ML090990054

Independent Review of Phase 1 Decommissioning Plan for the West Valley Demonstration Project, Dated March 25, 2009 (Enclosure 2)

ML090990054
Person / Time
Site:	West Valley Demonstration Project, P00M-032
Issue date:	03/25/2009
From:	Sharon Bennett, Fakundiny R, Garrick B, Neuman S, Potter T, Whipple C ENVIRON International Corp, State of NY, Geological Survey, University of Buffalo, Univ of Arizona, Univ of California - Los Angeles
To:	NRC/FSME, State of NY, Energy Research & Development Authority
References
Download: ML090990054 (65)

Text

Enclosure #2: Independent Review of the Phase 1 DecommissioningPlanfor the West Valley DemonstrationProject,dated March 25, 2009

Independent Review of the Phase 1 SDecommissioning Plan for the West Valley Demonstration Project Prepared for

New York State Energy Research and Development Authority West Valley, New York March 25, 2009

Independent Review of the Phase 1 Decommissioning Plan for the West Valley Demonstration Project Prepared by the Independent Expert Review Team B. John Garrick, Chairman Sean J. Bennett Robert H. Fakundiny Shlomo P. Neuman Chris G. Whipple Consultant to IERT Thomas E. Potter Prepared for New York State Energy Research and Development Authority West Valley, New York March 25, 2009

ACKNOWLEDGMENT This report was facilitated by the able assistance of the New York State Research and Development Authority staff. The review team especially acknowledges the excellent support provided by Dr. Paul L. Piciulo in providing source material and guidance for the reviews.

iii

CONTENTS Section Page 1 INTRODUCTION 1 2 APPROACH TO IERT REVIEW 2 3

SUMMARY

OF THE IERT REVIEW 3 Conceptual Model for Stream Sediments 4 Conceptual Model Underlying the Calculation of DCGLs and the Effect of Erosion in the Surface Soil 5 Conceptual Model Underlying the Calculation of DCGLs for the Subsurface Soil 5 Uncertainty and Sensitivity Analysis 6 Engineered Barriers 6 Dose Modeling and DCGL Development 7 Approach to Implementing As Low As Reasonably Achievable (ALARA) Dose Criteria 8 APPENDIX A - DETAILED IERT REVIEWS A-1 APPENDIX B - QUALIFICATION SUMMARIES OF THE MEMBERS OF THE INDEPENDENT EXPERT REVIEW TEAM B-1 APPENDIX C - ACRONYMS AND ABBREVIATIONS C-1 iv

SECTION 1 INTRODUCTION The U.S. Department of Energy (DOE) has prepared a Phase 1 Decommissioning Plan for the West Valley Demonstration Project (DP) based on Phase 1 of the Phased Decision Making alternative analyzed in the draft Decommissioning Environmental Impact Statement (EIS). The DP was submitted to the U.S. Nuclear Regulatory Commission (NRC) on December 5, 2008, and copies were provided to the New York State Energy Research and Development Authority (NYSERDA) at that time. Under their Memorandum of Understanding with NRC, DOE has prime responsibility to prepare the DP; NYSERDA will provide comments to NRC for consideration in their review of and comment on the DP. NYSERDA's comments will be based in part on input received by this review of the Phase 1 DP performed by an Independent Expert Review Team (IERT) made up of independent scientists and engineers expert in the disciplines of the DP.

The West Valley Demonstration Project Act (WVDP) of 1980 directed DOE to carry out numerous activities, including the decontamination and decommissioning of the tanks, facilities, materials, and hardware used in the project in accordance with requirements prescribed by the NRC. The NRC has prescribed the requirements in its License Termination Rules in the Code of Federal Regulations 10 CFR Part 20, Subpart E, to WVDP facilities as the decommissioning goal for the entire NRC-licensed site.

The scope of the DP is limited to selected facilities on the north plateau area of the site and to removal of one major facility on the south plateau, the Radwaste Treatment System Drum Cell, a former radioactive waste storage area. The focus of the IERT review is on decommissioning activities associated with subsurface soil contamination in Waste Management Areas 1 and 2 as well as surface soils and stream sediments in relation to contaminant transport and exposure. A phased decision-making approach is the preferred alternative in the Decommissioning EIS, and this alternative is used as the decommissioning approach described in the DP.

1

SECTION 2 APPROACH TO IERT REVIEW The approach taken in the IERT review was to assess the validity and defensibility of the technical basis for the DP focusing on the derived concentration guideline levels (DCGL) and the credibility of the radiation dose calculations associated with the proposed decommissioning plan. Contributing factors to the DCGLs and the dose calculations include the treatment of critical processes such as erosion and groundwater flow and transport, disruptive events (both natural and human initiated), the performance of engineered barriers, and the representativeness of the analytical models and their results.

Obviously, there are many uncertainties associated with the actual calculations of the DCGLs and the approach to uncertainty analysis must be a part of the review of the technical basis. Understanding the technical uncertainties is vital to planning a safe cleanup of the West Valley site.

It is not the purpose of the IERT review to assess compliance with the License Termination Rule or any of the specific licensing regulations. That's NRC's expertise and responsibility. The IERT review focuses only on the soundness of the science involved as a basis for the DP. For example, central to the dose calculations is the RESRAD computer code. In particular, does the chosen initial and boundary conditions concerning radionuclide source terms and transport, input parameters, and the approach to uncertainty analysis provide a reasonable basis for the assessment of the radiation dose levels associated with the different activities of the decommissioning plan? How have the dose calculations been augmented in those cases where RESRAD may not be applicable?

A summary of the IERT review follows in Section 3. The detailed reviews of the IERT members and their consultant are provided in Appendix A.

2

SECTION 3

SUMMARY

OF THE REVIEW The Phase 1 Decommissioning Plan for the West Valley Demonstration Project is very well presented in terms of presentation style, references, and links to other studies including the applicable regulations. The issues with the DP have to do with the supporting technical analysis for the plan. While not an excuse for inadequate supporting technical analysis, it is important to acknowledge the challenges of the proposed decommissioning plan and how it impacts the required analyses. For example, the dose calculations and the derivation of DCGLs for the Phase 1 Decommissioning Plan are complicated by (1) the complexity of the geology and hydrogeology of the site, its multiple radiological sources, and its multiple radionuclides, and (2) the sequencing of the decommissioning work in phases, but with specifically bounded units of the site to provide interim isolation until all phases are complete. That is, there are conflicts between the logic of defining management unit boundaries and the logic of the phasing of the decommissioning operations. The result is that portions (sub-units) of certain management units are decommissioned in each of the two phases and engineered barriers are required to assure the necessary interim isolation of management units or sub-units decommissioned in different phases. The result is major difficulties in finding suitable computational tools or adapting available computational tools to the required technical analyses and boundary conditions, including the derivation of DCGLs.

An example of such a complication is the need to compute separate sets of DCGLs for both surface soils and subsurface soils.' This need arises because in the decommissioning of sub-units of Waste Management Area 1 (WMA 1) and Waste Management Area 2 (WMA 2) contaminant sources are presumed to exist both within the top meter or so of surface soil and in unexcavated soils (bottoms and sides of excavations) at deeper levels (subsurface soils). It should be noted that calculation of the two sets of DCGLs assumes no contaminant sources between the surface soil layer and the deep soil layer, but the set of DCGLs computed for the deep soil layer is presumed to apply as well to any residual contaminated soils between the two layers.

The set of DCGLs computed for surface soils is presumed to apply to other Waste Management Areas, where subsurface radioactive material is not expected. However, such application of this DCGL set would require data supporting the underlying premise that radionuclides are confined only to the top one meter of soil.

Furthermore, the DCGL analysis is seriously limited by constraining analysis to only one exposure scenario evaluated for each of the two sources-a resident farmer using groundwater for extensive irrigation of crops and other uses. Because dilution and hence pumping rates become important in this scenario, one can envision the need for multiple exposure scenarios to better represent reality and the possible need for lower DCGLs for 1 Separate DCGL sets may not be required. An approach for computation of a single set of DCGLs applicable to both surface and subsurface soils is described in Appendix A, Section A. 1.4.

3

some radionuclides under some conditions. Conditions of particular interest in this regard would be those involving groundwater withdrawal rates much lower than assumed in the single exposure scenario analyzed. It is usuallythe case that a resident farmer scenario, the basis for the calculations, is a-conservative approach for cleanup of a site, because most exposure pathways are included. But that may not be true in this case for some nuclides.

In addition, it is important to account correctly for the impact of overall dilution in the application of the two sets of DCGLs defined and computed separately for surface and subsurface sources. As computed, the DCGLs for one source ignore any presence in well water of any nuclides from the other source. One way to accomplish this accounting would be to compute a single set of DCGL values applicable to both surface and subsurface sources. As noted earlier, a methodology for this approach using the RESRAD code is outlined in Appendix A, Section A. 1.4.

The residential and recreational exposure scenarios in this case are recognized to be not fully independent as it is assumed that the resident farmer also occasionally fishes and makes other use of the nearby creeks. However, independent application of the derived DCGLs for surface and subsurface soils, as implied by the derivation of two separate sets of DCGLs, would erroneously assume independence of exposure scenarios related to those two sources.

In addition to decommissioning to avoid exposures to existing site dontaminants, the plan considers such site altering events as gullying as a result of surface erosion and seismic activity. As noted below and especially in Appendix A, Section A.3.1, the treatment of surface erosion is inadequate.

The limited number of exposure scenarios evaluated and the assumed characteristics of the evaluated scenarios call into question the suitability of the calculated DCGL values for surface and subsurface soils. Adaptive application of the RESRAD code and selected offline calculations conducted in a way that better suits site conditions would be useful in the computation of more suitable DCGL values. Also, extending the analysis to include the probabilistic treatment of some key parameters would provide a basis for a meaningful uncertainty analysis. However, such adaptations and extensions would only make sense if fundamental changes are made in the conceptual models, particularly regarding the treatment of such phenomena as surface erosion and groundwater contamination.

A summary by topic of IERT concerns follows with reference to the appropriate sections in the detailed reviews documented in Appendix A of this report.

CONCEPTUAL MODEL FOR STREAM SEDIMENTS (See Section A.3.2)

The stream sediment conceptual model is invalid. It ignores other, equally viable pathways of contamination to the river corridor including surface hydrology. The model assumes an unrealistic role for the perimeter sediment as the sole source term, thus 4

ignoring fluxes of contaminated pore water, and it assumes an unrealistic static condition of the river channel perimeter for extended periods of time.

CONCEPTUAL MODEL UNDERLYING THE CALCULATION OF DCGLS AND THE EFFECT OF EROSION IN THE SURFACE SOIL (See Sections A.2.1.3, A.3.1, and A.4.1.1)

The premise that contaminants only are present in the top meter of the soil is not supported by data or analysis. Therefore, the calculated surface soil DCGL values are not a sound basis for making decisions on remediation strategies.

A feature of the West Valley site critically important to the transport and release of radionuclides is surface soil erosion. Unfortunately, surface soil erosion is completely ignored, which invalidates the current conceptual model. The development or presence of gullies would markedly alter the boundary conditions adopted in the RESRAD application and thereby the dose calculations both onsite and offsite. Erosion will enhance the infiltration rate of contaminated water from surface soils to the saturated zone and the farmer's well.

The model exaggerates the extent to which contaminants originating in the surface soil are diluted in the farmer's well by groundwater, erroneously considered to be uncontaminated even where contaminated subsurface soil is taken to exist. The reasons are (1) pumping would enhance the leaching of contaminants from a subsurface soil source into groundwater and their migration toward the pumping well, (2) the disregard of the impact that planned engineered systems would likely have on flow patterns around the farmer's well and on the well's inability to continue pumping indefinitely at its design rate, which in turn would reduce the well's potential for dilution, and (3) the disregard for possible pumping of water from the Kent Recessional Sequence that may be soil.

contaminated by waters originating in surface CONCEPTUAL MODEL UNDERLYING THE CALCULATION OF DCGLS FOR THE SUBSURFACE SOIL (See Section A.4.1.2)

The computation of DCGL values for subsurface soils ignores any dose contribution from residual contamination in subsurface soils other than a limited quantity brought to the surface as cistern cuttings, which are thenassumed to be mixed with unaffected soils and distributed on the surface. This ignores potential groundwater transport of nuclides from these soils to well water. Potential transport from residual soils in the permeable zone above the Lavery Till is ignored entirely. Potential transport of nuclides through Lavery Till soils is addressed qualitatively and dismissed, but is not evaluated quantitatively. In the absence of quantitative analysis, it is not clear to what extent the qualitative arguments are valid either for the evaluated exposure scenarios or for other reasonably likely scenarios not evaluated.

The hydrologic connections between the conceptual models employed, as per RESRAD, are physically unrealistic since they do not consider coupled surface-subsurface processes 5

and resultant release scenarios. The presence of actively-eroding gullies would greatly facilitate the communication of water downward into the subsurface or upward and outward onto the ground surface, and the intersection of any gully with any buried wastes would result in the exhumation of subsurface contaminated materials by a surface-based release mechanism. These coupled physical processes are not represented in either the surface or subsurface conceptual models.

UNCERTAINTY AND SENSITIVITY ANALYSIS (See Sections A.1, A.1.1.2, and A.4.2)

Uncertainty and sensitivity analysis becomes somewhat of a moot point in the presence of incorrect conceptual models. Even had the models been more applicable, there is no evidence of any probability based uncertainty analysis.

The DCGL analysis relies on point estimates for all parameter values. It is not clear that the point estimates used appropriately bounded the results of the analysis in which case an uncertainty analysis is necessary to have confidence in the results. The only credible point estimates that are not bounds are those that derive from probability distributions.

There is no evidence that the point estimates used were derived from any such analysis and are therefore assumed to be the analysts' 'best estimates', not bounding values.

Although the analysis is supported by substantial sensitivity analysis, that analysis varies only one parameter value at a time. Consequently, it does not adequately compensate for the limited number of scenarios evaluated or for joint uncertainty in' values for a number of important parameter values. This significant weakness in the analysis could be mitigated substantially by combined use of multiple exposure scenarios and probabilistic treatment of values for some key parameters, especially very important parameters with large uncertainties, such as Kds.

The absence of an uncertainty analysis is surprising since the RESRAD code includes a probabilistic interface capable of performing dose and risk analyses. Such results would be very beneficial in guiding additional site characterization planned for Phase 1 of the decommissioning process. However, performing uncertainty analysis based on incorrect conceptual models would not produce the desired results.

ENGINEERED BARRIERS (See Sections A.1, A.4.3 and A.5.1.5)

There are concerns about the effectiveness of the engineered barriers during decommissioning. A major concern is that at the interface of the barrier bottoms and the till, groundwater could seep back into the excavations of WMA 1 and 2, become contaminated and continue to contaminate the excavation surfaces and till floor.

Backfilling the excavation has the potential to trap contaminated water and thus become a source for future plumes of groundwater contamination.

There are a number of issues with the design and scheduling of construction of the impermeable hydraulic barriers. Design details are lacking such as the thickness of the barrier for WMA 2, the method of maintaining the necessary slope and support on the 6

excavation side of the barrier wall, and the consideration of possible seismic loads and severe storms on the excavated walls.

There are also concerns with the plan and schedule for the excavations associated with the DP as well as the design of the~engineered barriers. F6r example, the WMA 2 is scheduled to be excavated before the work on WMA 1. This schedule is questioned as it would be almost impossible to excavate WMA 2 before WMA 1, because the south end of the WMA 2 hydraulic barrier requires the presence of the hydraulic barrier of WMA 1 to prevent further movement of the groundwater plume from the source area.

DOSE MODELING AND DCGL DEVELOPMENT (See Sections A.1, A.2.1.2, A.2.1.3, A.2.1.4, A.4.1.1, and A.4.1.2)

The DP provides inadequate information to support key assertions affecting the dose calculations and DCGL development. Insufficient information is provided about the conceptual models underlying the DCGL calculations, the source and validity of some key input parameters derived from ancillary models, and details concerning these models.

As discussed in Appendix A (Section A.2) and illustrated in Table 1 of Section A.2, input values that differ significantly from RESRAD default values were used in the DCGL analysis. The technical basis for these changes are poorly documented, and in some cases (especially for Kd values), generic literature values appear to have been used where site-specific values were available. The sensitivity of the results to these changes can be significant, as illustrated by the discussion of the uranium Kd.

The failure to consider that contaminants may be present below the 1 m contaminated top layer is a major weakness of the conceptual model. In addition, the RESRAD model, when run in a nondispersive flow mode, relies on the assumption that "water withdrawal introduces only a minor perturbation in the water flow." However, with a well production rate that exceeds the total recharge rate, this assumption seems unwarranted.

The fact that the calculatedDCGLs do not change if the well production rate is dropped to one tenth the assumed value is troubling. All of these taken together suggests that a more robust approach to subsurface characterization, ground water flow, and contaminant transport is needed, an analysis partly within the capability of RESRAD with adaptations.

Other issues with the dose calculation include the need for better accountability of (1) the contaminants from surface and subsurface soils simultaneously, as both reach the farmer's well at the same time, (2) a realistic assessment of possible contaminated groundwater reaching the well from the saturated zone, and (3) better recognition of the limitations of RESRAD and the need for offline calculations to obtain results for individual contaminants. While it is possible to do some adaptation of the RESRAD code to address some of these issues, no attempts were made to do so.

