16
Volcanic Hazard Assessment for Disposal of High-Level Radioactive Waste

BRUCE M.CROWE

Los Alamos National Laboratory

ABSTRACT

Volcanic hazard studies for disposal of high-level radioactive waste pose a number of unique problems. These include the long time frame of hazard assessment (104 to 106 yr), the limited geologic record of volcanic activity at disposal sites and the political sensitivity of this national problem. The major variables affecting volcanic hazards are the structure of magma feeder systems at repository depths and the magma fragmentation and dispersal energy of eruptions. The latter is generally controlled by magma composition and the presence or absence of groundwater. Long-lived volcanic fields (>1 m.y.) provide the greatest potential risk for waste disposal, but these can be avoided by proper site selection. Short-lived volcanic fields are more difficult to avoid but are generally mafic in composition, which results in smaller disruption zones and explosive eruptions of lower energy than those of long-lived centers.

Volcanic hazards are evaluated through risk assessment, which is a product of probability and consequences. These studies have been completed for a potential waste disposal site in the Nevada Test Site (NTS). Cenozoic volcanism of the NTS region is divided into three distinct episodes. The youngest episode, 3.7 to 0.3 m.y., comprises scattered, monogenetic Strombolian centers of small volume (<1 km3). Rates of volcanic activity for the NTS region are estimated to be about 10−6 event/yr, based on vent counts through time and calculation of rates of magma production. The conditional probability of disruption of the possible waste disposal site at the NTS by basaltic volcanism is bounded by the range of 10−8 to 10−10 yr−1. Consequences, expressed as radiological release levels, were evaluated by assuming disruption of a repository by basaltic magmas fed along narrow dikes. Limits are placed on the volume of waste material incorporated in magma by analogy to the abundance of lithic fragments in basalt scoria and lava. These consequences would be increased if rising magma encountered water and produced magma/water vapor explosions, which can eject large volumes of country rock. Such a mechanism would be important only if the vapor explosions excavated a crater to repository depths (380 m)—an unlikely event, based on the dimensions of hydrovolcanic craters. The total expected release from disruption of a repository by basaltic magma for a 104-yr period is 1.8 Ci for spent fuel and 1.3 Ci for high-level waste.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 247
Active Tectonics: Studies in Geophysics 16 Volcanic Hazard Assessment for Disposal of High-Level Radioactive Waste BRUCE M.CROWE Los Alamos National Laboratory ABSTRACT Volcanic hazard studies for disposal of high-level radioactive waste pose a number of unique problems. These include the long time frame of hazard assessment (104 to 106 yr), the limited geologic record of volcanic activity at disposal sites and the political sensitivity of this national problem. The major variables affecting volcanic hazards are the structure of magma feeder systems at repository depths and the magma fragmentation and dispersal energy of eruptions. The latter is generally controlled by magma composition and the presence or absence of groundwater. Long-lived volcanic fields (>1 m.y.) provide the greatest potential risk for waste disposal, but these can be avoided by proper site selection. Short-lived volcanic fields are more difficult to avoid but are generally mafic in composition, which results in smaller disruption zones and explosive eruptions of lower energy than those of long-lived centers. Volcanic hazards are evaluated through risk assessment, which is a product of probability and consequences. These studies have been completed for a potential waste disposal site in the Nevada Test Site (NTS). Cenozoic volcanism of the NTS region is divided into three distinct episodes. The youngest episode, 3.7 to 0.3 m.y., comprises scattered, monogenetic Strombolian centers of small volume (<1 km3). Rates of volcanic activity for the NTS region are estimated to be about 10−6 event/yr, based on vent counts through time and calculation of rates of magma production. The conditional probability of disruption of the possible waste disposal site at the NTS by basaltic volcanism is bounded by the range of 10−8 to 10−10 yr−1. Consequences, expressed as radiological release levels, were evaluated by assuming disruption of a repository by basaltic magmas fed along narrow dikes. Limits are placed on the volume of waste material incorporated in magma by analogy to the abundance of lithic fragments in basalt scoria and lava. These consequences would be increased if rising magma encountered water and produced magma/water vapor explosions, which can eject large volumes of country rock. Such a mechanism would be important only if the vapor explosions excavated a crater to repository depths (380 m)—an unlikely event, based on the dimensions of hydrovolcanic craters. The total expected release from disruption of a repository by basaltic magma for a 104-yr period is 1.8 Ci for spent fuel and 1.3 Ci for high-level waste.

OCR for page 247
Active Tectonics: Studies in Geophysics INTRODUCTION Regardless of the current or future use of nuclear power reactors for the generation of electricity, one of the major national problems facing scientists across a range of disciplines is the safe disposal of high-level radioactive wastes. (High-level radioactive waste refers to heat-producing waste.) By current definitions, this includes two by-products of nuclear reactors—spent fuel assemblies and reprocessed spent-fuel assemblies. The current consensus of the scientific community is that the most viable means of waste isolation is through deep burial in selected rock formations. This is the present policy of the U.S. Department of Energy. The suitability of a number of sites for storage of high-level radioactive waste is now being actively investigated. Two of the sites that have undergone thorough geologic investigations are the Hanford site in southern Washington (waste burial in the Columbia River Basalt) and the Nevada Test Site (NTS) in southern Nevada (waste burial in ash-flow sheets of the Timber Mountain-Oasis Valley caldera complex). Geologic investigations leading to selection of sites for burial of radioactive wastes pose several problems. First, detailed characterization of a block of the Earth’s crust is required on a level that is unprecedented in the science. This characterization has unique requirements. First, the repository block must be defined with a high degree of confidence, and yet penetrations of the block by exploratory drilling or tunnels must be limited so the integrity of the block is not compromised. This restriction requires an emphasis on geophysical techniques as a primary means of site exploration. Second, the potential effects of construction of a repository tunnel complex and the thermal disturbance of the rock from the emplacement of heat-producing waste must be evaluated with respect to induced changes that could alter the isolation properties of the rocks. Third, the rates of operation of natural geologic processes on the block, such as groundwater movement and tectonic uplift or erosion, must be defined to evaluate the suitability of the block for containment of radioactive waste elements. Finally, possible future changes from naturally occurring but more catastrophic tectonic processes such as seismicity or volcanism must be evaluated for the required isolation period of high-level waste. This last topic, which falls in the regime of predictive geology is the focus of this paper. More and more frequently, geologists are being asked by other scientists and the public to make specific predictions on the future activity of geologic phenomena. Although some progress has been made in prediction of earthquake and volcanic activity, most predictions are valid for periods of months or at most a few tens of years. In contrast, geologic predictions required for waste disposal must encompass a period of thousands of years. We have limited experience in the geologic methods used to make such predictions and even less experience in communicating and defending decisions based on geologic predictions in a public forum. This chapter describes recent work concerned with prediction of tectonic processes for one specific problem—volcanic hazards related to permanent isolation of high-level radioactive wastes. Two topics are emphasized: (1) the special problems posed by volcanic hazard assessment for waste disposal and (2) the use of risk-assessment techniques to evaluate volcanic hazards for the storage of high-level radioactive waste in tuff in southern Nevada (Nevada Nuclear Waste Storage Investigations). SPECIAL PROBLEMS INHERENT IN VOLCANIC HAZARD ASSESSMENT FOR WASTE DISPOSAL SITES Perhaps the most novel aspect of volcanic hazard assessment for radioactive waste disposal is the length of time for which hazards must be forecast. The required containment period of high-level radioactive waste is 104 yr as defined in the draft version of the Environmental Standards for Management and Disposal of Spent Nuclear Fuel, High-Level and Transuranic Radioactive Wastes [U.S. Environmental Protection Agency (1982) 40 CFR 191]. This standard, which has not yet been formally approved, is based on a radiological comparison of the projected mass of radioactive waste [105 metric tons of heavy metal (MTHM)] with an equivalent amount of unmined uranium ore. A 104-yr period has been chosen as the required interval for radioactive waste to decay to a level where the risk is about the same as the smallest estimate of the risk from an equivalent amount of uranium ore (EPA 520/1 82–025, p. 29). However, Bredehoeft et al. (1978) noted that 104 yr provides adequate containment for the short-lived radionuclides such as 90Sr or 137Cs but allows for only a limited reduction in the potential hazards of long-lived radionuclides like 129I. Gera (1982) argued that the waste should be allowed to decay until the radiotoxicity levels are equal to the levels from the amount of uranium consumed in a reactor. This would require an isolation time of about 105 yr. Whatever the required isolation period of high-level waste, this period becomes the minimum length of time for which future volcanic hazards must be forecast. Both 104 and 105 yr are long compared to the rates of operation of volcanic processes. The long time frame of hazard assessment is a special problem of waste disposal. There is neither an estab-

