ML23017A087

Technical Review: Inventory for the 2020 Saltstone Disposal Facility Performance Assessment - CR

ML23017A087
Person / Time
Site:	PROJ0734
Issue date:	04/18/2023
From:	Christianne Ridge Division of Decommissioning, Uranium Recovery and Waste Programs
To:
Ridge A
Shared Package
ML23090A081	List: ML23017A083 ML23017A084 ML23017A085 ML23017A086 ML23017A087 ML23017A088 ML23017A089 ML23017A090 ML23017A113
References
eConcurrence 20230331-60018
Download: ML23017A087 (36)

Text

Enclosure Technical Review: Inventory for the Department of Energy 2020 Performance Assessment for the Saltstone Disposal Facility at the Savannah River Site Date

((date:long Reviewer Christianne Ridge, 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 radiological inventory used in the U.S. Department of Energy (DOE) 2020 Performance Assessment (PA) for the Saltstone Disposal Facility (SDF) at the Savannah River Site (SRS). The NRC staff performed this review 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. This technical review also supports NRC monitoring of the SDF under Monitoring Factor (MF) 1.01, Inventory in Disposal Structures, and MF 1.02 Methods Used to Assess Inventory, as detailed in the NRC staffs current plan for monitoring the SDF (NRCs Agencywide Documents Access and Management System [ADAMS] under Accession No. ML13100A113). The NRC staff reviewed both the radionuclide inventory projections the DOE used in models to support the 2020 PA for the SDF (SRR-CWDA-1019-00001, Rev. 0) (referred to as the 2020 PA in this document) and the processes the DOE used to develop the inventory. Those processes included radionuclide screening, calculation of best estimates of radionuclide activities, development of probabilistic distributions for inventory values, and performance of uncertainty and sensitivity analyses. The scope of the review included the concentrations and volumes of waste emplaced in the SDF, which directly affect the projected dose to a member of the public 100 meters (m) (330 feet) from the SDF boundary (referred to in this document as a member of the public). The radionuclide concentrations in saltstone are an input into the model the DOE used to project a dose to an individual who inadvertently intrudes into the SDF 100 years or more after site closure (referred to in this document as an inadvertent intruder). The NRC staff will review the geometric inputs to calculating the inventory encountered by an inadvertent intruder in a separate TRR because the geometric assumptions are part of the conceptual model for inadvertent intrusion.

2.0 Background

2.1 Overview This is the first NRC TRR on the SDF inventory. The NRC staff previously addressed the SDF inventory in the following documents: 2005 NRC Technical Evaluation Report (TER) for the SDF (ADAMS Accession No. ML053010225); 2005 NRC Request for Additional Information (RAI) on the 2005 DOE Draft Waste Determination for the SDF (ADAMS Accession No. ML051440589); 2007 NRC Monitoring Plan for the SDF (ADAMS Accession No. ML070730363); 2012 NRC TER for the SDF (ADAMS Accession No. ML121170309); 2012 NRC Letter from M. Satorius to T. Spears Acknowledging the DOE Response to the NRC Letter of Concern Regarding Disposal Activities at the SDF (ADAMS Accession No. ML12213A447); and 2013 NRC Monitoring Plan for the SDF (ADAMS Accession No. ML13100A076). The 2012 NRC TER for the SDF (referred to as the 2012 TER in this document) concluded that the NRC staff did not have reasonable assurance that the DOE disposal activities at the SDF would meeting the 10 CFR 61.41 performance objective, Protection of the General Population from Releases of Radioactivity. On April 30, 2012, the NRC issued a letter to the DOE and the South Carolina Department of Health and Environmental Control (SC DHEC) to inform those agencies of the NRC conclusion (referred to as the NRC Letter of Concern in this document). In response, the DOE provided the NRC with additional information that included revised projections of the inventory of technetium-99 (Tc-99) in Saltstone Disposal Structure (SDS) 1, 2 A, 2B, 3 A, 3B, 4, 5 A, and 5B. After the NRC issued the 2013 Monitoring Plan for the SDF (referred to as the Monitoring Plan in this document), the NRC staff reviewed the DOE Fiscal Year (FY) 2013 and FY 2014 Special Analyses for the SDF (ADAMS Accession Nos. ML14002A069 and ML15097A366). However, the NRC staff did not write TERs about those DOE analyses or address inventory in its RAI questions on those documents. Therefore, the 2012 TER and 2013 Monitoring Plan represent the most recent documented NRC staff review of the DOE SDF inventory. The remainder of this section provides key points from the documents listed above and the DOE FY 2013 and FY 2014 Special Analyses to provide background information on the DOE SDF inventory development. 2.2 The 2005 NRC TER, 2005 NRC RAI, and 2007 NRC Monitoring Plan Before the DOE began operating the Salt Waste Processing Facility (SWPF) at SRS in 2020, the DOE used two interim treatment processes to treat salt waste. The first process, called Deliquification, Dissolution, and Adjustment, (DDA) did not remove any dissolved radionuclides from the waste and also allowed a small amount of sludge that was entrained in the solid salt waste to be transferred in the salt waste and incorporated into saltstone. The second interim treatment process, which the DOE called the Actinide Removal Process/Modular Caustic Side Solvent Extraction Unit (ARP/MCU), had a decontamination factor of approximately 200 for cesium (Cs) and removed essentially all of the entrained sludge from the salt waste. The main issues related to inventory identified in the 2005 RAI and 2005 NRC TER were: (1) uncertainty in the contribution of entrained sludge from waste treated with the DDA process and (2) traceability of individual radionuclide inventory values. The uncertainty related to the contribution of entrained sludge decreased when the DOE finished treating waste with the DDA process and instead treated salt waste with the ARP/MCU. Although the DOE began using the ARP/MCU to treat waste in 2008, the DOE completed the final DDA treatment in September 2009. Therefore, inventory uncertainty due to potential sludge entrainment is no longer an active issue. The NRC staff monitors the traceability of individual radionuclide inventory values under MF 1.01 and MF 1.02 in the Monitoring Plan. In addition to sludge entrainment and data traceability, the 2007 Monitoring Plan also indicated that the staff would monitor removal efficiencies to assess whether doses were maintained as low as reasonably achievable, as required by the 10 CFR 61.41 performance objective for protection of the general population from releases of radioactivity. 2.3 The 2009 DOE PA, 2012 NRC TER, and 2013 NRC Monitoring Plan In the 2009 DOE SDF PA, the DOE identified Tc-99, iodine-129 (I-129), radium-226 (Ra-226), neptunium-237 (Np-237), and protactinium-231 (Pa-231) as key radionuclides. In response to an NRC RAI question, the DOE indicated that it had significantly overestimated the inventory of Ra-226 and its parent, thorium-230 (Th-230) because the actual concentrations of those radionuclides were below their detection limits (SRR-CWDA-2011-00044, Rev. 1). In the 2012 NRC TER, the NRC staff determined that it was acceptable for the DOE to develop the Ra-226 and Th-230 inventories based on the ingrowth from U-234 if there had not been significant inputs of thorium-bearing waste. The NRC Monitoring Plan indicates that the NRC staff will track the inventory of disposed waste into each disposal structure and evaluate the effects on dose from any exceedance of the projected inventories in the 2009 DOE PA. The structure-by-structure inventory comparison envisioned in the Monitoring Plan is not possible because the DOE significantly changed the number and design of the disposal structures at the SDF since the NRC issued the 2012 TER. Therefore, the NRC staff reviewed the changes in the projected SDF closure inventory in its entirety rather than on a structure-by-structure basis. Table 1 provides a comparison between the base case inventory the DOE used in its 2009 PA for the SDF and the most probable and defensible (MPAD) case the DOE used in the 2020 PA because the DOE identified both of those cases as the compliance case in their respective analyses. Table 1 includes both the original values provided in the 2009 SDF PA, which are based on an expected closure date of October 1, 2030, and values that reflect decay an ingrowth to the current projected closure date of January 1, 2037 to provide an equal basis for comparison to the 2020 PA inventory. The NRC staff calculated the revised inventory values using the GoldSim modeling platform and decay constants from the International Committee on Radiological Protection (ICRP) Report No. 107. The inventory estimates of several radionuclides increased by more than an order of magnitude between the 2009 and 2020 PA: americium-241 (Am-241), Am-242 m, californium-249 (Cf-249), chlorine-36 (Cl-36), curium-247 (Cm-247), potassium-40 (K-40), niobium-93 m (Nb-93 m), plutonium-238 (Pu-238), Pu-239, and strontium-90 (Sr-90) (see Table 1). Despite the significant increases in the modeled inventory of these radionuclides, none made a significant contribution to the modeled dose in any of the Central Scenario cases in the 2020 PA. The modeled inventory of three radionuclides decreased by an order of magnitude or more between the 2009 and 2020 PA: lead-210 (Pb-210), Ra-226, and Th-230. The decrease in the modeled inventory of those three radionuclides is related to changes the DOE made to its model of the inventory of the uranium decay series. That change is discussed further below, in the context of the FY 2013 Special Analysis. Table 1. Comparison of 2037 inventories the DOE used in the 2009 and 2020 PAs for the SDF. Radionuclide 2009 PA Base

Case, October 1, 2030 Closure Curies (Ci)a, b 2009 PA Base Case, Decayed to January 1, 2037 (Ci) a, c 2020 PA (Ci) a, d Realistic MPAD Pessimistic Ac-227 2.7x10-5 5.4x10-5 2.89x10-4 3.41x10-4 3.53x10-4 Al-26 1.3x101 1.3x101 2.52x101 3.65x101 3.92x101 Am-241 2.2x102 2.6x102 1.40 x104 2.06x104 2.21x104 Am-242 m 1.0x10-1 9.7x10-2 6.81 9.98 1.08x101 Am-243 4.2 4.2 6.17 8.84 9.45 Ba-137 m 2.8x105 (e)

(e) (e) (e) Bk-249 1.2x10-26 (e) (e) (e) (e) C-14 1.6x102 1.6x102 5.39x102 7.86x102 8.41x102 Ce-144 2.5x10-8 (e) (e) (e) (e) Cf-249 4.4x10-11 4.3x10-11 3.38x10-1 3.38x10-1 3.38x10-1 Cf-251 1.2 1.2 1.48x10-1 1.48x10-1 1.48x10-1 Cf-252 1.2x10-16 (e) (e) (e) (e) Cl-36 3.1x10-2 3.1x10-2 9.02x10-1 1.31 1.42 Cm-242 6.7x10-2 (e) (e) (e) (e) Cm-243 2.2x10-1 1.9x10-1 3.29x10-2 4.51x10-2 4.80x10-2 Cm-244 1.9x102 1.5x102 1.16x102 1.63x102 1.75x102 Cm-245 9.4x10-1 8.5x10-1 9.20x10-1 9.56x10-1 9.65x10-1 Cm-247 3.9x10-6 4.2x10-6 1.77x10-1 1.77x10-1 1.77x10-1 Cm-248 4.9x10-12 (e) (e) (e) (e) Co-60 3.9 1.7 7.54 1.11x101 1.19x101 Cs-134 5.2x10-1 (e) (e) (e) (e) Cs-135 5.4 5.4 3.12 3.67 3.81 Cs-137 3.0x105 2.6x105 3.44x105 4.47x105 4.71x105 Eu-152 6.4 4.6 7.88 1.16x101 1.24x101 Eu-154 1.3x102 7.8x101 7.32x101 1.07x102 1.15x102 Eu-155 9.0 (e) (e) (e) (e) H-3 2.2x103 1.5x103 3.07x103 4.50x103 4.84x103 I-129 2.5x101 2.5x101 1.59x101 1.66x101 2.44x101 K-40 3.1x10-2 3.1x10-2 9.03x10-1 1.32 1.42 Na-22 4.6 (e) (e) (e) (e) Nb-93 m 3.2x101 3.2x101 2.02x103 2.06x103 2.07x103 Nb-94 3.3x10-1 3.3x10-1 2.34x10-1 2.97x10-1 3.12x10-1 Ni-59 5.8 (e) (e) (e) (e) Ni-63 1.8x102 1.7x102 1.99x102 2.91x102 3.12x102 Np-237 3.8 3.8 1.23x101 1.77x101 1.89x101 Pa-231 1.6x10-4 2.5x10-4 4.51x10-4 5.81x10-4 6.17x10-4 Pb-210 no initial inventory 7.3x10-1 6.06x10-5 7.08x10-5 7.38x10-5 Pd-107 4.1x10-1 (e) (e) (e) (e) Radionuclide 2009 PA Base

Case, October 1, 2030 Closure Curies (Ci)a, b 2009 PA Base Case, Decayed to January 1, 2037 (Ci) a, c 2020 PA (Ci) a, d Realistic MPAD Pessimistic Pm-147 5.3 (e)