The DP indicates that modeling of flow around the engineered barriers in WMA 1 and WMA 2 was done using the three dimensional code STOMP. Given that this model was set up to calculate how the barriers would affect flow, and also that STOMP is a general purpose groundwater model that can calculate both flow and contaminant transport, it 7

would have been useful to use STOMP to calculate the contaminant concentrations in water in the extraction well for the case where the contaminated zone includes both the unsaturated zone and the upper layer of the saturated zone. If this calculation were made, RESRAD could be used to calculate doses from the use of the extracted water.

APPROACH TO IMPLEMENTING AS LOW AS REASONABLY ACHIEVABLE (ALARA) DOSE CRITERIA (See Sections A.2.2 and A.5.2.7)

As noted elsewhere, there are reasons to believe that the DCGLs are nonconservative.

For example, the surface soil DCGLs are derived based on the unsupported assumption that only the top meter of soil has radiological contamination. The concerns with the DCGLs notwithstanding, a reasonable ALARA analysis is presented. Analyses were performed on surface soil, the subsurface, and for creek sediments. The ALARA analyses are reasonable and credible providing the problems with the DCGLs are addressed.

8

APPENDIX A DETAILED IERT REVIEWS The approach taken in the IERT review was to assess the validity and defensibility of the technical basis for the DP focusing on' the derived concentration guideline levels (DCGL) and the credibility of the radiation dose calculations associated with the proposed decommissioning plan. Contributing factors to the DCGLs and the dose calculations include the treatment of critical processes such as erosion and groundwater flow and transport, disruptive events (both natural and human initiated), the performance of engineered barriers, and the representativeness of the analytical models and their results.

Obviously, there are many uncertainties associated with the actual calculations and part of the review of the technical basis is the approach taken to quantifying the uncertainties.

In general, the Independent Expert Review Team and their consultant had the following assignments.

B. John Garrick: Chaired the review and was responsible for integrating the results and leading the preparation of the report.

Chris G. Whipple: Reviewed overarching issues associated with the conceptual site models and the effects of RESRAD parameters on DCGL development.

Thomas E. Potter: A consultant to IERT collaborated with Chris Whipple on reviewing overarching issues, particularly those having to do with the radiation dose calculations.

Sean J. Bennett: Reviewed the role of erosion and surface transport conditions.

Shlomo P. Neuman: Reviewed groundwater flow and contaminant transport.

Robert H. Fakundiny: Reviewed geologic and tectonic issues.

A.1 REVIEW COMMENTS OF THOMAS E. POTTER A.1.1 OVERVIEW A number of factors complicate the DCGL derivation for Phase 1 of the West Valley Decommissioning Plan. The first is the complexity of the site itself, with its complex geology and hydrology, its multiple radiological sources, and its multiple nuclides. The second is the sequencing of the work in phases. While such an approach is often most logical for a complex site, establishing "bright lines" to provide interim isolation until all phases are complete is often difficult, and appears to be so in this case. For example, the logic of defining'management unit boundaries seems not entirely consistent with the logic of the phasing of the decommissioning operations. The result is that portions (subunits)

A-1

of certain management units are decommissioned in each of the two phases. In addition, engineered barriers are required to assure the necessary interim isolation of management units or subunits decommissioned in different phases. Finally, these factors complicate finding suitable computational tools or adapting available computational tools to the required DCGL analysis.

A.1.2 EVALUATION OF MULTIPLE EXPOSURE SCENARIOS Ideally, one first identifies a management unit, an area containing a single roughly uniform radiological source with boundaries between the source and the unaffected areas (portions of the unit with no radionuclides associated with operations present) beyond it defined by site characterization. One then calculates DCGLs for the affected soils. The DCGLs are evaluated by developing exposure scenarios based on reasonably likely land, water, and other resource uses and associated potential exposure pathways. For complex situations, a number of exposure scenarios may be developed to assure that DCGL minima are appropriately computed for each nuclide. Multiple scenarios may be required because one scenario may not fit all nuclides. The need for multiple scenarios or the adequacy of a single scenario can only be determined by scoping analysis. But, given the complex site and the complex mix of nuclides, analysis of multiple scenarios would seem likely to be needed for the DP.

In the proposed Phase 1 decommissioning of subunits of WMA 1 and WMA 2, sources are presumed to exist both within the top meter or so of surface soil (surface soils) and in unexcavated soils (bottoms and sides of excavations) at deeper levels (subsurface soils),

and separate DCGLs are computed for each. 2 However, only one exposure scenario is evaluated for each of the two sources-basically a resident farmer with intense irrigation by groundwater.

Intense groundwater pumping assumed for the analysis implies substantial dilution of recharge water flowing through the radiological source zones by water flowing from unaffected areas. While this might be appropriate for certain exposure scenarios, lower DCGL values might be computed for some nuclides for scenarios associated with much lower pumping rates, say a scenario in which a resident has a well for personal consumption only. In that case, recharge water flowing through the radiological source zones could constitute a substantial part of the pump flow, and, for some nuclides, a lower DCGL value could be computed.

A.1.3 SUBSURFACE SOIL DCGL The development in Section 5.2.1 of the exposure scenario for computation of DCGL values for subsurface soils, applicable only to the WMA 1 and WMA 2 subunits, seems 2 DCGL values are computed separately for surface and subsurface soils as though exposures to the two sources are decoupled. They need not be either in the exposure scenarios analyzed or in other plausible scenarios. The problem of addressing coupled exposure to multiple sources is not addressed for these two sources, although it is addressed in consideration of combined exposure to stream sediments and surface soils.

A-2

confused. Planned excavation is a complicating factor here. On one hand, the computed DCGL values are intended to apply to residual soils lower than about 1 meter below the surface left after excavation (bottom and sides of excavations, some portions below the water table, implying potential importance of groundwater pathways). On the other hand, the exposure scenario evaluated for DCGL computation considers soils between about 1 meter deep and the bottom of excavations, taken to be about 9 meters deep, entirely unaffected, presumably because of the plan to use unaffected area soils for backfill of excavations. So these intermediate residual soils are considered both affected and unaffected simultaneously. Unaffected soils do not need DCGL values. The computation of DCGL values for affected soils must include evaluation of the potential impact of those soils on dose.

The computation of DCGL values for subsurface soils ignores any dose contribution from residual subsurface soils other than a limited quantity brought to the surface as cistern cuttings, which are then assumed to be mixed with unaffected soils and distributed on the surface. This ignores potential groundwater transport of nuclides from these soils to well water. Potential transport from residual soils in the permeable zone above the Lavery Till is ignored entirely. Potential transport of nuclides through Lavery Till soils is addressed qualitatively and dismissed (See Attachment), but is not evaluated quantitatively. In the absence of quantitative analysis, it is not clear to what extent the qualitative arguments are valid either for the evaluated exposure scenarios or for other reasonably likely scenarios not evaluated.

A.1.4 RESRAD LIMITATIONS It should be noted that, to the extent that limitations in the design of the RESRAD code limited analysis of this question, as suggested in Section 5.2.4, adaptive application of the code could be helpful in analyzing the problem. A source in the saturated zone can be fairly well simulated in RESRAD by assuming a subsurface source just above the saturated zone in conjunction with use of the mass balance dilution model. (Assuming only a thin unsaturated zone with low Kd values provides little retardation and near immediate breakthrough from the contaminated zone to the saturated zone).

Incorporation of any justified dilution factor, the value of which would have to be derived outside of RESRAD, could be accommodated by judicious adjustment of infiltration flow parameter values and other parameter values. Application of RESRAD in this way would use the code as a simple calculational device rather than a more conventional simulation model. But such application could yield sound and useful results.

A similarly artificial RESRAD resident farmer run could be used to derive a single set of conservative DCGL estimates applicable to both surface and subsurface soils for any exposure scenario. This would require unrealistic input assumptions:

- a single source with an area greater than 15,000 m2 and a thickness greater than 5 m

- a very thin unsaturated zone with Kd for all nuclides less than 1

- mass balance groundwater model A-3

In such a model, Cw, the nuclide concentration in well water climbs to its maximum possible value, C, = C,/Kd, where C, is the concentration in the contaminated zone soil, at the earliest possible time, because infiltration through the contaminated zone accounts for all of the water reaching the well (due to the large area source and the mass balance model) and because nuclides are not retarded during flow through the very thin unsaturated zone. The thick source reduces the rate of source depletion, maximizing buildup of progeny nuclides in the contaminated zone, and matching as closely as possible the times of peak doses from groundwater related pathways with peak doses from pathways not related to groundwater. The large source area and thick source also maximize the doses from pathways not related to groundwater, without inflating them greatly because they far exceed the effectively infinite area and thickness.

DCGL values calculated on this basis are suitable for application to any soils in the unit, including soils in the saturated zone, provided appropriate Kd values are used. No combined exposure to surface and subsurface nuclides would lead to a dose greater than 25 millirem per year from soils containing any single nuclide (and its in-growing progeny) at concentrations less than the-DCGL so calculated.

This single run could be run in probabilistic mode, but the DCGL values would have to be computed offline from nuclide-by-nuclide dose results. The summary table DCGL (soil guideline) values are point estimates based on the standard RESRAD input, not the probabilistic input, and would, therefore, not be the results of interest.

As an alternate to running RESRAD in probabilistic mode, point estimate Kd values chosen from the low ends of the distributions (corresponding to higher-than-expected nuclide concentrations in infiltration water and well water) could be used in a standard RESRAD run, in which case, the summary results RESRAD DCGL table could be used as output.

A.1.5 UNCERTAINTY ANALYSIS The DCGL analysis relies on point estimates, primarily central tendency estimates, for all parameter values. The significance of uncertainties in the analysis is not thoroughly examined. Although the analysis is supported by substantial sensitivity analysis, that analysis varies only one parameter value at a time. Consequently, it does not adequately compensate for the limited number of scenarios evaluated or for joint uncertainty in values for a number of important parameter values. This significant weakness in the analysis could be mitigated substantially by combined use of multiple exposure scenarios and probabilistic treatment of values for some key parameters, especially very important

.parameters with large uncertainties, such as Kd.

A.1.6

SUMMARY

The limited number of scenarios evaluated and the questionable characteristics of the evaluated scenarios call into question the suitability of the calculated DCGL values for surface and subsurface soils. DCGL values for streambed sediments have not been A-4

examined here, but may also warrant reexamination. Adaptive application of the RESRAD code in a way that better suits site conditions and extension of the analysis to include probabilistic treatment of some key parameter values would be useful in the computation of more suitable DCGL values.

A.1.7 ATTACHMENT - DP DISCUSSION OF RESIDUAL SUBSURFACE CONSTITUENTS AFTER REMEDIAL ACTION A.1.7.1 From 5.2.1 Conceptual Models for DCGL Development Subsurface Soil Conceptual Model. The remainder of the contamination in the bottom of the excavation was not modeled as a continuing source to groundwater because this source is located below the assumed well pump intake depth and would not be expected to leach upward into the source of water available to the resident farmer. The potential dose contribution from this source has been determined to be small compared to the potential dose from contamination brought to the surface during installation of the hypothetical cistern. This matter is discussed further in Section 5.2.4.

A.1.7.2 From 5.2.4 Discussion of Sensitivity Analyses and Uncertainty Leaching of Residual Subsurface Contamination to Groundwater. The evaluation of DCGL radioactivity concentrations in the Lavery Till (that is, at the bottom of the WMA 1 and WMA 2 excavations) as a continuing source to groundwater could not be modeled using RESRAD, because the code does not provide for a site configuration with a source below the water table. Pore water concentrations estimated from the soil partition coefficients indicate that even assuming minimal dilution, the resulting well concentration would be low compared with the contribution from well cuttings leaching from the surface (see Appendix C). The uncertainty in neglecting this contribution to the overall dose is considered to be acceptable when considering the large percentage of the dose from pathways associated with subsurface soil cuttings spread on the surface compared to the potential dose from leaching of residual radioactivity at the bottom of the WMA 1 and WMA 2 excavations.

A.1.7.3 From Appendix C Consideration of Subsurface Lavery Till as a Continuing Source to Groundwater.

An evaluation of the potential for the Lavery Till to act as a continuing source to groundwater was conducted and' concluded the following (See Section 3.7 and Table 3-19 of the body of the plan):

• A well screened entirely in the Lavery Till could not produce enough groundwater for the resident farmer scenario.

• A well screened in both the sand and gravel unit and Lavery Till would likely pump mostly groundwater from the sand and gravel unit due to the much higher relative A-5

hydraulic conductivity and subsequent development of preferential flowpaths, and contain highly diluted contributions of contaminated groundwater from the Lavery Till.

• Advective movement from the Lavery Till to the overlying Sand and Gravel Unit is unlikely considering the vertical downward groundwater gradient.

- Diffusive movement from the Lavery Till to the Sand and Gravel Unit is unlikely considering the very low diffusion coefficients for radionuclides.

e Migration vertically upward from the till through the aquifer and into a well that is screened several meters above the till is unlikely.

A.2 REVIEW COMMENTS OF DR. CHRIS G. WHIPPLE A.2.1 CONCEPTUAL SITE MODELS A.2.1.1 Overview The conceptual site models used to determine requirements for decommissioning are based on resident farmer and recreationist scenarios. It is usually the case that a resident farmer scenario is a conservative basis for a cleanup, because most exposure pathways are included. These scenarios are not fully independent, in that it is assumed that the resident farmer also occasionally fishes and makes other use of the'nearby creeks. In addition to decommissioning to avoid exposures to existing site contaminants, the plan considers gullying and seismic events.

The RESRAD model is used to calculate derived concentration guideline levels (DCGLs) for three contaminated areas: surface soils, subsurface soils, and creek sediments.

Specific details of how these levels are calculated are discussed in the Dose Modeling and DCGL Development section of the report. This section mainly describes and reviews the conceptual models in a qualitative way. The three DCGLs are used to define what radionuclide concentrations need to be to meet the unrestricted release dose limit of 25 mrem per year.

A.2.1.2 Subsurface Soil DCGLs and Conceptual Model The subsurface DCGLs would apply in areas that would be fully excavated under the Decommissioning Plan (DP), WMA 1 and WMA 2. The conceptual model, described beginning on page 5-23, is that a cistern that is 2 m in diameter and 10 m deep is dug.

The top 9 meters of soil removed is assumed to be clean fill and the bottom 1 meter of soil is assumed to be radiologically contaminated. This soil is mixed and spread over an area of 100 m2 and with a thickness of 0.3 m. The 2 m diameter by 1 m thick cylinder is the entire source of contaminated material that is considered in the DCGL derivation.

The concentration of the 100 m2 area is one tenth that of.the contaminated subsurface soil.

A-6

In addition to the dose reduction due to the dilution of contaminated soil by clean soil the relatively small size of the contaminated area (a circular area of slightly over 11 mn diameter) affects the RESRAD calculation by allowing limited amount of food to be grown, limited dust inhalation, and limited direct exposure. For the parameters used in the modeling for precipitation (1.16 m/yr), evapotranspiration (0.55), irrigation (0.47 m/yr), and surface runoff (0.6), the resultant recharge rate through this contaminated area is 42 m 3/yr. For this analysis, a well extraction rate of 5,720 m3/yr is assumed. So in addition to the assumed dilution of the soil, the resultant leachate is diluted by a factor of 5,720/42 = 136. The combined dilution factors for soil and ground water reduce the concentration in well water by a factor of 1,360 compared to the concentration in the contaminated part of the saturated zone.

Section 5.2.4 includes a subsection titled "Leaching of Residual Subsurface Contamination to Groundwater." While the first paragraph of this section addresses residual subsurface contamination, the rest of it addresses the cistern scenario just described. This section notes:

The following conditions suggest that the dose associated with subsurface soil cuttings as a surface source does not warrant consideration in the overall combined dose assessment:

Even with conservative assumptions of a large cistern diameter and well depth, combined with a small thickness over wlhich the cuttings are spread, the result is a source area of approximately 1,000 square feet (100 square meters). When this source area is used in conjunction with the required area for a resident farmer of 100,000 square feet (10,000 square meters), the result is a large DCGL -for subsurface soil when compared with surface soil DCGLs (except in the case of Cs-137).

Dilution of contaminated well cuttings with overlying clean fill results in further reduction of overall dose from subsurface sources relative to surface sources.

Doses from potential surface soil sources are orders of magnitude greater than those from subsurface sources based on the resident farmer scenario.

In essence, this argues that the cistern scenario on which the subsurface DCGLs is based produces trivial doses and can be ignored. This makes sense, provided that some more important exposure pathways exist.

The failure to consider in situ subsurface contamination as an ongoing source of groundwater contamination is not well defended. Regarding the potential for residual subsurface contamination to migrate to the well, the DP (page 5-26) says:

A-7

The remainder of the contamination in the bottom of the excavation was not modeled as a continuing source to groundwater because this source is located below the assumed well pump intake depth and would not be expected to leach upward into the source of water available to the resident farmer. The potential dose contribution from this source has been determined to be small compared to the potential dose from contamination brought to the surface during installation of the hypothetical cistern. This matter is discussed further in Section 5.2.4.

