The Topopah Spring Member of the Paintbrush Tuff from the Nevada Test Site is being investigated by the Nevada Nuclear Waste Storage Investigations project (NNWSI) as a possible nuclear waste repository host rock. Changes with time of the permeability and fluid chemistry of the Topopah Spring Member have been measured in samples subjected to a temperature gradient. Maximum temperatures of the imposed gradients ranged from 90° to 250°C; minimum temperatures were 36° to 83°C. Confining and pore pressures simulated a depth of about 1.2 km, which is greater than the proposed repository depth, but chosen for comparison with previous studies at these pressures. Pore fluid used in the experiments was groundwater from the Nevada Test Site; the direction of pore-fluid flow was from the high- to the low-temperature side of the tuffs.

Initial permeabilities of the tuff samples ranged from 3 to 65 μdarcys, the wide range in values resulting from differences in the void and fracture geometries of the samples. Heating the tuffs produced no change in permeability in tne lowest temperature experiment and only small changes at higher temperatures. The fluids discharged from the tuffs were dilute waters of nearneutral pH that differed only slightly from the original groundwater composition.

Since proposed burial in the Topopah Spring Member would be in the unsaturated zone, the high initial permeabilities and the absence of permeability change with heating may be desirable, because downward-percolating waters would be able to drain into deeper formations and not collect at the repository level. in addition, any fluids that may come in contact with waste canisters wlll not have acquired any potentially corrosive characteristics through interaction with the tuff.

In appropriate aquifers the natural radionuclides of the U and Th decay series are important sources of information about the behavior of radwaste radionuclides in far-field aquifers. The Wolfcamp Carbonate, Pennsylvanian Carbonate and Granite Wash aquifers in the Palo Duro Basin of the Texas Panhandle are prime examples of such aquifers. Sampling and analysis for key radionuclides in the ground waters of these aquifers are quite feasible and have been accomplished. Key early results are: (1) Ra does not appear to be retarded by sorption, (2) Th appears to be strongly sorbed, (3) kinetics seem to be different on time scales of days to months than on ones of hundreds of thousands of years, and (4) U and Th behave similarily when the time scales (half-lives) are similar, leading to the suggestion that uranium is in the +4 valence state in these aquifers.

The sorption of americium(III) and neptunium(V) on geologic media has been measured under oxic conditions. The aqueous phase was generally a synthetic groundwater. The solid phase was a pure mineral, representative of crystalline granitic rocks. Altogether 40 minerals and rocks were used. The effects of parameters such as pH and ionic strength of the aqueous phase as well as cation exchange capacity and surface/mass ratio of the solid sorbent are discussed. The sorption mechanism in the environmental pH-range can be denoted as physical adsorption process, which is largely governed by pH and the speciation in the aqueous phase and to a less extent by the exchange capacity of the solid.

The site-screening process is designed to provide information on rock formations that will serve as the basis for recoximnding potentially suitable sites for locating a nuclear waste repository. While the application of criteria to extrinsic characteristics of the host rock, such as minimum depth and areal extent, are based on requiremnts associated with repository design and waste isolation, the significance of intrinsic characteristics (e.g., thermal conductivity, heat capacity, Young's modulus) is less firmly based. This is of particular importance with crystalline or grantitic-type rocks since they exhibit a wide range of canpositions and properties.

The range of crystalline rock property variations has been defined and results describing the impact of these variations on repository thermal, mechanical, chemical, and hydrological performance constraints obtained. Results also are presented for various repository design options (emplacement mode, thermal loading, package power/pitch). For the analysis, rock properties and repository design elements are treated as stochastic parameters by statistical methods. The output range of a performance parameter is determined with respect to the range of stochastic parameters to establish their importance in site selection.

We report our findings for two studies of Columbia River basalts that are relevant to United States radwaste disposal programs. A rhyodacite intrudes the Wanapum basalt and is exposed in a quarry along the Hood River, Oregon. Since the rhyodacite is different geochemically from the basalt, and was intruded at a high temperature, then this site can be used to evaluate the effects of the intrusion on the basalt. In such a study we assume the intrusive rock is a high temperature heat engine, and that cooling may be by convection, conduction, or both. Contact effects of the intrusion are not pronounced in the basalt, and are restricted to minor recrystallization within one meter of the contact, and to possibly some minor fracture filling. Traverses were made from the rhyodacite into the basalt, and samples at closely spaced intervals analyzed for K, U, Th by field and laboratory gamma spectrometry, and by neutron activation analysis for some 35 elements. Separate aliquots were analyzed by isotope dilution to inspect Sr isotopic variations. The chemical studies do not indicate migration of any elements from the rhyodacite into the basalt, nor for any in-basalt or in-rhyodacite migration. Despite a pronounced chemical gradient between the rocks, each has remained a closed, isochemical system. The Sr isotopic studies indicate that the rhyodacite possesses Sr (87/86) = 0.70407 ± 0.00008 while the intruded basalts possess Sr (87/86) = 0.7052 ± 0.0002. Even in samples of the contact, with intermingled rhyodacite and basalt, each rock preserves its own characteristic Sr (87/86) signature, thus again attesting to the closed system conditions. We conclude for this natural experiment that the Wanapum basalt is not readily penetrated by elements from a high temperature, molten source.