(e) (e) (e) Pr-144 2.5x10-8 (e) (e) (e) (e) Pt-193 8.1x101 7.4x101 8.13x101 1.12x102 1.21x102 Pu-238 2.0x104 1.9x104 1.43x105 2.10x105 2.26x105 Pu-239 1.3x103 1.3x103 8.88x103 1.30x104 1.40x104 Pu-240 3.8x102 3.8x102 1.91x103 2.78x103 2.98x103 Pu-241 5.1x103 3.8x103 1.68x104 2.47x104 2.65x104 Pu-242 1.1 1.1 8.94 1.03x101 1.07x101 Pu-244 1.7x10-2 1.7x10-2 3.91x10-2 4.55x10-2 4.71x10-2 Ra-226 4.1 4.1 1.65x10-4 1.94x10-4 2.02x10-4 Ra-228 5.6x10-3 5.0x10-2 2.56x10-1 3.77x10-1 4.0510-1 Rh-106 7.8x10-5 (e) (e) (e) (e) Ru-106 7.8x10-5 (e) (e) (e) (e) Sb-125 2.1x101 (e) (e) (e) (e) Sb-126 7.8x101 (e) (e) (e) (e) Sb-126 m 5.2x102 (e) (e) (e) (e) Se-79 1.4x102 1.4x102 1.00x102 1.42x102 1.52x102 Sm-151 3.8x103 3.6x103 3.86x103 5.66x103 6.09x103 Sn-126 5.3x102 5.3x102 3.56x102 5.18x102 5.59x102 Sr-90 2.4x105 2.1x105 3.60x106 5.30x106 5.70x106 Tc-99 3.5x104 3.5x104 2.24x104 3.29x104 3.55x104 Te-125 m 5.2 (e) (e) (e) (e) Th-229 2.8x101 2.8x101 3.71 3.73 3.74 Th-230 2.0x101 2.0x101 1.29x10-2 1.50x10-2 1.55x10-2 Th-232 9.0x10-2 9.0x10-2 2.56x10-1 3.77x10-1 4.06x10-1 U-232 6.4x10-2 6.0x10-2 1.86x10-1 2.02x10-1 2.06x10-1 U-233 2.7x101 2.7x101 2.51x101 3.04x101 3.18x101 U-234 3.5x101 3.5x101 3.12x101 4.01x101 4.23x101 U-235 6.7x10-1 6.7x10-1 5.67x10-1 7.84x10-1 8.38x10-1 U-236 1.8 1.8 1.16 1.65 1.77 U-238 7.0 7.0 1.53x101 2.24x101 2.41x101 Y-90 2.4x105 (e) (e) (e) (e) Zr-93 3.2x101 3.2x101 9.85x101 1.39x102 1.50x102 (a) To convert Ci to Becquerel (Bq), multiply by 3.7x1010. (b) From Table 3.3-7 of SRR-CWDA-2009-00017, Rev. 0. (c) The NRC staff decayed the values in column 1 to the current projected closure date using the GoldSim modeling platform with decay constants from the ICRP Report No. 107. (d) From the 2020 PA, Tables 3.3-5 (realistic), 3.3-6 (MPAD), and 3.3-7 (pessimistic). (e) Screened out of the 2020 PA (See Section 3.2 of this TRR). 2.4 The NRC 2012 Letter of Concern, 2012 DOE Response, and NRC 2012 Letter of Acknowledgment In the 2012 TER, the NRC staff concluded that it did not have reasonable assurance that the DOE disposal activities at the SDF would meet the 10 CFR 61.41 performance objective, Protection of the General Population from Releases of Radioactivity. On April 30, 2012, the NRC issued a Letter of Concern to the DOE and the SC DHEC to inform those agencies of the NRC conclusion. In response, the DOE provided the NRC with a revised model of SDF performance and a revised projection of the inventory of technetium-99 (Tc-99) in SDS 1, 2 A, 2B, 3 A, 3B, 4, 5 A, and 5B (SRR-CWDA-2012-00002, Rev. 0; SRR-CWDA-2012-00095, Rev. 1). In a letter dated August 31, 2012 (ADAMS Accession No. ML12213A447), (referred to as the NRC Letter of Acknowledgement in this document), the NRC staff indicated that the revised Tc-99 inventories the DOE provided for SDS 2 A, 2B, 3 A, 3B, 5 A, and 5B alleviated the NRC staff concern related to the 10 CFR 61.41 performance objective at that time, and that the NRC staff would continue to monitor several technical issues related to projected SDF performance. Specifically, the NRC stated that if the projected inventories for those disposal structures was correct, the disposal structures would be unlikely to cause an off-site peak dose exceeding the requirements of §61.41 (i.e., 0.25 mSv/yr (25 mrem/yr)). In the FY 2014 Special Analysis, the DOE increased the modeled inventories of Tc-99 in SDS 2B, 3 A, 3B, 4, 5 A, and 5B, as compared to the values the DOE projected in its response to the NRC Letter of Concern. In the transmittal letter for the FY 2014 Special Analysis (ADAMS Accession No. ML14322A259), the DOE stated: Disposal activities at the SDF during FY 2012 through FY 2014 have been at these lower Tc-99 inventory values [that the DOE projected in response to the NRC Letter of Concern] and DOE will continue to supply the quarterly Tank 50 sample results to NRC for your information. Based upon the results of the FY 2014 [Special Analysis], DOE has authorized disposal of Tc-99 at the analyzed inventory (see Table 3.4.1 of the FY 2014 [Special Analysis]). Table 2, below, provides the Tc-99 inventories from the DOE FY 2014 Special Analysis in comparison to the lower values the DOE projected in response to the NRC Letter of Concern and the values the DOE used in models supporting the 2009 and 2020 PAs. Section 2.6 of this TRR addresses the remaining inventory for the FY 2014 Special Analysis. The 2020 PA indicates that the DOE considers SDS 1, 2 A, 2B, 3B, 4, 5 A, and 5B to be operationally filled, and that SDS 3 A is 15 percent (%) full. Therefore, the Tc-99 inventories for those disposal structures in Table 2 represent emplaced saltstone and the inventory for SDS 3 A represents a new inventory projection. Because the inventories for all the disposal structures except for SDS 3 A are based on emplaced saltstone, they are the same in all the DOEs Central Scenario projections (i.e., the realistic, MPAD, and pessimistic projections). The value for SDS 3 A is from the MPAD projection. Table 2. Comparison of projected Tc-99 inventories in the DOE 2009 PA and 2012 DOE response to the NRC Letter of Concern with emplaced and projected inventories modeled in the 2020 PA for SDS 1, 2 A, 2B, 3 A, 3B, 4, 5 A, and 5B. Disposal Structure 2009 PA Base Case (Ci) (a) DOE Response to the NRC Letter of Concern (Ci) (a, b) DOE FY 2014 Special Analysis (Ci) (a, c) 2020 PA (Ci) (a, c) 1 110 55 55 49.3 2 A 540 105 to 120 120 114 2B 540 105 to 120 140 137 3 A 540 98 to 117 540 363 3B 540 98 to 117 540 380 4 580 530 640 634 5 A 540 98 to 117 540 175 5B 540 98 to 117 540 120 (a) To convert Ci to Bq, multiply by 3.7x1010. Because of the long half-life of Tc-99, the NRC staff did not account for differences in the assumed year of site closure in these values. (b) Values for SDS 1 and SDS 4 are from SRR-CWDA-2012-00002, Rev. 0. The remaining values are from SRR-CWDA-2012-00095, Rev. 1. (c) Except for the value for SDS 3 A, values represent emplaced saltstone in operationally filled disposal structures. The value for SDS 3 A is from the MPAD projection. 2.5 The DOE Fiscal Year 2013 Special Analysis In the FY 2013 Special Analysis, the DOE identified Tc-99, I-129, and Cs-135 as key radionuclides. The DOE FY 2013 Special Analysis used the same inventory as the DOE 2009 PA with four exceptions: Pu-238 decreased by 40% Ra-226 decreased by 99.96% Th-230 decreased by 96% U-234 decreased by 46%. The DOE provided the basis for the adjustment to the inventories of Ra-226 and its parents in response to an NRC RAI on the 2009 PA (SRR-CWDA-2011-00044, Rev. 1) and a supporting reference (SRR-CWDA-2011-00115, Rev. 0). The largest modeled decrease occurred for Ra-226 and Th-230 because those radionuclides had concentrations below detection limits in most salt waste samples. The DOE revised its calculated inventories for Ra-226 and Th-230 by assuming their inventories could be determined based on ingrowth from U-234 instead of by assuming Th-230 and Ra-226 were present at their detection limits (SRR-CWDA-2011-00044, Rev. 1). 2.6 The DOE Fiscal Year 2014 Special Analysis The modeled inventory of most radionuclides in the FY 2014 Special Analysis was significantly greater than the modeled inventories in the 2009 SDF PA, FY 2013 Special Analysis, or 2020 PA. For radionuclides that the DOE did not consider to be risk-significant in the initial phases of the FY 2014 Special Analysis, the DOE assigned the entire SRS Tank Farms (i.e., F-Tank Farm and H-Tank Farm) inventory in supernate and salt waste (i.e., all phases except sludge solids) to each of the 15 disposals structures modeled in that assessment with two exceptions: the DOE assumed a decontamination factor of 200 for Cs-137; and the DOE reduced the modeled disposal structure inventories of Pu-238, Pu-239, and Pu-240 so that the concentrations would be less than the Low-Level Waste Class C limit in 10 CFR Part 61 because the DOE would not allow the actual concentrations to exceed the Class C limit. In the FY 2014 Special Analysis, the DOE identified Tc-99, I-129, Ra-226, and K-40 as key radionuclides because they caused a projected dose greater than 0.0025 millisieverts per year (mSv/year) (0.25 millirem per year (mrem/year)) to a member of the public within 20,000 years of SDF closure. However, the DOE also indicated that it assigned overly conservative projected inventories to K-40 and Ra-226 because it did not consider them to be risk-significant during inventory development. In the analysis of results in the FY 2014 Special Analysis, the DOE excluded Ra-226 and K-40 from the list of risk significant radionuclides. The DOE instead listed Tc-99 and I-129 based on their projected dose to a member of the public and listed Cs-135 based on its projected dose to an inadvertent intruder. 3.0 Inventory Development for the 2020 PA 3.1 Overview For the 2020 PA, the DOE developed three inventory projections as part of its Central Scenario: realistic, MPAD, and pessimistic. The DOE described the realistic projection as its best estimate of the SDF inventory at the time of closure. The DOE described the MPAD inventory as more easily defensible than the realistic case, and used it in the deterministic compliance case in the 2020 PA. Finally, the DOE described the pessimistic projection as biased toward increasing dose results and maximizing defensibility. The NRC staff used the terms realistic, MPAD, and pessimistic throughout this TRR to facilitate comparison with cited DOE tables and figures. However, the purpose of this NRC staff review is to determine whether the MPAD inventory is acceptable for use in models supporting the 2020 PA. The NRC staff has not determined whether the MPAD inventory is the most probable and defensible inventory, or whether the projections the DOE refers to as the realistic and pessimistic are realistic or pessimistic, respectively. The DOE considered three sources of inventory: (1) the existing SDF inventory as of March 31, 2018, (2) the Tank Farms inventory in certain waste phases as of January 8, 2018, as modified by Cs removal, and (3) the projected transfers from H-Canyon from FY 2019 through FY 2026, as modified by Cs removal. For the realistic, MPAD, and pessimistic inventories, the DOE used the same projections of the existing SDF inventory and the projected contributions from H-Canyon. Therefore, all the difference between the realistic, MPAD, and pessimistic inventories results from differences in how the DOE treated the uncertainty in the projected contribution to the SDF from waste in the Tank Farms. Section 3.3 of this TRR addresses the general process the DOE used to develop radionuclide inventories for all radionuclides. Section 3.4 of this TRR addresses additional calculations the DOE used to refine the inventory projection for I-129 and Tc-99. The 2020 PA indicates that the DOE performed those calculations because I-129 and Tc-99 dominate the dose to a member of the public and an inadvertent intruder in the 2020 PA. 3.2 Radionuclide Screening To determine which radionuclides to include in detailed technical analyses in the 2020 PA, the DOE began with an initial list of 80 radionuclides based on process knowledge of the waste streams entering the tank farm and historical measurements. Figure 1 shows the DOE screening process to narrow down the list of 80 radionuclides into a list of 52 radionuclides of concern to include in detailed analyses (SRR-CWDA-2018-00044, Rev. 3). Figure 1. Flow chart of the DOE screening process for the 2020 PA (Figure 1 in SRR-CWDA-2018-00044, Rev. 3). As shown in Figure 1, the screening process considered the radionuclide half-life, progeny, volatility, and the potential dose in conservative screening calculations based on the expected inventory to determine which radionuclides to screen out from further analysis. The screening scenarios included scenarios relevant to both a member of the public and an inadvertent intruder. In those screening calculations (i.e., third and fourth diamonds from the top on the left side of Figure 1), the DOE assumed that an individual directly consumes, inhales, showers in, or is directly exposed to soil, water, air with radionuclide concentrations equal to saltstone grout. The DOE then compared the results to a 0.25 mSv (25 mrem) annual dose limit for a member of the public and a 5 mSv (500 mrem) annual dose limit for an inadvertent intruder. The DOE used the screening calculations for waste burial developed by the National Council on Radiation Protection and Measurement (NRCP) in NCRP Report No. 123 to identify radionuclides to deliberately screen in and did not use the NRCP by itself to screen out any radionuclides (see Figure 1). In addition to the screening process described above, the DOE screened out molybdenum-93 (Mo-93) based on analyses with the GoldSim model that supported the FY 2014 Special Analysis for the SDF that showed low dose projections for a member of the public and an inadvertent intruder (SRR-CWDA-2015-00020, Rev. 0). Finally, the DOE omitted palladium-107 (Pd-107) from detailed analyses although it did not meet the screening criteria. However, based on a comparison of the inventory, dose conversion factors, and transport properties of Pd-107 and I-129, the DOE expects the projected dose to either a member of the public or an inadvertent intruder from Pd-107 would be at least four orders of magnitude less than the projected dose from I-129 (SRR-CWDA-2021-00047, Rev. 0). Table 3 shows the result of the result of the screening process. Table 3. Screening decisions in the 2020 PA for radionuclides the DOE determined could potentially be in saltstone (adapted from Tables 1, 2, and 3 of SRR-CWDA-2018-00044, Rev.3). Radionuclide Screening Decision H-3 C-14 Al-26 Cl-36 K-40 Co-60 Ni-63 Se-79 Sr-90 Zr-93 Nb-93 m Nb-94 Tc-99 Pd-107(a) Sn-126 I-129 Cs-135 Cs-137 Sm-151 Eu-152 Eu-134 Pt-193 Pb-210 Ra-226 Ac-227 Ra-228 Th-229 Th-230 Pa-231 Th-232 U-232 U-233 U-234 U-235 U-236 Np-237 Pu-238 U-238 Pu-239 Pu-240 Pu-241 Am-241 Pu-242 Am-242 m Am-243 Cm-243 Cm-244 Cm-245 Pu-244 Cm-247 Cf-249 Cf-251 Screened in. Na-22 Ru-106 Rh-106 Te-125 m Sb-125 Sb-126 Cs-134 Pr-144 Ce-144 Pm-147 Lu-174 Am-242 Screened out because of short half-life without significant progeny. Cm-242 Bk-249 Cf-252 Screened out because of short half-life and insufficient inventory (compared to progeny inventory) to significantly contribute to progeny inventory. Radionuclide Screening Decision Y-90 Ba-137 m Sb-126 m Screened out because radionuclide is accounted for in modeling by assuming it is in secular equilibrium with parent. Ca-41 Ni-59 Sm-147 Gd-152 Eu-155 Cm-248 Screened out because of low dose potential in screening scenarios based on available inventory, absence of significant progeny, and lack of volatility. Ag-108 m Bi-210 m Cm-246 Screened out because no inventory data was available and because one or more members of the decay chain are tracked in the radionuclide or chemical inventory. Mo-93 Screened out because of low dose potential in calculations performed for the FY 2014 Special Analysis. (a) Although the DOE initially screened Pd-107 in, the DOE later determined Pd-107 had low dose potential (SRR-CWDA-2021-00047, Rev. 0). 