Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Deep Burial of Toxic Wastes INTRODUCTION STANLEY N. DAVIS University of Arizona AB STRACT Deep burial of toxic wastes provides several advantages over disposal in surface structures or by shallow burial. The primary advantage of deep burial is the high degree of physical isolation that it provides. Some of the hydrogeologic advantages of deep burial are (1) increase in length of flow path of contaminants that may become dissolved in groundwater; (2) increased protection of waste against weathering and erosion; (3) for some wastes and waste containers, elimination of free oxygen that may mobilize certain constituents; and (4) for plutonic rocks and to some extent all rocks, reduction of permeability with depth. Primary disadvantage of deep burial is the high cost of exploration, development, and monitoring of deep disposal systems. Reduction of permeability with depth in metamorphic and platonic igneous rocks is well defined to depths of about 300 m. Further reduction of permeability with depth probably takes place but is difficult to quantify using present data. Most of the reduction in permeability with depth in sedimentary rocks is within the upper 100 m. The reduction of permeability is generally at least three orders of magnitude from near the land surface to depths of 100 m for all types of indurated rocks. Based on hydrogeologic criteria alone, many different rock types should provide safe waste repositories at depths greater than 100 m. Deep repositories in most locations will eventually fill with water. However, if zones of significant ground- water circulation are either avoided or grouted, repositories at depths of more than 300 m in granitic rocks should take several hundred years to fill with groundwater once they are closed. In a well-placed repository, several thousand years may be needed to accomplish a simple piston-flow displacement of all water in the flooded repository. Even this length of isolation may not allow enough time for the degradation of all wastes to take place, but a long isolation time will most commonly mean a slow release of mobile contaminants into the biosphere, which, in turn, suggests that dilution will be more effective than in the case of a fast release of contaminants from shallow burial sites. Humans have produced toxic wastes since before recorded history. At least four factors, however, have made the problem of toxic waste disposal far more acute within the past 100 yr than at any previous period. First, and most obvious, is the 78 human population explosion and the fact that most of the ad- ditional people have accumulated in urban areas. The second is the development of numerous industrial processes that have toxic wastes as a by-product. The third is the development of the ability to concentrate and produce radionuclides. The fourth is the public awareness of the toxic waste problem, which has,
Deep Burial of Toxic Wastes paradoxically, made the solution both possible and at the same time more difficult. Public awareness has forced the formation of the legal frame- work to control the wastes and has stimulated funding to help solve the technical problems of disposal. However, concern about the hazards has given rise to an almost universal resis- tance to the creation of repositories for hazardous wastes. Com- monly, the same groups that have lobbied for the formation of strict laws that force the creation of special waste repositories join forces with local citizens to prevent the actual construction of repositories at any specific site. Partly as a consequence of this resistance, in some regions the actual disposal of waste is so difficult, owing to great distances to approved disposal sites or to excessive disposal costs, that illegal and highly dangerous dumping of wastes has been practiced. Even though the location, authorization, and preparation of disposal sites for hazardous wastes have been slow partly owing to the public's "not in my backyard" reaction, the additional time made available due to the delay is being used beneficially. Analytical methods are being developed that will allow a more precise characterization of hazardous wastes; alternative meth- ods of packaging, transportation, and disposal are being stud- ied; legal and administrative procedures are being perfected; and time for a balanced public education program is available. The present production of hazardous wastes in the United States is measured in several tens of millions of cubic meters per year (United States Congress, 19797. The exact amount varies according to the classification system used. Unfortu- nately, only a small fraction of this waste is currently treated or disposed of in a satisfactory manner (Council on Environ- mental Quality, 19811. Compaction, evaporation, and chemical treatment of the waste may eventually transform much of the waste to an innocuous form and reduce the volume of the remaining hazardous material to only a few million cubic meters per year, which will still require disposal in highly engineered structures. Even if this optimistic goal is achieved, the disposal of this volume of material still requires a major national effort. Because the principal mechanism for escape of hazardous waste to the biosphere is through groundwater flow, involvement of the geologic profession in this effort is essential for the proper solution of the problem (NRC Committee on the Geological Aspects of Industrial Waste Disposal, 1982) Wastes can be hazardous because of the presence of (a) toxic chemicals, (b) chemicals that are initially innocuous but turn toxic on degradation or reaction with other chemicals, (c) ex- plosive or combustible materials, (d) sharp objects that will cut or puncture, (e) large objects that will collapse or tumble, (f) radioactivity, and (g) pathogenic organisms. Unless noted otherwise, the remainder of this chapter will be confined to a discussion of only the more stable toxic chemicals and materials contaminated with radionuclides having half-lives in excess of several decades. Although the broad topic of the disposal of radioactive waste is not treated in detail, the similarity of the programs of disposing of transuranic and some low-level ra- dioactive material to problems of hazardous chemical wastes makes it convenient to combine all hazardous chemical and radioactive wastes in the discussion. 79 METHODS OF DISPOSAL A large number of methods of disposing of toxic wastes exist. Large concrete mausoleums, slurry injection in deep wells, trenches in desert regions, and repositories mined in bedrock are some of the options of disposal. From a purely hydrogeo- logic standpoint, trench burial in dry permafrost, which would receive prefrozen wastes, would be the most acceptable method of disposal. The delicate nature of the Arctic environment, the physical difficulties of operating in such a harsh climate, and its remoteness from points of origin of the waste, however, suggest that extensive use will not be made of dry permafrost for the disposal of hazardous wastes. The eventual choice of a method and geographic location will be a function of waste form, level of hazard present, duration of the hazardous char- acteristics, availability of a proper natural setting, and complex economic and legal factors. Although a number of disposal methods will be used eventually, this chapter will discuss only deep repositories mined in consolidated rocks. For this chap- ter, "deep burial" will be considered to start at depths of 100 m. The existence of alternative disposal methods is assumed, so that only highly compacted wastes that will remain hazardous for thousands of years are considered for deep burial. Actual burial of waste will be assumed to be in mined cavities within rocks of