Final Environmental Impact Statement for the Tank Waste Remediation System, Hanford Site,

F.2.0 METHODOLOGY AND APPROACH

The approach and steps taken to assess potential impacts to the groundwater system are provided in this section. The alternatives considered in this assessment are as follows:

Tank Waste

No Action
Long-Term Management
In Situ Fill and Cap
In Situ Vitrification
Ex Situ Intermediate Separations
Ex Situ No Separations
Ex Situ Extensive Separations
Ex Situ/In Situ Combination 1
Ex Situ/In Situ Combination 2
Phased Implementation

Cesium and Strontium Capsules

No Action
Onsite Disposal
Overpack and Ship
Vitrify with Tank Waste

These alternatives are described in detail in Volume Two, Appendix B.

As shown on Figure F.2.0.1, the groundwater assessment is divided into three major subtasks: source characterization, vadose zone modeling, and groundwater modeling. These subtasks are discussed in the following sections.

F.2.1 SCREENING ALTERNATIVES

The initial step in screening the alternatives was to determine which alternatives could impact groundwater and eliminate alternatives from rigorous numerical modeling that have little or no potential of impacting groundwater. The following alternatives were screened from numerical modeling analysis:

Cesium and Strontium Capsules

No Action
Onsite Disposal
Overpack and Ship
Vitrify with Tank Waste

Figure F.1.0.2 Area of Potential Groundwater Impact

Figure F.2.0.1 Groundwater Impacts Assessment Approach

The remaining alternatives considered for numerical modeling are the tank waste-related alternatives listed previously. The following sections provide the rationale for screening each alternative for inclusion or exclusion from detailed groundwater modeling. Results of the vadose zone, groundwater flow, and transport simulations used to assess the groundwater impacts of each alternative are provided in subsequent sections.

F.2.1.1 No Action Alternative (Tank Waste)

This alternative would potentially impact groundwater because no remediation would be performed and all waste would remain in the tanks. Releases to the groundwater system would be from the waste in the tanks. During the 100-year institutional control period, tank waste management operations would continue; however, no additional measures, such as the construction of additional tanks, would be implemented to manage the waste. Waste releases to the vadose zone were assumed to occur at the end of institutional control.

F.2.1.2 Long-Term Management Alternative

This alternative would potentially impact groundwater because no remediation would be performed. Double-shell tanks (DSTs) would be retanked every 50 years and all waste would remain in the tanks. During the 100-year institutional control period, tank waste management operations would continue and two DST retanking campaigns would be completed. Releases to the groundwater system would be from the waste in the tanks.

F.2.1.3 In Situ Fill and Cap Alternative

Under this alternative the tanks would be filled with gravel, and a low permeability multi-layer earthen cover (Hanford Barrier) would be placed over the tanks. Potential releases to the groundwater system would be from the contaminants in the waste tanks. The form and inventory of the waste would be similar to the No Action alternative. Waste releases to the vadose zone would occur. The total mass of waste entering the vadose zone and ultimately reaching the groundwater would be the same as for the No Action alternative. However, the releases would occur at a slower rate because the Hanford Barrier restricts the amount of precipitation that would infiltrate into the tanks and carry the waste downward into the vadose zone. While the gravel fill would structurally stabilize the tanks by supporting the tank dome, it would not otherwise reduce infiltration or retard contaminant transport.

F.2.1.4 In Situ Vitrification Alternative

Under this alternative, all tank waste solids would be vitrified in situ (in tank). A Hanford Barrier would then be placed over the vitrified waste. Potential releases to the groundwater system would be associated with the contaminants in the vitrified waste, but the form of the waste and inventory differ from that of the No Action alternative. The In Situ Vitrification alternative would require the addition of materials for forming glass. Organic and other volatile materials present in the No Action alternative inventory would be destroyed or vaporized.

F.2.1.5 Ex Situ Intermediate Separations Alternative

Under this alternative, waste would be retrieved from the tanks, high-level waste (HLW) would be separated from the LAW, and both HLW and LAW would be vitrified. The HLW would then be shipped to a potential geologic repository and the LAW would be disposed of onsite in shallow subsurface burial vaults. A Hanford Barrier would be placed over the tanks and vaults. Potential releases to the groundwater system would be associated with 1) releases during retrieval from the waste tanks; 2) releases from residuals; and 3) releases from the LAW disposal facility.

F.2.1.6 Ex Situ No Separations Alternative

Under this alternative, waste would be retrieved from the tanks, vitrified or calcined, and shipped to a potential geologic repository for disposal. A Hanford Barrier would be placed over the tanks. Potential releases to the groundwater system would be associated with 1) releases during retrieval from the waste tanks; and 2) releases from residuals remaining in the tanks. The vitrified or calcined waste would not have a potential groundwater impact because they would be shipped to the potential geologic repository. The groundwater impacts for this alternative would be the same as the retrieval and residual releases estimated for the tank waste Ex Situ Intermediate Separations alternative.

F.2.1.7 Ex Situ Extensive Separations Alternative

This alternative is similar to the Ex Situ Intermediate Separations alternative with the major difference being that a more extensive separations process would be used. Under this alternative, waste would be retrieved from the tanks, HLW would be separated from the LAW, and both HLW and LAW would be vitrified. The extensive separations process would result in less waste volume and more activity (i.e., curies) shipped offsite to a potential geologic repository and a smaller contaminant source (i.e., curies) associated with the LAW vaults as compared to the Ex Situ Intermediate Separations alternative. A Hanford Barrier would be placed over the tanks and vaults. Potential releases to the groundwater system would be associated with 1) releases during retrieval from the waste tanks; 2) releases from residuals; and 3) releases from the LAW disposal facility. Groundwater impacts associated with retrieval and residual releases would be the same as for the Ex Situ Intermediate Separations alternative. The groundwater impacts resulting from releases from the LAW vaults would be lower than those from the Ex Situ Intermediate Separations alternative LAW vaults because the source term would be smaller.

F.2.1.8 Ex Situ/In Situ Combination 1 Alternative

Under this alternative, approximately half of the waste would be processed as described for the Tank Waste In Situ Fill and Cap alternative. Releases associated with this group of tanks would occur as described previously. Tanks selected for fill and cap processing would contain relatively small amounts of radioisotopes technetium-99 (Tc-99), carbon-14 (C-14), iodine-129 (I-129), and the uranium (U) series, compared to the other tanks. The waste in the remaining tanks would be retrieved and separated into LAW and HLW. The LAW would be placed into shallow subsurface LAW burial vaults in the 200 East Area and the HLW would be shipped offsite for disposal at the potential geologic repository. A Hanford Barrier would be placed over the tanks and vaults.

F.2.1.9 Ex Situ/In Situ Combination 2 Alternative

Under this alternative, 25 tanks would be selected for retrieval and the remaining 152 tanks would be remediated in situ. The retrieved waste would be separated into LAW and HLW. The LAW would be placed into shallow subsurface LAW burial vaults in the 200 East Area and the HLW would be shipped offsite for disposal at the potential geologic repository. A Hanford Barrier would be placed over the tanks and vaults.

This alternative is a variation of Ex Situ/In Situ Combination 1 alternative, and was designed for the ex situ treatment of the largest contributors to long-term risk (i.e., Tc-99, C-14, I-129, and U-238) while limiting the total amount of waste to be retrieved and processed. In the Ex Situ/In Situ Combination 2 alternative, 25 tanks (approximately 30 percent of the total waste in the tanks) would be retrieved compared to 70 tanks (approximately 50 percent of total tank waste) that would be retrieved for Ex Situ/In Situ Combination 1 alternative. Out of the 30 percent of total tank waste retrieved, approximately 85 percent of Tc-99, 80 percent of C-14, 80 percent of I-129, and 50 percent of U-238 would be retrieved rather than 90 percent as with the Ex Situ/In Situ Combination 1 alternative.

F.2.1.10 Phased Implementation Alternative

Phase 1

Under the first phase, waste from the DSTs would be retrieved, vitrified and stored temporarily onsite. There would not be any groundwater impacts under this phase because 1) waste loss is assumed not to occur for retrieval from DSTs which contain primarily liquid and their double containment would catch any releases; and 2) the storage of the vitrified waste is temporary and under controlled conditions.

Total Alternative

In the second phase of this alternative, the plants constructed for Phase 1 would continue to operate and in addition the remainder of the tank waste would be retrieved and treated in the same way as in the Ex Situ Intermediate Separations alternative. Potential releases to the groundwater and contamination from those releases would be the same for the second phase of the Phased Implementation alternative and the Ex Situ Intermediate Separations alternative.

F.2.1.11 Effluent Treatment Facility

This facility would potentially impact groundwater because treated effluent from the ETF will be discharged to a State-approved land disposal site (SALDS) located immediately north of the 200 West Area. All tank alternatives considered, except No Action, would contribute wastewater to the ETF through either periodic operation of the 242-A Evaporator to manage waste volume, DST retanking campaigns, or liquid effluent collected from the process facility. The SALDS would consist of a piping manifold used to infiltrate treated effluent into vadose zone soil and deeper groundwater beneath the site. The primary contaminant present in the treated effluent would be tritium, with other organic, inorganic, and radiogenic contaminants having been removed during the treatment process. Waste releases to the vadose zone beneath the SALDS would be assumed to only occur during the 100-year institutional control period.

F.2.1.12 No Action Alternative (Capsules)

Under this alternative, the capsules would be stored temporarily for a period of 10 years in the Waste Encapsulation and Storage Facility (WESF) until a decision is made to store the waste elsewhere or put the waste to a beneficial use. There would be no groundwater impacts.

F.2.1.13 Onsite Disposal Alternative

Under this alternative, the capsules would be placed in 0.3-meter (m) (1.0-foot [ft]) canisters surrounded by a 0.76-m (2.5-ft) diameter sand backfill. There would be 672 drywells on 5-m (16-ft) center-to-center spacing with a 30-m (98-ft) buffer around the facility. The overall area of the facility would 195 by 195 m (639 by 639 ft) or 38,000 square meters (m²) (410,000 square feet [ft²]). The drywell depth would be 4.6 m (15 ft) belowground surface.

It is estimated that as of December 31, 1995, the total inventory of the Onsite Disposal Facility would be as follows:

Constituent	Inventory	Inventory
	(Curies)	(Grams)
Cs-137	5.22E+07	604,167
Sr-90	2.25E+07	161,583

Because these storage wells are belowgrade in an arid environment, they are assumed to remain intact beyond 500 years past the end of the institutional control period, after which the contents of the drywells are assumed to be released to the vadose zone at a depth of 4.6 m (15 ft) belowground surface.

The potential impact to the groundwater associated with this alternative is not a concern until the time of release. The half-lives of Cs-137 and Sr-90 are 30.17 years and 28.60 years, respectively. The inventory of Cs-137 and Sr-90 remaining when release would occur would be greatly reduced from the present quantities, because of these relatively short half-lives. The estimated inventory 600 years into the future (approximately 20 half-lives) is as follows:

Constituent	Inventory	Inventory
	(Curies)	(Grams)
Cs-137	54	0.625
Sr-90	10.98	0.079

Both Cs-137 and Sr-90 are very immobile in earth systems at the Hanford Site. The distribution coefficients (K_d) for these isotopes are as follows:

Constituent	Distribution Coefficient Kd
	(mL/gram)
Cs-137	51
Sr-90	24

Vadose zone and groundwater flow, and transport simulations performed for the No Action alternative for tank waste (Section F.3.1) indicate that it would take a nominal period of 1,200 years for a constituent with a K_d of one to reach the groundwater. The Cs-137 and Sr-90 are transported much slower because of their greater K_d. Thus, high K_d values of Cs-137 and Sr-90 coupled with their relatively short half-lives mean that no measurable amount of either Cs-137 or Sr-90 would reach the groundwater within the 10,000-year period of interest.

The Cs-137 decays to barium-137 (Ba-137), a stable isotope with an estimated K_d of over 100 (Droppo et al. 1991). This constituent would not be expected to reach the groundwater within the 10,000-year period of interest. The Sr-90 decays to zirconium-90 (Zr-90), also a stable isotope, with an estimated K_d of 8.2 (Droppo et al. 1991). This constituent could reach the groundwater in very low concentrations, 5,000 to 9,000 years from the present, based on simulations performed for the No Action alternative. No further groundwater assessment is provided for this alternative.

F.2.1.14 Overpack and Ship Alternative

Under this alternative, capsules would be removed from temporary storage, overpacked, and shipped offsite. Releases to the vadose zone would not occur and groundwater impacts would not be expected for this alternative. No further groundwater assessment is provided for this alternative.

F.2.1.15 Vitrify with Tank Waste Alternative

Under this alternative, capsules would be removed from temporary storage and vitrified with the tank waste. Releases to the vadose zone would not occur and groundwater impacts would not be expected for this alternative. No further groundwater assessment is provided for this alternative.

F.2.2 SOURCE CHARACTERIZATION

The remainder of this appendix explains the approach (Figure F.2.0.1) and results of assessing groundwater impact by simulating flow and transport through the vadose zone and underlying unconfined aquifer. A series of vadose zone, groundwater flow, and transport simulations for each selected alternative were performed using a combined flow and transport model, VAM2D. The items that potentially impact groundwater are associated with the following common elements.

The contaminant sources are associated with near-surface waste forms (e.g., residuals in tanks and vitrified LAW).
The physical AOI is the unconfined aquifer bounded laterally by the Rattlesnake Hills in the west and southwest, by the Columbia River in the north and east, and by the Yakima River to the south. The bottom of the unconfined aquifer is a confining bed in the bottom of the Ringold Formation.
Contaminants (tank saltcake, sludge, and vitrified waste) are assumed to be released by their desorption and dissolution into pore fluids, and then moved by advection and diffusion from the waste source into the surrounding natural material (Kincaid et al. 1993). Once in the natural material, the contaminants are assumed to move downward by advection with infiltration from precipitation and liquid leakage from tanks. The flow of water and transport of contaminants in the vadose zone is principally in the vertical direction because of the hydraulic gradient, and the geologic structure or layering in the vadose zone is assumed insufficient to result in extensive lateral spreading.
The 40 inactive miscellaneous underground storage tanks associated with the tank farms contain less than one-half of 1 percent of the tank waste and are located in close proximity to the farms. Therefore, the inactive miscellaneous underground storage tank inventory is assumed to be contained within the tank vaults, and potential releases to the groundwater were not modeled separately.