The first paragraph of part of the Section 5.2.4 titled "Leaching of Residual Subsurface Contamination to Groundwater" (page 5-41) is quoted below.

The evaluation of DCGL radioactivity concentrations in the Lavery Till (that is, at the bottom of the WMA 1 and WMA 2 excavations) as a continuing source to groundwater could not be modeled using RESRAD, because the code does not provide for a site configuration with a source below the water table. Pore water concentrations estimated from the soil partition coefficients indicate that even assuming minimal dilution, the resulting well concentration would be low compared with the contribution from well cuttings leaching from the surface (see Appendix C). The uncertainty in neglecting this contribution to the overall dose is considered to be acceptable when considering the large percentage of the dose from pathways associated with subsurface soil cuttings spread on the surface compared to the potential dose from leaching of residual radioactivity at the bottom of the WMA 1 and WMA 2 excavations.

Given that RESRAD is not capable of modeling "a site configuration with a source below the water table," it is difficult to understand how "The potential dose contribution from this source has been determined to be small compared to the potential dose from contamination brought to the surface during installation of the hypothetical cistern." The DP does not describe how such a determination was made. It is particularly difficult to understand since they assert, as quoted above, that "the dose associated with subsurface soil cuttings as a surface source does not warrant consideration in the overall combined dose assessment."

In addition to diffusive mixing of radionuclides in the saturated zone, the assumed extraction rate of 5,720 m 3/yr exceeds the recharge rate for a nominal 10,000 m 2 area.

Based on the parameters cited above, the annual recharge, including that from irrigation, would be 4,200 in 3 . For this situation to exist in steady state, an additional 1,500 m3/yr of water is needed. This additional water can come from a combination of lateral flow and upward flow of groundwater from deeper in the saturated zone. To the extent that such deeper water flows to the well, subsurface contaminants could be mobilized.

Appendix D discusses the modeling of groundwater flow in WMA 1 and WMA 2 as affected by the proposed subsurface barriers. The details of this modeling are presented in figures on page D-8. It is not clear from the description provided what was done.

A-8

However, the STOMP model does have the capability to model flow and transport of contaminants from a source in the saturated zone. It should be possible to set up and run the STOMP model for the case of some zone of contamination remaining at the top of the unexcavated Lavery Till layer and for the well location and extraction rate used in the RESRAD runs. If this were modeled, one would have a prediction of the radionuclide concentration in the water being extracted versus time. With such results, one could then use RESRAD to calculate the doses from the relevant water-related pathways.

Summary: no analytical basis is provided for dismissing potential exposures from subsurface radionuclides via extraction from a well, and the arguments in support of ignoring the contributions of subsurface contamination to water-related exposures are unconvincing. The situation that was analyzed regarding excavation of a cistern involves the relocation of so little contaminated material that it is not interesting nor does it provide meaningful DCGLs for subsurface soils in WMA 1 and WMA 2.

A.2.1.3 Surface Soil DCGLs and Conceptual Model The surface soil DCGLs "apply to the areas of the site where deep excavation is not planned. The conceptual model used to derive these DCGLs is that, initially, the top 1 m of an area of 10,000 m3 is contaminated. All deeper soil and groundwater are pristine. As implemented in RESRAD, unit concentrations (i.e., 1 pCi/g) of relevant radionuclides are assumed to be present, and RESRAD is run with all pathways turned on except for aquatic foods and radon. The approach is consistent with how RESRAD is often used to derive DCGLs. However, this approach does not consider whether existing subsurface and groundwater contamination exists. If such contamination exists, the RESRAD analysis used to derive surface soil DCGLs will not calculate how such contaminants would affect future dose rates.

As with the subsurface DCGLs discussed above, a modeling approach could be applied to the areas of the site where major excavation is not planned. Such modeling should explore the contribution to future doses from existing contamination of subsurface soils and groundwater. Where data are available for current subsurface and groundwater concentrations, calculations should be made to determine whether dose rates exceeding 25 mrem/yr are likely to occur in the future. If and where this is the case, additional contaminant removal would likely be required. Where existing contamination would produce future doses below 25 mrem/yr, the dose rates from groundwater and from subsurface and surface soil should be considered in combination. While RESRAD cannot perform this calculation directly, multiple RESRAD runs can separately produce estimates for future dose rates versus time, and superposition can be used to estimate the total dose rate.

Summary: In areas where the plan is to use the surface soil DCGLs to determine where soil removal is needed or not, the premise that contaminants only are present in the top meter needs to be supported by data. If subsurface contaminants are present, the proposed method for determining DCGLs is not appropriate.

A-9

A.2.1.4 Comments on How RESRAD is used for the DCGL Calculations As noted inthe comments above regarding the underlying conceptual models, contributions to future doses from contaminants at depth are not properly modeled. Other aspects of how the analysis was performed are discussed here.

Basis for Parameter Selection. The input and output files for the RESRAD runs are included. A useful feature of the RESRAD input file is that it clearly indicates what parameter values were used in the run, and also what the authors of RESRAD consider to be reasonable default values. In general, site-specific values are preferable to generic values. Nonetheless, the basis for many of the parameters used that were different from the default values is not explained. The list of parameters that differ from the default values is presented in Table 1. The bases for many of these values are not explained.

Sensitivity of Uranium DCGL to Kd. Comments concerning the uranium DCGLs in Section 5.2.4 of the DP are particularly interesting:

The initial model runs produced inordinately low DCGLs for uranium radionuclides in surface soil. The calculated DCGLs for U-238, for example, was 1.0 pCi/g, slightly above measured background concentrations in surface soil shown in Table 4-11 of this plan.

The next iteration involved changes to radionuclide distribution coefficients. Evaluation of the basis for the original distribution coefficients and sensitivity analysis results led to the conclusion that some distribution coefficients used were inappropriate. These distribution coefficients were changed. The resulting distribution coefficients are based either on site specific data for the sand and gravel layer or, where site-specific data are not available, values for sand from Sheppard and Thibault 1990, as shown in Table C-2.

These model changes produced higher DCGL values for uranium radionuclides, e.g., 4.8 pCi/g for U-238. However, these values were still low compared to uranium DCGLs for unrestricted release developed at other sites. Further evaluation showed that the main reason for the low uranium DCGLs was the conservative use of the RESRAD mass balance model. After considering the results of the sensitivity analysis that evaluated use of the non-dispersion model, and RESRAD Manual guidance, it was determined to be more appropriate to use the non-dispersion model in the surface soil analysis and this was done.

A-10

Clearly a DCGL for uranium of 1 pCi/g is unreasonable, given that the average uranium concentration of US soil is about 1 pCi/g. 3 The approach taken was to cite a report (Sheppard and Thibauld) that reported literature values for Kds for various isotopes. The specific chemical conditions or form of uranium at the site was not considered. In a major report on Kds by EPA and DOE,4 it is noted "For example, uranium Kd values can vary over 6 orders of magnitude depending on the composition of the aqueous and solid phase chemistry (see Volume II, Appendix J)."

As noted in Table 1 and in the above quote, the Kd for uranium was lowered from 50 cm 3/g to 35 cm 3/g and RESRAD was switched from the mass balance mode-to the non-dispersion mode. In rerunning these cases, lowering the Kd for uranium also lowered the uranium DCGL. For the surface soil case using the RESRAD default Kd of 50 cm 3/g and running the model in mass balance mode, the calculated DCGL for U-238 is 6.86 pCi/g.

Lowering the Kd to 35 cm 3/g results in a drop in the DCGL to 4.8,pCi/g. It makes sense that reducing the Kd for uranium would also reduce the DCGL because over 98% of the dose from uranium is through water-dependant pathways (these include drinking water and food). By lowering the Kd, the uranium is more likely to partition into water and to be washed out of the surface soil. If all else were equal, this would mean that uranium concentrations in groundwater and well water would be increased in comparison to the higher Kd case.

So the explanation above - that the first thing tried to resolve an excessively low DCGL for uranium was to reduce the distribution coefficient - seems to imply that when the Kd was lowered, the DCGL increased. In fact, the reverse is true; lowering the Kd also lowers the DCGL.

The other change to the uranium analysis noted above was the switch from a mass balance model to a nondispersive model. The RESRAD User's Manual describes the difference between the two transport models:

The water transport parameters for radionuclide i are the breakthrough time Ati (the time following the radiological survey at which radionuclides first appear in the water at the point of use), the rise time 6tir (the time following the breakthrough time for the radionuclide concentration in the water to attain a maximum value), and the dilution factor fir (the ratio between the concentration in the water at the point of use to the concentration 3 T. E. Myrick, B. A. Berven and F. F. Haywood, Determinationof Concentrationsof Selected Radionuclides in Surface Soil in the U.S., Health Physics, Vol. 45, No. 3 (September), pp. 631-642, 1983.

4 US EPA Office of Air and Radiation, Understanding Variation in Partition Coefficient, Kd, Values, 2 Volumes, EPA 402-R-99-004A, August 1999.

A-11

Table 1. RESRAD Parameters for Surface Soil DCGLs, Where Different from Defaults Parameter Value Used Default Value Units*

Contaminated zone thickness 1 2 M Density of contaminated zone 1.7 1.5 g/cm_

Contaminated zone erosion rate 0 1E-3 m/yr Contaminated zone total porosity 0.36 0.4 Contaminated zone hydraulic conductivity 140 10 rn/yr Contaminated zone b parameter 1.4 5.3 Average annual windspeed 2.6 2 m/sec Evapotranspiration coeff 0.55 0.5 Precipitation 1.16 1 m/yr Irrigation 0.47 0.2 m/yr Runoff coefficient 0.6 0.2 Watershed area for nearby stream or pond 1.37E+07 1.OE+06 m2 Density of saturated zone 1.7 1.5 g/cm Saturated zone total porosity 0.36 0.4 Saturated zone effective porosity 0.25 0.2 Saturated zone hydraulic conductivity 1,400 100 m/yr Saturated zone hydraulic gradient 0.03 0.02 Water table drop rate 0 0.001 m/yr Well pump intake depth below water table 5 10 M Well pumping rate 5,720 250 mJ/yr Unsaturated zone thickness 2 4 M Distribution coefficients for Am-241, 243 1,900 20 cm /g Distribution coefficients for C-14 5 0 cml/g Distribution coefficients for Cs-137 280 4,600 cm'/g Distribution coefficients for 1-129 1 0.1 cmj/g Distribution coefficients for Np-237 2.3 -1 cm3/g Distribution coefficients for Pu (all isotopes 2,600 2,000 cm3/g Distribution coefficients for Sr-90 6.16 30 cm /g Distribution coefficients for Tc-99 0.1 0 cm /g Distribution coefficients for U (all isotopes) 35 50 cm /g Mass loading for inhalation 2.5E-05 1.OE-04 g/m3 Exposure duration 1 30 Yr Shielding factor, external gamma 0.273 0.7 Fraction of time spent indoors 0.66 0.5 Fruits, vegetables and grain consumption 178 160 kg/yr Leafy vegetable consumption 24.6 14 kg/yr Milk consumption 101 92 L/yr Meat and poultry consumption 65 63 kg/yr Soil ingestion rate 18.3 36.5 g/yr Drinking water intake 730 510 L/yr Livestock fodder intake for meat 27.3 68 kg/day Livestock fodder intake for milk 64.2 55 kg/day Mass loading for foliar deposition 4.OE-04 1.OE-04 g/m Wet weight crop yield for Non-Leafy 1.75 0.7 kg/mz Wet foliar interception fraction for leafy 0.67 0.25 Weathering removal constant for vegetation 18 20

• dimensionless unless specified A-12

in the infiltrating water as it leaves the unsaturated zone). Two models are used for calculating these parameters: a mass-balance (MB) model and a nondispersion (ND) model. In the MB model, it is assumed that all of the radionuclides released from the contaminated zone are withdrawn through the well. In the ND model, it is assumed that the dispersivity is nil, the unsaturated zone consists of one or more horizontal homogeneous strata, the saturated zone is a single homogeneous stratum, and water withdrawal introduces only a minor perturbation in the water flow. These assumptions lead to a pattern of flow lines from which the dilution factor can be estimated by geometric considerations.

The user has the option of selecting which model to use. Usually, the MB model is used for smaller contaminated areas (e.g., 1,000 m2 or less) and the ND model is used for larger areas. The breakthrough times are the same for both models; the rise times and dilution factors are different.

For the parameters set as reported in the DP, RESRAD, switching from the nondispersive mode to mass balance mode resulted in-sharp drops in the DCGLs. For the surface soil case, again using U-238 as the example and using the Kd of 35 cm 3/g as in the DP analysis, the switch to mass balance mode reduced the DCGL from 24 pCi/g to 4.8 pCi/g.

Regarding the assumption for the nondispersive case that "water withdrawal introduces only a minor perturbation in the water flow," as noted above, the assumed extraction rate (5,720 m3/yr) exceeds the recharge rate for a 10,000 m2 footprint. So it is highly likely that withdrawal introduces major perturbations in the flow field. As a check on the sensitivity of the calculation to the assumed extraction rate, the surface soil base case (which uses the nondispersive mode) was rerun with the well production rate reduced from 5,720 m 3/yr to 572 m 3/yr. There was no change to the DCGLs associated with this parameter change.

Summary: The failure to consider that contaminants may be present below the 1 m contaminated top layer is a major weakness of the conceptual model. In addition, the RESRAD model, when run in a nondispersive flow mode, relies on the assumption that "water withdrawal introduces only a minor perturbation in the water flow." However, with a well production rate that exceeds the total recharge rate, this assumption seems unwarranted. The fact that the calculated DCGLs do not change if the well production rate is dropped to one tenth the assumed value is troubling. All of these taken together suggests that a more robust approach to subsurface characterization, groundwater flow, and contaminant transport is needed, compared with what RESRAD can do.

A.2.1.5 Specific Comments

- Iodine- 129 On page 5-37, in reference to the surface soil DCGLs, a bulleted list includes "Reducing the irrigation/well pump rate increased the DCGL for 1-129 most significantly. Similarly, A-13

increasing the pump rate decreased the DCGL for 1-129. This is because reducing the pumping rate results in a lower dilution factor, and increasing the pumping rate results in more radionuclide inventory available for exposure."

The comment that "reducing the pumping rate results in a lower dilution factor" suggests that with less dilution, doses should go up and DCGLs should go down. But the above also says that the reverse happened. As the quote above states, if both irrigation and the well pumping rate are cut to one tenth of their baseline values, the 1-129 DCGL roughly doubles. If only the well pumping rate is modified from the DP value of 5,720 m 3/yr to one tenth that volume, the DCGL for 1-129 does not change. This was also the case above for U-238.

- Plutonium-241 In reference to a 200 year groundwater travel time through the Unweathered Lavery Till, the DP says (page 5-17) "Short-lived radionuclides (Sr-90, Cs-137, and Pu-241) will have decayed away during these time frames." While it is true that Pu-241, with a half-life of 14.4 years, will be essentially gone after 200 years, this statement ignores the fact that when Pu-241 decays, Am-241 (half-life 432 years) is formed. And Am-241 decays to Np-237 (half-life is essentially forever). RESRAD can calculate the ingrowth of decay products, but if the above statement suggests that Pu-241 was not included in some analysis because it is likely to decay away before the time period of interest, that would be incorrect.

- Fish Consumption Regarding the stream sediment DCGLs, the DP says (page 5-30) "The hypothetical recreationist is assumed to eat venison from deer whose flesh is contaminated with radioactivity from contaminated stream banks, such as from grazing on grass, and ingesting stream water." Are there fish in the stream, and if so, would these fish contain higher concentrations of radionuclides than deer?

- Hot Spots The DCGLEMc values are calculated based on elevated measurement concentration locations of 1 m2 . However the sampling plan to investigate a site thought to have A-14

multiple hot spots, according to MARSSINM and on Gilbert 5 is not described. As a rule, many more samples are required for sites with hot spots and where the radionuclides are not uniformly distributed. The sampling plan associated with site decommissioning is not described.

- Time at Which Future Doses Occur For each radionuclide, the RESRAD model was used to derive a DCGL based on a unit concentration of 1 pCi/g. The DCGL for each radionuclide is the concentration that would produce a 25 mrem/yr dose rate at the time when the radionuclide-specific dose rate peaks. The time at which the peak dose rate occurs is not the same for all radionuclides; the dose rate is highest at time zero for Am-241, C-14, Cm-243, Cm-244, Cs-137, Pu-238, Pu-239, Pu-240, and Sr-90, but for other radionuclides, the dose rate peaks at a later time. A sum-of-fraction approach would add the fraction of a measured radionuclide to the allowed peak concentration. Because the allowed peak concentrations do not all occur at the same time, the use of the DCGLs derived from each radionuclide's time of peak dose would be conservative. To implement the DCGLs once site data are available, RESRAD could be run with the actual surface soil concentrations. This would eliminate the use of radionuclide-specific DCGLs calculated at different times.

A.2.2 REVIEW OF THE ALARA ANALYSIS Chapter 6 of the Decommissioning Plan presents an ALARA analysis. This analysis begins with the assertion that DCGLs were derived in a conservative way (bullet on the top of page 6-4 and bullet in Section 6.2.2. No specifics are provided regarding conservative assumptions or analytical methods, other than the statement "Section 5 provides examples of this conservatism." As noted in the discussion of the conceptual models, there are good reasons to think that the DCGLs are in fact nonconservative. In particular, the surface soil DCGLs are derived based on the assumption that only the top meter of soil has radiological contamination.