OCR for page 247
Active Tectonics: Studies in Geophysics lished methodology nor scientific experience to draw on to test data conclusions. Because the regulatory guidelines for volcanic hazards are general, it is difficult to decide what types of data best satisfy licensing requirements. This ambiguity has led to increased pressure to quantify the interpretations because numeric data are more easily judged in licensing decisions. However, no matter how the arguments are constructed, these calculations have large uncertainties that are difficult to evaluate within the constraints of the existing general guidelines for site licensing. Moreover, the development of safe methods to permanently isolate radioactive waste is closely linked to the future use of nuclear power reactors, a highly sensitive national issue. Studies dealing with the suitability of a site for waste disposal are subjected to intense public scrutiny. Another problem that applies to most tectonic processes and to volcanism in particular is the relatively limited record used to forecast future activity. Licensing requirements of the U.S. Nuclear Regulatory Commission (document 10 CFR 60) specify that a site under consideration for disposal of high-level radioactive waste must have limited Quaternary igneous activity. Although this is logical for siting considerations, it presents a dilemma for forecasting future rates of volcanism. Either subjective assumptions must be made to estimate numerical rates of volcanism from the limited Quaternary record, or the time scale of the volcanic assessment must be lengthened to include more volcanic events. Neither approach is satisfactory and both increase the possibility that the data do not adequately represent possible future volcanic activity at a repository site. FORECASTING VOLCANIC HAZARDS: NATIONAL PERSPECTIVE Examination of the distribution of volcanic rocks in the United States shows that all Quaternary and, in fact, all major Cenozoic sites of volcanic activity are restricted to the western conterminous United States (for example, Suppe et al., 1975; Smith and Luedke, 1984). Giletti et al. (1978) noted that for the area east of the Rocky Mountains “…the probability for a magmatic incursion into a waste repository during the next million years has been taken as astronomically low, but numerically uncertain.” The obvious conclusion, therefore, is that siting a waste repository east of the Rocky Mountains eliminates volcanism as a licensing issue. However, all current potential waste disposal sites, with the exception of an unspecified site in domal salt in Louisiana or Mississippi, are west of the Rocky Mountains and within the broad region identified by Smith and Luedke (1984) as having the potential for future volcanic eruptions. Crowe (1980) reviewed the major factors controlling the disruption of a repository by volcanism. He noted that the significant variables are the structure of magmatic intrusions or feeder systems at repository depths (>300 m) and the eruptive style of surface activity. The latter is generally controlled by magma composition and the presence or absence of groundwater. Magma feeder systems range from narrow dikes to diapirs; the controlling variables are the viscosity and yield strength of the melt and the rate of heat exchange between the wall rock and the magma. It must be possible for the magma to ascend to the surface before falling temperatures or crystallization prevent eruption (Wilson and Head, 1981). Basaltic magmas ascend along linear dikes because of their low viscosities and Newtonian fluid behavior. Calculated ascent rates based on the presence of high-density nodules, geophysical data, and theoretical considerations are about 1 to 10 cm/sec (Crowe et al., 1983b). In contrast, silicic magmas must ascend as diapirs as a result of their high viscosity. The rate of rise is controlled by the diapir size and the viscosity of the wall rock; ascent velocities are 10−2 to 10−6 cm/sec (Marsh, 1984). Monogenetic volcanic centers are single-cycle volcanoes with eruptive durations of days, months, or years. Magmas associated with these centers are generally of basaltic composition with variations toward andesite. Field studies of the roots of monogenetic centers show that they are fed by narrow dikes (aspect ratios of 10−2 to 10−3), which is consistent with their composition (Crowe et al., 1983b). Eruptive activity in these centers is transitory, and the flux of magma is insufficient to maintain a magma chamber at shallow crustal levels (Fedotov, 1981). The disruption zone associated with monogenetic volcanic centers is limited, therefore, to the feeder dikes. Magma along these dikes must directly intersect a repository to disrupt and disperse waste radionuclides. In contrast, shield volcanoes, stratavolcanoes, or large silicic centers, such as caldera complexes, are long-lived features with multiple eruption cycles and life spans of several million years. Field and geophysical evidence shows that they are fed from shallow-reservoir magma chambers. The presence of shallow magma chambers, multiple feeder systems, and hydrothermal circulation systems above a magma chamber means that the potential effect of repository disruption by these magmas is much greater than that by magmas from monogenetic volcanic centers. The potential effect of eruption mechanisms on disruption and dispersal of the inventory of a waste repository is illustrated in Figure 16.1. This figure, which is

