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.
Overview and Recommendations A burst of earthquake tremors with associated ground deformation and fumarolic activity at Mammoth Lakes in the Long Valley caldera, California, prompted the U.S. Geological Survey (USGS) to issue a notice of a potential volcanic hazard on May 25, 1982. Mam- moth Lakes is a major outdoor recreation area that at times accommodates 30,000 or more vacationers. The focus of the earthquake activity is about 3 km from the center of population and at the time of the alert was about 3 km below the surface. The USGS notice (1982) states that "neither the probability nor the nature, scale, or timing of a possible eruption can be determined as yet." Field studies in the Long Valley area suggest that such an eruption might take a variety of forms but would most likely include explosions of groundwater superheated by contact with magma, ash falls, and minor pyroclastic flows, terminating with the extrusion of a rhyolite dome. More harrowing is the possibility of eruption of a major pyroclastic flow of the magnitude of the Bishop Tuff, erupted from Long Valley 700,000 years ago. Such an eruption could devastate an area of hundreds of square miles, cause great loss of life, and possibly interrupt the water supply from the Owens Valley aqueduct to Los Angeles. This situation is analogous to the scene at Mount St. Helens immediately following the onset of seismic activity on March 20, 1980. That the eruption of Mount St. Helens could be forecast in considerable detail is a remarkable achievement. That the prediction did not include more than the possibility of a lateral blast at Mount St. Helens, however, and the fact that considerable uncertainty surrounds an expectation of events at Mam- moth Lakes, is a symptom of our still imperfect understanding of volcanic processes. The eruption at Mount St. Helens has profoundly heightened public awareness of volcanism. Such awareness has long existed among the residents of Hawaii and Alaska. Today those who live in the conterminous United States no longer think of volcanic eruptions as being remote in time or space. Curiousity about volcanic events and a desire to minimize loss of life in future eruptions have stimulated research. The intel- lectual and social climate for fostering advances in volcanology is excellent. Some ave- nues for research that appear especially promising are discussed in this volume. Although explosive eruptions are usually associated with andesitic volcanoes, they can occur in any volcanic system. The energy for explosive eruptions is commonly
Overview and Recommendations supplied by expansion of volatiles exsolved from magma; the eruption of magmas that are relatively poor in volatiles is normally not accompanied by violent explosions. Nevertheless, interactions between magmas and groundwater, lakewater, or seawater also can produce explosive eruptions. About 1 percent of the eruptions at Kilauea Volcano in Hawaii have been explosive, with attendant loss of life, although this volcano is noted for relatively gentle, effusive outpouring of lava (see Chapter 9). The problem of explosive volcanism is thus neither separate nor distinct from the more general problem of volcanism, its causes, manifestations, and effects on society. A discussion of the process of volcanism must begin with the generation of magma. Magma-forming processes are intimately connected with tectonics since pronounced differences exist in the nature of volcanism between regions of compression and sub- duction and regions of rifting or broader extension. The causes of these differences are incompletely understood. The cyclic or episodic nature of volcanic activity forms a basis for predicting eruptions and can be a source of information on the rates at which magma and energy are introduced to volcanic systems. Improved understanding of the physics of volcanic eruptions is an exciting goal and is vital to progress in hazard evaluation. Development of the physics of planetary volcanism with reference to the Moon, Mars, Venus, and especially Jupiter's moon, lo, has added a new dimension to research in this field. The study of explosive volcanism also must include an appreciation of the severe social problems that are caused by erupting volcanoes. None is of greater urgency than planning for a crisis. Progress in research on these aspects of explosive volcanism is sketched in this Overview and discussed in detail in the chapters that follow. TECTONISM AND VOLCANISM A better understanding of the interrelationships of tectonism and volcanism is funda- mental to progress in understanding the chemical aspects of magmatic processes. These interrelations are especially complex and varied in western