low porosity and permeability. Although ex- . . . cavat~on anct Placement ot wastes may be accomplished by remotely controlled methods, all shafts and cavities will be large enough for human entry for various purposes including testing, inspection, and monitoring. Deep burial will require consid- erable investment of effort and money for site preparation and sinking of shafts; consequently, large volumes of waste, perhaps as much as 105 m3 of radioactive waste or 106 m3 of chemical waste, would probably need to be buried in a single repository in order to justify the overall investment in such an undertak- ing. Even with multilevel placement of drifts, the excavation alone would most likely underlie at least 100 hectares. The concept of multibarriers to the migration of wastes, which has been discussed widely in connection with radioactive wastes (U.S. Department of Energy, 1979; Davis, 1982; NRC Board on Radioactive Waste Management, 1983), should be applied to all repositories. An attempt should be made to convert the waste into a chemically inert, nonpermeable form. The waste should then be packaged in strong, resistant containers to fa- cilitate handling and to retard contact with groundwater after burial. Drifts and shafts should be backfilled with material having a low permeability combined with a high capacity to sorb water-transported contaminants. The host rock should be grouted along fractures to reduce further the already low nat- ural permeability. The host rock should be chosen for its me- chanical stability, for its low permeability, and for its uniformity of mechanical and hydrogeologic properties. The repository should be placed in an area that would favor a long flow path of the groundwater prior to the groundwater's emerging at the surface or into a large body of water such as the ocean (Bre- dehoeft and Maini, 1981~. This placement would allow time for chemical and radioactive decay as well as dilution of the contaminants in the groundwater. Thus, the multiple barriers
80 would be waste form, packaging material, backfill, engineering modification of host rock, host rock, and, finally, the long mi- gration path prior to the emergence of contaminants at the land surface. ADVANTAGES AND DISADVANTAGES OF DEEP BURIAL Deep burial of toxic wastes has several advantages over shallow burial or surface storage systems. One of the most important advantages is that contaminants that may become dissolved in groundwater will not migrate directly to the land surface. The increased length of the groundwater flow path will allow time for the decay of radionuclides, the decomposition of unstable chemical compounds, and the dilution of toxic materials by dispersion. Deep burial will commonly place wastes in zones near or adjacent to natural saline or brackish water, which, in contrast with surface water, has little or no practical value and, if contaminated, would not represent a large loss. Deep burial will also afford protection against the possibility that the haz- ardous materials will be exposed at the surface through slow processes of erosion. Finally, and perhaps most important, deep burial will make human intrusion less likely. The depth of burial that may be chosen is related generally to rock permeability so that the deeper the burial, the lower the permeability. As will be shown, however, the reduction of permeability is small beyond depths of about 300 m. The re- duction in the first 50 m, in contrast, is most commonly at least one order of magnitude. Generally, pH, salinity, and alkalinity of natural water in- crease with depth, but concentrations of nitrate and dissolved oxygen decrease (Davis, 1981~. For some wastes, the chemical characteristics of deep groundwater may have some advantages over the chemical characteristics of shallow groundwater. For example, many of the transuranic elements will be less mobile in the deeper groundwater, where chemically reducing con- ditions prevail. The most serious disadvantage of deep burial is economic. Deep repositories will be expensive to construct; they will be difficult to monitor; and if errors are made or unexpected flaws in the repository are uncovered, the removal of the waste in order to place it in a better location will be costly. Estimates of costs (in 1982 dollars) of constructing deep re- positories for high-level nuclear waste generally range from about $1.5 billion to $1.7 billion (Waddell et at., 1982~. Such repositories would be designed to receive about 7.0 x 104 metric tons of high-level waste in about 1.25 x 105 containers of various types having a total volume of about 1.1 x 105 m3. The cost of the repository includes site acquisition, site im- provements, receiving facilities, excavation of underground workings, and ventilation systems. It does not include opera- tion costs nor waste preparation facilities. If one assumes equal costs for a mined repository for hazardous chemical wastes, then it becomes clear that such a structure can be justified only for the most hazardous wastes. For packages of nuclear waste, the cost will be in excess of $1O,OOO per m3 of packaged waste. If operating and processing costs are added, the total costs of STANLEY N. DAVIS disposal will be more than $30,000 per m3 of waste (Waddell et al., 1982~. Similar cost estimates are not available for a hy- pothetical deep repository constructed to receive only chemical wastes. However, the costs should be considerably less because the heat dissipation problem of high-level waste, which pre- vents close packing of the waste, will not be present. Therefore, the volume of chemical waste accommodated should be much larger than for high-level wastes, perhaps by an order of mag- nitude. A guess of possible costs might range from $5000 to $10,000 per m3 for chemical wastes, which would include con- struction of the repository and operating costs. Even with extra close packing of waste containers and relaxed problems of waste handling, it is hard to visualize a total cost of less than $1000 per m3 (or $1 per liter) for deep burial with isolation require- ments similar to that of nuclear waste. Costs of burial in shallower mined cavities would be much less. Estimates of costs in mined space in salt together with conventional facilities to handle packages of low-level radio- active waste suggest costs of about $100 (1978 dollars) per m3 for a volume of 1.5 x 10 m3 of waste to be placed in the repository each year (Wacks, 1979~. In conclusion, the cost of waste disposal will probably range from $100 to $200 per m3 for not-so-hazardous materials to a few thousand dollars per cubic meter for very hazardous chem- ical wastes and as much as several tens of thousands of dollars for high-level radioactive waste. Exact costs are most sensitive to variations in the density of waste placement in the subsur- face; requirements for special containers and packaging pro- cedures; and methods of transporting, storing, and handling the materials prior to subsurface disposal (Clark and Cole, 1982~. Total costs for repositories for high-level wastes are not highly sensitive to variations of depth and geology, which help de- termine the costs of the excavation of subsurface space, because processing, packaging, and handling together comprise the largest part of the total cost. A HYPOTHETICAL REPOSITORY Deep geologic repositories of various designs have been pro- posed for the disposal of radioactive wastes (St. John, 1982~. If the economic factors become favorable, similar repositories will undoubtedly be proposed also for highly toxic chemical wastes. From the standpoint of scientific factors, no general reason exists that properly packaged chemical wastes, particularly if they are in solid form, could not be placed in the same repos- itory with transuranic and low-level radioactive wastes. In the United States, however, institutional arrangements for the su- pervision and control of hazardous chemical wastes are gen- erally separated from radioactive wastes. Construction of a joint- use repository would seem to be unlikely for the next few decades. Consequently, the hypothetical repository described will be assumed to be only for the purpose of receiving toxic chemical wastes. The selection of a repository site will undoubtedly follow many of the criteria developed for locating sites for nuclear waste (Bredehoeft et al., 1978; NRC Panel on Geologic Site
