the Jacobian coefficient matrix is composed of the partial derivatives of the governing equations with respect to the primary variables.

The algebraic expressions are evaluated using upwind interfacial averaging for fluid density, mass fractions, and relative permeability. Specified weights (arithmetic, harmonic, geometric, upwind) are applied to the remaining terms of the flux equations. For the simulations described in this paper, harmonic averages were used. The maximum number of Newton–Raphson iterations was 16, with a convergence factor of 10–6. Secondary variables, those parameters not directly computed from the solution of the governing equations, are computed from the primary variable set through constitutive relations. The relative permeability–saturation–capillary pressure (k–S–P) relations consist of the Brooks and Corey (1964) S–P relations in combination with the k–S relations derived from the Burdine (1953) or Mualem (1976) model. In these relations, the effects of fluid entrapment and residual saturation formation have been included.

The computational domain, with lengths in the west–east, south–north, and top–bottom directions of 440, 540, and 159 m, respectively, was discretized into 27 x 34 x 85 = 78,030 grid blocks. The model was refined near the disposal site and the low-permeability Cold Creek unit. The footprint of the 216-Z-9, with approximate dimensions of 9.1 x 18.3 m, was located near the center at the top of the site-specific domain (Fig. 1). The Basalt unit, at an elevation of 42 m, was assumed to form an impermeable bottom boundary. Because the STOMP-WOA (water–oil–air) mode was used, the number of nodes translates into 3 x 78,030 = 234,090 unknowns.

At the top boundary, atmospheric gas pressure was assumed in conjunction with a 0.5 cm yr–1 water flux (Rockhold et al., 1995. For the south, north, west, and east boundaries, fluctuating water table boundary conditions were imposed for the water mass balance equation below the water table, and zero-flux boundary conditions were applied above the water table. The time variant boundary conditions for the water mass balance equation at the south and north boundary were derived from observed water levels in wells 299-W15-5 and 6 (Oostrom et al., 2004). The resulting boundary conditions yielded a groundwater flow direction from south to north. Neumann boundary conditions were imposed for water and DNAPL discharges at the 216-Z-9 trench area during the years when these liquids were disposed (details on the discharge rates and volumes are presented in the next section). Dense nonaqueous phase liquid was allowed to move freely across all boundaries. The initial gas and aqueous phase pressure distributions in the domain at 1954 were obtained by conducting a 10,000-yr simulation using the interpolated 1954 water levels at the south and north boundary and a recharge rate of 0.5 cm yr–1. It was assumed that in 1954 no DNAPL was present in the domain.

Soil vapor extraction from wells in the vicinity of the 216-Z-9 was included in the simulations from 1993 to 2005. The details of the field SVE campaign through 2001 were described by Rohay (2002). The SVE operation information from 2002 to 2005 was obtained from V.J. Rohay (personal communication). During the active SVE campaign, soil vapor was extracted simultaneously from multiple wells open above and/or below the Cold Creek unit. A top view of the well locations within the STOMP computational domain is shown in Fig. 4 . Overall, 23 well intervals are available for SVE, with 14 of these open above and 9 open below the Cold Creek unit. Active SVE operations began at the 216-Z-9 well field in March 1993, and the system began operating 24 h d–1 and 7 d wk–1 on 19 Oct. 1994 (Rohay 2002). Not all of the 23 intervals in the 216-Z-9 well network were connected to the SVE system during all periods of operation, and some of the wells were also modified during the period of SVE operations. The data from Rohay (2002) were used in conjunction with information on the current screened intervals for each well in the 216-Z-9 well field to generate time-averaged gas flow rates that were applied as sink terms to represent SVE in the STOMP model. Several modifications were made to STOMP to allow it to simulate the process of SVE more accurately. These modifications included the addition of a fully coupled SVE well model to the code (White and Oostrom, 2006) and incorporation of a model for a gas-slip phenomenon known as the Klinkenberg effect (Klinkenberg, 1941).


Figure 4
View larger version (32K):
[in this window]
[in a new window]

  		FIG. 4. Plan view of the STOMP computational domain with model grid and locations of soil vapor extraction wells. The domain is 440 m in the west–east direction and 540 m in the south–north direction. Well 299-W15-223 is a slant-borehole.