5In the standard reference book "Statistical Methods for Environmental Pollution Monitoring," by Richard Gilbert, 1987, Chapter 10 is titled "Locating Hot Spots," and the first paragraph says "Chapters 4 through 9 have discussed sampling designs for estimating average concentrations or total amounts of pollutants in environmental media. Suppose, however, that the objective of sampling is not to estimate an average but to determine whether "hot spots" or highly concentrated local areas are present." The chapter concludes, "This chapter gives methods for determining grid spacing when the primary objective is to search for circular or elliptical hot spots. The grid spacings are obtained so that the consumer's risk is held to an acceptable level. ....Since grid spacing must be small to have a high probability of finding small hot spots, the cost of sampling can be high. For that reason judgment is necessary to decide in advance where hot spots are most likely to lie and to concentrate sampling in those areas. Larger grid spacing can be used in areas where hot spots are less likely to be present."

A-15

On page 6-8, Section 6.3.3, Surface Soil Preliminary ALARA Analysis, a useful analysis is presented. It is assumed that surface soil has been removed, based on the DCGLs, and that the dose rate to an individual living on the site is 25 mrem/yr. The analysis considers that marginal cost of further soil removal. For the assumptions presented (a population density of 0.0004 people per m 2,6 a 3% discount rate, a value of $2,000 per person-rem avoided, and the assumption that further soil removal would take the annual dose rate from 25 mrem to zero, it would be cost effective to remove additional soil if it could be done for $0.67 per in 2 . Additional details are provided regarding disposal costs to suggest that additional soil removal at the specific unit cost would not be possible. Given this result, there was no need to examine other effects of additional. soil removal such as transportation risks.

Similar analyses are provided for the subsurface and for creek sediments. In both cases, the analysis indicates that contaminated soil above the DCGLs is performed such that the future dose rate does not exceed 25 mrem/yr, then further soil removal would not be cost effective.

This analysis seems reasonable and credible provided the problems with the DCGL derivation are fixed.

A.3 REVIEW COMMENTS OF DR. SEAN J. BENNETT A.3.1 CONCEPTUAL MODEL FOR SURFACE SOIL DCGL DEVELOPMENT The fundamental assumption used in the Conceptual Model for Surface Soil (Figure 5-7),

and implemented in RESRAD (Yu, et al., 2001), is that the erosion rate (m/yr) of the contaminated zone (the upper 1 m of soil) is zero (i.e., the central portion of the north plateau is assumed to be generally stable over the next 1,000 years). This ensures that:

(1) no contaminated soil leaves the site via surface transport processes (erosion by overland or sheet flow or erosion by concentrated flow in rills and gullies), thus the source signal can decrease with time in situ; (2) the integrity of the contaminated soil layer and all contaminated pore waters remain intact, thus hydrologic and geomorphic boundary conditions remain constant in time and space; and (3) the hydrologic communication between surface water and subsurface water (unsaturated or saturated) remain disconnected, thus no mass transfer of water or mixing of contaminated and uncontaminated sources occur. In addition, it is noted by the preparers of the DP that this assumption of no sediment erosion is considered an appropriate simplification since it provides a conservative estimate of dose based on source depletion (no loss of source via erosion).

The assumption of zero surface soil erosion is based primarily on two lines of evidence:

(1) field observations of soil erosion rates from the north plateau and nearby environs ranging from 0 to a maximum of 0.0 12 m over an 11-year period (or 1 mm/yr; 6A footnote on page 6-9 notes that a population density of 0.0004/M2 is equivalent to 1,040 people per square mile. The actual population density in Cattaraugus County in 2000 was 64 people per square mile.

A-16

Table 3-13); and (2) numerical results obtained using the landscape evolution models SIBERIA and CHILD, as presented in Appendix F of the DEIS.

This conceptual model is invalid for the following reasons.

1. As described in the Independent Review of the DEIS (Garrick, et al., 2008), the greatest at-a-point surface erosion rates at this iocation are due to advancing gullies (0.4 to 0.7 m/y), which is one to two orders of magnitude greater than soil erosion due to overland or sheet flow. It is noted here that such gully erosion is a discrete process in time and space, that gullies are rather small topographic features (decimeters to meters in scale), and that both down-cutting and widening of the gully would occur concomitantly during headward migration.

The DEIS expanded further the possibility of gully erosion occurring at the West Valley site in Appendix G (Models for Long-term Performance Assessment) and Appendix H (Long-term Performance Assessment Results), wherein the authors introduced a simple gully erosion model. This model used the following characteristics: the gully was triangular in shape, and it had an initial advance rate of 0.4 miyr, a down-cutting rate of 0.058 miyr, and stable sideslopes of 210.

Moreover, the time-variation of advance rate was quantified using a decay function for gully length, which assumed that migration rate decreased asymptotically with time.

Gully advance rates (0.4 to 0.7 miyr) as described herein are considered to be on the very low side of published results. Rates of gully headcut advance should be a function of overland flow rate conditioned by the soil's erodibility and the characteristics of the scour pool (Alonso, et al., 2002). Because of this dependency, gully headcut advance rates can vary widely. For example, Gordon, et al. (2008) modeled ephemeral gully advance rates in Belgium, Georgia, Mississippi, and Iowa, and these simulated rates could reach as much as 200 m per year. Nachtergaele, et al. (2002) reported gully advance rates in Belgium ranging from 8 to 23 m/yr. In rill erosion studies where headcuts were observed under steady overland flow, migration rates for these features ranged from about 0.1 to 2.0 mmi/s depending on headcut height and flow rate (Bryan and Poesen, 1989; Bryan, 1990; Slattery and Bryan, 1992; Bennett, 1999; Bennett, et al.,

2000). Depending on overland flow rates and durations, rill and gully headcut advance rates of several meters per year would not be atypical for the West Valley site and nearby environs. That is, it is highly likely that gully advance rates could be significantly greater than the 0.4 m/yr reported here.

At present, more than 20 major and moderate-sized gullies have been identified in this area, as shown in Figure F-5 of the DEIS, and the DP acknowledges that "with unmitigated erosion, gullies could eventually extend into the areas of Lagoons 1, 2, and 3 during the 1,000-year evaluation period" (p. 5-13).

A-17

With regard to the erosion results using landscape evolution models, as described in detail elsewhere (Garrick, et al., 2008), several highly critical comments were made on the parameters used, the disconnect between model parameterization and onsite hydrologic and geomorphic characteristics, the complete lack of verification of model components or comparisons with onsite field data, and the calibration scheme used, among others. It was the opinion then that the predictions from these models could not be accepted or ratified.

While the purpose of the current discussion is not to debate the gully erosion or landscape evolution models above or that gully erosion remains the primary surface erosion threat to West Valley site, the point here isthat this widely recognized and universally accepted surface soil erosion process is completely ignored in the conceptual model (Figure 5-7) as implemented in RESRAD, and that the rate of gully erosion may be significantly underestimated.

2. The authors do not recognize the role of seepage (exfiltration) on gully erosion initiation and upstream migration, and its potential effects on local hydrology.

Evidence of surface seepage processes at the West Valley Site is pervasive, and exfiltration has been shown to cause, catalyze, and significantly enhance headcut erosion and gully development in cohesive sediment (Huang and Laflen, 1996; Fox, et al., 2007).

It is likely that the presence of gullies, formed by seepage oi concentrated flow erosion, would greatly facilitate the communication of water either downward into the subsurface (enhanced infiltration or recharge) or upward and outward onto the ground surface (enhanced exfiltration). Such hydrologic alteration due to gully development would greatly modify both the boundary conditions implicit in the Conceptual Model for Surface Soil and the connectivity of the surface and subsurface hydrologic regimes.

In summary, the Conceptual Model for Surface Soil DCGL Development is invalid because it completely ignores surface soil erosion processes, specifically gully erosion, as an important mechanism for the release, of the radionuclides to the environment. Gully erosion remains the primary surface erosion threat to the West Valley site, a point widely recognized and universally accepted amongst all agencies and consultants involved in the assessment of the site. Yet this surface soil erosion process is completely ignored in the current conceptual model. Moreover, the development or presence of gullies would markedly alter the boundary conditions adopted in this conceptual model and employed in RESRAD, and would markedly alter dose calculations both onsite and offsite.

A.3.2 STREAMBED SEDIMENT CONCEPTUAL MODEL In this conceptual model, a 333-m long stream channel reach is contaminated by radionuclides adsorbed to sediments to a depth of 1 m along the channel perimeter (Figure 5-9). According to RESRAD (p. E-26; Yu, et al., 2001), the surface water pathway consists of an onsite groundwater pathway segment that extends from the edge A-18

of the contaminated zone to a location where surface seepage occurs (i.e., stream recharge by contaminated groundwater; see Figure 2-10, Yu, et al., 2001). It is further assumed that (a) the sediment within the stream channel perimeter retains, sequesters, or adsorbs these contaminants, (b) radiation is then released to the stream channel via determination of breakthrough times and dilution factors, and (c) the contaminated sediment remains in place during the requisite time period. That is, sediment is neither eroded from nor deposited into this stream channel bed (this physical process is currently unavailable within RESRAD). Thus, the contamination source term degrades with time in situ.

Subsurface radioactive materials can enter stream channels in a number of ways, which include:

a. release and transport of aqueous (dissolved) radionuclides via surface processes (overland flow, gullies, stream flow, etc.);

b. release and transport of radionuclides adsorbed to silt- and clay-sized sediments or particulate organic matter via surface processes;

c. discharge of aqueous radionuclides directly into the stream channel via exfiltration (e.g., seepage from stream banks); and

d. discharge of aqueous radionuclides directly into the stream via groundwater recharge.

The transport of mass and solutes through stream systems over time at-a-point typically is modeled using a one-dimensional mass continuity equation for non-reactive solutes that includes downstream advection, vertical diffusion and mixing, and transient storage (Bencala and Walters, 1983). A similar approach would be adopted for the transport of suspended sediment with sorbed radionuclides (Julien, 1998).

This conceptual model is invalid for the following reasons.

1. At present, RESRAD can only address mechanism "d." above. There is no justification provided to consider only a single source of contamination to the stream channels, as it would appear that other pathways are equally justifiable (i.e., surface transport and exfiltration).

2. During recharge of contaminated groundwater to the stream channel, it is not clear if the perimeter sediment retards or completely eliminates transfer of aqueous radionuclides to the stream flow directly, or if the contaminated groundwater can recharge the stream directly. It would appear that no direct hydrologic communication exists; the perimeter sediment becomes the source term rather than the contaminated pore waters in transport, and radiation is then transferred to the stream channel via these breakthrough times and dilution factors.

3. It is unrealistic to assume that the perimeter of a river channel remains completely' static for any extended period of time, simply by consideration of the temporal variations of river channel hydraulics, bank stability criteria, and a sediment continuity equation (e.g., Parker, et al., 2000; Langendoen and Simon, 2008).

A-19

Assuming that the concept of dominant discharge (channel-forming discharge) is applicable in these stream systems (e.g., Wolman and Miller, 1960), one can assume significant (decimeter-scale) erosion and deposition within any natural stream corridor to occur every 1 to 3 years. As such, release of radionuclides to the environment would be punctuated by relatively high doses discrete in time, rather than attenuated (ever-diminishing), continuous doses as employed here.

In summary, the Stream Sediment Conceptual Model is invalid because it ignores other, equally viable pathways of contamination to the river corridor including surface hydrology, it assumes an unrealistic role for the perimeter sediment as the sole source term and thus ignoring fluxes of contaminated pore water, and it assumes an unrealistic static condition of the river channel perimeter for extended periods of time.

A.

3.3 REFERENCES

Alonso, C.V., S.J. Bennett, and O.R. Stein, Predicting headcut erosion and migration in concentrated flows typical of upland areas, Water Resources Research, 38, 39-1 to 39-15, 2002.

Bencala, K.E., and R.A. Walters, Simulation of solute transport in a mountain pool-and-riffle stream - a transient storage model, Water Resources Research 19, 718-724, 1983.

Bennett, S. J., Effect of slope on headcut growth and migration in upland concentrated flows, Geomorphology, 30, 273-290, 1999.

Bennett, S.J., C.V. Alonso, S.N. Prasad, and M.J.M. Rbmkens, Experiments on headcut growth and migration in concentrated flows typical of upland areas, Water Resources Research, 36, 1911-1922, 2000.

Bryan, R.B., Knickpoint evolution in rillwash. In Soil Erosion-Experiments and Models, edited by R.B. Bryan, Catena Supplement 17, pp. 111-132, 1990.

Bryan, R.B., and J. Poesen, Laboratory experiments on the influence of slope length on runoff, percolation, and rill development, Earth Surface Processes and Landforms, 14, 211-231, 1989.

Fox, G.A., G.V. Wilson, A. Simon, E.J. Langendoen, 0. Akay, and J.W. Fuchs, Measuring streambank erosion due to ground water seepage: correlation to bank pore water pressure, precipitation and stream stage, Earth Surface Processes and Landforms, 32, 1558-1573, 2007.

Garrick, B.J., J.T. Bell, S.J. Bennett, R.H. Fakundiny, S.P. Neuman, F.L. Parker, M.T.

Ryan, P.N. Swift, C.G. Whipple, M.P. Wilson, 2008, Independent Review of the Draft Environmental Impact Statement for Decommissioning and/or Long-Term Stewardship at the West Valley Demonstration Project and Western New York Nuclear Service Center, A-20

prepared for New York State Energy Research and Development Authority West Valley, New York, September 23, 2008, 164 pp.

Gordon, L.M., S.J. Bennett, C.V. Alonso, and R.L. Bingner, Modeling long-term soil losses on agricultural fields due to ephemeral gully erosion, Journal of Soil and Water Conservation, 63, 173-181, 2008.

Huang, C.-H., and J.M. Laflen, Seepage and soil erosion for a clay loam soil, Soil Science Society of America Journal, 60, 408-416, 1996.

Julien, P.Y., Erosion and Sedimentation, Cambridge University Press, New York, 1998.

Langendoen, E.J., and A. Simon, Modeling the evolution of incised streams. II:

Streambank Erosion, Journal of Hydraulic Engineering, 134, 905-915, 2008.

Nachtergaele, J., J. Poesen, A. Sidorchuk, and D. Torri, Prediction of concentrated flow width in ephemeral gully channels, Hydrological Processes, 16, 1935-1953, 2002.

Parker, G., C. Paola, and S. Leclair, Probabilistic Exner sediment continuity equation for mixtures with no active layer, Journal of Hydraulic Engineering, Vol. 126, No. 11, November, 2000.

Slattery, M.C., and R.B. Bryan, Hydraulic conditions for rill incision under simulated rainfall: a laboratory experiment, Earth Surface Processes and Landforms, 17, 127-146, 1992.

Wolman M.G., and J.P. Miller, Magnitude and frequency of geomorphic processes, Journal of Geology 68, 57-74, 1960.

Yu, C., A.J. Zielen, J.-J. Cheng, D.J. LePoire, E. Gnanapragasam, S. Kamboj, J. Amish, A. Wallo III, W.A. Williams, and H. Peterson, User's Manual for RESRAD Version 6, ANLIEAD-4, prepared by Argonne National Laboratory, 9700 South Cass Avenue, Argonne, IL, 60439, 2001.

A.4 REVIEW COMMENTS OF DR. SHLOMO P. NEUMAN A.4.1 CONCEPTUAL SITE MODELS UNDERLYING DCGL DEVELOPMENT A.4.1.1 Surface Soil (SS)

DCGLs for Surface Soil are computed by means of the non-dispersion (ND) RESRAD model (pp. 5-20 through 23). According to this model, all contaminated water reaching a well in the sand and gravel aquifer is derived from surface soil by downward percolation through the unsaturated zone. In the well, this water mixes with uncontaminated water drawn by the well laterally from the saturated portion of the aquifer, bringing about A-21

dilution. The well can supply water to the farm indefinitely, and its operation has no impact on the ambient groundwater flow regime, which takes place in a single horizontal direction.

The conceptual model underlying the calculations of DCGLs for SS is inadequate for the following reasons:

a) The model ignores groundwater contamination in the saturated zone by virtue of its contact with contaminated Subsurface Soil (SB), regardless of whether or not such contamination is present. The DP does not distinguish clearly and convincingly between areas in which contaminated SS is present with and without contaminated SD. While areas over the Sr-90 groundwater plume with and without surface contamination may exist, DOE has specifically excluded these areas from the scope of the DP. Definitive data are not presented to indicate that there may be other areas of the Project premises at which contaminated SS is present while contaminated SB is absent.

Where contaminated SB exists, it is dismissed in the DP (p. 5-26): "... contamination in the bottom of the excavation was not modeled as a continuing source to groundwater because this source is located below the assumed well pump intake depth and would not be expected to leach upward into the source of water available to the resident farmer. The potential dose contribution from this source has been determined to be small compared to the potential dose from contamination brought to the surface during installation of the hypothetical cistern" in the SB scenario.