Deep Burial of Toxic Wastes 81 Criteria, 1978; U. S. Department of Energy, 1979, 1983~. The erosion would generally be slower than with most other rocks following are general criteria as adapted from St. John (1982~: given equivalent geographic locations. (a) Site geometry (b) Host-rock properties Adequate depth, thickness, and lateral extent of host rock Low permeability and mechanical strength sufficient to allow for stable excavations; chemical composition of waste will not react adversely with rock (c) Hydrology Low groundwater velocities; long dis- tance from points of potential contami- nation to points of use of potable groundwater Uplift or subsidence at rates low enough not to be a threat to the longer-term stability of the site Located away from active faults that would threaten operational safety or long-term containment; not located in areas of major historical earthquakes Site would avoid areas of Quaternary (d) Tectonic stability (e) Seismic activ- ity and faulting <t Volcanism volcanism (g) Natural As far as practical, located away from resources areas that invite exploration for natural resources Similar criteria for the siting of near-surface disposal of haz- ardous wastes have been proposed by Longmire et al. (1981~. Redistribution of toxic material by erosion, human intrusion, deep plant roots, or burrowing animals is, however, potentially more troublesome in near-surface disposal than in deeper bur- ial in bedrock. An attractive compromise between shallow bur- ial and deep repositories has been proposed by Winograd (1974, 1981) for radioactive waste. He has pointed out that the non- saturated parts of alluvial fans in the deserts of the Southwest are commonly more than 50 m thick and provide an environ- ment generally free from percolating water. Moreover, many of the fans are either not being eroded or are still growing so that future uncovering by erosion would not take place for many thousands of years. A large number of geologic settings are being considered for repositories for radioactive wastes. Many of these will un- doubtedly meet the requirements. A discussion of the various strengths and weaknesses of all of these geologic settings is beyond the scope of the present review. I have chosen only one general type of setting for a more detailed discussion of a potential repository for toxic chemical wastes. This setting is within metamorphic and plutonic igneous rocks. This setting is considered a likely candidate because (1) large regions of the world have vast areas underlain by these rocks at shallow depths; (2) mining of these rocks is rapid and yields vaults that are stable for long periods; (3) areas having deposits of economic interest are identified and avoided more easily than in sedi- mentary rocks that may contain gas and oil; (4) unplanned human intrusion into the repository is less likely than in softer rocks; (5) the rocks generally have a low permeability and, where fractures are present, grouting can be completed effi- ciently; and (6) uncovering the contents of the repository by Conceptually, a repository will start at a depth of about 300 m and will be covered by at least 200 m of bedrock. An aquifer may be near the surface. As pointed out by Bredehoeft and Maini (1981) (Figure 5.1), this aquifer may actually be a fa- vorable feature, particularly if the aquifer discharges ground- water directly into the ocean or a large lake. Groundwater in the aquifer would serve to carry the small amount of escaping toxic material away from the site and would serve as a dilution mechanism, which is described later in this chapter. A repository shaft starting in metamorphic and igneous rocks at the surface will penetrate weathered material at varying depths (Figure 5.24. Joints and fracture zones tend to be weath- ered to the greatest depths. The joints and fractures commonly enclose large residual blocks of unweathered rock. Unusual care is needed, therefore, in this transition zone with its highly variable rock conditions to ensure proper sealing of the shaft during closure of the repository. HYDROGEOLOGIC PROPERTIES OF ROCKS AT DEPTH General Data from Wells and Test Holes Most permeability data for igneous plutonic and metamorphic rocks deeper than 100 m come from the interpretation of packer tests in drill holes (Snow, 1968; Lindblon, 1977; Davison, 1981~. A small number of aquifer tests using data from flowing or pumping wells in deep bedrock have also been completed (Ma- rine, 19677. For depths of less than 100 m, abundant data from both water wells and exploratory test holes are available (Davis and Turk, 1964; Stewart, 1966; Landers and Turk, 1973; Mundi and Wallace, 1973; Uhl, 1976; Gale, 1982~. The fact that the specific capacity per unit thickness of sat- urated rock penetrated by a water well decreases rapidly with depth has been noted by hydrogeologists for almost 100 yr. This trend is most marked in shale, limestone, metamorphic rocks, and plutonic igneous rocks. Typical graphs of data from different areas are shown in Figure 5.3. Two possible expla- nations for these trends exist. First, and most obvious, the permeability decreases with depth. Second, the data may be only an artifact of the way in which water wells are drilled and may only indicate lateral variations of permeability. In other words, where permeable rocks are encountered, drilling stops at a shallow depth and a high near-surface permeability is sug- gested. In contrast, where rocks of low permeability exist, wells are drilled deeper in search of water and the apparent overall permeability of the rocks is low. Although the type of bias in the data that I have suggested above may have some effect, it must be small. Davis and Turk (1964) and later-investigators (Snow, 1968; Lindblon, 1977; Johnson, 1981) have looked at data obtained during the explo- ration for dams and tunnels, some of which should have a bias in the opposite direction. Specifically, many deep holes have been drilled in order to search for permeable zones. Hence, data from a large number of drill holes would probably have a
82 ~1 Ventilation shelf t (Exhaust) i MAT a, . (around - water f low , 11 ~1 // / Waste receiving, inspection, and final packaging ~ I _ Monitoring ~ n n ~ hi/ ~ ~ ~ , ~~,,~-,,^~l -~ 7~~ -' ~ ! _ . _ · Double liner on shaft to be filled and , grouted on closure . ~o~ Water bearing fractures grouted ~~ Bedrock /)< ~ a/ \ Massive metamorphic or platonic igneous rocks 3uikhcad closure Zeolite tuft ~ Grout <= backfill ~ bib ~1 Of' Shaft t~for waste _ _ _ · . ~ Alluvial aquifer a Inner casing into upper part of bedrock / STAN LE Y N. DAVI S Service shaft l _£ i,~ _ - - O Ground-watcr f low .4 . ~° Grout back of inner casing ~ // /' ,/ / Subsurface control station S ~ '' containers for solid waste - Lr~ Inclined drif ~ for sea led canIstcre of liquid waste FIGURE 5.1 Type of repository development that may be possible for toxic chemical wastes. Features shown are not to scale, and all shafts and drifts are not shown. Full development would involve one or more levels having a gridlike network of drifts. Mining, placement of waste, and eventual closure would follow a sequence that is not indicated in the figure.