 
Base Case Simulation
Dense nonaqueous phase liquid fluid properties were measured in the laboratory based on an assumed average fluid composition of 8.8% (v/v) tributyl phosphate, 14.7% dibutyl butyl phosphonate, 2.9% lard oil, and 73.6% CT (Last and Rohay, 1993). The obtained values are listed in Table 1. Retention parameters, porosities, and hydraulic conductivities for the units were obtained from Khaleel et al. (2001) and Khaleel and Freeman (1995) and are listed in Table 2. A permeability anisotropy ratio of 10:1 was used. The parameters for the relative permeability–saturation–pressure van Genuchten (1980)–Mualem (1976) models were converted to equivalent parameters for the Brooks-Corey (Brooks and Corey, 1964)–Burdine (1953) models, using the method described by Lenhard et al. (1989). A linear sorption model with a partitioning coefficient, Kd, of 0.2 mL g–1 (Rohay et al., 1994) was applied to all porous media.


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

  		TABLE 1. Laboratory-determined DNAPL properties.

 

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

  		TABLE 2. Saturated hydraulic conductivity (Ks), porosity (n), and retention parameter values (Brooks-Corey [1964] air-entry pressure head hd, pore geometry factor {lambda}, and irreducible water saturation, srl) of hydrostratigraphic units.

 
Sensitivity Analysis Simulations
Twenty-eight sensitivity analysis simulations were conducted. The parameters that were changed from the base case for each sensitivity simulation are described below.
  1. Increased DNAPL volume to 407 m3, according to the maximum disposal volume estimate of Last and Rohay (1993). The volume is distributed proportionally to the base case distribution (Table 3).
  2. Decreased DNAPL volume to 113 m3 according to the minimum disposal volume estimate of Last and Rohay (1993). The total volume is distributed over the years proportionally to the base case distribution (Table 3).
  3. The fluid properties of disposed DNAPL are equal to properties of pure CT. Density = 1594 kg m–3; viscosity = 9.7 x 10–4 Pa s; vapor pressure = 11950 Pa; surface tension (air–DNAPL) = 26.2 dyne cm–1; interfacial tension (water–DNAPL) = 40.8 dyne cm–1.
  4. Properties of DNAPL reflecting minimal CT content in DNAPL (50%). Fluid properties: density = 1260 kg m–3; viscosity = 1.357 x 10–3 Pa s; vapor pressure = 8250 Pa; surface tension (air–DNAPL) = 24.2 dyne cm–1; interfacial tension (water–DNAPL) = 11.8 dyne cm–1.
  5. Dense nonaqueous phase liquid infiltration into subsurface through an area 20% of 216-Z-9 footprint.
  6. Dense nonaqueous phase liquid infiltration rate increased by a factor of 4 for 1 wk mo–1, with no infiltration for the rest of each month.
  7. Combination of simulations 5 and 6: DNAPL infiltration through an area 20% of 216-Z-9 footprint and DNAPL infiltration rate increased by a factor of 4 for 1 wk mo–1, with no infiltration for the rest of each month.
  8. Use of equivalent van Genuchten (1980)–Mualem (1976) parameter values.
  9. Increased the base case air-entry pressure head (Brooks and Corey [1964] hd) by a factor of 2.
  10. Decreased the base case air-entry pressure head (Brooks and Corey [1964] hd) by a factor of 2.
  11. Increased the base case pore geometry parameter (Brooks and Corey [1964] {lambda}) value by a factor of 2.
  12. Decreased the base case pore geometry parameter (Brooks and Corey [1964] {lambda}) value by a factor of 2.
  13. Increased the base case saturated hydraulic conductivity by a factor of 10.
  14. Decreased the base case saturated hydraulic conductivity by a factor of 10.
  15. An anisotropy ratio of 1:1.
  16. An anisotropy ratio of 20:1.
  17. Laboratory-measured maximum residual DNAPL saturation for the Cold Creek unit silt (0.13), Hanford sand (0.10), Lower Gravel (0.05), and Ringold E material (0.11). For the other materials, a maximum residual of 0.1 was applied.
  18. Measured and assumed maximum residual DNAPL saturation in simulation 17 multiplied by 1.25.
  19. The base case porosity multiplied by a factor of 1.25.
  20. The base case porosity multiplied by a factor of 0.75.
  21. A linear Kd partitioning coefficient of 0.1 mL g–1.
  22. A linear Kd partitioning coefficient of 0.4 mL g–1.
  23. Reduced the hydraulic conductivity of the Cold Creek unit carbonate by a factor of 10.
  24. Reduced the hydraulic conductivity of the Cold Creek unit carbonate by a factor of 100.
  25. Reduced the hydraulic conductivity of only the Cold Creek unit carbonate by a factor of 10 and increased the air-entry pressure head by a factor of {surd}10. The reduction of the hydraulic conductivity and the increase in entry pressure head are consistent with the Miller and Miller (1956) scaling theory.
  26. Reduced the hydraulic conductivity of only the Cold Creek unit carbonate by a factor of 100 and increased the air-entry pressure head by a factor of {surd}100 = 10. The reduction of the hydraulic conductivity and the increase in entry pressure head are consistent with the Miller and Miller (1956) scaling theory.
  27. Divided the vapor pressure and the aqueous solubility of the DNAPL mixture by a factor of 4.
  28. Divided the aqueous phase solubility of the DNAPL mixture by a factor of 4.


