ML23017A083

Technical Review: Percolation Through and Potential Erosion Near the Closure Cap for the 2020 Saltstone Disposal Facility Performance Assessment - Ha

ML23017A083
Person / Time
Site:	PROJ0734
Issue date:	04/18/2023
From:	Hans Arlt NRC/NMSS/DDUWP/RTAB
To:
Arlt H, Ridge A
Shared Package
ML23090A081	List: ML23017A083 ML23017A084 ML23017A085 ML23017A086 ML23017A087 ML23017A088 ML23017A089 ML23017A090 ML23017A113
References
eConcurrence 20230331-60017
Download: ML23017A083 (53)

Text

Technical Review: Percolation Through and Potential Erosion near the Closure Cap of the U.S. Department of Energy 2020 Performance Assessment for the Saltstone Disposal Facility at the Savannah River Site

Date

April 18, 2023

Reviewer

Hans Arlt, Sr. Risk Analyst, U.S. Nuclear Regulatory Commission

1.0 Purpose and Scope

The purpose of this U.S. Nuclear Regulatory Commission ( NRC) staff Technical Review Report (TRR) is to document the NRC staff review of the U.S. Department of Energy (DOE) analyses of percolation rates through, and potential erosion near, the c losure cap at the DOE Saltstone Disposal Facility (SDF) at the Savannah River Site (SRS). This TRR is closely related to another NRC staff TRR entitled Site Stability (ML23017A114), which addresses erosion of the closure cap at the SDF.

The DOE included analyses related to percolation through and erosion near the SDF closure cap in the DOE 2020 SDF Performance Assessment (2020 SDF PA) (the DOE document SRR-CWDA-2019-00001) and other related documents. The 2020 SDF PA is available in the NRCs Agencywide Documents Access and Management System under Accession No. ML20190A056.

The NRC staff evaluated the 2020 SDF PA and other relevant DOE documents to support a future decision about whether the DOE has demonstrated that radioactive waste disposal activities at the SDF are in compliance with the performance objectives of Title 10 of the Code of Federal Regulations (10 CFR) Part 61, Licensing Requirements for Land Disposal of Radioactive Waste, at the SRS SDF pursuant to Section 3116(b) of the Ronald W. Reagan National Defense Authorization Act for Fiscal Year 2005.

This technical review also supports the NRC monitoring of the SDF under Monitoring Factor (MF) 2.01(Hydraulic Performance of Cl osure Cap) and MF 2.02 (Erosion Control of the SDF Engineered Surface Cover and Adjacent Area) under Monitoring Area (MA) 2 (Infiltration and Erosion Control). Those monitoring factors and monitoring area are described in the 2013 NRC NDAA-WIR SDF Monitori ng Plan (NRC SDF Monitoring Plan, (ML13100A113). This TRR includes NRC staff recommendations for changing the NRC SDF Monitoring Plan after the NRC issues a Technical Evaluation Report (TER) for the 2020 SDF PA. The basis for all the NRC staff recommended changes to the NRC SDF Monitoring Plan appear in Section 4.0, NRC Staff Evaluation". This TRR also lists each recommendation in Section 7.0, Conclusions.

2.0 Background

The NRC staff previously addressed percolation through and potential erosion near the SDF closure cap in the following documents:

• 2005 NRC SDF TER (ML053010225);

• 2012 NRC SDF TER (ML121170309);

Enclosure

- 2 -

• 2013 NRC SDF Monitoring Plan (ML13100A113); and

• 2018 NRC SDF TRR entitled Hydraulic Performance and Erosion Control (ML18002A545).

The NRC staff concluded in the 2012 NRC SDF TER (2012 TER) that more model support is needed for the long-term infiltration -limiting performance of the closure cap and infiltration-limiting engineered layers above the SDF disposal structures. The NRC staff made changes to the NRC SDF Monitoring Plan based on the analyses documented in the 2012 TER.

Subsequently, the NRC staff reviewed both the DOE Fiscal Year (FY) 2013 and FY 2014 SDF Special Analysis Documents (ML14002A069 and ML15097A366). In the 2020 SDF PA, the DOE changed the assumptions in those documents about infiltration and percolation through the closure cap, specifically with regard to the performance of the high-density polyethylene (HDPE) geomembrane in combination with a geosynthetic clay liner (GCL) (referred to as HDPE/GCL composite barriers or composite barriers) and the lateral sand drainage layers.

Previous NRC TRRs have not addressed erosion in the areas adjacent to the future closure caps; however, the NRC staff addressed erosion in areas near the closure cap in a January 2017 SDF Onsite Observation Visit (OOV) ( ML17054C453). The DOE has also provided additional information since the NRC staff c ompleted that OOV. This TRR will discuss the results of the NRC staffs evaluation of the erosion in areas adjacent to the Z-Area and revisions that the NRC staff recommends for MF 2.02 (Erosion Control of the SDF Engineered Surface Cover and Adjacent Area). However, erosion on the closure cap itself and site stability are discussed in the NRC staff TRR entitled Site Stability (ML23017A114).

3.0 Percolation Through and Potential Erosion near the Closure Cap:

Information from the 2020 SDF PA and Relevant DOE Documents

3.1 Overview of Percolation Through the Closure Cap

Terminology and Conceptual Model

This TRR uses terminology related to water flow as shown in Figure 1, below. In some cases, the terminology that the DOE used in the 2020 SDF PA and supporting documents differed from common usage. This section describes those differences and clarifies how the terms are used in this TRR.

Percolation, as discussed in this TRR, is the water movement through the unsaturated zone, which extends from the ground surface (e.g., surface of the closure cap) to the aquifer water table. For the natural system, infiltration is the flux of water from the atmosphere (e.g., rain) into the soil. However, in the 2020 SDF PA and supporting documents, the DOE described the subsurface water flux from the bottom of the C losure Cap Model to the top of the Vadose Zone Model as infiltration. To differentiate between water on the ground surface infiltrating into the surface soil and water flux from the c losure cap to the lower backfill, this TRR will refer to the former as shallow infiltration and the latter as deep infiltration.

The deep infiltration rate is the rate of water flow from the closure cap flowing towards the disposal structures. This rate will usually be much greater than the groundwater recharge directly below the disposal structures (water flux from the vadose or unsaturated zone directly under the disposal structure to an aquifer) since the disposal structure itself may shed water off to the side (e.g., from the lower lateral drainage layer).

- 3 -

Percolation encompasses both shallow and deep infiltration. Rainwater falling on an engineered surface cover can become part of the shallow infiltration. That part of precipitation that does not infiltrate into the ground runs off the surface cover and is termed surface runoff. Part of the shallow infiltration into a cover will be returned to the surface by the forces of evaporation and transpiration (both processes together are called evapotranspiration [ET]). Additional shallow infiltration water can be diverted and leave the system due to subsurface lateral drainage, primarily within sand drainage layers, which were designed for this purpose. The remaining water flux is the deep infiltration and eventually the groundwater recharge.

Based on the descriptions above:

• Precipitation = surface runoff + shallow infiltration

• Shallow infiltration = evaporation & transpiration (ET) + subsurface drainage + deep infiltration

• Therefore, precipitation = surface runoff + ET + subsurface drainage + deep infiltration

• Deep infiltration rate above SDF disposal structures groundwater recharge rate below SDF disposal structures

For larger areas without SDF disposal structures, deep infiltration rate = groundwater recharge rate

Figure 1. Surface and Subsurface Water Balance Components for an Engineered Surface Cover (NRC staff drawing - not to scale)

- 4 -

Codes Used

The DOE revised the Closure Cap Model used to estimate shallow infiltration rates, since the 2009 SDF PA, to incorporate updated assumptions and to use WinUNSAT H software instead of the Hydrologic Evaluation of Landfill Performance (HELP) model. WinUNSAT H is a Windows implementation of the UNSAT H code. UNSAT H is a variably saturated flow code that is used to predict unsaturated flow, water redistribution, and atmospheric interactions for the design of landfill and burial covers (i.e., closure caps). WinUNSAT H has been used for a diverse range of climates and has been verified using data sets from large scale field experiments (SRRA107772 000009 (ML18170A244)). For the 2020 SDF PA, the DOE used WinUNSAT H to determine the shallow infiltration rates for the closure cap for two climate conditions: average climate conditions based on typical weather conditions over the past 50 years and wetter climate conditions based the wettest 10 year period over the past 50 years. The DOE then coupled these intermediate results from WinUNSAT H with the Giroud Houlihan analytical solution to estimate deep infiltration rates through the composite barrier and engineered closure cap. The DOE then used the calculated deep infiltration rates as inputs to the subsequent flow and contaminant transport PORFLOW models.

3.1.1 Shallow Infiltration Calculations

In the 2020 SDF PA, the DOE discussed that WinUNSAT H employs an atmospheric boundary condition consisting of evaporation and shallow infiltration. S ince WinUNSAT H simulations are one-dimensional, runoff is computed as the difference between applied precipitation and the shallow infiltration. The shallow infiltration rate is controlled by the rate at which precipitation is applied and by the moisture capacity of the cover. The moisture capacity of the cover i s defined by the hydrologic conditions in the cover during shallow infiltration. WinUNSAT H also simulates vegetative water uptake by distributing the potential transpiration throughout the root zone in proportion to the relative root density.

The Closure Cap Model for the 2020 SDF PA was based on the original SDF closure cap design, which only assumed a shallow 1.5 percent (%) slope at the surface. In the 2020 SDF PA, the DOE states that the one-dimensional WinUNSAT H model may be underestimating the influence of runoff since the new closure cap design includes a 3% surface slope to meet closure requirements specified in the South Carolinas New Consolidated Solid Waste Landfill Regulation ( Regulation 61-107.19, May 2008).

Within WinUNSAT H, a modified form of the Richards equation was used to focus on the terms used to define the shallow infiltration rate (i.e., Equation ( Eq.) 4.4 2 in the 2020 SDF PA where shallow infiltration is referred to as percolation). That equation served as the primary differential equation solved by WinUNSAT H and describes changes in water storage, isothermal redistribution of liquid water, and water uptake by plants.

3.1.2 Deep Infiltration Calculations

The 2020 SDF PA used three deterministic deep infiltration rates to support the compliance case and various sensitivity cases. Tho se deterministic deep infiltration rates were defined by the DOE as a Most Probable and Defensible (MPAD) infiltration rate, a best estimate infiltration rate, and a conservative estimate infiltration rate. The DOE used those rates for what DOE

- 5 -

termed the compliance case, the realistic case, and the pessimistic case, respectively, as well as other sensitivity cases. The DOE referred to the compliance, realistic, and pessimistic cases collectively as the Central Scenario cases. The NRC staff used the same terms for the Central Scenario cases throughout this TRR to facilitate comparison with cited DOE tables and figures.

However, the use of those terms in this TRR does not indicate an NRC staff judgment regarding whether the cases are realistic, probable and defensible, or pessim istic.

In the Central Scenario cases, the most significant barriers to infiltration within the closure cap (i.e., the upper lateral drainage layer (ULDL), the HDPE, and the GCL) were expected by the DOE to perform as designed for the first 500 years after SDF closure. After 500 years, the compliance and pessimistic cases assume that the ULDL becomes less permeable (i.e., can drain less water laterally away from the disposal structures) while the realist ic case does not make this assum ption. A t 2,000 years, the compliance and pessimistic cases assume that holes in the HDPE increase in size as a result of antioxidant depletion while the realistic case does not make this assumption. Also at 2,000 years, in the pessimistic case, the GCL becomes an order of magnitude more permeable as the result of assumed cation exchange in the pessimistic case. To partially address the risks associated with long-term uncertainties in the deep infiltration rates, a number of sensitivity cases within the 2020 SDF PA used higher deep infiltration rates than the pessimistic case does.

In the 2020 SDF PA, the DOE assumed that deep infiltration into the lower backfill would be equivalent to the leakage rate through the HDPE/GCL composite barrier in the closure cap. In the 2020 SDF PA, the DOE used the Giroud-Houlihan equation to predict the average depth of flow within a drainage barrier with Girouds empirical solution for calculating leak rates for composite barriers (DOE document SRRA107772 000009). The Giroud Houlihan analytical solution is based on Darcys Law and mass balance principles. The equation for the leakage rate from a single defect in the composite barrier based on an average depth of flow (Eq. 4.4 5 in the 2020 SDF PA) showed that the estimated leakage rate was a function of the slope and the slope length of the closure cap, the hydraulic properties of the ULDL and the GCL, and assumptions about the size of the HDPE defects.

In the 2020 SDF PA, the DOE expected that the impact on flow into the lower backfill underlying the composite barrier would be localized given the limited size of the assumed defect so that only a limited area within proximity to the footprint of the defect would be wetted by the water passing through the defect. However, as a modeling simplification, the DOE implicitly assumed that the wetted area would encompass the entire area beneath the closure cap by multiplying the volumetric leakage rate per defect by the number of assumed defects per unit area and then dividing that value by the unit area, resulting in an average leakage rate. In the 2020 SDF PA,

the DOE stated that (1) given the uncertainties in the potential locations for the assumed defects and (2) given that localized flow would result i n localized releases, the simplifying assumption that the entire area under the closure cap was wetted was expected to result in an overestimate of potential releases and transport for the assumed number and size of HDPE defects.

3.1.3 Inputs and Assumptions for the Deterministic Closure Cap Model

Precipitation

In the 2020 SDF PA, the DOE assumed that meteorological conditions would remain approximately the same as current conditions in the indefinite future. Certain DOE sensitivity analyses demonstrated that it was a risk-significant assumption for the DOE Central Scenario in the 2020 SDF PA. The DOE based its assumption about the stability of meteorological

- 6 -

conditions on the paleoclimate record along river channels in eastern Georgia and in the Carolinas, which show ed that the most recent climate transition was a warming event that occurred approximately 8, 000 years ago, with climate conditions remaining relatively stable since then. In the 2020 SDF PA, the DOE assumed that, because the local environment has not shown significant changes over the past 8, 000 years, it was reasonable to expect conditions will continue to be similar into perpetuity for modeling purposes. The DOE also considered wetter climate conditions as an alternative modeling scenario to provide insights into how the assumed climate conditions may influence the performance of the system.

Meteorological conditions used by the DOE for the WinUNSAT H modeling were based on SRS precipitation data from 1964 to 2016. Maximum annual rainfall occurred in 1964 with 183.0 centimeters/year (cm/yr) (72 inches/year (in/yr)), the minimum occurred in 2010 with 72.4 cm/yr (29 in/yr), and the 50-year average rainfall rate was 119.6 cm/yr (47.1 in/yr). The DOE also selected t ypical years with actual data sets that exhibit natural variability and are not affected by the smoothing inherent in long term averages: the year 2002 with 103.0 cm/yr (40.6 inches/yr) and the year 2004 with 119.1 cm/yr (47.1 inches/yr). For the alternative wetter climate condition, the wettest 10 year period (from 1989 to 1998) was used to obtain the wetter conditions with an annual average rainfall of 133.0 cm/yr (52.4 inches/yr ) (i.e., 11.2% more the 50-year average).

Evapotranspiration

In the 2020 SDF PA, the DOE selected grass as the assumed vegetati on growing on the closure cap. The DOE indicated that selecting grass would ensure greater defensibility in the estimated shallow infiltration rates because grass transpires less than bamboo or pine trees, leaving more of the water to flow towards the wasteform. The DOE document WSRC -STI -2008-00244 (ML083400069) provided ET rate estimates from various studies which range from 74 to 91 cm/yr (29 to 36 in/yr) with a median of 79.3 cm/yr (31.2 in/yr). With these rates, ET dominates the water balance distribution of precipitation at the SRS.

Figure 4.4 9 from the 2020 SDF PA presented the simulated water balance for typical years and from the wettest ten-year period on record. In the 2020 SDF PA, the DOE stated that the ET rate of 650 millimeters ( mm) (26 in) showed relatively little variability because ET is near its potential due to the humid climate of the SRS.

- 7 -

Closure Cap Layer Thicknesses

Table 1 below provides the modeled thicknesses for each layer of the closure cap.