A number of tests have been performed, under controlled conditions, with the SWIFT model in order to complete its quality-assurance records. Some of these tests verify the numerical algorithms employed by the code implementation through comparison with analytical data. Other tests validate the processes addressed by the model through comparison with field data. These tests show good agreement between SWIFT and the reference data. In this paper several of the tests are highlighted.

The health risks from radioactive waste may be expressed as a sum of products of transfer factors that characterize the causal chain of events between disposal of radionuclides in a waste field and the consequent health effects. “Model” estimates for the transfer factors are commonly obtained by modeling transport and other mechanisms in the subsystems that form the links in the causal chain. “Natural” estimates of some transfer factors for naturally occurring radionuclides can be obtained from data on the concentrations of naturally occurring radionuclides in soil, food, and the human body. These “model” and “natural” estimates can be used with scaling procedures to estimate the uncertainties and to obtain better estimates of the transfer factors. The scaling procedures take into account the differences in the source characteristics for radionuclides in a waste field of limited size and for radionuclides generally distributed in the natural environment. The ratios of the natural estimates to the model estimates for several transfer factors and several radionuclides belonging to the U-238 decay series have been determined. These ratios range from 1/8 to 4/1 for food-concentration/sourceconcentration transfer factors for the food pathways and from 1 to 77 for dose-rate/source-concentration transfer factors for the internal radiation dose pathways to various organs.

Ground water samples (brines) from deep wells in the Palo Duro Basin, Texas are being analyzed for noble gases in an attempt to obtain radiometric ages for these brines. The brines contain radiogenic 4He and 40Ar produced from the radioactive decay of U, Th, and K. Consideration of hydrochemical data for the brines, various isotopic, chemical, and mineralogical data for the aquifer rocks and noble gas production rates allow estimating the age of the brines to be about 130 million years at two wells. At a third well interaquifer mixing has occurred and the age is presently indeterminate.

Comparison of trace element signatures of country rocks as a function of distance from the contact with two pegmatites, Tin Mountain and Etta, in the Black Hills of South Dakota, suggests that some elements such as K, Li, Rb, Cs, As, Sb, Zn and Pb, have migrated to distances of 4 to 40 meters during contact metamorphism. The relative degree of migration varies depending on the element. On the other hand, there is virtually no migration of rare earth elements (REE), Al, Sc, Cr, Hf, U, and Th. Biotite and muscovite are effective trace element traps for Li, Rb and Cs. Biotite has a greater affinity for Rb, Cs and Li than muscovite.

Nuclear waste emplacement in devitrified volcanic tuff at Yucca Mountain will raise the temperature of surrounding rock for a geologically significant period of time. This study evaluates the susceptibility of an underlying 50 ft-thick vitrophyre to thermal alteration by examining alteration that occurred in the rock as it cooled after deposition. A 10°C temperature rise should have no mineralogical effects on the vitrophyre, but an increase of 60° or more is likely to result in alteration. Expected mineralogic changes in the vitrophyre caused by this amount of thermal loading include crystallization of zeolites and smectite. Alteration will be concentrated in a thin interval near the top of the vitrophyre and along fractures. Adsorbed water and water in preexisting hydrous minerals and in glass may contribute to hydrothermal alteration of underlying vitrophyre. Bulk porosity change would be slight and local porosity increase would probably be restricted to the upper part of the vitrophyre. Although some fracture filling could occur, such a minor sealing effect would be balanced by development of secondary porosity. Zeolites and smectite, newly-crystallized along fluid flow paths below the waste repository, could provide an enhanced sorptive barrier to radionuclide migration.