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

  		TABLE 3. Dense nonaqueous phase liquid (DNAPL) and aqueous phase disposal rates and volumes for the 216-Z-9 disposal site. The infiltration area (footprint) of the site is 167.2 m2.
 
Moment Analysis
Normalized spatial moments of the simulated DNAPL saturations in 1993 were calculated to provide a quantitative basis for comparing the results of the different simulation cases, before start-up of SVE operations (Freyberg 1986). The ijkth moment of the NAPL mass distribution in space was defined as

Formula 1[1]
where {rho}n is the mass density of the NAPL, {phi} is the porosity, Sn is the NAPL saturation, and x, y, and z are the spatial coordinates. The integrals in Eq. [1] were evaluated over the extent of the three-dimensional model domain.

The zeroth, first, and second (i + j + k = 0, 1, or 2, respectively) spatial moments of the NAPL plumes were computed. These moments provide measures of the total NAPL mass, the location of the center of mass, and spread about the center of mass, respectively. The zeroth moment, M000, is equal to the total mass in the domain. The first moment, normalized by the zeroth moment, defines the location of the center of mass (xc, yc, zc):

Formula 2[2]
The second moment about the center of mass defines a spatial covariance tensor (Freyberg 1986):

Formula 2


   Results and Discussion
TOP
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results and Discussion
Conclusions and Conceptual Model...
 REFERENCES
 
Base Case
A total volume of 315 m3 of DNAPL infiltrated at the 216-Z-9 site from July 1955 through June 1962, together with almost 3800 m3 of aqueous phase waste liquid. For this simulation, the first kilogram of DNAPL arrived at the water table in 1964, corresponding to a travel time of approximately 9 yr (Table 4). By 1993, before the start of SVE operations near the disposal site, almost 27,000 kg of CT DNAPL had moved into the saturated zone and 43% of the original DNAPL inventory was still present in the vadose zone as a DNAPL (Table 4). Dense nonaqueous phase liquid saturations in 1993 (Fig. 5 ) show the highest saturations in the Cold Creek unit, while most DNAPL had drained from the Hanford 1 unit. Due to capillary action, DNAPL had spread the most in the vadose zone part of the Ringold E. Below the water table, the DNAPL body thinned out with depth due to aqueous phase dissolution. No DNAPL reaches the bottom of the computational domain.


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

  		TABLE 4. Total dense nonaqueous phase liquid (DNAPL) mass inventory, percentage in vadose zone for year 1993, time for DNAPL to reach water table, and DNAPL mass that has moved across the water table for year 1993.

 

Figure 5
View larger version (40K):
[in this window]
[in a new window]

  		FIG. 5. Dense nonaqueous phase liquid saturations in 1993 for the base case simulation. The stratigraphic units in the legend denote Backfill, Hanford 1a (h1a), Hanford 1 (h1), Hanford 2 (h2), Lower Gravel (low_grvl), Lower Sand (low_sand), Cold Creek unit silt (pplz), Cold Creek unit carbonate (pplc), Upper Ringold (up_ring), Ringold E (ring_e), Ringold Lower Mud (low_rmud), Ringold A (ring_a), and Basalt.