Table 1. Modeled Thicknesses for SDF C losure Cap Layers (Table 4.4-1 in the DOE 2020 SDF PA)

The designed thicknesses of the middle backfill vary based on the location relative to the apex and the perimeter of the c losure cap. During the development of the WinUNSAT H model, the DOE examined the effect middle backfill thickness variations would have and found annual shallow infiltration rates differed by less than 2% (SRRA107772-000009). Based on that evaluation, the DOE conducted all subsequent simulations with the middle backfill layer set at a thickness of 305 mm (12 in). The geotextile fabric and geotextile filter fabric layers within the designed closure cap we re not credited for any hydraulic performance according to the 2020 SDF PA and they are not included in the closure cap modeling.

Closure Cap Erosion

In the 2020 SDF PA, the DOE assumed that erosion will not significantly impact the estimated shallow infiltration rates. Sensitivity analyses were presented to provide ins ights relative to the potential risks associated with erosion; however, the DOE indicated that erosion, gullying, and slope stability will be addressed as part of the final c losure cap design. Although the DOE discussed increased erosion potentially increasing shallow infiltration rates into the closure cap over time, the DOE concluded that such an increase would not be significant within the first 1,000 years after SDF closure. The DOE used the U.S. Department of Agricultures (USDAs)

Revised Universal Soil Loss Equation (RUSLE) from USDA-HDBK-703 ( USDA, 1997). Using the RUSLE approach (Eq. 1, below), the erosion rate was calculated by the DOE using values related to surface runoff, slope, vegetative cover factor, among others. Table 4.4-6 in the 2020 SDF PA provided the assumed values for tho se factors and the basis for the assumed parameter values for the RUSLE. The NRC TRR entitled Site Stability (ML23017A114) describes potential erosion on the planned c losure cap in more detail. Figure 2, below, shows the results of the erosion calculations from the 2020 SDF PA.

runoff soil slope slope cover erosion Erosion erosivity erodibility length steepness management control Eq. 1 Rate = rainfall practice factor factor factor factor factor factor

- 8 -

Figure 2. Potential C hange to the T hickness of the Upper Backfill Layer over T ime -

the Upper Backfill is D esigned to be 76 cm ( 30 in) Thick (from Figure 4.4-12 in the DOE 2020 SDF PA )

The DOE performed an evaluation during the development of the WinUNSAT H model (SRRA107772 000009) related to the varying thickness of the middle backfill (see Figure 3).

That evaluation is relevant to determining what impact erosion of the upper backfill and topsoil layers will have on shallow infiltration rates through the closure cap.

Figure 3. Conceptual C onfiguration of the SDF C losure Cap (from Figure 3.2-32 in the DOE 2020 SDF PA)

- 9 -

The results of the DOE evaluation indicated that the middle backfill thickness will not have much impact on the shallow infiltration rates relative to other parameters because the erosion barrier was described by the DOE as acting as a choke in the cover profile, thereby forcing water redistribution to occur almost entirely in the upper backfill layer and the topsoil. T he modeled upper and middle backfill layers and erosion barrier all have the same s aturated hydraulic conductivity value, but the modeled erosion barrier has a lower porosity then the backfills (Table 2, below). This is due to the fact that the granite stones within the erosion barrier layer are considered impermeable and nonporous while the remaining void space between the stones, filled with sandy soil, is designed to be a little more than one-third of the total volume. The resultant porosity becomes that of the sandy soil porosity times that of the void space (i.e.,

roughly 40% times a little more than one-third). Since approximately two-thirds of the erosion barrier volume is composed of granite, the DOE stated that the remaining volume of fill material was usually simulated as being saturated, thereby restricting the modeled flow rate to the middle bac kfill. Although the c losure caps upper backfill layer and the middle backfill layer were separate and discrete features. The DOE indicate d that, given that the thickness of the middle backfill layer does not have a strong influence over shallow infiltration rates, it was reasonable to expect that changes to the thickness of the upper backfill layer would also exhibit a minimal influence since both upper and middle backfill layers have identical material properties and serve similar functions, and that the assumption that erosion will not significantly influence the SDF performance was reasonable.

Closure Cap Hydraulic Properties

In the 2020 SDF PA, the DOE provided an overview of the hydraulic properties for the SDF closure cap (see Table 2, below).

Table 2. Modeled Hydraulic Properties for SDF Closure Cap L ayers (from Table 4.4-2 in the 2020 SDF PA)

The upper and middle backfill saturated hydraulic conductivity values are slightly higher than the value the DOE used to model the lower backfill material due to the assumption that soil layers within the c losure cap will undergo pedogenesis and los e compaction since upper and m iddle backfill layers are closer the surface and more susceptible to these processes.

- 10 -

The DOE modeled value for the sand in the ULDL was the minimum saturated hydraulic conductivity procurement requirement for the sand within the drainage layers (WSRC STI 2008 00244). A lower (bounding) value was also assumed by the DOE since closure cap performance (i.e., the ability of the closure cap to decrease infiltration) was expected to decrease with a decreasing hydraulic conductivity of sand in the ULDL. For the GCL, a lower value and a higher value (or bounding condition) were assumed for the saturated hydraulic conductivity. Other property values were obtained from the DOE document WSRC STI 2006 00198 (ML101600380), while the erosion barrier porosity is from WSRC STI 2008 00244. In the 2020 SDF PA, the DOE indicated that the fill material remains an open design issue; however, the DOEs current plan is to use coarse grained sand to fill in the void spaces between the larger stones of the erosion barrier. In the DOE document SRR -CWDA -2021-00072 (ML21321A087), the DOE expected that sand to have equivalent properties to the sand used in the ULDL.

In addition to the modeled saturated hydraulic conductivity of each modeled layer in the closure cap, Table 2, above, also provides two shape parameters, and n, for the soil moisture characteristic curves (also called water retention curves). In a soil mois ture characteristic curve, and n affect the modeled water content as a function of suction head (also cal led matric potential, ). The DOE indicated the values of and n in Table 2, above, are similar to the upper bound for long-term properties of ear then covers that are recommended in NUREG/CR 7028 (ML12005A110).

In general, m oisture characteristic curves are used to determine the unsaturated hydraulic conductivity for materials that are not fully saturated. In the 2020 SDF PA, the DOE used the van Genuchten Mualem model (SRRA107772 000009) to define the relationship between water saturation content and matric suction. Without providing a clear basis for assuming an initial matric head of 1000 kPa throughout the profile for the first day of the first year, more realistic initial conditions for the simulations were created by the DOE running a 5-yr WinUNSAT H simulation using the meteorological record for typical (Year 2002) as input for each year. The WinUNSAT H modeling for wetter years assumes an initial matric potential based on the final value from simulations of the typical years reflecting hydrologic processes occurring within the cover profile.

Subsurface Lateral Drainage in Combination with Geomembranes

The NRC TRR entitled Performance of the Composite Barrier Layers and Lateral Drainage Layers (ML23017A089) discusses in depth characteristics, assumptions, and performance of the ULDL and the HDPE/GCL composite barrier layer below the ULDL. Basic characteristic and assumptions related to these components include:

• the geotextile fabric and geotextile filter fabric layers associated with the drainage layers and composite barriers are not credited for any hydraulic performance

• the hydraulic conductivity of the ULDL assumes that minimum procurement specifications (i.e., 5.0 x 10 2 cm/s [2.0 x 10-2 in/s]) will be met

• the assumed service life of the HDPE is 2,000 years

• the assumed diameter of each HDPE defect is 2 mm (initially) and 10 mm (after the end of service life)

- 11 -

• the assumed number of initial defects is 5 holes per hectare

• the composite barrier and drainage layer have a 4% slope

• the contact factor 0.21 is based on assuming good contact between the HDPE and GCL layers

Other Inputs and Assumptions

As currently designed by the DOE, the leakage rate through the composite barrier layer above the lower backfill in the planned closure cap is based, in part, on the 4% slope of the underlying lower backfill layer and foundation layer (the latter layer is not credited for any hydraulic performance). The composite barrier layer above the middle backfill will be installed on the underlying middle layer with a 3% slope (see Figure 3, above) while previous designs of the closure cap included a cover surface slope of 1.5%. The development of the C losure Cap Model, as described in SRRA107772 000009, was based on the original conceptual design drawings (Figures 5, 6, 7, and 8 in WSRC STI 2008 00244). The new design includes a revised cover slope of 3%, which will promote more surface runoff so that the resulting maximum slope length would be 356.6 m (1,170 ft). The contact factor (0.21) was based on assuming good contact between the HDPE and GCL layers. The three reasons that the DOE assumed there would be good contact were: (1) the DOE plans to perform grading and compaction to create a relatively smooth surface upon which to place the GCL, (2) the DOE will rely on quality control standards to minimize the occurrence of wrinkles and surface, and (3) the DOE assumes that there will be constant pressure on the HDPE due to a minimum of six feet overburden and an expected hydraulic head.

3.1.4 Deterministic Closure Cap Model Results

Closure Cap Water Content and Shallow Infiltration Rates

Intermediate results in the DOE analysis showed that modeled ET rates vary little, so that shallow infiltration varies mainly as a function of precipitation and runoff. In addition, Figure 4.4-10 in the 2020 SDF PA showed the shallow infiltration rate correlated to the precipitation rate, although some non linearity exists due to temporal variability and because durations of individual rainfall events within each simulated year influenced the runoff and shallow infiltration.

For the two typical years, shallow infiltration was estimated to be approximately 400 mm/yr (16 in/yr). For the four highest precipitation years, shallow infiltration rates were all close to 650 mm/yr (26 in/yr). Therefore, the DOE selected 400 mm/yr (16 in/yr ) as a representative shallow infiltration rate for expected climate conditions and 650 mm/yr (26 in/yr ) was selected as a representative shallow infiltration rate for wetter climate conditions. As support for these rates, a range of cover infiltration rates was estimated in Groundwater Flow Simulation of the Savannah River Site General Separations Area, ( the DOE document SRNL STI 2017 00008, Rev. 1) as being between 250 mm/yr (10 in/yr) and 410 mm/yr (16 in/yr). The DOE indicated that that range provided confidence in the shallow infiltration rate of 400 mm/yr (16 in/yr) used in the 2020 SDF PA.

Deep Infiltration Rates

The DOE calculated the leakage rates through the composite barrier layer with multiple sets of

- 12 -

input conditions. These rates served as the basis for the deep infiltration rates assumed in the Vadose Zone Flow Model. Table 3 summarizes the saturated hydraulic conductivities, HDPE defect diameters, climate states, and resulting deep infiltration rates.

Table 3. Summary of I nputs and C alculated Deep Infiltration Rates (from Table 4.4-14 in the DOE 2020 SDF PA)

Modeled Deep Infiltration Rates

In addition to the infiltration rates the DOE used in the 2020 SDF PA, Figure 4 below shows historical results from the previous DOE HELP modeling, which relied on assumptions different from those in the 2020 SDF PA. For example, the 2020 SDF PA assumed no silting in of the ULDL and no tree roots puncturing holes through the composite barrier (SRRA107772 000009).

No matter the assumed climate state, the DOE expected the ULDL, the HDPE, and the GCL to perform as designed for the first 500 years. At 500 years after SDF closure, the DOE assumed that the ULDL would become modestly less permeable during both climate states of the compliance and pessimistic cases. At 2,000 years, the same cases assume that holes in the HDPE increase in size as a result of antioxidant depletion. In addition, the pessimistic case also assumes the GCL becomes 10 times more permeable as the result of assumed cation exchange. In the DOE document SRR CWDA 2018 00030, the DOE discussed corrections that were needed to the original calculations. Therefore, the values in the 2020 SDF PA were not identical to those reported in SRRA107772 000009.

- 13 -

Figure 4. DOE Deep Infiltration Rates Based on the C alculated Leakage Rates from the C losure Cap Model (from Figure 4.4-11 in the 2020 SDF PA)

3.1.5 Development of the Probabilistic SDF Closure Cap Model

In the 2020 SDF PA Closure Cap Model, the DOE calculated deep infiltration rates using an integrated set of deterministic sub-models that simulated environmental conditions and variably saturated flow conditions. However, the model and the results are not probabilistic (i.e., the model did not support statistical distribution sampling and multiple realizations to address uncertainties). To support a fully probabilistic model to estimate uncertainty in long-term deep infiltration rates, the DOE developed a probabilistic SDF C losure Cap Model using GoldSim, a software platform for probabilistic analysis.

The DOE based the deterministic shallow infiltration ( referred to in SRRA107772-000009 as percolation) model and rates used in the 2020 SDF PA on work documented in SRRA107772-000009. However, a more recent DOE document that analyzed the F-Tank Farm ( FTF) and H-Tank Farm (HTF) closure caps (SRRA162682-000002 (ML21179C265)), which were very similar to the SDF closure cap design, was used to provide the results for the abstraction associated with the probabilistic modeling because it included updated input parameters (e.g.,

percent-slope of the closure cap). Therefore, the DOE used the shallow infiltration models of SRRA162682-000002 and their results to evaluate long-term deep infiltration uncertainties.

The DOE abstracted the shallow infiltration models of SRRA162682-000002 into an empirical analytical formula dependent on erosion rates and precipitation rates. With the abstracted

- 14 -

empirical formula incorporated in the GoldSim -based probabilistic model, the shallow infiltration rates were estimated dynamically using parameter distributions for inputs (SRR-CWDA -2021-00040 (ML21160A064)). A range of shallow infiltration rates was estimated by independently lessening the soil thickness above the erosion barrier and varying precipitation rates in the deterministic models. The DOE then statistically analyz ed the relationship between the two parameters and the resulting shallow infiltration rates. The DOE used those s tatistical relationships to abstract the deterministic shallow infiltration models into an empirical formula that could be dynamically applied to a probabilistic C losure Cap Model.

The DOE then developed the probabilistic GoldSim-based SDF C losure Cap Model and used it to generate a range of deep infiltration rates and to address other uncertainties associated with the SDF closure cap. Output from the probabilistic SDF Closure Cap Model (i.e., deep infiltration rates) was then used as input for the Vadose Zone Flow Model. In the DOE document SRR-CWDA-2021-00040, the DOE evaluated long-term deep infiltration rate uncertainties and discussed the probabilistic SDF Closure Cap Model results. While the implementation of the shallow infiltration abstraction has already been discussed in this TRR, the following paragraphs discuss the development and implementation of the other parameter ranges of the probabilistic model documented in SRR-CWDA -2021-00040.

To evaluate the influence of precipitation uncertainties, the range of precipitation values from the probabilistic SDF Closure Cap Model were sorted into probability percentiles (see Table 4.2-1 in SRR-CWDA -2021- 00040). To limit the number of percolation models needed, only low, intermediate, and high precipitation rates were evaluated for each of the drier, the current, and the wetter climate conditions (discussed in Section 3.2.1.3 i n this TRR). For the low and high precipitation rate values, the 10th and the 90th percentile values were selected and the 50 th percentile was selected to represent the intermediate precipitation rate values. This limited the number of percolation models s imulated to nine modeling cases; however, the total number of modeling cases were expanded to 36 modeling cases when various stages of erosion condition were included (see Table 4.2-4 in the DOE document SRR -CWDA-2021-00040).

The DOE varied modeled saturated hydraulic conductivities of the ULDL using probabilistic distributions developed in DOE document SRR-CWDA-2021-00031, (ML21160A061). Table 4, below, presents the DOEs initial and final ULDL hydraulic conductivity distributions for the probabilistic model. As shown in Table 5, the DOE defined minimum, mode, and maximum values of the probabilistic distribution for the final (i.e., degraded) hydraulic conductivity of the ULDL by using sampled values from other distributions. For example, the DOE used the sampled value for the initial hydraulic conductivity of the ULDL as the maximum of the distribution for the final hydraulic conductivity of the ULDL. For the minimum (i.e., most degraded) value for the ULDL hydraulic conductivity, the DOE sampled from a distribution of values for backfill (i.e., the DOE assumed that the most degraded possible state of the ULDL hydraulic conductivity would be equivalent to the value for backfill). Table 6 below shows the DOEs distribution for modeling the time when the ULDL becomes completely silted in and no longer performs as intended.