OCR for page 247
Active Tectonics: Studies in Geophysics FIGURE 16.1 Classification diagram for pyroclastic eruptions (after Walker, 1973; Wright et al., 1980). The F is the weight percent of material finer than 1 mm on the axis of dispersal, where it is crossed by the 0.1Tmax isopach. The D is the area inclosed by the 0.01Tmax isopach. used for classification of eruption types based on the properties of the pyroclastic component (Walker, 1973; Wright et al., 1980), can be viewed as a measure of the explosiveness of an eruption. Positive displacement along the x axis represents increased particle dispersal, and height along the y axis represents increased particle fragmentation. Volcanic eruptions of basaltic composition have low dispersal and fragmentation values because of their low water contents (0.1 to 2 wt. %), lower viscosities (102 to 103 P), and Newtonian fluid properties. These eruptions fall into the Hawaiian and Strombolian fields in the lower left part of Figure 16.1. The limited dispersal and fragmentation characteristics of basaltic eruptions result in the smallest potential for disruption effects should a repository be penetrated by basaltic magma. The only important exception is hydrovolcanic activity, when basalt magma mixes at shallow levels with water and is fragmented and dispersed by water/magma vapor explosions. These eruptions have greatly increased particle fragmentation and somewhat increased dispersal and fall within the Surtseyan field of Figure 16.1. In marked contrast, andesitic to silicic volcanic activity spans the sub-Plinian, Plinian, Vulcanian, and Ultraplinian fields of Figure 16.1. These eruptions have orders of magnitude greater dispersal of pyroclastic debris than that of basaltic eruptions. Silicic eruptions have the potential for disrupting the entire area of a waste repository and dispersing waste radionuclides on a global scale (Crowe, 1980). We thus have two end members of volcanic activity with very different potential effects on a buried radioactive waste repository: (1) monogenetic volcanoes of mafic composition, which have limited dispersal and disruptive effects, and (2) long-lived silicic volcanoes, which have large dispersal and disruptive effects. The potential risk from the latter type of volcanic activity can and should be greatly reduced by careful selection of sites for a repository. Large-volume, long-lived volcanic centers are generally confined to distinct volcanic zones. For example, continental arc volcanoes form aligned volcanic chains that are located above active subduction zones. The long-term existence of the arc itself and the host of tectonic features associated with an active subduction zone make arc volcanism relatively easy to recognize and avoid (for example, the Cascade volcanic arc in the Pacific Northwest). Similarly, we now recognize that many if not all of the major silicic volcanic fields of the western United States exclusive of the Cascade chain (for example, the Jemez field in New Mexico, the Yellowstone field in Wyoming, and the Long Valley field in California) occur in distinct volcanic lineaments (Smith and Luedke, 1984). These volcanic zones appear to form from melting of the lower crust within areas of upwelling of mantle-derived basaltic magmas (Smith, 1979;

OCR for page 247
Active Tectonics: Studies in Geophysics Hildreth, 1981). The large flux of basaltic magma required to raise crustal rocks to their melting temperatures and the apparent need for extensional deformation to provide storage space for the basalts mean that sites of silicic magmatism are zones of high heat flow and tectonic activity. Again, these are relatively easy to recognize and avoid in selection of sites for waste storage. The problem of choosing a site for waste disposal in the western United States and minimizing potential effects of future volcanism becomes a matter of determining the risk of future volcanic eruptions in recognized volcanic zones. SITE-SPECIFIC TECHNIQUES OF VOLCANIC HAZARD ASSESSMENT There are two methods for assessing the hazards of volcanism at a site being investigated as a repository for high-level radioactive waste. The first and generally more traditional approach is to study the past record of volcanism at and around a site by using the current tools available to geologic sciences (field mapping, geochronology, petrology, geochemistry, and geophysics). Whereas this approach provides essential data for understanding the past record of volcanism, it does not provide a direct means for deciding whether volcanism represents a significant threat to waste disposal. The second method, which is used increasingly in geological studies for waste disposal, is risk assessment (for example, Crowe and Carr, 1980; D’Alessandro et al., 1980). In this sense, risk is defined as the product of probability and consequences. Probability determinations provide a means of specifying levels of risk as a function of time, and consequences give a perspective for judging acceptable probability levels. In the following sections, the use of risk assessment is described for a potential waste disposal site on and adjacent to the NTS. Volcanic Geology of the Nevada Test Site Region The NTS is located in southcentral Nevada approximately 80 km northwest of Las Vegas (Figure 16.2). The current proposed site for burial of high-level radioactive wastes is Yucca Mountain, a linear range upheld by a thick accumulation of ash-flow and air-fall tuff and associated volcaniclastic rocks derived from the Timber Mountain-Oasis Valley caldera complex (Christiansen et al., 1977). The identified volcanic hazards for the Yucca Mountain site are twofold: the hazards of future silicic volcanism and the hazards of basaltic volcanism. The hazards of silicic volcanism are considered to be negligible for a number of reasons (Crowe et al., 1983a): There has been no silicic volcanism within the NTS region for at least the last 7 million years (m.y.). The youngest silicic volcanic activity (7 to 8 m.y.) was associated with the Black Mountain caldera, which is located approximately 80 km north-north west of Yucca Mountain. There has been a dramatic regional decrease, and in most areas a cessation, of silicic volcanic activity within the southern Great Basin during the last 10 to 20 m.y. (Stewart et al., 1977). Silicic volcanic activity of Quaternary age is restricted entirely to the margins of the Great Basin. The hazards of basaltic volcanism are much more difficult to define and have been the subject of detailed studies (Crowe and Carr, 1980; Vaniman and Crowe, 1981; Crowe et al., 1982, 1983a,b). The NTS region is in a zone of active volcanism referred to as the Death Valley-Pancake Range (DV-PR) volcanic zone. This zone extends from the Lunar Crater volcanic field in central Nevada south-southwestward to southern Nevada and adjoining areas of eastern California (Figure 16.2). The zone merges at its southern end with the Western Cordilleran rift zone of Smith and Luedke (1984). The DV-PR volcanic zone has been active since the cessation of silicic volcanism in the southern Great Basin, about 8 m.y. ago. At that time, basaltic volcanism surpassed silicic volcanism in erupted volume. Centers of active basaltic and minor silicic volcanism shifted progressively to the margins of the southern Great Basin, whereas intermittent activity continued within the DV-PR volcanic zone (Crowe et al., 1983a). Centers of Quaternary volcanic activity in the volcanic zone include the Cinder Hill scoria cone in southern Death Valley (0.7 m.y.), the Lathrop Wells scoria center (0.3 m.y.) located 20 km south of the waste disposal site at Yucca Mountain, four basalt centers (1.2 m.y.) aligned along a north-northeast trending arc in Crater Flat immediately west of Yucca Mountain, two small scoria cones named the Sleeping Butte cones (0.3 m.y.) located 10 km south of Black Mountain caldera, and at least 10 centers in the Lunar Crater volcanic field. Two major types of volcanic fields are present in the DV-PR volcanic zone. Type-I fields include long-lived volcanic fields that are active over a period of several million years. Mafic centers in these fields are clustered, with vent densities of 10−1 to 102 vents/km. Magma volumes of individual centers are approximately 1 to 2 km3, and the total volume of the fields exceeds 10 km3. The range of rock types in type-I fields includes basaltic andesite and trachyte that probably evolved through fractionation from basalt and bimodal basalt-rhyolite as-