Deep Burial of Toxic Wastes o 20 40 , - 60 - I c:~80 _ 100 _ 120 _ \38 `" \ "" _ ~ `_ I ~ I_""~to~ \ \ ~ 26 20. \ '; _ 140 0 20 40 60 80 100 PERCENT UNWEaTHERED GRANITIC ROCK FIGURE 5.2 Percentage of unweathered rock as a function of depth at the Folsom Dam site in California. The total number of drill holes that supplied data for each depth interval is given next to each plotted point. Data are from unpublished reports of the U.S. Army Corps of Engineers and were compiled by Davis (1981). bias in favor of a greater permeability at depth. However, the location of the dam or tunnel most commonly favors areas of sound rock, so permeabilities of rocks within the area in general would tend to be lower than adjacent areas. The general con- clusion is that test-hole data also show a decrease of permea- bility with depth that is similar to water-well data (Figures 5.4, 5.5, and 5.6~. In general, the well-defined decreases in perme- ability are caused by at least the following five factors (Davis and Turk, 1964~: 1. A decrease in the effects of surficial weathering with an increase in depth (Figure 5.2~. 2. An increase in distance between joints, particularly sheet- ing joints, with depth. 3. An increase of lithostatic pressures with depth that tends to close fractures at depth. 4. A decrease with depth of fractures related to mass wast- ing. 5. A decrease with depth of the effect of topography on those stress patterns that might help contribute to localized rock failure. 83 ,000_0 ~ ~ ,oO.O - o - ~ at _ ~ - _ ,.0 0.001 0.01 0.1 1.0 10.0 100.0 l W.~°.\o \~ 0 x Statesville Area, North Carolina o Howard and Montgomery Counties, Maryland ~ Satpura Region, Central India Llano Area, Central Texas · Sierra Nevada, California ~ Eastern United Status ~~~ of 0 \,.~~~;' a---. ·. \ WELL YIELD (liters per minute per motor) FIGURE 5.3 Depth-yield relationship for water wells in metamor- phic and platonic igneous rocks. With the exception of the one line labeled "median," data used are from mean specific capacities of wells of different depths. Diagram by Johnson (1981). Even though a large number of mines in consolidated rocks are virtually dry, the fact that many actually produce large amounts of water (Cook, 1982) suggests that those sites where proper conditions exist for geologic isolation of wastes may be difficult to locate. As Cook has pointed out, much of the perme- ability, which allows an initial inflow to excavated openings, is local and has been caused by strains associated with the mining itself. If this permeability is potentially too large, repository extraction ratios must be kept low. Nevertheless, it should be remembered that most mines from which data are available are in hydrologically anomalous regions where various types of geologic discontinuities should favor natural zones of locally high permeabilities that contain water and would be inter- cepted by zones of artifically high permeability near the mines. Data on rock permeabilities from mines in general should, therefore, be considered as representing the higher extremes to be expected. .e solo 0 E - 1.0 Various locations in Swedish bedrock. ~ Crystalline rock in Oroville, California. - Auburn Dam site, California. Metamorphic rocks. lose 10-5 HYDRAULIC CONDUCTlVlTY (mesers per second) FIGURE 5.4 Results of packer tests in drill holes penetrating met- amorphic and platonic igneous rocks. Trend lines are based on the mean hydrualic conductivities at various depths Johnson, 1981).
84 500 - - 100 50 _ 1 ,/,l~ o f. o / ·/ ~ o a/ /- /o / 10 50 100 PERCENT ZERO WATER TAKE FIGURE 5.5 Diagram showing the increase in nonpermeable zones with depth in metamorphic (open circles) and platonic igneous rocks (black dots) of the Sierra Nevada, California. Data are from U. S. Bureau of Reclamation drill holes. Individual points represent data from at least 30 packer tests at various depth intervals. The percentage of packer tests that were unable to inject water are plotted for each depth interval (Davis and Turk, 1964). n 1 two Ann 500 , I HI 1 . ,,, 200 _ ,, ~ .RM / 7(170)/ ~ 1.0a '' 7(168,: 7(171' ; /6 11(237) ''' /. 11 (244) ''' 11(240)/ ,, _ 1 1(2~3)/ ~~ ~ / 11(244)/ 0 10 20 30 40 50 l 7(59) 8(187) / PERCENTAGE OF 2-m PACKER TESTS INDICATING HYDRAULIC CONDUCTIVITY OF 1 X 10-8 m/s (PERMEABI LITY OF 1.03 MD at 20°C) OR GREATER FIGURE 5.6 Diagram showing the reduction in the number of permeable zones with depth in metamorphic and plutonic igneous rocks of Sweden. Percentage values for each 50-m depth interval are plotted. Small numbers indicate the number of test holes in which the packer tests were performed, and the larger number in parentheses gives the total number of packer tests for each depth interval (Davis, 1981). STANLEY N. DAVIS Many permeable zones encountered in mines and tunnels drain rather rapidly, and little water is produced from these zones after a few days even though the zones may be a long distance beneath the regional water table. This general obser- vation suggests that many permeable fractures in metamorphic and platonic igneous rocks have a limited extension. This qual- itative conclusion is supported by aquifer tests and other tests completed on test holes in metamorphic rocks underlying the Savannah River Plant (Marine, 1981~. Geochemical Evidence Several aspects of water chemistry suggest that much of the water in the deep subsurface is almost static, and, under natural gradients, the movement of potential contaminants in this water toward the surface would be so slow that it would be negligible. One of the most important aspects of repository design is to ensure that this remains so. The most basic argument for the conclusion that the water is nearly static comes from the high chloride content of almost all deep water. Because significant amounts of chloride do not come from the dissolution of min- erals except in obvious cases where evaporites are present, the most important reason for the high chloride content in the water would be either the presence of ancient formation water or the concentration of chlorides through ion filtration. In the case of ancient formation water, if an active circulation system is connected with the land surface, the chloride would have been flushed out long ago by infiltrating surface water that has a chloride concentration from three to four orders of magnitude less than would be present in today's subsurface brines. In contrast, if subsurface brines are the end product of chlo- ride concentrating by means of ion filtration, water circulation would need to be quite vigorous in order to account for the total mass of chloride present in the aquifer. For example, if a sandstone outcrop, which is also the intake area of an aquifer, is 3 km wide and 1 m of water having 10 mg/L of chloride enters the aquifer each year, then for every 1.0 m along the strike of the aquifer, a total of 3 x 107 mg of chloride enters the subsurface each year. If the same sandstone has a porosity of 20 percent and extends downdip 50 km with an average thickness of 20 m and is saturated with water having 20,000 mg/L of chloride, then a total of 2 x 10~2 mg of chloride is present in each 1-m strip perpendicular to the strike. Thus, about 7 x 104 yr would be needed to accumulate this amount of chloride if the ion filtration is 100 percent efficient. The velocity of groundwater needed (750 m/yr) in this example would be very large for a natural system that might also be near a waste repository. The correct interpretation of the origin of the chloride in the groundwater is, therefore, critical to the safety evaluation of the waste-disposal system. Studies of the quality of water from deep drill holes and mines in metamorphic and plutonic igneous rocks generally show an increase of salinity with depth (Jacks, 1973, 1978; Marine, 1976; Davis, 1981; Frape and Fritz, 1981). In general, water contains less than 1000 mg/L total of dissolved solids at depths of less than 100 m. Concentrations increase to more than 10,000 mg/L as bedrock is penetrated to depths of more than 1000 m. Isotopic composition of the deeper water suggests