3.3 General Process for Inventory Development The first contribution to the projected SDF closure inventory is the existing inventory in the SDF. The DOE tracks the waste disposed of in the SDF with the Waste Inventory Disposed Estimator (SDF-WIDE) model (SRR-CWDA-2015-00003, Rev. 0). Section 3.7 of this TRR provides more information about that DOE model. For the 2020 PA, the DOE revised its estimate of the emplaced inventory in the SDF by reevaluating some of the alternative methods the SDF-WIDE model uses to calculate radionuclide concentrations when direct concentration measurements are not available (e.g., ingrowth and decay calculations, ratios to other radionuclides). For the 2020 PA, the DOE used the SDF-WIDE model with historical information about waste transfer volumes, historical measured radionuclide concentrations, and revised alternative methods to recalculate the emplaced SDF inventory as of March 31, 2018 (SRR-CWDA-2018-00041, Rev. 3). As noted in Section 3.1 of this TRR, the DOE used the same estimate of the existing inventory in the SDF for the realistic, MPAD, and pessimistic inventory projections. The second source of waste that the DOE considered was the existing inventory in the Tank Farms. As in the 2009 SDF PA, the DOE used the Waste Characterization System (WCS) as the primary source of information about the Tank Farms inventory. The WCS uses a combination of sample data, materials accounting, assumptions, and special calculation methods to track radionuclide inventories in each phase of waste (i.e., supernate, interstitial liquid, salt, and sludge) in each tank. Like the alternative methods used in the SDF-WIDE model, the special calculations used in the WCS include using detection limits, calculating ingrowth from measured ancestor radionuclides, using solubility information to estimate supernate concentrations, applying partitioning between sludge and supernate observed in some waste tanks to other tanks, or using correlations to more easily detectable radionuclides. For the 2020 PA, the DOE used data taken from the WCS on January 8, 2018. The DOE supplemented that information with reports of waste transfers to, within, and from the Tank Farms through March 1, 2018 (see SRR-CWDA-2018-00041, Rev. 3, Table A-2). Unlike the SDF-WIDE model, which uses data from samples of waste that have undergone Cs removal, the WCS tracks radionuclide inventories in untreated tank farm waste. Therefore, to use the information in the 2020 PA, the DOE applied a decontamination factor of 200 to the projected inventories of all isotopes of Cs. That assumption is consistent with operating experience from the ARP/MCU and the DOE expects that it will be conservative for waste treated in the SWPF. Although the DOE indicated that it did not assume any solids filtering would occur during ARP/MCU or SWPF treatment (SRR-CWDA-2018-00041, Rev. 3), the DOE also did not include any sludge inventory in the contributions to the SDF inventory for radionuclides other than I-129 and Tc-99. Instead, the DOE considered the radionuclide inventory the soluble phases of tank waste (i.e., supernate, interstitial liquid, and dry salt) because those are the phases of waste that the DOE plans to send to the SDF. However, for both I-129 and Tc-99, the DOE included a contribution from sludge in addition to the soluble phases (see Section 3.4 of this TRR). As noted in Section 3.1 of this TRR, the difference between the realistic, MPAD, and pessimistic inventory projections results from differences in the DOE projection of the contribution from the existing Tank Farms waste. For radionuclides other than I-120 and Tc-99, the DOE based the realistic, MPAD, and pessimistic inventory projections on the following assumptions about the contributions from existing waste in the Tank Farms (SRR-CWDA-2018-00041, Rev. 3): the realistic projection includes the DOE estimate of the inventory in supernate, interstitial liquid, and dry salt, as modified by cesium (Cs) removal; the MPAD projection includes the DOE estimate of the inventory in interstitial liquid plus 1.5 times the DOE estimate of the inventory in supernate and solid salt, as modified by Cs removal; and the pessimistic projection is 1.6 times the realistic projection. The final component of the projected closure inventory for the SDF was the contribution to the SDF (i.e., via the Tank Farms) from future effluents from H-Canyon. The DOE estimated that contribution by multiplying projected radionuclide concentrations in H-Canyon waste by the anticipated volume of annual transfers from H-Canyon to the Tank Farms. For the 2020 PA, the DOE used the same projected radionuclide concentrations in H-Canyon waste that the DOE used in the 2009 SDF PA. As it did for the calculation of the SDF inventory contribution from the existing Tank Farms waste, the DOE applied a decontamination factor of 200 to the projected contribution from H-Canyon to account for salt waste treatment. The DOE projected a total transfer volume of 9,270 cubic meters (m3) (2.45 x 106 gallons), based on the annual transfer volume estimates in the DOE Liquid Waste System Plan (SRR-LWP-2009-00001, Rev. 20) for FY 2018 through FY 2026, when the DOE anticipated transfers from the H-Canyon to the Tank Farms would end. After the DOE established the projected inventory for the SDF for use in the 2020 PA, the DOE issued a revision of the Liquid Waste System Plan (SRR-LWP-2009-00001, Rev. 21), which extended the duration of transfers from H-Canyon to the Tank Farms until FY 2030. Accounting for the lower-than-projected transfers from H-Canyon to the Tank Farms in FY 2018 through FY 2020 (SRR-CWDA-2021-00147, Rev. 1), the revised Liquid Waste Treatment Plan increased the projected volume transferred from H-Canyon to the Tank Farms from FY 2018 to FY 2030 to 11,800 m3 (3.11x106 gallons) (i.e., a 27% increase). 3.4 Additional Calculations for I-129 and Tc-99 In addition to the general approach for inventory development described in the previous section, the DOE performed additional calculations for I-129 and Tc-99. The 2020 PA indicates the DOE performed those additional calculations because I-129 and Tc-99 dominate the projected dose to a member of the public and an inadvertent intruder within 10,000 years of site closure. To reduce the uncertainty in the projected SDF closure inventory of I-129 and Tc-99, the DOE re-evaluated its algorithms for I-129 and Tc-99 concentrations in Tank Farms waste. The DOE has many more measurements of Cs-137 concentrations in Tank Farms waste than it has of either I-129 or Tc-99 concentrations. In addition, I-129 and Tc-99 are sometimes present in tank waste at concentrations below their respective detection limits. To reduce uncertainty in the I-129 and Tc-99 inventories, the DOE developed relationships between Cs-137 and both I-129 and Tc-99 to supplement measured concentrations of I-129 and Tc-99. In addition to using the relationship with Cs-137 to estimate I-129 and Tc-99 concentrations in samples in which the DOE measured Cs-137 but did not analyze for I-129 or Tc-99, the relationships were used to estimate I-129 and Tc-99 concentrations in samples of tank farm waste when I-129 and Tc-99 concentrations were less than the detection limit. Prior to the 2020 PA, the DOE calculated both I-129 and Tc-99 concentrations as constant fractions of measured Cs-137 activity. However, because Cs-137 has a much shorter half-life than either I-129 or Tc-99 does (i.e., 30 years for Cs-137, 15.7 million years for I-129, and 112,000 years for Tc-99), the relationships between Cs-137 and the other two radionuclides changed as the waste aged. In 2014, the DOE adjusted the relationship between measured Cs-137 and estimated concentrations of the other two radionuclides; however, the DOE still represented the I-129 and Tc-99 concentrations as a constant fraction of Cs-137 concentration and did not account for Cs-137 decay. To revise the relationships between measured Cs-137 and projections of I-129 and Tc-99 concentrations in soluble tank farm waste for the 2020 PA, the DOE re-analyzed data from tank farm samples taken between 1976 and 2017 (SRR-CWDA-2015-00077, Rev. 2 and SRR-CWDA-2015-00123, Rev. 2). The DOE used samples with detectable concentrations of either I-129 or Tc-99 to develop relationships with measured Cs-137 concentrations that could be used to project I-129 or Tc-99 concentrations in samples in which I-129 or Tc-99 was not measured or when the I-129 or Tc-99 concentrations were below their detection limits. Using a different process than it did in previous analyses, the DOE first calculated decayed values for all the measured Cs-137 concentrations accounting for decay to October 1, 2032. By accounting for the effects of Cs-137 decay before evaluating the relationship between I-129 and Cs-137, the DOE should be able to avoid the need for further adjustments to the relationship between I-129 and Cs-137 unless tank waste characteristics change (e.g., introduction of different waste streams with significantly different I-129, Tc-99, or Cs-137 concentrations). Before quantifying the relationships between I-129, Tc-99, and the decayed values of Cs-137, the DOE removed four types of outliers from the data: concentrations in samples from Tank 50, which had been treated to remove Cs-137 but not I-129 or Tc-99; data from samples taken after bulk waste removal (i.e., residual waste samples), which the DOE expected did not accurately represent waste that will be sent to the SDF; measurements from samples with chemical properties that the DOE did not expect would represent waste to be sent to the SDF, including waste that had undergone chemical treatment and waste in which I-129 might have been bound to mercury; and data with reporting issues, including values that did not appear to account for sample dilution or that did not match the underlying reference. After removing the outliers, the DOE made three additional adjustments to the data: removal of values that represented either I-129 or Tc-99 detection limits instead of measured concentrations; averaging values for samples that represented multiple measurements of essentially the same waste (i.e., samples of the same waste phase from the same tank on or near the same date); and removal of very low concentration samples. The DOE removed data points that represented I-129 or Tc-99 detection limits because they did not represent the relationship between the true concentration of those radionuclides in the samples and the measured Cs-137 concentration. The DOE averaged values that represented samples of the same phase of waste from the same tank on or near the same day to avoid giving those samples undue weight in the relationship. Finally, the DOE removed several samples with Cs-137 concentrations below 3.7x105 becquerel per liter (Bq/L) (1x107 picocuries per liter [pCi/L]) for two reasons: (1) the dilute samples represent waste that will not add significantly to the SDF inventory, and (2) the data lowered the projected values of I-129 and Tc-99 for a given Cs-137 concentration (i.e., eliminating those values caused a conservative increase in projected I-129 and Tc-99 inventories). Although the DOE observed one additional apparently anomalous Tc-99 data point, the DOE retained the value in its analysis because the underlying reference showed no clear justification for removing it and retaining it was conservative (i.e., increased