 
The CT vapor phase plume in 1993 is shown in Fig. 6 . The plume is extensive because of enhanced gas phase advection caused by the density difference between gas containing CT (~1.9 kg m–3 for air saturated with CT at a vapor pressure of 10,820 Pa) and ambient soil gas (~1.2 kg m–3 at room temperature). As shown in Fig. 6, the shape of the CT vapor plume is greatly influenced by the Cold Creek unit and the saturated zone. The density differences cause the vapors containing CT to move downward relatively fast, while the lower permeability Cold Creek unit and the saturated zone force the vapors to move laterally because of the reduced gas permeability associated with higher water contents. During this relatively rapid spreading process, CT partitions into the upper regions of the groundwater through phase partitioning. The fast lateral spreading of the CT vapors and the subsequent contamination of the groundwater could potentially affect large volumes of water. The resulting CT vapor plume in the Ringold E above the water table is approximately 150 m in all directions. It is important to note that this phenomenon can only be investigated through a simulator recognizing an mobile gas phase, including density-driven advection. Three-phase simulators using a passive gas phase or ignoring density-driven advection are not appropriate to simulate behavior of a volatile DNAPL such as the organic liquids disposed of at the Hanford Site.


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

  		FIG. 6. Carbon tetrachloride gas concentrations (g L–1) in 1993 for the base case simulation. The stratigraphic units in the legend denote Backfill, Hanford 1a (h1a), Hanford 1 (h1), Hanford 2 (h2), Lower Gravel (low_grvl), Lower Sand (low_sand), Cold Creek unit silt (pplz), Cold Creek unit carbonate (pplc), Upper Ringold (up_ring), Ringold E (ring_e), Ringold Lower Mud (low_rmud), Ringold A (ring_a), and Basalt.

 
Figure 7 shows the CT mass distribution over the phases and sorbed to the solid phase through 2005. Until the SVE was activated in 1993, the total CT mass in the system remained fairly constant. The simulation indicates that from the end of the infiltration period in 1962 until 1993, about 180,000 kg of CT transferred from the DNAPL to the aqueous and gas phases and to sorbed CT. The sorbed mass grew especially rapidly, from approximately 40,000 kg in 1962 to 170,000 kg in 1993. The mass distributions are heavily influenced by volatilization, subsequent vapor transport, and the equilibrium assumption. The influence of the SVE process is clearly visible by the rapidly decreasing total mass and the mass in all phases after 1993. Within a few years of operation, the simulation predicted the removal of almost 400,000 kg from the system by SVE, which is considerably more than the 53,000 kg that had actually been removed from SVE wells near the 216-Z-9 site (Fluor Hanford, 2006). There are several possible reasons for the discrepancy between field observations and simulation results. The differences could be the result of uncertainties in flow rates, fluid-media properties, and disposal history (e.g., volumes, rates, infiltration, and timing). In addition, the differences may also be the result of the equilibrium-phase partitioning used in the STOMP simulations. Currently, STOMP does not account for rate-limited (kinetic) interfacial mass transfer effects.


Figure 7
View larger version (21K):
[in this window]
[in a new window]

  		FIG. 7. Carbon tetrachloride mass (kg) distribution in the dense nonaqueous phase liquid (DNAPL), aqueous, and gaseous phases, and sorbed to the solid phase for the base case simulation. (VOC, volatile organic compound.)

 
Figure 8 shows the DNAPL mass distribution over the hydrostratigraphic units used in the simulations. The lines show that by 1993, most DNAPL had disappeared from both the Hanford 1 and Hanford 2 units due to drainage, volatilization, and dissolution. Before the SVE was started, the DNAPL was primarily located in the Lower Sand (Hanford 4) unit, the Cold Creek unit, and in the Ringold E. The DNAPL mass in Cold Creek unit silt and carbonate is almost constant from 1962 until 1993, indicating that these units are able to retain DNAPL better than the other units. The DNAPL in the Ringold E is situated both in the vadose and saturated zones. After 1993 the DNAPL in all units decreased rapidly due to SVE operations. Only the Cold Creek unit carbonate and the Ringold E units showed some resistance to DNAPL removal by SVE. The Cold Creek unit carbonate was able to retain the DNAPL longer due to its higher water contents and lower gas phase permeability, relative to the other units in the vadose zone. The SVE process removed DNAPL from the unsaturated Ringold E rapidly within the first 2 yr of operation from 1993 to 1995 but had no influence on the DNAPL below the water table. The slower decrease from about 1995 onward for this unit is directly associated with aqueous phase dissolution.


Figure 8
View larger version (25K):
[in this window]
[in a new window]

  		FIG. 8. Dense nonaqueous phase liquid (DNAPL) mass (kg) distribution among the hydrostratigraphic units for the base case simulation.