This section addressed source characterization, which involves the following:

Aggregating the many potential sources into common source areas;
Grouping contaminants into categories based on their mobility; and
Developing the source term (i.e., mass flux and fluid flux release as a function of time) for each source area.

The contaminant source for each alternative selected for detailed groundwater analysis is then characterized, based on contaminant inventory data (i.e., mass or activity of each constituent) (Pelton-Davis 1995).

F.2.2.1 Aggregate Source Areas

The 178 potential sources (i.e., each of the 177 tanks and the proposed LAW disposal facility) have been aggregated into 9 discrete source areas based on waste inventory and proximity (Figure F.1.0.1). The criteria used for these groupings are as follows.

The LAW disposal facility would be located in the 200 East Area and would be considered one source area consisting of many vaults. Vault spacing would be approximately 30 m (100 ft). Four alternatives have vaults as a component. The alternative and the continuous area required for the vaults are: Ex Situ Intermediate Separations, Ex Situ Extensive Separations, and Phased Implementation (17 hectares (ha) [42 acres (ac)] each), and Ex Situ/In Situ Combination (9 ha [22 ac]). The vaults would be covered with one continuous Hanford Barrier, and the contents of each vault would be assumed to have the same composition.
The 177 tanks are divided into eight source areas based on tank configuration, tank proximity, and groundwater flow direction. Two types of tanks, single-shell tanks (SSTs) and DSTs have been constructed to hold high-level radioactive waste at the Hanford Site. The 149 SSTs and 28 DSTs are all located in the 200 Areas. Within each type of tank, inventories are assumed to be similar (i.e., contaminants and relative concentrations are similar for all SSTs). It was necessary to first group tanks by type. Tanks were constructed in tank farms ranging from two to 12 tanks in a single tank farm. The tank farms were typically located near waste-generating facilities such as the Plutonium-Uranium Extraction (PUREX) Plant, and thus the tank farms themselves are generally grouped together. The next level of grouping is by tank farm, combining nearby tank farms of same-type tanks into a single source area. Finally, a check was made to determine that source areas did not cross major hydrogeologic features such as groundwater divides.

Table F.2.2.1 provides the tank waste source area designations, a brief description of each source area, and the equivalent area of each source area.

Table F.2.2.1 Tank Waste Source Area Designations and Descriptions

F.2.2.2 Contaminant Groups

The tanks contain more than 100 radioactive and nonradioactive contaminants that could potentially impact groundwater. The approach used for this assessment was to group the contaminants based on their mobility in the vadose zone and underlying unconfined aquifer. The contaminant groups are used rather than the individual mobility of each contaminant primarily because of the uncertainty involved in determining the mobility of individual contaminants. The groups were selected based on relatively narrow ranges of mobility and, where there was uncertainty, contaminants were placed in the more mobile group.

Some of the contaminants such as I and Tc move at the rate of water whether they are in the vadose zone or in the underlying groundwater. The movement of others in water, such as americium (Am) and Cs, are slowed or retarded because they are absorbed onto and with particles within the ground. The VAM2D flow and transport model was able to account for the retardation of contaminant movement with the parameter referred to as K_d, which is the distribution coefficient (milliliters/gram [mL/g]). This parameter is a measure of sorption and is the ratio of the quantity of the adsorbate adsorbed per gram of solid to the amount of the adsorbate remaining in solution (Kaplan et al. 1994). Values of K_d for the contaminants range from 0 (in which the contaminant's movement in water is not retarded) to more than 100 (in which the contaminant moves much slower than water).

The waste inventory was grouped and modeled according to each contaminant's reported or assumed K_d. These groups are defined as follows:

Group 1 - Contaminants are modeled as nonsorbing (i.e., K_d = 0). Contaminant movement is unretarded in water. Contaminant K_d values in this group ranged from 0 to 0.99 mL/g.
Group 2 - Contaminants are modeled as slightly sorbing (i.e., K_d = 1). Contaminant K_d values in this group ranged from 1 to 9.9 mL/g.
Group 3 - Contaminants are modeled as moderately sorbing (i.e., K_d = 10). Contaminant K_d values in this group ranged from 10 to 49.9 mL/g.
Group 4 - Contaminants are modeled as strongly sorbing (i.e., K_d = 50). Contaminant K_d values in this group are 50 mL/g or greater.

The contaminants and associated groupings (based on K_d values) for each alternative are provided in the following sections.

F.2.2.2.1 No Action Alternative (Tank Waste)

Under this alternative, no remediation would be performed, and all waste would remain in the tanks. Potential releases to the groundwater system are all associated with the waste in the tanks. The waste inventory, which is a list of the radioactive and nonradioactive contaminants and their mass or activity for the waste tanks, is provided in Volume Two, Appendix A. For the radioactive contaminants, the mass was estimated for each isotope based on the decay of that isotope as of December 31, 1995. Provided in Table F.2.2.2 is a listing of the contaminants modeled, their estimated half-life (if radioactive) and their K_d group. Several radionuclides in the tank inventory are of little concern and are not considered further in the groundwater assessment. Table F.2.2.3 provides a list of these contaminants and the rationale for eliminating them from further analysis.

F.2.2.2.2 Long-Term Management Alternative

Under this alternative, minimal remediation would be performed, and all waste would remain in the tanks. The two retanking campaigns for the DSTs would delay the release of contaminants from the DSTs for approximately 100 years. Potential releases to the groundwater system are all associated with the waste in the tanks.

The waste inventory, which lists the radioactive and nonradioactive contaminants and their mass or activity for each of the eight tank source areas, is presented in Volume Two, Appendix A. For the radioactive contaminants, the mass estimated for each isotope is based on calculating the decay of that isotope to December 31, 1995.

The contaminants associated with this alternative are the same as those for the No Action (Tank Waste) alternative (Table F.2.2.2). The locations of the new DSTs are assumed to be within the same source area, adjacent to the old DSTs. Because the entire waste inventory is eventually released for this alternative, the portion left as residual in the old DSTs does not affect the model results.

F.2.2.2.3 In Situ Fill and Cap Alternative

Under this alternative, the tanks would be filled with gravel, and a Hanford Barrier would be placed over the tanks. Potential releases to the groundwater system are associated with the contaminants in the waste tanks. The form of the waste and inventory are similar to the No Action alternative. The gravel fill would structurally stabilized the tanks by supporting the dome to prevent collapse. The gravel fill would not otherwise serve to reduce infiltration or retard contaminant transport. The contaminants associated with this alternative are the same as for the No Action alternative. The contaminant inventory is provided in Volume Two, Appendix A. The contaminants, their half life, and mobility are provided in Table F.2.2.2.

Table F.2.2.2 Inventory of Contaminants for the No Action Alternative (Tank Waste)

Table F.2.2.3 Radionuclides Excluded from Vadose Zone and Groundwater Modeling

F.2.2.2.4 In Situ Vitrification Alternative

Under this alternative, all tank waste would be vitrified and disposed of in tank. A Hanford Barrier would then be placed over the tanks. Potential releases to the groundwater system would all be associated with the contaminants in the waste tanks, but the waste form and inventory differ from the No Action alternative. The In Situ Vitrification alternative would volatize certain materials that are present in the No Action alternative inventory. The resulting estimated components of the glass, the mass of these components, and the ionic form for the stable isotopes are provided in Table F.2.2.4.

For groundwater modeling purposes, it was assumed that the vitrified waste form had the same composition as that produced for the Ex Situ No Separations alternative, discussed in Section F.2.2.2.6 (WHC 1995c). The exact composition has not been determined. Contaminants in Table F.2.2.4 are also grouped based on their K_d value.

Table F.2.2.4 Inventory of Contaminants for the In Situ Vitrification Alternative

F.2.2.2.5 Ex Situ Intermediate Separations Alternative

Under this alternative waste would be retrieved from the tanks, HLW would be separated from the LAW and both HLW and LAW would be vitrified. The HLW would then be shipped to a potential geologic repository. The LAW would be disposed of onsite, and a Hanford Barrier would be placed over the tanks. Potential releases to the groundwater system are associated with 1) releases during retrieval from the waste tanks; 2) releases from waste not removed from the tanks (residuals); and 3) releases from the LAW disposal facility. The list of potential contaminants associated with the tank retrieval and residual releases is the same as that provided for the No Action alternative (Table F.2.2.2) and the contaminant inventory is provided in Volume Two, Appendix A. The mass associated with tank residual for each contaminant released is 1 percent of that shown in Appendix A (i.e., 99 percent of the initial inventory is assumed to be retrieved). The amount and type of waste that would remain in the tanks after retrieval is uncertain. The Hanford Federal Facility Agreement and Consent Order (Tri-Party Agreement) (Ecology et al. 1994) set a goal of no more than 1 percent residuals and the ex situ alternatives have been developed to attempt to achieve that goal. However, achieving this level of tank waste retrieval may require extraordinary efforts and cost and it may not be practical to achieve 99 percent retrieval. Conversely, the contaminants that are not recovered are likely to be those that are insoluble in water since substantial quantities of water would be used in an attempt to dissolve or suspend the waste in water during retrieval. Since neither of these issues can be resolved a conservative assumption was made to bound the impacts of the residual waste. For purposes of this analysis it was assumed that 99 percent recovery would be achieved but that the residual would contain 1 percent of all the contaminants including the water soluble contaminants.

The contaminants and groupings for the eight tank source areas are the same as discussed in Section F.2.2.2.1. Based on the above assumptions, the estimated mass released during retrieval operations at each source area is provided in Table F.2.2.5.

The LAW disposal facility would contain LAW, which has been retrieved from the tanks, vitrified, placed in disposal vaults, and capped. The vitrification process requires adding materials for glass formers. Also, the organic and other volatile materials present in the retrieved waste would be destroyed or vaporized during vitrification. The estimated components of the glass, the mass of these components, and their ionic form for stable isotopes are provided in Table F.2.2.6 (WHC 1995j and Jacobs 1996).

F.2.2.2.6 Ex Situ No Separations Alternative

Under this alternative, both HLW and LAW waste would be retrieved from the tanks, vitrified or calcined, and shipped to a potential geologic repository. Potential releases to the groundwater system are associated with 1) releases during retrieval from the waste tanks; and 2) releases from residuals that cannot be removed from the waste tanks (residuals). The list of potential contaminants associated with the retrieval and residual releases is the same as that provided for the No Action alterative (Table F.2.2.2). The mass associated with the residual for each contaminant released is 1 percent of that shown in Volume Two, Appendix A (i.e., 99 percent of the initial inventory is assumed to be retrieved).

The contaminants and groupings for the eight tank source areas are the same as discussed in Section F.2.2.2.1. The estimated mass released during retrieval operations at each source area is provided in Table F.2.2.5.

F.2.2.2.7 Ex Situ Extensive Separations Alternative

This alternative is similar to the Ex Situ Intermediate Separation alternative, with the major difference being that a more extensive HLW and LAW separations process would be used. Under this alternative, waste would be retrieved from the tanks, HLW would be separated from the LAW, and both HLW and

LAW would be vitrified. The extensive separations processes would result in less waste volume and more activity (i.e., curies) being shipped offsite to a potential geologic repository, and a smaller contaminant source would be associated with the LAW vaults. A Hanford Barrier would be placed over the tanks and the vaults. Potential releases to the groundwater system are associated with 1) releases during tank waste retrieval; 2) releases from residuals; and 3) releases from the LAW disposal facility.

Table F.2.2.5 Inventory of Contaminants Released During Retrieval for the Ex Situ Intermediate Separations Alternative

The list of potential contaminants associated with the retrieval and residual releases is the same as that provided for the No Action alternative (Table F.2.2.2). The mass associated with residual for each contaminant released is 1 percent of that shown in Volume Two, Appendix A (i.e., 99 percent of the initial inventory is assumed to be retrieved). The contaminants and groupings for the eight tank source areas are the same as discussed in Section F.2.2.2.1. The estimated mass released during retrieval operations at each source area is provided in Table F.2.2.5.

The LAW vault contaminant inventory would be less than that associated with the vault contaminants from the Ex Situ Intermediate Separations alternative because of the more extensive separations. Table F.2.2.7 lists the LAW vault constituents by K_d group, and the initial inventory.

F.2.2.2.8 Ex Situ/In Situ Combination 1 Alternative

Under this alternative, tanks would be selected for one of two types of remediation based on their contents. Waste would be removed from the tanks containing most of the mobile radionuclides using the same process as described for Ex Situ Intermediate Separations Alternative. The remaining tanks would be remediated in situ as described for the In Situ Fill and Cap alternative.

Contaminants for selecting tanks for waste removal were identified based on the following criteria:

High mobility in the groundwater;
Persistence in the environment; and
Toxicity.

Table F.2.2.6 Inventory of Contaminants for the Low-Activity Waste Vaults - Ex Situ Intermediate Separations Alternative

The contaminants that met these criteria are Tc-99, C-14, I-129, and the isotopes of the uranium series. Reviewing the tank waste inventory identified a total of 70 tanks (60 SSTs and 10 DSTs) that contain approximately 90 percent of these contaminants. Under this alternative, these 70 tanks would be remediated as described for the Ex Situ Intermediate Separations alternative, where the waste would be retrieved, separated into HLW and LAW, the HLW vitrified and transported to an offsite potential geologic repository, and the LAW vitrified and disposed of in onsite LAW vaults. The mass of waste disposed of in vaults would be approximately 49 percent of the mass associated with LAW vaults for the Ex Situ Intermediate Separations Alternative (Table F.2.2.6). The potential contaminants associated with the retrieval and residual releases for this group of tanks are provided in Table F.2.2.8, and the overall contaminant inventory is provided in Volume Two, Appendix A. The 1 percent residual waste assumed remaining in tanks that would be retrieved was added to the inventory that would result from the eventual leaching of the tanks that are remediated in situ. The contaminants and groupings for the eight tank source areas are the same as discussed in Section F.2.2.2.1.