OCR for page 247
Active Tectonics: Studies in Geophysics FIGURE 16.2 Generalized geologic map of the Death Valley-Pancake Range volcanic zone. SWM: Stonewall Mountain; BM: Black Mountain; TM-OV: Timber Mountain-Oasis Valley caldera complex; YM: Yucca Mountain exploration block; DV: Death Valley; and CR: Coso Range. semblages. Type-I volcanic fields in the DV-PR volcanic zone include the Greenwater and Black Ranges in southern Death Valley (8 to 4 m.y.), the oldest episode of basaltic volcanism in the NTS region (11 to 8.5 m.y.), the Reveille Range (6 to 5 m.y.), and the Lunar Crater volcanic field (4 m.y. to Recent). Type-II volcanic fields include small-volume (<1 km3) monogenetic centers, which consist of clusters of scoria cones and small-volume lava flows. Vent densities of these fields are low and average about 10−3 to 10−4 vent/km2. Lava compositions are predominantly hawaiite with lesser amounts of alkalic olivine basalt. The hawaiites tend to be of evolved composition (Vaniman et al., 1982) with Mg numbers [Mg/(Mg+Fe2+)] generally <0.55. Parental basalts (Mg number >0.68) are rare. Type-II volcanic fields in the DV-PR volcanic zone include southernmost Death Valley, the NTS region, and Kawich Valley (Figure 16.2). Crowe et al. (1983a) divided the basaltic rocks of the NTS area surrounding the Yucca Mountain site into three episodes; each episode spans several million years and includes basalt from many eruptive centers. The volume relations of these episodes versus time are shown on Figure 16.3. The oldest episode of basaltic volcanism includes the basalts of the silicic episode, which erupted during the waning phase of silicic volcanic activity (11 to 8.5 m.y.). These basalts form bimodal basalt-rhyolite centers that crop out in a northwest-trending zone extending from the south moat zone of the Timber Mountain caldera to Stonewall Mountain (Figure 16.2). Volumes of basaltic magma associated with the basalts of the silicic episode are large, and individual centers exceed 10 km3 in volume. The basalts of the silicic episode were replaced gradationally in time by the older rift basalts. The older rift basalts are either distinctly younger than or spatially separate from the silicic volcanic cen-

OCR for page 247
Active Tectonics: Studies in Geophysics FIGURE 16.3 Plot of magma volume versus time for volcanic episodes of the NTS region. Volumes for the silicic episode are average estimations for a 1 m.y. period. Volumes of the rift basalts are for individual volcanic cycles. The identification of the volcanic hiatus is based on extensive age determinations of volcanic rocks in the NTS region. ters. They consist of small-volume (<1 km3) monogenetic Strombolian centers that erupted during the probable peak of extensional faulting. Magma volumes during this episode declined drastically, then stabilized at low but uniform rates (Figure 16.3). There was a brief but important pause in volcanic activity between 6.5 and 3.7 m.y. This hiatus was followed by eruption of the younger rift basalts. These basalts form widely scattered, small-volume centers identical to those of the older rift basalts. They are separated from the older rift basalts on the basis of their age (3.7 m.y. to Recent) and their enrichment in incompatible trace elements, with the exception of Rb (Vaniman et al., 1982). The younger rift basalts record an interval of declining magma production rates (Figure 16.3). The decline may coincide with a declining rates of tectonism in the NTS region and increased activity in the adjoining southwestern areas of the Great Basin (Death Valley and Owens Valley, Figure 16.2). Figure 16.4 is a geologic map of the Lathrop Wells center. This center illustrates a number of characteristic features of Strombolian centers of both the younger and older rift basalts. Each center consists of a main scoria cone flanked by several smaller satellite cones. The satellite cones are generally older and are located south or southeast of the main scoria cone. This suggests a north to northeastward migration of active vents during an eruption cycle. The number of vents per center averages 2 to 3 for the Quaternary centers of the NTS region (Crowe et al., 1983b). Blocky aa lava flows vented from the flanks of the main scoria cone and traveled short distances. Measured flow lengths of Quaternary flows in the NTS region range from 0.6 to 1.9 km with a mean length of 1.1 km. It is inferred that scoria fall sheets extended downwind from the main scoria cone. These are entirely removed by erosion at all but the youngest centers in the NTS region. The only exception is the 270,000-yr-old scoria fall sheet of the Lathrop Wells center, and only minor remnants of this sheet are preserved. The cumulative volume of individual centers in the NTS region (dense rock equivalent) was calculated assuming a scoria fall sheet-to-cone ratio of 5:1. Magma volumes for the Quaternary cones of the NTS region range from 105 to 108 m3 with a mean of 3×107 (Crowe et al., 1983b, p. 269). PROBABILITY AND CONSEQUENCE ASSESSMENT Probability Several aspects of the history of basaltic volcanism in the NTS region provide justification for a probabilistic approach to volcanic hazards. First, there has been a consistency or decline in the rates of basaltic volcanism for the past 8.5 m.y. Second, all basalt centers formed during the past 8.5 m.y. are small-volume Strombolian centers. Third, the petrology of the erupted basalts is broadly similar throughout this period, which is indicative of similar conditions of magma genesis. The consistency in rates, eruptive style, and petrology for this extended period provides a basis for forecasting rates of future volcanism for the required containment period of high-level radioactive waste. The probability of disruption of a repository by basal-