North America. During the Tertiary a compressional subduction system was partly replaced by extensional tecton- ism. Broad changes in the compositions of erupted magmas from those that include andesites to a combination of basalt with associated large rhyolitic magma bodies oc- curred, with change in the style of extension. Remnants of the compressional system include a subducted plate, perhaps containing a giganic hole, that underlies the Basin and Range and an active subduction system in the Pacific Northwest that underlies the Cascades. The portion of the Juan de Fuca plate beneath the Cascades has recently been shown to be a source of earthquakes whose foci delineate a subplanar zone that dips about 60° between the depths of 40 and 70 km (Crosson, 1983). Studies of the Andean subduction system have shown major variations in the sub- duction angle both down dip and along strike (see Chapter 3). Beneath central Peru the initial dip of the Nasca plate is a normal 30°, but at a depth of about 100 km it flattens for 300 km beneath the continent before subducting again at a steeper angle. There is no volcanism in the area where there is no asthenosphere between the sub- ducted plate and the overlying, rigid continental lithosphere. Equivalent complexities may well be present beneath western North America. Tectonism influences magma generation through convective flow in the mantle and may influence the eruption of magma through fracturing in the lithosphere and crust. Magma erupts from vents that form a well-defined arc in the subduction system of the Cascades. In contrast, volcanic loci form a complex network of rectilinear zones in the Basin and Range, where extensional tectonics prevail (see Chapter 4). Many of these loci have migrated with time over the past 15 million years (m.y.), perhaps in response to differential movements of the crust and mantle. The migrations are to the east and 4
Overview and Recommendations northeast in the area that lies east of central Nevada and Idaho and to the west in California and western Oregon. The tectonic setting of a magma's ascent from a mantle source to the surface influences whether it rises directly or is stored temporarily at intermediate depth. The nature of magmatic movements will in turn control the opportunities for chemical interaction between melts and rocks forming conduit walls and the possible development of sec- ondary, more silicic melts. Some alkali basalts contain abundant inclusions of mantle rocks, indicating rapid transport from mantle sources with little opportunity for inter- action with wall rocks. The mantle sources for alkali basalts may, however, have been metasomatically altered during or prior to partial fusion (see Chapter 1). Some basalts, such as those in the Columbia Plateau, have chemical and isotopic compositions that are commensurate with assimilation of crustal rocks (see Chapter 2). Major eruptions of rhyolite in areas such as Yellowstone are believed to be due to melting of deep crustal rocks by basaltic magma, and in those circumstances the basalts may have acted more as thermal than chemical parents. Formation of calderas after major eruptions of rhyolite is clear evidence that the magma chambers were at a relatively shallow depth. Seismic data can be interpreted to show that magma is present beneath some parts of Yellowstone and that the top of a low-velocity body, 2 to 5 km deep, may represent a zone of partial melt or a large vapor-dominated hydrothermal system (see Chapter 7). Yellowstone rhyolites are more likely to have been formed by partial fusion of high-grade metamorphic rocks of the lower crust than of once-sedimentary, upper crustal rocks (see Chapter 6). The formation of rhyolite in large volume probably involves a combination of deep and shallow magma chambers. The absence of well-defined collapse features associated with flood basalt eruptions suggests that storage of these lavas, possibly accompanied by assimilation of wall rocks, was deeper than is characteristic for near-surface rhyolite chambers. Eruptive configurations that maximize opportunities for chemical interaction between magma and wall rocks or that maximize opportunity for generation of anatectic magmas also maximize heat loss. Such configurations might include blocks broken from conduit walls and interlacing networks of dikes. Freezing or block choking in such systems could be inhibited by episodic injection of fresh magma. Basaltic magma may produce up to an equal volume of rhyolite melt in the course of crystallization of the basalt (see Chapter 5). But how do basaltic magmas contaminated by interaction with crustal rocks (e.g., Columbia Plateau basalts) erupt as phenocryst-free lava? Active seismic studies in the Yellowstone region