Table F.2.2.7 Inventory of Contaminants for the LAW Vaults - Ex Situ Extensive Separations Alternative

The remaining 107 waste tanks would be processed as described for the tank waste In Situ Fill and Cap alternative. Because of the selective nature of this alternative, the inventory of Tc-99, C-14, I-129, and the isotopes of the uranium series within the 107 tanks would be 10 percent of the original inventory of all 177 tanks. The potential contaminants associated with the tanks remediated in situ (including the 1 percent from tanks retrieved) are provided in Table F.2.2.9.

F.2.2.2.9 Ex Situ/In Situ Combination 2 Alternative

Under this alternative, 25 tanks would be selected for retrieval (thirteen SSTs and twelve DSTs) and 152 tanks would be remediated in situ. The retrieval process and associated released waste would be as described for the Ex Situ Intermediate Separations alternative.

The remaining 152 tanks would be remediated as described for the In Situ Fill and Cap alternative. This variation of the Ex Situ/In Situ combination 1 alternative achieves a high percentage retrieval of the long-term risk contributors (e.g., from 80 to 85 percent of the C-14, I-129, and Tc-99 and 50 percent of the U-288 are retrieved) with a low percentage of total waste retrieval. Thirty percent of the total tank waste would be retrieved for this alternative compared to approximately 50 percent for the Ex Situ/In Situ Combination 1 alternative. The waste retrieved would be separated into HLW and LAW, and the HLW would be vitrified and transported to an offsite potential geologic repository. The LAW would be vitrified and disposed of in onsite LAW vaults.

Table F.2.2.8 Inventory of Contaminants Released During Retrieval - Ex Situ/In Situ Combination 1 Alternative

The inventory of contaminants potentially released during retrieval is provided in Table F.2.2.10. The inventory of contaminants that would be remain in the tanks remediated in situ plus the estimated 1 percent residual of the initial tank contents from tanks from which retrieval would occur is provided in Table F.2.2.11. The mass of contaminants that would be disposed of in the LAW vaults is provided in Table F.2.2.12. The contaminants and groupings for the eight tank source areas are the same as discussed in Section F.2.2.2.1.

F.2.2.2. 10 Phased Implementation

Phase 1

In Phase 1 of this alternative, waste from the DSTs would be retrieved, vitrified, and stored temporarily onsite. There are no potential contaminants from this phase since it is assumed that there will be no retrieval losses from the DSTs due to the nature of their construction. In addition, the storage of the vitrified waste is temporary.

Table F.2.2.9 Inventory of Contaminants in Tanks Remediated In Situ - Ex Situ/In Situ Combination 1 Alternative

Table F.2.2.10 Inventory of Contaminants Released During Retrieval - Ex Situ/In Situ Combination 2 Alternative

Table F.2.2.11 Inventory of Contaminants in Tanks Remediated In Situ - Ex Situ/In Situ Combination 2 Alternative

Table F.2.2.12 Inventory of Contaminants for Low-Activity Waste Vaults - Ex Situ/In Situ Combination 2 Alternative

Total Alternative

In Phase 2 of this alternative, the plants constructed for Phase 1 would continue to operate, and in addition the remainder of the tank waste would be retrieved and treated in the same way as in the Ex Situ Intermediate Separations alternative. Potential releases to groundwater are associated with retrieval, residuals remaining in tanks, and the LAW disposal facility. The list of potential contaminants from retrieval is provided in Table F.2.2.5. The list of contaminants for tank residuals is 1 percent of the amounts given in Table F.2.2.2., and the contaminant inventory is provided in Volume Two, Appendix A. The inventory of contaminants associated with the LAW vaults is provided in Table F.2.2.6.

F.2.2.3 Source Terms

The numerical modeling approach used to assess groundwater impacts requires understanding and quantifying when, what, and how much (mass or activity) contaminants are released. The quantification of this information is the source term. Also included in this section is the water flux into the vadose zone. Flux is the time-variable volume of water that enters the vadose zone and may be from natural sources such as infiltrating precipitation or artificial sources, such as releases from the tanks during retrieval operations.

The model requires a consistent set of units for the input parameters and geometry. Units of meters, days, and grams were selected. There are parameters, such as infiltration rate, that are not commonly reported in the units used for the modeling. For such cases, the commonly used units are reported first followed by the consistent units used for modeling.

F.2.2.3.1 No Action Alternative (Tank Waste)

The source term for this alternative is derived from the eventual failure of the tanks and release of their inventory into the vadose zone. Water flux versus time diagrams are provided for this alternative and for the In Situ Vitrification alternative only, to illustrate diagrammatically when and how much water is estimated to be passing through the waste, which ultimately impacts contaminant concentrations in the underlying groundwater.

Developing the source term requires an understanding of the expected operating conditions at the eight tank source areas and is discussed in the following text. The discussion will first focus on the water flux, followed by the estimation of contaminant concentrations.

Institutional control is assumed to be maintained for a 100-year period. During this period, waste management operations would continue such that necessary repairs would be performed, but there would be no scheduled tank replacement. The tank facilities are assumed to be maintained in their current condition (e.g., no vegetation around tank farms). Drainable liquid would continue to be removed from the SSTs during this period. Infiltration from precipitation is assumed to be 5.0 centimeters/year (cm/year) (1.36E-4 meters/day [m/day]) at both the SST and DST source areas during institutional control and for the period afterwards, based on ranges reported in the literature (Kincaid et al. 1995, Wood et al. 1995, Gee 1987, Gee et al. 1992, and Rockhold et al. 1990).

Contaminant release from the SSTs and DSTs is assumed to begin at the end of institutional control. Figure F.2.2.1 provides a diagrammatic representation of the water flux used in modeling contaminant transport from source area 1WSS. The duration of the release is based on a congruent dissolution model. In this model, all constituents in the waste inventory are assumed to be released in proportion to the most abundant material in the waste inventory, nitrate, and at the rate of nitrate dissolution. Thus, the duration of release for each source area is based on the solubility of nitrate, which is assumed to be 360 grams/liter (g/L) (Serne-Wood 1990), volumetric water flux (area of source times 5.0 cm/year (1.36E-04 m/day [2.0 in./year])), and the initial mass of nitrate in the inventory. It should be noted that the source term developed under this alternative is overly conservative for many of the contaminants modeled because solubility controls in groundwater of neutral pH (7.0 to 8.0) and relatively oxidizing conditions (EH of 300 to 400 mv SHE) would cause the contaminants to be leached at a rate less than nitrate (NO₃¹), or because the contaminants would be insoluble under these conditions. A simple example of the congruent dissolution model follows:

Figure F.2.2.1 Water Flux Versus Time for Tank Source 1WSS, No Action Alternative (Tank Waste)

Given the following data:

Source area of 100 m² (=1,080 ft²);
Infiltration of 5.0 cm/year (1.36E-04 m/day) into the waste; and
Solubility of nitrate = 360 g/L (3.6E+05 g/m³).

The volumetric flux into the waste is 1,000,000, cm² 5.0 cm/year = 5,000,000 cm³/year. The time required to dissolve the inventory is 7.2E+08 g/(0.360 g/cm³ 5.0E+06 cm³/year) = 400 years. In this example, the rate of release for nitrate and technetium is 7.20E+08/400 = 1.8E+06 g/year and 3.6E+04/400 = 90 g/year, respectively. Note that chromium (Cr), a potentially high-risk contaminant, has a solubility limit that controls its dissolution rate. Its solubility is substantially lower than would be calculated by the congruent dissolution release model. For this case, the solubility of Cr was used in predicting release rather than that of the congruent dissolution release model.

The release durations and total mass of nitrate released for each of the eight source areas are provided in Table F.2.2. 13 for the No Action alternative. The contaminant concentrations for each of the eight source areas are provided in Table F.2.2. 14 . Contaminant concentrations are provided for the K_d groups that reach groundwater within the period of interest (i.e., K_dGroups 1 and 2).

Table F.2.2.13 Contaminant Releases Modeled for the No Action Alternative (Tank Waste)

F.2.2.3.2 Long-Term Management Alternative

The source term for the Long-Term Management alternative is nearly identical to that described for the No Action alternative. The only difference is that contaminant release for the DSTs begins 100 years later than the time assumed for the No Action alternative. This would occur because under the Long-Term Management alternative, the DSTs would be replaced with new tanks every 50 years, with the last group of new DSTs being completed in 100 years. A 100-year effective life is assumed for the new DSTs. The release durations and total mass of nitrate released for the simulations performed for each of the eight source areas are provided in Table F.2.2. 15 . The contaminant concentrations are the same as provided for the No Action alternative (Table F.2.2.11).

Table F.2.2.14 Concentration of Contaminants Released for the No Action Alternative (Tank Waste)

Table F.2.2.15 Contaminant Releases Modeled for the Long-Term Management Alternative

F.2.2.3.3 In Situ Fill and Cap Alternative

The source term for this alternative is a result of releases from the waste stored within the tanks. There is no retrieval from the tanks, thus the initial contaminant inventory is the same as assumed for the No Action alternative (Volume Two, Appendix A). The tanks would be filled with gravel to provide structural stability, and a Hanford Barrier would be constructed over the tanks to reduce the infiltration of precipitation.

Initially, there would be a 28-year construction phase in which the tanks would be structurally stabilized (i.e., filled with gravel) and a Hanford Barrier would be installed over each source area. Activities at the Site such as removing snow, diverting runoff, and protecting the open tanks from incident rainfall are assumed to lower infiltration from the base value of 5.0 cm/year (1.36E-04 m/day) to 0.5 cm/year during this 28-year period. Contaminant releases from the tanks during this 28-year period are assumed not to occur, thus, the total water flux is from infiltrating precipitation and is 0.5 cm/year (1.36E-05 m/day).

For the next 1,000 years, infiltration through the Hanford Barrier is assumed to be approximately 0.05 cm/year (1.37E-06 m/day). As a human-made structure, the Hanford Barrier is not expected to maintain its design functions indefinitely. A simplifying assumption used in this analysis is that the infiltration through the Hanford Barrier would double to approximately 0.1 cm/year (2.74E-06 m/day), 1,000 years after the Hanford Barrier was initially constructed. Infiltration is assumed to remain at this level for the remainder of the period of interest, 10,000 years from the present. For all tanks, releases to the vadose zone from the waste are assumed to begin 500 years after completing the Hanford Barrier. The waste inventory and constituent concentrations are the same as those for the No Action alternative (Volume Two, Appendix A and Table F.2.2. 11 ). The principal constituent of the waste is nitrate, and the congruent dissolution release model is used to estimate release from the waste, which is the same approach as described for the No Action alternative. The dissolution rate of nitrate is assumed to remain constant at 360 g/L (360,000 mg/L) (Serne-Wood 1990), regardless of the water flux. For 500 years, the water flux through the intact Hanford Barrier is limited to 0.05 cm/year (1.37E-06 m/day). The mass flux is estimated as follows:

1.37E-6 m/day · 360,000 mg/L= 0.49 g/day-m².

After 500 years, when the water flux is assumed to double to 0.1 cm/year (2.74E-06 m/day), the dissolution rate remains constant at 360,000 mg/L, resulting in a doubling of the mass flux to approximately 0.98 g/day-m² until the mass of nitrate has been depleted. Table F.2.2. 16 provides the contaminant release durations and the total mass of nitrate released for each of the eight source areas. The source term developed for the In Situ Fill and Cap alternative is very conservative for many of the contaminants modeled because solubility constraints in groundwater of neutral pH (7.0 to 8.0) and relatively oxidizing conditions (EH 300 to 400 mv SHE) will cause the contaminants to be either leached at a rate less than nitrate, or to be insoluble.

Table F.2.2.16 Contaminant Releases Modeled for the In Situ Fill and Cap Alternative

F.2.2.3.4 In Situ Vitrification Alternative

The source term for this alternative is a result of release from in-tank vitrified waste. Both SSTs and DSTs are discussed together in the following text because the vitrified waste inventory is similar (i.e., the contaminants and their relative concentrations are assumed to be approximately the same in each tank).

Developing the source term requires understanding the expected operating conditions at the eight tank source areas. The following discussion will first focus on the water flux, followed by the estimate of contaminant concentrations. Initially, there is a 38-year construction phase where the in situ vitrification equipment is tested and set up, vitrification takes place, and a Hanford Barrier is installed over each source area (Figure F.2.2.2). Activities at the Hanford Site such as removing snow, diverting runoff, and protecting the open tanks from incident rainfall are assumed to reduce the infiltration from the base value of 5.0 to 0.50 cm/year (1.36E-04 to 1.36E-05 m/day) during this 38-year period. During the 38-year period, the total water flux from infiltrating precipitation is assumed to be 0.5 cm/year (1.36E-05 m/day). For the next 1,000 years, infiltration through the Hanford Barrier is assumed to be 0.05 cm/year (1.36E-06 m/day). As a human-made structure, the Hanford Barrier is not expected to maintain its design function indefinitely. An assumption used is that the infiltration through the Hanford Barrier increases by a factor of two 1,000 years after Hanford Barrier construction to 0.10 cm/year (2.74E-06 m/day). Infiltration is assumed to remain at this level for the remainder of the period of interest. All tank releases to the vadose zone from the vitrified waste form are assumed to begin 500 years after completing the Hanford Barrier, in accordance with the Nuclear Regulatory Commission (NRC 1994).

The vitrification process requires adding materials to make glass. Also, the organic and other volatile materials initially present in the waste inventory would be destroyed or vaporized. The release model for the vitrified mass was based on a constant total mass loss rate of 1E-03 g/m² day (Shade et al. 1995). This mass loss rate is independent of the water flux from recharge. The composition of the vitrified mass was assumed to be identical to the soda-lime glass, which is formed in the Ex Situ No Separations Alternative (WHC 1995c). The concentration of the contaminants released is then assumed to be proportional to their concentration in the soda-lime glass. Because the total mass loss rate is constant, the composition of the released solution is unaffected by the recharge rate. As the recharge rate doubles after 1,000 years, the mass flux increases proportionately. The low value of the total mass loss rate, combined with the very large quantity of vitrified mass results in a release time measured in millions of years.