One reason for ignoring the impact of SB-related groundwater contamination on DCGLs under the SS scenario is the inability of RESRAD to model this impact: (p. 5-

41) "The evaluation of DCGL radioactivity concentrations in the Lavery Till (that is, at the bottom of the WMA 1 and WMA 2 excavations) as a continuing source to groundwater could not be modeled using RESRAD, because the code does not provide for a site configuration with a source below the water table."

This is unfortunate because in reality, the farmer's pumping well would draw water toward it from all depths within the aquifer, including from contaminated SB at the aquifer bottom, and from all directions.

The argument that "this source is located below the assumed well pump intake depth and would not be expected to leach upward into the source of water available to the resident farmer" is invalid: Regardless of where the intake of the pumping well is located, it would generate a hydraulic gradient drawing water toward the well from all depths (including the aquifer bottom and the potentially contaminated shallow Lavery Till) and from all lateral directions.

The argument that "the potential dose contribution from this source has been determined to be small compared to the potential dose from contamination brought to the surface during installation of the hypothetical cistern" is based on equilibrium A-22

partitioning of contaminants between the SB source and groundwater. This argument fails to consider that the leaching of contaminants from the SB into groundwater is affected by pumping, which causes their concentration in groundwater to increase toward the well. That such increase must in fact take place is demonstrated mathematically for the SB scenario in our Attachment. By disregarding this SB-derived groundwater concentration in the SS scenario, the DP exaggerates the extent to which contaminants originating in SS are diluted (by supposedly uncontaminated groundwater) in the farmer's well.

b) Even if and where contaminated SB is not present, the analysis of SS-derived contaminant dilution in the farmer's well is still problematic.

The analysis of dilution is one-dimensional, based on average groundwater flow conditions prior to pumping; the effect of pumping on groundwater flow patterns near the well is disregarded. There is no a priorireason to assume that pumping by the farmer would not significantly alter the groundwater flow regime around the well. To demonstrate that the well does not significantly impact this flow regime would require a three-dimensional groundwater flow model that accounts adequately for pumping on and off the ariea being analyzed jointly with the impact of other engineered site features. No such modeling is reported in the DP.

Disregarding the impact of pumping on groundwater flow patterns around the farmer's well is especially problematic considering that flow toward the well is (1) impeded by hydraulic barriers such as those planned for WMA 1 and 2, and (2) modified by a French Drain planned for WMA 1 and by a dewatering system planned for WMA 3. Impedance of flow toward the farmer's well due to hydraulic barriers is expected to reduce the flow of groundwater from uncontaminated portions of the sand and gravel aquifer toward the well, thereby reducing dilution. If and where contaminated SB is present, such impedance is expected to increase the flow of SB-contaminated groundwater from the bottom of the sand and gravel aquifer and/or the top of the Lavery Till, thereby further reducing dilution and increasing contaminant concentration in the well, as explained in a). No modeling of these three-dimensional effects on the flow and contaminant transport regime around the farmer's well is reported in the DP, 'which fails to consider the problem.

Contrary to assumption, there is no guarantee that the farmer's well would be able to continue pumping indefinitely at its design rate, given impedance of flow toward it by the hydraulic barriers. This too remains unrecognized in the DP.

c) Possible pumping of contaminated water from the Kent Recessional Sequence (KRS) is not considered. Such contamination may originate in part in Surface Soil.

A.4.1.2 Subsurface Soil (SB)

DCGLs for Subsurface Soil are computed (pp. 5-23 through 28) by assuming that (5-23)

"exposure to the subsurface radioactivity occurs following intrusion and surface dispersal A-23

when installing a water collection cistern." p. 5-25: "The exposure occurs when the subsurface radioactivity is deposited on the ground surface where it can result in exposure to members of the critical group through various pathways." p. 5-26: "For conservatism the hypothetical well is assumed to have a large diameter."

As already stated, "The remainder of the contamination in the bottom of the excavation was not modeled as a continuing source to groundwater because this source is located below the assumed well pump intake depth and would not be expected to leach upward into the source of water available to the resident farmer. The potential dose contribution from this source has been determined to be small compared to the potential dose from contamination brought to the surface during installation of the hypothetical cistern."

The conceptual model underlying the calculations of DCGLs for SB is invalid for the following reasons:

a) The model ignores groundwater contamination in the saturated zone by virtue of its contact with contaminated Subsurface Soil.

One reason for ignoring the impact of SB-related groundWater contamination on DCGLs under the SB scenario is the inability of RESRAD to model this impact: (p.

5-41) "The evaluation of DCGL radioactivity concentrations in the Lavery till (that is, at the bottom of the WMA 1 and WMA 2 excavations) as a continuing source to groundwater could not be modeled using RESRAD, because the code does not provide for a site configuration with a source below the water table."

This is unfortunate because in reality, the farmer's pumping well would draw water toward it from all depths within the aquifer, including from contaminated SB at the aquifer bottom, and from all directions.

The argument that "this source is located below the assumed well pump intake depth and would not be expected to leach upward into the source of water available to the resident farmer" is invalid: Regardless of where the intake of the pumping well is, located, it would generate a hydraulic gradient drawing water toward the well from all depths (including the aquifer bottom and the potentially contaminated shallow Lavery Till) and from all lateral directions.

The argument that "the potential dose contribution from this source has been determined to be small compared to the potential dose from contamination brought to the surface during installation of the hypothetical cistern" is based on equilibrium partitioning of contaminants between the SB source and groundwater. This argument fails to consider that the leaching of contaminants from the SB into groundwater is affected by pumping, which Causes their concentration in groundwater to increase toward the well. That such increase must in fact take place in the considered SB scenario is demonstrated mathematically in our Attachment. By disregarding this SB-derived groundwater concentration in the SB scenario, the DP underestimates the overall contribution of SB-derived contamination under this scenario.

A-24

b) Possible pumping of contaminated water from the KRS is not considered. Such contamination may originate in part in Subsurface Soil.

A.4.2 SENSITIVITY AND UNCERTAINTY ANALYSES The DP presents a limited number of sensitivity analyses which it considers to represent uncertainty assessment. In fact, the DP does not present any probabilistic or statistical analyses of results. This is puzzling considering that according to the RESRAD manual (p. xi) the "code can perform uncertainty/probabilistic analyses with an improved probabilistic interface. It uses a preprocessor and a postprocessor to perform probabilistic dose and risk analyses ... The results of an uncertainty analysis can be used as a basis for determining the cost-effectiveness of obtaining additional information or data on input parameters (variables)." The latter could be used to guide additional site characterization planned, according to the DP, for Phase I of the decommissioning process.

A.4.3 ADDITIONAL OBSERVATIONS

1. Title Page box: "The proposed decommissioning approach described in this plan is based on the preferred alternative in the Revised Draft Environmental Impact Statement for Decommissioning and/or Long-Term Stewardship at the West Valley Demonstration Project and Western New York Nuclear Service Center, which is referred to as the Decommissioning EIS."

Hence some of the issues the IERT identified in the context of the DEIS may be relevant to the DP.

As pointed out in Appendix A of the IERT report, long-term impacts of the Phased Decisionmaking Alternative have not been subjected to performance assessment (PA) in the DEIS. It is therefore difficult to see how the results of the PA in Section 2.6 of the DEIS could define sharply the issues and provide a clear basis for choice among the alternatives, as stated. The Phased Decisionmaking Alternative has been preselected as the preferred alternative, rendering any such comparison mute. The rationale given in support of this preferred alternative in Section 2.7 does not rely in any way on long-term PA. Considering further that only two alternatives are analyzed in the DEIS (No Action and Close-in-Place, p. E-55), the IERT asks how will this allow assessing the preferred phased alternative?

2. Executive Summary: "The environmental impacts of the proposed approach described in this plan are being analyzed in the Environmental Impact Statement on Decommissioning and/or Long-Term Stewardship of the WVDP and Western New York Nuclear Service Center, hereafter referred to as the Decommissioning EIS."

In fact, such impacts were not analyzed for the Phased Decision-Making Approach in the DEIS.

A-25

3. p. 3-12: "PermeableTreatment Wall and PermeableReactive Barrier.According to the DP, a full-scale passive permeable treatment wall and a permeable reactive barrier are expected to be installed before Phase 1 of the decommissioning to mitigate the offsite migration of Sr-90 contaminated groundwater in the sand and gravel unit in the north plateau. According to NYSERDA the DOE recently announced that they will not install the permeable reactive barrier in the ditch on the north plateau. It is not known if there will be a substitution.

The permeable treatment wall will be located in WMA 2 immediately south of the Construction Demolition and Debris Landfill in WMA 4 approximately perpendicular to the flow path of the north plateau groundwater plume. It will be approximately 400 feet long in a northwest-southeast direction with two 50-foot long lateral sections extending off of each end of the 400-foot long section to the west and south. The permeable treatment wall will be two to four feet thick, extend down into the underlying Unweathered Lavery Till, and composed of granular zeolite to reduce Sr-90 concentrations in groundwater through ion-exchange.

The last bullet of the 1ERT findings under the section dealing with groundwater flow and transport reads:

A permeable treatment wall and a permeable reactive barrier are planned to mitigate further North Plateau Groundwater Plume Migration. No discussion or analysis of the design or effectiveness of such a wall and barrier are provided in the 2008 PDEIS. 7 Given mixed.results 8 with the existing passive treatment system at the site, it is not clear to us that the planned wall and barrier can be counted on to perform as assumed in the 2008 PDEIS.

The same finding applies to the DP.

4. p. 3-70: The precise source of the hydraulic conductivity values in Table 3-19 is unclear; where in the DEIS do these numbers appear? Values cited for the Unweathered Lavery Till appear to be too low.

5. p. 3-7 1: Not clear what is the source of information on Water Budget within the Unsaturated Zone on this page.

6. p. 3-72: Section 3.7.5 Description of Unsaturated Zone Monitoring Stations does not provide any such description.

Garrick, B.J., et al., Independent Review of the Draft Environmental Impact Statement for Decommissioning and/or Long-Term Stewardship at the West Valley Demonstration Project and Western New York Service Center, September 23, 2008.

8 West Valley Nuclear Services Company, Inc., Supplemental Hydrogeologic Investigation of the North Plateau Pilot Permeable Treatment Wall: Performance Assessment and Evaluation of Potential Enhancements. Nov. 2002 A-26

7. pp. 3-72 and 3-73: Section 3.7.7 Numerical Analysis Technique provides no context or reference; what model(s) is it referring to? What is the purpose of these models?

How are they related to the DP?

8. p. 3-73: What is the evidence for the assertion that "The Lavery till ... does not absorb significant amounts of Sr-90?"

9. p. 3-73: "The Kd is defined as the ratio of the concentration (or activity in the case of radionuclides) of a species sorbed on the soil, divided by its concentration (or activity) in solution under steady-state conditions."

The definition of Kd is not restricted to steady state flow or transport; perhaps the writer meant to say thermodynamic equilibrium?

10. p. 5-1: Section 5.2 is said to describe the three conceptual models and the mathematical model (RESRAD) used to develop DCGLs. In fact, I found it impossible to understand any of these models without referring to the RESRAD manual.

11. pp. 5-16 through 17: Potential Impact on the Kent Recessional Sequence is assessed on the basis of work published during or before 1993. Why is the DP not relying on DOE models developed as part of the DEIS?

I am not convinced, based on this discussion, that "The potential for impacts on groundwater in the Kent Recessional Sequence from any residual radioactivity that might remain in the bottom of the WMA 1 and WMA 2 excavated areas has been evaluated and found to be very low."

12. p. 5-19: DCGLs have been developed for 18 radionuclides. "These radionuclides were selected based on screening of simplified groundwater release and intrusion scenarios for north and south plateau facilities. The screening indicated that other radionuclides would in combination contribute less than one per cent of potential dose impacts at the individual facility."

No information is given about this screening process.

13. Figures 5-7 and 5-8 show shale bedrock under the Lavery Till; in fact, the latter is underlain by the Kent Recessional Sequence and other glacial outwash units.

Correspondingly, the DP fails to consider the possibilities that (a) the KRS could serve as a potential source of water for a resident farmer, and (b) contaminants could reach the KRS by vertical migration through the Lavery Till.

14. p. 5-42: The bottom half of this page suggests that both the model and its parameters were manipulated to obtain desired DCGLs for uranium.

A-27

15. p. 5-43: The Overall Conclusions are suspect because they were derived on the basis of invalid conceptual models.

A.4.4 ATTACHMENT - PROOF THAT CONTINUOUS SOURCE OF CONTAMINATION IN SATURATED ZONE CAUSES CONCENTRATION TO INCREASE TOWARD PUMPING WELL The following is a mathematical proof that if Subsurface Soil (SB) in the saturated zone acts as a continuous source of groundwater contamination, and one disregards dispersion (as in the No-Dispersion model of RESRAD) then, under steady state flow and transport in the absence of groundwater recharge, contaminant concentrations increase monotonically toward the pumping well.

Consider a well screened throughout the saturated thickness of the sand and gravel aquifer pumping groundwater at a constant volumetric rate Q. Assume for simplicity that the aquifer has uniform isotropic hydraulic properties and a constant saturated thickness

b. Then under steady state flow in the absence of groundwater recharge the velocity vr of groundwater toward the well is given by Q

2,rrbo where 0 is porosity. Consider a cylinder of infinitesimil thickness dr having height b, inner radius r and outer radius r + dr. The time rate of mass outflow from the cylinder across its inner wall is J, = 2Zrrbvrcr =-QCr* (2) and the time rate of mass inflow across the outer wall of the cylinder is Jr+dr = 2)(r +dr)bVr+drCr+dr Cr+dr Cr + 3r dr) (3)

The last term in parentheses represents the leading terms of an expansion of Cr+dr in a Taylor series about cr. Designate byf the time rate at which contaminated soil within the saturated zone leaches contaminant mass into groundwater per unit horizontal area. Then the rate at which contaminant mass leaches into the cylinder is f 2;rrdr. (4)

The rate at which contaminant mass accumulates inside the cylinder is given by A-28

2rrbdraCr (5) at Mass balance in the cylinder requires that Rate of mass inflow - Rate of mass outflow + Leaching rate = Rate of accumulation.

Mathematically this translates into Q cr + Cr dr-Qcr + f2Zrrdr= 27crbdr r*c (6)

Dividing throughout by 2nTrbdr and simplifying yields Q aCr +f - aCr (6) 2zrrob ar b at Under steady state transport the time derivative on the right hand side vanishes, Q cr+f =0. (7) 2)rrqrb ar b Since the leaching ratef from SB is positive (f > 0), (7) implies that dCr 0 (8) ar In other words, the concentration cr of contaminants must decrease monotonically with radial distance r from the pumping well or, equivalently, must increase monotonically toward the well. Q.E.D.

A.5 REVIEW COMMENTS OF DR. ROBERT H. FAKUNDINY Provided below are comments on the Phase 1 Decommissioning Plan for the West Valley Demonstration Project (DP) (Revision 0).

A.5.1

GENERAL COMMENT

S The DP has a wide spectrum of quality. Some sections seem to address adequately the items in Appendix A, Decommissioning Plan Annotated Checklist, while others do not.

A.5.1.1 The Executive Summary does a fair job of guiding the reader through the DP.

Some improvements could be made, however, such as referring each discussed item to the relevant sections that contain the details of that item.

A-29

A.5.1.2 The Introduction does a fair job of providing some of the background for the DP. Some questions remain, however.

A.5.1.3 Section 2, Facility Operating History, is mostly beyond my expertise, but I have a few questions and comments.

A.5.1.4 Section 3, Facility Description, appears to be a group of discussions put together from various convenient sources to satisfy the requirements of the Decommissioning Plan Annotated Checklist of Appendix A. Its quality, therefore, varies from excellent to poor, depending probably on the sources.

Other than the questions and comments given in the attachment, several items are of concern. The geological descriptions do not consider that the Olean Till may be present in the subsurface and possibly in the shallow subsurface near the bedrock walls of the valley. Also, the bedrock is fractured and weathered in many places below the Lavery Till. These factors could be relevant to groundwater modeling. Erosion rates are presented in a table that shows many diverse numbers, some of which are suspect, but are not discussed as to how they will be used in erosion-rate modeling. The regional structure and tectonics is a hodge podge of ideas and speculations. The topic could have been satisfied mostly by a more complete discussion of the Clarendon-Linden fault system (CLFS) with its important southwest branch.

The section on seismology does not have a focus other than to satisfy the checklist, which does not include a requirement for a discussion of a seismic-hazard analysis. It uses Wong, et al. (2004), which relied upon the EPRI-SOG exercise of 1989 (Risk Engineering, Inc., 2007). The evaluation of seismic hazards presented in the DP does not consider advancements in seismic engineering analyses and modeling and should not, therefore, be used to support any future safety decisions regarding earthquake effects at the Western New York Nuclear Service Center (WNYNSC) site and to the future integrity of the radioactive wastes onsite.