- 15 -

Table 4. Distributions for the Saturated Hydraulic Conductivity of the U LDL in the Probabilistic Closure Cap Model (A dapted from Tables 5.2 -2 and 5.2-3 in SRR -CWDA-2021-00040 )

Parameter Distribution Parameter Parameter Value (cm/s)

Mean 1.5 x 10-1 Initial hydraulic Standard Deviation 8.0 x 10-2 conductivity Minimum 5.0 x 10-2 Maximum 1.29 Shape Truncated Beta Minimum Sampled value for backfill (see Table 5)

Final hydraulic Mode Sampled initial value for sand divided by 10 conductivity Maximum Sampled initial value for sand Shape Log-Triangular

Table 5. Distribution for the Saturated Hydraulic Conductivity of B ackfill in the Probabilistic Closure Cap Model (from Table 5.2-1 in SRR-CWDA-2021-00040)

Distribution Parameter Parameter Value (cm/s)

Mean 4.10 x 10-5 Standard Deviation 7.59 x 10-5 Minimum 2.0 x 10-5 Maximum 1.4 x 10-4 Shape Truncated log normal

Table 6. Distribution for the T ime to C omplete ULDL Degradation from Silting In (from Table 5.2-4 in SRR-CWDA-2021-00040)

Distribution Parameter Years Minimum 300 Mode 3,000 Maximum 30,000 Shape Log-Triangular

The DOE varied the modeled properties of the HDPE and GCL layers in the composite barrier of the closure cap using probabilistic distributions developed in the DOE document SRR-CWDA-2021-00033 (ML21160A062). The DOE modeled the HDPE with three potential end states: complete HDPE failure (33.3% likelihood), partial HDPE failure (33.4% likelihood), and no HDPE failure (33.3% likelihood). The DOE stated that, while it might be plausible for the entire HDPE material within the closure cap to completely degrade, it was highly unlikely that this would occur throughout the entirety of the HDPE within the closure cap and that HDPE was more likely to degrade locally with areas of HDPE still providing some limited barrier to flow. For tho se reasons, the DOE expected that the most likely failure mode for the HDPE was the partial failure condition, although the probabilistic analysis gives the same weighting to the partial failure condition as to the complete failure condition.

The DOE modeled several GCL properties with probabilistic distributions developed in SRR-CWDA-2021-00033. Modeled distributions for the GCL thicknesses, hydraulic conductivities, and the GCL degradation multiplier were provided (Section 5.3 in SRR -CWDA-2021-00040).

Implementation of the SDF slope geometry required two GoldSim data elements: a maximum slope length of 311 m (1,020 ft) and 4% slope of the composite barrier layer within the closure cap.

- 16 -

3.1.6 Probabilistic Closure Cap Model Results

Probabilistic Analyses

Figure 5 below shows the statistical time history of infiltration rates for a probabilistic model run that randomly sampled the HDPE failure condition (i.e., complete, partial, no failure). The DOE indicated it would consider using two of these time histories for future PA modeling (i.e., they were not used in modeling to support the 2020 SDF PA). Specifically, the authors of the report (SRR-CWDA-2021-00040) recommended that the DOE use the mean infiltration rate (show n in red in Figure 5) as the infiltration rate for compliance cases in future P As. The authors of the report also recommended that the DOE used the median infiltration rate (shown in blue in Figure 5) as the infiltration rate for realistic cases in future PAs.

Figure 5. Statistical Time History of I nfiltration Rates from the Probabilistic SDF Closure Cap Model that R andomly Sampled the HDPE F ailure Condition (from Figure 6.1 -6 in the SRR-CWDA-2021-00040)

Table 4 below shows the results of the DOE probabilistic sensitivity analysis for randomly sampled HDPE failure conditions. The most significant uncertainty associated with the greater than 1,000-year infiltration rates was the HDPE failure condition (i.e., no failure, partial failure, or complete failure) (SRR-CWDA-2021-00040). Failure condition also provided the most risk-significant results as measured by peak dose. The model projections for which the deep infiltration rate peaks near the maximum within 10,000 years were primarily due to the complete failure of the HDPE. Projections that peaked after 10,000 years were predominantly attributed to partial failure of the HDPE because the degradation of the HDPE continues to increase over

- 17 -

time. General insights the DOE gathered from the uncertainty analysis results (SRR-CWDA -

2021-00040), included :

• there was a wide range of variability in the projected deep infiltration rates, indicating that there was significant uncertainty associated with the long-term performance of the SDF closure cap

• the sampled HDPE failure condition has a significant influence on the final deep infiltration rates

• deep infiltration rates that result from modeling the complete H DPE failure condition were generally orders of magnitude higher than the infiltration rates associated with the no HDPE failure or partial HDPE failure conditions

• with no HDPE failure, the 95th percentile deep infiltration rate peaked at 1.2 cm/yr (0.46 in/yr)

• with partial HDPE failure, the 95th percentile deep infiltration rate peaked at 15 cm/yr (5.8 in/yr)

• with complete HDPE failure, the mean and median deep infiltration rates were approximately 33 cm/yr (13 in/yr)

Because of the large effect of the modeled HDPE failure condition o n the projected deep infiltration rate, the DOE performed two types of probabilistic sensitivity analyses for the deep infiltration rate. First, the DOE conducted a probabilistic sensitivity analysis that selected the HDPE failure condition randomly. Then, to gain further insights, the DOE conducted separate probability sensitivity analyses holding the HDPE failure condition fixed at each of its three values: complete failure, partial failure, and no failure. This TRR includes the DOE results for sensitivity analyses conducted with a randomly sampled HDPE failure condition and for a fixed condition of partial HDPE failure in tables and figures below.

Table 7 below presents results from the stepwise ranked regression coefficient (SRRC) analysis for the model runs that randomly sampled the HDPE failure condition. The first eight variables for the 1,000-year period after SDF closure are shown, but it is the first four or five variables that effectively dominate any influence over projected deep infiltration rates at the time analyzed (i.e., 1,000 years). The initial hydraulic conductivity of the GCL and the number of initial defects per hectare both have a strong positive correlation to the deep infiltration, while the length of time for the ULDL sand to degrade and the final hydraulic conductivity of the ULDL both have a strong negative correlation to the deep infiltration. The HDPE failure conditions become the most dominant driver of uncertainty in the deep infiltration rate after 1,000 years, which indicated for the DOE that an improved understanding of the long-term performance of the HDPE material likely will reduce uncertainty.

- 18 -

Table 7. Top Eight SRRC R esults for the D eep Infiltration Rates from the Probabilistic SDF Closure Cap Model that R andomly Sampled the HDPE F ailure Condition (from Table 6.2-1 in SRR -CWDA-2021-00040)

Table 8 below presents results from the SRRC analysis for the final deep infiltration rate from the simulation that assumed the partial failure condition of the HDPE. The 1,000-year results are similar to Table 7 except that the SRRC analyses are able to establish higher cumulative R 2 values because the HDPE failure condition is not being sampled. The 10,000-year result is now missing the HDPE failure condition, but other than that i s similar to Table 7 with the top ranked variable showing the importance of the composite barrier/drainage layer combination.

Table 8. Top Eight SRRC R esults for the D eep Infiltration Rates from the Probabilistic SDF Closure Cap Model with a Partially Failed HDPE (from Table 6.2-3 in SRR -CWDA-2021-00040)

The top ten partially ranked correlation coefficients (PRCCs) for the final deep infiltration rate from the simulation that assumed the partial failure condition of the HDPE are depicted in Figure 6 below and show how the importance of parameters changes over time. Some parameters change significantly between the time of closure and 1,000 years (e.g., the final hydraulic conductivity value of the ULDL) while others change relatively little over 10,000 years (e.g., the initial size of the defects in the HDPE ).

- 19 -

Figure 6. Top Ten PRCCs for the D eep Infiltration Rates from the Probabilistic SDF Closure Cap Model with a Partially Failed HDPE (From Figure 6.2-3 in SRR -CWDA-2021-00040)

Deep Infiltration Rates for Future SDF Modeling

The results from the DOE document SRR-CWDA -2021-00040 show ed that even 1,000 years after SDF closure, the variability in the possible deep infiltration rates can span more than eight orders of magnitude. In SRR-CWDA-2021-00040, the DOE concluded that the uncertainty analysis results indicate that a wide range of deep infiltration rates could be considered for future SDF PA modeling until parameters that affect the projected deep infiltration rate can be further refined. Figure 7 below shows deterministic and probabilistic infiltration rates that the DOE report SRR-CWDA-2021-00040 indicated that the DOE would consider for use in future SDF vadose zone modeling. The DOE also indicated that, for future probabilistic simulations, to limit the number of parametric vadose zone flow simulations needed to support probabilistic analyses, the three HDPE failure conditions (complete, partial, and no HDPE failure) could be modeled separately to represent the ranges of deep infiltration uncertainties with each of the failure conditions.

In the DOE document SRR-CWDA -2021-00066 (ML21217A083), the DOE indicated that (1) the deep infiltration rates from SRR-CWDA -2021-00040 and (2) the projected hydraulic conductivity of saltstone as a function of time for the 10,000 years after SDF closure provided in DOE document SRR-CWDA -2021-00056 (ML21217A081) were both applied to the Vadose Zone Flow Model as inputs to generate a set of probabilistic flow fields. The DOE then used tho se probabilistic flow fields generated with the Vadose Zone Flow Model as input in the probabilistic SDF GoldSim Model.

- 20 -

Figure 7. Deterministic and Probabilistic Deep Infiltration Rates for Potential Use in Future SDF Vadose Zone Modeling (from Figure 6.3-1 in SRR-CWDA-2021-00040)

While Figure 7 above presents deterministic and probabilistic deep infiltration rates the DOE indicated it could use as input to future SDF vadose zone modeling (SRR-CWDA -2021-00040),

the DOE did not revise the deep infiltration rates used to simulate performance and dose in the 2020 SDF PA. Figure 8 below shows the potential future deep infiltration rates compared to the deep infiltration rates the DOE used in the 2020 SDF PA for the realistic case, the compliance case, and the pessimistic case using a log-log scale. The potential future and the currently used rates are similar for the first 300 years after SDF closure, but then diverge by orders of magnitude after that for the compliance and the pessimistic cases.

- 21 -

Figure 8. Potential Future Modeled Deep Infiltration Rates Compared to the Modeled Deep Infiltration Rates used in the 2020 SDF PA (adapted from Figure 6.3-4 in SRR -CWDA-2021-00040)

3.2 Overview of Potential Erosion near the Closure Cap

This section provides an overview of the DOE evaluation of potential erosion that could occur in the areas near and adjacent to the future SDF closure cap. This section does not present an overview of potential erosion on the future closure cap itself because that is addressed in the NRC staff TRR entitled Site Stability (ML23017A114).

An overview of the local drainage areas or watersheds adjacent to the SRS Z-Area are provided in Figure 9 below. Watersheds contribute to local stream flows, which affect erosional processes such as runoff and local stream discharge. The Z -Area and the surrounding topography fall within the Upper Three Runs (UTR) watershed, which was further subdivided into the McQueen Branch watershed, Crouch Branch watershed, and other minor UTR watersheds. As shown in Figure 9 below, the area surrounding the Z -Area is divided between the McQueen Branch watershed (as shown with a pink dotted line) and the other minor watersheds that drain directly to the UTR between the McQueen Branch and the Crouch Branch watershed (as shown with a green dashed line). The average annual flow at the mouth of the McQueen Branch is estimated to vary from approximately 0.06 m 3/s (2.0 ft3/s) to 0.2 m3/s (7.0 ft3/s), with a mean on the order of 0.1 m3/s (5.0 ft3/s). The DOE did not provide the average annual flow estimates for the mi nor UTR watersheds between the McQueen and Crouch Branches in the 2020 SDF PA or in SRR-CWDA-2021-00036.

- 22 -

Figure 9. Topographic Map with W atersheds Surrounding the Z -Area (from Figure 2.4-3 in S RR-CWDA-2021-00036)

3.2.1 Inputs and Assumptions for Erosion Rate Calculations

3.2.1.1 Water Balance and Climate

Processes related to water that affect erosion in some manner include: (1) precipitation, which together with gravity provides the energy that drives erosion; (2) ET, which removes most of rainwater from the system; (3) shallow infiltration, which removes even more rainwater; and (4) runoff, through which the remaining rainwater is available to erode the surface under the correct conditions.

The DOE document SRRA107772-000009 estimated annual precipitation at SRS ranging from 72.9 cm (28.7 in) to 183.0 cm (72 in) with a mean of 120 cm (47 in) using precipitation data from 1964 through 2016 from one SRS weather station. SRR -CWDA -2021-00036 expanded the data used to 13 SRS weather stations (11 currently operating) using data collected over differing periods providing a wider distribution range relative to the dataset used in SRRA 107772-000009. The resulting estimated annual precipitation at SRS ranged from 72.9 cm (28.7 in) to 199 cm (78.4 in) with a mean of 125 cm (49.4 in).

A brief literature review of ET rates at or near SRS provided in SRR-CWDA -2021-00036 showed an annual range from 66.8 cm (26.3 in) to 91.7 cm (36.1 in) with a mean of 81 cm (32 in). In WSRC-STI -2008- 00244 (ML083400069), the DOE also completed a literature review of ET and reported an annual range from 76 cm (30 in) to 85.1 cm (33.5 in) with a median estimate of 79.3 cm (31.2 in). In addition, WSRC -STI -2008-00244 performed long-term water balance modeling for an earlier closure cap design at the SDF with results showing an annual range of ET from 82.7 cm (32.57 in) to 85.4 cm (33.62 in). A water balance model in SRRA107772-

- 23 -

000009 used data from two typical years (2002 and 2004) and the wettest 10-year period (1989 to 1998) resulting in an annual ET of approximately 65.0 cm (25.6 in). SRR -CWDA -2021-00036 summarized the ET rates from each of the values discussed above and calculated a mean and median of value 82.0 cm/yr (32.3 in/yr ). The DOE indicated that that rate did not include the rate given in SRRA107772-000009 because the DOE considered that value to be an outlier. The DOE used the estimated ET rate of 82.0 cm/yr (32.3 in/yr) as the mean for water balance modeling. The estimate uncertainty around that mean is given in Table 9 below. The technical bases for the DOE ET recommendation can be found in Table 3.2-4 of SRR -CWDA -

2021-00036.

Table 9. Distribution for M odeling Annual ET (Adapted from Table 3.2-4 in SRR-CWDA-2021-00036)

Distribution Parameter ET Distribution Values Mean 82.0 cm/yr (32.3 in/yr)

Standard Deviation 6.71 cm/yr (2.64 in/yr)

Minimum 66.8 cm/yr (26.3 in/yr)

Maximum 93.7 cm/yr (36.9 in/yr)

Shape Normal Correlation to Precipitation 0.84

Table 3.3-1 in SRR -CWDA -2021-00036 summarized a literature review of surface runoff rates at or near SRS. The estimates were a mixture of measured and simulated values of surface runoff estimates from natural areas and from engineered surface covers of the SRS region.

Tho se surface runoff rates in the Z -Area and surrounding area ranged from 0 cm/yr to 11 cm/yr (0 in/yr to 4.4 in/yr) with an average rate of 5 cm/yr (2 in/yr). The DOE assumed that years with greater rainfall will result in higher runoff and years with less r ainfall will result in lower runoff.

The DOE distribution for modeling annual runoff from SRR -CWDA -2021- 00036 was summarized in Table 10 below. The DOE bases for the runoff values the DOE used in the probabilistic analyses the DOE developed in response to the NRC staff Request for Supplemental Information (RSI) can be found in Table 3.3 -2 of SRR -CWDA -2021-00036.