Figure F.2.2.2 Water Flux Versus Time for All Tank Sources, In Situ Vitrification Alternative

The release durations and total mass of silicon dioxide (SiO₂)released for each of the eight source areas are provided in Table F.2.2.17 for the In Situ Vitrification alternative. The initial contaminant concentrations for each of the eight source areas are provided in Table F.2.2. 18 . Contaminant concentrations are provided for K_d Group 1. The other K_d groups do not reach groundwater within the period of interest.

Table F.2.2.17 Contaminant Releases Modeled for the In Situ Vitrification Alternative

Table F.2.2.18 Concentrations of Contaminants for the In Situ Vitrification Alternative in K_d Group One (K_d = 0)

F.2.2.3.5 Ex Situ Intermediate Separations Alternative

The source term for this alternative is a result of releases from SSTs, DSTs, and the LAW disposal facility. Each of these potential sources are discussed in the following text.

Single-Shell Tanks

The source term for the SSTs is based on contaminant releases from two events:

Releases to the vadose zone during retrieval of waste from the SSTs; and
Releases to the vadose zone from residual contaminants in the SSTs.

Developing the source term requires understanding the expected operating conditions at the eight tank source areas. The following discussion on operating conditions will focus on the water flux and later the estimated contaminant concentrations.

The waste would be retrieved from the SSTs over a 15-year period. Work is assumed to be ongoing at all of the eight sites during this period. The infiltration rate is assumed to decrease from 5.0 to 0.5 cm/year (1.36E-04 to 1.36E-05 m/day) during the 15-year retrieval period because of construction and retrieval activities (e.g., removing snow and diverting runoff).

The retrieval operations within the SSTs are assumed to result in relatively early contaminant releases to the vadose zone. The release volume per tank is assumed to be 15,000 L (4,000 gal). Using source area 1WSS as an example, the water flux is calculated as follows:

Give the following data:

Numbers of tanks = 40
Combined area of tanks = 14,900 m²
Period over which release occurs = 15 years
Release volume per tank = 15,000 L.

The water flux due to the retrieval release at 1WSS is estimated as:

Total volume released = 15,000 L/tank 40 tanks 1 m³/1,000 L = 600 m³
Flux from release = 600 m³/(14,900 m² 15 years 365.25 day/year) = 7.42E-06 m/day.

Thus during the 15-year retrieval period, the water flux from tank releases at 1WSS is approximately 7.4E-06 m/day. The flux from infiltrating precipitation during this period is assumed to be 0.5 cm/year (1.4E-05 m/day). The total water flux infiltrating into the vadose zone during this period is the sum of the releases from the tanks and natural infiltration, or 2.1E-05 m/day.

Following the retrieval period is a 14-year construction period when the Hanford Barrier would be built over the source areas. Water flux into the vadose zone in the vicinity of the source area is assumed to be approximately 0.5 cm/year (1.36E-05 m/day) during this period for the same reasons as used for the retrieval period. For the next 1,000 years infiltration through the Hanford Barrier is assumed to be approximately 0.05 cm/year (1.36E-06 m/day). As a man-made structure, the Hanford Barrier is not expected to maintain its design functions indefinitely. A simplifying assumption used herein is that the infiltration through the Hanford Barrier increases at some point in time. Thus, infiltration through the Hanford Barrier is assumed to double to approximately 0.1 cm/year (2.74E-06 m/day) 1,000 years after the Hanford Barrier was initially constructed. Infiltration is assumed to remain at this level for the remainder of the period of interest.

The other release that impacts groundwater is from waste tank residual contaminants. The retrieval process is assumed to be 99 percent effective, leaving within the tanks 1 percent of the waste inventory. Major assumptions used in developing the source term for the tank residuals are:

The residual materials are assumed in a relatively saturated state;
The Hanford Barrier limits the potential for infiltrating precipitation to mobilize the residuals;
The solubility of each contaminant is proportional to the solubility of nitrate;
Release to the vadose zone begins 500 years after the Hanford Barrier has been installed (NRC 1994); and
The residuals are present in the same proportion as the initial inventory.

The release of residuals in the tanks is assumed to begin 500 years after barrier completion for all the SSTs. The congruent dissolution release model is used to estimate retrieval and residual releases of contaminants into the vadose zone following the same approach as described for the No Action alternative. Also, contaminant concentrations for residual release are the same as those used for the No Action alternative (Table F.2.2. 14 ). The release durations and masses associated with retrieval and residuals for each source area are summarized in Table F.2.2. 19 .

Double-Shell Tanks

The source term for DSTs is based on contaminant releases to the vadose zone from residual contaminants. Releases are not expected from the DSTs during the retrieval process because of their double-shell construction and associated leak-capture systems.

The timing of the residual releases are similar to the SSTs. Release from residuals in the tanks begins 500 years after barrier completion for all the DSTs. The duration of the releases are for each DST source area and are also summarized in Table F.2.2. 19 . Again, the congruent dissolution release model and the No Action alternative contaminant concentrations are used (Table F.2.2. 14 ).

Low-Activity Waste Disposal Facility

The Ex Situ Intermediate Separations alternative has a groundwater source associated with LAW disposal in addition to the eight tank source areas. The source term for the LAW disposal facility is a result of releases from the waste, which has been vitrified and placed in vaults.

Table F.2.2.19 Contaminant Releases Modeled for the Ex Situ Intermediate Separations Alternative

Waste retrieval from the tanks, waste separation processes, waste vitrification, vault construction, and Hanford Barrier construction are assumed to occur over a 39-year period.

Activities at the LAW disposal site (such as removing snow and diverting runoff) are assumed to lower the infiltration from the base value of 5.0 to 0.5 cm/year (1.36E-04 to 1.36E-05 m/day) during this 39-year period.

For the 1,000-year period after the Hanford Barrier is constructed, infiltration through the Hanford Barrier is assumed to be approximately 0.05 cm/year (1.4E-06 m/day). As a human-made structure, the Hanford Barrier is not expected to maintain its design functions indefinitely. A simplifying assumption used herein is that the infiltration through the Hanford Barrier doubles to approximately 0.1 cm/year (2.7E-06 m/day), 1,000 years after the Hanford Barrier is initially constructed. Releases to the vadose zone from the vitrified waste form are assumed to begin 500 years after completing the Hanford Barrier (NRC 1994).

The vitrification process requires adding materials for glass makeup. Also, organic and other volatile materials initially present in the waste inventory would be destroyed or vaporized. The release model for the glass waste form was based on a constant corrosion rate of 3E-06 cm/year (8.16E-11 m/day) (Jacobs 1996). The corrosion rate is independent of the water flux from recharge. The composition of the LAW glass is taken from the engineering date package for this alternative (WHC 1995j). The release concentration of the contaminants is assumed to be proportional to their concentration in the LAW glass. Because the total mass loss rate is constant, the composition of the released solution is unaffected by the recharge rate. As the recharge rate doubles after 1,000 years, the mass flux increases proportionately. The low value of the corrosion rate, combined with the very large quantity of vitrified mass, results in a calculated release time of 170,000 years.

The release duration for the LAW disposal facility goes beyond the period of interest. The contaminant concentrations for each of the eight source areas are provided in Table F.2.2. 20 . Contaminant concentrations are provided for K_d Group 1 (K_d = 0), which is the only K_d group that reaches groundwater during the period of interest.

Table F.2.2.20 Concentrations of Contaminants for the LAW Vault Source for K_d Group One (K_d = 0) Ex Situ Intermediate Separations Alternative

F.2.2.3.6 Ex Situ No Separations Alternative

The source term for this alternative is as described for the Ex Situ Intermediate Separations alternative, except that all retrieved waste is disposed of offsite, thus there is no source from a LAW facility on the site. The source term for the SSTs is based on contaminant releases from two events:

Releases to the vadose zone during retrieval of waste from the SSTs; and
Releases to the vadose zone from residual contaminants in the SSTs.

The source term for the DSTs is based on contaminant releases to the vadose zone from residual contaminants. Releases would not be expected from the DSTs during the retrieval process because of their double-shell construction and associated leak-capture systems. The details and assumptions associated with the source terms for both the SSTs and DSTs are provided earlier in the tank waste Ex Situ Intermediate Separations alternative section

F.2.2.3.7 Ex Situ Extensive Separations Alternative

The source term for this alternative is as described for the Ex Situ Intermediate Separations alternative, except waste inventory for the LAW facility is smaller because of the more extensive separations process. The source term for the SSTs is based on contaminant releases from two events:

Releases to the vadose zone during retrieval of waste from the SSTs; and
Releases to the vadose zone from residual contaminants in the SSTs.

The source term for the LAW facility involves the same construction sequencing and physical parameters (e.g., infiltration rates, contaminant release mechanisms, and vadose zone hydrogeologic setting) as would apply to the Ex Situ Intermediate Separations alternative, except that the waste inventory in the LAW vaults would be smaller (Table F.2.2.7). The initial concentration of contaminants in the LAW vaults is provided in Table F.2.2. 21 . The details and assumptions associated with the source terms for both the SSTs and DSTs are provided earlier in the discussion on the Ex Situ Intermediate Separations alternative.

Table F.2.2.21 Concentration of Contaminants in K_d Group One (K_d = 0) For the Ex Situ Extensive Separations Alternative (LAW Vaults)

F.2.2.3.8 Ex Situ/In Situ Combination 1 Alternative

The source term for this alternative has three components:

Releases to the vadose zone during retrieval from 60 SSTs. Waste from 10 DSTs would be retrieved, however, it is assumed there would be no retrieval losses because of the DST construction.
Releases to the vadose zone from waste in the 107 tanks that would be remediated in situ using the fill and cap technology.
Releases from a LAW disposal facility.

As with the Ex Situ Intermediate Separations Alternative, there is an assumed residual mass (1 percent of the original waste inventory) left in tanks where waste is retrieved. This residual waste has been added to the waste inventory of tanks that would be remediated in situ.

Retrieval Releases

Tank waste released during retrieval is based on the same rationale as was used for the Ex Situ Intermediate Separations Alternative. Retrieval releases are assumed to occur only from the 60 SSTs that would be selected for retrieval. Waste retrieved would be separated into HLW and LAW. Both would be vitrified with the HLW being sent to a potential geologic repository and LAW disposed of onsite in vaults. The initial concentrations of those contaminants assumed to be lost during retrieval are provided in Table F.2.2.22 . The contaminant release periods for the source areas and associated masses on which the release periods are based are provided in Table F.2.2. 23 .

In Situ Remediation Releases

The source term for this component of the alternative is the result of waste that is leached out of the 107 tanks that are remediated in situ. These tanks would be filled with gravel and covered with a Hanford Barrier as described for the In Situ Fill and Cap alternative.

The gravel fill would provide structural stability and a Hanford Barrier would be constructed over the tanks to reduce infiltration of precipitation. Eighty-nine SSTs and 18 DSTs would be remediated in this fashion.

Initially, there is a 28-year construction phase in which the tanks are structurally stabilized (i.e., filled with gravel) and a Hanford Barrier is installed over each source area. Activities at the Site such as removing snow, diverting runoff, and protecting the open tanks from incident rainfall are assumed to have the effect of lowering the infiltration from the base value of 5.0 to 0.50 cm/year (1.36E-04 to 1.36E-05 m/day) during this 28-year period. Contaminant releases from the tanks during this 28-year period are assumed not to occur; thus, the total water flux is from infiltrating precipitation, which is 0.5 cm/year (1.36E-05 m/day).

Table F.2.2.22 Concentration of Contaminants in K_d Group 1 (K_d = 0) Associated with Waste Retrieval- Ex Situ/In Situ Combination 1 Alternative

Table F.2.2.23 Contaminant Releases Modeled for the Ex Situ/In Situ Combination 1 Alternative

For the next 1,000 years, infiltration thorough the Hanford Barrier is assumed to be approximately 0.05 cm/year (1.36E-06 m/day). The Hanford Barrier, as a human-made structure, is not expected to maintain its design functions indefinitely. A simplifying assumption used herein is that the infiltration through the Hanford Barrier doubles to approximately 0.10 cm/year (2.74E-06 m/day), 1,000 years after the Hanford Barrier was initially constructed. Infiltration is assumed to remain at this level for the remainder of the period of interest, 10,000 years from the present. For all tanks, releases to the vadose zone from the waste are assumed to begin 500 years after completing the Hanford Barrier.

The waste inventory was provided in Section F.2.2.2.8. The estimated constituent concentrations are provided in Table F.2.2. 24 . The principal constituent of the waste is nitrate, and the congruent dissolution release model is used to estimate release from the waste, which is the same approach as described for the No Action alternative. The dissolution rate of nitrate is assumed to remain constant at 360 g/L (360,000 mg/L) (Serne-Wood 1990), regardless of the water flux. From the 500-year period when the water flux through the intact Hanford Barrier is limited to 0.05 cm/year (1.36E-06 m/day), the mass flux for an area encompassing 1 m² is estimated as follows: 1.36E-06 m³/day 360,000 g/m³ = 0.49 g/day. The release from this component has been adjusted upward by 1 percent of the waste retrieved to account for contaminants that might be left in the tanks and not removed during the retrieval process.

Table F.2.2.24 Concentration of Contaminants in K_d Group One (K_d=0) Associated with Waste Tanks Remediated In Situ for the Ex Situ/In Situ Combination 1 Alternative

After 1,000 years, when the water flux is assumed to double to 0.1 cm/year (2.74E-06 m/day), the dissolution rate remains constant at 360,000 mg/L. This results in a doubling of the mass flux for a 1 m² area to approximately 0.98 g/day, until the mass of nitrate is released for each of the eight source areas.

Low-Activity Waste Disposal Facility

The source term for the LAW disposal facility would be the result of releases from the waste, which has been vitrified and placed in vaults.