The DP does not present a probabilistic seismic-hazard analysis (PSHA). It depends instead on Wong, et al. (2004) for PSHA. The DP only presents some peak horizontal ground accelerations and some hazard curves that are taken from Wong, et al. (2004). A combined probabilistic and deterministic seismic-hazard analysis along with a seismic-risk analysis is the most appropriate approach to assuring the safety of the site once it has been decommissioned (Krinitsky, 1998). The approach of Wong, et al. (2004) relies heavily upon the technique of the EPRI-SOG exercise performed in the late 1980s (EPRI, 1989) that tried to establish a connection between seismic-source zones and tectonic structures with local seismicity, which could then be used in hazard analyses for the northeastern United States (Risk Engineering, 2007). This approach is questionable, because no such associations have been made between seismic-source zones and earthquakes and few between tectonic structures and earthquakes for the northeastern United States (Ebel and Tuttle, 2002; Fakundiny, 2003). Wong, et al. (2004) use, instead of the EPRI-SOG results, Adams and Halchuck (2003) without justification or A-30

explanation for their source zones. Adams and Halchuck (2003) place the WNYNSC in their eastern Great Lakes seismic-source zone, a zone not,recognized by the eight teams of the EPRI-SOG exercise. Most of the members of the teams in the EPRI-SOG exercise, and Adams and Halchuk are not tectonic specialists, but seismologists whose designations of seismic-source zones are not justified by cited references, nor by the known geology of the northeastern United States. Wong, et al. (2004) seismic-source zones and little, if any, tectonic sense (see Fakundiny' and Pomeroy, 2002, for another interpretation). The analysis by Wong, et al. (2004) does not place a maximum credible earthquake at the southwestern end of the southwest branch of the Clarendon-Linden fault system, even though they provide some detail about the new finding of this branch (Bay Geophysical, 2001). Doing so would be more conservative and justifiable, because the few data we have available on recurrence rates and locations of earthquakes for the Clarendon-Linden fault system is from the work of Tuttle, et al. (2002). Wong, et al.

(2004) apply a number of steps to their seismic-hazard-analysis logic tree that are questioned by many engineering geologists (Krinitsky, 1998). These include: (1) the b-value approach; (2) paleoseismic studies, such as those of Tuttle, et al. (2002); (3) logic trees themselves; (4) slip rates on faults (Wong, et al., 2004); and (5) using expert opinions to derive seismic-source zones (the EPRI-SOG approach) (Risk Engineering, Inc. (2007a), among others. Krinitsky states "Engineering design must be done deterministically if one is to have seismic safety coupled with good engineering judgement. However, there is a need for probability.. .Probability is needed to obtain operating basis earthquakes, to perform risk analyses to prioritize projects, and for assigning recurrence estimates to deterministic earthquakes.. .The design for critical structures.. .such as hazardous waste repositories.. .must be based on maximum credible earthquakes, obtained by deterministic procedures ...(Krinitsky, 1998). Krinitsky is critical also of Gaussian smoothing, which is used by Wong, et al. (2004). Many new approaches to seismic-risk analyses are being developed for the nuclear power industry, which may be applicable to the decommissioning of the site and the integrity of engineered barriers, remaining structures, filled holes in surficial deposits, and other facilities. Topics include: cumulative absolute velocity (CAV) (actually the cumulative absolute acceleration x time) (Risk Engineering, Inc., 2007) and seismic-wave coherency (NRC, 2007), among others. Soil-structure interaction studies (Ostaden, 2007; EPRI, 2007) should be considered for the final design of in-ground structures left at the end of the decommissioning process.

A seismic-risk analysis is required providing information including how expected seismicity will affect any structures, facilities, engineered barriers, and surficial ground inhomogeneities, such as filled pits or trenches and berms remaining at the WNYNSC after decontamination and decommissioning. The prudent approach would be to undertake complete probabilistic and deterministic seismic-hazard analyses (PSHA and DSHA) and seismic-risk evaluation, similar to those used in the design of nuclear power plants. These were not presented in the DP or for that matter the SDEIS.

However, it should be pointed out that some engineering geologists are not convinced that the probabilistic approach is appropriate to stand alone in a seismic-hazard analysis, but rather should be coupled with a deterministic approach, especially for critical A-31

structures, such as nuclear facilities (Krinitzsky, 1998; Krinitzsky, 2002a; Krinitzsky, 2002b; Wang, et al., 2003). Wang, et al. (2003) state that the probabilistic seismic-hazard analysis (PSHA) does not "...provide the intended uniform protection against seismic risk,... [it] is either over-conservative in some areas or not conservative enough in other areas..." Four of their concerns are: (1) "...there is not consensus on exactly how to select seismological parameters and assign weights in PSHA..."; (2) "...the ground motion derived from PSHA does not have a clear physical meaning and should not be compared to ground motion from any individual earthquake..."; (3) "...PSHA cannot define the worst-case ground-motion scenario..."; and (4) "...PSHA provides... infinite choices for the users and decision makers." This highlights the need for a statement of how the results are assumed to be used.

Seismic-risk analysis requires a description of the structures that may be affected by the hazard. Different results will be derived from different scenarios, i.e., whether the high-level tanks will be left in ground, whether the waste-tank vaults will be left in ground, and whether the voids will be filled with backfill, water, or other material. Also of concern is the stability of high-standing impermeable hydraulic barriers with free faces of tens of feet during rains and earthquakes. These and other questions become relevant, if soil-structure interaction (SSI) studies are made for the final design of in-ground structures left at the end of the decommissioning process. These SSI studies can provide valuable information about how impinging seismic waves may disturb the integrity of below-ground features through ground collapse, collapse affecting erosion rates, groundwater flow paths, and the influence of one facility upon another, such as berms affecting the shaking within pits, among other features.

Several issues that are crucial to a complete analysis are ignored or given little discussion. A section is needed on triggering of earthquakes, such as crustal unloading, deep-well injection or withdrawal of fluids, and far-field seismic events, especially away from the Clarendon-Linden fault system.

A.5.1.5 Appendix D, Engineered Barriers and Post-Remedial Activities The description of the sheet piles and impermeable hydraulic barriers leaves a lot of questions about their design and the ability for them to withstand the forces that will be exerted on them when they are excavated to free-standing walls. The two foot anchor of sheet piles into the Lavery Till does not make sense, especially if the till becomes saturated as also would be the case with the hydraulic barriers. No consideration for seismic loading is given except for the "slurry wall."

The DP seems to rely on the engineered barriers to hold back groundwater seepage from the non-source area into WMA 1 and WMA 2 during and after excavation, which seems to be contradictory to the data in the DEIS, Appendix E, where it is stated that hydraulic conductivities have been measured in the Unweathered Lavery Till as great as 3 ft/day.

A-32

A.5.1.6 Other Concerns The Cs-137 prong is omitted from consideration in many parts of the DP.

The possible contamination of the Demolition Debris Landfill is not considered in the DP, except for its brief discussion in the Executive Summary, Sections 1 and 3, and Appendix D. Confusing statements exist throughout the DP where in some parts it says that any work in WMA 4 is not part of Phase 1, but in other parts it says that soil characterization, monitoring, and maintenance will be apart of Phase 1.

The proximity of Erdman Brook to WMA 1 and WMA 2 is not considered in the groundwater modeling or in the excavation plans for WMA 2.

A.5.2 SPECIFIC QUESTIONS AND COMMENTS BY SECTION AND PAGE A.5.2.1 The Executive Summary The Executive Summary does a fair job of guiding the reader through the DP. Some improvements could be made, however, such as referring each discussed item to the relevant sections that contain the details of that item.

P. ES-4, Figure ES-2: The oblique aerial photograph does not identify the lagoons and the location of the groundwater plume.

P. ES-9, Subsurface Conditions, Bullet 5: States that shale bedrock underlies Lavery Till, but fails to indicate that the bedrock has a weathered and fractured upper surface, and that other units might underlie the Lavery Till including the Kent Till and overlying recessional sediments, and possibly Olean Till. The items listed under the bullets in the section "Subsurface Conditions," page ES-9, do not convey the 3-dimensional relations wherein more than one of the glacial units may overlap the bedrock, a concept that may be important to groundwater modeling. The section also fails to include the Olean till as part of the glacial sequence, which may also be an issue in groundwater modeling, since it too overlaps the bedrock in places. The fifth bullet fails to mention that the bedrock is weathered and fractured in places.

P. ES-13, Figure ES-5: fails to identify the pale blue areas in WMA 4, WMA 5, and the area north of WMA 4 (presumably wetlands).

P. ES-15, WMA 4: is not discussed in Section 7 (at least I can't find it), so there is no discussion of soil characterization, monitoring, or maintenance of Cs-137 or Sr-90 contamination. Bullet 2: radioactivity in drainage ditches and surface soil are not depicted on Figure ES-5.

P. ES-15, WMA 5. Waste Storage Area, Bullet 3: mentions Cs-137 in surface soil, but no mention is made in Section 7.6, even though Figure ES-5 implies that the Cs-137 prong may be removed from WMA 4 and WMA 5.

A-33

P. ES-20, Para 1, Last sentence: The details of the 36 wells, depth of screens, etc., and 59 piezometers are not provided on P. D-15, Appendix D, topic 2.2.1.

A.5.2.2 Introduction The Introduction does a fair job of providing some of the background for the DP. Some questions remain, however.

P. 1-15, Figure 1-2: This figure depicts the water bodies onsite, but does not indicate any wetlands. One has to go to other sections, such as Figure B-3, in Appendix B, P. B-8. It would seem that these areas should be described, since a few may be in the area of the Cs-137 prong, and one linear zone (ditch?) is at the terminus of the fourth-quarter 2007 gross beta contour of the groundwater plume on Figure 4-14 (P. 4-66). These then appear to drain into Franks Creek. I do not recall any part of the DP that addresses either the characteristics of these bodies, or contamination that might be in these areas. How deep are they? What kind of heads could they provide to the groundwater situations especially those on the sand and gravel unit? The drafting of this map has some flaws, especially where the label "WMA 2" is positioned over Lagoon 5, and obscures it.

A.5.2.3. Section 2, Facility Operating History An evaluation of much of Section 2 is beyond my expertise, but I have a few questions and comments.

P. 2-22, Figures 2-3 and 2-4: These figures are labeled a bit incorrectly, they do not depict contaminated surface areas. Neither the groundwater plume on the north plateau, nor the spill in the swale between the SDA and the NDA are indicated.

P. 2-33, 2.3.1, North Plateau Groundwater Plume, Para. 1, 1st sentence: the bifurcation is not evident on Figure 2-6. It appears to me to depict a widening of the front of the plume.

P. 2-34, Figure 2-6: The southeastern side of the groundwater plume has contours of gross beta that are migrating to the east through time. Is this considered in the groundwater modeling?

P. 2-37, 2.3.3, The Cesium Prong, Para. 2, last sentence: when will the characterization of the prong outside the site be analyzed for possible cleanup? I am told by Paul Piciulo that NYSERDA will likely be responsible for characterizing soil offsite. This should be stated in the DP. Will the soil in the area north of WMA 4 be characterized, monitored, and maintained as WMA 4 presumably might be?

P. 2-38, Figure 2-7: Have any newer surveys been made? I am told by Paul Piciulo that no new aerial surveys have been made, although soil samples have been taken. This should be stated in the DP.

A-34

P. 2-39 through 2-42, Tables 2-17 and 2-18: These spills should be depicted on one or more maps.

P. 2-44: 1st Para (Not complete), 1st line: "...degraded process solvent absorbed on suitable solid medium..." This process actually used barrels filled with vermiculite, a heated and expanded mica (kitty litter), then filled the barrels with radioactive fluid, topped off the barrels with vermiculite and placed covers on them. Saturated vermiculite is not what I consider a suitable solid medium.

P. 2-44, 2nd full Para: States that the swale supposedly contains an unknown amount of low-level radioactive contamination, with "gamma readings" five to seven times above background. It is not clear that cleanup of the swale is part of the decommissioning of the NDA hardstand. Paul Picuilo says that the swale is not included in Phase 1. Why isn't it a part of the soil characterization, monitoring and maintenance process during the interim end of Phase 1?

P. 2-47, 2.4.3, Construction and Demolition Debris Landfill, Para 4: States that the landfill could be contaminated by the groundwater plume. A more detailed plan is required to evaluate the surface soil contamination in this area. Indeed, the plan for the soil characterization, monitoring, and maintenance of WMA 4 and the area just north of WMA 4 is not present in Section 7.

A.5.2.4 Section 3.0 Facility Description Section 3 appears to be a group of discussions put together from various convenient sources to satisfy the requirements of the Decommissioning Plan Annotated Checklist of Appendix A. Its quality, therefore, varies from excellent to poor.

3.1.3 Facility Description P. 3-10, WMA 2, Low-level Waste Treatment Area, Lagoon 4 and 5: It should be noted that NFS suspected that tritium may have leaked into the groundwater of the sand and gravel unit. This situation might require more testing under these two lagoons than planned.

P. 3-16, WMA 4: This section should recognize that the landfill may be contaminated by the groundwater plume, and requires close monitoring.

P. 3-22, NDA Hardstand: This section does not mention the possibility of contamination down slope from the hardstand in the swale.' The contamination should be monitored during the interim end period of Phase 1.

A-35

3.5 Geology and Seismology 3.5.1, P. 3-45, Regional Physiology, Para 1, 1st Sentence: The Appalachian Plateau in the region of the site is not considered to be a maturely dissected upland region, but rather an immaturely eroded region dominated by glacial deposits and physiography.

2nd Sentence: The region is not bounded on the east by the Tug Hill Plateau, but rather the Catskill Mountains. The Tug Hill Plateau is located to the northeast beyond the Mohawk River lowlands.

3.5.2, Site Stratigraphy, Para. 2. 2nd Sentence: The statement about the undeformed nature of the bedrock does not mention that this is true at the surface, but that strata at depth may be deformed by Alleghanian stresses.

P. 3-48, Surficial Sand and Gravel Unit, 1st complete Para, 2nd. Sentence: Two of the gravel pits are depicted on Figure 3-9, not 3-8.

Sentence 3: The third (Ashford Town) gravel pit is not depicted on Figure 3-9, nor any other figure, as far as I can tell.

P. 3-48. Lavery Till, last Para, 1st Sentence: The borrow pit is not depicted on Figure 3-9.

3.5.3 Site Geomorphology Para. 1, Sentence 1: This statement is only true for surface bedrock. Carbonate rocks do occur at depth beneath the site.

Sentence 2: "Natural subsidence" as used here should be made clearer, because landsliding and slumping, which does occur at the WNYNSC, is considered by most to be types of land subsidence.

P. 3-49, Channel Incision, Para. 1, Sentence 1: Buttermilk Creek would not be characterized as V-shaped, but steep-walled with a flat bottom.

P 3-50, 1st full Para., Sentence 2: Headward advance can also be initiated by other types of base-level changes.

P. 3-50, Slope Movement, Para. 1: This description of slope movement is not complete and a bit simplistic to imply that all erosion is dominated by slump-block movement, since soil creep is also a dominant form of movement.

P. 3-50, Para. 2: These 5 slump blocks should be identified on a map.

P. 3-50, Para. 3: This statement about rates of downslope movement of slump blocks should be referenced. Other estimations should be provided also.

A-36

P.3-50, Gullying: The term "gullying: should be defined, to distinguish this concept, which I presume means formation of gullies along creek walls, from the processes that formed the creeks themselves. Commonly the creeks would be called gullies.

P. 3-51, Para. 2, sentence 1: These 20 major and moderate-sized gullies should be depicted on a map.

P. 3-51, Table 3-13, item 1: Sheet and Rill Erosion was measured with erosion frames.

Those frames observed by Mike Wilson, Sean Bennett, and me were not located at sites of typical erosion, but rather on slumping ground or other unstable conditions. Also, it was never explained why some frames revealed deposition on slopes; could they have been affected by frost heave? These data are suspect.

Item 5: The U.S. Geological Survey (USGS) report provides terrace (?) ages from 15 to 17 Kya (thousand years ago), an age of 15 Kya for Cattaraugus Creek, and an age of 21 Kya for Connoiseraully Creek. It would have been worthwhile to have had these localities depicted on a map, because the dates given in USGS Table 1 make no sense within the standard concepts of the terminal glaciation of the site occurring after these dates. Terraces, mostly slump blocks (?), could not have withstood the Lavery glacial event. Buttermilk Creek is believed to have started eroding after the last ice retreat.

These dates, and the erosion rate derived from them, are therefore, greatly suspect.

3.5.4, Regional Structure and Tectonics P. 3-52, Para. 1, Sentence 2: Underlying beds may have been deformed by orogenies prior to the Alleghanian.

P. 3-55, Bedrock Fractures Para. 3, Sentences 1 and 3: These joint sets (N45E, N60E) are not prominent on Figure 3-54, if at all there.

Para. 3, last sentence: What is meant by joint sets being "cells." This is a new term to me.

Para. 4, last sentence: The timing of the imposition of joint sets upon the bedding is conjectural and has little relevance to the DP, whereas the orientation and prevalence of bedrock joints might.

P. 3-55, Regional Northwest Trending Lineaments and Structures: This part of Section 3.5.4 does not discuss how it is to be used in an analysis of regional structure or tectonics.,

These ideas are interesting and have been speculated upon by many tectonics researchers, including me (Fakundiny and Pomeroy, 2002, and other papers in Fakundiny, et al.,

A-37

2002). The most impressive geophysical lineament in western New York is the one that defines the CLFS in the basement.

