Review of the U.S. Geological Survey's Volcano Hazards Program (2000)

The committee hopes that if a major eruption were to occur in the United States in the year 2010, the USGS VHP would be prepared to respond in a manner more similar to Prologue 2 than to Prologue 1. This chapter discusses in greater detail the committee’s vision for the VHP exemplified in Prologue 2.

Most of the technology and understanding described in this vision exist today, and others are extrapolated from current research. The scenario is optimistic but realistic. For this country to have a VHP capable of saving tens of thousands of lives and greatly limiting property damage and economic disruption, investment in technology, people, and basic research is required. Although the hypothesized eruption could occur tomorrow, in 2010, in 2110, or in any future year, the decisions made today will greatly affect VHP’s ability to forecast, monitor, predict, and minimize the effects of devastating volcanic events. The consequences of failing are great, as outlined in Prologue 1. The difference may range from a few fatalities to thousands of deaths, from major damage to structures to complete regional economic collapse.

The first sign that Mount Rainier was reawakening came from crustal deformation measurements at the CVO. A dense array of permanent GPS receivers in the area picked up subtle ground movements nearly a year before the eruption. The motions were so small (<1 mm) that they would

not have been recognized by looking at the data from an individual instrument. Only sophisticated computer programs especially designed to integrate numerous data sets and to search for patterns of ground deformation diagnostic of magma migration were able to detect the early warning signs.

Data from a constellation of orbiting InSAR satellites were also critical. Interferometric maps of ground motion automatically generated from these data, with a horizontal spatial resolution of 10 meters and a vertical resolution of 0.1 mm, also showed subtle indications of inflation of a deep magma body. Because these InSAR systems employed longer-wavelength signals than the radars used at the end of the twentieth century, they were less affected by vegetation growth and other surficial processes. Imaging from multiple satellites allowed three-dimensional vector displacements to be determined and compared directly with the GPS measurements. The analysis software detected inflation of part of a deep magma chamber 15 km beneath the earth’s surface.

The deformation alert triggered several immediate actions. After presenting the results to the scientist in charge of CVO, the deformation group began an intensive series of tests to check the validity of the data and the automated computer modeling. The scientist in charge brought up on a computer screen a three-dimensional, interactive hologram, showing version 15.3 of the hazard assessment for Mount Rainier and called in the heads of the other scientific groups. The scientist in charge did not have to be reminded that Mount Rainier posed the highest risk of any volcano in the continental United States. Rainier had been a high-priority volcano for study since the 1990s, and hazard assessments had been revised and updated numerous times since then. The nominal annual probability of an eruption with a Volcanic Explosivity Index (VEI) greater than or equal to 2 was 4 percent. (Such an eruption would be classified as explosive and involve roughly one million cubic meters of ash.) The new information would increase this probability significantly.

All major Cascade volcanoes had updated hazard assessments based on extensive field mapping by teams of experienced geologists. Field mapping, combined with subsurface imaging using ground penetrating radar and shallow seismic methods, allowed the hazard assessment team to develop complete three-dimensional maps of deposits from past eruptions, debris flows, and landslides and to assemble detailed

descriptions of the volcano’s eruptive style. Advances in geochronology during the previous 1.0 years allowed the team to date eruptive products with unprecedented precision and to derive sophisticated models of processes active within the magma chambers deep below the mountain. The three-dimensional mapping combined with accurate dating and dramatically improved models of volcanic eruption processes allowed the team to assign probabilistic estimates for impending eruption scenarios. The scientist in charge was thus able to update the likelihood of these scenarios in near real time, as monitoring data from satellites and remote locations on the volcano streamed into the observatory. The scientist in charge scanned three-dimensional displays showing the extent of a worst-case scenario directed blast and evaluated the probabilities.

At this point, none of the other monitoring systems reported anomalous signals. Seismicity was at background level and no unusual gas emissions had been detected. The scientist in charge asked for updated seismic and electromagnetic images of the subsurface. The permanent broadband seismic network was augmented with portable stations and electromagnetic sensors brought to the area by university scientists working with researchers from the USGS VHP. Together these scientists were able to create high-resolution images of subsurface structure and time-dependent changes in rock and fluid properties. Laboratory calibrations were used to interpret the images of seismic velocity, attenuation, and electrical resistivity in terms of temperature, composition, and extent of partial melting. Based on these results, the team reported to the scientist in charge that the deep magma chamber beneath the volcano had indeed swelled in size and changed in shape. Given the available data, they were able to estimate the size, shape, depth, and location of the magma body. The potential for an eruption was identified, but the future behavior was still unclear.