Waste retrieval from the tanks, waste separation processes, waste vitrification, vault construction, and Hanford Barrier construction are assumed to occur over a 39-year period. Activities at the LAW disposal site such as heavy equipment yards, parking lots, snow removal, and runoff diversion are assumed to have the effect of lowering the infiltration from the base value of 5.0 to 0.50 cm/year (1.36E-04 to 1.36E-05 m/day) during this 39-year period. The vitrified waste is placed day thus during this 39-year period, the total water flux from infiltrating precipitation is assumed to be 0.5 cm/year (1.36E-05 m/day).

For the 1,000-year period after the Hanford Barrier has been constructed, infiltration through the Hanford Barrier is assumed to be approximately 0.05 cm/year (1.36E-0.6 m/day). The Hanford Barrier, as a human-made structure, is not expected to maintain its design functions indefinitely. A simplifying assumption used herein is that the infiltration through the Hanford Barrier doubles 1,000 years after the Hanford Barrier was initially constructed to approximately 0.10 cm/year (2.74E-06 m/day). Infiltration is assumed to remain at this level for the remainder of the period of interest, 10,000 years from the present. Releases to the vadose zone from the vitrified waste form are assumed to begin 500 years after completing the Hanford Barrier.

The vitrification process requires the addition of materials for glass make-up. Also, the organic and other volatile materials initially present in the waste inventory are destroyed or vaporized. The release model for the glass waste form was based on a constant corrosion rate of 3E-06 cm/year (8.16E-11 m/day). This corrosion rate is independent of the water flux from recharge. The composition of the low-activity glass is taken from the engineering data package for this alternative (WHC 1995j). The release concentration of the contaminants is assumed to be proportional to their concentration in the low-activity glass. Because the total mass loss rate is constant, the composition of the released solution is unaffected by the recharge rate. As the recharge rate doubles after 1,000 years, the mass flux increases proportionately. The low value of the corrosion rate, combined with the very large quantity of vitrified mass, results in a calculated release time of approximately 83,000 years. The mass of contaminants placed in the LAW vaults is approximately 49 percent of the mass shown for the Ex Situ Intermediate Separations Alternative (Table F.2.2.6). The initial waste concentrations are assumed to be the same as the concentrations from the vaults for the Ex Situ Intermediate Separations alternative (Table 2.2. 20 ).

It should be noted that the source term developed under this alternative may be overly conservative for many of the contaminants modeled because solubility controls in groundwater of neutral pH (7.0 to 8.0) and relatively oxidizing conditions (EH of 300 to 400 mv SHE) will cause the contaminants to be leached at a rate less than nitrate, or because the contaminants would be insoluble under these conditions. This can be expected to effect the final results by increasing the maximum concentrations calculated in groundwater and narrowing the spread of the contaminants distribution with time.

F.2.2.3.9 Ex Situ/In Situ Combination 2 Alternative

The source term for this alternative has three components:

Releases to the vadose zone during retrieval from 13 SSTs. There would be no retrieval losses from the retrieved from the 12 DSTs because of their construction.
Releases to the vadose zone from the waste in the 152 tanks that would be remediated in situ using the fill and cap technology.
Releases from a LAW disposal facility.

As with the Ex Situ Intermediate Separations alternative, there is an assumed residual mass of 1 percent of the original waste that would remain in the 25 waste tanks after the waste had been retrieved. This 1 percent residual waste has been added to the inventory of tanks that would be remediated in situ.

The calculated impacts for this alternative are based in part on the calculated impacts from the Ex Situ/In Situ Combination 1 alternative, which have been appropriately scaled for the contaminant inventory. The retrieval and LAW components of the two alternatives are sufficiently similar to allow for scaling of the results. This was accomplished on a source area by source area basis by multiplying the Ex Situ/In Situ Combination 1 alternative calculated groundwater concentrations for a specific contaminant at each time frame by the ratio of the mass of that specific contaminant in the Ex Situ/In Situ Combination 2 alternative to that in the Ex Situ/In Situ Combination 1 alternative.

The impacts from the in situ remediation component of this alternative was calculated using the same modeling approach as was used for the In Situ Fill and Cap alternative.

Retrieval Releases

Tank waste released during retrieval is based on the same rationale as was used for the Ex Situ Intermediate Separations Alternative. Retrieval releases are assumed to occur only from the 13 SSTs that would be selected for retrieval. Waste retrieved would be separated into HLW and LAW. Both would be vitrified with the HLW being sent to a potential geologic repository and LAW disposed of onsite in vaults. The initial concentrations of those contaminants assumed to be lost during retrieval are provided in Table F.2.2.25. The contaminant release periods for the source areas and associated masses on which the release periods are based are provided in Table F.2.2.26.

In Situ Remediation Releases

The source term for this component of the alternative is the result of waste that is leached out of the 152 tanks that are remediated in situ. These tanks would be filled with gravel and covered with a Hanford barrier as described for the In Situ Fill and Cap alternative.

The gravel fill would provide structural stability and a Hanford Barrier would be constructed over the tanks to reduce infiltration of precipitation. One hundred thirty six SSTs and 16 DSTs would be remediated in this fashion.

For the next 1,000 years, infiltration through the Hanford Barrier is assumed to be approximately 0.05 cm/year (1.36E-06 m/day). The Hanford Barrier, as a human-made structure, is not expected to maintain its design functions indefinitely. A simplifying assumption used herein is that the infiltration through the Hanford barrier doubles to approximately 0.10 cm/year (2.74E-06 m/day), 1,000 years after the Hanford Barrier was initially constructed. Infiltration is assumed to remain at this level for the remainder of the period of interest, 10,000 years fro the present. For all tanks, releases to the vadose zone from the waste are assumed to begin 500 years after completing the Hanford Barrier.

The waste inventory was provided in Section F.2.2.2.9. The estimated constituent concentrations for tanks remediated in situ are provided in Table F.2.2.27. The principal constituent of the waste is nitrate, and the congruent dissolution release model is used to estimate release from the waste, which is the same approach as described for the No Action alternative. The dissolution rate of nitrate is assumed to remain constant as 360 g/L (360,000 mg/L) (Serne-Wood 1990), regardless of the water flux. From the 500-year period when the water flux through the intact Hanford Barrier is limited to 0.05 cm/year (1.37E-06 m/day),, the mass flux for an area encompassing 1 m² is estimated as follows: 1.37E-06 m³/day - 360,000 g/m³ = 0.49 g/day. The release from this component has been adjusted upward by 1 percent of the waste retrieved to account for contaminants that might be left in the tanks and not removed during the retrieval process. After 1,00 years, when the water flux is assumed to double to 0.1 cm/year (2.74E-06 m/day), the dissolution rate remains constant at 360,000 mg/L.

Table F.2.2.25 Concentration of Contaminants in K_d Group 1 (K_d = 0) Associated with Waste Retrieval- Ex Situ/In Situ Combination 2 Alternative

Table F.2.2.26 Contaminant Releases Modeled for the Ex Situ/In Situ Combination 2 Alternative

Table F.2.2.27 Concentration of Contaminants in K_d Group 1 (K_d = 0) Associated with Waste Tanks Remediated In Situ for the Ex Situ/In Situ Combination 2 Alternative

Low-Activity Waste Disposal Facility

The source term for the LAW disposal facility would be the result of releases for the waste, which has been vitrified and placed in vaults. This component of the alternative is very similar to the Ex Situ/In Situ Combination 1 alternative, which is described in Section F.2.2.3.8. The groundwater impacts for LAW vaults component of this alternative is calculated on a contaminant by contaminant basis by taking the product of the calculated contaminant concentration for the Ex Situ/In Situ Combination 1 alternative and the ratio of the LAW vault mass for each contaminant of the Ex Situ/In Situ Combination 2 alternative by the mass of the Ex Situ/In Situ Combination 1 alternate mass. The resulting ratio is less than one for the long-term risk contributor contaminants (e.g. C-14, I-129, Tc-99, and U-238) because there is less mass of these contaminants in Ex Situ/In Situ Combination 2 alternative compared to the mass in Ex Situ/In Situ Combination 1 alternative.

F.2.2.3. 10 Phased Implementation Alternative

There would be no contaminant release nor source of groundwater contamination under the first phase of this alternative as explained in Section F.2.1. 10. The source term for Phase 2 of this alternative would be the same as that for the Ex Situ Intermediate Separations alternative, discussed in Section F.2.2.3.5.

F.2.3 VADOSE ZONE MODELING

The approach used to predict contaminant transport through the vadose zone was to perform one-dimensional modeling through the vadose zone at each of the eight tank source areas and the LAW disposal facility. One-dimensional modeling through a uniformly porous media is a conservative approach that does not allow for lateral spreading in the vadose zone and tends to reduce the calculated time that it takes contaminants to reach the water table. This approach requires reducing the volumetric flux rates at the surface to one dimension by dividing by the area of the waste source. The corresponding model node(s) in the groundwater model were later assigned the appropriated area to allow the groundwater transport model to receive the volumetric flux for the source area.

Conceptual models were developed for each of the source areas, which included a Site-specific diagram of the model stratigraphy, the upper and lower boundaries, and a table of material units and corresponding flow and transport parameters. The conceptual model was used to guide the setup of the numerical model.

The first phase of the modeling effort entailed establishing the initial flow field (based on an assumed infiltration rate of 5.0 cm/year [2.0 in./year]) to be used to determine the initial velocity values throughout the vadose zone column. This was accomplished by performing a steady-state flow simulation through the one-dimensional column at each site. The initialization file represents Site conditions in the year 1995 (time equals zero), and was used as a startup file for each alternative.

Once the initial flow modeling was performed, the input file for each site was copied and modified to perform combined transient flow and transport modeling for each of the alternatives considered. The file at each site was modified appropriately to represent the transient fluid flux and contaminant source conditions conceptualized for each alternative. One node at the base (i.e., vadose zone and groundwater contact) of the model representing 1 m² (11 ft²) in area was defined as an observation node.

The concentration and fluid flux exiting this node was tracked through time to generate a graph of concentration and fluid flux over time. From the graph, the contaminant mass entering the groundwater system from the entire source area was calculated and used to construct input records for the groundwater model.

F.2.3.1 Vadose Zone Conceptual Models

A conceptual model was developed for each of the eight tank waste sites and the LAW vault site. The conceptual model consists of Site-specific geometry, waste release information, and hydrostratigraphy, including the hydraulic parameters. The conceptual models were used to construct and run numerical transport simulations through the vadose zone. The results of the numerical modeling included the solute concentration and fluid flux released from the vadose zone to the top of the unconfined aquifer through time.

F.2.3.1.1 Hydrostratigraphy and Soil Properties

The Site-specific stratigraphy and subsequent model domain geometry were obtained from isopleth maps and boring logs contained in the 200 East Groundwater Aggregate Area Management Study Report (DOE 1993a) and 200 West Groundwater Aggregate Area Management Study Report (DOE 1993b). The basic hydrostratigraphy and hydraulic parameters for the 200 West Area sites have been extracted principally from Wood et al. (Wood et al. 1995). Hydrostratigraphy and hydraulic parameters for the 200 East Area sites and LAW vaults are based principally on Kincaid et al. (Kincaid et al. 1993). Depth and diameter values for the base of the tank were obtained from the document entitled Tank Characterization Reference Guide (WHC 1994f). Tank depths ranged from 11 to 17 m (37 to 57 ft). In cases where depths differed within a source area, the largest tank depth was used.

The basic hydrostratigraphic units (material types) consist of the following units:

Material Type	200 East Area	200 West Area
1	Hanford Formation, sandy sequence	Hanford Formation
2	Hanford Formation, gravel sequence	Early Palouse Soil
3	Ringold Formation	Pliocene/Pleistocene Unit
4	Not applicable	Ringold Formation

Depths of tanks and major hydrostratigraphic sequences in the 200 Areas are provided in Table F.2.3.1 and Table F.2.3.2.

Table F.2.3.1 Tank Depths and Material Property Intervals Within the Vadose Zone at 200 East Source Areas

Table F.2.3.2 Tank Depths and Material Property Intervals Within the Vadose Zone at 200 West Source Areas /P>

For each of the source areas, a vertical grid spacing of 0.1 m (3.9 in.) was used. Figures F.2.3.1 and F.2.3.2 depict the system geometry used for modeling each of the source areas at 200 Areas source sites.

Figure F.2.3.1 Conceptual Profiles of the Vadose Zone for Source Areas in the 200 West Area

Figure F.2.3.2 Conceptual Profiles of the Vadose Zone for Source Areas in the 200 East Area

F.2.3.1.2 Flow Properties

Input parameters required by VAM2D for the variably saturated flow modeling include:

Infiltration rate;
Porous medium properties;
Constitutive relationships for variably saturated flow; and
Initial and boundary conditions.

These parameters and the values used for the vadose zone modeling effort are described in the following subsections.

Infiltration Rate

Flow through the vadose zone is primarily controlled by the degree of water saturation in the pore space and the unsaturated hydraulic conductivity function, which is in turn affected by the quantity of infiltration (recharge) coming from the surface. As used here, the infiltration rate is the amount of precipitation that enters the soil, is not removed by evaporation or plant transpiration, and eventually reaches the groundwater table. This input of water to the model is also referred to as fluid flux.

The annual infiltration rate assumed at the Hanford Site is 5.0 cm/year (2.0 in./year) based on work reported by Gee et al. (Gee et al. 1992) and Rockhold et al. (Rockhold et al. 1990). Previous studies indicate that infiltration rates at the Hanford Site vary from 0 to 10 cm/year (0 to 2.74E-04 m/day) depending primarily on precipitation amounts and vegetative cover. As discussed in a recent report (Rockhold et al. 1990), gravel-covered lysimeters designed to simulate tank farm conditions on the 200 Areas plateau drained approximately 4.3 cm (0.043 m) of water from an initially dry condition for the previous year. The assumed infiltration rate of 5.0 cm/year (1.36E-04 m/day) is also consistent with vadose zone modeling for the performance assessment for the disposal of LAW in the 200 West Area (Wood et al. 1995). For additional discussion on infiltration rate, refer to Sections F.4.3.5 and Section F.4.4.