Table 10. Modeled Annual Surface Runoff Distribution (Adapted from Table 3.3-2 in SRR -CWDA-2021-00036)

Distribution Parameter Surface Runoff Distribution Values Mean 5.1 cm/yr (2.0 in/yr)

Standard Deviation 2.3 cm/yr (0.9 in/yr)

Minimum 0 cm/yr (0 in/yr)

Maximum 11 cm/yr (4.4 in/yr)

Shape Normal Correlation to Precipitation 0.78

Table 3.4-1 in SRR-CWDA -2021-00036 summarized a literature review of shallow infiltration and groundwater recharge rates at or near SRS. As a simplification, SRR-CWDA -2021-00036 assumed that sh allow infiltration rates were equal to groundwater recharge rates as in the natural system, that is, there is no disposal system that intersects infiltrating water from the surface to the groundwater aquifer (see Figure 1 in this TRR). Tho se combined rates ranged from 37.3 to 42.9 cm/yr (14.7 to 16.9 in/yr) with an average rate of 39.6 cm/yr ( 15.6 in/yr). An input parameter distribution was not necessary for shallow infiltration since parameter distributions for precipitation, ET, runoff had been defined in SRR-CWDA -2021-00036, and DOE relied on the sum of surface runoff and ET subtracted from precipitation to equal shallow

- 24 -

infiltration. The DOE applied this approach with GoldSim probabilistic modeling software and simulated shallow infiltration rates for 10,000 realizations. The results were generally consistent with the values presented in Table 3.4 -1 in SRR -CWDA-2021-00036 (median rate equal to 38.9 cm/yr (15.3 in/yr)). In addition, the model runs showed that shallow infiltration rates were closely correlated to the precipitation values.

3.2.1.2 Probable Maximum Precipitation (PMP) Estimates for the SRS

Previous estimates of the SRS-specific PMP for drainage areas ranging from 2.6 to 2,600 square kilometers (1 to 1, 000 square miles) and rainfall durations from 5 minutes to 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> were made in SRNL-ATG -2005-00022 (ML17170A185), presented in WSRC-STI -2008-00244, and used in the 2009 SDF PA (SRR-CWDA-2009-00017 (ML101590008)). The DOE defined a PMP as the theoretically greatest depth of precipitation for a give duration that is physically possible over a given storm size area at a particular geographic location. Tho se estimates were partially summarized in Table 11 below. The DOE based the PMP estimates for rainfall durations less than 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> and for a 2.6 square kilometer ( 1 square mile) area on procedures outlined in HMR-52 ( U.S. Army Corps of Engineers, 1982). The DOE based the 1-hour duration rainfall over storm areas from 2.6 to 2,600 square kilometers ( 1 to 1, 000 square miles) on interpolation from the standard PMP isohyetal maps, which are maps that show lines connecting points where the same amount of rainfall occurs in a given period. The DOE used a dditional maps presented in HMR-52 to obtain SRS -specific scaling factors that the DOE then applied to the 1-hour PMP value to determine 5 and 15-minute PMP values (SRNL-ATG -2005-00022).

Table 11. PMP for the SRS used in the 2009 SDF PA (Adapted from Table A -1 in WSRC -STI -2008-00244)

Duration 2.6 sq kilometer [km] 26 sq km 518 sq km 2,690 sq km (1 square mile (sq mi)) (10 sq mi) (200 sq mi) (1,000 sq mi) 5 minutes 16 cm (6.2 in) 13 cm (5.1 in) 7.4 cm (2.9 in) Not available 15 minutes 25 cm (9.7 in) 20 cm (8.0 in) 12 cm (4.6 in) Not available 60 minutes 48.8 cm (19.2 in) 139.9 cm (5.7 in) 23 cm (9.1 in) 13 cm (5.1 in)

Although not used in previous SDF PAs, the DOE discussed in SRR-CWDA -2021-00036 that the earlier PMP estimates at SRS were based on the Eliasson Method. The Eliasson Method is a statistical model for extreme precipitation which relies on the mean, standard deviation, and high annual observed precipitation, known as the annual maximum series (AMS), corresponding to each weather station during each year of record. The Eliasson M ethod for estimating the PMP starts by selecting the highest AMS (highest observed precipitation over a given period) for each calendar year recorded. Any precipitation values that are not associated with the AMS were not included in this analysis. The AMS values can be determined from precipitation records collected from the National Oceanic and Atmospheric Administration (NOAA) National Climatic Data Center (NCDC). The NOAA NCDC precipitation records are available at various weather stations for 15-minute, hourly, and daily records.

SRR-CWDA-2021-00036 also used an alternative approach for estimating PMP values that was developed by NOAA. The NOAA Precipitation Frequency Estimates (PFEs) use spatial analysis of historic rainfall data to generate isohyetal maps. The DOE used NOAAs PFE website, which allows location specific PFE data to be retrieved by selecting a state, town, region, or entering coordinates. The NOAA website then generated plots of precipitation depth or PFE for either partial duration or annual maximum time series, including all precipitation amounts for a specified duration at a given station above a pre-defined threshold regardless of year. The

- 25 -

NOAA website also generated depth duration frequency plots based on the r ainfall accumulation intervals and the average recurrence intervals out to 1,000 years. One of the limitations of the NOAA PFE values is that return periods of greater than 1,000 years are not provided. As seen in Figure 10 below, the 15-minute precipitati on depths are considerably less than the depth presented in Table 11 above, although both results were based on interpolation from standard maps of generalized, all -season isohyets of PMP presented in NOAA hydrometeorological reports.

Figure 10. Depth Duration Frequency Curves Based on A verage Recurrence Interval from the NOAA Website (from Figure 3.5-4 SRR-CWDA-2021-00036)

Additional SRS estimates of PMP were presented and discussed by the DOE in SRR-CWDA -

2021-00036. The results of tho se estimates, the Eliasson Method, and the NOAA PFE method discussed above were shown in Table 3.5-8 in SRR -CWDA -2021-00036, and the DOE selected and used the highest values for each time duration, regardless of the reference or method used.

The results are shown in Table 12 below. The PMP results from WSRC-STI-2008-00244, used in the 2009 SDF PA, were not included in the 2020 SDF PA or in the DOE response to the NRC RSI.

- 26 -

Table 12. The DOE PMP Values Used in Erosion Calculations (from Table 3.5-9 in SRR-CWDA-2021-00036)

3.2.1.3 Potential Climate Change Impacts

In the 2020 SDF PA, the DOE assumed that current climate conditions at SRS will remain unchanged throughout the 10,000-year performance period for the SDF. Among other documents, the DOE relied on Leigh (2008), which documented a study on paleoclimates. Leigh documented the long-term climate of the Atlantic Coastal Plain based on pollen and paleochannel records and determined that the local climate transitioned from a cooler, dryer period to the current state 8, 200 years ago and has remained relati vely unchanged ever since.

Despite the described climate stability, the DOE did include an alternative future scenario that included long-term climate change occurring within 20,000 years of SDF closure in the 2020 SDF PA. Leighs paleoclimate summary stated that changes in precipitation occurred approximately 11,000 years ago and approximately 16,000 years ago. The DOE stated in SRR -

CWDA-2021-00036 that:

These transitions are equivalent to an annual probability of 9.1E -05/yr or 1.3E -04/yr, respectively. As a simplifying assumption, one transition will be assumed to occur during the period between 100 years after closure and 10,000 years after closure. This transition will be implemented as a step change and the timing of the step change will be a randomly sampled year, using a uniform distribution between 100 years (i.e., the end of the institutional control period) and 10,000 years (i.e., the end of the performance period).

A distribution for modeling climate change transition time is presented in Table 13 below. The bases for the climate change transition time can be found in Table 3.6-1 in SRR -CWDA -2021-00036. After obtaining that distribution, the DOE then assumed that a decrease in precipitation was as likely as an increase due to precipitation having been historically lower than the current climate. Therefore, there are two climate conditions: drier and wetter climates. Based on the observed transitions by Leigh (2008), the DOE further assumed that any change to the climate at SRS will change the mean precipitation by approximately +/-400 mm/yr (+/-16 in/yr).

- 27 -

Table 13. Modeled Distribution for M odeling Climate Change Transition Time (Adapted from Table 3.6-1 in SRR -CWDA-2021-00036)

Distribution Parameter Climate Change Transition Distribution Values Minimum 100 years Maximum 10,000 years Shape Uniform

To simulate the assumed climate change transition, the model multiplied the water balance parameters for precipitation, ET, and surface runoff by climate change factors (see Table 14 below) after the model reached the sampled time that a climate change transition occurs. The DOE stated that they did not appl y a climate change factor to shallow infiltration due to this value being determined by equations described in Section 3.2.1.1 in this TRR. In addition, no climate change factor wa s applied to the PMP values. The DOE basis for not changing the PMP involves the difference between the 1,000-ye ar and 10,000-year PMP values being on the order of 20% to 30% for storm durations of less than 1 day and that difference having already been factored into the precipitation rate (i.e., the values were changed equivalent to the potential climate change). Once applied, the climate change multipliers would change the runoff rate of 5.1 cm/yr (2.0 in/yr) to either 2.5 cm/yr (1.0 in/yr) for drier conditions or 7.6 cm/yr (3.0 in/yr) for wetter conditions. The resulting shallow infiltration rate would change from a current mean of 38.4 cm/yr (15.1 in/yr) to either 25.9 cm/yr (10.2 in/yr) for drier conditions or 50.8 cm/yr (20.0 in/yr) for wetter conditions.

Table 14. The DOE Change Factors for Selected Water Balance Parameters (Table 3.6-3 in the SRR -CWDA-2021-00036)

3.2.1.4 Estimating Potential Gully Erosion and Sheet and Rill Erosion

The four general surface areas considered in the DOE erosion analysis included the SRS Z-Area hill slopes, the Z-Area hilltops, the SDF closure cap side slopes, and the SDF closure cap top surface. This TRR addresses erosion of the Z-Area hill slopes and hilltops. The NRC staff evaluated erosion of the SDF closure cap side slopes and closure cap top surface in a separate TRR on site stability (ML23017A114).

- 28 -

Estimating Gully Erosion

Gullying was described by the DOE in SRR-CWDA -2021-00036 as being one of the most severe types of erosion that can occur on a cover system, leading to the exposure and possible degradation of internal cover components. Gullying can be caused by surface runoff and the speed at which it flows. The DOE relied on equations for estimating gullying from WSRC-STI -

2008-00244. The DOE also relied on the basic premise that each slope is considered stable relative to the risk of gullying if the actual velocity of flow over a given surface with a defined slope during a PMP event is less than the permissible velocity of flow over a given surface with a defined slope during a PMP event. The equations for calculating the actual velocity are in Section 6.2.1 of SRR-CWDA -2021- 00036 and involve the following parameters and coefficients:

the flow concentration factor, the runoff coefficient, the rainfall intensity, the drainage area, and the flow depth.

It is general practice to use flow concentration factor s to adjust for uncertainty in the geometric influence of tributary discharge. NUREG/CR-4651, Vol. 2 (ML20154Q812) gave possible values for the flow concentration factor ranging from 1 for overland sheet flow, 2 for a high probability of concentrated flow, or 3 for a high probability of channelized flow. Although other values could be used if appropriately justified, according to NUREG -1623 (Step 3 on page A -4) [ML022530043],

the default recommended value should be the most conservative value, which is 3, and DOE assumed that value for all surfaces.

The runoff coefficient was calculated based on the following watershed characteristics: relief, shallow or soil infiltration, vegetation coverage, and surface storage. Runoff coefficients were determined based on information found in Table 15.

Table 15. Runoff C oefficients for R ural Watersheds (from Table 6.2-1 in SRR-CWDA-2021-00036)

- 29 -

The rainfall intensity is a function of the PMP value in addition to the following factors and parameters: the slope (in percentage) and the maximum drainage length. The slope varies significantly in and around the Z -Area. The DOE chose to model slopes with the maximum or near the maximum slope value for the applicable areas. For example, the Z-Area hilltop has a minimum slope of 0% (no slope) and a maximum slope of 2.8%, so the DOE assumed a slope of 3% for the hilltop. Similarly, the Z-Area hill slope has a minimum slope of 5.2% and a maximum slope of 18.3%, so the DOE assumed a slope of 18% for the hill slope. The maximum drainage length can vary based on the component area being studied. Table 6.2-5 in SRR -

CWDA-2021-00036 provided maximum drainage length values for the 2020 SDF PA. The drainage area is equal to the maximum drainage length divided by 43,560 (a unit conversion factor). The flow depth wa s calculated based on Equation A-3 in NUREG -1623 ( ML022530043) and shown again with Equation 6-7 in Section 6.2.1 of SRR -CWDA-2021-00036.

The equation the DOE used to calculate the permissible velocity was in Section 6.2.2 of SRR-CWDA-2021-00036 and involved both the maximum permissible velocity and the velocity correction factor (i.e., correction factors, based on the computed depth of flow, are applied to determine the permissible velocity). The DOE developed the velocity correction factor by interpolating from values provided on page A-5 of NUREG -1623. The velocity correction factor relationship to flow depth was presented with Figure 6.2-1 in SRR -CWDA -2021-00036. Table 6.2-7 in SRR -CWDA -2021-00036 provided the maximum permissible velocities for channels lined with vegetation, as adapted from the USDA Handbook of Channel Design for Soil and Water Conservation (SCS-TP -61) (USDA,1966 ). Table 16 below presents the velocity values for maximum permissible velocities for each of the four general surface areas that were considered in the DOE erosion analysis. The bases for the maximum permissible velocities assigned were presented in Table 6.2-8 in SRR -CWDA -2021-00036.

Table 16. Maximum Permissible Velocities A ssigned for the SRS Z-Area Erosion Analyses (Adapted from T able 6.2-8 in SRR-CWDA-2021-00036)

Surface Being Evaluated Maximum Permissible Velocities SDF Closure Cap Top Surfaces 1.524 m/s (3.871 cm/s)

SDF Closure Cap Side Slopes 2.286 m/s (5.806 cm/s)

Z-Area Hilltops 2.134 m/s (5.420 cm/s)

Z-Area Hill Slopes 0.914 m/s (2.322 cm/s)

Estimating Sheet and Rill Erosion

The average annual rate of soil loss due to erosion without gullying is usually referred to as sheet and rill erosion and commonly estimated using a form of the USDA RUSLE equation, as described in Section 6.3 of SRR -CWDA -2021-00036 and discussed in Section 3.1.3 of this TRR, which contains the rainfall-runoff erosivity factor, the soil erodibility factor, the slope length factor, the slope steepness factor, the cover -management factor, and the erosion control practice factor.

The rainfall-runoff erosivity factor is estimated as a function of rainfall, but more specifically, the value for the rainfall-runoff erosivity factor (350) was selected from isoerodent maps, such as the map shown in Figure 6.3-1 in SRR -CWDA -2021-00036. The influence of climate change on the rainfall-runoff er osivity factor is addressed by the climate change factors for precipitation from Table 1 4 in this TRR (i.e., the rainfall -runoff erosivity factor is multiplied by the climate change factors). For the soil erodibility factor, Table 4.2-4 contributed to the selection of these

- 30 -

values in SRR-CWDA -2021-00036 by identifying erosion properties for selected soils identified at the Z -Area. Weighting the values from Table 4.2-4 based on area, the DOE estimated that the average values were 0.08 for the soil erodibility factor for surface soils and 0.22 for soils weighted by depth. The DOE combined the slope length factor and the slope steepness factor into a single unitless factor. The DOE used Equation E -4 from NUREG -1623 to estimate the combined slope length and slope steepness factor and provided the following results: 0.94 for the Z-Area hilltops and 10.8 for the Z -Area hill slopes. Although the slope steepness factor may decrease over time as soil mass is transported from higher areas to lower areas, the DOE did not take credit for this process. The cover -management factor was used for predicting long-term soil losses at SRS. WSRC-STI -2008- 00244 provided a succession timeline for the growth of pine trees at SRS, which the DOE expected would begin growing approximately 140 years after SDF closure and become mature 460 years after SDF closure. The projected succession timeline yielded the following c over management factors:

• from 0 to 140 years after closure, cover management factor = 0.05 to 0.15

• from 140 to 460 years after closure, cover management factor = 0.003 to 0.05

• at 460 years after closure and beyond, cover management factor = 0.0001 to 0.003

The range of cover management values were then applied as a function of the fractional abundance of bare soil versus ground cover, which the DOE assumed to have a uniform distribution from 0 to 1, yielding a mean of 0.5 (i.e., 50% of the surface is vegetated). As for the erosion control practice factor, the DOE assume d no credit for any potential erosion control practices and thereby assumed a value of 1. The DOE used the parameter values for RUSLE from Table 6.3-3 in SRR -CWDA -2021-00036 to determine the av erage soil loss per unit area.