The room-temperature thermal conductivities of two kinds of tuff from the Nevada Test Site have been measured on a linear-heat-flow thermal comparator. The results are the basis for an empirical model of the conductivity of these rocks in the dry and water-saturated conditions as a function of porosity. Tuff is one medium being considered for nuclear waste disposal. Devitrified, non-zeolitized tuffs with 6-22% porosity, and non-welded, zeolitized tuffs with 24-35% porosity were examined. The empirical geometric mean model is used to characterize the porosity-dependent thermal conductivity. We consider the rock matrix conductivity and the effective fluid conductivity as adjustable parameters in fitting the model to data for saturated and dehydrated samples, separately. For each rock type, the fitted value of matrix conductivity is the same in both the saturated and dehydrated cases; 2.3 and 1.1 W/(m.K) for the non-zeolitized and zeolitized tuffs, respectively. The fitted values for fluid conductivities are different for the the zeolitized and non-zeolitized tuffs, but the ratio of fitted water conductivity to fitted gas conductivity is very nearly the same for both rock types. This permits the use of a single equation to predict with good accuracy the ratio of saturated to dry rock conductivity.

A tool has been developed to assist investigators in exploring common problems found in radionuclide transport work. This tool consists of an integrated series of programs known as the Simplified Codes for Performance Evaluation (SCOPE) system. The heart of the system is a set of programs which make use of analytical solutions to the onedimensional solute transport for up to three-member decay chains. Longer chains are approximated by combining the results of the one-, two-, and three-member chain solutions with the effects of short lived members.

SCOPE also provides programs that allow the user to evaluate the results by integrated radionuclide releases or by dose-to-man. In the first case, the integrated release passing through the boundary of an aquifer over time is compared to a given set of criteria. The criteria normally used are those developed by EPA (Radiation Protection Program 1982) [1] for estimating hazard. To provide the user with the option of calculating dose-to-man, two existing programs (ALLDOS and PABLM) were integrated into the system.

The system is not intended to solve all classes of radionuclide transport problems. However, it can provide the user with a good “first cut” understanding of many geohydrologic systems by allowing the problem to be looked at first as one-dimensional flow in a homogeneous geology.

Sorption of 89Sr, 58Co, 134Cs and 241Am in rock and mineral samples was determined by the batch method using crushed sanmples and by autoradiographic method using rock thin sections. The rock samples were tonalite and rapakivi granite, representing the Olkiluoto and Loviisa nuclear power plant sites, respectively. The autoradiographic method was found to be applicable sorption of radionuclides in rock and mineral samles. The 8orption order of nuclides in rock samples was: KaAM> KaCs>KaCo> KaSr the sorption order of the different minerals was biotite> hornblende >plagioclase> potassium feldspar> quartz. Distribution coefficients (Ka) obtained by the batch and autoradiographic methods were of the same order of magnitude. Ka-values of Sr, Co, Cs and Am for tonalite were 5, 20, 190 and 40 × 10-4 m3/m2 by the batch method and 1, 9, 130 and 170 × 10-4 m3/m2 by the autoradiographic method.

The rate of dissolution of low-solubility species from waste forms in a geologic repository can be calculated from a theoretical analysis of the time-dependent rate of mass transfer by diffusion and convection Into the groundwater surrounding the waste, assuming a concentration at the waste-form surface eanal to the solubility of the radioelement. The analytical solutions for time-dependent mass transfer are reduced to asymptotic steady-state approximations valid over specified ranges of repository conditions. The predicted steady-state dissolution rates are considerably below those observed in laboratory leaching experiments with borosilicate glass and with other waste forms, indicating that the solid-liquid chemical reaction rates measured in the laboratory experiments are greater than the rates of diffusive-convective mass transfer in the concentration boundary layer surrounding the waste form in a geologic repository. It is shown that the time to reach steady-state dissolution can be as short as a few years, if convective mass transfer in the concentration boundary layer is important, to many thousands of years, If mass transfer is mainly by diffusion in the groundwater. The steady-state mass transfer rate can be Increased, and the time to reach steady state decreased, by sufficiently short half lives of the dissolving species. The theory has been extended to Include the effect of backfill on the steady-state mass transfer from the waste-form surface into moving groundwater.

The mass-transfer theory has been extended to include the effect of time-dependent solubilities, diffusion coefficients, and retardation coefficients, which provides a means of calculating the time-dependent dissolution of lowsolubility species from waste exDosed to groundwater during the Period of repository heating. The effects of timetemperature transient dissolution on far-field concentrations and cumulative release are calculated.

The transient and steady-state diffusion of radionuclides through a finite backfill layer separating a finite waste solid and porous rock has been analyzed, Including the effects of radioactive decay. Previous backfill analyses have neglected radioactive decay and have assumed an infinite amount of backfill. The results show that the break-through time and rate of radionuclide release depend on properties of the backfill and surrounding rock and on the waste form dimensions.