An initial steady-state velocity field was established for each site modeled. This velocity field was based on an infiltration rate of 5.0 cm/year (1.36E-04 m/day) and was used to represent initial conditions for each alternative at time equals zero years. Infiltration conditions throughout the 10,000-year period of interest varied according to the fluid flux source term developed for the alternative, as described in Section F.2.2.3.

Porous Medium Properties

Input parameters describing porous medium properties include the vertical component of saturated hydraulic conductivity (K_s) residual saturation (S_wr) and the saturated and residual water content, (_s and _r, respectively). Saturated hydraulic conductivity, K_s, is defined as the rate of flow of water through a unit cross-sectional area of the aquifer under a unit hydraulic gradient at the prevailing temperature and density of the water (Walton 1985). The saturated water content, _s, (also referred to as the total porosity), is defined as a percentage, representing the volume of a soil or rock occupied by void spaces (pores) divided by the total volume of the soil or rock (Freeze-Cherry 1979). Residual saturation, S_wr, (also called specific retention) is a measure of the water retaining capacity of the rock and is expressed quantitatively as the percentage of the total volume of rock occupied by groundwater that will be retained in interstices against the force of gravity (Walton 1985). Residual water content, _r, is defined as the water content that remains under a relative permeability of zero. In other words, the water content that cannot be removed even under extreme levels of suction.

These parameters are related according to the following equation:

Input values for each of these parameters for each material type in the 200 East Area were primarily obtained or calculated from Kincaid et al. (Kincaid et al. 1993). These values are presented in Table F.2.3.3. Input values for the 200 West Area (Table F.2.3.4) were obtained from Wood et al. (Wood et al. 1995).

Table F.2.3.3 Porous Medium Properties for Each Material Type in the 200 East Area

Table F.2.3.4 Porous Medium Properties for Each Material Type in the 200 West Area

Kincaid et al. (1993) specified a value of 0.498 for _s in the Ringold Formation in the vicinity of the 200 East Area. This value was considered unrealistically high; therefore _s and the related S_wr and _r values were changed to the values reported for the Ringold Formation in the 200 West Area. The value reported for _r in Wood et al. (Wood et al. 1995) was 0.0. It was assumed that the reported value was below detection and was reported as a zero. Therefore, a small number (0.001) was assumed in its place to maintain the relationship between _s and S_wr stated previously.

Constitutive Relationships for Variably Saturated Flow

Two alternative functional expressions are used to describe the unsaturated hydraulic conductivity function (relationship of relative permeability versus moisture content). These functions are the Brooks-Corey relation and the van Genuchten relations. The van Genuchten relations , , and , were used for the vadose zone modeling. These parameters were selected over the Brooks-Corey parameters because they were available in the published literature at the Site.

The van Genuchten curve shape parameters (, , and ) are used to characterize the hysteretic saturation-pressure head relation (i.e., hysteresis). The relation is bounded by wetting and drying curves; thus , , and and have different values for wetting and drying. The parameter is an empirical value, defined as the inverse of the air entry pore water pressure. The and parameters are dimensionless empirical shape factors for the wetting/drying curve.

By using several simplifying assumptions (e.g., the wet and dry values for are equal, the wet and dry values for are equal, and = 1-1/), only five parameters are necessary to characterize hysteresis; porosity, residual saturation, and shape parameters (wet), (dry), and (Huyakorn et al. 1991).

Input values for the vadose zone modeling effort at the 200 Areas source areas are provided on Tables F.2.3.5 and F.2.3.6. For the 200 East Area, and were obtained from Table 3.42 in Kincaid et al. (Kincaid et al. 1993) and was calculated using the following equation, obtained from Indirect Methods for Estimating the Hydraulic Properties of Unsaturated Soils (van Genuchten-Leij 1989):

= 1 - (1/)

Table F.2.3.5 Van Genuchten Parameters for Each Material Type in the 200 East Area

Table F.2.3.6 Van Genuchten Parameters for Each Material Type in the 200 West Area

For the 200 West Area, and were obtained from Table F.19 in Wood et al. (Wood et al. 1995) and was calculated using the preceding equation.

Initial and Boundary Conditions

Initial and boundary conditions required to define the flow field for the vadose zone simulation include:

Initial distribution of pressure head, _o;
Prescribed values of pressure head at the water table, ; and
Prescribed values of nodal fluid flux at the surface, Q.

Pressure head distribution was initialized at 0.0 for the first steady-state simulation, which assumes the soil column is fully saturated. The VAM2D model developers recommended using full saturation as an initial condition and allowing the model to adjust pressure to achieve steady-state unsaturated conditions.

Prescribed (fixed) values of 0.0 pressure head were assumed for node values at the water table (i.e., the bottom of the soil column). As stated previously, a prescribed fluid flux (recharge) of 5 cm/year (2.0 in./year) was assumed for the initial steady-state runs. This fluid flux value was also used throughout the period of interest for the No Action alternative. However, fluid flux conditions varied through time with the other alternatives, consistent with activities expected at the Hanford Site.

F.2.3.1.3 Contaminant Transport Properties

Transport parameters required by VAM2D for the vadose zone modeling effort include:

Free water molecular diffusion coefficient, D_m;
Longitudinal dispersivity, _L:
Effective porosity, ;
Bulk density, _B;
Distribution coefficient, K_d;
Darcy velocity components of the fluid phase considered, v₁ and v₂; and
Prescribed values of solute flux at boundary nodes, q_c.

Decay of radioisotopes was accounted for during post processing. Daughter products were not considered. A brief description of these parameters and the initial values used for the vadose zone modeling effort is provided as follows.

Free Water Molecular Diffusion Coefficient

Diffusion is the process where ionic or molecular constituents move under the influence of their concentration gradient from zones of high concentrations to zones of lower concentrations, even in the absence of groundwater flow. As a result of diffusion, concentrations will tend to equalize in all parts of the aquifer system over time (Walton 1985). For the VAM2D model, the free water molecular diffusion coefficient, D_m, is input as a porous medium property with the units of length squared per time.

Molecular diffusion coefficients depend on the solute, solute concentration, and temperature. For major ions in water (e.g., Na⁺, K⁺, Mg²⁺, Ca²⁺, Cl^-, CO₃^2-, SO₄^2-), diffusion coefficients range from 1E-09 to 2E-09 m²/s at 25 C (Walton 1985). Diffusion coefficients in porous materials are commonly 0.5 to 0.01 times the values of the diffusion coefficient in water, thus they typically range from 5E-11 to 1E-13 m²/s (4.3E-06 to 8.6E-09 m²/day) (Walton 1985). A value of 4.0E-06 m²/day was selected to be used for transport modeling in the vadose zone.

Longitudinal Dispersivity

Mechanical dispersion is the process of the individual groundwater particles and chemical constituents traveling at variable velocities through irregular-shaped interstices and meandering streamlines. The result of this movement is the arrival of the chemical constituents at an earlier time than predicted by groundwater flow velocity alone. The VAM2D model requires longitudinal dispersivity, _L, with units of length, to be input as porous medium properties. Because one-dimensional modeling was performed in the vadose zone, transverse dispersivity was not relevant.

Values of longitudinal dispersivity are best determined by field studies at a particular site. A discussion of longitudinal dispersivity in the unsaturated zone at Hanford is presented in the environmental setting data document (Schramke et al. 1994). This document recommends that if no value is available from the site data, the estimate to be used for longitudinal dispersivity in the vadose zone is based on the following equation:

_L = 0.01 (Th)

Where: Th is the thickness of the layer (material unit).

Longitudinal dispersivity values used for vadose zone modeling (Table F.2.3.7) were taken from Schramke et al. (Schramke et al. 1994). These values have been calculated by Schramke (Schramke et al. 1994) and appear to be based on the reported thickness of the material unit and not determined by field studies.

Table F.2.3.7 Longitudinal Dispersivity Values for the 200 East and 200 West Areas

Effective Porosity

Total porosity is defined as a percentage, representing the volume of a soil or rock occupied by void spaces (pores) divided by the total volume of the soil or rock (Freeze-Cherry 1979). Effective porosity, _e, is the percentage volume of soil through which flow occurs and is often quantified as the specific yield, representing the volume of water that will drain from the pore spaces of a saturated soil or rock material (aquifer). Effective porosity is typically somewhat less than the total porosity due to adhesion of water molecules to the aquifer material and cohesion of water molecules to one another and the presence of dead end pore spaces. Effective porosity was estimated from saturated and residual water content data using the relationship _e=_s-_r. Calculated values are presented in Tables F.2.3.8 and F.2.3.9.

Table F.2.3.8 Effective Porosity Values for the 200 East Area

Table F.2.3.9 Effective Porosity Values for the 200 West Area

Bulk Density

Bulk density, _B, is the weight per unit volume of a dry soil mass. Estimates of bulk density for the 200 Areas are provided in the Table F.2.3.10.

Table F.2.3.10 Bulk Density Values for the 200 East and 200 West Areas

Distribution Coefficient

The movement of chemical species is retarded within the aquifer due to sorption, which may include the processes of adsorption on aquifer materials, ion exchange, colloid filtration, reversible precipitation, and irreversible mineralization (Walton 1985). The distribution coefficient, K_d, quantifies the sorption process and is the slope of the curve representing the amount of solute in the solid phase to the concentration of solute in solution as follows (Walton 1985):

K_d = mass of solute on the solid phase per unit mass of solid phase
concentration of solute in solution

Higher values of K_d indicate lower mobility of the solute. Radionuclides were grouped into categories according to mobility (represented by the distribution coefficient) to limit the number of simulations. K_d groups are summarized in Section F.2.2.2. For additional discussion on distribution coefficient, refer to Section F.4.3.5 and F.4.4.

Darcy Velocity Components of the Fluid Phase

For the vadose zone, the Darcy velocity refers to the rate of flow of the solute through a cross-sectional area of a porous medium (the vadose zone) in response to differences in pressure. This pressure is the sum of chemical, capillary, and gravitational forces. The designations v₁ and v₂ correspond to the Darcy velocities in the x and y dimensions, respectively. Units of Darcy velocity are length per time. Because the vadose zone model was one-dimensional, only the y (vertical) component of Darcy flow was used. The average linear velocity (the average velocity of unretarded contaminant migration) can be calculated by dividing the Darcy velocity by the effective porosity.

The initial values of Darcy velocity (used as input for the combined contaminant flow and transport model) were obtained by the steady-state flow simulation through the vadose zone for the assumed infiltration rate of 5.0 cm/year (2.0 in./year) (refer to Section F.2.3.1.2). The values of Darcy velocity are written to an output file from the steady-state flow model. This output file is subsequently used as input for the transient, combined flow and transport model. The model calculates Darcy velocity at each timestep for the transient model run, based on the fluid flux source term developed for each specific alternative modeled.

Prescribed Values of Solute Flux at Boundary Nodes

Solute flux refers to the mass of solute entering the model at boundary nodes. Solute flux has the units of mass of solute per volume per time. Input for solute flux for the vadose zone model was developed for each alternative modeled. Section F.2.2.3 discusses the development of source terms for each alternative.

Decay Coefficient

Radioactive decay is the spontaneous disintegration of radionuclide atoms into new nuclides (also called daughter products), which may be stable or undergo further decay until a stable nuclide is finally created. Contaminant concentrations were adjusted in post processing as described in Section F.2.4.2 to account for radioactive decay. This was done because each radionuclide decays at a different rate. Accounting for radioactive decay during modeling would have necessitated a separate model run for each radioactive constituent.

F.2.3.2 Post-Processing for Groundwater Model Input

Vadose zone modeling results included a graph of concentration and flux values at the vadose zone and groundwater interface throughout the 10,000-year period of interest. This information was subsequently post-processed in the following manner:

A graph was generated to represent the solute flux exiting the vadose zone for the 1-m² (11-ft²) area represented by the observation node. This was compared to modeling results for other alternatives at the Hanford Site and results from nearby source areas for the same alternatives to ensure the modeled results appeared to be valid. In particular, the times of first arrival of solute flux at the water table and the peak concentration of contaminants were evaluated.
The concentration and fluid flux with time were imported into a spreadsheet and mass balance was calculated to verify the results adequately represented the modeled scenario. In most cases, the mass of solute flux exiting the vadose zone at the observation node was within 1 percent of the calculated mass entering the top of the vadose zone column from the source area. Mass balance could not be verified for several of the alternatives because solute was still present within the vadose zone column at the end of the time period modeled.
The vadose zone model results generally included more than 10,000 values representing the concentration and flux at each time step. From this information, up to 12 values (Figure F.2.3.3) were selected to represent a step function for input to the groundwater model. The total mass represented by these selected values was calculated to ensure they represented the total mass exiting the vadose zone. Generally, the selected values slightly underestimated the total mass, therefore two additional time values were selected before and after the peak of the curve to flatten out the peak and achieve 100 percent of the total mass. Figure F.2.3.3 illustrates the process of choosing the values to represent input to the groundwater model.
Based on the time versus concentration points selected previously, the mass representing 1 m² (11 ft²) of the source area was multiplied by the total area, defined as the sum of tank bottom areas, to determine the total mass entering groundwater from the source. This mass was next divided by the number of nodes in the groundwater model used to represent the source area (one to four nodes, depending on the size and geometry of the source). Transient flux records were then generated for input to the groundwater model based on this information.

F.2.4 GROUNDWATER MODELING

Contaminant transport through the saturated unconfined aquifer was simulated with the VAM2D model at each of the eight tank source areas and the LAW disposal facility.

Figure F.2.3.3 Example of Groundwater Model Input Development from Vadose Model Results

A conceptual model was developed for the unconfined aquifer that included stratigraphy, the upper and lower aquifer boundaries, and a table of material units and corresponding flow and transport parameters. The conceptual model was used to guide the setup of the numerical model. A grid spacing of 250 m (820 ft) was established for the Hanford Site and overlain onto a Site map containing physical features and the source area boundaries. Node numbers of model boundaries (e.g., basalt outcrop and subcrop areas, river nodes, wastewater effluent discharge points, the eight tank source areas, and the LAW disposal facility) were determined to allow numerical representation of these features for the modeling effort.