 
Cumulative DNAPL and other fluxes are depicted in Fig. 9 . The figure shows that the difference between the DNAPL flux into and out the Cold Creek unit is nearly constant for the simulation period, indicating a constant holding capacity of about 120,000 kg. The volatilized mass that had moved out of the domain was heavily affected by the SVE process. Before 1993 about 25,000 kg disappeared from the domain through the side boundaries. After 1993 CT vapor mass was removed through SVE.


Figure 9
View larger version (26K):
[in this window]
[in a new window]

  		FIG. 9. Cumulative transported dense nonaqueous phase liquid (DNAPL) and dissolved mass for the base case simulation.

 
Table 5 shows the zeroth and first order moments of the DNAPL plumes. Table 6 presents an overview of the standard deviations. The moments are computed for the year 1993. The moment analysis for the base case simulation shows that the center of mass for DNAPL in the subsurface in 1993 had only shifted 0.29 m in the x direction and 0.23 m in the y direction from the centerline of the 216-Z-9 trench (Table 5). The center of mass in the vertical direction is located in the Cold Creek unit silt, although Fig. 8 shows that the Lower Sand (Hanford 4) and Ringold E units contain more mass than the Cold Creek unit at that point in time. The second order moment analysis shows that 95% of the mass was expected to be within 11.6, 13.6, and 22.8 m of the center of mass in the x, y, and z directions, respectively (Table 6). These distances are twice as high as the standard deviations reported in Table 6. This analysis suggests that most of the liquid DNAPL in the vadose zone is located directly beneath the footprint of the disposal facility, with the majority of the mass located within the Lower Sand (Hanford 4) and Cold Creek units.


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

  		TABLE 5. Zeroth order moment, M000, and center of DNAPL mass (subscript c) in the x, y, and z direction for the year 1993. (The center of 216-Z-9 trench is at x = 0 m, y = 0 m, and z = 195 m. The Cold Creek unit below the trench is located approximately between z = 158 m and z = 169 m.)

 

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

  		TABLE 6. Standard deviations of the center of dense nonaqueous phase liquid mass in the x, y, and z directions for the year 1993.

 
Sensitivity Analysis Cases
The main results of the sensitivity simulations are summarized in Tables 4, 5, and 6. Table 4 presents a comparison of some pertinent DNAPL statistics for the vadose zone in 1993, before SVE initiation at the site, including the percentage of total inventory present in the vadose zone, the time it took for the first kilogram of DNAPL to reach the water table after the DNAPL disposal ceased, and the total DNAPL mass that had moved across the water table by 1993. Table 5 shows the zeroth and first order moments, and Table 6 shows the standard deviations. A detailed discussed of each sensitivity simulation, including figures similar to Fig. 5–9GoGoGoGo, can be found in Oostrom et al. (2004, 2006). In this section, only the main results will be presented.

In all cases except for simulations 2, 12, and 16, substantial DNAPL volumes reached the water table by 1993. Simulation 2 had a DNAPL inventory that was only 28% of the base case inventory. Somewhat surprising was the effect of a decrease in the Brooks-Corey (Brooks and Corey, 1964) model pore geometry factor {lambda} in simulation 12. Apparently, this reduction causes considerable additional DNAPL spreading in the vadose zone, preventing DNAPL from penetrating into the water table. For simulation 16, the increased porosity resulted in more storage of DNAPL in the vadose zone and less downward migration. Interestingly, none of the simulations with a reduction in permeability (simulations 14, 23, 24, 25, and 26) resulted in complete prevention of DNAPL moving across the water table. Even a reduction of the Cold Creek unit carbonate permeability by a factor of 100 and an increase of the entry pressure by a factor of 10 allowed more than 8000 kg CT to penetrate into the saturated zone.

The simulated DNAPL mass in the vadose zone as a percentage of the initial inventory ranged between 19 and 65%. The lowest percentage is associated with simulation 13, in which the base case permeability was increased by a factor of 10. This increase had the largest effect on CT transport in the gas phase, as volatilized CT quickly moved away from the source area in downward and lateral directions. As a result, the DNAPL mass in the domain decreased rapidly, and even less DNAPL moved across the water table compared with the base case. The largest percentage of DNAPL retained in the vadose zone occurred in simulation 14, where the base case permeability decreased by a factor of 10. In this case, removal of volatilized CT was much slower and CT primarily remained in the DNAPL, which moved relatively slowly through the subsurface.