OCR for page 247
Active Tectonics: Studies in Geophysics FIGURE 16.4 Geologic map of the Lathrop Wells basalt center located 20 km south of Yucca Mountain. From Vaniman and Crowe (1981). tic magma can be formulated as a case of conditional probability (16.1) where Pr is the probability of repository disruption, E1 is the rate of occurrence of volcanic events, and E2 is the probability of intersection of a repository by magma, given E1. This probability can be expressed as (Crowe et al., 1982) (16.2) where λ is the rate of volcanic activity and p is the probability that an event is disruptive. This formulation assumes that the time between events is exponentially distributed, and the number of events in time intervals of length t has a Poisson distribution with a mean of λt. The p can be estimated as the ratio a/A, where a is the area of a repository or of an assigned volcanic disruption zone (whichever is larger) and A is some minimal area that encloses the repository and the volcanic events used to describe λ. For the case of basaltic volcanic activity, a is the area of the repository. It is estimated to be about 6 to 8 km2, based on the size of the exploration block at Yucca Mountain (Crowe et al., 1982). The A can be selected in a number of ways. Crowe and Carr (1980) defined A as the area enclosed by circles of 25- and 50-km radii with centers at the Yucca Mountain exploration block. This approach, although computationally simple, does not consider the spatial trends in the distribution of volcanic centers that are controlled by tectonic features. Crowe et al. (1982) developed a numerical approach to find the minimum area circle and the minimum area ellipse that contain the volcanic centers of interest and the repository site. Differing combinations of ellipses and circles were obtained using geologic assumptions concerning the tectonic setting of volcanic centers. Disruption probability values are calculated and tabulated in table form (Crowe et al., 1982), which allows a range of values p to be used in Eq. (16.2). The preferred definition of p is based on the distribution of volcanic centers in the DV-PR volcanic zone. The most difficult parameter to estimate for probability calculations is λ, the rate of volcanic activity. Ideally, rates should be determined from an understanding of the controlling processes of volcanic activity. Although this is not now possible, a number of studies suggest that rates of volcanism are related in a direct but as yet poorly understood way to the regional stress-strain pattern. That is, rates of volcanism, similar to regional

OCR for page 247
Active Tectonics: Studies in Geophysics seismicity, reflect the degree of tectonic activity of a region. For example, Shaw (1980) related rates of magma production at Hawaii to seismic moment—a measure of the volume change represented by the release of seismic energy. He assumed that the potential energy available in a tectonic system may be released through both seismic radiation and the generation and ascent of magma. The demonstration of a connection between the release of seismic energy per year and the annual rate of magma production suggests that the two tectonic processes are closely related and that, by inference, magmatism may be an important process in regulating mantle stress levels (Shaw, 1980, p. 259). Similarly, Feigensen and Spera (1981) recognized a correlation between the occurrence of tholeiitic and incompatible-element-enriched alkalic magmas and eruption frequency in Hawaiian lavas. They related magma production to shear-stress deformation (viscous deformation of Shaw, 1973) and suggested that a reduction in shear stress caused a decrease in the degree of partial melting and an increase in the time between eruptions. Bacon (1982) documented a time-predictable relation among the basalts and high-SiO2 rhyolites of the Coso volcanic field. He suggested that the magma production rate for the field is controlled by a constant strain rate. A minimum strain value equal to the failure strength of the crustal rocks must be exceeded to allow ascent of magma. These studies, although promising, are not yet sufficiently developed to allow estimates of volcanism rates, particularly for relatively inactive regions such as the NTS. We thus are forced to forecast future rates of volcanic activity on the basis of past activity. Crowe et al. (1982) estimated rates of volcanic activity in two ways: through counts of age-controlled volcanic centers and through variations in magma volume versus time. The first technique involves field mapping vent zones for Quaternary volcanic centers. Each vent is treated as one volcanic event, and time control is provided by the K-Ar age and the magnetic polarity of the volcanic center. A somewhat similar approach, with less rigorous age controls, was used to determine the probability of faulting for a potential waste disposal site in Italy (D’Alessandro et al., 1980). Estimated rates of volcanic activity are calculated by counting vents within the areas defined in the determination of A. These rates during Quaternary time (<1.8 m.y.) are about 8×10−6 volcanic events per year for the NTS region, 7×10−5 volcanic events per year for the DV-PR volcanic zone, and 8×10−5 volcanic events per year for the southern Great Basin (Crowe et al., 1982). The latter two calculations are strongly effected by the high cone density of the Lunar Crater volcanic field at the northern end of the DV-PR volcanic zone. The calculation of volcanic rates by using rates of magma production is based on Figure 16.5, an expanded version of Figure 16.3. Figure 16.5 shows magma volume (dense rock equivalent) versus time for the past 3.7 m.y. The basaltic eruptions show a linear decrease in magma volume with time. Regression analysis, treating magma volume as the dependent variable, gives a coefficient of determination of 0.8 and a rate of magmatic production of 210 m3/yr. This magma production rate is used to calculate the times required to generate magma volumes that are equal to representative volumes of past magmatic cycles in the NTS region. These times, corrected for the time since the last basaltic FIGURE 16.5 Plot of magma volume (dense rock equivalent) versus time for the younger rift basalts. Error bars for time are the standard deviation (1) of replicate age determinations. Error bars for the volume are estimates of errors in volume calculations. From Crowe et al. (1982).