Deep Burial of Toxic Wastes that the major dissolved constituents are unrelated to modern surface water and that they probably are not derived from ancient seawater (Fritz and Frape, 1981; Nordstrom et al., 1982~. The salinity may be caused by the slow diffusion of small amounts of ions from original metamorphic and magmatic water in micropores in the dense rock into larger fractures that pen- etrate into the subsurface (Nordstrom et al., 1982~. The ability of minor amounts of interstitial brines to increase significantly the salinity of water in fractures has been demonstrated in the Hot Dry Rock Energy Extraction test in New Mexico. Here, water of low salinity was injected into large artificial fractures made within nonpermeable geologically young plutonic rock. The first return water from this injection showed large increases in total dissolved solids as well as a number of key ions. For example, chloride concentrations in one of the injection cycles increased from less than 50 mg/L in the injected water to 1750 mg/L in the initial water circulated back to the surface from the fresh fracture (Smith and Ponder, 1982~. To summarize the chemical evidence, in metamorphic rocks and plutonic igneous rocks the salinity of water in fractures increases with depth. If ion filtration can be discounted as a mechanism for concentrating these dissolved solids, then it is suggested that the chemical composition reflects water within fractures in these rocks, which approaches stagnation at depths of from 300 to 1000 m in most regions. This is, then, a further argument for the safety of deep burial of wastes within these rocks. PROBLEMS OF EVALUATING THE HYDROLOGIC HAZARDS OF A WASTE REPOSITORY Flow of groundwater through a waste repository is the most likely natural mechanism for the transportation of hazardous materials from the subsurface into the biosphere. Several ques- tions must be posed concerning this potential source of con- tamination. These can be generalized as follows: 1. If contamination reaches the surface, where will this take place? 2. What will be the predicted concentrations of these con- taminants in the water when it reaches a point of use? 3. What will be the total amount of contaminants to reach the surface? 4. How long will the contaminants take to reach the surface? Each of these questions will be discussed in a general way in relation to the hypothetical repository. Trajectory of Contaminants The trajectory of a contaminated plume of groundwater from the repository will be particularly difficult to predict within a few hundred meters of the repository. After an initial non- steady-state flow, which will be controlled by the repository geometry and construction and may last a few months, the subsequent pseudo-steady-state flow- of water will be controlled primarily by major fracture zones in the bedrock. Within the 85 major fractures, flow could be locally almost at right angles to the trajectory of the water, which might be predicted on the basis of regional hydraulic head measurements in the overlying aquifer. The location of the major fractures that form potential conduits is aided by geologic and geophysical methods but probably will never be precise enough to allow an exact defi- nition of the details of groundwater flow in the vicinity of the repository. Once the contaminated water reaches the more permeable upper part of the bedrock and the base of the over- lying aquifer, however, it will move downgradient in the same general direction as the bulk of the groundwater in the aquifer. After a few kilometers of migration, transverse dispersion should work the contaminated water into the upper part of the aquifer where it will eventually reach the biosphere through down- gradient wells, springs, or diffuse seepage into large bodies of water. The assumed upward diffusion into the aquifer, never- theless, could be inhibited almost indefinitely if the density of the contaminated water is much larger than normal ground- water and if the aquifer is horizontal with abundant clay lenses, which would have the effect of producing a strongly anisotropic flow in the horizontal direction. Concentration of Contaminants No method exists to predict concentrations of contaminants in groundwater passing through a repository because of the un- certainties in the source term. Containers for waste will be built to last as long as practical, and their rate of failure will be unknown. Once the container is breached, the contents will be removed slowly by water. Some rough estimates of the rates of this removal can be made based on the results of laboratory tests. Backfill around containers, however, will be designed to sorb as many of the contaminants as is practical. The rate of migration of contaminants through the irregularly shaped back- ffll probably can only be bounded in a broad way by generalized calculations. The usual method of handling the source-term problem of predicting concentrations is to assume some physically reason- able rate of removal of the repository contents and arbitrarily inject this hypothetical amount into a mathematical model of the moving body of groundwater. Sophisticated transport models then are used to predict downgradient migration rates and concentrations. The transport models, unfortunately, lend an air of authenticity that is rarely justified because of the arbitrary assumptions made in the source terms. The movement and dispersion of contaminants once the groundwater leaves the artificial cavities and enters the sur- rounding host rock will be very difficult to define in great detail. As the water moves into the regional flow field, nevertheless, useful approximations of contaminant concentrations can be made for any hypothetical source term. Total Amount of Contaminants that Reach the Surface A concern for the total amount of contaminants that eventually reach the surface is particularly acute in relation to the burial of radioactive wastes. This concern exists because, for lack of better evidence, adverse health effects from radiation are as-