outline another puzzle involving magma ascent. The lower crust beneath Yellowstone appears to be homogeneous and similar to thermally undis- turbed, surrounding crust (see Chapter 7), whereas one might expect a complex eruptive plumbing system at high temperature. Yellowstone is the only major rhyolite caldera system in North America that has been examined in detail by active seismic methods. VOLCANIC PERIODICITY The periodic nature of some volcanic activity is an important characteristic for both theoretical studies and hazard assessments. The interval since the last eruption clearly is important in estimating the imminence of hazardous activity for volcanoes with a characteristic period. Moreover, data on periodicity will reveal information on the rates of energy and mass input to volcanic systems. Some large volcanic systems, particularly those with rhyolitic magma chambers, undergo episodes of periodic, cyclic eruptivity separated by intervals of quiescence. The cyclic nature of the activity can involve both progressive changes in magmatic composition and variations in eruptive mode. Eruption of the Yellowstone rhyolites, for example, began 2 m.y. ago and has occurred in three episodes, each with a similar eruptive cycle of ash flows and caldera collapse. 5
Overview and Recommendations The Smithsonian Institution maintains historical records of eruptive activity in a computer data bank (see Chapter 8). The sequence and age of prehistoric eruptions can be determined by stratigraphic studies coupled with age determinations using such methods as potassium-argon, carbon 14, and secular magnetic variations. Records for historic eruptions in compressional arcs (such as the Cascades) comprise 80 percent of the, data for 530 volcanoes presently known to be active and more than 95 percent of the over 5000 eruptions tabulated. These dominantly andesitic volcanoes tend to erupt explosively, and their common location in belts along the borders of the Pacific Ocean put them near many population centers, making them a special danger to people. The average time interval between eruptions for those volcanoes that have had multiple eruptions in historic times is about 5 yr. The most catastrophic eruptions generally occur after the longest quiescent intervals. Data also show that there is no established pattern of timing of the paroxysmal phase of an explosive eruption. The paroxysmal phase can occur at the beginning of an eruptive cycle or at any time up to several years after the beginningâposing a serious problem in forecasting dangerous eruptions. Some volcanologists have speculated that the rise of magma from subduction zones beneath arc volcanoes takes place in diapirs. The explosive nature of these andesitic volcanoes, in combination with the characteristic, relatively short eruptive period, suggests the presence of magma chambers at intermediate depth. The form of such chambers is not known; their detection by geophysical techniques has to date been equivocal. The rate of eruption at volcanic loci in extensional tectonic regimes may be influenced by stresses that can provide avenues for magma rise and by convective movements in the mantle that supply both the heat and at least part of the melt. Cooling with partial crystallization and buildup of volatile pressure also must be important, especially for silicic volcanic systems with near-surface magma chambers. ERUPTIVE MECHANICS The mechanisms of volcanic eruptions can become better understood through theoretical and experimental studies as well as by more sophisticated field observations. New insights can be gained by consideration of data for volcanic eruptions on other planets and satellites, particularly Jupiter's moon, lo. Problems that are important from the viewpoints of both scientific progress and hazard evaluation include the interrelation- ships of physical factors that determine the explosive character of an eruption and the identification of sources of volatiles involved in explosive eruptions. The variety of sizes and shapes of volcanic jets associated with violent eruptions reflects differences in the geometric shapes of near-surface parts of the volcanic systems and, secondarily, the differences in thermodynamic properties and rheology of the erupting fluids. The most substantial accelerations in a volcanic eruption occur when the volatiles, either juvenile or meteoric, have become a vapor or vapor-particulate mixture. A ther- modynamic equation of state of the erupting fluid can be modeled by a pseudo-gas equation in which the vapor-particulate mixture is approximated by a gas of heavy molecular weight and low isentropic exponent (see Chapter 11). In volcanic systems the erupting fluid will accelerate to sonic conditions in narrow parts of a conduit, exiting with pressures that are high enough to erode surface craters. Pressure-balanced jets result if the pressure approximately equals ambient when the jet emerges at the surface. Overpressured or underexpanded jets occur if the jet pressure greatly exceeds atmo- spheric. Plinian eruption columns are typical terrestrial balanced jets because they commonly emerge through craters; on lo, jet 3 (Prometheus) may be an equivalent. The lateral blast of May 18, 1980, at Mount St. Helens, however, was an overpressured 6