OCR for page 247
Active Tectonics: Studies in Geophysics eruption (2.7×105 yr) give predicted occurrence times or rates of future activity (Crowe et al., 1982). The conditional probability for the combined occurrence of a future basaltic event and that event to directly intersecting a buried radioactive waste repository at Yucca Mountain is calculated using Eq. (16.2) and the data tables for λ and p. Because of the large uncertainties in the probability calculations, little emphasis is placed on individual calculations. Rather, a range of probability values is assembled by using a matrix of area ratios and rate determinations. Probability bounds are established by the extremes of the calculated probabilities. These bounds are 5×10−8 to 3×10−10 yr−1 for the Yucca Mountain site (Crowe et al., 1982). Consequence Consequence analysis involves identification of volcanic processes that could lead to failure of a waste-isolation system and calculation of the results of failure expressed as radiological levels of released waste elements. Crowe et al. (1983b) suggested that the most likely event to affect a buried repository at Yucca Mountain would be intrusion of the repository by rising basaltic magma, followed by Strombolian eruptions of waste-contaminated magmas. They argued, based on field evidence, that the most likely intrusion structure for basalt at repository depths (>300 m) would be linear dikes, that there would be 2 to 3 dikes per volcanic event, and that the dike widths and lengths would be 0.3 to 4 m and 0.5 to 5 km, respectively. Using these background parameters, the important questions become: (1) How much waste material would be incorporated in the basalt magma, and (2) how would the waste material be dispersed with the magma at the surface? The mechanisms of intrusion of basalt magma into a waste repository are difficult to predict. Magma could completely flood a repository tunnel or intrude as narrow dikes with the only magma/waste contact being along the dike walls. In either case, the number of waste canisters that would be contacted by magma and potentially carried to the surface ranges from 0 to more than 400 (Logan et al., 1982). It is possible to establish limits on the volume of waste material incorporated in basalt magma by assuming that the repository is intersected by a dike of random orientation and tracing the pathways of waste incorporated in magma through analogy with the distribution of lithic (country-rock) fragments in basalt scoria. Studies of the distribution of these fragments in basaltic rocks revealed that they are derived from shallow levels and are concentrated in the pyroclastic component (Crowe et al., 1983b). This latter conclusion is based on the fact that particles in a magma exsolving volatiles are preferential sites of bubble nucleation (Sparks, 1978) and that lithic fragments are present in small amounts in cone scoria but are extremely rare in lava. Two mechanisms are proposed for the incorporation of lithic fragments in basaltic magma: (1) during magma fragmentation when the gas-to-magma ratio reaches 0.75 and the magma disrupts (Wilson and Head, 1981) and (2) during hydrovolcanic explosions associated with a predominantly Strombolian eruptive cycle. Assuming that waste particles are dispersed in an eruption in the same pattern as country-rock lithic fragments would be, Crowe et al. (1983b) calculated that 600 to 1100 m3 of a repository inventory would be dispersed in a pyroclastic eruption. Of that inventory, 15 to 20 percent will be deposited in the scoria cone (less than 1 percent will be exposed at the surface during a time period of 104 yr), 50 to 80 percent will be deposited in the scoria fall sheet, and 2 to 5 percent will be regionally dispersed with the fine-grained particle component (windborne). These calculations are based on an average lithic fragment abundance of 0.032 percent by volume in studied scoria deposits of the NTS region and the San Francisco volcanic field in northern Arizona (Crowe et al., 1983b). An additional eruptive process in basaltic eruptions is hydrovolcanic vapor explosions. This type of eruption occurred at three basalt centers in the NTS region during the past 10 m.y. The increased dispersal distance and the greater particle fragmentation of hydrovolcanic explosions, as noted previously, could greatly increase the consequences of repository disruption. Crowe and Carr (1980) originally argued that hydrovolcanic activity was unlikely at Yucca Mountain. This was based on the depth to the groundwater table at Yucca Mountain (>300 m below the surface) and the lack of surface water as a result of the arid climate and steep drainage gradients. Moreover, the moisture content of the unsaturated zone at Yucca Mountain is insufficient to allow initiation or maintenance of hydrovolcanic explosions except under extreme saturation conditions (Crowe et al., 1983b). New data on hydrovolcanic activity have caused re-examination of this question. Although field evidence indicates that lithostatic load has an inhibiting effect on the formation of hydrovolcanic activity (probably because the vapor phase of water is suppressed), this may not always be the case. In the geologic setting of Yucca Mountain, groundwater may mix in a supercritical state with magma below the repository level. As the magma/water mix ascends and the pressure drops, vapor explosions may occur throughout an interval extending from the water table through the repository horizon to the surface. Studies of fuel-coolant interactions and experimental work using thermite melt mixed with

OCR for page 247
Active Tectonics: Studies in Geophysics water to simulate volcanic eruptions (Wohletz and McQueen, 1984) show that the major control of hydrovolcanic explosions is the mass ratio of water to magma. Eruptions may vary between Surtseyan and Strombolian, depending on this ratio. The optimum ratio for development of Surtseyan explosions is about 0.3 (Wohletz and McQueen, 1984). The occurrence of lithic fragments in basalt scoria of the Lathrop Wells center may be due, as noted above, to hydrovolcanic explosions. Such explosions would require a water-to-magma ratio of less than about 0.3 so that the eruptions would have been predominantly Strombolian with only a minor hydrovolcanic component. This is suggested by the irregular distribution of fragments in scoria (intermittent explosions) and the fact the initial eruptions of the center were hydrovolcanic (Vaniman and Crowe, 1981). If it is assumed that the fragments were produced by a hydrovolcanic component, possible effects of hydrovolcanic explosions were included in the calculations of lithic fragment abundance. In this case, the consequences of a hydrovolcanic component can be shown to be small (Logan et al., 1982). A more extreme case, having major negative consequences for waste isolation, would be the possibility of exhumation of a buried repository during crater-forming hydrovolcanic explosions. This possibility is tested by data summarized in Figure 16.6, which is a relative-frequency diagram of crater depth for hydrovolcanic centers (maars and tuff rings) based on the data of Pike and Clow (1981). The crater-depth data are positively skewed, and the mean depth is 91±67 m. The average depth to the repository horizon at Yucca Mountain is about 380 m (the depth varies throughout the exploration block because of the 6 to 8° eastward dip of the currently favored repository horizon). A depth of 380 m is greater than 4 standard deviations from the mean depth of hydrovolcanic craters and exceeds the maximum listed crater depth in the data catalog of Pike and Clow (1981). The low probability of repository disruption (10−8 to 10−10 yr−1), coupled with the low probability of exhumation of a repository by explosive cratering, suggests that the risk of hydrovolcanic explosions is of limited concern at Yucca Mountain. The radiologic consequences of basaltic volcanism have been calculated for the disruption of two hypothetical repositories, both located at Yucca Mountain (Logan et al., 1982). The first contained unreprocessed spent fuel and the second reprocessed waste. The calculations assumed a repository storage capacity of half the estimated volume of spent fuel or reprocessed waste generated by commercial power reactors through the year 2000. It was estimated that the repository area would be 1640 m2 and would contain waste aged for 10 yr at the time of emplacement. A range of geometric arguments is used to determine the number of canisters that would be intersected by a linear basalt dike. The preferred approach was to assume a random orientation of the dikes FIGURE 16.6 Relative-frequency diagram of crater depth for hydrovolcanic craters. The data are positively skewed with a mean of 91 m. Repository depth is the estimated average depth to the repository horizon at Yucca Mountain. Maximum crater depth is the deepest measured crater. Data on crater depths are from Pike and Clow (1981).