The DOE then converted the average soil loss per unit area into a depth of soil loss per year by dividing the value by the average bulk density of the soil (the DOE assumed a nominal value of 1.56 g/cm3).

3.2.2 Gully Erosion and Sheet and Rill Erosion Results

The DOE did not perform slope stability analyses for the natural areas surrounding the SDF.

Gully Erosion Results

According to the DOE, the gully erosion results show ed whether erosion can be expected or not for the four general surface areas of the Z-Area hill slopes, the Z-Area hilltops, the SDF closure cap side slopes, and the SDF closure cap top surface. As discussed in the introduction to Section 3.2.1.4, this TRR addresses erosion of the Z-Area hill slopes and hilltops. As stated in Section 3.2.14 of this TRR, if calculated actual velocity rates of runoff flow over these surfaces are greater than the permissible velocity rates during a storm event, erosion will likely occur.

Four tables were presented in SRR-CWDA-2021-00036 with the last row of these tables stating either Erosion or No Erosion. Of these four tables, one table provided a summary of various parameters and analysis results from the gullying erosion evaluation and used a 1, 000-year PMP. A second table used a 10,000-year PMP. Two additional tables used the same input values but were modified (see Section 7.1 in SRR -CWDA -2021-00036) to account for runoff accumulation (i.e., surface runoff flowing from a higher general area (SDF c losure cap top surface) to a lower general area (Z-Area hill slopes)). The modifications had increased both the estimated flow depths and the calculated actual velocities due to the accumulation of surface runoff. However, although all four tables had diff erent calculated actual and permissible velocity

- 31 -

rates, all four tables had the same results with regard to erosion. That is, the DOE determined that only the Z -Area hill slopes are at risk of gullying erosion. Table 17 below shows the results from the 10,000-years PMP return period with accumulating surface runoff. The calculated actual velocity nears the permissible velocity for the Z-Area hilltops 1.0 < 1.2 m/s (3.28 < 3.83 ft/s), but for the Z-Area hill slopes the calculated actual velocity is over the permissible velocity 1.33 > 0.51 m/s (4.36 > 1.68 ft/s) so that gully erosion is likely in those areas.

Table 17. Gullying Erosion Results Using the 10,000-year PMP R eturn Period with Accumulation of R unoff from Preceding Sur faces (r esults obtained from Table 7.1-4 in SRR -CWDA-2021-00036)

Parameters SDF Closure SDF Closure Z-Area Hilltops Z-Area Hill Cap Top Cap Side Slopes Surfaces Slopes Actual Flow 0.56 m/s 1.12 m/s 1.00 m/s 1.33 m/s Velocity (Va) (1.84 ft/s) (3.69 ft/s) (3.28 ft/s) (4.36 ft/s)

Permissible 0.76 m/s 1.22 m/s 1.17 m/s 0.51 m/s Velocity (Vp) (2.50 ft/s) (4.01 ft/s) (3.83 ft/s) (1.68 ft/s)

Will Erosion No Erosion No Erosion No Erosion Erosion Occur? Va < Vp Va < Vp Va < Vp Va > Vp

Sheet and Rill Erosion Results

The range of evaluation results for sheet and rill erosion (for the same four general surface areas that were evaluated for gully erosion) are shown in Figure 7.2-1 and Figure 7.2-2 from SRR-CWDA-2021-00036. The dashed lines in the figures represent the mi nimum ground cover by vegetation (i.e., 0%) while the solid curves show maximum ground cover by vegetation (i.e.,

100%). In SRR-CWDA -2021-00036, the DOE stated that the 1,000-year results assumed that no major climate change transitions occurred. In addition, based on the results seen in Figure 11 below, it appears that the DOE did not include climate change transitions for the next 10,000-year evaluation either.

The DOE results for the first thousand years after SDF closure showed that the greatest depth of sheet and rill erosion is projected to occur along the side slope of the SDF closure cap. Since the Z-Area hilltop and the SDF closure cap top both have similar assumptions for slope and vegetation, thei r sheet and rill erosion results are similar. The Z-Area hilltop, SDF closure cap top, and SDF side slope all are projected to become forested over time and their respective erosion rates are projected to slow down. The erosion rates for the steeper -sloped Z-Area hill slopes were projected to be initially slower due to the dense vegetation but eventual ly to obtain the highest rate of sheet erosion by 10,000 years. The DOE stated that these sheet and rill erosion results represent higher-than -expected estimates, but that they also highlight the need to evaluate the riprap material prior to construction to ensure that the selected materials will provide an appropriate level of defense against long-term sheet and rill erosion.

- 32 -

Figure 11. Depths of Sheet Erosion for 10,000 years B ased on RUSLE A ssumptions (from Figure 7.2-2 in SRR-CWDA-2021-00036)

4.0 NRC Staff Evaluation

Overview

The NRC staff reviewed the DOE analyses of projected percolation rates through, and potential erosion near, the planned SDF closure cap based on information included with the DOE 2020 SDF PA and supporting documents. In addition, the NRC staff used insights from this review to develop recommended changes to the NRC SDF Monitoring Plan. Section 7 in this TRR provides a summary of the recommended changes.

4.1 Overview of Percolation Through the Closure Cap

4.1.1 Shallow Infiltration Calculations

The NRC staff found the use of the one-dimensional WinUNSAT-H code by the DOE to calculate shallow infiltration to be acceptable because the code and the calculation methodology were well documented and supported. The NRC staff had asked the DOE to provide additional information with regard to one-dimensional versus two-dimensional modeling because the one-dimensional WinUNSAT-H calculates runoff by subtracting the infiltration from the precipitation.

If the percent-slope becomes too great, the one-dimensional model may underestimate the influence of runoff. The DOE referred to the DOE document SRRA162682-000002 in their DOE Response to NRC Request for Additional Information ( RAI) Question IEC-9, which had stated that two-dimensional simulations were considered unnecessary because the slope of the cover was less than or equal to 4.7%. T he NRC staff found DOEs explanation acceptable, that the DOE expected that using the one-dimensional WinUNSAT-H model will provide runoff estimates that may be slightly lower than what may be predicted using a more complex two-dimensional modeling software, but the differences are likely not to be risk significant.

- 33 -

4.1.2 Deep Infiltration Calculations

The NRC staff found the approach of us ing three separate deterministic deep infiltration rates for the compliance case, the realistic case, and the pessimistic case to be acceptable because tho se cases allow insights to performance to be gained based on degradation of potentially risk -

significant barriers. The NRC staff found the use of t he Giroud Houlihan analytical solution to be acceptable because it is based on Darcys Law and mass balance principles. The Girouds empirical solution for calculating leak rates for composite barriers (the DOE document SRRA107772 000009) is widely used for that purpose and was found to be acceptable by the NRC staff.

4.1.3 Inputs and Assumptions for the 2020 SDF PA Closure Cap Model

Precipitation

The DOE discussed SRS precipitation data and estimates in SRR -CWDA -2021-00036. The rainfall analysis in that document was an updated version of the analysis described in the 2020 SDF PA, which used precipitation data from one weather station at SRS from 1964 through 2016. SRR-CWDA-2021-00036 included data from 13 additional SRS weather stations up through to 2020. Although the technical basis for the results for the analysis in SRR -CWDA -

2021-00036 were stronger, the rainfall range and mean annual rainfall did not change significantly (i.e., from 1196 mm [47 in] in the 2020 SDF PA to 1, 255 mm [49.4 in] in SRR-CWDA-2021-00036) so that the NRC staff found the 2020 SDF PA precipitation estimates to be acceptable.

Evapotranspiration

The NRC staff found the selection of grass as the initial cover vegetation to be acceptable because grass transpires less than bamboo or pine trees so that more water is potentially available to infiltrate at the surface. The NRC staff found the ET rate of 650 mm (26 in) presented in the 2020 SDF PA not acceptable because no technical basis was provided by the DOE, and the value is significantly lower than given in WSRC-STI -2008-00244 (i.e., 793 mm/yr

[31.2 in]). The 2020 SDF PA stated that ET shows relatively little variability since the ET value of 650 mm (26 in) is near the maximum potential ET. However, in SRR-CWDA -2021- 00036, the DOE described a subsequent analysis of ET and stated that the ET rate of 650 mm (26 in) differs significantly from other observations. Although an ET value of 650 mm/yr (26 in/yr) could be considered a conservative rate if it had an influence on percolation rates, the DOE stated in the 2020 SDF PA that the c alculated percolation rate varied mainly as a function of the precipitation and runoff and so for the deterministic percolation calculations of the 2020 SDF PA, the ET rate is not a risk-significant factor due to its lack of variability. For the specific DOE models discussed here, this may be true; however, in reality, the NRC staff determined that the ET rate within engineered surface cover s, especially ET surface covers, is a risk-significant factor with regard to influencing percolation rates. Section 4.1.4 in this TRR discusses this aspect in more detail.

Closure Cap Layer Thicknesses and Erosion

The NRC staff found the thicknesses of the closure cap layers modeled to be acceptable because the modeled thicknesses are the same thicknesses as the layers of the current preliminary closure cap design based on the DOE document WSRC STI 2008 00244.

- 34 -

In response to the NRC RSI, the DOE discussed SRS erosion related data and erosion estimates in SRR-CWDA -2021-00036. That erosion analysis was an updated version of the analysis described in the 2020 SDF PA, which did not implement a vegetation succession timeline for the growth of grass, transitional woodland shrub, and then pine trees. The timing of the erosion depths (see Figure 11 in this TRR) from SRR-CWDA -2021- 00036 varies considerably from that of the 2020 SDF PA (see Figure 2 in this TRR) within 1,000 years of SDF closure. The analysis in SRR -CWDA-2021-00036 was more detailed than the analysis documented in the 2020 SDF PA and had a stronger technical basis; however, there are still uncertainties with regard to the PMP and climate change transitions which can influence erosion rates and future closure cap thicknesses. Due to those concerns, the calculated erosion depth estimates for determining future closure cap thicknesses for climate conditions other than that of the current climate state are not acceptable to the NRC staff. Tho se concerns are discussed in this TRR and in the TRR entitled Site Stability (ML23017A114) which includes a recommendation for monitoring and addressing this technical issue under a new medium -

priority monitoring factor.

Section 4.4.1.4.1 of the 2020 SDF PA stated that the thickness of the middle backfill layer is insignificant hydraulically because the erosion barrier acts as a choke in the cover profile due to its low porosity, forcing water redistribution to occur almost exclusively in the upper backfill layer and the topsoil. Further, the rate at which water can flow through the erosion barrier is limited by its saturated hydraulic conductivity. Since the DOE stated that the middle backfill thickness will not have much impact on the shallow infiltration rates relative to other parameters and the upper and middle backfills both have identical material properties and serve similar functions, the DOE also stated that it was reasonable to expect that changes to the thickness of the upper backfill layer would also exhibit a minimal influence (Section 4.4.1.5 of the 2020 SDF PA), and that the assumption that erosion will not significantly influence the SDF performance was reasonable. As discussed in the following two paragraphs, the NRC staff does not agree with either one of tho se DOE assessments and is recommending that this issue be monitored within the new MF 2.07 as proposed in Section 4.1.5 in this TRR.

Appendix A in the DOE document SRR-WDA -2021-00081 presented water balance estimates from probabilistic model realizations with the highest cumulative volumetric flows through the saltstone an d with the highest doses. Precipitation rates were generally on the higher end of the range with some precipitation rates as high as 2, 093 mm/yr (82.4 in/yr). A higher rate of flow should activate the choke if the rate at which water can flow through the erosion barrier is limited by its lower saturated volumetric water content ( see DOE Response to Clarifying Comment (CC)-6 in SRMC -CWDA-2022-00016). However, the tables in Appendix A from SRR -

WDA-2021- 00081 did not show that effect. Table 18 below presents simulated water balance flow components, including that for the upper backfill and the erosion barrier. No chokepoint effect was detected during the DOE simulations despite the higher rainfall rates. However, although the NRC staff sees no evidence of a chokepoint in the water balance simulations, the NRC staff is open to the possibility that a chokepoint could exist in the future closure cap.

- 35 -

Table 18. Realization #302 W ater Balance Summary with R elatively High Rainfall and No Evidence of a C hokepoint between the U pper Backfill and the Erosion Barrier (from Table A-6 in SRR-CWDA-2021-00081)

The NRC staff also does not agree with the DOE determination that it was reasonable to expect that changes to the thickness of the upper backfill layer would not significantly influence the SDF performance. It is not reasonable because m ost of the water for ET would be provided from the topsoil and the upper backfill. Any chan ge in that thickness will likely change the dynamics of the water balance components including ET, as is also described in the DOE Response to CC-6 in SRMC -CWDA -2022-00016. For example, the concept of the ET cover system is to use the water storage capacity of the soil to promote water evaporation and transpiration at some later time after a rainfall event. The greater the storage capacity and ET properties, the lower the potential for percolation through the cover system. The lower the soil storage capacity (e.g., the upper backfill), the higher the potential for percolation through the cover system (EPA, 2003).

Closure Cap Hydraulic Properties

The NRC staff found hydraulic property values assigned by the DOE to the topsoil, upper backfill, erosion barrier, and middle backfill to be acceptable because the technical basis for the value of each component was robust. The NRC staff agrees with the DOE assumption that the upper soil layers within the closure cap will undergo pedogenesis and los e compaction so that their saturated hydraulic conductivity values will increase. The NRC staff found the value the DOE used in the 2020 SDF PA for the fill material between the stones of the erosion barrier (i.e., equivalent properties to sand) acceptable for modeling purposes; however, the DOE stated that the final choice of material for the infill between the riprap stones has not been made ( see the DOE Response to RAI Question IEC-4 in SRR -CWDA -2021-00072). The rate of flow through the erosion barrier also affects the amount of water available above the erosion barrier, which affects the rate of runoff and er osion in those layers. Therefore, the NRC staff recommends that the NRC monitor information related to the design and simulated performance of the erosion barrier (i.e., the simulated flow rate th rough the erosion barrier and the potential erosion and stability of the layers above the erosion barrier) under a new monitoring factor in MA 2 (Infiltration and Erosion Control) under bo th 10 CFR 61.41 and 10 CFR 61.42.

Recommendation PECC-1 The NRC staff recommends opening a new medium-priority monitoring factor entitled Long-Term Erosion Barrier Performance under MA 2 (Infiltration and Erosion Control) under PO §61.41 and PO §61.42 and expects to close the new monitoring factor after the NRC determines that there is a sufficiently strong technical basis to support the simulated hydraulic performance above, below, and thr ough the erosion barrier.

- 36 -

The NRC staff found simulated hydraulic property values assigned by the DOE to the ULDL, the HDPE, and the GCL below the HDPE not acceptable because the technical basis supporting their initial performance over time to be weak (i.e., lack of support that the degradation processes had minimal or nonexistent impacts in the compliance simulations). The NRC staff TRR entitled Performance of the Composite Barrier Layers and Drainage Layers (ML23017A089) discusses th ose cover components, the NRC staffs evaluation of their performance in detail, and includes the new monitoring factor to address monitoring of the values the NRC staff did not find to be acceptable entitled Long-Term HDPE/GCL Composite Barrier and Drainage Layer Degradation.

As for the moisture characteristic curves used to determine the unsaturated hydraulic conductivity for materials that are not fully saturated, the NRC staff found the DOE WinUNSAT H modeling assumption of an initial matric suction or potential of 1, 000 kPa throughout the profile on the first day of the first year to be acceptable based on the results seen in Figure IEC-9.4 in the DOE Response to RAI Question IEC-9 in SRMC -CWDA -2022-00016.

Although it is clear that 1, 000 kPa (applied as 10,000 cm of water suction) as the initial condition is not as close to the equilibrium state as 100 kPa, both sets of models showed that, aside from the first-year iteration, the matric suction values achieved equilibrium conditions for matric suction quickly. The initial assignments were shown not to be significant as long as a sufficient number of years were repeatedly run to achieve this equilibrium state (i.e., the closer the initial matric suction assignments are to the equilibrium conditions, the faster the models can reach equilibrium).