The first phase of the modeling effort entailed establishing the steady-state flow field that was consistent with previous Site-wide groundwater flow simulations (Wurstner-Devary 1993). This was accomplished by adopting, as closely as possible, the hydraulic parameters from the previous effort. This was necessary to generate the velocity field for subsequent contaminant transport simulations. The steady-state results with the VAM2D model clearly matched results previously reported. The steady-state flow field, which is one of the principal bases for the groundwater impacts assessment, was developed using December 1979 sitewide water level measurements because it was determined (Wurstner-Devary 1993) that this data set was most representative of steady-state conditions. Using this data set also meant that the mounding from U Pond and B Pond would be evident. The mounding was recognized as a present-day condition that may dissipate over the next several decades with changes in the Site waste management practices. It is conservative from an overall groundwater concentration and risk perspective to determine groundwater impacts with the mounds in place because the vadose zone would be thinner in the 200 West and 200 East Areas and contaminant travel times would be faster to the groundwater, resulting in higher concentrations in groundwater and higher risk. The travel time in the unconfined aquifer to the Columbia River would not be materially affected by the groundwater mounds, compared to the vadose zone travel time. The approach based on the December 1979 water level data provides conservative, comparable results for each alternative, especially in light of the uncertainties of waste disposal practices and how it would affect the present groundwater mounds, future land use such as irrigation to the west of the site and on the site, uncertainty in the depth of contamination in the unconfined aquifer, and climate change.

Once the initial flow modeling was completed, input files were developed to perform transient transport modeling from each source area for each of the alternatives. The results of the vadose zone modeling were used to develop input records for the groundwater model. Consequently, each groundwater simulation calculated contaminant levels in the unconfined aquifer resulting from a single source area. These were later combined during post-processing to represent contaminant levels from all source areas.

The approach of performing separate contaminant transport simulations for each source area and each K_d group and later combining the results during post-processing allowed one model simulation to represent all contaminants with similar mobility from one source area. This significantly decreased the number of model runs needed to assess each alternative.

F.2.4.1 Groundwater Conceptual Model

Previous groundwater modeling efforts at the Hanford Site formed much of the basis for developing the conceptual model of the unconfined aquifer. One such study is the recent modeling effort and ongoing study for the Ground-Water Surveillance Project, performed by Pacific Northwest National Laboratory (PNL). This ongoing project includes two-dimensional (2-D) modeling of regional groundwater flow using the Coupled Fluid, Energy, and Solute Transport (CFEST) code. Several documents describing this effort have been published, the most recent being the Hanford Site Ground-Water Model: Geographic Information System Linkages and Model Enhancements, FY 1993, hereafter referred to as the CFEST model document (Wurstner-Devary 1993).

A second published document specifies modeling parameter data to be used for modeling efforts in support of the U.S. Department of Energy (DOE) Programmatic Environmental Impact Statement (PEIS). This document is the Hanford Site environmental setting data developed for the Unit Risk Factor Methodology in Support of the PEIS, hereafter referred to as the environmental setting data document (Schramke et al. 1994). This document was used primarily as a source of information for the contaminant transport parameters.

F.2.4.1.1 Geology

The geology of the Hanford Site is described in detail in a number of reports (Thorne-Chamness 1992 and Tallman et al. 1979). Detailed geologic information can be obtained from these reports. Information for the following summary was obtained primarily from the environmental setting data document (Schramke et al. 1994) and the CFEST Model document (Wurstner-Devary 1993).

The Hanford Site is located on the Columbia Plateau within a structural depression known as the Pasco Basin (Schramke et al. 1994). Structural features within the Pasco Basin include two synclinal areas know as the Cold Creek and Dry Creek valleys (Kincaid et al. 1993) and three anticlinal structures known as the Rattlesnake Hills, Yakima Ridge, and Gable Mountain structures (Wurstner-Devary 1993).

The stratigraphic column, as described by various authors, is presented in Figure F.2.4.1. Local formations, from oldest to youngest, include the Columbia River Basalt Group, overlain by the Ringold Formation, glaciofluviatile and fluviatile deposits known as the Hanford Formation, and recent alluvial and eolian sediments. These include the following:

Columbia River Basalt Group. Flood basalts with associated Ellensburg Formation sedimentary interbeds, deposited 6 to 17 million years ago (Tertiary Period).
Ringold Formation. A thick sequence of coarser-grained (gravel, sand, and silt) migrating channel deposits and the finer-grained overbank deposits of ancestral river systems.
Hanford Formation. A complex series of coarse and fine-grained layers deposited by cataclysmic floods during the last ice age.
Recent deposits. Recent alluvial and eolian deposits, primarily reworked Hanford Formation sediments.

Figure F.2.4.1 Stratigraphic Column for the Hanford Site Showing Nomenclature From Previous Investigations

F.2.4.1.2 Hydrology

In general, the unconfined aquifer is located within the Ringold and Hanford formations, in consolidated to semi-consolidated sediments overlying the Columbia River Basalt. Because of deposition in a structural depression, the Ringold Formation is up to 366 m (1,200 ft) thick within the Pasco Basin. The Ringold formation is up to 38 m (125 ft) thick at the 200 East Area and up to 84 m (274 ft) thick at the 200 West Area. In addition, the upper portion of the aquifer is more transmissive than the finer-grained lower portion.

Historically, the unconfined aquifer was located almost exclusively in the Ringold Formation, except for a few areas near the Columbia River. A confining bed at the base of the Ringold Formation serves as an aquitard and inhibits the vertical migration of contaminants downward from the unconfined aquifer. However, wastewater discharges occurring since 1944 (Kincaid et al. 1993) have raised the water table, causing water levels to enter the Hanford Formation in the 200 East Area and in a wider area near the Columbia River (Wurstner-Devary 1993). Because of the increase in groundwater elevation, the water table is now in the Hanford Formation over much of the eastern portion of the Hanford Site (Thorne-Newcomer 1992). In general, water levels have increased at least 15 m (49 ft) in the vicinity of the 200 West Area and 5 m (16 ft) in the vicinity of the 200 East Area. The groundwater mounding created a vertical downward gradient in the areas of wastewater discharge. However, this downward gradient does not extend to the area between Gable Butte and Gable Mountain where there is an erosional window in the aquitard.

The change of the water table elevation is important to the modeling effort because the Hanford Formation is 10 to 100 times more permeable than the Ringold Formation (Wurstner-Devary 1993). Groundwater mounds of approximately 28 and 9 m (90 and 30 ft) have developed under wastewater discharge areas at the 200 Areas. Although more water has been discharged at the 200 East Area, the mound is higher at the 200 West Area because of a lower aquifer hydraulic conductivity.

F.2.4.1.3 Flow Properties

Groundwater in the unconfined aquifer generally flows from recharge areas on the western boundary of the region east and north towards the Columbia River. Groundwater recharge occurs primarily in the Cold Creek, Dry Creek, and Yakima River valleys and in wastewater discharge areas. Groundwater discharge occurs along the Columbia River.

For the modeling effort it was assumed that no interaction exists between the unconfined aquifer and the underlying confined aquifer. There is a potential for leakage between these systems in areas of increased vertical hydraulic conductivity, such as the area northeast of the 200 East Area (Wurstner-Devary 1993). Although limited quantitative information exists on these areas, adequate flow system calibration was obtained without including these areas in the model (Wurstner-Devary 1993). Flow in the Columbia River Basalts is not considered in this study because the basalts are hydraulically isolated from the Ringold Formation in areas where downward hydraulic gradients would have the potential to cause contaminants to move into the basalts. The vertical gradients resulting from groundwater mounds in the wastewater discharge areas will dissipate within a short period compared to the time frame of interest.

The modeling effort for all TWRS alternatives assumed steady-state flow conditions for December 1979, consistent with the CFEST modeling effort. The CFEST modeling effort generated water elevation contours as part of a steady-state 2-D model run. This information is presented in Figure F.2.4.2. Water elevation contours in Figure F.2.4.2 are based on conditions observed in December 1979 (Wurstner-Devary 1993).

Although fluid flux volumes (based on infiltration) change at the eight source sites in accordance with Hanford Site activities for each alternative, these changes in infiltration are not important to groundwater elevations and flow velocities at the Site. For example, a steady-state run based on an infiltration of 0.5 cm/year (0.2 in./year) at the tank source areas locally affected groundwater elevations by approximately 1.0E-05 m (3.2E-05 ft), compared with no infiltration at the source areas.

Transmissivity and saturated water content were based on values used for the CFEST modeling effort. Transmissivity values ranged from 5.5 to 6.5E+03 m/day (18 to 2.1E+04 ft/day). Transmissivity values used for CFEST are presented in Figure 4.2.8 of the CFEST model document (Wurstner-Devary 1993). Saturated water content was set at 0.5.

Boundary Conditions

The conceptual flow model includes several areas defined as no flow, fixed head, and fixed flux boundaries. These conditions reflect the physical conditions at the Site affecting flow. Physical boundaries include the Rattlesnake Hills, Yakima Ridge, Umtanum Ridge, and the Columbia and Yakima rivers. The boundary conditions for the modeling effort are consistent with previously published groundwater modeling efforts performed by PNL (Wurstner-Devary 1993).

The Rattlesnake Hills, Yakima Ridge, and Umtanum Ridge are outcrop areas of the Columbia River Basalt. These three features follow the axes of anticlines (the Rattlesnake Hills Structure, Yakima Ridge Structure, and Gable Mountain Structure, respectively, see Volume Five, Appendix I). The Columbia River Basalt where it occurs as an outcrop or subcrop acts as a flow barrier. Consequently, the model boundary adjacent to these features is defined as a no-flow boundary. The two synclinal areas between these structures, known as the Cold Creek and Dry Creek valleys (Kincaid et al. 1993), recharge the aquifer. To achieve model calibration with the CFEST model, the Cold Creek drainage was defined as a constant head boundary, and the Dry Creek drainage was defined as a fixed flux boundary.

The Yakima River recharges the unconfined aquifer in the southern part of the AOI, creating a hydraulic gradient in this area from west to east (Wurstner-Devary 1993). The model boundary adjacent to this river is set as a constant head boundary. The Columbia River, located along the northern and eastern perimeter of the AOI, drains the unconfined aquifer and is set as a constant head boundary.

Figure F.2.4.2 Results of CFEST Steady-State Simulations for December 1979

Four interior boundaries are defined by outcrops of the Columbia River Basalt. These consist of Gable Butte, Gable Mountain, and two unnamed basalt outcrop areas south of Gable Mountain. These areas are defined as no flow boundaries.

The average wastewater discharge quantities for 1979 were used as part of the flow conceptual model. These discharges influence flow conditions in the 200 Areas. Wastewater discharge areas are defined as fixed flux boundaries. Fluid flux quantities are summarized in Table F.2.4.1.

Table F.2.4.1 Summary of Fixed Flux Boundaries from CFEST Model

Additional fixed flux boundary conditions were established to allow contaminant input from the source areas. Although these boundary conditions are not defined as part of the CFEST model, their effect on groundwater elevations is inconsequential. The source area boundary conditions assumed an infiltration rate of 0.5 cm/year (0.20 in./year) would originate from the tank areas. The volumetric fluid fluxes were calculated in m³/day by multiplying the infiltration rate by the area of the source. Fluid flux quantities of the source areas are summarized in Table F.2.4.2.

F.2.4.1.4 Saturated Zone Contaminant Transport Properties

Once contaminants move through the vadose zone and enter the unconfined aquifer, they migrate in the groundwater until they are intercepted by a well or discharged to the Columbia River. Generally, contaminants will move from source locations at the 200 East Area towards the east, and from source locations at the 200 West Area towards the north and east, eventually discharging to the Columbia River or one of the springs located adjacent to the river.

Table F.2.4.2 Source Area Fixed Flux Boundaries

Previous simulations of contaminant transport in the unsaturated zone and unconfined aquifer have shown the time of travel within the vadose zone is much greater than in the aquifer (Kincaid et al. 1993). Existing tritium contaminant plumes originating in the 200 East Area reached the Columbia River in 25 to 30 years.

Saturated zone contaminant transport parameters required by VAM2D for the modeling effort include:

Free water molecular diffusion coefficient, D_m;
Longitudinal and transverse dispersivity, _L and _T;
Effective porosity, _e;
Bulk density, _B;
Distribution coefficient, K_d;
Darcy velocity components of the fluid phase considered, v₁ and v₂; and
Prescribed values of solute flux at boundary nodes, q_c.

Contaminant concentrations were adjusted in post processing (Section F.2.4.2) to account for radioactive decay. A brief description of each of the above parameters is provided in the following text.

Free Water Molecular Diffusion Coefficient

Groundwater flow across the Hanford Site is sufficient to make a molecular diffusion coefficient value in the range of 2E-9 to 1E-11 m²/second insignificant to contaminant transport simulation. Therefore, this value was set to 0.0 for the saturated zone model runs. A more detailed discussion of the molecular diffusion coefficient is provided in Section F.2.3.1.3.

Dispersion Parameters

A discussion of longitudinal and transverse dispersivity in the saturated zone is presented in the environmental setting data document (Schramke et al. 1994). If no value is available from the site data,

the recommended estimate to be used for longitudinal dispersivity is based on the following equation (Walton 1985):

_L = 0.1 (Tr)

Where:

Tr is the length of the travel path (plan view) from the center of the waste site to the receptor point.

Transverse dispersivity should be calculated as 1/5 of the longitudinal dispersivity (Walton 1985):

_T = 0.2 (_L)

Walton (Walton 1985) states that the equation for _L applies to mean travel distances less than about 305 m (1,000 ft). However, the actual relationship is not linear. Consequently, the equation may not be valid for transport simulation across the Hanford Site and should only be considered an upper bound to dispersivity.