OCR for page 247
Active Tectonics: Studies in Geophysics encountering the repository tunnel complex. Accordingly, magma would intersect about seven canisters of spent fuel and about one cannister of high-level waste—equivalent to an inventory fraction in both cases of about 8×10−5 (Logan et al., 1982, p. 32). There are no data on the behavior of a waste canister in rising and vesiculating magma. Therefore, it was assumed that all waste contacted by magma would be incorporated uniformly in the magma and carried to the surface to be erupted preferentially with the pyroclastic component. Radiological dose was examined in two forms: the dose resulting from the waste entrained in the basalt eruption column and the dose resulting from exposure to waste-bearing scoria deposits, resuspended particles, and radionuclides entering the food chain. Population, agricultural, and meteorological data from Yucca Mountain were used to calculate doses from the airborne component. The worst case for the airborne dose resulted in a cumulative total of 25 health effects in 106 yr (Logan et al., 1982, p. 6). Dose effects were calculated from scoria deposits that formed the scoria cone and the scoria fall deposits downwind of a hypothetical eruption and include effects of erosional transport during a period of 106 yr. The worst case here yields about 2800 health effects in 106 yr. Finally, Logan et al. (1982) calculated the total activity that would be transported to the surface, normalized to 103 MTHM. The values vary for the type of waste and the time of eruption relative to emplacement and closure of the repository. Stepwise integration of the release fraction by radionuclide with the annual probability of volcanic disruption of a repository using the computer code AMRAW (Logan and Berbano, 1978) gives total expected releases as a function of time. The total expected release for 104 yr is 1.8 Ci or 0.038 Ci/ 103 MTHM for spent fuel and 1.3 Ci or 0.034 Ci/103 MTHM for high-level waste (Logan et al., 1982, p. 163). DISCUSSION A variety of evidence shows that there is a finite risk of volcanism with respect to storage of radioactive waste at Yucca Mountain. In particular, the site is located in a zone of active volcanism—the Death Valley-Pancake Range volcanic zone and five Quaternary volcanic centers (1.2 to 0.3 m.y.) are present within a range of 8 to 20 km of the exploration block. Geologic data show that there has been a consistency in the rates, eruptive style, and petrology of basaltic volcanism in the NTS region for the past 8 m.y. Volcanic activity in the region was characterized by formation of scattered small-volume Strombolian scoria cones and lava flows. Rates of volcanism have been low, and there is some evidence of a decline in the rate of magma production during the past 3.7 m.y. The consistency in the geologic record provides some degree of confidence that future volcanic patterns can be projected for periods of 104 to 105 yr. These circumstances provide a reasonable basis for applying risk-assessment techniques to define the magnitude of volcanic hazards. Probability estimates of the likelihood of the combined occurrence of a future volcanic event and the likelihood that the volcanic event will intersect a buried repository at Yucca Mountain are very low—10−8 to 10−10 per year. The consequences of penetration of a repository by basaltic magma followed by eruption of the magma at the surface are limited, perhaps surprisingly limited. This is due to the small subsurface area of basalt feeder dikes, the probable limited intrusion effects of these dikes, and the limited surface dispersal of basaltic particulates in Strombolian eruptions. The combination of the low probability and limited consequences indicate that the risk of volcanism at this particular site is low. The suitability or unsuitability of the Yucca Mountain site will be decided by the U.S. Nuclear Regulatory Commission if the Department of Energy recommends the site as a formal candidate for a waste repository. Certainly many other criteria other than volcanism will be considered in any licensing decisions; this paper emphasizes the methods followed to define the magnitude of volcanic hazards. In the Nevada case, geologic, geochronologic, petrologic, geochemical, and geophysical data are used in combination with techniques of risk assessment to evaluate volcanic hazards. Some but not all of these methods may be useful for hazard assessment at other sites under consideration for disposal of high-level radioactive waste. How reliable are the geologic data used for risk assessment? Certainly the field, geophysical, and laboratory data have been gathered as carefully as possible, and the reliability of this work must be judged by research standards established in the current geologic literature. Research data have been evaluated at review meetings where hazard data were presented to a panel of scientists. The volcanic hazard work has also been published in the scientific literature to provide a wide exposure to the scientific community. An additional item of concern is the built-in bias in the work. That is, the purpose of the Nevada Nuclear Waste Storage Investigations is to attempt to find and prove a site for disposal of radioactive waste. As studies of a site progress and increased funding is invested, there is mounting pressure to “prove” the site. The presentation of both positive and negative evidence of the hazards of volcanism (Crowe et al., 1983b) is an attempt to overcome this bias. There is a continuing controversy over the use of