projected Tc-99 concentrations). After normalizing the data as described above, the DOE used the remaining values to derive relationships between the measured Cs-137 and I-129 or Tc-99 concentrations in aqueous phases (i.e., supernate and interstitial liquid). The DOE found the data varied logarithmically and used linear regression between the log-transformed concentrations to determine the best-fit lines to project either I-129 or Tc-99 as a function of measured Cs-137. For I-129, after determining a best-fit line with Cs-137, the DOE found that the relationship tended to systematically underestimate the I-129 concentrations in higher concentration samples. To avoid underestimating the I-129 in the more concentrated samples and to provide a margin for uncertainty, the DOE multiplied the projected I-129 values by a factor of 1.25. The resulting relationship is given in Equation 1 below (Equation 3-6 in SRR-CWDA-2015-00077, Rev. 2): projected I-129 concentration = 1.25 (3.64 x 10x y.) where y is the measured Cs-137 concentration decayed to October 1, 2032. For Tc-99, the DOE did not observe any systematic variation in the fit of the data to the regression line and, therefore, used the relationship given in Equation 2 below (Equation 2-4 in SRR-CWDA-2015-00123, Rev. 2) without further adjustment: projected Tc-99 concentration = 0.0115 x y. The DOE used Equations 1 and 2 above to project values for I-129 and Tc-99 in liquid phases (i.e., supernate and interstitial liquid) for all samples with measured Cs-137 values. For samples of aqueous tank waste that had measured I-129 or Tc-99 concentrations, the DOE then replaced the projected I-129 values with the measured values. The relationships in Equations 1 and 2 do not apply to I-129 or Tc-99 in solid salt because differences in the solubility of all three radionuclides change their relative concentrations. For I-129, Equation 1 does not apply to I-129 in solid salt because I-129 is more soluble than Cs-137, and the DOE expects there to be very little I-129 in solid salt. For the inventory estimate for the 2020 PA, the DOE assumed that all solid salt has an I-129 concentration of 96 Bq/L (2.6 pCi/mL) because I-129 is highly soluble and the DOE does not expect there to be a Equation 1 Equation 2 significant contribution of I-129 from the dry salt. For Tc-99, Equation 2 does not apply because Tc-99 is less soluble than Cs-137 and Equation 2 would underestimate the Tc-99 concentration in the solid salt waste. For the 2020 PA, the DOE calculated the Tc-99 concentration in salt waste by using the relationship between Tc-99 and Cs-137 in aqueous phases given by Equation 2 and increasing the Tc-99 concentration by a factor of 1.5: projected Tc-99 concentration = 1.5(0.0115 x y.) Finally, the DOE used the relationships developed in this section, with measured I-129 and Tc-99 concentrations when available, to develop inventories of I-129 and Tc-99 in the Tank Farms for use in projecting closure inventories for the SDF. For I-129, the DOE used the following contributions from the Tank Farms to the projected SDF inventory (SRR-CWDA-2015-00077, Rev. 2): 5.55x1011 Bq (15.0 Ci) for the realistic projection (i.e., 5.51x1011 Bq [14.89 Ci] in aqueous phases, 4.07x109 Bq [0.11 Ci] in solid salt); 5.81x1011 Bq (15.7 Ci) for the MPAD projection (i.e., 5.51x1011 Bq [14.89 Ci] in aqueous phases, 4.07x109 Bq 0.11 Ci in solid salt, 2.66x1010 Bq [0.72 Ci] in sludge); and 8.73x1011 Bq (23.6 Ci) for the pessimistic projection (i.e., the MPAD projection increased by 50%). For Tc-99, the DOE accounted for the waste phases slightly differently. For Tc-99, the DOE used the following contributions from the Tank Farms to the projected SDF (SRR-CWDA-2015-00123, Rev. 2): 7.84x1014 Bq (2.12x104 Ci) for the realistic projection (i.e., the inventory in supernate, interstitial liquid, and salt); 1.17x1015 Bq (3.17x104 Ci) for the MPAD projection (i.e., the realistic value multiplied by 1.5); and 1.27x1015 Bq (3.43x104 Ci) for the pessimistic projection (i.e., the realistic value plus half the sludge inventory). The two other sources of I-129 and Tc-99 in the projected SDF inventory are the existing inventory in emplaced saltstone and projected future additions from H-Canyon. For both radionuclides, the DOE updated the emplaced SDF inventory by retrieving data from the SDF-WIDE model that reflected transfers to the SDF through September 2017. The DOE calculated that 3.2x1010 Bq (0.85 Ci) of I-129 and 4.59x1013 Bq (1.24x103 Ci) of Tc-99 had been emplaced in the SDF by that time. Because of the long half-lives of I-129 and Tc-99, those value can be used without adjustment as the contributions of emplaced saltstone to the projected SDF inventories of I-129 and Tc-99. In response to an NRC RAI question, the DOE indicated that the inventories for I-129 and Tc-99 used in the 2020 PA inadvertently omitted future effluent from H-Canyon (SRR-CWDA-2021-00047, Rev. 0). Since the DOE developed the Tank Farms inventory as of the end of 2017, the DOE expected the SDF inventory should include eight years of future transfers from H-Canyon, from FY 2018 through FY 2025. The DOE provided two different estimates of the potential I-129 and Tc-99 transfers from H-Canyon. The first estimate used the projected radionuclide concentrations and transfer volumes provided in Tables 4.2-1 and 4.2-1 of SRR-CWDA-2018-00041, Rev. 3. That DOE document provides an estimate of 1,140 m3/year (300,000 gallons/year) of effluent from H-Canyon to the Tank Farms with an estimated I-129 concentration of 19.4 Bq/mL (5.23 x 102 pCi/mL) and estimated Tc-99 concentration of 4.33x103 Bq/mL (1.17x105 pCi/mL). Using those volume and concentration values, the DOE calculated that the projected SDF inventory should increase by 1.79x1011 Bq (4.85 Ci) I-129 and 4.03x1013 Bq 1.09x103 Ci Tc-99. For the MPAD case, those additions would be a 29% increase in the total I-129 inventory and a 3.3% increase in the total Tc-99 inventory. The DOE based the second estimate on revised projections of the volume of future H-Canyon transfers and the concentrations of I-129 and Tc-99 in those transfers (SRR-CWDA-2021-00047 Rev. 0). To revise the volume estimate, the DOE provided a history of annual transfers from H-Canyon to the tank farm and demonstrated that the average annual volume transferred from 2010 through 2019 was 473 m3 (125,000 gallons) rather than 1,140 m3 (300,000 gallons/year). The DOE stated that the smaller volume would be a reasonable assumption for future transfers from H-Canyon. In addition, the DOE indicated that the concentrations of I-129 and Tc-99 that it assigned to H-Canyon effluents in SRR-CWDA-2018-00041, Rev. 3 also were conservative estimates. Based on measurements from Tank 39, which receives H-Canyon effluents, the DOE indicated that a more realistic I-129 concentration for future H-Canyon effluents would be 1.1 Bq/mL (30 pCi/mL) rather than 19.4 Bq/mL (523 pCi/mL) and a more realistic Tc-99 concentration would be 910 Bq/mL (2.46x104 pCi/mL) rather than 4.33x103 Bq/mL (1.17x105 pCi/mL). Based on those revised volume and concentration projections, the DOE indicated that the projected additions from H-Canyon would increase the projected SDF inventory of I-129 by less than 3% and the increase in the Tc-99 inventory would be negligible. Therefore, the DOE did not change the realistic, MPAD, or pessimistic projections of the I-129 or Tc-99 inventories (Table 1). 3.5 Probabilistic Sensitivity and Uncertainty Analyses The DOE assessed the effect of inventory assumptions on the projected dose to a member of the public and an inadvertent intruder with both probabilistic and deterministic analyses. This section addresses the development of probabilistic distributions to represent uncertainty in inventory and the results of the probabilistic analyses. Section 3.6 addresses the deterministic sensitivity analyses related to inventory in the 2020 PA. In the 2020 PA, the DOE created three truncated, log-normal distributions to represent uncertainty in inventory: one for I-129, one for Tc-99, and one representing the uncertainty in the inventory of all other radionuclides (see Table 4 below). Because the uncertainty in the volumes of each soluble waste phase in each tank in the Tank Farms was much smaller than the uncertainty in the concentration data, the DOE based the uncertainty distributions on variability in the concentration data and assumed the waste volumes did not contribute to the uncertainty in inventory. To apply distributions of concentration data to inventory, the DOE normalized the concentration distributions by dividing by the mean concentration for the appropriate radionuclides. That step created distributions of unitless values. The DOE then modeled uncertainty in inventory by multiplying the realistic inventory values in Table 1 by the appropriate distribution of unitless multipliers from Table 4. Table 4. Parameters defining the truncated log-normal distributions for the unitless multiplier applied to the realistic inventory values in probabilistic analyses in the 2020 PA (from Tables 3, 6, and 12 of SRR-CWDA-2018-00076, Rev. 0). Radionuclide Minimum Mean Standard Deviation Maximum I-129 0.079 1.0 0.26 9.78 Tc-99 0.13 1.0 0.11 13.71 All other radionuclides 0.10 1.19 3.11 14.4 For I-129 and Tc-99, the DOE based the distributions on the measured and calculated concentration values referred to in Section 3.4 of this TRR. The DOE used the percentile function in Microsoft Excel to generate 1st, 5th, 10th, 25th, 50th, 75th, 90th, 95th, and 99th percentiles for the concentration data, which created an empirical probability distribution for each radionuclide. The DOE then tested the fit of uniform, triangular, normal, log-normal, and gamma distributions to the portion of the empirical distribution between the 10th and 90th percentiles of the empirical distribution. Figure 2 shows an example of the fit of different distribution shapes to the empirical distribution for I-129 concentration values. For both I-129 and Tc-99, the DOE determined that a log-normal distribution provided the best-fit to the empirical distribution (SRR-CWDA-2018-00076, Rev. 0). Figure 2. Comparison of fits of different distribution types to the empirical distribution (dotted line) of I-129 concentration data (Figure 2 of SRR-CWDA-2018-00076, Rev. 0). The DOE used the mean and standard deviation from fitting the log-normal distribution to the empirical distributions to define the probabilistic distributions for inventory uncertainty for I-129 and Tc-99 (Table 4). The DOE normalized both the mean and standard distribution by dividing by the mean concentration (i.e., either of I-129 or Tc-99) to create a distribution of unitless multipliers. Similarly, the DOE also calculated unitless minimum and maximum multipliers by taking the ratio of the smallest or largest concentration values for I-129 and Tc-99 and dividing each by the appropriate mean concentration value. The DOE used a similar process to develop the probabilistic distribution for inventories of radionuclides other than I-129 and Tc-99, except that the process was slightly more complex because the DOE first took steps to combine concentration data for different radionuclides. The DOE considered the 18 radionuclides that made the greatest contribution to dose to a member of the public in the FY 2014 and FY 2016 Special Analyses for the SDF. The DOE evaluated concentration data for those 18 radionuclides from samples of salt waste, primarily taken from salt batch qualification reports (SRR-CWDA-2018-00009, Rev. 0). Of the 18 radionuclides, two (I-129 and Tc-99) were considered individually, six had an inadequate number of samples to reliably quantify variability, and three had concentrations that were all below the detection limits. The DOE used the variability in the data from the remaining seven radionuclides (Am-241, C-14, Np-237, Pu-238, Pu-239, Pu-241, and U-234) as the basis for the distribution to represent the uncertainty in the inventory all radionuclides other than I-129 and Tc-99. The DOE first decayed all the concentration data to October 1, 2032. That step allowed the DOE to remove variation in measured concentrations caused by sampling on different dates from the true variability of the waste. The DOE then normalized the concentration data from the seven radionuclides by dividing the concentrations of each radionuclide by the median concentration for that radionuclide (SRR-CWDA-2018-00076, Rev. 0). That step allowed the DOE to compare the variability in the data from different radionuclides that had very different median concentrations. Using the scaled data, the DOE determined that C-14, Np-237, and U-234 showed much less variation than Am-241, Pu-238, Pu-239, and Pu-241 (SRR-CWDA-2018-00076, Rev. 0). To avoid underestimating the variability