86 sumed to be linearly related to the radiation dose received by humans. Thus, even very low levels of radiation widely dis- persed in the environment, if extended over a long period of time and if enough people are exposed, may cause the same total number of health effects as would a highly concentrated dose to which only a few people were exposed. One stategy for reducing the total amount of radionuclides reaching the biosphere is, therefore, to increase the time of isolation to allow for radioactive decay. The problem of isolation of hazardous chemical wastes has not been handled in the same way as radioactive wastes. The long-term stability of many compounds under conditions of burial are poorly understood. Certainly, some complex com- pounds will break down with time so that isolation renders the wastes less hazardous. Many compounds and elements such as cadmium and arsenic, in contrast, are stable for an infinite period, so that the time of isolation alone will not affect the total amount of contaminants that will eventually reach the surface. The seriousness of the eventual movement of trace amounts of these stable materials to the Earth's surface has not been studied in detail. Can we assume a linear relationship between the hazardous chemical and resultant adverse health effects? Probably not. For example, arsenic, in milligram amounts, is hazardous. However, current evidence suggests that arsenic in nanogram amounts is essential to human health (Mertz, 19811. Thus, simple dilution of wastes containing ar- senic could eliminate all adverse health effects from this con- taminant. In this example, the total amount of a contaminant reaching the surface is, therefore, of much less importance than the projected maximum concentrations, which would control the dose to individuals. Travel Time Estimates of the time that might be taken for a contaminant to be transported by groundwater from a subsurface repository to the land surface should be made for most repositories. In general, the longer the travel time is, the more favorable the site will be; although some hydrogeologic situations certainly exist where additional travel time does not necessarily mitigate potential hazards from toxic chemicals. Additional time in most settings will allow for dilution by molecular diffusion, decay of radiaoctive components, and decomposition of hazardous com- pounds. Most important, long travel times will help with the public acceptance of a site. Contaminants that might reach the land surface after thousands of years may be of little direct concern to the average citizen. The calcuation of travel time, unfortunately, is commonly accomplished only by assuming "reasonable" values for a num- ber of critical factors. Using different assumptions, hydrogeol- ogists can calculate travel times for waste from the same re- pository that may vary from one another by an order of magnitude or more. This is particularly true of irregularly fractured met- amorphic and plutonic igneous rocks. The following equation can be used to calculate the incre- ment of time, At, for the average time that a particle of ground- water takes to traverse a given distance, AL: STANLEY N. DAVIS Ne(/`L)2 At = (5.1) in which K is a measure of hydraulic conductivity of the rock, Ah is the head drop over the distance AL, and Ne is the effective (interconnected) porosity. The value of K can be estimated from field tests, and the values of AL and Ah are commonly deter- mined to some extent by known boundary conditions. The value of Ne however, for fractured rocks is rarely determined with accuracy and may easily vary from about 0.5 for weathered rock to 0.05 for highly fractured rock to less than 0.005 for dense, sparsely fractured rock. Calculated travel times would vary in the same way and are more often based on porosity values that are assumed rather than measured. Chemical variables introduce even greater uncertainties in travel-time estimates than do problems of defining the effective porosities of the rocks. Most chemical species dissolved in water will be sorbed to some extent on the solid matrix of the water- bearing materials. Even though Resorption takes place, most chemical species will partition strongly onto the solid matrix. This means that the chemical species that is a potential con- taminant will usually travel at only a small fraction of the ve- locity of the groundwater. Theoretically, the relative velocities of the chemical species and the groundwater can be measured by laboratory experiments. In practice, however, only an order- of-magnitude estimate is commonly possible. The chemical processes involved in the transport phenomena are a complex function of pH, chemistry of the solid surfaces of the rocks, nature of the dissolved species, temperature, water velocity, total volume of water flowing, other dissolved species in the water, and the relative concentrations of those species. These variables must be specified in time and space for the complex natural setting in order to estimate the velocities of contaminant migration. An early error of chemists dealing with this problem was to conduct laboratory experiments with artificially crushed rock and consider only mineralogical properties of the rocks, whereas the thin natural coatings along fractures in the rocks, which would not show up in bulk analyses, will actually be the most important control in the sorption process. A process similar to sorption in fractured media is that of molecular diffusion into the micropores in the otherwise solid rock. This process is important if the migration of the contam- inant is slow, which would allow time for diffusion to take place. Molecular diffusion would, therefore, serve to slow down the velocity of contaminant movement by allowing these contam- inants to migrate through the rock as well as through the more open fractures. FLOW OF GROUNDWATER THROUGH A CLOSED REPOSITORY Initial Inflow After the repository is filled with waste and access shafts are sealed, the repository should become saturated with water if it is below the water table. The rate of inflow of groundwater, however, should be quite small, particularly if fractures leaking
Deep Burial of Toxic Wastes significant amounts of water have been grouted during exca- vation of the repository. The process of saturation may take from decades to thousands of years, depending on the perme- ability of the rock and whether gases are easily expelled from the closed repository and also depending on the depth of the repository beneath the water table. Once the repository is saturated, water will move slowly into the structure and then drift out the downgradient side. If waste is in an insoluble form or if waste containers are watertight, the initial water flowing through the repository should not become contaminated. However, containers in the repository will ultimately fail, and "insoluble" material will dissolve to some extent so that water flowing through the repository will eventually be contaminated. The length of time taken for con- taminated water to start to move out from the repository after the repository has become saturated could be in the range of hundreds to thousands of years, depending primarily on the time that it takes to saturate the repository, waste form, and construction of waste containers. Steady-State Flow Conditions The quantity of water moving through the repository per unit time under steady-state conditions will be very small as illus- trated by the following example. The excavated region in the repository is assumed to measure 1500 m x 1500 m x 10 m and has an internal permeability much greater than the sur- rounding bedrock so that it intercepts a steam of water twice the width and height of the repository (Figure 5.7~. Assuming a hydraulic conductivity of 1O-4 m/day (see Figure 5.4Col- orado Front Range and Auburn data) and a regional hydraulic gradient of 10-3, then only 6 x 10-3 m3/day will move through the repository each day. If the repository has an overall porosity of 10 percent when pillars and other nonexcavated portions as well as backfill are considered, then the total pore volume in the respository is 2.25 x 106 m3. If the flow through the system is 6 x 10-3/ ._ ~ - Repository `'=~ - ,`~3 - E: - - - E o o o or K (bedrock) ~ 10-4 m/day Gradient ( L ) = 10-3 AN 6 x 104 m2 0=10-4(6 x 104)10-3-6 x 10~3m3/day 0~6 1 iters /day FIGURE 5.7 Map of a hypothetical repository showing groundwater flowing through the repository. The width of the zone of groundwater diversion is only approximate. The height of the repository, measured perpendicularly to the map, has been assumed to be 10 m. 87 day, then a once-through, pistonlike displacement of all the water