Subsurface Lateral Drainage in Combination with Geomembranes

The NRC staff TRR entitled Performance of the Composite Barrier Layers and Drainage Layers (ML23017A089) discussed in-depth characteristics, assumptions, and performance of the ULDL the HDPE/GCL composite barrier layer below the ULDL, and the NRC staff evaluation of that information. Findings and recommendations from that TRR were as follows:

• Observing monitoring the installation of the HDPE/GCL composite barriers and the repairs to cuts and defects in the HDPE under a new monitoring factor under MA 2 (Infiltration and Erosion Control) under both POs 10 CFR 61.41, Protection of the general population from releases of radioactivity, and 10 CFR 61.42, Protection of individuals from inadvertent intrusion.

• Monitoring the development of additional information relevant to HDPE degradation in the heat affected zones near welded seams and at edges, HDPE degradation due to root penetration, GCL degradation due to HDPE defects, and drainage layer degradation due to diminishing hydraulic conductivity under a new monitoring factor under MA 2 (Infiltration and Erosion Control) under both POs 10 CFR 61.41 and 10 CFR 61.42.

• Observing and monitoring the cutting of the HDPE geomembrane to remove wrinkles and waves under a new monitoring factor under MA 2 (Infiltration and Erosion Control) under both POs 10 CFR 61.41 and 10 CFR 61.42.

• Observing whether the initial GCL hydraulic conductivity value used in the 2020 SDF PA was comparable with an independently validated GCL value originally defined by the manufacturer under a new monitoring factor under MA 2 (Infiltration and Erosion Control) under both POs 10 CFR 61.41 and 10 CFR 61.42.

- 37 -

• Monitoring Quality Assurance/Quality Control to ensure future emplacements and installations of HDPE/GCL composite layers are managed and executed at a very high standard and reduce or bound risk -significant uncertainty under a new monitoring factor under MA 2 (Infiltration and Erosion Control) under both POs 10 CFR 61.41 and 10 CFR 61.42.

• Monitoring information and planned activities related to degradation of the ULDL barrier, which includes the GCL, under a new monitoring factor under MA 2 (Infiltration and Erosion Control) under both POs 10 CFR 61.41 and 10 CFR 61.42.

• Monitoring information and activities associated with fine-particle transport and deposition within the ULDL under a new monitoring factor under MA 2 (Infiltration and Erosion Control) under both POs 10 CFR 61.41 and 10 CFR 61.42.

• Monitoring the modeled flow rate to and through the Lower Lateral Drainage Layer (LLDL) for significantly larger infiltration rates to determine if the LLDL modeling approach within the Vadose Zone Flow Model needs to be re-evaluated under a new monitoring factor under MA 2 (Infiltration and Erosion Control) under both POs 10 CFR 61.41 and 10 CFR 61.42.

The NRC staff found the DOE assumed contact factor between HDPE and GCL of 0.21 to be acceptable because the NRC staff monitors processes and features related to the contact factor such as ensuring a relatively smooth surface upon which to place the GCL, good quality control standards to minimize the occurrence of wrinkles and surface, and constant pressure applied to the HDPE. Tho se monitoring activities are associated with a new monitoring factor entitled Confidence in Quality Assurance/Quality Control for HDPE/GCL Composite Barrier and Drainage Layer Installation under MA 2 (Infiltration and Erosion Control).

4.1.4 Deterministic Closure Cap Model Results

Shallow Infiltration Rate Estimates

As previously discussed in Section 4.1.3 in this TRR, the NRC staff disagrees with the DOE on the risk significance of the role ET has on surface covers and on the role of an ero sion barrier acting as a chokepoint in the cover profile. In addition, there is little evidence or model support for the DOE statements that water redistribution occurs almost exclusively in the upper backfill layer and the topsoil. However, the NRC staff does not recommend an additional monitoring activity to address whether the erosion barrier acts as a chokepoint because the assumption did not appear to affect the DOE model results (i.e., the model results do not show that the erosion barrier acted as a chokepoint in the 2020 SDF PA). In addition, the NRC staff considers an ET rate of 650 mm/yr (26 in/yr) a conservative rate with regards to determining shallow infiltration.

The NRC staff agrees with shallow infiltration e stimates of 400 mm/yr (16 in/yr) for expected climate conditions and 650 mm/yr (26 in/yr) for wetter climate conditions as used in the 2020 SDF PA because the model results of the D OE document SRNL STI 2017 00008, Rev. 1 provided support for those estimates.

Deep Infiltration Rates

- 38 -

The NRC staff found the deep infiltration rate estimates the DOE used not to be acceptable because the technical bases for the long-term performance of risk-significant barrier layers lacked confirmatory evidence and model support to match the risk -significance of these cover layers. It is the combined function of the ULDL and HDPE/GCL composite barrier layer that reduces deep infiltration rates to a small fraction of the shallow infiltration rates. For example, as taken from Table 4.4-4 in the 2020 SDF PA, 400 mm/yr [16 in/yr] of shallow infiltration is reduced to a deep infiltration of 0.13 mm/yr [0.005 in/yr], or deep infiltration is 0.0325% of the shallow infiltration rate. Because tho se components provide such significant performance, the technical basis supporting their performance needs to be strong. It is the ability of these cover components to maintain hydraulic property values in the long term that guarantees such significant performance which is in question. As the NRCs TRR entitled Performance of the Composite Barrier Layers and Drainage Layers (ML23017A089) demonstrated, the uncertainty with regard to long-term performance of the ULDL and HDPE/GCL composite barrier remains too high for the 2020 SDF PA deterministic values to be acceptable. The NRC staff addressed these issues with recommended monitoring activities related to the composite barriers and drainage layers, as provided in Section 4.1.3 in this TRR.

4.1.5 Development of the Probabilistic SDF Closure Cap Model

The purpose of the DOE GoldSim-based probabilistic SDF Closure Cap Model was to estimate uncertainty in long-term deep infiltration rates, especially in association with the ULDL, the cover HDPE/GCL composite barrier, and potential long-term impacts of erosion. Neither th at probabilistic model, nor the probabilistic model documented in the DOE document SRR -CWDA -

2021-00066 was intended to provide reasonable assurance that disposal at the SDF will comply with any specified performance objectives, but rather to provide insights into influences of uncertainties. Thus, the results of the review and evaluation by the NRC staff are in reference to the acceptability, or not, of the DOE probabilistic modeling ex ercises to estimate uncertainties and obtain risk insights and not in reference to obtaining reasonable assurance with regard to safety standards.

The 2020 SDF PA excluded the features, events, and processes (FEP) 2.7.08, Climate Change, which included the effects of long-term change in global climate and shorter -term change in regional and local climate. The DOE screened out the FEP because more discrete FEPs were identified which more accurately captured the relevant FEPs. The DOE performed sensitivity analyses with wetter conditions; however, the NRC staff found tho se DOE analyses did not fully represent potential conditions if the climate in the SRS region changes in the future since the analyses essentially used wetter years from the current climate, which underrepresents potential precipitation and associated effects if the climate chan ges. However, the DOE did include and analy ze climate change with the p robabilistic SDF Closure Cap Model (SRR-CWDA-2021-00040, Section 3.2) which is discussed in this TRR in more detail in Section 4.2.1.

The NRC staff found the DOE response to the NRC RSI with regard to the sand drainage layer (i.e., the ULDL) uncertainties to be acceptable because the technical bases provided in SRR-CWDA-2021-00031, Rev. 1 were sound and the distributions for the saturated hydraulic conductivity of the sand layers were sufficiently large to allow risk insights associated with saltstone degradation and dose to be gained. However, some of the technical bases from a stability aspect may be weak, that is, the NRC staff has concerns related to the stability of the ULDL due to perched water on top of the HDPE/GCL composite barrier layer and potential confined conditions in the ULDL. Specific concerns such as differential settlement, sand entrainment, and erosion and deposition within the ULDL are addressed in the NRC staf f TRR

- 39 -

entitled Site Stability (ML23017A114).

The NRC staff found the DOE response to the NRC RSI with regard to the initial defect frequencies, HDPE service life, and initial HDPE defect diameters to be acceptable because the technical bases provided in SRR-CWDA -2021- 000331, Rev. 1 were sound. The DOE probability density function for the HDPE failure condition was sufficient to allow risk insights associated with saltstone degradation and dose to be gained.

The NRC staff found the GCL uncertainty distributions for the probabilistic model the DOE provided in response to the NRC RSI to be acceptable because the technical bases provided in SRR-CWDA-2021-00033, Rev.1 were sound, and the distributions were sufficiently large to allow risk insights associated with saltstone degradation and dose to be gained.

The NRC staff found the implementation of shallow infiltration abstraction for the probabilistic SDF Closure Cap Model not to be acceptable due to multiple factors discussed below. Table 19 below summarizes the deterministic shallow infiltration rates and the nine modeling cases derived from the probabilistic model as discussed in Section 3.1.5 in this TRR.

Table 19. Shallow I nfiltration Rates for D ifferent Climates and D ifferent Moisture Conditions within those C limates (Based on information from the 2020 SDF PA (Section 4.4.1.4.2) and SRR-CWDA-2021-00040 ( from Table 4.4-1 in Section 4.4))

Deterministic Shallow Infiltration Shallow Infiltration Shallow Infiltration Model Rates Under Drier Rates Under Average Rates Under Wetter Precipitation Precipitation Precipitation Conditions Conditions Conditions

Drier Climate NA NA NA Present Climate NA 40.0 cm/yr (16 in/yr) 65.0 cm/yr (26 in/yr)

[based on the years 1991, 1992, 1997 and 1998]

Wetter Climate NA NA NA

Modeling Cases Shallow Infiltration Shallow Infiltration Shallow Infiltration from Rates Under Drier Rates Under Average Rates Under Wetter Probabilistic Precipitation Precipitation Precipitation Model Conditions Conditions (mean) Conditions Drier Climate 11 cm/yr (4.2 in/yr) 16 cm/yr (6.4 in/yr) 31.0 cm/yr (12.2 in/yr)

Present Climate 19 cm/yr (7.3 in/yr) 38.6 cm/yr (15.2 in/yr) 52.3 cm/yr (20.6 in/yr)

Wetter Climate 35.8 cm/yr (14.1 in/yr) 55.6 cm/yr (21.9 in/yr) 78.2 cm/yr (30.8 in/yr)

• NA = not available

There is a large uncertainty with regard to the long-term performance of the erosion barrier (see the NRC staff TRR entitled Site Stability (ML23017A114)). Although there appears to be no indication that a simulated chokepoint exists within the DOEs current models of flow, it is a possibility that a chokepoint could exist in the future, as-yet-to-be-constructed closure cap. As discussed in the NRC staff TRR entitled Performance of the Composite Barrier Layers and Drainage Layers (ML23017A089) and this TRR, it is possible that the existence and influence of the erosion barrier can have a large impact on the stability of the cover and on the shallow infiltration. The possibility exists that the tap roots of many future longleaf pines may extend into

- 40 -

to the erosion barrier and affect its cohesion over time. Tho se tap roots are capable of growing deeper than 2 m (6 ft), and successive forestation is expected to occur as loblolly pine and longleaf pine from the surrounding environment advance over the SDF closure cap ( see DOE document SRMC-CWDA -2022-00016).

Further, the NRC staff asked the DOE to provide additional information regarding the DOE claim that evaporation rates increase with erosion on the closure caps. In the DOE Response to RAI Question IEC-8 (SRMC-CWDA-2022-00016), the DOE stated that evaporation rates do not increase with increased erosion on the closure caps. The DOE indicated that it changed the position based on new insights having been gathered leading to a revised interpretation of the model results described in both the DOE documents SRR-CWDA -2021- 00040 and SRR-CWDA-2021-00081. The DOE model data (see Table 20 in this TRR) showed precipitation, soil thicknesses, and erosional depths have little influence on evaporation rates, although they do influence the rate of transpiration. However, it is not clear to the NRC staff what does affect the modeled evaporation rates. T he DOE stated in the DOE Response to the RAI Question IEC-8, that the WinUNSAT-H modeling software cannot extract data from the models showing the depth of evaporation (e.g., the percentage of evaporation that has its source deeper than the erosion barrier). As the DOE stated in their response, evaporation rates remain essentially unchanged although the rates do vary from precipitation case to precipitation case. As seen in this TTRs Table 20, evaporation rates are lower (20.0 in [ 51 cm] < 24.4 in [62 cm ]) when the precipitation rate is moderate, that is, not higher than 49 in or 124 cm (IC P49) and not lower than 44 in or 112 cm (HD P44). This appears to the NRC staff as a modeling artifact because there are no know natural phenomenon that could explain such an irregular correlation between evapo ration and precipitation rates.

- 41 -

Table 20. Summary of WinUNSAT -H Modeling Cases from a DOE R esponse to RAI Question Showing Shallow Infiltration (Percolation)

(from Table IEC-8.1 in DOE document SRR-CWDA-202 2-00016)

In the DOE Response to RAI Question IEC-8, the DOE stated that,

Instead of attributing the decreasing percolation rates as a function of increasing erosion to this concept of enhanced evaporation, it is clearly a function of the reduced soil storage capacity combined with an increase to surface runoff (as described further in the response to CC-6).

The DOE Response to CC-6 in SRMC -CWDA-2022-00016 also claimed that more water was being removed from the system via reduced soil storage capacity and enhanced surface runoff.

However, that explanation does not have a strong technical basis for two reasons. First, s urface

- 42 -

runoff, as the DOE had stated, is not a rate derived during the simulations but rather the difference that remains when the precipitation is greater than shallow infiltration. That is, evaporation is a function of the capacity of the soil/backfill to support infiltration and is not dependent on simulated surface runoff rates. Second, the concept of the ET cover was discussed previously in Section 4.1.3 in this TRR; the soil is there to store water from rainfall events and then promote water evaporation and transpiration during and after rainfall events.

That is, the ET cover system uses the water storage capacity of the soil to decrease the rate of shallow infiltration (EPA, 2003). Based on that concept, the less soil there is to store water, the less water is available to be removed by ET. While Table 20 above shows transpiration is decreasing with decreasing soil thickness, the rate of evaporation remains steady, even when the topsoil and upper backfill have been completely removed. It is not clear to the NRC staff what the driving factors are for the constant evaporation rates considering that there is less soil to hold the water from which the evaporation process would be able remove the water and prevent it from percolating downward. It is also not clear to the NRC staff why evaporation rates do not decrease when the riprap of the erosion barrier is exposed (i.e., complete er osion )

because NUREG/CR-7200 demonstrated that a layer of riprap will reduce the rate of evaporation from soil below it.

The DOE Response to CC-6 in SRMC -CWDA-2022-00016 also described a second chokepoint between the topsoil and the upper backfill where the difference in the hydraulic conductivity from the topsoil to the upper backfill layer (from 3.0E -05 m/s [10E-05 ft/s] to 5.0E -07 m/s [16E -07 ft/s]) acts to slow down the movement of the water from leaving the topsoil layer. In addition, the DOE stated that:

While the decreasing thicknesses also mean that the layers will require less water to become fully saturated, it also means that fully saturated conditions will exert a smaller amount of hydraulic head on the underlying materials. Hydraulic head can be an effective driver for flow, so decreasing the magnitude of the hydraulic head will slow down the flow rates.

The NRC staff saw no supporting evidence or basis for that DOE conceptual model. That concept would indicate that saturated conditions were the predominant driver of shallow infiltration so that stronger precipitation events would be required to produce s aturated conditions and thus produce shallow infiltration.

Although the features and process that control shallow infiltration rate are not as critical ly significant as the features and process that control deep infiltration, they are sufficiently risk -

significant to warrant further observation. Therefore, the NRC staff recommends monitoring the DOE method for calculating shallow infiltration, including the risk-significant processes and features that determine shallow infiltration rates, under a new monitoring factor in MA 2 (Infiltration and Erosion Control) under both POs 10 CFR 61.41 and 10 CFR 61.42.

Recommendation PECC-2 The NRC staff recommends opening a new medium -priority monitoring factor entitled Shallow Infiltration under MA 2 (Infiltration and Erosion Control) under PO §61.41 and PO §61.42. The NRC staff expects to close the new monitoring factor after the NRC determines that there is a sufficiently strong technical basis to support (1) the DOE method for estimating shallow infiltration rates and (2) the DOE understanding of the interrelationships and interdependencies between the processes that determine shallow infiltration rates such as evaporation, transpiration, runoff, and precipitation and relevant features such as the vegetation, topsoil, erosion barrier, middle backfill, and upper

- 43 -

backfill including the influence the upper backfill and its thickness has on shallow infiltration rates.