for all radionuclides, the DOE combined the normalized data for C-14, Np-237, and U-234. The DOE then had five data sets: one each for Am-241, Pu-238, Pu-239 and Pu-241 and one additional data set representing C-14, Np-237, and U-234. The DOE used the percentile function of Microsoft Excel to find the 1st, 5th, 10th, 25th, 50th, 75th, 90th, 95th, and 99th percentiles of each of the five normalized data sets. Then, the DOE used the geometric mean of the normalized data for each percentile value to combine the data into one empirical probability distribution. The DOE set the minimum and maximum multiplier values for radionuclides other than I-129 and Tc-99 (Table 4) to the 10th and 90th percentile values of that combined empirical distribution (Tables 10 and 12 of SRR-CWDA-20018-00076, Rev. 0). To develop a probabilistic distribution, the DOE tested the fit of normal, log-normal, gamma, uniform, and triangular distributions to the 10th, 25th, 50th, 75th, and 90th percentiles of the combined empirical distribution. The DOE determined that the log-normal distribution had the best-fit and used the mean and standard deviation from fitting the log-normal distribution to the empirical distribution to define the probabilistic distribution for inventory uncertainty for radionuclides other than I-129 and Tc-99 (Table 4). Section 5.7 of the 2020 PA addresses the results of probabilistic uncertainty and sensitivity analyses for both a member of the public. Section 6.6 of the 2020 PA focuses on the results of probabilistic analyses for an inadvertent intruder who drills or uses a well 1 m from a disposal structure. Section 5.7.3.3 demonstrates that high inventory multipliers for I-129 appear to have a significant effect on dose in several probabilistic realizations with relatively high projected doses to a member of the public. The other two inventory multipliers (i.e., for Tc-99 or the remaining radionuclides) did not appear to significantly affect those realizations. Similarly, Section 5.7.4.2.4 of the 2020 PA indicates that the projected dose to a member of the public was sensitive to the inventory multiplier for I-129 in SDS 9. The projected dose to a member of the public was not sensitive to the inventory multiplier for Tc-99 or the other radionuclides and the projected dose to an inadvertent intruder was not sensitive to any of the inventory multipliers. 3.6 Deterministic Sensitivity Analyses In addition to the probabilistic analyses discussed in the previous section, the DOE used the following deterministic sensitivity analyses to evaluate the effects of alternative inventory assumptions on the projected dose to a member of the public: two Central Scenario cases that evaluated the effects of using the realistic and pessimistic inventories shown in Table 1 for all radionuclides; one case assuming the same total I-129 inventory with an alternative distribution among disposal structures; two Extreme Inventory Values cases that used the minimum and maximum multiplier values in Table 5 multiplied by the realistic inventory in Table 1 to create minimum and maximum inventories for all radionuclides except for Tc-99; and four cases that considered the realistic, pessimistic, minimum, and maximum values for the Tc-99 inventory alone. For each of sensitivity cases listed above, the values of all other (non-inventory related) parameters were set to their MPAD values. Section 5.8.5.1 of the 2020 PA provides the results of the DOE sensitivity cases that used the realistic and pessimistic inventories. Figure 3 below (Figure 5.8-55 of the 2020 PA) shows the results of those two sensitivity cases with three other cases for context. The two dotted lines show the dose projections for the cases that used the realistic and pessimistic inventory values from Table 1 with MPAD values of all other model parameters (e.g., parameters related to infiltration, wasteform performance, far-field flow, and biosphere). The solid black line represents the compliance case, which uses the MPAD inventory for all radionuclides and MPAD values for all other model parameters. The two additional solid lines represent the overall realistic and pessimistic Central Scenario cases, in which the DOE used realistic and pessimistic values for uncertain parameters throughout the model. As shown in Figure 3, the variation in inventory accounts for a small fraction of the variation in the dose projections for the overall realistic and pessimistic cases. For the cases that only varied the inventory projections, there is little variation between the projected dose in the MPAD and realistic cases, and approximately a 50% increase in dose in the case that used the pessimistic inventory as compared to the MPAD inventory. That pattern is similar to the difference between the realistic, MPAD, and pessimistic inventories of I-129, even though there is more difference between the realistic and MPAD inventories of other radionuclides, because the I-129 inventory has a greater effect on the projected dose than the inventories of other radionuclides. Figure 3. Comparison of the projected dose to a member of the public in sensitivity cases that vary inventory alone (dotted lines) versus cases that vary all uncertain parameters between the realistic, compliance (MPAD), and pessimistic cases (Figure 5.8-55 of the 2020 PA). Section 5.8.5.2 of the 2020 PA provides the results of an analysis that used the MPAD total I-129 inventory (i.e., 6.14x1011 Bq [16.6 Ci]) with alternative assumptions about the spatial distribution of that inventory in individual disposal structures. The DOE based the alternative assumptions on a preliminary system planning estimate. In general, the alternative assumptions shift inventory away from SDS 9-12 and into SDS 6-8 (see Table 5 below; Table 5.8-17 in the 2020 PA). That case resulted in an approximate 5% decrease in the projected dose within 6,000 years of site closure and an approximately 20% increase in the projected dose within 10,000 years of site closure (see Figure 5.8-56 in the 2020 PA). Section 5.8.5.3 of the 2020 PA provides the results from cases that used values the DOE referred to as extreme inventory values. For that case, the DOE used the realistic inventory values in Table 1 multiplied by the minimum and maximum values in Table 4 to generate upper and lower bounding values for the inventory of all radionuclides except Tc-99. Tc-99 was tested separately, as described below. As shown in Figure 4 (Figure 5.8-57 of the 2020 PA), varying the radionuclide inventories to the maximum bounding values in Table 4 above causes approximately as much of an effect on the projected dose to a member of the public as changing parameters to their pessimistic case values throughout the model. In contrast, using the minimum inventory multipliers in Table 4 has a significantly smaller effect than changing parameters to their realistic values through the model (note the log scale of the y-axis in Figure 4). The smaller difference between the dose projections for the realistic and MPAD inventory cases as compared to the difference between the MPAD and pessimistic inventory cases is consistent with the differences between the realistic, MPAD, and pessimistic inventory projections for I-129 (Table 1). Although other radionuclides have a greater difference between the realistic and MPAD cases than I-129 does, the I-129 inventory has a greater effect on dose projections, as addressed in Section 3.5 of this TRR. Table 5. Alternative arrangement of I-129 inventory in the SDF (Table 5.8-17 in the 2020 PA). Disposal Structure MPAD Case I-129 Inventory (Ci) (a) Alternative Case I-129 Inventory (Ci) (a) 1 0.20 0.20 2 A 0.07 0.07 2B 0.07 0.07 3 A 0.19 0.19 3B 0.19 0.19 4 0.28 0.28 5 A 0.14 0.14 5B 0.09 0.09 6 2.20 2.81 7 2.20 3.28 8 2.20 3.11 9 2.20 1.19 10 2.20 1.63 11 2.20 1.24 12 2.20 1.39 Total 16.6 16.6 (a) To convert Ci to Bq, multiply by 3.7x1010. Figure 4. Comparison of the projected dose to a member of the public in sensitivity cases that vary inventory alone (dotted lines) to bounding values versus cases that vary all uncertain parameters between the realistic, compliance (MPAD), and pessimistic cases (Figure 5.8-57 of the 2020 PA). Section 5.8.5.4 of the 2020 PA provides the results from cases that varied the inventory of Tc-99 alone. This group included four sensitivity cases: two cases that used the realistic and pessimistic values for Tc-99 in Table 4 and two cases that used the realistic inventory values for Tc-99 in Table 1 multiplied by the minimum and maximum multipliers for Tc-99 in Table 4. As shown in Figure 5 (Figure 5.8-59 of the 2020 PA), the change in the inventory values for Tc-99 caused a much smaller change in the projected dose. For example, the maximum multiplier for Tc-99 in Table 4 is 13.71, but the peak dose within 10,000 years in the maximum bounding case that used that inventory multiplier increases by approximately 15%. The 2020 PA indicates that the sub-linear relationship between an increase in modeled inventory and the projected peak dose occurs because Tc-99 is solubility limited when it is chemically reduced, so that increasing Tc-99 inventory does not increase its release rate from chemically reduced parts of saltstone. Figure 5. Comparison of the projected dose to a member of the public in sensitivity cases that vary Tc-99 alone while using MPAD values for all other parameters (Figure 5.8-59 of the 2020 PA). 3.7 The DOE Methods for Tracking SDF Inventory Since 2015, the DOE has tracked the emplaced inventory in the SDF with the SDF-WIDE model (SRR-CWDA-2015-00003, Rev. 0). The SDF-WIDE model calculates the inventory inputs to the SDF based on measured or calculated radionuclide concentrations in each batch of salt waste and the volume of waste in the batch. When possible, the model uses measured radionuclide concentrations from the feed tank for the Saltstone Production Facility (i.e., currently Tank 50). For radionuclide concentrations that are below detection limits, the model uses alternative calculations, including using detection limits, calculating ingrowth from measured ancestor radionuclides, using solubility information to estimate supernate concentrations, applying partitioning between sludge and supernate observed in some waste tanks to other tanks, or using correlations to more easily detectable radionuclides. The SDF-WIDE model also calculates radionuclide ingrowth and decay. 4.0 NRC Staff Evaluation 4.1 Overview The NRC staff conducted a risk-informed review that focused on the development of the inventories of I-129 and Tc-99 because those two radionuclides dominate the projected dose to both a member of the public and an inadvertent intruder in the 2020 PA. Sensitivity analyses and uncertainty analyses in the 2020 PA demonstrated that the peak projected dose to a member of the public is more sensitive to the inventory of I-129 than it is to any other radionuclide, including Tc-99 (see Sections 3.5 and 3.6 of this TRR). That result is consistent with NRC staff expectations because the Tc-99 release rate is limited by solubility when saltstone is chemically reducing. However, the NRC staff focused on Tc-99 as well as I-129 because the inventory of Tc-99 affects the duration of the peak release from saltstone. In addition to reviewing the development of the inventory of I-129 and Tc-99 in detail, the NRC staff reviewed the following DOE processes: (1) screening radionuclides for detailed analysis, (2) developing the inventory for all radionuclides, (3) developing uncertainty distributions for all radionuclides, (4) developing sensitivity and uncertainty analyses, and (5) tracking the SDF inventory. The NRC staff conducted this review to assess the radionuclide inventories the DOE used in modeling to support the 2020 PA. In addition, the NRC staff will use insights from this technical review to develop recommended changes to the Monitoring Plan. In this section, the NRC staff has included recommended additions to the Monitoring Plan in italics. The NRC staff summarized those additions in the Conclusions section of this TRR (i.e., Section 7). As described in the current Monitoring Plan, the NRC staff routinely monitors several documents related to SDF inventory, including the SDF PA Annual Review, Saltstone Permit Website Reporting Data (http://sro.srs.gov.saltstone.htm), and the Quarterly Tank 50 Waste Acceptance Criteria Sample Analysis (see the Monitoring Plan, Table 1-1). Therefore, the NRC staff recommended additional specific monitoring activities in this section with the expectation that they would supplement the ongoing routine monitoring activities the NRC staff performs under MF 1.01 and MF 1.02 of the Monitoring Plan. The Monitoring Plan includes a table that provides the DOE expected inventory of the SDF at the time of closure. The DOE updated that inventory in the 2020 SDF PA. Therefore, INVT-01 The NRC staff recommends updating Table A-1 of the Monitoring Plan to reflect the MPAD inventory in the 2020 PA under MF 1.01. 