Of the 29 simulations, the median, mean, and standard deviation for the simulated mass of CT in the form of DNAPL that had moved across the water table by 1993 were 21,932 kg, 33,345 kg, and 36,353 kg, respectively. For most of these simulations, DNAPL migration across the water table had been reduced to less than 100 kg yr–1. In contrast, the median, mean, and standard deviation of dissolved CT mass migrating across the water table (data not shown) were only 4345, 5521, and 3101 kg, respectively. The mean simulated value of the combined DNAPL and aqueous CT mass that had moved across the water table is about 39,000 kg. The aqueous CT that moved across the water table combines both the CT that was transported downward through the aqueous phase and the CT that transferred into the aqueous phase from the gas phase near the water table.

Nine of the simulations reported a cumulative DNAPL mass transfer across the water table of more than 60,000 kg. In simulation 2, the increase compared to the base case was due to the larger initial inventory. Of the additional 130,000 kg disposed of at the site, 42,000 kg was predicted to move into the saturated zone. The combination of a smaller footprint and DNAPL disposal for 1 wk mo–1 (simulation 7) yielded the largest DNAPL mass moving across the water table. This simulation produced relatively high DNAPL saturations and associated relative permeability underneath the disposal site, followed by a rapid downward movement. Simulation 8, using the van Genuchten (1980)–Mualem (1976) constitutive relations, also predicted a large CT mass movement into the water table. This simulation, however, is not considered to be as realistic since the van Genuchten (1980) saturation–pressure relationship does not recognize a distinct nonwetting fluid entry head. Simulation 10, using a decrease in the nonwetting entry head, also resulted in considerable transport across the water table, consistent with simulation 8. Another unrealistic result was obtained with simulation 15, in which an anisotropy ratio of 1:1 was used. Based on arguments provided by Zhang et al. (2003), the assumption of isotropic conditions for Hanford vadose zone porous materials is not appropriate. An increase in movement to the water table was also used in simulation 20, where the porosity was reduced by 25%. This reduction led to less storage capacity for the DNAPL, increased DNAPL relative permeability values, and enhanced downward movement. Simulations 27 and 28, with decreased CT partitioning into the gas and aqueous phases, also resulted in relatively large masses of DNAPL moving across the water table by 1993. In these simulations, more CT remained in the DNAPL phase, decreasing the extent of the CT vapor plume and increasing the vertical penetration of the DNAPL.

On the basis of recent groundwater CT concentration data, including the consideration of continual CT degradation by hydrolysis, Murray et al. (2006) estimated that approximately 100,000 kg of CT may have entered into the aquifer. This mass estimate is for the entire Hanford 200 West Area subsurface and includes CT from all three major Hanford Site DNAPL disposal sites (216-Z-1A, 216-Z-18, and 216-Z-9). Assuming that, in terms of CT transport in the aqueous phase, the two other sites yield results similar to the 216-Z-9 site, it can be estimated that about 15,000 kg of dissolved phase CT had transported across the water table. This estimate shows that the accumulated mass in the aquifer would be significantly lower than the mass of CT in the groundwater estimated by Murray et al. (2006) if only aqueous and vapor phase CT and no DNAPL phase entered the groundwater. This assessment suggests that DNAPL CT likely entered the groundwater.

The analysis of first moments shows that the center of mass in the x and y directions for all simulations did not significantly shift from the center of the 216-Z-9 trench (Table 5). This means that in the horizontal plane, the center of mass is located directly underneath the disposal site. The center of mass in the vertical direction for most simulations is located in or slightly above the Cold Creek unit silt. This analysis indicates that most of the liquid DNAPL in the vadose zone is expected to be located underneath the footprint of the disposal facility with the majority of the mass within the Lower Sand and Cold Creek unit. The standard deviations of the DNAPL center of mass, listed in Table 6, indicate that in the horizontal direction 95% of the mass can be found within 12 m in the x-direction and 16 m in the y-direction for all simulations except simulation 8, in which the van Genuchten (1980)–Mualem (1976) constitutive relations were used. The numbers show that the vast majority of the mass is likely to be located directly underneath the footprint of the 216-Z-9. In the vertical direction, 95% of the simulated DNAPL mass is within a distance of about 8 to 60 m, with the largest extents related to simulations where most CT remained in the DNAPL phase.

