The LLNL near-field hydrologic source term (HST) model is based on a mechanistic approach to radionuclide retardation-that is, a thermodynamic description of chemical processes governing retardation in the near field, such as aqueous speciation, surface complexation, ion exchange, and precipitation The mechanistic approach allows for radionuclide retardation to vary both in space and time as a function of the complex reaction chemistry of the medium. This level of complexity is necessary for near-field HST transport modeling because of the non-linear reaction chemistry expected close to the radiologic source. Large-scale Corrective Action Unit (CAU) models-into which the near-field HST model results feed-require that the complexity of the mechanistic approach be reduced to a more manageable form (e.g. Linear, Langmuir, or Freundlich sorption isotherms, etc). The linear sorption isotherm (or K{sub d}) approach is likely the most simple approach for large-scale CAU models. It may also be the most appropriate since the reaction chemistry away from the near field is expected to be less complex and relatively steady state. However, if the radionuclide retardation approaches in near-field HST and large-scale CAU models are different, they must be proved consistent. In this report, we develop a method to link the near-field HST and large-scale CAU model radionuclide retardation approaches.

Reliable quantitative prediction of contaminant transport in subsurface environments is critical to evaluating the risks associated with radionuclide migration. As part of the Underground Test Area (UGTA) program, radionuclide transport away from selected underground nuclear tests conducted in the saturated zone at the Nevada Test Site (NTS) is being examined. In the near-field environment, reactive transport simulations must account for changes in water chemistry and mineralogy as a function of time and their effect on radionuclide migration. Unlike the Kd approach, surface complexation reactions, in conjunction with ion exchange and precipitation, can be used to describe radionuclide reactive transport as a function of changing environmental conditions. They provide a more robust basis for describing radionuclide retardation in geochemically dynamic environments. In a companion report (Zavarin and Bruton, 2004), a database of radionuclide surface complexation reactions for calcite and iron oxide minerals was developed. In this report, a second set of reactions is developed: surface complexation (SC) and ion exchange (IE) to aluminosilicate minerals. The most simplified surface complexation model, the one-site non-electrostatic model (NEM), and the Vanselow IE model were used to fit a large number of published sorption data and a reaction constant database was developed. Surface complexation of Am(III), Eu(III), Np(V), Pu(IV), Pu(V), and U(VI) to aluminum oxide, silica, and aluminosilicate minerals was modeled using a generalized approach in which surface complexation to aluminosilicate>SiOH or>AlOH reactive sites was considered equivalent to the reactivity of aluminum oxide and silica reactive sites. Ion exchange was allowed to be mineral-dependent. The generalized NEM approach, in conjunction with Vanselow IE, was able to fit most published sorption data well. Fitting results indicate that surface complexation will dominate over ion exchange at pH>7 for the rare earth and actinide ions examined here. Ion exchange is effectively suppressed due to aqueous speciation at high pH which tends to result in neutral or negatively charged aqueous species that are less likely to undergo ion exchange. The resulting set of average NEM and Vanselow IE constants provides a consistent set of constants for use in reactive transport simulations. The average NEM and Vanselow IE constants were used to predict single-mineral K{sub d}s under conditions similar to K{sub d} measurements reported by the Yucca Mountain site characterization program. In most cases, predicted Kds were consistent with measured K{sub d}s. In some cases, differences could be explained by surface area, mineralogy, or redox state. The NEM and Vanselow IE constants described here are an attempt to arrive at a consistent simplified database of reaction constants to be used in reactive transport simulations in chemically and mineralogically heterogeneous environments. The accuracy of these reaction constants is limited by the quality and quantity of available sorption data and the limitations of the NEM and Vanselow IE approach used. The reactivity and accessibility of natural minerals is complicated and cannot be assumed to behave ideally. Thus, the validity of the NEM and Vanselow IE constants must always be examined for the sediment of interest. For example, Triay et al. (1997) suggested that the weak sorption of Np(V) on tuff containing small amounts of hematite may indicate that the iron oxide mineral is passivated. Thus, the reactive surface area of hematite in these samples may be lower than expected. On the other hand, a limited comparison of NEM and Vanselow IE constants determined here and K{sub d}s reported by Wolfsberg (1978) for alluvium from Frenchman Flat, NTS, suggests that the reaction constants and reactive surface areas developed here would provide a conservative estimate of radionuclide retardation in Frenchman Flat alluvium.