4.2 Radionuclide Screening The NRC staff determined that the DOE process for radionuclide screening is acceptable because it began with a comprehensive list of radionuclides that was consistent with process knowledge and it used technically defensible reasons for screening out radionuclides. The DOE considered the potential dose from longer-lived progeny of radionuclides with half-lives less than five years before screening out the short-lived parents. The DOE also accounted for short-lived radionuclides that it assumed to be in secular equilibrium with a parent. The NRC staff agreed with the DOE that it is appropriate to screen radionuclides based on conservative stylized dose calculations. Although the DOE used the whole dose limit (i.e., either 0.25 mSv/year (25 mrem/year) for a member of the public or 5 mSv/year (500 mrem/year) for an inadvertent intruder) instead of using a fraction of the dose limit for screening, the NRC staff found the radionuclide concentration assumptions that the DOE used in its screening dose calculations were sufficiently conservative that the calculations were appropriate for screening. Although Pd-107 did not meet the formal screening criteria, the NRC agreed with the DOE conclusion that omitting it did not impact the 2020 PA results. The NRC staff agreed with the DOE technical basis for determining that it would have a projected dose at least four orders of magnitude less than that of I-129 for both a member of the public and an inadvertent intruder. However, the NRC staff expects it will be included in future DOE analyses for completeness unless it meets the screening criteria in future analyses. The DOE screened out Ag-108m, Bi-210m, and Cm-246 because there was no inventory data available for those radionuclides. All three radionuclides have half-lives greater than five years. The NRC staff determined that lacking inventory data is not a sufficient technical basis to screen out the radionuclides because it does not limit the potential dose from those radionuclides. However, the DOE did not include any of those three radionuclides in previous analyses of the SDF that the NRC staff reviewed. Furthermore, the DOE did not indicate that new process information indicated that those radionuclides should be evaluated. Instead, the DOE indicated that it added the radionuclides for consistency with the SDF-WIDE model (SRR-CWDA-2018-00044, Rev. 3). Therefore, the NRC staff does not expect that Ag-108m, Bi-201m, or Cm-246 will make a significant contribution to the projected dose to an offsite member of the public or an individual who inadvertently intrudes on the SDF. 4.3 General Process for Inventory Development The NRC staff determined that the DOE process for inventory development was acceptable because it accounted for the major projected contributions to the SDF (i.e., the existing SDF inventory, the existing Tank Farms inventory, and future transfers from H-Canyon) and used technically defensible methods to determine the contributions from those sources. The NRC staff calculated that the existing SDF inventory makes a relatively small contribution to the projected SDF inventory for most, but not all, radionuclides (Table 6). Two exceptions for which the existing SDF inventory contributes more than 15% of the projected SDF closure inventory are cesium-135 (Cs-135) and Cs-137. The NRC staff expects those isotopes are exceptions because the ARP/MCU and SWPF remove Cs from salt waste, in contrast with the DDA process, which did not. Therefore, it is reasonable that Cs isotopes would have a significant fractional contribution from the waste already emplaced in the SDF. Other exceptions include isotopes of actinium (Ac), californium (Cf), curium (Cm), lead (Pb), niobium (Nb), plutonium (Pu), protactinium (Pa), radium (Ra), thorium (Th), and uranium (U). That observation is consistent with the NRC staff expectation because those elements are associated with sludge, and, as discussed in Section 3.2, the transfer of sludge entrained in solid salt waste to the SDF diminished significantly after completion of the DDA process. The DOE process to calculate the existing SDF inventory of radionuclides other than I-129 and Tc-99 is acceptable for two reasons: (1) whenever possible, the DOE based the inventory calculation on direct measurements of radionuclide concentration in the tank that receives treated salt waste prior to incorporation into saltstone (i.e., Tank 50) and (2) for radionuclides that have concentrations below detection limits in Tank 50, the DOE used technically justified alternative methods to estimate the concentrations. As noted in Section 3.3, those alternative methods included using detection limits, calculating ingrowth from measured ancestor radionuclides, using solubility information to estimate supernate concentrations, applying partitioning between sludge and supernate observed in some waste tanks to other tanks, or using correlations to more easily detectable radionuclides. In general, the NRC staff found the results of special methods agreed with available detected concentrations or detection limits for each radionuclide in Tank 50 (see SRR-CWDA-2018-00041, Rev. 3, Appendix B). The NRC staff also found that the DOE special methods used in the development of the existing SDF inventory appropriately accounted for the different characteristics of the wastes generated by the DDA and ARM/MCU processes. Table 6. Comparison of the DOE inventory projections at closure (i.e. January 1, 2037) compared to the inventory in the SDF on March 31, 2018 decayed to January 1, 2037. Radionuclide 2020 PA MPAD Inventory (Ci) (a, b) SDF Inventory as of 3/31/2018 Decayed to 1/1/2037 Projected Contribution from H-Canyon (Ci) (a, c) Percent of MPAD Inventory(d) (Ci) (a, e) Percent of MPAD Inventory(d) Ac-227 3.41x10-4 1.78x10-4 52 (f) (f) Al-26 3.65x101 1.24 3.4 (f) (f) Am-241 2.06x104 2.28x101 0.11 3.34x101 0.16 Am-242 m 9.98 3.39x10-2 0.34 (f) (f) Am-243 8.84 5.41x10-1 6.1 5.57 63 C-14 7.86x102 2.02x101 2.6 1.31 0.17 Cf-249 3.38x10-1 3.38x10-1 100 (f) (f) Cf-251 1.48x10-1 1.48x10-1 100 (f) (f) Cl-36 1.31 1.92x10-2 1.5 (f) (f) Cm-243 4.51x10-2 7.10x10-3 16 (f) (f) Cm-244 1.63x102 1.59x101 10 1.75x101 11 Cm-245 9.56x10-1 8.45x10-1 88 (f) (f) Cm-247 1.77x10-1 1.77x10-1 100 (f) (f) Co-60 1.11x101 7.97x10-3 0.072 3.09x10-2 0.28 Cs-135 3.67 1.92 52 2.29x10-3(g) 6.2x10-4 Cs-137 4.47x105 1.26x105 28 4.91(g) 1.1x10-5 Eu-152 1.16x101 2.87x10-2 0.25 (f) (f) Eu-154 1.07x102 6.78x10-1 0.63 2.09 1.9 H-3 4.50x103 2.91x101 0.49 9.46 0.21 I-129 1.66x101 8.80x10-1 5.3 1.19x10-1(h) 0.72 K-40 1.32 2.02x10-2 1.5 (f) (f) Nb-93 m 2.06x103 8.84x102 19 (f) (f) Nb-94 2.97x10-1 9.98x10-2 34 1.34x10-1 45 Ni-63 2.91x102 3.01 1.0 1.95 0.67 Np-237 1.77x101 1.01 5.7 2.50 14 Pa-231 5.81x10-4 1.65x10-4 28 (f) (f) Pb-210 7.08x10-5 3.70x10-5 53 (f) (f) Pt-193 1.12x102 1.34x101 12 (f) (f) Radionuclide 2020 PA MPAD Inventory (Ci) (a, b) SDF Inventory as of 3/31/2018 Decayed to 1/1/2037 Projected Contribution from H-Canyon (Ci) (a, c) Percent of MPAD Inventory(d) (Ci) (a, e) Percent of MPAD Inventory(d) Pu-238 2.10x105 3.87x102 0.18 5.68x101 0.027 Pu-239 1.30x104 6.29x101 0.48 7.48x101 0.57 Pu-240 2.78x103 7.70x101 2.8 7.50x101 2.7 Pu-241 2.47x104 5.50x101 0.22 2.74x102 1.1 Pu-242 1.03x101 5.96 58 7.46x10-2 0.72 Pu-244 4.55x10-2 2.53x10-2 56 (f) (f) Ra-226 1.94x10-4 1.02x10-4 53 2.39x10-5 12 Ra-228 3.77x10-1 2.87x10-4 0.076 (f) (f) Se-79 1.42x102 1.08x101 7.6 4.83x10-1 0.34 Sm-151 5.66x103 1.77x101 0.31 1.54x102 2.7 Sn-126 5.18x102 7.93 1.5 4.66x101 9.0 Sr-90 5.30x106 1.94x103 0.037 8.53x102 0.016 Tc-99 3.29x104 1.27x103 3.9 9.78x101(h) 0.30 Th-229 3.73 3.66 98 (f) (f) Th-230 1.50x10-2 8.49x10-3 57 3.54x10-6 0.025 Th-232 3.77x10-1 2.87x10-4 0.076 3.54x10-6 0.0010 U-232 2.02x10-1 1.50x10-1 74 (f) (f) U-233 3.04x101 1.36x101 45 1.05x10-1 0.34 U-234 4.01x101 1.21x101 30 4.18 10 U-235 7.84x10-1 1.01x10-1 13 8.78x10-2 11 U-236 1.65 1.21x10-1 7.4 5.68x10-1 35 U-238 2.24x101 2.04x10-1 0.91 2.12x10-3 0.0095 Zr-93 1.39x102 1.02x101 7.3 (f) (f) (a) To convert Ci to Bq, multiply by 3.7x1010 (b) Inventory projections from the 2020 PA, Table 3.3-6 (c) The NRC staff calculated the inventories as of the closure date (January 1, 2037) based on the sum of the individual disposal structure inventories in SRR-CWDA-2018-00041, Rev.3, Table 3.5-1. Staff decayed the totals for each radionuclide with the GoldSim modeling platform using the decay constants from the ICRP Report No. 107. (d) Precents calculated by the NRC staff based on the inventory values in this table. (e) Unless otherwise noted, the NRC staff calculated the projected contribution using the radionuclide concentrations and projected transfer volumes provided in Tables 4.2-1 and 4.2-2 of SRR-CWDA-2018-00041, Rev. 3, with the equation in Section 4.2.1 of that document. (f) Concentrations to calculate inventories were not included in SRR-CWDA-2018-00041, Rev.3, Table 4.2-1. (g) Staff applied a decontamination factor of 200 to Cs isotopes, per the DOE description in SRR-CWDA-2018-00041, Rev. 3, Section 4.2.1. (h) The DOE calculated alternative H-Canyon contributions for I-129 and Tc-99 (Section 3.4). In the Monitoring Plan, the NRC staff indicated that it was acceptable for the DOE to base Ra-226 and Th-230 concentrations on ingrowth from U-234 if the Tank Farms did not receive significant inputs of thorium-bearing waste. The NRC staff found the assumption to be acceptable for the emplaced saltstone because Th-230 measurements (i.e., both detectable concentrations of Th-230 and upper bounds determined by detection limits) were consistent with the values calculated from ingrowth from U-234 (SRR-CWDA-2018-00041, Rev. 3, Appendix B). Furthermore, because the DOE has been able to detect Th-230 in some samples at concentrations similar to the detection limit, the NRC staff expects that waste with significantly greater Th-230 concentrations would have measurable Th-230 concentrations. The NRC staff therefore expects that it would observe any increases in Th-230 contributions to the SDF as part of its routine monitoring processes. The NRC staff will continue to monitor the addition of all radionuclides to the SDF by reviewing the DOE salt waste acceptance criteria sample analysis reports, which the DOE issues approximately quarterly (see Table 1-1 in the Monitoring Plan). INVT-02 The NRC staff recommends monitoring additions to the SDF of radionuclides that are at or near the total projected SDF inventory, including thorium-229 (Th-229), curium-245 (Cm-245), Cm-247, californium-249 (Cf-249), and Cf-251 under MF 1.01. The NRC staff recommends assessing the impact on dose to a member of the public and an inadvertent intruder if the SDF inventories exceed the MPAD inventory in the 2020 PA. The NRC staff found the development of the projected contribution of the existing tank farm inventory to the future SDF inventory to be acceptable for two reasons: (1) whenever possible, the DOE based the inventory calculation on direct concentration measurements of samples from supernate and solid salt waste and volumes of those phases in waste tanks and (2) for radionuclides that have concentrations below detection limits in Tank 50, the DOE used technically justified alternative methods to calculate the concentrations. As noted in Section 3.3, those methods were essentially the same as the alternative methods the DOE used in the SDF-WIDE model to track the disposed inventory in the SDF. Because the inventory in the Tank Farms represents future disposals to the SDF, the NRC staff will assess the agreement between the inventory additions to the SDF from the Tank Farms and the DOE projections by monitoring periodic reports the DOE produces related to SDF inventory (see Table 1-1 of the Monitoring Plan). The NRC staff will continue to monitor radionuclide concentrations in Tank 50 (see, e.g., SRR-STI-2021-00165, Rev. 0) to ensure that the concentrations are consistent with or less than the concentrations the DOE assumed in the calculation of tank farm contributions to the SDF. If radionuclide concentrations in Tank 50 are significantly greater than the concentrations the DOE assumed in the development of the projected SDF inventory, the NRC staff will assess whether the differences affect the conclusions of the 2020 PA. To calculate the projected contribution to the SDF of radionuclides other than I-129 and Tc-99 from future H-Canyon transfers to the Tank Farms, the DOE used the same radionuclide concentration estimates that the DOE used to represent H-Canyon waste in the 2009 PA, multiplied by the projected volumes of transfers from FY 2018 to FY 2026 in the DOE Liquid Waste System Plan (SRR-LWP-2009-00001, Rev. 20). Based on the DOE projected radionuclide concentrations and transfer volumes provided in SRR-CWDA-2018-00044, Rev. 3, the NRC staff calculated that transfers from H-Canyon accounted for a small percentage of the DOE projected SDF MPAD inventory for most radionuclides (Table 6). The two exceptions are Am-243 and Nb-94, for which the projected contributions from H-Canyon account for 63% and 45% of the MPAD SDF inventory, respectively. As discussed in Section 3.3, the most recent revision of the DOE Liquid Waste System Plan (SRR-LWP-2009-00001, Rev. 21) projects 29% more waste being transferred from H-Canyon to the Tank Farms than the DOE anticipated when developing the inventory for the 2020 PA. Despite that increase, the NRC staff determined the volume estimate the DOE used to develop the inventory for all radionuclides other than I-129 and Tc-99 to be acceptable because the annual transfer volumes that the DOE used to calculate the inventory contribution from future H-Canyon transfers overestimated the actual average transfer volume from FY 2010 to FY 2019 by a factor of 2.4. INVT-03 The NRC staff recommends monitoring significant changes in waste characteristics and annual volumes of waste sent from H-Canyon to the Tank Farms to assess the potential effects on SDF inventory under MF 1.02. 