Overview and Recommendations jet because it emerged directly at high pressure from the face of the mountain; on lo, jet 2 (Loki), which apparently emerges through a fissure, may be an equivalent. The accurate prediction of hazards from volcanic jets will require a detailed understanding of the pressures within jets, in terms of volcanic system geometry and fluid thermo- dynamic properties. In some eruptions the volatiles are juvenile and exsolve as pressure on the magma is reduced. In terrestrial volcanic systems, however, meteoric water can be combined with lava in eruptive events either below the surface or subsequent to eruption. Ex- periments to model subsurface interactions of meteoric water and magma have been made with thermite and water encased in steel vessels (see Chapter 12). The thermite forms a basalt-like melt and its interaction with water can be controlled by varying the contact geometry of the container. Eruption modes ranging from strombolian and surt- seyan to passive chilling with pillow formation have been produced. The explosivity has been found to be a function of both the water-to-melt mass ratio and the confining pressure. The May 18, 1980, eruption of Mount St. Helens was recorded by infrared sensors on military satellites. Study of these records has shown that the eruption started with two major explosions about 2 min apart. The first explosion, near the summit, occurred when a giant avalanche relieved confining pressure on the intruding magma. The second explosion occurred 8 km to the north near the Toutle River and may have been caused by interaction of hot dacite with surface water in the river valley and possibly Spirit Lake (see Chapter 10). Mud flows and floods are a special hazard for areas surrounding major volcanoes, where unstable terranes are produced by a combination of steep slopes, immense quantities of water (including snow and ice), and large masses of unconsolidated tephra and hydrothermally altered rock. These hazards persist long after the eruption. Sub- stantial success in evaluating flood hazards for the environs of Mount St. Helens has been achieved by computer modeling of two danger zones (Jennings et al., 1981). The interface between volcanology and hydrology offers important research opportunities, and vigorous pursuit of these opportunities will help alleviate risks to life and property. EMERGENCY PLANNING Most of the potentially dangerous volcanic loci in the western United States have been identified and can be monitored. Some volcanic systems other than Mount St. Helens, possibly including loci outside the Cascades, may become active in the next few decades. The youngest volcano in the conterminous United States outside of the Cascade Moun- tains is an alkali basalt cinder cone and associated flow that erupted only 390 yr ago in the eastern Mohave Desert, 180 km from Los Angeles (see Chapter 1). Eruptions from such vents are more likely to be moderately explosive or nonexplosive than catastrophic, but there is still a significant risk that a major tuff-flow eruption, like the Bishop tuff, might occur or that an ash fall might blanket crops and affect population centers in a number of states. Clearly we should plan for such emergencies. Recent experiences with volcanic crises in the Caribbean (see Chapter 13) and at Mount St. Helens (Miller et al., 1981) have provided insights into problems that can arise. A common problem in all such crises is the effective flow of information from informed scientists to the public. This difficulty is compounded by a number of factors, including the following: ⢠Conflicting information about risks or conflicting predictions about future events undermines confidence in crisis leadership and promotes either panic or apathy. These