P. 3-54 and 3-55, first Para of section: This paragraph refers to the parallel Tyrone-Mt.

Union and Lawrenceville-Attica lineaments, depicted on Figure 3-56, which were presented in Diment, et al. (1972, 1980), and are irrelevant to any of the seismic-hazard analyses that was a part of the old EPRI-SOC exercise that will be commented on below.

P. 3-56, 3rd full Para.: The Georgian Bay lineament was proposed by Wallach (1990) and Wallach and Mohajer (1990). They should be referenced, if this lineament were relevant to the seismic hazard analysis. The Georgian Bay lineament was not used in the 1989 EPRI-SOC exercise (Risk Engineering, Inc., 2007), because of the timing of publications.

P. 3-56, Clarendon-Linden Fault System: This part of Section 3.5.4 has no references, but should, because the CLFS is the most important structural feature in western New York, and is relevant to seismic-hazard analyses.

Para. 1: The CLFS was depicted in the early stages of its study to have at least 5 faults, but only 3 extended from Lake Ontario to the southern tier counties. This was by Van Tyne (1975). Later studies cast doubt on such a simple structure (Fakundiny and Pomeroy, 2002).

P. 3-57, 2nd full Para., Sentence 1: The CLFS has been active at least since the Proterozoic (Fakundiny, et al., 1978; Wheeler, et al., 2001).

3.5.5, Historical Seismicity: This section is confusing, since it uses different data sources for the list and Figure 3-55.

P., 3-57, WVDP Seismic Reflection Survey, Para. 3, Sentence 3: typographic error: 8945 degrees must be 89 degrees, since dips are not recorded any more precise; the most critical part of this sentence is that it states that the southwest branch of the CLFS could have relative vertical displacement of up to 300 ft. The cross section-interpretation of these lines should be included as figures.

P. 3-58, Para. 2, sentence 1: While the list has about 45 events up to 2003, at least 24 are not depicted on Figure 3-55, since their epicenters were in Ohio or Pennsylvania. Even without depicting these 24 events, the figure still has at least 53 epicenters located. The important earthquake for seismic hazard analysis of the site is the 1929 Attica event with a body-wave magnitude of 5.2.

3.5.6, Evaluation of Seismic Hazard: A discussion of seismic hazards at the site is not required by Appendix A, Decommissioning Plan Annotated Check List. This may be why the section is brief, incomplete, and does not present more than three tables of spectral accelerations: peak ground accelerations at return periods of 500, 1,000, and 2,500 for hard rock, and 500 and 2,500 return periods for the north and south plateaus, A-38

and 3 graphs: seismic hazard curves for peak horizontal acceleration, for 1.0 second horizontal spectral acceleration, and the speculative seismic source contributions to mean peak horizontal acceleration. This section references Wong, et al. (2004) from which these 5 paragraphs, 3 tables, and 3 figures were taken. Nothing is presented about the magnitude and distance contributions to the mean peak acceleration hazard and the magnitude and distance contributions to the mean 1.0 second horizontal spectral acceleration hazard at 500, 1,000, and 2,500 years, nor' the sensitivity analysis of acceleration to attenuation relations, hazard to recurrence models for the CLFS among others. The seismic-hazard curve in Figure 3-58 is somewhat different from-the one currently used for the region by the USGS, which show lower annual exceedance probability at lower accelerations and higher exceedance probability at higher accelerations. The USGS frequency of exceedance versus peak acceleration curve shows about. 145 g for 10,000 years exceedance frequency.

The DP does not consider the seismic loading to the side walls of excavations during the removal of contaminated soil. No consideration is given to induced seismicity, such as might be produced by mine collapse of the kind that occurred at the Retsof mine at Himrod, NY, to the northeast of the site, or by injection of fluids into the subsurface, such as that produced at the brine fields near Dale, NY, just east of Attica. Another concern is the possibility of liquefaction of the sand and gravel unit or the Lavery Till during excavation of WMA 1 and WMA 2.

3.6, Surface Hydrology 3.6.1, P. 3-62, Hydrologic

Description:

This section does not mention the water bodies depicted in WMA 4, WMA 5, WMA 6, and south of WMA 9. The water bodies in WMA 4 and WMA 5 lie within the area of the surface contamination of the Cs-137 plume.

P. 3-64, Para. 6, 1st Sentence: Where are these located on a map? Figure 3-12 is a photograph of the construction of the process building.

3.7, Groundwater Hydrology 3.7.1, Description of the Saturated Zone, P. 3-67, Para. 2, last Sentence: This statement mentions detailed geologic cross sections using data from bore holes installed since 1961.

Where is a map depicting the location of these cross sections, and where are these cross sections themselves presented? These cross sections would be extremely useful.

P. 3-69, No discussion is given of any groundwater in the Olean Till beneath the Kent Recessional deposits.

P. 3-69, Shale Bedrock: Here we have one of the few statements that the bedrock is weathered and fractured, a condition not depicted on cross sections, nor discussed as a possible groundwater pathway. The bedrock valley cross section is more V-shaped than U-shaped.

A-39

3.7.5 Description of Unsaturated Zone Monitoring Stations. P. 3-72; Reference Figure 3-62 as the map of these well locations.

3.8.2, Mineral Resources P. 3-80, Sand, Gravel, and Clay, Para. 3: The number of active sand and gravel pits comes from the New York State Department of Environmental Conservation, and only includes those mines that produce more than 10,000 tons per year. Many more pits producing less most likely exist.

3.8.3, Water Resources:

P. 3-81, 5th full Para, Sentence 1: The concept of regional aquifers, such as the Cattaraugus Creek Basin is a poor scientific notion. Many of the local aquifers are not hydrologically connected, and thus, should be considered different aquifers.

Section 3 Figures:

Figure 3-5, Security Fence Around WVDP Premises Boundary: Why is there a security fence around the area just north of WMA 4? There are water bodies here that are connected by surface drainage to WMA 4, and some of the Cs-137 plume exists here.

Should this area have been considered part of WMA 4?

Figure 3-6, North Plateau Geologic Cross Section: Not depicted is any possible Olean Till, or fractured and weathered zones within the bedrock.

Figure 3-7, South Plateau Geologic Cross Section: Same comment as for Figure 3-6.

On which map are these two cross sections located? A convenient map would be Figure 3-8, WMAs 1 through 10.

Figure 3-9, WMAs 11 and 12: The Town of Ashford gravel pit is not depicted.

Figure 3-51, Regional Physiographic Map: This should be titled Regional Physiographic Province Map. Many mistakes exist on the eastern side of this map, east of the WNYNSC, but these are irrelevant to the DP.

Figure 3-52, Bedrock and Glacial Stratigraphy of the WVDP: Is not drawn to scale, and should be noted so. Where is the Defiance-Lake Escarpment deposits? Olean Till is not presented beneath the Olean Ice Deposits.

Figure 3-53, Surface Geology of the Project Premises and the SDA: This map should be enlarged and better drawn. The source is Wong, et al. (2004). The location of the sand pits and borrow pits should be depicted, as well as Rock Springs Road.

A-40

Figure 33-54, Fold and Selected Joint Trends.... This figure should provide a reference, probably Engelder and Geiser (1980).

Figure 3-55, Seismo-Tectonic Map of Western New York Showing Selected Regional Structures: This is not a seismotectonic map, and the title also misleads by indicating that all of the lines are structures. This is actually a map of selected satellite-obtained linears (lines on images rather than lineaments, which are topographic) with selected epicenters projected onto it. The map has no scale or longitude and latitude designations, which are needed to identify the locations of the earthquakes in Table 3-15. The map fails to locate the epicenters of 27 earthquakes listed in Table 3-15, although several are situated beyond the boundaries of this map. The source of the map is Jacobi (2002). The Georgian Bay Linear Zone is a speculation by Wallach, Mohajer, and Thomas (1998).

Figure 3-56, Major Northwest Trending Lineaments in New York and Pennsylvania:

Many deficiencies exist on this map, including no mention that the contours are in milligals, the type of aeromagnetic map it is, that the hatched contours have negative values, and that the source is Diment, et al. (1972) and Diment, et al. (1980). The map and these lineaments have little value for the DP, because the approach that was referenced by Wong, et al. (2004) did not use these lines.

Figure 3-57, Location of Seismic Lines WVN1 and BER 83-2A: The location of these reflection-seismic lines would be useful, if we also had the interpretation of the data that came from them.

Figures 3-58, 3-59, and 3-60 (Seismic-hazard curves): Wong, et al. (2004) should be given as the source.

Figure 3-62, Groundwater Elevation Contours of the Sand and Gravel Unit, First Quarter, 2008: Notice that the walls of Erdman Brook influence the water table.

Figure 3-63, Groundwater Elevation Contours of the Weathered Till, First Quarter, 2008:

The base map is obscure; this needs to be better located.

Figure 3-65, Groundwater Elevation Contours of the Kent Recessional Sequence, First Quarter, 2008: Does groundwater exist in the KRS or on the north plateau?

Figure 3-69, Locations of Natural Gas and Oil Wells in Western New York: This map should have a scale and longitude and latitude markings. What is the source?

Figure ,3-70, Locations of Natural Gas and Oil Wells in the Vicinity of the WVDP:

Requires longitude and latitude marks. What is the source?

A.5.2.5 Section 4, Radiological Status of Facility P'. 4-7, Figure 4-1, Location of Impacted and Non-Impacted Facilities: This map does not distinguish between impacted and non-impacted facilities.

A-41

P. 4-34, Figure 4-8, Cross-section of Sr-90 Concentrations Versus Depth in Subsurface Soil on the North Plateau: This diagram is not a standard cross section. It projects wells GP75, GP78, GP7898, and BH21A into the cross-section line depicted on the bottom map. As such, it is less than useful. I have tried to contour the concentrations on the cross section and can make no sense of it. Also, the data were collected from three different years over a 5-year period while the plume is moving, yet the implication of such a diagram is that the concentrations are static through this period.

P. 4-66, Figure 4-14, North Plateau Groundwater Plume: Same comments as given for Figure 2-6 above.

A.5.2.6 5.0 Dose Modeling P. 5-5, Figure 5-2, Conceptual Cross Section View of WMA 1 Excavation With Representative Data on SR-90 Concentrations: The scale is not given, nor is the vertical exaggeration. We do not find this until Figure 7-7, which is the same, but with a different title. Figure 7-7 has a vertical exaggeration of 10:1. The diagram does not indicate that the cross section is depicted on Figure 7-6, and is not straight, but curves through the most contaminated part of WMA 1. Implied here is that the barrier wall, at 38 ft high with a free face will be able to withstand the pressure from the sand and gravel on the outside. Such a free-standing wall, made from material with almost the same density as the natural ground outside, could not provide the strength to counteract the shear stresses produced in such a configuration.

P. 5-6, Figure 5-3, Conceptual Cross Section View of WMA 2 Excavation With...: The location of this cross section is not given either, but is depicted on Figure 7-10. It is difficult to measure slopes where vertically exaggerated cross sections do not give the horizontal scale. The two steepest slopes, however, appear to be about 20 degrees each, close to failure angles. The slopes for the west and east sides of the excavation are not provided. One would presume that since the groundwater plume extends through the area of the long part of the barrier, the excavation would have to go down to the Lavery Till at its base, again establishing a free wall to hold back the sand and gravel and groundwater.

Also not discussed nor depicted is form and construction of the east wall of WMA 2.

What provisions will be made for slope stability there?

P. 5-11, Figure 5-4, Sources at the Conclusion of Phase 1 of the Proposed Decommissioning: The water table on the south has contours that appear to not respond to the barrier walls. It appears that flow would move around the south end of WMA 1.

P. 5-13, 5.1.4, Potential Impact of Long Term Erosion: The data from the erosion frames (P. 5-14, 2nd full Para.) are suspect, especially those that indicated buildup (deposition).

P. 5.1.5, Potential Changes in Groundwater Flow Fields, Para. 1: The contentions in this statement are conjectural, and need to be confirmed by adequate groundwater modeling.

A-42

5.1.6, Seepage of Groundwater, P. 5-16, 3rd full Para., Bullet 2: Seepage from the west side of and under the impermeable barrier of WMA 2 will most likely occur. And seepage from WMA 2 toward Erdman Brook may likely occur also.

P. 5-16, Potential Impacts of the Kent Recessional Sequence, last Para.: The statement that the vertical hydraulic barriers are expected to substantially reduce the contamination from the non-source area is not consistent with a footing of only 2 feet and the conductivities, as presented in the DEIS, Appendix E, with a range from 1OE-9 ft/day to 3+ ft/day for Unweathered Lavery Till. Groundwater could drain into the excavations rapidly, especially if a sand and gravel pod is encountered that connects both sides of the Till under the barrier foot.

P. 5-17, 1st full Para., last Sentence: Groundwater reaching the Kent Recessional Sequence (KRS) flows "...at an average velocity of 0.40 ft/yr..." What is the source of this statement?

A.5.2.7 6.0 ALARA Analysis P. 6-4, 6.2.1, Phase 1 Proposed Decommissioning Strategy promotes ALARA, Para. 1:

Bullet 1: Where is the planned location of this canister interim storage facility, which is also mentioned on P. 7-14, 7.3.2, last bullet?

P. 6-4, Bullet 6: Parts of the DP are not clear as to whether soil characterization, monitoring and maintenance will take place in WMA 4 and WMA 5 if necessary.

A.5.2.8 7.0 Planned Decommissioning Activities 7.1, Conditions at the Beginning of the Phase 1 Decomnmissioning, P. 7-2 and 7-3, Table 7-1, Facility and Area Conditions at the Beginning of Phase 1: For WMA 4 and WMA 5 there is no mention of soil characterization, monitoring, and maintenance.

7.3.8, Removing the Underground Structures and Equipment and the Plume Source Area, P.7-21, Figure 7-6: The 1,000 contour and where I eyeball the 500 contour of pCi/L appear to trend into and through WMA 2. If so, the placement of the barrier walls may have to be relocated to capture the concentrations they desire.

P. 7-23, Figure 7-8: Same comments as for Figure 5-5, only here the scales are given, which is a 10:1 vertical to horizontal.

P. 7-23, Figure 7-8. Need scales. Presumably the vertical exaggeration is 1:1.

P. 7-30, 7.4.3, Decommissioning the Lagoons, Figure 7-10: This should have a cross section from NW to SE that extends down the slope of Erdman Brook to illustrate how much sand and gravel remains between the excavation wall and the brook's wall.

A-43

P. 7-31, Figure 7-11, Conceptual Arrangement of WMA 2 Excavation, Cross Section: It would be useful to have the groundwater table depicted. The barrier wall for WMA 1 and the barrier on the northeast side are not illustrated.

P.7-32, Hydraulic Barrier Wall Installation, Para, 3, last sentence: This statement about not requiring sheet piles on the southeast side of WMA 2 should be supported by groundwater modeling.

7.5.4, Monitoring and Maintenance, P. 7-36. Where is the discussion of WMA 4? What would be the monitoring plan, as mentioned in Section 3, P. 3-16?

P. 7-49, Figure 7-15. Conceptual Schedule of Phase 1 Proposed"Decommissioning Activities: schedules the removal of WMA 2 lagoons and other facilities in years 4 and 5, and schedules the work on WMA 1 for years 5 to 8. It also omits the scheduling of the installation of the barrier wall for WMA 2.

A.5.2.9 Appendix A Appendix A, II.d, Spills, P. A-6, last Bullet: Figures 2-3 and 2-4 do not depict spills.

P A-7, III.e, Geology and Seismology: This section does not require a seismic-hazard evaluation, although the DP has a section that purports to do that.

P. A-8, III.f, Surface Water Hydrology, Bullet 4: The DP does not describe the surface water bodies other than the lagoons.

P.A-8, Bullet 5: The wording of this requirement is ambiguous. Does it include the buried groundwater control barriers on the north Plateau, which are not described in the DP.

P. A-9, Surface Hydrology, last Bullet: There should be a discussion of the effect of the membrane on the NDA on whether it could affect flooding or not.

P. A-31, XIV.b, Characterization Surveys, Bullet 5: Figure 2-1 does not distinguish areas that are impacted or not.

P. A-31 and A-32. Bullets 6 and 8: I could not find a justification for not characterizing the area north of WMA 4 inside the security fence.

A.5.2.10 Appendix D Appendix D, Engineered Barriers and Post-Remediation Activities.

P. D-2 and D-3, last Para. of P. D-2: The sheet piles would only be driven two feet into the Lavery Till, and would be done so without knowing whether sand and gravel pods or open cracks exist there. The excavation could possibly proceed right up to some of the A-44

piles, which would leave the ability for them to withstand the pressure from the upgrade sand and gravel and groundwater to that two feet of Lavery Till. If a storm should occur at that time, the Lavery could turn to mud.

P. D-5, Permanent Downgradient Hydraulic Barrier Wall, 4th Para.: It is difficult for me to imagine a 13 ft thick wall, with a free face that is 50 feet high, made out of soil, bentonite, and cement, holding back the sand and gravel and groundwater to the north, especially if rain occurs or a seismic event occurs, or both at the same time.