4.4 Additional Calculations for Inventory Development for I-129 and Tc-99 The NRC staff determined the DOE overall approach to developing the projected inventory for I-129 and Tc-99 is acceptable because it accounted for the major projected sources of I-129 and Tc-99 (i.e., the existing SDF inventory, the existing Tank Farms inventory, and future transfers from H-Canyon) and used technically defensible methods to determine the contributions from those sources. The DOE process to calculate the existing SDF inventory of I-129 and Tc-99 is acceptable for two reasons: (1) it was based on measurements from the Saltstone Production Facility feed tank when possible and (2) the DOE expects the existing inventory of both I-129 and Tc-99 is 5% or less of the projected inventory of those radionuclides. As noted in Section 2.4 of this TRR, in the NRC Letter of Acknowledgment (ADAMS Accession No. ML12213A447), the NRC staff determined that the lower Tc-99 inventories in SDS 2B, 3 A, 3B, 4, 5 A, and 5B that the DOE projected in response to the NRC Letter of Concern alleviated the NRC staff concern regarding Tc-99 releases from those disposal structures. After that NRC evaluation, the DOE increased the authorized Tc-99 inventory in SDS 2B, 3 A, 3B, 4, 5 A, and 5B based on the results of the FY 2014 Special Analysis. As indicated in Section 2.1 of this TRR, the NRC staff did not write a TER or make conclusions based on the FY 2014 Special Analysis. The NRC staff will evaluate the effects of the emplaced Tc-99 inventories in SDS 2B, 3 A, 3B, 4, 5 A, and 5B in the context of the revised models the DOE used to support the 2020 PA. The DOE process for developing the projected contributions of I-129 and Tc-99 from the current Tank Farms inventory demonstrated a risk-informed focus on the largest projected source of I-129 and Tc-99 and a focus on the radionuclides that cause the majority of the projected dose for a member of the public and an inadvertent intruder. The data analysis the DOE used to improve the I-129 and Tc-99 inventory estimates is acceptable for four reasons: (1) the sources of data the DOE used to develop the relationship between Cs-137 and I-129 or Tc-99 were appropriate and traceable, (2) the process the DOE used to develop a relationship between measured Cs-137 and I-129 or Tc-99 was clear and appropriately accounted for radioactive decay of Cs-137, (3) the DOE provided a clear and justified technical basis for removing any data points that were not included in the derived mathematical relationships, and (4) the application of the relationship between Cs-137 concentrations and I-129 or Tc-99 was clear and technically justified. The DOE proposed realistic, MPAD, and pessimistic inventory estimates for I-129 and Tc-99 are acceptable because they are based on the best available data and they adequately represent the uncertainty in the existing Tank Farms inventory of I-129 and Tc-99. Although the inventory recommendations only included uncertainty in the current tank farm inventories of I-129 and Tc-99 (i.e., the DOE treated the current SDF inventory and H-Canyon inventory as certain), the NRC staff determined the recommendations were acceptable for two reasons: (1) the NRC staff calculated that 95% or more of the DOE MPAD projected SDF inventory of both radionuclides is in the current Tank Farms inventory (see Table 6 above); and (2) for both radionuclides, the proposed variation in the realistic, MPAD, and pessimistic inventory estimates is larger than the entire current inventory of the SDF and the revised projected inventory from projected transfers from the H-Canyon. For I-129 and Tc-99, the NRC staff found the calculation of the inventory contribution from projected additions of waste to the tank farm from H-Canyon to be acceptable because the concentration estimates were based on recent measured values from the tank that receives H-Canyon waste (i.e., Tank 39) and the volume estimates are equal to the average transfer volume from FY 2010 to FY 2019 (SRR-CWDA-2021-00047, Rev. 1). However, the NRC staff also determined that the revised estimates of the H-Canyon transfer volumes and radionuclide concentrations provided in the DOE document SRR-CWDA-2021-00047, Rev. 0 appeared to remove the allowance for uncertainty that was included in the original volume and concentration measurements. The DOE also could require some of that margin to account for the additional years of H-Canyon inputs to the tank farm and 29% increased volume of H-Canyon contributions discussed in Section 3.3. Therefore, INVT-04 The NRC staff recommends monitoring significant changes in waste characteristics and annual volumes of waste sent from H-Canyon to the Tank Farms to assess the potential effects on I-129 and Tc-99 inventory in the SDF under MF 1.01. 4.5 Probabilistic Distributions The probabilistic distributions the DOE developed for I-129 and Tc-99 inventories are acceptable for two reasons: (1) the data the DOE used to develop the distributions was traceable and represented the tank farm waste that the DOE plans to send to the SDF and (2) the method the DOE used to develop the distributions was clear and technically justified. The NRC staff determined that the concentration data the DOE based the distributions on were representative of the soluble phases of tank farm waste because the data included direct measurements of I-129 and Tc-99 concentrations in supernate from approximately half of the waste tanks (see Table 3-1 of SRR-CWDA-2015-00077, Rev. 2) supplemented with concentrations calculated from Cs-137 measurements using Equations 1 and 2 (above) from all of the operating waste tanks (see Table 3-2 in SRR-CWDA-2015-00077, Rev. 2). The NRC staff found that the lower number of concentration measurements from samples of dry salt were acceptable because both I-129 and Tc-99 are highly soluble and the NRC staff expects the dry salt phase to make a very small contribution to the SDF inventory. Similarly, the NRC staff determined that the DOE choice to base the distributions on the variability of I-129 and Tc-99 concentrations in soluble tank farm waste was acceptable because the existing tank farm inventory is the largest input to the projected SDF inventory (i.e., 95% of the projected I-129 inventory and 96% of the projected Tc-99 inventory) and (2) the NRC staff expects the SDF disposal structure inventory to be less variable than the tank farm inventory because the DOE intentionally combines lower and higher activity waste in salt waste batches to moderate radionuclide concentrations in saltstone. The DOE process for developing probabilistic distributions for the inventory for radionuclides other than I-129 or Tc-99 is acceptable for four reasons: (1) the DOE based the distributions on the variability of measured radionuclide concentrations in salt batch verification reports, which the NRC staff expects to represent waste that will be emplaced in saltstone, (2) the DOE used a technically justified process to select the radionuclides that it used to develop the distribution, (3) the NRC staff expects the SDF disposal structure inventory to be less variable than the tank farm inventory because the DOE combines lower and higher activity waste in salt waste batches to minimize variability, and (4) the lower dose contributions of radionuclides other than I-129 and Tc-99 do not justify a more detailed approach. 4.6 Deterministic Sensitivity Analyses The scope of the DOE deterministic sensitivity analyses for inventory is acceptable because the cases are well-suited to demonstrate key features of the 2020 PA, including: (1) the relative importance of realistic, MPAD, and pessimistic assumptions about inventory as compared to realistic, MPAD, and pessimistic assumptions made throughout the model, (2) the effects of bounding assumptions about inventory, and (3) the relative importance of the inventory of I-129, Tc-99, and all other radionuclides combined. The deterministic analyses are acceptable because the bases for the alternative inventory assumptions is transparent and the range of inventory assumptions the DOE used covers plausible inventory values. The NRC staff agrees the values the DOE refers to as extreme inventory values are likely to bound the SDF inventory projection because the values represent the extremes of large sets of concentration measurements (and calculated values) for tank farm waste, and it is unlikely that the average concentration of any radionuclide in tank farm waste would be more extreme than the variation seen in individual values. In the deterministic Central Scenario cases, the DOE considered the uncertainty in the inventory in emplaced saltstone to be negligible compared to the uncertainty in projected additions to the SDF from the Tank Farms. That approach differs from the approach the DOE took in the probabilistic analyses described in Section 3.5, in which the same probabilistic distribution of unitless multiplier values was applied to the realistic inventory values for all radionuclides other than I-129 and Tc-99, irrespective of how much of the inventory had already been emplaced. In the deterministic Central Scenario cases, the projected inventories of radionuclides that are attributable mainly to the existing SDF inventory have little or no variation between the realistic, MPAD, and pessimistic estimates. For example, by decaying the current SDF inventory values that the DOE provided in the 2020 PA to the expected closure date of January 1, 2037, the NRC staff calculated that the DOE has already emplaced 100% of the MPAD closure inventory of Cf-251 in the SDF (see Table 6 of this TRR) and Table 1 of this TRR shows identical values for the realistic, MPAD, and pessimistic projections for Cf-251. Similarly, Table 1 shows only slight variation in the realistic, MPAD, and pessimistic projections for U-232, for which Table 6 shows that 74% of the projected inventory has been emplaced in the SDF. In contrast, Table 1 shows a factor of 1.6 difference between the pessimistic and realistic projections for Cl-36, for which the DOE estimates less than 3% of the projected inventory has been emplaced in the SDF. Those results are consistent with the DOE assumption that the emplaced saltstone inventory does not contribute uncertainty to the projected closure inventory. The NRC staff determined that assumption is acceptable in the Central Scenario cases for two reasons: (1) radionuclide concentrations and waste volumes in salt waste in the feed to the Saltstone Production Facility is significantly less uncertain than radionuclide concentrations and waste volumes in each waste phase in each tank in the Tank Farms; and (2) the DOE conducted additional sensitivity analyses to demonstrate the effects of using bounding inventory values for all radionuclides. Similarly, the DOE assumed that future contributions to the SDF inventory from H-Canyon transfers of waste to the Tank Farms would not contribute uncertainty to the inventory in the Central Scenario cases. In contrast with the inventory emplaced in the SDF, future inputs from H-canyon appear to be uncertain. For example, as noted in Section 3.3 of this TRR, the most recent Liquid Waste System Plan (SRR-LWP-2009-00001, Rev. 21) increases the projected transfer of H-Canyon waste to the Tank Farms by 25% compared to the volume the DOE assumed when developing the inventory for the 2020 PA. In addition, the radionuclide concentrations in the waste