4.1.6 Probabilistic Closure Cap Model Results

Probabilistic Analyses

Although the NRC staff found some of the DOE efforts associated with the implementation of shallow infiltration abstraction for the probabilistic SDF Closure Cap Model not to be acceptable, and shallow infiltration affects deep infiltration, the NRC staff found the deep infiltration rates that the DOE presented in Section 6.1 of SRR-CWDA -2021- 00040 to be acceptable for use in the probabilistic analysis because the rates obtained were sufficiently broad to allow uncertainties to be estimated and risk insights to be obtained. As an example, at 1,000 years, the 2020 SDF PA had deep infiltration rates of 0.0091 cm/yr ( 0.0036 in/yr) for the compliance or MPAD case and 0.022 cm/yr (0.0087 in/yr) for wetter conditions while the deep infiltration rates at 1,000 years for the probabilistic SDF Closure Cap Model with partial HDPE failure had a mean rate of approximately 0.3 cm/yr (0.1 in/yr) for the compliance or most probable and defensible case and a maximum rate of approximately 38 cm/yr (15 in/yr).

Deep Infiltration Rates for Future SDF Modeling

Section 6.3 in the DOE document SRR-CWDA -2021-00040 presented conclusions and probabilistic distributions for parameter values that the DOE referred to as recommended values for future DOE modeling. In SRR-CWDA -2021-00040, the DOE concluded that the uncertainty analysis results indicate that a wide range of deep infiltration rates could be considered for future SDF PA modeling until parameters that affect the deep infiltration rates can be further refined. The NRC staff agrees with that DOE conclusion. Figure 7 and Figure 8 in this TRR show deep infiltration rates that the DOE referred to as recommended values for future modeling. Figure 8 shows the potential deep infiltration rates for future modeling of the compliance, realistic, and pessimistic cases; the equivalent deep infiltration rates used in the 2020 SDF PA; the equivalent deep infiltration rate used in previous SDF PAs (labeled as HELP); and the natural infiltration rate at the SRS.

The values that the DOE referred to as recommended deep infiltration rates for future modeling are orders of magnitude higher than those rates used in the 2020 SDF PA after circa 300 years. In addition, for the pessimistic case, the curve lies between the equivalent deep infiltration rate used in previous SDF PAs and the natural infiltration rate of the SRS after 100 years. For the compliance case, the curve comes very close to that of the previous SDF PAs after roughly 700 years. The NRC staff therefore found the deep infiltration rates that the DOE referred to as recommended rates in SRR-CWDA -2021-00040 to be acceptable for estimating uncertainties and obtaining risk insights associated with long-term deep infiltration rates.

Note that the values the DOE referred to as the recommended rates in Figures 6 and 7 are rates the DOE could potentially use for future modeling, and not the rates the DOE used to demonstrate reasonable assurance of compliance (dashed curves from Figure 6). Tho se latter rates were not found acceptable by the NRC staff as discussed in Section 4.1.4 in this TRR.

- 44 -

4.2 Overview of Potential Erosion near the Closure Cap

4.2.1 Inputs and Assumptions for Erosion Rate Calculations

4.2.1.1 Water Balance and Climate

The NRC staff found the approach and calculations by the DOE to determine a range of values for precipitation, ET, surface runoff, and shallow infiltration for the purpose of evaluating erosion in the area surrounding the SDF to be acceptable because the technical bas es for the values obtained were well supported. Relevant data were assembled and analyzed, previous studies and modeling exercises were evaluated and compared, assumptions made were supported, and bases were provided for the DOE parameter values and distributions.

4.2.1.2 PMP Estimates for the S RS

The NRC staff did not find the DOE basis for not determining PMPs for the wetter climate condition to be acceptable, as d iscussed furth er below. The NRC staff found the evaluation by the DOE to determine PMP rates for the unchanged climate condition (used in the Central Scenario) not acceptable as discussed in the TRR entitled Site Stability (ML23017A114).

As discussed in Section 3.2.1.2 in this TRR, the DOE did not apply the climate change factor to the PMP estimates for a wetter climate at the SRS because the DOE expected that it would be sufficient to apply the 10,000-year PMP value to assess uncertainty. The NRC staff asked the DOE to provide additional information regarding the DOE rationale for not using climate change factors to estimate PMP values for the wetter climate condition. The DOE provided additional information in the DOE document SRMC -CWDA-2022-00016. In the DOE Response to RAI Question IEC-1, the DOE stated that the current PMP approach already accounted for future precipitation uncertainty because the PMP estimates are not based solely on current records of precipitation, but instead the values are estimated based on an approach that implicitly accounts for uncertainties associated with geographic area and long-term return periods. That is, scaling PMP values to account for the potential influences of climate change would be effectively equivalent to double counting the potential influences of the uncertainty. The NRC staff disagrees with that DOE assessment because the two uncertainties discussed by the DOE are two different types of uncertainty. As discussed in Section 4.2.1.3 in this TRR below, the DOE assumed that any change to the climate at SRS will increase or decrease the mean annual precipitation by 400 mm (16 in) and multiplying the mean precipitation rate (125. 5 cm/yr [49.4 in/yr]) by the climate multiplier of 1.25 (see Table 14 in this TRR) results in a precipitation rate of 157 cm/yr (61.8 in/yr) for a wetter climate state. This rate compares to 133 cm/yr (52.4 in/yr) which uses observed rainfall data from 1989 to 1998 to represent a wetter condition or phase of the current climate state. Section 3.5.1 in SRR -CWDA -2021- 00036 discusses how the Eliasson Method for estimating PMP uses mean, standard deviation, high, and highest annual observed precipitation rates corresponding to each station during each year of record to obtain PMP at SRS. The alternative approach for estimating PMP described in Section 3.5.2 in SRR -CWDA -

2021-00036 is the National Oceanic and Atmospheric Administration Method for e stimating PMP values which includes spatial analyses of historic rainfall data to generate isohyetal maps.

In both cases, the PMP methodologies use actual observed or historical data of the current climate state to obtain the PMP. Although the DOE stated that PMP estimates are also based on an approach that implicitly accounts for uncertainties associated with geographic area and long-term return periods, the data relied upon explicitly is the observed data from the current climate state in recent times and not from observed data from a hypothetical future climate state which not possible. Uncertainty related to long-term return periods is associated with uncertainty

- 45 -

of the large precipitation events (i.e., infrequent, but unusually powerful rainstorms) in the future of the current climate state. However, if that climate does change ( i.e., becomes wetter or drier),

the uncertainty associated with the frequency of relatively large precipitation events in that future climate state will differ. The current PMPs used by the DOE are associated with an unchanged climate, not a changed climate, even if it does implicitly account for uncertainty associated with long-term return periods of the current climate state.

The DOE does consider sheet erosion under a changed climate, as can be seen in Table 3.2-7 of SRR-CWDA -2021-00040, whereby one of the RUSLE par ameters was multiplied by the climate change factor for precipitation (either 0.75 or 1.25). In addition, the DOE did consider wetter and drier conditions during wetter and drier climate states for shallow infiltration rates (see Table 19 in this TRR).

With regard to determining PMP rates for the unchanged climate condition, Section 3.2.1.2 in this TRR discusses the PMP results from WSRC -STI -2008- 00244 and used in the 2009 SDF PA. Tho se PMP results were not included in the discussion in the 2020 SDF PA. The DOE approach to estimate the PMP used in the current PA differ ed from the previous DOE approach.

The results between the two varied significantly. In the NRC staff TRR entitled Site Stability (ML23017A114), the NRC staff discussed this topic in more detail.

Relatively large precipitation events occurring during a plausible wetter climate state should be considered in some form due to their potential to cause significant erosion and in order to gain insights into potential future gullying. Therefore, the NRC staff intends to monitor the projected physical stability of the area surrounding the SDF under Central Scenario conditions as well as wetter and drier climate conditions to confirm that the projected level of erosional degradation is not expected to significantly impede the performance of the disposal facility. (See new monitoring factor presented in Section 4.2.2 in this TRR).

4.2.1.3 Potential Climate Change Impacts

The SRR-CWDA -2021- 00036 analysis also considered potential climate change impacts and, based on both that analysis and the observed climate transitions discussed (see Section 3.2.1.3 in this TRR), the DOE assumed that any change to the climate at SRS will increase or decrease the mean precipitation by 400 mm (16 in). Given the mean precipitation rate of 125 cm/yr (49.4 in/yr), using the wetter climate multiplier of 1.25 results in a precipitation rate of 156 cm/yr (1.25 x 125 cm/yr), or 61.8 in/yr (1.25 x 49.4 in/yr ). This compares with the typical years of 103 cm/yr (40.6 in/yr) and 119 cm/yr (47.1 in/yr) and the wetter condition of 133 cm/yr (52.4 in/yr) (based on rainfall from 1989 - 1998) as used in the 2020 SDF PA.

The NRC staff found the approach and calculations the DOE used in response to the NRC RSI (SRR-CWDA-2021-00036) to determine distributions for modeling climate change conditions and transition times for the purpose of evaluating erosion in the area surrounding the SDF to be acceptable because the parameter distributions allowed the probabilistic model to provide insights into the potential influences of combined uncertainties. The NRC staff found the DOE approach and calculations in the response for determining climate change factors for selected water balance parameters for the purpose of evaluating erosion in the area surrounding the SDF to be acceptable; however, additional support for the bas es (presented in Table 14 in this TRR) for each of the multiplier factors given in Table 14 in this TRR could result in values that better match data presented p revious studies and modeling exercises discussed in this TRR. In addition, the basis for the ET multiplier factor (see Table 14 in this TRR) indicated a linear relationship between ET and precipitation; however, Section 4.4.1.4.2 in the 2020 SDF PA

- 46 -

seems to contradict that basis by stating that the simulated SDF ET rate showed relatively little variability because ET was near its potential due to the humid climate of the SRS.

Nevertheless, the distributions and factors presented in the Section 3.6.3 in SRR-CWDA -2021-00036 are acceptable to the NRC staff for the uncertainty analyses because the rates obtained for the water balance parameters were sufficiently broad to allow uncertainties to be estim ated and risk insights to be obtained.

4.2.1.4 Estimating Potential Gully Erosion and Sheet and Rill Erosion

Estimating Gully Erosion

The NRC staff found the approach and calculations by the DOE to determine a range of parameter values used in the equations for calculating the actual velocity and the permissible velocity for the purpose of evaluating gully erosion in the area surrounding the SDF to be acceptable (with the exception of one parameter ) because the technical bases for the values obtained were supported. Tho se parameters included the flow concentration factor, the runoff coefficient, the drainage area, the flow depth, the maximum permissible velocity, and the velocity correction factor.

The parameter value that the NRC staff found not acceptable was the rainfall intensity, which is the PMP divided by the time of concentration. One of the parameters used to determine the time of concentration is percent-slope. The slope for the Z-Area hill slope is given as 18% in Table 6.2-4 form SRR -CWDA -2021- 00036. The DOE basis for the percent-slope were assumptions made in SRR-CWDA -2021-00036 based on Table 4.2-2, Figure 4.2-4 in that document, the discussions of the Z-Area cross sections in Section 5.6 in this TRR, and information in Table 6.1-1 in SRR-CWDA -2021- 00036. Tho se references did not provide the necessary technical support for the assumed percent-slope because they show that there are areas surrounding the SDF that have greater slopes than the values the DOE modeled.

Table 6.1-1 in SRR-CWDA -2021-00036 provides average gradients of the potential erosion pathways and not the gradient specific for the hill slopes. Section 5.6 in this TRR does not discuss hill slopes along the UTR, which have higher percent-slopes than hill slopes in other locations near the SDF. For example, Figure 4.2-4 in SRR-CWDA -2021- 00036 indicates that the hill slopes along the UTR have gradients ranging from 15% to 45%, so that an 18% slope would lie on the lower end of that range. Table 4.2-2 in that document also gave a similar range for these hill slopes: 15% to 40% (map unit symbol TuE and TuF). Therefore, the values that DOE picked to represent the average slope for the equations to determine gullying was not conservative and no explanation was provided.

The hill slopes along the UTR can undergo gullying in the current climate, as documented in the Waterfall Creek Area section of the DOE document SRR-CWDA-2019-00040 and by Photographs # 8J-6 and #8J -7 in that document. A lthough the Waterfall Creek Area is nearer to the F-Tank Farm and not directly adjacent to the SDF, the NRC staff determined erosion in the Waterfall Creek Area is relevant to erosion adjacent to the SDF because of the topographic and hydrogeologic similarities to the two areas along the UTR. Photograph # 8J-7 (mentioned above) shows a rock weir consisting of five steps built to minimize gully erosion. Photograph

1. 8D-10 of the Waterfall Creek Area (Figure 12, below) shows recent erosional activity that has increased the dimensions of an existing gully.

For these reasons, t he NRC staff recommends monitoring the assumed percent slope and rainfall intensity due to its importance when calculating gully erosion (see recommended monitoring activity in the NRC staff TRR entitled Site Stability TRR (ML23017A114 )).

- 47 -

Figure 12. Photograph from March 18, 2019, of an Existing Gully in the Waterfall Creek Area in the SRS. (Photograph # 8D-10 in SRR-CWDA-2019-00040)

Estimating Sheet and Rill Erosion

The NRC staff found the approach and calculations by the DOE to determine a range of parameter values used in the equation for calculating the spatial soil loss for the purpose of evaluating sheet and rill erosion in the area surrounding the SDF to be acceptable because the technical bases for the values obtained were supported. Tho se parameters and factors included the rainfall-runoff erosivity factor, the soil erodibility factor, the slope length factor, the slope steepness factor, the cover-management factor, and the erosion control practice factor. The NRC staff asked the DOE to provide additional information regarding the basis supporting the soil erodibility values in Table 4.2-4 and in Figure 4.2-5 in SRR -CWDA -2021-00036. The DOE provided additional information in the DOE document SRMC -CWDA -2022-00016. In the DOE Response to RAI Question IEC-3, the DOE stated that a single soil erodibility factor of 0.22 was assumed for the entire SDF based on the areas attributed to each soil unit and the respective soil erodibility factors reported in Table 4.2-4 of SRR -CWDA -2021-00036. The DOE calculated that value would change from 0.22 to 0.23 i f all the areas covered by the Troup and Lucy sands are assumed to have the same soil erodibility factors as other disturbed soils (0.28). The DOE stated that the resulting area average increase (i.e., from 0.22 to 0.23) would likely have negligible influence on the estimates for sheet and rill erosion. It is not clear to NRC staff if that would be the case because Section 3.4 in SRR-CWDA-2021-00040 show ed that the soil

- 48 -

erodibility factor exhibits strong positive corr elations to the side slope and top surface soil loss es on the SDF closure cap. In addition, the DOE did not show that weighing the soil erodibility values from Table 4.2-4 in SRR -CWDA -2021-00036 based on area was necessary and if this averaging significantly changes the degree of soil loss. Therefore, the NRC staff recommends monitoring information on parameter and factor estimates such as rainfall intensity and the soil erodibility factor for calculating soil loss due to gully, and sheet and rill, erosion. Discussions relating to new monitoring factors developed in the NRC staff TRR entitled Site Stability (ML23017A114) provide more detailed information on the monitoring activities associated with these parameters and factors.

4.2.2 Erosion Rate Results

Gully Erosion Results

The NRC staff found the DOE calculations providing gully erosion results for areas adjacent to SDF closure cap to be acceptable (with the exception of rainfall intensity parameter) ; however, the NRC staff found the results to be incomplete because the results for wetter and drier climate conditions were not presented by the DOE. In addition, as discussed in Section 4.2.1.4 of this TRR, the NRC staff will monitor DOE support for (1) the expected gullying that will take place on the hill slopes near the Z-Area (see Section 3.2.2 and Table 17 in this TRR) and (2) the DOE determination that gully erosion is not expected to significantly impede SDF performance.