Overview and Recommendations conflicts can arise because individuals or groups of scientists disagree or appear to disagree about the probable future course of events. Differences of opinion are inevi- table. However, in an emergency interested scientists who lack up-to-date information should refrain from making comments to the press. ⢠Confusion results from multiple official voices. Problems have developed during volcanic crises through a variety of governmental organizations issuing press releases. Even if all their releases are virtually the same, differences in emphasis and wording can lead to confusion and distrust. ⢠Conflicting reports in crises arise because of distortions or misinterpretations by the press. These also must be regarded as inevitable, but their damage can be minimized if the public and the press develop confidence in a scientific team that has the official responsibility for monitoring the eruption. RECOMMENDATIONS 1. Physical and chemical processes interact in the generation and eruption of magma. The diversity and extent of these interactions, however, is poorly understood. In western North America there is a broad correlation of magma composition with tectonism in that andesites are primarily erupted in zones of tectonic compression, whereas eruption of basalt, sometimes accompanied by rhyolite, is characteristic of broad areas under tectonic extension. Why these correlations exist, however, is a matter of speculation. Basalt is now generally believed to be the thermal parent of major extrusions of rhyolite. In what degree basalt is also the chemical parent of these rhyolites is a matter that has been debated for decades and remains an unsolved problem. The mechanics of formation of large silicic magma chambers with basaltic roots extending into the mantle are not known. Partial fusion and assimilation of crustal rocks are widely believed to be involved in the generation of these silicic lavas, but in such cases what is the physical nature of the interface between the basaltic magma and the crustal rocks? How is the geometric form of a major magmatic system influenced by tectonic forces? Partial fusion of wall rocks would be unlikely in the absence of volatiles, but what is the source of the volatiles and how do they migrate during formation of large magma chambers in the crust? These are important petrologic problems, all of which have physical as well as chemical aspects. Research on petrologic processes, especially those that involve reactions between magma and crustal or mantle rocks, should be undertaken on the physical as well as the chemical aspects of the processes. We suggest that progress in understanding such processes may be enhanced if geochemists and geophysicists collaborate in a consid- eration of the physical as well as the chemical nature of magmatic evolution. 2. Precursory measurements of both a geophysical and a geologic nature are essential for reliable assessment of volcanic hazards. An inadequate base of such measurements for the Soufriere Volcano on Guadeloupe in 1976 resulted in great difficulty both for scientists who were called on to forecast the course of that eruption and for a government concerned with the continuing evacuation of 70,000 people (see Chapter 13). The catastrophic eruption of Mount St. Helens on May 18, 1980, was preceded by a 2- month period during which seismicity was high and minor explosions were intermittent. This period provided an opportunity for installation of geophysical instruments and development of a rudimentary measurement base. Of equal importance in assessing the Mount St. Helens hazard was the detailed field study of the nature and sequence of prehistoric eruptions (Crandall and Mullineaux, 1978). These authors forecasted eruptive events at Mount St. Helens with remarkable accuracy. A sustained effort should be made to obtain a precursory geophysical and geologic 8
Overview and Recommendations data base for potentially dangerous volcanic loci. Without such a data base trustworthy forecasts are not possible. Such data should include identification of the nature and extent of prehistoric flows and tephra deposits, construction of hazard maps, hydrologic characterization, and geodetic monitoring (trilateration as well as leveling observations). Passive seismic mon- itoring should be maintained in volcanic areas where the presence of magma is sus- pected. The data base at Kilauea Volcano has allowed fairly accurate short-term (days) predictions to be made; a similar capability for forecasting events at Mount St. Helens is now being obtained. Active seismic studies, such as those carried out in Yellowstone (see Chapter 7), are important for theoretical as well as hazard assessment. Such studies may be of great value in helping us understand the physical form of magma chambers. Active seismic studies together with electrical and other geophysical measurements (e.g., heat flow, magnetics, groundwater behavior) may provide information on subsurface magma ac- cumulation and should be actively pursued. 