OCR for page 247
Active Tectonics: Studies in Geophysics probabilistic studies in geology. Indeed, the special problems of hazard prediction in waste disposal have brought this question to the forefront (for example, Marsily and Merriam, 1982). Questions have been raised whether geologic process are truly random and suitable for probabilistic analyses (D’Alessandro and Bonne, 1982). Certainly, with respect to the Nevada work, the data used for probability calculations are limited at best. The number of data points for rate calculations are small, and it is difficult to include effects of tectonic setting in localizing sites of volcanism. What is the time sensitivity of the data? For the Nevada site, future rates of volcanism are forecast on the basis of the geologic history of the region during the past 3.7 m.y. The isolation period of radioactive waste is 3 to 0.3 percent of this time, and, thus, it is not clear how sensitive this approach is to short-term variations in tectonic processes. Because of these data uncertainties, the probability is presented as a range spanning 3 orders of magnitude and the data assumptions are biased toward the worst case to provide conservative estimates (Crowe et al., 1982). Similar questions are raised about the results of consequence analyses. The detailed effects of intrusion of a repository by volcanism are not clear. Variations in the strike of linear dikes intersecting a repository result in differences in the number of contacted cannisters, ranging from 0 to more than 400 (Logan et al., 1982). Equally, we can only speculate how waste could be incorporated in a magma. It may be maintained in its cladding, may be fragmented as a particulate, or may be geochemically dispersed in the magma. Any of these possibilities results in major changes in the calculated radiological dose levels through time. ACKNOWLEDGMENTS Much of the work on the volcanic hazards of the NTS site was completed with Will Carr (U. S. Geological Survey) and Dave Vaniman (Los Alamos National Laboratory); probability studies were completed with Mark Johnson and Richard Beckman (Los Alamos National Laboratory). I have been greatly aided by their cooperation and numerous discussions. The manuscript was reviewed critically by B.D.Marsh, G.D.Robinson, D.T.Vaniman, and K.H.Wohletz. Funding for the NTS work was provided by the Nevada Nuclear Waste Storage Investigations and the Office of Basic Energy Sciences through the U.S. Department of Energy. REFERENCES Bacon, C.R. (1982). Time-predictable bimodal volcanism in the Coso Range, California, Geology 10, 65–69. Bredehoeft, J.D., W.W.England, D.B.Stewart, N.J.Trask, and I.J.Winograd (1978). Geologic disposal of high-level radioactive wastes—earth science perspective, U.S. Geol. Surv. Circ. 779. Christiansen, R.L., P.W.Lipman, W.J.Carr, F.M.Byers, Jr., P.P.Orkild, and K.A.Sargent, (1977). The Timber Mountain-Oasis Valley caldera complex of southern Nevada, Geol. Soc. Am. Bull. 88, 943–959. Crowe, B.M. (1980). Disruptive event analysis: Volcanism and igneous intrusion, Battelle Pacific Northwest Laboratory Rep. PNL-2882. Crowe, B.M., and W.J.Carr (1980). Preliminary assessment of the risk of volcanism at a proposed nuclear waste repository in the southern Great Basin: U.S. Geol. Surv. Open-File Rep. 80–375, 15 pp. Crowe, B.M., M.E.Johnson, and R.J.Beckman (1982). Calculation of the probability of volcanic disruption of a high-level radioactive waste repository within southern Nevada, USA, Radioactive Waste Management 3, 167–190. Crowe, B.M., D.T.Vaniman, and W.J.Carr (1983a). Status of volcanic hazard studies for the Nevada Nuclear Waste Storage Investigations, Los Alamos National Laboratory Rep. LA-9325-MS, 47 pp. Crowe, B.M., S.Self, D.Vaniman, R.Amos, and F.Perry (1983b). Aspects of potential magmatic disruption of a high-level radioactive waste repository in southern Nevada, J. Geol. 91, 259–276. D’Alessandro, M., and A.Bonne (1982). Fault-tree analysis for probabilistic assessment of radioactive-waste segregation: An application to a plastic clay formation at a specific site, in Predictive Geology with Emphasis on Nuclear-Waste Disposal, G.Marsily and D.F. Merriam, eds., Pergamon Press, New York. D’Alessandro, M., M.Murray, C.N.Bertozzi, and G.F.Girardi (1980). Probability analysis of geologic processes: A useful tool for the safety assessment of radioactive waste disposal, Radioactive Waste Management 1, 25–42. Environmental Protection Agency (1982). Draft Environmental Impact Statement 40 CFR Part 191: Environmental Standards for Management and Disposal of Spent Nuclear Fuel, High Level and Transuranic Radioactive Wastes, 228 pp. Fedotov, S.A. (1981). Magma rates in feeding conduits of differing volcanic centers, J. Volcanol. Geotherm. Res. 9, 379–394. Feigensen, M.D., and F.J.Spera (1981). Dynamic model for temporal variation in magma type and eruption interval at Kohala volcano, Hawaii, Geology 9, 531–533. Gera, F. (1982). Geologic predictions and radioactive waste disposal: A time limit for the predicted requirements, in Predictive Geology with Emphasis on Nuclear Waste Disposal, G.Marsily and D.F. Merriam, eds., Pergamon Press, New York. Giletti, B., R.Siever, J.Handin, J.Lyons, and G.Pinder (1978). State of geological knowledge regarding potential transport of high-level radioactive waste from deep continental repositories, Environmental Protection Agency Rep. 520/4–78–004, 53 pp. Hildreth W. (1981). Gradients in silicic magma chambers: implications for lithospheric magmatism, J. Geophys. Res. 86, 10153–10192. Logan, S.E., and M.C.Berbano (1978). Development and application of risk assessment method for radioactive waste management, Environmental Protection Agency Rep. 520/6–78–005. Logan, S.E., R.Link, H.S.Ng, F.A.Rockenbach, and K.J.Hong (1982). Parametric studies of radiological consequences of basaltic volcanism, Sandia National Laboratories Rep. SAND81–2375, 219 pp. Marsh, B.D. (1984). Mechanics and energetics of magma formation and ascension, in Explosive Volcanism: Inception, Evolution, and

OCR for page 247
Active Tectonics: Studies in Geophysics Hazards, Geophysics Study Committee, National Research Council, National Academy Press, Washington, D.C., pp. 67–83. Marsily, G., and D.F.Merriam, eds. (1982). Predictive Geology with Emphasis on Nuclear Waste Disposal, Pergamon Press, New York, 206 pp. Pike, R.J., and G.D.Clow (1981). Revised classification of terrestrial volcanoes and catalog of topographic dimensions, with new results on edifice volume, U.S. Geol. Surv. Open-File Rep. 81–1038, 40 pp. Shaw, H.R. (1973). Mantle convection and volcanic periodicity in the Pacific: Evidence from Hawaii, Geol. Soc. Am. Bull. 84, 1505–1526. Shaw, H.R. (1980). The fracture mechanisms of magma transport from the mantle to the surface, in Physics of Magmatic Processes,R. B.Hargraves, ed., Princeton University Press, pp. 201–264. Smith, R.L. (1979). Ash-flow magmatism, in Ash-Flow Tuffs, C.E. Chapin and W.E.Elston, eds., Geol. Soc. Am. Spec. Paper 180, pp. 5–27. Smith, R.L., and R.G.Luedke (1984). Potentially active volcanic lineaments and loci in western conterminous United States, in Explosive Volcanism: Inception, Evolution and Hazards, Geophysics Study Committee, National Research Council, National Academy Press, Washington, D.C., pp. 47–66. Sparks, R.S.J. (1978). The dynamics of bubble formation and growth in magmas: A review and analysis, J. Volcanol. Geotherm. Res. 3, 1–37. Stewart, J.H., W.Moore, and I.Zietz (1977). East-west patterns of Cenozoic igneous rocks, aeromagnetic anomalies, and mineral deposits, Nevada and Utah, Geol Soc. Am. Bull. 88, 67–77. Suppe, J., C.Powell, and R.Berry (1975). Regional topography, seismicity, Quaternary volcanism, and the present-day tectonics of the western United States, Am. J. Sci. 275-A, 397–436. Vaniman, D., and B.M.Crowe (1981). Geology and petrology of the basalts of Crater Flat: Applications to volcanic risk assessment for the Nevada Nuclear Waste Storage Investigations, Los Alamos National Laboratory Rep. LA-8845-MS. Vaniman, D., B.M.Crowe, and E.S.Gladney (1982). Petrology and geochemistry of hawaiite lavas from Crater Flat, Nevada, Contrib. Mineral. Petrol. 80, 341–357. Walker, G.P.L. (1973). Explosive volcanic eruption—A new classification scheme, Geol. Rund. 62, 431–466. Wilson, L., and J.W.Head (1981). Ascent and eruption of basaltic magma on the Earth and Moon, J. Geophys. Res. 86, 2971–3001. Wohletz, K.H., and R.G.McQueen (1984). Experimental studies of hydromagmatic volcanism, in Explosive Volcanism: Inception, Evolution, and Hazards, Geophysics Study Committee, National Research Council, National Academy Press, Washington, D.C., pp. 158–169. Wright, J.V., A.L.Smith, and S.Self (1980). A working terminology of pyroclastic deposits, J. Volcanol. Geotherm. Res. 8, 315–336.