Laboratory tests have been designed to validate these theoretical analyses of diffusive-convective mass transfer with solubility-limit boundary conditions. Preliminary tests are now underway at the Battelle Northwest Laboratory.

Peak far-field concentrations of more soluble radionuclides such as cesium-135, with suitably long radionuclide transport times and sufficiently large axial dispersion, are shown to he insensitive to dissolution rate. Equivalent phenomena occur in fracture-flow radionuclide transport.

Spent nuclear fuel buried in deep geologic repositories may eventually be wetted by water. The alfa-radiation will radiolyse the water and produce hydrogen and oxidizing agents, mainly hydrogen peroxide and oxygen. The hydrogen will escape by diffusion and the oxidizing agents may attack the canister materials, oxidize the uranium oxide matrix or diffuse out and oxidize reducing agents in the surrounding rock.

The rate of radiolysis has been computed recently within the Swedish nuclear fuel safety projects KBS. It is strongly influenced by the amount of available water and by the presence of dissolved iron. The movement of the oxidizing agents out from the canister and their reaction with the reducing agents (mainly ferrous iron) in the Swedish crystalline rock has been modelled as well as the movement of the radionuclides within and past the redox front. Some substances such as uranium, neptunium and technetium will precipitate at the redox front and will be withdrawn from the water to a considerable extent.

In near-field regions of nuclear waste repositories, thermodynamically coupled flows of heat and matter can occur in addition to the independent flows in the presence of gradients of temperature, hydraulic potential, and composition. The following coupled effects can occur: thermal osmosis, thermal diffusion, chemical osmosis, thermal filtration, diffusion thermal effect, ultrafiltration, and coupled diffusion. Flows of heat and matter associated with these effects can modify the flows predictable from the direct effects, which are expressed by Fourier's law, Darcy's law, and Fick's law. The coupled effects can be treated quantitatively together with the direct effects by the methods of the thermodynamics of irreversible processes. The extent of departure of fully coupled flows from predictions based only on consideration of direct effects depends on the strengths of the gradients driving flows, and may be significant at early times in backfills and in near-field geologic environments of repositories. Approximate calculations using data from the literature and reasonable assumptions of repository conditions indicate that thermal-osmotic and chemical-osmotic flows of water in semipermeable backfills may exceed Darcian flows by two to three orders of magnitude, while flows of solutes may be reduced greatly by ultrafiltration and chemical osmosis, relative to the flows predicted by advection and diffusion alone. In permeable materials, thermal diffusion may contribute to solute flows to a smaller, but still significant, extent.

The physical processes modeled and the mathematical and numerical methods employed in a simulator for non-isothermal flow of water, vapor, and air in permeable media are briefly summarized. The simulator has been applied to study thermohydrological conditions in the near vicinity of high-level nuclear waste packages emplaced in unsaturated rocks. The studies reported here specifically address the question whether or not the waste canister environment will dry up in the thermal phase.

The unsaturated, fractured tuff of Yucca Mountain in the Nevada Test Site is one of the potential sites for geologic storage of high-level radioactive waste. A modeling study of flow and transport in this geologically complex site is presented. Numerical models of mass and heat flow in conjunction with analytical solutions are being used for sensitivity and pathway analysis studies and to aid in design and interpretation of laboratory and field flow and transport tests in tuff.

The shear behavior and normal closure behavior of rock joints have a major influence on the changes in joint conductivity to be expected during the life of a repository. Methods are described for obtaining the joint parameters required to model these conductivity perturbations. They include the joint roughness and the wall strength, which can both be obtained from simple tests on jointed drill core. Examples are shown that illustrate the shear stress-displacement, dilation and conductivity coupling that occurs when a joint of given size is subjected to shear under various levels of effective normal stress. Major increases in conductivity may occur. Examples are also shown that illustrate the stress-closure-conductivity coupling. This comprehensive joint behavior model indicates that important size-effects exist in shear behavior which are related to the joint spacing or effective in situ block size. The discrete modelling of near-field joint and fracture performance is essential for predicting the long-term integrity of a nuclear waste repository.

The radiological hazard from the disposal of high-level nuclear waste deep within the ocean sediments is evaluated, on a preliminary basis, for locations in the central North Pacific and in the northwestern Atlantic. Radionuclide transport in the sediment and water column and by marine organisms is considered. Peak doses to an individual are approximately five orders of magnitude below background levels for both places. Sensitivity analyses for several imporLant aspects of the post-emplacement systems model are described.