in the repository would take 3.7 x 1O# days or about 106 yr. Even if the assumed grouting of larger cracks in the host rock is not effective and the average permeability is an order of magnitude larger, the time taken for piston displacement will still probably be at least 104 yr. This is hardly the picture one obtains when reading the literature where calculations of the dissolution rates of hazardous materials are based on lab- oratory batch tests or flow-through tests where simulated waste is exposed to periodically replaced water or to a constant stream of water greatly undersaturated with respect to the waste. A more exact field analogy of these types of laboratory experi- ments would be the placement of waste canisters directly in the Mississippi River rather than in a body of groundwater moving at velocities of much less than a millimeter/day and water that is already close to chemical equilibrium with respect to the repository contents. To be sure, a repository will not have pure piston displacement of water. However, simple cal- culations such as those given above but taking into account various degrees of dispersion will all show that the water moves very slowly in the repository and that thousands, if not hundreds of thousands, of years will be needed to flush out the water saturating the repository just once. Isolated Leaks The slow drift of small amounts of groundwater through the entire repository is not expected to present a significant health problem, at last for periods of many thousands of years. A more likely problem would be created by an isolated leak that could develop (1) along a major fracture, (2) through a poorly sealed shaft, or (3) by a borehole drilled into the repository. Stress release associated with the construction of the repository could also open existing small fractures enough to.provide significant leakage. The maximum amount of water cloning through an isolated leak is limited by the low permeability of the rock as a whole, which would not transmit much water to a permeable zone. It is further limited by the very low permeability of the backfill around waste cannisters, which would not allow sig- nificant water circulation past the Bannisters. Under extreme conditions, a properly constructed repository might be able to feed a maximum of 1 or 2 L/day to a fracture from a groundwater system driven by a normal hydraulic gradient. This water could then contaminate an aquifer, or, if an aquifer does not overlie the bedrock, the water moving upward could form a small contaminated spring. The fact that only a small volume of con- taminated water is involved in these processes is important. A well penetrating the contaminated aquifer overlying the re- pository simply would not be developed unless it had a yield of at least several hundred or, more commonly, several thou- sand liters per day. Therefore, at least a thousandfold dilution of the water would be expected before it would be used for drinking (Figure 5.8~. In addition to the fracture leakage shown in Figure 5.8, a poorly sealed shaft leading into a repository is also a possible source of leakage (Figure 5.9~. If a natural hydraulic gradient is assumed to be driving water through the repository and even if the repository offers no resistance to water movement, it is
Municipal = More than supply wel 1 200,000 L /day Domestic well= 2,OOO L/day i l C =~`J' Vl ~ U- ''.1 Contaminated water . is, ~ x~: I ~~ >;~= of// Dilution by a factor any' of at least 103. / \~3 Waste Repository L FIGURE 5.8 Even though small amounts of contaminated water may leave the repository through isolated fractures, this water will most probably be diluted before use, because water wells and springs are rarely developed unless they can yield several hundred to several thousand gallons per day. Although a leak of as much as 2 L/day, as shown in this figure, is considered unlikely, the contaminated water would probably be diluted by a factor of at least 103 before use. difficult to imagine more than a liter per day being circulated through the filled shafts. In fact, if the shafts were sealed by packing them only with silt having a hydraulic conductivity of 10-2 m/day (Davis, 1969>, the total leakage would be proably much less than 1.0 L/day (Figure 5.9~. Shafts will certainly be sealed with material much less permeable than silt, so, in gen- eral, shaft sealing is not a major problem unless unusual hy- draulic gradients exist between the repository and waters in the more accessible parts of the environment. DISCUSSION AND CONCLUSIONS The Federal Nuclear Waste Policy Act of 1982 and subsequent documents such as the U. S. Department of Energy's proposed general guidelines for the location of repositories for high-level radioactive waste (U.S. Department of Energy, 1983) have assumed that the first permanent storage of high-level waste will be in geologic repositories specifically mined for that pur- pose. If political and social problems can be overcome, there seems to be little doubt that such repositories will be con- structed. I have tried to make the case that similar repositories for highly toxic chemical wastes should be considered. Given identical geologic settings and construction methods, mined repositories for chemical wastes will probably be cheaper per unit volume of waste stored than for high-level radioactive STANLEY N. DAVIS //r/ / /} Fil led shafts Q Ah=1 m L by _ - - - Q it ~ ~ . , ~ ~ _ _ ~ ~ A=10 mama ~ ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,/, 777 K=~ crepes tOrY ;;t ~ A=IOm2 c 0 Q = K A L L= (500 + 500) m A= 10 m2 Ah 1 -3 = = 10 L 1000 K= 10~2m/day Q = (10-2)(10)(10 3)=10~4 m3/day Q = 100 ml /day 1 //~/ / Fi l led r shaft FIGURE 5.9 The small potential erect of poor shaft sealing is shown in this hypothetical example of a repository that has an infinite hydraulic conductivity so that the head drop is entirely within the 1000 no of the two filled shafts. Even though the shaft is "sealed" with a semiperme- able material, only 100 mL/day of contaminated water flows out of the respository. wastes because most chemical wastes will not generate large amounts of heat after being packaged for the repository. Wastes that might have strong exothermic chemical reactions after closure of the repository are assumed to be excluded from the repository. In contrast, all high-level wastes will generate sig- nificant amounts of heat, and close packing of waste in repos- itories must be avoided in order to prevent very high tem- peratures from building up in the storage area. In general, however, extraction ratios, mining methods, rock stability, and other factors may be more important than heat dissipation in determining overall construction costs. Even with the lower costs, however, storage of chemical wastes in large mined repositories will still probably range from a few hundred to a few thousand dollars per cubic meter of waste. Clearly, waste storage in a mined repository will be cost effective only for exceptionally hazardous materials. As is also true of repositories for high-level radioactive wastes, the long-term confinement of chemical waste in deep reposi- tories will be threatened most by human intrusion and by the transport of chemicals in solution through groundwater migra- tion. Deep burial in metamorphic and platonic igneous rocks probably will provide the most protection of any geologic ma- terials against human intrusion. The same geologic material will probably also be hydrogeologically satisfactory at depths greater than 300 m, although salt and shale are two geologic materials that may have lower permeabilities (Davis, 1969).