All sensitivity simulations with the layered model included SVE for the time period 1993 to 2005. The actual SVE process in the field has extracted approximately 53,000 kg of CT. Application of the averaged extraction rates over the periods when the extraction system was active yielded an almost-complete removal of the CT distributed over all the phases for most sensitivity cases. Although removal was more difficult to obtain from the Cold Creek unit, the imposed vapor extraction rates were, in most cases, also sufficient to deplete DNAPL from these layers during the extraction periods. However, for simulations in which the permeability of the Cold Creek unit carbonate was lowered (simulations 14, 24, and 25), CT removal from the Cold Creek unit was significantly reduced. For these cases, not only was the porous medium permeability lowered, but the relative permeability of the gas phase in the Cold Creek unit was reduced due to increased water saturations. The combined effect of reduced permeability in the Cold Creek carbonate and relative gas phase permeability in both the Cold Creek unit silt and carbonate caused the SVE operations to be less effective for such conditions. For the sensitivity cases with the reduced vapor pressure and solubility (simulations 27 and 28), SVE was still able to remove a considerable amount of CT from the more porous materials, while the Cold Creek unit was able to retain the liquid CT located in these layers.

The extracted mass of CT for most simulations was significantly higher than the amount removed in the field. There are several possible reasons for the discrepancy between observed and simulated results, including uncertainties in flow rates, fluid-media properties, and disposal history (e.g., volumes, rates, and timing). The differences may also result from the current simulations being based on equilibrium phase partitioning, meaning that simulations do not account for any rate-limited (kinetic) interfacial mass transfer effects.


   Conclusions and Conceptual Model Update
TOP
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results and Discussion
 Conclusions and Conceptual Model...
REFERENCES
 
Simulations using the multifluid flow simulator STOMP (White and Oostrom, 2006) were conducted to estimate how disposed DNAPL may have migrated in the vadose zone beneath the 216-Z-9 trench as a function of the properties and distribution of subsurface sediments and of the properties and disposal history of the waste. A total of 29 three-dimensional simulations were conducted to examine infiltration and redistribution of CT in the subsurface before and after implementation of SVE. The simulations consisted of one base case simulation and 28 sensitivity analysis simulations.

Based on the presented modeling results, Fig. 2 has been updated, and a revised overall conceptual model is shown in Fig. 10 . The updates primarily focus on items 1 and 2 in the caption of Fig. 10. Item 4 in the caption of Fig. 10 has been addressed by Oostrom et al. (2006), who concluded that DNAPL movement across the water table from undocumented sources is unlikely. The key revisions, as shown in Fig. 10, include the following:

  1. Lateral movement of DNAPL from the disposal site to underneath the Plutonium Finishing Plant (Fig. 1) is unlikely.
  2. The zones of persistent CT mass in the vadose zone are primarily in the Cold Creek unit.
  3. Large vertical and lateral density-driven vapor movement has likely occurred in the past.
  4. Some DNAPL penetration to groundwater is likely to have occurred in the past.
  5. The phase distribution of CT changes over time due to volatilization, interaction of gas-phase CT with pore water and aqueous-phase CT with the solid phase, DNAPL dissolution in groundwater, and the impact of soil vapor extraction.


Figure 10
View larger version (52K):
[in this window]
[in a new window]

  		FIG. 10. Updated carbon tetrachloride subsurface flow and transport conceptual model for the Hanford Site.

 
Figure 10 illustrates how CT in DNAPL and other phases may have been distributed through the subsurface at a given time, but it does not necessarily represent an overview of changes in CT distribution over time. In addition to a static picture of the conceptual model for CT shown in Fig. 10, the modeling results provide information about the variation in CT distribution over time, and the revisions to the previous conceptual model include a temporal component to interpreting CT distribution in the subsurface. A conceptual depiction of temporal variation in the CT distribution over time for the 216-Z-9 site based on the base case simulation is shown in Fig. 11 for the years 1966, 1974, 1993, and 2005. In 1966 DNAPL has moved below the Cold Creek unit but has not yet reached the water table. In 1974 DNAPL is moving into the water table and shows some spreading in the Cold Creek unit. The situation in 1993, before start-up of SVE operations, is similar to that in 1974. In 2005 the SVE operations have greatly reduced the DNAPL in the vadose zone. The remaining DNAPL can be found in the Cold Creek unit and the upper portion of the unconfined aquifer, based on the base case simulation. Figures 11a, b, and c show that the CT vapor plume continued to grow over time, with the ability to contaminate large areas of the aquifer. Mass transfer limitations and slow aqueous phase diffusion have limited the vertical penetration of dissolved CT originating from the CT in the vapor phase. The vapor phase plume in 2005 was greatly reduced due to the ongoing SVE operations.