In the field, dispersivity approaches a maximum asymptotic value (Walton 1985), and the equation used to estimate longitudinal dispersivity is:

Where:

A_d	=	asymptotic or maximum dispersivity (L)
B_a	=	mean travel distance corresponding to A_d/2 (L)
L_d	=	mean travel distance (L).

Walton (Walton 1985) also presents a graph depicting field measurements of _L versus the mean travel distance of the plume (Figure F.2.4.3). In this graph, the maximum dispersivity value approaches approximately 125 m (400 ft). Due to the large travel distances modeled at this site, the maximum dispersivity value presented by Walton (Walton 1985) was selected for the groundwater modeling effort.

A second factor affected by the value used for dispersivity is the Peclet number, which is defined as:

Where

= (longitudinal) dispersivity

= largest grid dimension in model

In groundwater modeling, the local Peclet number criterion should not exceed a value of 4, or, in cases where the flow is steady-state, it should not exceed 2 or 3 (Huyakorn et al. 1985). Applying this formula to fix the Peclet number at 2 and the longitudinal dispersivity at 125 m (400 ft) yields a maximum grid of 250 m (820 ft). Thus, the selected value of 125 m (410 ft) satisfies the Peclet number and is consistent with values observed in the field.

Figure F.2.4.3 Longitudinal Dispersivity Values Observed in the Field

Effective Porosity

Effective porosity values estimated for the Hanford Site were presented in the environmental settings data document (Schramke et al. 1994). These values are presented by environmental setting areas, (Figure F.2.4.4) defined within the document. Table F.2.4.3 presents effective and total porosity values reported in a number of tables within Appendix B of the environmental settings data document. These data were recommended by Schramke (Schramke et al. 1994) to be used for the saturated zone at each area.

Figure F.2.4.4 Hanford Environmental Settings Areas

Table F.2.4.3 Effective Porosity Values Recommended for the Hanford Site

Bulk Density

Bulk density values estimated for the Hanford Site were presented in the environmental setting data document (Schramke et al. 1994). These values are presented by environmental setting areas, defined within the document and shown in Figure F.2.4.4. Table F.2.4.4 presents bulk density values reported in a number of tables within Appendix B of the environmental settings data document. These data were recommended by Schramke (Schramke et al. 1994) to be used for the saturated zone at each area.

Table F.2.4.4 Bulk Density Values Recommended for the Hanford Site

Distribution Coefficient

To limit the number of modeling runs, radionuclides and nonradiologic tank constituents were grouped according to mobility (represented by the distribution coefficient [K_d]). These contaminant groups are summarized in Section F.2.2.2.

Darcy Velocity Components of the Fluid Phase

The Darcy velocity refers to the rate of groundwater flow through a cross-sectional area of a porous medium (the aquifer) in response to differences in hydraulic head. The designations of v₁ and v₂ correspond to the Darcy velocities in the x and y dimensions, respectively. For the groundwater modeling effort, the orientation can be thought of as an areal view, with the x component of flow oriented in an east to west direction and the y component of flow oriented north to south. This orientation differs from the vadose zone model, where the y component of flow represented vertical flow and the x component (representing lateral flow) was not used.

For the groundwater modeling effort, the flow was modeled as steady-state; therefore Darcy velocity remains constant over time. These values were obtained by performing a steady-state flow simulation of the unconfined aquifer using VAM2D. The values of Darcy velocity are written to an output file from the steady-state flow. This output file is subsequently used as input for the transient transport model.

Prescribed Values of Solute Flux at Boundary Nodes

Input for solute flux for the groundwater model was developed from the vadose zone model results at each source area for each alternative model. Section F.2.3.2, Post-Processing for Groundwater Model Input, discusses the development of the source term for groundwater.

F.2.4.2 Post-Processing Groundwater Results for Risk Assessment

The contaminants were grouped based on their mobility as represented by K_ds in the vadose zone and underlying unconfined aquifer. The contaminant groups were used rather than the individual mobility of each contaminant primarily because of the uncertainty involved in determining the mobility of individual contaminants. Where there was ambiguity, contaminants were placed within the more mobile group.

This approach required post-processing to determine contaminant concentrations for each constituent in the group, perform other adjustments as appropriate, and combine the results of each source area. The approach is conservative in that it will result in a somewhat higher overall estimation of concentration and mobility compared to an approach that uses each contaminant's estimated K_d values.

In summary, post-processing was performed in two phases. The first phase entailed reducing the data from multiple files (generally eight) at a 250-m (820-ft) grid spacing, into one file representing the desired 1-km (0.62-mi) grid spacing, for each time step of interest and K_d group. One file for each time of interest was input into the ARC/INFO geographic information system (GIS). An INFO program was written to enable GIS to perform the second phase of post-processing.

The second phase of post-processing for each of these values included adjusting the eight raw concentration data values for aquifer thickness, initial calculated concentration for each constituent, and radioactive decay. The eight adjusted values were then added to predict a single concentration value for each constituent within the K_d group at each 1-km (0.62-mi) grid node. The results of the ARC/INFO post-processing program were exported into Surfer format files for each constituent at each time of interest. Additional details for each phase are provided in the following subsections.

F.2.4.2.1 Reducing Data Results to 1-km (0.62-mi) Grid

The VAM2D model stores all predicted contaminant concentrations for each grid node for each time period of interest consecutively in one file. Thus, each output file would contain up to six data arrays, each representing one of the specified time periods of interest (e.g., 300, 500, 1,000, 2,500, 5,000, and 10,000 years from the model initiation time). Each data array contained 32,768 lines, 1 for each node of the groundwater model. Each line contained the x-coordinate, y-coordinate, and calculated concentration value. The output file for each source area was split into separate data files, each representing a concentration data array for one time period of interest. These matrix files were given a name representing the appropriate source area and time.

For each time period of interest, the eight data matrix files (one for each source area, which include 1WSS, 2WSS, 3WDS, 1ESS, 2ESS, 3EDS, 4ESS, and 5EDS) were combined into one file containing the x-coordinate, y-coordinate, and eight consecutive concentration values. Additionally, the data were reduced from the 250-m (820-ft) model grid spacing to the desired (1-km [0.62-mi]) grid spacing. This reduced the number of concentration values representing each source area from 32,768 to 2,173.

A FORTRAN program was written to address combining the eight files (one for each source area) into one file. The program performed the following tasks:

First, the program read in the x-coordinate and y-coordinate and calculated concentration value for the first site (1WSS) and stored these into three data arrays. Next, the program read the remaining seven files and stored only the concentration values into data arrays.
Once all of the data were stored into arrays, the program determined if there were any negative concentration values at 1-km (0.62-mi) node points. (Negative concentration values are caused by numeric dispersion within the model. This usually only occurs early in time and at the leading edge of the contaminant plume where concentration values are low.) Negative concentration values were adjusted as follows: The program looked for two nodes on either side of the negative node that were both positive concentration values and adjusted the negative value to be the average of the two. The program looked at orthogonal values (i.e., east to west, or north to south), followed by diagonal values. If the negative value could not be resolved in this manner, it was changed to its absolute value.
Next, values reported in the model below 1.0E-12 were changed to zero values. The VAM2D model numerically estimates concentrations at all node points for each timestep; consequently, even at the beginning timesteps of the model, nodes tens of kilometers away show minuscule numbers such as 1.23E-370. These numbers obviously are not valid predictions of contaminant concentrations; therefore, a determination was made as to where the concentration should be considered "zero." The value 1.0E-12 was chosen because it was more than 20 orders of magnitude less than the initial concentration.
Finally, the program wrote an output file for import into the GIS. Every fourth node on the finer grid corresponded directly to the 1-km grid spacing. Therefore, only every fourth grid point and corresponding concentration was output to the GIS file. This output file contained 2,173 lines of data, each representing a 1-km (0.62-mi) node. Each line contained the x-coordinate, y-coordinate, and eight consecutive calculated concentration values corresponding to the eight sites, respectively.

F.2.4.2.2 Adjusting Raw Data to Constituent Concentrations

Once the data were reduced to a more manageable number of values, the resulting file was imported as a raw data table into ARC/INFO for additional post-processing. Each raw data table generally contained eight values (one for each tank source area) for each 1-km (0.62-mi) grid node. The steps used for this process are described in the following text.

Step One - Adjust Raw Data for Aquifer Thickness for Each Site
Two-dimensional contaminant transport modeling results in predictions of contaminants distributed uniformly throughout the thickness of the aquifer. However, Hanford Site data indicated that the majority of contaminants are concentrated within the upper portion (approximately 6 m [20 ft]) of the aquifer. The unadjusted results from the 2-D model effectively diluted concentration predictions.

To compensate for the dilution of calculated contaminant concentrations throughout the aquifer, the model code was modified by HydroGeoLogic to track cumulative mass per unit thickness (meter) as well as the cumulative mass retained within the aquifer. From this information, the average thickness of the aquifer within the area of the contaminant plume was determined and a corresponding concentration factor was calculated. The raw value at each 1-km (0.62-mi) grid node was multiplied by the concentration factor to re-distribute the contaminant into the upper 6 m (20 ft) of the aquifer.

Step Two - Adjust Results for the Initial Concentration of Each Constituent in the K_d Group
Within any one of the K_d groups, the calculated concentration of each waste constituent at any location and time within the aquifer is scalable from the concentration used in the transport simulation.

A linear relationship exists between the unit concentration used in the transport simulation and the resulting calculated concentration in groundwater for all other constituents within any K_d group. For example, if an initial concentration of 100 g/L at the source results in a concentration of 25 g/L at a given node and time, then a contaminant with an initial concentration of 10 g/L will result in a concentration of 2.5 g/L at that node and time. This relationship allows the model results for one contaminant to be post-processed for all of the other contaminants within the K_d group.

Step two entailed adjusting the raw data values at each node by the ratio of the concentration of the contaminant of interest to the concentration modeled. This step resulted in a matrix of concentration values for each constituent for the particular K_d group. For example, the concentration simulated for the No Action alternative for the K_d group 1 (K_d = 0) was based on nitrate and set at 400 g/L. For this alternative, the initial concentration of U-238 was 70.036 g/L. Initial concentration values for each of the constituents for the various alternatives are provided in Section F.2.2.3. To predict the concentration values of U-238 (also in K_d group 1) at site 1WSS, the result at each node was adjusted by multiplying the calculated concentration by 70.036/400 = 0.17509.

Step Three - Adjust Radionuclide Constituents for Decay at the Time of Interest
The concentration of each radioisotope was then adjusted for decay for each time of interest. The relationship used for this adjustment is A(t) = A(0)e^-kt, where k = ln 2/ half-life of the radioisotope of interest in days, and t = the time of interest (days), A(t) is the decayed concentration value at time t, and A(0) is raw data concentration value prior to decay.

For example, the half-life of U-238 is 2.34E+07 years, or 8.55E+09 days. To determine the adjustment factor for decay at 27.4 years (10,000 days) the following calculations were performed:

k = ln 2 / 8.55E+09 days = 8.11E-11/day
kt = 8.11E-11/day 10,000 days = 8.11E-7
e^-kt = e^-8.11E-07 = 9.999E-01

To determine the concentration at each node, the raw data concentration would then be multiplied by 9.999E-01 to determine the final result at each node as follows, assuming an initial concentration value of 5.500 mg/L.

A(t) = 5.500 mg/L (9.999E-01)
A(t) = 5.499 mg/L

Ingrowth of progeny was not calculated.

Step Four - Combine Results for Each Constituent
The first three steps determined calculated concentrations in groundwater for discrete source areas at each 1-km (0.62-mi) node for each constituent at each time of interest. Once this information was obtained, the eight concentration values at each node (associated with each tank source) were added to provide a single calculated concentration in groundwater from all sources. This information was stored in the INFO database and exported to an ASCII text file for final processing. Each file was then run through a program to change the format of the file so that it could be read directly into Surfer for the risk assessment task. The ASCII file contained a list of the concentration values sorted by grid location. The Surfer file was required to be in the following format:

Line 1:	id (4 characters)
Line 2:	nx,ny (where nx=number of grid lines along X axis, ny=number of grid lines along Y axis)
Line 3:	xlo,xhi (where xlo=minimum x-coordinate of grid, xhi=maximum x-coordinate of grid)
Line 4:	ylo,yhi (where ylo=minimum y-coordinate of grid, yhi=maximum y-coordinate of grid)
Line 5:	grid row 1 (concentration values organized in row order)
Line 6:	grid row 2
Line 7:	grid row 3
.
.
.
Line 57:	grid row 53

NEWSLETTER

Join the GlobalSecurity.org mailing list

Weapons of Mass Destruction (WMD)

F.2.0 METHODOLOGY AND APPROACH

F.2.1 SCREENING ALTERNATIVES

F.2.1.1 No Action Alternative (Tank Waste)

F.2.1.2 Long-Term Management Alternative

F.2.1.3 In Situ Fill and Cap Alternative

F.2.1.4 In Situ Vitrification Alternative

F.2.1.5 Ex Situ Intermediate Separations Alternative

F.2.1.6 Ex Situ No Separations Alternative

F.2.1.7 Ex Situ Extensive Separations Alternative

F.2.1.8 Ex Situ/In Situ Combination 1 Alternative

F.2.1.9 Ex Situ/In Situ Combination 2 Alternative

F.2.1.10 Phased Implementation Alternative

F.2.1.11 Effluent Treatment Facility

F.2.1.12 No Action Alternative (Capsules)

F.2.1.13 Onsite Disposal Alternative

F.2.1.14 Overpack and Ship Alternative

F.2.1.15 Vitrify with Tank Waste Alternative

F.2.2 SOURCE CHARACTERIZATION

F.2.2.1 Aggregate Source Areas

F.2.2.2 Contaminant Groups

F.2.2.3 Source Terms

F.2.3 VADOSE ZONE MODELING

F.2.3.1 Vadose Zone Conceptual Models

F.2.3.2 Post-Processing for Groundwater Model Input

F.2.4 GROUNDWATER MODELING

F.2.4.1 Groundwater Conceptual Model

F.2.4.2 Post-Processing Groundwater Results for Risk Assessment