3. Remote monitoring data obtained by branches of the U.S. Department of Defense (DOD) have been made available in a few instances to volcanologists for application to problems of general geophysical interest. An exciting example of such collaboration was the use of infrared data of satellite origin to determine the initial sequence of events during the cataclysmic eruptions of Mount St. Helens on May 18, 1980 (see Chapter 10). Another example was the use of underwater sound data to detect and locate a number of active submarine volcanoes in the Pacific Ocean (Norris and Johnson, 1969). A related circumstance in which use of remote monitoring might potentially be useful is in the occasional occurrence of pumice rafts in the Pacific Ocean. These are produced in submarine eruptions, but in some cases the locations of the sources are difficult to identify because the pumice rafts drift intact for vast distances. Triangulation with data from hydrophone arrays might, however, result in identification of these vents. Another kind of data that would be useful to volcanologists is high-resolution aerial photography. Volcanoes exist for which no published photographs are available because of their remote location and frequent cloud cover; Reventador Volcano in Ecuador is an example. Data obtained in the course of routine remote monitoring for the DOD may be retained for only short periods of time. Moreover, reduction of such data for volcan- ological purposes can be a time-consuming process for those few persons who have access to the raw data. For these reasons applications to theoretical problems will be most easily brought about through collaboration of research workers with interested scientists engaged in DOD operations. Volcanologists in research positions should make every practical effort to build in- tellectual bridges with colleagues whose related work is within the DOD establishment. We urge the Department of Defense to develop policies wherein maximum use, including volcanological use, is made of data acquired during routine geophysical monitoring. 4. The USGS is the only organization in the United States with the available man- power, support, and expertise to assume the task of monitoring dangerous volcanic eruptions, of warning the public, and of providing analyses of events. A commendable job was done by the USGS during the Mount St. Helens crisis. Operational decisions to block access to an area or to evacuate people, however, must be the responsibility of other organizations. These will normally be local, either state or municipal, bodies or local representatives of federal agencies, including the Forest Service and the National Park Service. The Federal Emergency Management Agency (FEMA) is responsible for contingency planning and coordinating responses of all federal agencies to disasters; in the past FEMA has cooperated effectively with the USGS. It would be very helpful during a volcanic crisis if the USGS team responsible for 9
Overview and Recommendations monitoring the event and providing information was able to establish communications with local officials and the press in advance of the crisis. The nature of volcanic eruptions is such that there will normally be ample time to lay a communications groundwork. Efforts to provide such a groundwork are under way at Mammoth Lakes. The USGS should continue and extend these communication efforts with local, state, and federal agencies wherever signs develop of potential volcanic activity that could be a danger to the public. REFERENCES Crandall, D. R., and D. R. Mullineaux (1978). Potential hazards from future eruptions of Mount St. Helens, Washington, U.S. Geol. Surv. Bull. 1383-C, 26 pp. Crosson, R. S. (1983). Review of seismicity in the Puget Sound region from 1970 through 1978, in Proceeding of Workshop XIV: Earthquake Hazards of the Puget Sound Region, Washington, J. C. Yount and R. S. Crosson, eds., U.S. Geol. Surv. Open-File Rep. 83-19, pp. 6-18. Jennings, M. E., V. R. Schneider, and P. E. Smith (1981). Computer assessments of potential flood hazards from breaching of two debris dams, Toutle River and Cowliz River systems, in The 1980 Eruptions of Mount St. Helens, Washington, P. W. Lipman and D. R. Mullineaux, eds., U.S. Geol. Surv. Prof. Paper 1250, U.S. Government Printing Office, Washington, D.C., pp. 829-836. Miller, C. D., D. R. Mullineaux, and D. R. Crandall (1981). Hazard assessments at Mount St. Helens, in The 1980 Eruptions of Mount St. Helens, Washington, P. W. Lipman and D. R. Mullineaux, eds., U.S. Geol. Surv. Prof. Paper 1250, U.S. Government Printing Office, Washington, D.C., pp. 789-802. Norris, R. A., and R. N. Johnson (1969). Submarine volcanic eruptions recently located in the Pacific by sofar hydrophones, /. Geophys. Res. 74, 650-664. U.S. Geological Survey (1982). Notice of potential volcanic hazard issued for eastern California, press release, May 25. 10
BACKGROUND