For the last several years, the Underground Test Area (UGTA) program has funded a series of studies carried out by scientists to investigate the role of colloids in facilitating the transport of low-solubility radionuclides in groundwater, specifically plutonium (Pu). Although the studies were carried out independently, the overarching goals of these studies has been to determine if colloids in groundwater at the NTS can and will transport low-solubility radionuclides such as Pu, define the geochemical mechanisms under which this may or may not occur, determine the hydrologic parameters that may or may not enhance transport through fractures and provide recommendations for incorporating this information into future modeling efforts. The initial motivation for this work came from the observation in 1997 and 1998 by scientists from Lawrence Livermore National Laboratory (LLNL) and Los Alamos National Laboratory (LANL) that low levels of Pu originally from the Benham underground nuclear test were detected in groundwater from two different aquifers collected from wells 1.3 km downgradient (Kersting et al., 1999). Greater than 90% of the Pu and other radionuclides were associated with the naturally occurring colloidal fraction (

Temperature can significantly affect radionuclide transport behavior. In simulation of radionuclide transport originating from an underground nuclear test, temperature effects from residual test heat include non-isothermal groundwater flow behavior (e.g. convection cells), increased dissolution rates of melt glass containing refractory radionuclides, changes in water chemistry, and, in turn, changes in radionuclide sorption behavior. The low-yield (0.75 kiloton) Cambric underground nuclear test situated in alluvium below the water table offers unique perspectives on radionuclide transport in groundwater. The Cambric test was followed by extensive post-test characterization of the radionuclide source term and a 16-year pumping-induced radionuclide migration experiment that captured more mobile radionuclides in groundwater. Discharge of pumped groundwater caused inadvertent recirculation of radionuclides through a 220-m thick vadose zone to the water table and below, including partial re-capture in the pumping well. Non-isothermal flow simulations indicate test-related heat persists at Cambric for about 10 years and induces limited thermal convection of groundwater. The test heat has relatively little impact on mobilizing radionuclides compared to subsequent pumping effects. However, our reactive transport models indicate test-related heat can raise melt glass dissolution rates up to 10{sup 4} faster than at ambient temperatures depending on pH and species activities. Non-isothermal flow simulations indicate that these elevated glass dissolution rates largely decrease within 1 year. Thermally-induced increases in fluid velocity may also significantly increase rates of melt glass dissolution by changing the fluid chemistry in contact with the dissolving glass.

Most long-lived radionuclides associated with an underground nuclear test are initially incorporated into melt glass and become part of the hydrologic source term (HST) only upon their release via glass dissolution (Pawloski et al., 2001). As the melt glass dissolves, secondary minerals precipitate. The types of secondary minerals that precipitate influence the water chemistry in and around the melt glass. The secondary minerals also provide a sorption sink to the released radionuclides. The changing water chemistry affects the rate of glass dissolution; it also affects the sorption behavior of the released radionuclides. This complex nature of glass dissolution and its important role in defining the HST requires a thorough understanding of glass dissolution and secondary mineral precipitation. The identity of secondary minerals formed at temperatures from 40 to 200 C are evaluated in this report to assist in that understanding.

The goals of this task were to evaluate the availability of published temperature-dependent thermodynamic data for radionuclides and sorbing minerals and to evaluate the applicability of published estimation methods for temperature-dependent aqueous complexation, radionuclide mineral precipitation, and sorption. This task fills a gap in the hydrologic source term (HST) modeling approach, which, with few exceptions, has neglected the effects of temperature on radionuclide aqueous complexation, using 25 C complexation data for all temperatures without evaluating the consequences of this assumption. In this task, we have compiled thermodynamic data available in the literature and evaluated the options and benefits of applying temperature-dependent radionuclide speciation to future HST modeling. We use the recent experience of HST modeling at Cheshire (Pawloski et al., 2001) to focus our evaluation. Our literature search revealed that few thermodynamic data or extrapolation methods could be used to define the temperature-dependent speciation of key HST radionuclides Np, Pu, Am, and U, particularly for the higher valence-state (e.g., 5+ and 6+), the oxidation states most pertinent to NTS groundwater conditions at Cheshire. This suggests that using 25 C data for all temperatures may be the best modeling approach currently available. We tested established estimation techniques such as the Criss-Cobble method and other correlation algorithms to calculate thermodynamic parameters needed to extrapolate aqueous complexation data to higher temperatures. For some reactions, the isocoulombic method does allow calculation of free energy data and equilibrium values at higher temperatures. Limitations in algorithms and input data for pentavalent and hexavalent cations prevent extending temperature ranges for reactions involving radionuclides in these oxidation states and their complexes. In addition, for many of the radionuclides of interest, carbonate complexes appear to be the dominant complexes formed in NTS groundwaters, and data for these types of complexes are lacking for radionuclides as well as analog species. For the few species where enough data are available, the effect of temperature on radionuclide aqueous complexation has been calculated. These calculations allow partial estimation of the potential error that may be involved in ignoring speciation changes as a function of temperature, as was done in the Cheshire HST model (Pawloski et al., 2001). In some cases, differences between the most recent 25 C data available in the literature and data used in Pawloski et al. (2001) were more significant than calculated speciation changes as a function of temperature. To incorporate radionuclide speciation as a function of temperature, a robust set of temperature-dependent reaction constants is necessary. Based on our literature search and the few reactions that could be extrapolated to higher temperatures, the change in dominant complexes with temperature cannot be adequately addressed at this time. However, the effect of temperature on speciation can be qualitatively examined. In general, the log K values for radionuclide complexation reactions considered here increase with increasing temperature, suggesting that increasing temperature may enhance radionuclide aqueous complexation. However, complexation reactions often involve H{sup +} and reactant species such as carbonate which exhibit their own temperature-dependent speciation. Thus, any change in the value of a radionuclide complexation log K may be offset or enhanced by temperature effects on pH and carbonate speciation. In addition, sorption processes that involve surface complexation change with increasing temperature, and these reactions may enhance or negate the mobility effects of any increase in aqueous complexation with temperature. While increasing temperature may increase complexation, it also may reduce or increase ligand concentrations through shifts in speciation. Similarly, higher temperatures may favor or reduce sorption and/or co-precipitation in mineral phases. Consequently, the net effect on radionuclide mobility of increasing temperature depends on the effects of temperature on a number of geochemical processes. Thus, it is even difficult to make qualitative assumptions about the direction much less the magnitude of temperature effects on radionuclide mobility. Until sufficient data become available in the literature to precisely capture the effects of temperature on radionuclide complexation, it appears unwarranted to invest in complex estimation techniques based on extrapolations from available data.