Deep Burial of Toxic Wastes AC KN OWLE D G M E NTS The presentation in this paper has been improved significantly by many constructive suggestions of N. G. W. Cook. The work of L. l. Turk and K. L. Johnson, former students and associates, has been most useful. These individuals, however, should be disassociated from my snore simplistic calculations and free- wheeling remarks. REFERENCES Bredehoeft, J. D., and T. Maini (1981). Strategy for radioactive waste disposal in crystalline rocks, Science 213, 293-296. Bredehoeft, J. D., et al. (1978). Geologic disposal of high-level radio- active wastes, U.S. Geol. Sure. Circ. 779, 15 pp. Clark, L. L., and B. M. Cole (1982). An analysis of the cost of mined geologic repositories in alternative media, Richland, Washington, Battelle Pacific Northwest Laboratory, Publ. PNL-3949, UC-70, 57 PP Cook, N. G. W. (1982). Groundwater problems in open-pit and un- derground mines, Geol. Soc. Am. Spec. Pap. 189, pp. 397-405. Council on Environmental Quality (1981). Contamination of Ground- water by Toxic Organic Chemicals, Executive Office of the Presi- dent, U.S. Government Printing Office, Washington, D.C., 84 pp. Davis, S. N. (1969). Porosity and permeability of natural materials, in Flow Through Porous Media, R. J. M. DeWiest, ea., Academic Press, New York, pp. 53-86. Davis, S. N., ed. (1981). Workshop on hydrology of crystalline base- ment rocks, Los Alamos National Laboratory, Rep. LA-8912-C, 63 PP Davis, S. N. (1982). Hydrogeology of radioactive waste isolation, the challenge of a rational assessment, Geol. Soc. Am. Spec. Pap. 189, pp. 389-396. Davis, S. N., and L. J. Turk (1964). Optimum depth of wells in crys- talline rocks, Ground Water 2(2), 6-11. Davison, C. C. (1981). Physical hydrogeologic measurements in frac- tured crystalline rock- summary of 1979 research programs at WNRE and CRNL, Inland Waters Directorate, Environment Canada, Publ. TR-161, 108 pp. Frape, S. K., and P. Fritz (1981). A preliminary report on the occur- rence and geochemistry of saline groundwaters on the Canadian shield, Atomic Energy of Canada Ltd., Tech. Rep. TR-136, 68 pp. Fritz, P., and S. K. Frape (1981). Saline groundwaters in the Canadian shield, a first review, Chem. Geol. Gale, J. E. (1982). Assessing the permeability characteristics of frac- tured rock, Geol. Soc. Am. Spec. Pap. 189, pp. 163-181. Jacks, G. (1973). Chemistry of some groundwaters in igneous rocks, Nordic Hydrol. 4, 207-236. Jacks, G. (1978). Groundwater chemistry at depth in granites and gneisses, KBS Tech. Rep. 88, 28 pp. Johnson, K. L., (1981). Permeability-depth relationships in crystalline rocks with applications to low-level waste repositories, unpublished M.S. thesis, U. of Arizona, Tucson, 112 pp. Landers, R. A., and L. J. Turk (1973). Occurrence and quality of groundwater in crystalline rocks of the Llano area, Texas, Ground Water 11(1), 5-10. Lindblon, U. E. (1977). Geological and geotechnical conditions in ground- water movements around a repository, in KBS Tech. Rep. 54(1), H. Stille et al., eds., pp. 1-49. Longmire, P. A., B. M. Gallaher, and J. W. Hawley (1981). Geological, geochemical and hydrological criteria for disposal of hazardous wastes in New Mexico, New Mexico Geol. Soc. Spec. Publ. No. 10, pp. 93- 102. Marine, I. W. (1967). The permeability of fractured crystalline rock at the Savannah River Plant near Aiken, South Carolina, U.S. Geol. Surv. Prof. Pap. 575-B, pp. 203-211. Marine, I. W. (1976). Geochemistry of groundwater at the Savannah River Plant, Savannah River Laboratory Publ. DP-1356, 102 pp. Marine, I. W. (1981). Comparison of laboratory, in situ, and rock mass measurements of the hydraulic conductivity of metamorphic rock at the Savannah River Plant near Aiken, South Carolina, Water Resour. Res. 17, 637-640. Mertz, W. (1981). The essential trace elements, Science 213, 1332- 1338. Mundi, E. K., and J. R. Wallace (1973). On the permeability of some fractured crystalline rocks, Assoc. Eng. Geol. Bull. 10, 299-312. NRC Board on Radioactive Waste Management (1983). A Study of the Isolation System for Geologic Disposal of Radioactive Wastes, Na- tional Research Council, National Academy Press, Washington, D.C., 345 pp. NRC Committee on the Geological Aspects of Industrial Waste Dis- posal (1982). Geological Aspects of Industrial Waste Disposal, Na- tional Research Council, National Academy Press, Washington, D.C., 44 pp. NRC Panel on Geologic Site Criteria (1978). Geological Criteria for Repositories for High-Level Radioactive Wastes, National Research Council, National Academy Press, Washington, D.C., 19 pp. Nordstrom, D. K., P. Fritz, R. J. Donahoe, and J. Ball (1982). Recent investigations of the major element, trace element, and isotopic geochemistry of deep granitic groundwaters at the Stripa test site, Sweden, Geol. Soc. Ann. Abstr. Programs 14, 577. St. John, C. M. (1982). Repository design, Underground Space 6, 247- 258. Smith, M. C., and G. M. Ponder, eds. (1982). Hot dry rock geothermal energy development program, Annual Report Fiscal Year 1981, Los Alamos National Lab. Publ. LA-9287-HDR, 140 pp. Snow, D. T. (1968). Hydraulic character of fractured metamorphic rocks of the front range and implications for the Rocky Mountain arsenal wells, Q. J. Colo. School Mines 63(1), 167-199. Stewart, G. W. (1966). Drilled water wells in New Hampshire, New Hampshire Mineral Resources Survey, Part 20, 58 pp. Uhl, V. W., Jr. (1976). The occurrence of groundwater in the Satpura Region of Central India, unpublished M.S. thesis, U. of Arizona, Tucson, 135 pp. U. S. Congress (1979). Hazardous waste disposal, U. S. House of Rep- resentatives, 96th Congress, 1st Session, Committee on Interstate - and Foreign Commerce Print 96-lFC31, 82 pp. U. S. Department of Energy (1979). Management of commercially gen- erated radioactive waste, draft environmental impact statement, U.S. Department of Energy Publ. DOE/EIS-0046-D. U. S. Department of Energy (1983). Nuclear Waste Policy Act of 1982; proposed general guidelines for recommendation of sites for nuclear waste repositories, Federal Register 48(26), 5670-5682. Wacks, M. E. (1979). Alternatives to shallow land burial for the disposal of low-level wastes, U. of Arizona Engineering Experiment Station, Quarterly Rep. on Contract L 28-9766F-1, 137 pp. Waddell, J. D., D. G. Dippold, and T. I. McSweeney (1982). Projected costs for mined geologic repositories for disposal of commercial nu- clear wastes, Columbus, Office of NWTS Integration, Publ. ON1- 3, Battelle Memorial Institute, Columbus, Ohio, 55 pp. Winograd, I. J. (1974). Radioactive waste storage in the arid zone, EOS 55, 884-894. Winograd, I. J. (1981). Radioactive waste disposal in thick unsaturated zones, Science 212, 1457-1464.