P. D-6, Durability of Engineered Barriers: This discussion is confusing because they are using the term slurry wall as if they mean the impermeable hydraulic barrier. The discussion of the construction of the impermeable hydraulic barriers implied that the slurry would be used to hold the trench open as they dug it, and would be displaced by a soil-bentonite-cement mixture that would set in the trench and serve as the barrier. The statement in Para 4 on P. D-6 discusses the seismic resistance of the soil-bentonite slurry wall. Do they mean the slurry in the trench or the barrier wall without support on the inside of the excavation?

P. D-8, Figure D-2: Why is the northeast direction scale different from the northwest direction on the map of the current condition? Is this to accommodate some quirk in the modeling program? The Tank Farm on the map of the current condition does not seem to affect the flow lines very much. Neither model considers that the gully slope to Erdman Brook is within the area depicted on these maps, but does not affect'the flow lines.

P. D-10, Figure D-3: Other maps that depict the flow of groundwater around the barrier walls of WMA 1 indicate that flow would impinge on the south side of the excavation of WMA 2 near the solvent dike and interceptors where barriers are not planned. This situation is alluded to on P. D-11, 1.2.4 Engineered Barriers and Groundwater Flow, Para. 3: but no solution to this situation is provided.

P. D-14, Figure D-4: It would seem important that another seepage monitoring station be established just to the northeast of the northern barrier wall to WMA 2 to monitor the groundwater flowing around that end of WMA 2 from the west.

A.

5.3 REFERENCES

Adams, J., and Halchuck, S., 2003. Fourth generation seismic hazard maps of Canada:

Values for over 650 Canadian localities intended for the 2005 National Building Code of Canada. Geological Survey of Canada, vol. 155, 48p.

Bay Geophysical, 2001. Report-Seismic reflection survey to identify subsurface faults near the West Valley Demonstration Project. Bay Geophysical, Traverse City, Michigan.

A-45

Diment, W. H., Muller, D. H., and Lavin, Pd. M., 1980. Basement tectonics of New York and Pennsylvania as revealed by gravity and magnetoc studies. in Wones. D. R.

(ed.) Proceedings: The Caledonides in the USA. IGCP, Project 27, Caledonide Orogen, 1975 Meeting, Blacksburg, VA, p. 221-227.

Diment, W. H., Urban, T. C., and Revetta, F. A., 1972. Some geophysical anomalies in the eastern United States, in Robinson, E. C. (ed.), The nature of the solid earth.

McGraw-Hill, NY, p. 544-574.

Ebel, J. E., and Tuttle, M., 2002. Earthquakes in the Eastern Great Basin from a regional perspective, in Fakundiny, R..H. et al. (eds.) Neotectonics in the Eastern Great Lakes Basin, Tectonophysics, vol. 353, nos. 1-4, p. 17-30.

Engelder, T. and Geiser, P. A., 1980. On the use of regional joint sets as trajectories of paleo-stress fields during the development of the Appalachian Plateau, New York.

Journal of Geophysical Research, vol. 85, p. 6319-6341.

EPRI (Electric Power Research Institute), 2007. Demonstration of SSI (soil-structure interaction) effects, EPRI Technical Presentation on Seismic Site Response Analysis-Workshop on Seismic Issues, at U.S. Nuclear Regulatory Commission, Rockville, MD, 8/29/07, Palo Alto, CA.

EPRI (Electric Power Research Institute), 1989. Seismic hazard methodology for the Central and Eastern United States, prepared by Risk Engineering, Inc. and others, EPRI NP-4726, 3 volumes.

Fakundiny, R. H., 2003. Seismic-risk evaluation in cities of New York and surrounding regions: issues related to all intraplate cities, in Heiken, G., Fakundiny, R., and Sutter. J.

(eds.) Earth Science in the City, American Geophysical Union, Washington, DC., p.75-119.

Fakundiny, Robert H. and Pomeroy, Paul W., 2002. Seismic-reflection profiles of the central part of the Clarendon-Linden fault system of western New York in relation to regional seismicity, in Fakundiny, R. H, Jacobi, R. D., and Lewis, C. F. M., (eds.),

Neotectonics and seismicity in the Eastern Great Lakes Basin. Tectonophysics, vol. 353, nos. 1-4, p. 17 3 -2 1 3 .

Fakundiny, R. H., Pomeroy, P. W., Pferd, J. W., and Nowak,, T. A., 1978. Structural instability features in the vicinity of the Clarendon-Linden fault system, western New York and Lake Ontario. Proceedings of the 12th Canadian Rock Mechanics Symposium, University of Waterloo, Solid Mechanics Division Study, vol. 13, no. 4, p. 121-178.

Jacobi, Robert, D., 2002. Basement faults and seismicity in the Appalachian Basin of New York State, in Fakundiny, R. H., Jacobi, R. D., and C. F. M. Lewis (eds.)

Neotectonics and seismicity in the eastern Great Lakes Basin. Tectonophysics, vol. 353, nos. 1-4, p.75-113.

A-46

Krinitsky, E. L., 2002a. Epistematic and aleatory uncertainty: a new shtick for probabilistic seismic hazard analysis. Engineering Geology, vol. 66, p. 157-159.

Krinitsky, E. L., 2002b. How to obtain earthquake ground motions for engineering design. Engineering Geology, vol. 65, p. 1 - 1 6 .

Krinitsky, E. L., 1998. The hazard in using probabilistic hazard analysis for engineering, Environmental and Engineering Geology vol. IV, no. 4, p. 425-443.

NRC (Nuclear Regulatory Commission), 2007. Technical Presentations on Seismic Site Response Analysis-Work shop on PSHA, U.S. Nuclear Regulatory Commission, Rockville, MD, 8/28/07 to 8/29/07.

Ostaden, F., 2007. Methods of SSI analysis, Session 2, Technical Presentation on Soil Structure Interaction (SSI) by Bechtel Corporation at Workshop on Seismic Site Response at U.S. Nuclear Regulatory Commission, Rockville, MD, 8/29/08.

Risk Engineering, Inc., 2007. Probabilistic seismic hazard analyses: EPRI-SOG sources, Technical Presentation on seismic site response analyses-workshop on PSHA, site response and site spectra, at U.S. Nuclear Regulatory Commission, Rockland, MD, 8/27/07, Boulder, CO.

Tuttle, Marticia, P., Dyer-Williams, Kathleen, and Barstow, Noel L., 2002.

Paleoliquefaction study of the Clarendon-Linden fault system, western New York State, in Fakundiny, R. H. et al. (eds.) Neotectonics in the Eastern Great Lakes Basin, Tectonophysics, vol. 353, nos. 1-4, p. 263-286.

Wallach, J. L., 1990. Newly discovered geological features and their potential impact on Darlington and Pickering. Atomic Energy Control Board of Canada, INFO-0342, p. 1-20.

Wallach, J. L. and Mohajer, A. A., 1990. Integrated geoscientific data relevant to assessing seismic hazard in the vicinity of the Darlington and Pickering nuclear power plants. Prediction and Performance of Geotechnique. Canadian Geotechnical Conference Proceedings, October, 1990, Quebec City, p. 679-686.

Wallach, J. L., Mohajer, A. A., and Thomas, R. L., 1998. Linear zones, seismicity, and the possibility of a major earthquake in the intra-plate western Lake Ontartio area of eastern North America. Canadian Journal of Earth Science, vol. 35, p. 762-786.

Wang, Z., Woolery, E., Shi, Baoping, and Kiefer, John D., 2003. Communicating with uncertainty: a critical issue with probabilistic hazard analysis. EOS (American Geophysical Union), vol. 84, no. 56, 50in 506, 508.

A-47

Wheeler, R. L., Trevor, N. K., Tarr, A. C., and Crone, A. J., 2001. Earthquakes in and near Northern United States. U.S. Geological Survey Geologic Investigations, Series 1-2737, map.

Wong, I., et al., 2004. Seismic hazard evaluation for the Western New York Nuclear Service Center, New York. Wong, I, et al., URS Corporation, Oakland, California, June 24, 2004.

Van Tyne, A. M., 1975. Clarendon-Linden structure, western New York. New York State Geological Survey Open File. 1-10.

A-48

APPENDIX B QUALIFICATION SUMMARIES OF THE MEMBERS OF THE INDEPENDENT EXPERT REVIEW TEAM Dr. B. John Garrick - Chairperson of the Independent Expert Review Team - Dr.

Garrick has a Ph.D. in Engineering and Applied Science and an M.S. in Nuclear Engineering from the University of California, Los Angeles; graduate from the Oak Ridge School of Reactor Technology; and a B.S. in Physics from Brigham Young University. He is an executive consultant on the application of the risk sciences to complex technological systems in the space, defense, chemical, marine, transportation, and nuclear fields. He was appointed as Chairman of the U.S. Nuclear Waste Technical Review Board on September 10, 2004, by President George W. Bush. He served for 10 years (1994-2004), 4 years as chair, on the U.S. Nuclear Regulatory Commission's Advisory Committee on Nuclear Waste. His areas of expertise include risk assessment and nuclear science and engineering. A founder of the firm PLG, Inc., Dr. Garrick retired as President, Chairman, and Chief Executive Officer in 1997. Before PLG's acquisition and integration into a new firm, it was an international engineering, applied science, and management consulting firm.

Dr. Garrick was elected to the National Academy of Engineering in 1993, President of the Society for Risk Analysis 1989-90, and recipient of that Society's most prestigious award, the Distinguished Achievement Award, in 1994. He has been a member and chair of several National Research Council committees, having served as vice chair of the Academies' Board on Radioactive Waste Management and as a member of the Commission on Geosciences, Environment, and Resources. He recently chaired the National Academy of Engineers Committee on Combating Terrorism. Among other

.National Academy committees he has chaired are the Committee on the Waste Isolation Pilot Plant, the Committee on Technologies for Cleanup of High-Level Waste in Tanks in the DOE Weapons Complex, and the Panel on Risk Assessment Methodologies for Marine Systems. Other Academy committee memberships included space applications, automotive safety, and chemical weapons disposal. He is a member of the first class of lifetime national associates of the National Academies.

Dr. Garrick has published more than 250 papers and reports on risk, reliability, engineering, and technology, author of the book "Quantifying and Controlling Catastrophic Risks" (September 2008), written several book chapters, and was editor of the book, The Analysis, Communication, and Perceptionof Risk.

Dr. Sean J. Bennett - Dr. Sean J. Bennett is a Professor in the Geography Department at the State University of New York at Buffalo. He holds a Ph.D., M.A., and B.S. in Geology. Dr. Bennett has extensive experience in physical and numerical modeling of gully erosion and river processes. His current research interests seek to quantify flow and sediment transport processes in watersheds and to determine the impact of these processes on soil losses, river form and function, water quality and ecology, landscape evolution, and watershed infrastructure and integrity. Prior to joining the State

,B-I

University of New York, he served as a Research Geologist with the U.S. Department of Agriculture, Agricultural Research Service, National Sedimentation Laboratory in Oxford, MS, and was a Research Fellow in the School of Earth Sciences at the University of Leeds.

Dr. Bennett has served as Guest Editor for the International Journal of Sediment Research (WASER), Assistant Editor for The Professional Geographer (AAG), Associate Editor for Water Resources Research (AGU), Associate Editor for the Journal of Hydraulic Engineering (ASCE), and Co-editor for Sedimentology (IAS). Dr. Bennett has published two edited books and authored over 100 journal publications, conference proceeding papers, and technical reports.

Dr. Robert H. Fakundiny - Dr. Robert H. Fakundiny is the New York State Geologist Emeritus. He holds a Ph.D., M.A., and B.A. in Geology. He served as the New York State Geologist and Chief of the New York State Geological Survey for 26 years before his retirement in 2004. Among other honors, he is a Fellow of the Geological Society of America, the American Association for the Advancement of Science, the New York Academy of Sciences, the Geological Society of Canada, and the Geological Society (London). He is a Past President of the American Institute of Professional Geologists, Past President of the Association of American State Geologists and Past Chair of the North American Commission on Stratigraphic Nomenclature. He authored numerous scientific papers on the structure and tectonics of New York State, and is the author of highly recognized work on the Clarendon-Linden fault system.

Dr. Fakundiny was one of the principal investigators and conducted or managed extensive research on the geology, hydrology and geomorphology of the Western New York Nuclear Service Center during the 1970s and 1980s. He was a member of NYSERDA's Independent Radioactive Waste Technical Review Group during the 1990s, and he served as a member of the 2005-2006 West Valley EIS Performance Assessment Peer Review Group.

Dr. Shlomo P. Neuman - Dr. Shlomo P. Neuman is Regents Professor in the Department of Hydrology and Water Resources at the University of Arizona in Tucson. He holds a Ph.D. and a M.S. in Engineering Science, and a B.S. in Geology. Dr. Neuman's fields of specialization are subsurface hydrology and contaminant transport. He has made seminal contributions to the areas of pumping test design and analysis, flow in multilayered geologic media, finite element simulation of subsurface flow and transport, estimation of aquifer parameters, fractured rock hydrology, peat hydrology, geostatistics, and stochastic analysis of heterogeneous geologic media. He is a Member of the National Academy of Engineering, a Fellow of the American Geophysical Union, and a Fellow of the Geological Society of America. He holds honorary professorships at the University of Nanjing and the Hydraulic Research Institute in China.

Dr. Neuman has received numerous awards and citations during his career, including the 2003 Robert E. Horton Medal of the American Geophysical Union, and is a former Birdsall Distinguished Lecturer of the GSA. Dr. Neuman has served on various national B-2

and international advisory panels including the Scientific Review Group for high-level nuclear waste disposal in Canada. Dr. Neuman is Associate Editor of Water Resources Research and a member of the Editorial Board of Stochastic Hydrology and Hydraulics.

He is the author of over 300 publications, and he served on the 2005-2006 West Valley EIS Performance Assessment Peer Review Group.

Dr. Chris G. Whipple - Dr. Chris G. Whipple is a Principal with ENVIRON International Corporation in Emeryville, CA. He holds a Ph.D., M.S., and B.S. in Engineering Science. He is a Member of the National Academy of Engineering and is a Designated National Associate of the National Academies. He chaired and served on the National Academy of Sciences Board On Radioactive Waste Management, andhe chaired the Peer Review of the Yucca Mountain Total System Performance Assessment.

He has been a consultant to the U.S. Nuclear Regulatory Commission's Advisory Committee on Nuclear Waste, to the U.S. Nuclear Waste Technical Review Board, and to the Swedish Radiation Protection Institute. He is a Member of the National Council on Radiation Protection, and a Charter Member, Fellow, and Former President of the Society for Risk Analysis.

Dr. Whipple has served on a number of national and international review boards and oversight committees, and he is the author of numerous publications on risk assessment, risk management, and risk communication. Dr. Whipple chaired the 2005-2006 West Valley EIS Performance Assessment Peer Review Group.

CONSULTANT TO IERT Thomas E. Potter - Thomas E. Potter holds a Master of Science degree in environmental science (emphasis in radiation protection) from the University of Michigan, and a Bachelor of Science degree in chemistry from the University of Pittsburgh. He is an independent radiation protection consultant. His consulting experience exceeds 30 years and is pre-dated by 7 years of experience in nuclear materials processing, operational health physics, and nuclear materials licensing. His consulting work has included a broad range of radiation protection matters, mostly for private U.S. Nuclear Regulatory Commission licensees. Projects included environmental radiation dose assessments of operations, accidents, and decommissioning actions; assistance in formulation of licensee positions and comments on developing regulations; design of radiation protection programs and environmental radiation monitoring programs; audits and management reviews of radiation protection programs; and litigation support. Mr. Potter lectured and conducted computer workshops in Cairo as part of a course on environmental radiation dose assessment sponsored by the International Atomic Energy Agency for the Egyptian government. As a consultant at Pickard, Lowe and Garrick, Mr. Potter participated in the design and development of the CRACIT code for the assessment of consequences from severe power reactor accidents, and participated in the consequence assessment portions of a number of full-scope probabilistic risk assessments for power reactors. He also participated in a comprehensive assessment of offsite radiation from the Three Mile Island accident. 0-B-3

APPENDIX C ACRONYMS AND ABBREVIATIONS ALARA As Low As Reasonably Achievable Center Western New York Nuclear Service Center CLFS Clarendon-Linden Fault System Codes CHILD, RESRAD, SIBERIA DCGL Derived Concentration Guideline Level DEIS Draft Environmental Impact Statement DOE U.S. Department of Energy DP Phase 1 Decommissioning Plan for the West Valley Demonstration Project DSHA Deterministic Seismic-Hazard Analysis EIS Environmental Impact Statement IERT Independent Expert Review Team KRU Kent Recessional Unit ND No Dispersion NDA NRC-Licensed Disposal Area NRC U.S. Nuclear Regulatory Commission NFS Nuclear Fuel Services NYSERDA New York State Energy Research and Development Authority OLS Olean Recessional Sequence PA Performance Assessment PDEIS Preliminary Draft of the West Valley Decommissioning Environmental Impact Statement PHGA Peak Horizontal-Ground Acceleration PSHA Probabilistic Seismic-Hazard Analysis SB Subsurface Soil SDA State-Licensed Disposal Area SS Surface Soil SSI Soil Structure Interaction ULT Unweathered Lavery Till USGS U.S. Geological Survey WMA Waste Management Area WNYNSC Western New York Nuclear Service Center WVDP West Valley Demonstration Project C-1