In the Yucca Flat basin of the Nevada Test Site (NTS), 747 shaft and tunnel nuclear detonations were conducted primarily within the tuff confining unit (TCU) or the overlying alluvium. The TCU in the Yucca Flat basin is hypothesized to reduce radionuclide migration to the regional carbonate aquifer (lower carbonate aquifer) due to its wide-spread aerial extent and chemical reactivity. However, shortcuts through the TCU by way of fractures may provide a migration path for radionuclides to the lower carbonate aquifer (LCA). It is, therefore, imperative to understand how radionuclides migrate or are retarded in TCU fractures. Furthermore, understanding the migration behavior of radionuclides once they reach the fractured LCA is important for predicting contaminant transport within the regional aquifer. The work presented in this report includes: (1) information on the radionuclide reactive transport through Yucca Flat TCU fractures (likely to be the primary conduit to the LCA), (2) information on the reactive transport of radionuclides through LCA fractures and (3) data needed to calibrate the fracture flow conceptualization of predictive models. The predictive models are used to define the extent of contamination for the Underground Test Area (UGTA) project. Because of the complex nature of reactive transport in fractures, a stepwise approach to identifying mechanisms controlling radionuclide transport was used. In the first set of TCU experiments, radionuclide transport through simple synthetic parallel-plate fractured tuff cores was examined. In the second, naturally fractured TCU cores were used. For the fractured LCA experiments, both parallel-plate and rough-walled fracture transport experiments were conducted to evaluate how fracture topography affects radionuclide transport. Tuff cores were prepared from archived UE-7az and UE-7ba core obtained from the USGS core library, Mercury, Nevada. Carbonate cores were prepared from archived ER-6-1 core, also obtained from the USGS core library, Mercury, Nevada.

Underground nuclear tests conducted at the Nevada Test Site (NTS) are characterized by abundant fission product and actinide source terms. Included are {sup 99}Tc and other soluble radionuclides ({sup 3}H, {sup 14}C, {sup 36}Cl, {sup 85}Kr, and {sup 129}I), which are presumably mobile in groundwater and potentially toxic to down-gradient receptors. NTS provides the Office of Civilian Radioactive Waste Management (OCRWM) with an analog of the release of these radionuclides from a nuclear waste repository in the absence of engineered barriers. The investigation described in this report synthesizes a substantial body of data collected on the identity and distribution of soluble radionuclides at field scales over distances of hundreds of meters, for durations up to 40 years, and under hydrogeologic conditions very similar to the proposed geological repository at Yucca Mountain. This body of data is complemented by laboratory transport studies and a synthesis of recent modeling investigations from the NTS, with an emphasis on the ongoing Yucca Mountain Program (YMP) efforts. Overall, understanding the controls of radionuclide mobility associated with these nuclear tests will provide insight into the repository's future performance as well as bounds and calibrations for the numerical predictions of long-term radionuclide releases and migration.