Sheet and Rill Erosion Results

The NRC staff found the DOE calculations providing sheet and rill erosion results for areas adjacent to SDF closure cap to be acceptable; however, the NRC staff found the results to be incomplete. Although the results for wetter and drier climate conditions for the SDF closure cap were presented in Section 3.4 in the DOE document SRR-CWDA -2021- 00040, the results for wetter and drier climate conditions for the Z-Area hilltops and Z -Area hill slopes were not presented. Therefore, the NRC staff recommends modifying the text in the NRC SDF Monitoring Plan that describes when the NRC staff expects to close MF 2.02, Erosion Control of the SDF Engineered Surface Cover and Adjacent Area, in addition to increasing the priority of the monitoring factor due to the results of the evaluation caried out in NRCs TRR entitled Performance of the Composite Barrier Layers and Lateral Drainage Layers (ML23017A089).

Recommendation PECC-3 The NRC staff recommends updating the title of MF 2.02 to Erosion Control of the SDF Engineered Surface Cover and Adjacent Area under MA 2 (Infiltration and Erosion Control). The NRC staff recommends updating the priority of MF 2.02 from low to high under PO §61.41 and PO §61.42. The NRC staff recommends updating the text of MF 2.02 to indicate that the NRC staff expects to close MF 2.02 after the NRC determines that the projected level of erosional degradation of the SDF closure cap and the area adjacent to the SDF closure cap, is not expected to significantly impede the performance of the disposal facility under Central Scenario climate conditions as well as possible future wetter and drier climate states. Given the importance of construction activities on the performance of the final engineered surface cover, MF 2.02 will not be closed prior to construction of the closure cap.

5.0 Teleconference or Meeting

There were no teleconferences or meetings with the DOE related to this TRR.

- 49 -

6.0 Follow-up Actions

Besides recommendations to revise the NRC SDF Monitoring Plan, there are no specific Follow-up Actions related to this TRR.

7.0 Conclusions

The NRC staff evaluated the DOE analyses of percolation rates through, and potential erosion near, the SDF closure cap at the SRS. In addition, the NRC staff evaluated the technical bases and assumptions associated with these analyses.

The NRC staff found the long-term performance of the erosion barrier to be risk -significant for modeling both infiltration and erosion, and the currently assigned value for the fill material between the stones of the erosion barrier to be acceptable for modeling purposes. However, because the DOE has not yet selected the final material to fill the voids between the stones, the NRC staff recommended monitoring the simulated flow rate through the erosion barrier and the potential erosion and stability of the layers above the erosion barrier. T he NRC staff found the shallow infiltration estimates used in the 2020 SDF PA to be acceptable for deterministic modeling. In contrast, the NRC staff found the deep infiltration rate estimates not to be acceptable for determining compliance because the technical bases for the long-term performance of risk-significant cover barrier layers lacked confirmatory evidence and model support to match their risk-significance.

The purpose of the DOE GoldSim-based probabilistic SDF Closure Cap Model was to estimate uncertainty in long-term deep infiltration rates rather than to provide a compliance demonstration. With this clarification in mind, the results of th e review and evaluation by the NRC staff are in reference to the acceptability, or not, of the DOE probabilistic modeling exercises to estimate uncertainties and obtain risk insights, and not in reference to obtaining reasonable assurance with regards to s afety standards. The NRC staff found DOEs responses to the NRC RSI with regard to uncertainty distributions for the probabilistic mode involving the sand drainage layer, the initial defect frequencies, HDPE service life, initial HDPE defect diameters, and the GCL to be acceptable because the technical bases were sound, and the distributions were sufficiently large to allow risk insights associated with saltstone degradation and dose to be gained. However, some of the technical bases from a stability aspect may be weak with regard the ULDL, as discussed in detail in the NRC staff TRR entitled Site Stability (ML23017A114).

The NRC staff found the implementation of the shallow infiltration abstraction (also called percolation by the DOE) for the probabilistic SDF Closure Cap Model not to be acceptable due to the large uncertainty in evaporation rates, the long-term hydraulic performance of the erosion barrier, and multiple other factors. Therefore, the NRC staff recommended monitoring the DOEs conceptual model for shallow infiltration, including the risk-significant processes and features that determine shallow infiltration rates.

Although the NRC staff found some of the DOE efforts associated with the implementation of shallow infiltration abstraction for the probabilistic SDF Closure Cap Model not to be acceptable, and shallow infiltration affects deep infiltration, the NRC staff found the deep infiltration rates provided in response to the NRC RSI for future modeling to be acceptable for uncertainty analyses because the range included values that are orders of magnitude higher than deep infiltration rates used in the 2020 SDF PA (from circa 300 years after SDF closure). In addition,

- 50 -

for the pessimistic case, the curve after 100 years lies between the equivalent deep infiltration rate used in previous SDF PAs and the natural infiltration rate of the SRS. For the compliance case, the curve after roughly 700 years comes very close to that of the previous SDF PAs. The NRC staff therefore found these rates to be acceptable for estimating uncertainties and obtaining risk insights associated with long-term deep infiltration rates.

The NRC staff found the approach and calcula tions the DOE used to evaluate erosion near the SDF to be acceptable with two exceptions. The NRC staff found the range of parameter values the DOE used in the equations for calculating the actual velocity and the permissible velocity for the purpose of evaluating gully erosion in the area surrounding the SDF to be acceptable (with the exception of the rainfall intensity parameter) because the technical bas es for the values obtained were supported. Similarly, the NRC staff found the approach and calculations by the DOE to determine a range of parameter values used in the equation for calculating the spatial soil loss for the purpose of evaluating sheet and rill erosion in the area surrounding the SDF to be acceptable (with the exception of the soil erodibility factor) because the technical bas es for the values obtained were supported. The NRC staff recommended monitoring information related to parameters and factors such as rainfall intensity and soil erodibility factor for calculating the soil loss due to gully, and sheet and rill, erosion. Discussions over the new monitoring factors developed in the NRC staff TRR entitled Site Stability (ML23017A114) provide more detailed information on the monitoring activities associated with these parameters and factors.

Although the NRC staff found the calculations providing gully erosion results to be acceptable (with the exception of the rainfall intensity parameter), the NRC staff found the results to be incomplete because the DOE did not present results for wetter and drier climate conditions.

Similarly, the NRC staff found the calculations providing sheet and rill erosion r esults to be acceptable (with the exception of the soil erodibility factor) ; however, the NRC staff found the results to be incomplete. Therefore, the NRC staff recommended modifying MF 2.02, Erosion Control of the SDF Engineered Surface Cover and Adjacent Area, to include monitoring projected physical stability of the final engineered surface cover and the adjacent area under Central Scenario conditions as well as wetter and drier climate conditions.

The NRC staff recommends making the following changes to the NRC SDF Monitoring Plan after the Technical Evaluation Report is complete:

Recommendation PECC-1 Long-Term Erosion Barrier Performance

• The NRC staff recommends opening a new medium-priority monitoring factor entitled Long-Term Erosion Barrier Performance under MA 2 (Infiltration and Erosion Control) under the performance objectives of §61.41 and §61.42. The NRC staff expects to close the new monitori ng factor after the NRC determines that there is a sufficiently strong technical basis to support the simulated hydraulic performance above, below, and through the erosion barrier.

Recommendation PECC-2 Shallow Infiltration

• The NRC staff recommends opening a new medium-priority monitoring factor entitled Shallow Infiltration under MA 2 (Infiltration and Erosion Control) under the performance objectives of §61.41 and §61.42. The NRC staff expects to close the new monitoring factor after the NRC determines that there is a sufficiently strong technical basis to support (1) the DOE method for estimating shallow infiltration rates and (2) the DOEs

- 51 -

understanding of the interrelationships and interdependencies between the processes that determine shallow infiltration rates such as evaporation, transpiration, runoff, and precipitation and relevant features such as the vegetation, topsoil, erosion barrier, middle backfill, and upper backfill including the influence the upper backfill and i ts thickness has on shallow infiltration rates.

Recommendation PECC-3 Updating Erosion Control of the SDF Engineered Surface Cover and Adjacent Area

• The NRC staff recommends updating the title of MF 2.02 to Erosion Control of the SDF Engineered Surface Cover and Adjacent Area under MA 2 (Infiltration and Erosion Control). The NRC staff recommends updating the priority of MF 2.02 from low to high under the performance objectives of §61.41 and §61.42. The NRC staff recommends updating the text of MF 2.02 to indicate that the NRC staff expects to close MF 2.02 after the NRC determines that the projected level of erosional degradation of the SDF closure cap and the area adjacent to the SDF closure cap, is not expected to significantly impede the performance of the disposal facility under Central Scenario climate conditions as well as possible future wetter and drier climate states. Given the importance of construction activities on the performance of the final engineered surface cover, MF 2.02 will not be closed prior to construction of the closure cap.

8.0 References

Leigh, D.S., Late Quaternary Climates and River Channels of the Atlantic Coastal Plain, Southeastern USA, Geomorphology Vol. 101 (1), October 2008 DOI: 10.1016/j.geomorph.2008.05.024 (Copyright) https://www.researchgate.net/publication/252421337_Late_Quaternary_climates_and_river_cha nnels_of_the_Atlantic_Coastal_Plain_Southeastern_USA

South Carolina Department of Health and Environmental Control, Regulation 61-107.19 Solid Waste Management: Solid Waste Landfills and Structural Fill, May 2008.

U.S. Army Corps of Engineers, HMR -52: Application of Probable Maximum Precipitation Estimates: United States East of the 105th Meridian, 1982

[1]

U.S. Department of Agriculture (USDA), SCS-TP -61, Rev. 2, Handbook of Channel Design for Soil and Water Conservation, August 1966.

https://www.nrcs.usda.gov/wps/PA_NRCSConsumption/download?cid=nrcseprd1804858&ext=p df

___, USDA-HDBK-703, Rev. 0, Predicting Soil Erosion by Water: A Guide to Conservation Planning with the Revised Universal Soil Loss Equation (RUSLE). January 1997 https://www.ars.usda.gov/ARSUserFiles/64080530/RUSLE/AH_703.pdf

U.S. DOE, SRNL-ATG -2005-00022, Rev. 0, Memorandum: Probable Maximum Precipitation for the Z-Area Saltstone Facility, September 2005. ML17170A185

___, SRNL STI 2017 00008, Rev. 1, Groundwater Flow Simulation of the Savannah River Site General Separations Area, September 2017. ML18081A304

___, SRR CWDA 2009 00017, Rev. 0, 2009 Performance Assessment for the Saltstone

- 52 -

Disposal Facility at the Savannah River Site, October 2009. ML101590008

___, SRR-CWDA -2013- 00062, Rev. 2, F iscal Year 2013 Special Analysis for the Saltstone Disposal Facility at the Savannah River Site, October 2013. ML14002A069

___, SRR-CWDA -2014- 00006, Rev. 2, Fiscal Year 2014 Special Analysis for the Saltstone Disposal Facility at the Savannah River Site, September 2014. ML15097A366

___, SRR CWDA 2018 00030, Rev. 1, Recommended Values for Percolation Rates to Support Future SDF Modeling, November 2018. ML23058A051

___, SRR-CWDA -2019- 00001, Rev. 0, 2020 Performance Assessment for the Saltstone Disposal Facility at the Savannah River Site, March 2020. ML20190A056

___, SRR-CWDA -2019- 00040, Presentation: Savannah River Site F and T Tank Farms for NRC Onsite Observation Visit on March 18-19, 2019, Photos - GSA Streams Field Observation 3/18/2019, March 2019. ML19116A225

___, SRR-CWDA -2021- 00031, Rev. 1, Closure Cap Model Parameter Evaluation: Saturated Hydraulic Conductivity of Sand, May 2021. ML21160A061

___, SRR-CWDA -2021- 00033, Rev. 1, Closure Cap Model Parameter Evaluation: High Density Polyethylene (HDPE) and Geosynthetic Clay Liner (GCL) Composite Barrier Performance, May 2021. ML21160A062

___, SRR-CWDA -2021- 00036, Rev. 0, Evaluation of the Potential for Erosion in the Vicinity of Z-Area, June 2021. ML21160A063

___, SRR-CWDA -2021- 00040, Rev. 0, Evaluation of the Uncertainties Associated with the SDF Closure Cap and Long-Term Infiltration Rates, June 2021. ML21160A064

___, SRR-CWDA -2021- 00056, Rev 0, Evaluation of the Uncertainties Associated with Long-Term Saltstone Degradation, July 2021. ML21217A081

___, SRR-CWDA -2021- 00066, Rev. 0, Evaluation of the Combined Uncertainties Associated with the Long-Term Perf ormance of Saltstone Disposal Facility Flow Barriers, August 2021.

ML21217A083

___, SRR-CWDA -2021- 00072, Rev. 1, Comment Response Matrix for the Second Set of U.S.

NRC Request for Additional Information on the Performance Assessment for the Saltstone Disposal Facility at the Savannah River Site, November 2021. ML21321A087

___, SRR-CW DA-2022- 00016, Rev. 0, Comment Response Matrix for the Fourth Set of U.S.

NRC Request for Additional Information on the Performance Assessment for the Saltstone Disposal Facility at the Savannah River Site, April 2022. ML22118A297

___, SRRA107772-000009, Rev. 0, Predicting Long-Term Percolation from the SDF Closure Cap, Report No. GENV-18-05, University of Virginia School of Engineering, April 2018.

ML18170A244

___, SRRA162682-000002, Rev. 0, Predicting Long-Term Percolation from the HTF and FTF

- 53 -

Closure Caps, Report No. GENV-20 -09, University of Virginia School of Engineering, June 2020. ML21179C265

___, WSRC-STI -2006-00198, Rev. 0, Hydraulic Property Data Package for the E-Area and Z-Area Soils, Cementitious Materials, and Waste Zones, September 2006. ML101600380

___, WSRC-STI -2008-00244, Rev. 0, Saltstone Disposal Facility Closure Cap Concept and Infiltration Estimates, May 2008. ML083400069

U.S. Environmental Protection Agency (EPA), EPA 542-F 0 01, Fact Sheet on Evapotranspiration Cover Systems for Waste Containment Solid Waste and Emergency Response (5203P), February 2011. https://www.epa.gov/remedytech/fact-sheet-evapotranspiration-cover -systems-waste-containm ent

U.S. NRC, NUREG/CR-4651, Vol. 2, Development of R iprap Design Criteria by R iprap Testing in Flumes: Phase II, September 1988. ML20154Q812

___, NUREG-1623, Design of Erosion Protection for Long-Term Stabilization, Se ptember 2002.

ML022530043

___, Technical Evaluation Report for U.S. D OE Savannah River Site Draft Section 3116 Waste Determination for Salt Waste Disposal, December 2005. ML053010225

___, NUREG/CR 7028, Vol. 1, Rev. 0, Engineered Covers for Waste Containment: Changes in Engineering Properties and Implications for Long Term Performance Assessment, December 2011. ML12005A110

___, Technical Evaluation Report for the 2009 Performance Assessment for the Saltstone Disposal Facility at the Savannah River Site, Rev. 1, April 2012. ML121170309

___, NDAA-Waste Incidental to Reprocessing Monitoring Plan for the Savannah River Site Saltstone Disposal Facility, Rev. 1, September 2013. ML13100A113

___, NUREG/CR-7200, Rev. 0, Influence of Coupling Erosion and Hydrology on the Long-Term Performance of Engineered Surface Barriers, May 2016. ML16125A124

___, Report for J anuary 25, 2017, Onsite Observation Visit for the Savannah R iver Site Saltstone Disposal Facility, May 2017. ML17054C453

___, Technical Review: Hydraulic Performance and Erosion Control of the Planned Saltstone Disposal Facility Closure Cap and Adjacent Area, January 2018. ML18002A545

___, Technical Review: Performance of the Composite Barrier Layers and Lateral Drainage Layers for the 2020 Performance Assessment for the Saltstone Disposal Facility at the Savannah River Site, Rev. 1, April 18, 2023. ML23017A089

___, Technical Review: Site Stability for the 2020 Performance Assessment for the Saltstone Disposal Facility at the Savannah River Site, Rev. 1, April 18, 2023. ML23017A114