Figure 11
View larger version (62K):
[in this window]
[in a new window]

  		FIG. 11. Conceptual distribution of carbon tetrachloride from waste disposed at the 216-Z-9 site for (a) 1966, (b) 1974, (c) 1993, and (d) 2005. These figures are based on the results of the base case simulations. Note that some sensitivity simulations show significantly different results. The figure for the year 2005 shows the conceptual impact of soil vapor extraction remediation operations.

 
The results and conclusions from these simulations of CT distribution through the vadose zone and into the groundwater provide useful information to assess potential vadose zone remediation approaches. Estimates for the phase distribution of CT in the vadose zone and the type of continuing CT flux to the groundwater, along with a comparison to the site characterization data, can be used to establish appropriate targets for vadose zone remediation and assessment of appropriate technologies for these target volumes and the physical state of CT within these target volumes. As shown in the simulation results for SVE, however, additional work is needed to improve simulation of how SVE impacts CT within the vadose zone. Of particular importance may be an improved understanding of how the CT retained within the Cold Creek unit volatilizes, migrates in the vapor phase over time, and is impacted by SVE. Coupled vadose zone and groundwater simulations are also a potentially important component for implementing remedies for the vadose zone and groundwater to assess the overall fate of CT and the impact of the CT flux from the vadose zone to the groundwater. For instance, if the primary flux of CT from the vadose zone to the groundwater is via vapor phase interaction at the water table, it may be important to quantify this flux relative to the attenuation capacity of the aquifer and the capacity and longevity of any groundwater remediation technologies.

Based on the simulation results outlined in this paper, the following recommendations are made for future research:

The modeling has been conducted without the consideration of small-scale heterogeneities, which would be expected in the Hanford and Ringold formations. These heterogeneities may enhance the retention of CT in the vadose zone and may increase the relative mass of CT in the vapor, in the vadose zone pore water, and adsorbed to vadose zone sediments compared to mass in the DNAPL phase. Models are needed that describe field subsurface heterogeneities through upscaling and parameterization of porous media using effective parameters for multiphase systems. Parameterization techniques such as pedotransfer function and similar media scaling methods, as explored by Oostrom et al. (2006), need to be further developed for three-phase systems.


   ACKNOWLEDGMENTS
 
Simulations reported in this paper were conducted on the MPP2 supercomputer at the Molecular Science Computing Facility (MSCF) in the William R. Wiley Environmental Molecular Sciences Laboratory (EMSL), a national scientific user facility sponsored by the U.S. Department of Energy's Office of Biological and Environmental Research and located at the Pacific Northwest National Laboratory (PNNL). These simulations were performed under a Computational Grand Challenge Application project, "Multifluid Flow and Multicomponent Reactive Transport in Heterogeneous Subsurface Systems." The residual DNAPL saturations were obtained in the Subsurface Flow and Transport Experimental Laboratory at the EMSL. Development of the experimental procedures to determine residual DNAPL saturations was funded by the Remediation and Closure Science Project through the U.S. Department of Energy's Richland Operations Office. PNNL is operated by Battelle for the U.S. Department of Energy.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results and Discussion
 Conclusions and Conceptual Model...
 REFERENCES
 




This article has been cited by other articles:


Home page
 Vadose Zone JHome page
S. R. Nellis, H. Yoon, C. J. Werth, M. Oostrom, and A. J. Valocchi
Surface and Interfacial Properties of Nonaqueous-Phase Liquid Mixtures Released to the Subsurface at the Hanford Site
Vadose Zone J., April 14, 2009; 8(2): 343 - 351.
[Abstract] [Full Text] [PDF]

Home page
 Vadose Zone JHome page
M. D. White, M. Oostrom, M. L. Rockhold, and M. Rosing
Scalable Modeling of Carbon Tetrachloride Migration at the Hanford Site Using the STOMP Simulator
Vadose Zone J., May 27, 2008; 7(2): 654 - 666.
[Abstract] [Full Text] [PDF]

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]

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 (1)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Oostrom, M.
Right arrow Articles by Rohay, V. J.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Oostrom, M.
Right arrow Articles by Rohay, V. J.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Oostrom, M.
Right arrow Articles by Rohay, V. J.
Related Collections
Right arrow Flow
Right arrow Multi-Fluid Systems
Right arrow Multiphase Fluid Flow

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