from H-Canyon may depend on process inputs that may not all be known at this time. Despite this uncertainty, the NRC staff determined that the DOE assumption that future waste transfers from H-Canyon will not add uncertainty to the Central Scenario cases to be acceptable for two reasons: (1) the DOE projected that the contribution of inventory from H-Canyon to the SDF will be small (see Table 6 of this TRR) and the DOE does not expect that the two exceptions (i.e., Am-243 and Nb-94) will cause a significant dose to a member of the public or an inadvertent intruder; and (2) the DOE conducted additional sensitivity analyses to demonstrate the effects of using bounding inventory values for all radionuclides. As noted in Section 4.3, the NRC staff recommends monitoring significant changes in waste characteristics and annual volumes of waste sent from H-Canyon to the Tank Farms to assess the potential effects on SDF inventory. As noted in Section 3.6 of this TRR, the alternative spatial arrangement for I-129 that the DOE modeled based on a preliminary system planning estimate resulted in an increase of approximately 20% in the projected dose to a member of the public within 10,000 years of site closure. Therefore, INVT-05 The NRC staff recommends monitoring the inventory of I-129 in individual disposal structures in comparison to the projected inventories in Table 5 of this TRR under MF 1.01. The NRC staff recommends assessing the impact on the dose to a member of the public and an inadvertent intruder if one or more individual disposal structures exceed the MPAD inventory in the 2020 PA. 4.7 The DOE Process for Tracking SDF Inventory The NRC staff monitors the DOE processes for tracking the SDF inventory under MF 1.02, Methods Used to Assess Inventory. The DOE currently uses the SDF-WIDE model (SRR-CWDA-2015-00003, Rev. 0) to track the SDF inventory. That model tracks inventory disposed in the SDF based on radionuclide concentrations and volumes of treated salt waste in the feed tank for the Saltstone Production Facility (i.e., currently Tank 50). The NRC staff finds use of the SDF-WIDE model to be acceptable for the purpose of tracking SDF inventory for three reasons: (1) the model accounts for decay and ingrowth, which limits the potential for calculation and transcription errors, (2) the model facilitates tracking of data references, which improves data traceability, and (3) the DOE routinely updates the model with data from samples from Tank 50. As documented in the June 2008 Onsite Observation Visit Report (ADAMS Accession No. ML081290367), the NRC staff previously found the DOE methods for sampling and analyzing waste from Tank 50 to be acceptable (ML081290367). The NRC staff will continue to evaluate the sampling and analysis methods to determine if there are changes due to the recent start of operations at the SWPF. Staff will evaluate those methods both during Onsite Observation Visits and by monitoring the description of sampling and analysis methods in the DOE quarterly Tank 50 Salt Solution Sample reports (e.g., SRR-STI-2021-00165, Rev. 0). The NRC staff also will continue to use the quarterly Tank 50 Salt Solution Sample reports to track the inventory added to the SDF in comparison to the values in Table 1. The NRC staff will evaluate whether values that exceed the MPAD values the DOE used in the 2020 PA significantly affect the conclusions of the PA. 5.0 Teleconference or Meeting There were no teleconferences or meetings with the DOE related to this TRR. 6.0 Follow-up Actions There are no specific Follow-up Actions related to this TRR. The NRC staff will continue to monitor the radionuclide inventory in emplaced saltstone under MF 1.01 and the DOE methods to determine inventory under MF 1.02 until the DOE completes saltstone emplacement. The NRC staff recommends updating the Monitoring Plan as described in the Conclusions. 7.0 Conclusions The NRC staff concluded that the radionuclide inventories in the 2020 PA are acceptable for modeling the projected dose from the SDF for the purpose of the DOE demonstrating compliance with the 10 CFR 61.41 performance objective, Protection of the General Population from Releases of Radioactivity. In addition, the NRC staff determined that the resulting radionuclide concentrations in saltstone are acceptable as one input to the inventory for modeling the dose to an individual who intrudes on the SDF for the purpose of the DOE demonstrating compliance with the 10 CFR Part 61.42 performance objective, Protection of Individuals from Inadvertent Intrusion. The NRC staff did not change the status (open) or priority (periodic) of MF 1.01, Inventory in Disposal Structures, based on this TRR. The NRC staff also did not change the status (open) or priority (medium) of MF 1.02, Methods Used to Assess Inventory, based on this TRR. The NRC staff will continue to monitor the radionuclide inventory in emplaced saltstone under MF 1.01 and the DOE methods to determine inventory under MF 1.02 until the DOE completes saltstone emplacement. The NRC staff will continue the following monitoring activities, which are included in the Monitoring Plan (ADAMS Accession No. ML13100A003): review waste sampling reports and other documents the DOE issues periodically related to inventory, as listed in Table 1-1 of the Monitoring Plan (MF 1.01 and MF 1.02); monitor waste sampling methods and frequency to ensure reported radionuclide concentrations in salt waste are representative of waste emplaced in the SDF (MF 1.02); monitor the methods the DOE uses to track inventory emplaced in the SDF (MF 1.02); and monitor inputs of thorium-bearing waste to ensure additions of Th-230 are small compared to the ingrowth of Th-230 from U-234 (MF 1.01 and MF 1.02). The NRC staff recommends updating the Monitoring Plan to add the following monitoring activities after the NRC staff completes the Technical Evaluation Report: INVT-01 The NRC staff recommends updating Table A-1 of the Monitoring Plan to reflect the MPAD inventory in the 2020 PA under MF1.01. INVT-02 The NRC staff recommends monitoring additions to the SDF of radionuclides that are at or near the total projected SDF inventory under MF 1.01. The NRC staff recommends assessing the impact on dose to a member of the public and an inadvertent intruder if the SDF inventories exceed the MPAD inventory in the 2020 PA; INVT-03 The NRC staff recommends monitoring significant changes in waste characteristics and annual volumes of waste sent from H-Canyon to the Tank Farms to assess the potential effects on SDF inventory under MF 1.02; INVT-04 The NRC staff recommends monitoring significant changes in waste characteristics and annual volumes of waste sent from H-Canyon to the Tank Farms to assess the potential effects on I-129 and Tc-99 inventory in the SDF under MF 1.01; INVT-05 The NRC staff recommends monitoring the inventory of I-129 in individual disposal structures in comparison to the projected inventories in Table 5 of this TRR under MF 1.01. The NRC staff recommends assessing the impact on the dose to a member of the public and an inadvertent intruder if one or more individual disposal structures exceed the MPAD inventory in the 2020 PA. 8.0 References National Council on Radiation Protection (NCRP) Report No. 123, Screening Models for Releases of Radionuclides to the Atmosphere, Surface Water, and Ground, January 1996. U.S. Department of Energy (DOE), SRR-CWDA-2011-00115, Rev. 0, Saltstone Disposal Facility Case K Inventory Determination, August 16, 2011. ML113320433 ___, SRR-CWDA-2011-00044, Rev 1, Comment Response Matrix for NRC RAI-2009 Second Request for Additional Information on the Saltstone Disposal Facility Performance Assessment at the Savannah River Site, August 2011. ML113320303 ___, SRR-CWDA-2012-00002, Rev. 0, Revised Methodology for Determination of Inventories in SDF Vaults 1 and 4 through 9/30/2011, January 2012. ML12171A395 ___, SRR-CWDA-2012-00095, Rev. 1, Projected Technetium-99 Inventory in Saltstone Disposal Facility Units 2, 3, and 5, July 2012. ML1219A8307 ___, SRR-CWDA-2013-00147, Rev. 0. SDF Inventory Estimates for Transport Modeling, Savannah River Site, Aiken, SC, December 9, 2013. ML21181A117 ___, SRR-CWDA-2014-00006, Rev. 2, Letter from J. Folk to L. Camper Transmitting the DOE Fiscal Year 2014 Special Analysis for the Saltstone Disposal Facility at the Savannah River Site, October 29, 2014. ML14322A259 ___, 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-2015-00003, Rev. 0, Saltstone Disposal Facility Waste Inventory Disposed Estimator Model Report, January 2015. ML15335A099 ___, SRR-CWDA-2015-00020, Rev. 0, Memorandum: Consideration of Mo-93 as a Constituent in SDF Inventory, February 17, 2015. ML21187A295 ___, SRR-CWDA-2016-00072, Rev. 0, Fiscal Year 2016 Special Analysis for the Saltstone Disposal Facility, October 2016. ML18081A262 ___, SRR-CWDA-2015-00003, Rev. 0. Saltstone Disposal Facility Waste Inventory Disposed Estimator Model Report, January 2015. ML20206L216 ___, SRR-CWDA-2015-00020, Rev. 0. Consideration of Mo-93 as a Constituent in SDF Inventory, February 2015. ML21187A295 ___, SRR-LWP-2009-00001, Rev. 20. Liquid Waste System Plan, January 2016. ML21201A040 ___, SRR-CWDA-2015-00077, Rev. 2, Evaluation of I-129 Concentration Data to Improve Liquid Waste Inventory Projections, February 2018. ML20206L219 ___, SRR-CWDA-2015-00123, Rev. 2, Evaluation of Tc-99 Concentration Data to Improve Liquid Waste Inventory Projections, March 2018. ML18170A279 ___, SRR-CWDA-2018-00009, Rev. 0, Memorandum: Recommended Sampling Distribution for SDF Performance Assessment Inventory, May 2, 2018. ML21176A027 ___, SRR-CWDA-2018-00076, Rev. 0, Memorandum: Recommended Implementation of Inventory Sampling Distributions for the Saltstone Disposal Facility Performance Assessment GoldSim Model, October 24, 2018. ML20206L249 ___, SRR-CWDA-2018-00072, Rev. 0, Determination of the SDF Inventory through 9/30/2018, January 7, 2019. ML19179A042 ___, SRR-LWP-2009-00001, Rev. 21, Liquid Waste System Plan, January 2019. ML21201A046 ___, SRR-CWDA-2018-00044, Rev. 3. Memorandum: Inventory Screening Methodology and Application to the FY2019 Saltstone Disposal Facility (SDF) Performance Assessment (PA) Inventory, February 14, 2019. ML20206L241 ___, SRR-CWDA-2018-00041, Rev. 3. Determination of Inventory for FY 2019 PA Modeling, July 2019. ML20206L240 ___, SRR-CWDA-2019-00001, Rev. 0, Performance Assessment for the Saltstone Disposal Facility at the Savannah River Site. March 2, 2020. ML20190A056 ___, SRR-CWDA-2021-00047, Rev. 0. Comment Response Matrix for the First Set of U.S. Nuclear Regulatory Commission Staff Requests for Additional Information on the Performance Assessment for the Saltstone Disposal Facility at the Savannah River Site, May 2021. ML21147A104 ___, SRR-STI-2021-00165, Rev. 0. Results for the First Quarter Calendar Year 2021 Tank 50 Salt Solution Sample, May 2021. ML21148A002 U.S. Nuclear Regulatory Commission (NRC), U.S. Nuclear Regulatory Commission Technical Evaluation Report for the U.S. Department of Energy Savannah River Site Draft Section 3116 Waste Determination for Salt Waste Disposal, December 2005. ML053010225 ___, Request for Additional Information on the Draft Section 3116 Determination for Salt Waste Disposal at the Savannah River Site, May 26, 2005. ML051440589 ___, U.S. Nuclear Regulatory Commission Plan for Monitoring the U.S. Department of Energy Salt Waste Disposal at the Savannah River Site in Accordance with the National Defense Authorization Act for Fiscal Year 2005, May 3, 2007. ML070730363 ___, Onsite Observation Visit Report for the Savannah River Site Saltstone Production and Disposal Facilities, June 5, 2008. ML081290367 ___, Technical Evaluation Report for the Revised Performance Assessment for the Saltstone Disposal Facility at the Savannah River Site, Rev. 1, April 2012. ML121170309 ___, Letter from M. Satorius to T. Spears Acknowledging the DOE Response to the NRC Letter of Concern Regarding Disposal Activities at the SDF, August 31, 2012. ML12213A447 ___, U.S. Nuclear Regulatory Commission Plan for Monitoring Disposal Actions Taken by the U.S. Department of Energy at the Savannah River Site Saltstone Disposal Facility in Accordance with the National Defense Authorization Act for Fiscal Year 2005, Revision 1, September 2013. ML13100A113 ___, U.S. Nuclear Regulatory Commission Staff Comments and Requests for Additional Information on the Fiscal Year 2013 Special Analysis for the Saltstone Disposal Facility at the Savannah River Site, SRR-CWDA-2013-00062, Rev. 2, June 13, 2014. ML14148A153 ___, U.S. Nuclear Regulatory Commission Staff Request for Additional Information on the Fiscal Year 2014 Special Analysis for the Saltstone Disposal Facility at the Savannah River Site, SRR-CWDA-2014-00006, Rev. 2, June 26, 2015. ML15161A541}}