By this time, civil defense officials, as well as state, county, and municipal authorities had been briefed on the changes taking place beneath the mountain. Although there were no visible signs of unusual activity, these officials were well aware of the potential hazards posed by the volcano. VHP scientists had been working closely with Washington State emergency management personnel. At this point the scientist in charge reported that it was too early to determine whether this increase in the size of the magma chamber would lead to an eruption. The USGS

issued an updated probabilistic assessment based on the new geophysical data. The Washington State Emergency Management Agency developed an action team to deal with a worst-case Mount Rainier eruption scenario. The action team reviewed specific plans and assignments developed over the past decade in the following areas:

• identification and mapping of the hazard zones; registering of valuable movable property;

• identification of safe refuge zones to which the population could be evacuated;

• identification of evacuation routes, their maintenance, and clearance;

• identification of assembly points for persons awaiting transport for evacuation;

• means of transport and traffic control;

• shelter and accommodation in the refuge zones;

• inventory of personnel and equipment for search and rescue;

• hospital and medical services for treatment of injured persons;

• security in evacuated areas;

• the formulation of alert, warning, and evacuation procedures;

• relocation and recovery activities; and

• provisions for revising and updating the plan.

During this time, a university team brought in an array of absolute gravity meters, capable of measuring the earth’s gravitational force with an accuracy of 1 part in 100 million. They detected a slight increase in the gravity field that, when combined with the GPS, InSAR, seismic, and electromagnetic imaging results, helped constrain estimates of the density and composition of the magma. The results were not encouraging. The magma was in all probability dacitic, the same composition as the devastating eruption of Mount St. Helens in 1980. At the same time, field teams began monitoring the volcanic edifice in response to the subtle changes in strain. Special care was taken to monitor local strain rates in well-known alteration zones high on the flanks of the volcano in order to record the response of the shallow hydrothermal system to the new activity.

Meanwhile VHP scientists used geophysical and geochemical observations to initialize eruption models that predict the volume of erupted material, height of ash plumes, size and distribution of pyroclastic flows, and related hazards. Based on recent improvements in understanding magma rheology, chemistry, and volatile kinetics, it was possible to integrate the physical and thermodynamic governing equations forward in time to predict magma flow, eruption potential, and behavior during eruptions. Although these models were not sufficiently well constrained to accurately predict the exact time and magnitude of the eruption, they did reveal a range of outcomes that were then considered by disaster planners. The predictive capability improved as more data were collected and the volcanic activity increased.

Indeed, much of the improvement in the ability to assess and forecast volcanic hazards had come from an improved understanding of volcanic processes. These advances were derived from basic research in theoretical, numerical, and laboratory studies, along with knowledge gained from global monitoring of volcanic systems. Advances in understanding physical and chemical processes improved monitoring capability and led to better methods for interpreting data. At the same time, the integrated data sets collected at active volcanoes and in the course of hazards assessment studies provided the basis for testing concepts about processes within active systems. One practical result of this research was the construction of engineered barriers around Mount Rainier, capable of diverting flows of mud and debris away from critical facilities and population centers. Structures were strengthened to withstand anticipated ash accumulations.

At the same time, other scientists were analyzing the stability of the glacier-clad flanks of the volcano and the possibility of devastating debris flows. Computer models of hot ash accumulation onto Rainier’s snow and ice fields estimated the maximum possible runout distances and flow volumes. Acoustic flow monitors, which detect the ground vibrations due to fast-moving mudflows, were double-checked, and people in the path of possible mudflows were alerted and evacuation procedures reviewed. These models and monitoring devices had been developed in part based on sophisticated studies of debris flows using a large-scale experimental facility designed and run by USGS VHP scientists.

Six months after the first signs of magma chamber swelling, the first unusual seismic events were recorded. A dense, broadband, high-dynamic range, seismic network jointly operated by CVO and a consortium of university groups detected an intense swarm of earthquakes at a depth of 8 km. This was considerably shallower than magma sources previously noted and was the first indication that magma was migrating upward. Within a few weeks, the first long-period seismic events were recorded, providing further evidence of magma flow. High-resolution earthquake locations using full seismic waveforms yielded resolution of a few meters in near realtime, allowing for precise imaging of seismically active structures. The earthquakes outlined three fingers of magma migrating upward through the crust. Sophisticated source modeling helped seismologists locate constrictions in the magma conduit that caused pressure increases followed by episodic discharges of magma. Shortly thereafter, the first recordings of harmonic tremor, a low-frequency oscillation detected by seismic instruments, were reported. Research by VHP postdoctoral scientists working with VHP theoreticians had led to well-tested models of harmonic tremor. Arrays of seismic instruments were able to locate the source of the tremor, and the amplitude information was used to calibrate the size of the conduit and the volume flux of melt.

The VDAP team installed a network of five borehole tilt and strainmeters to 100-meter depths using microdrilling methods. The VDAP team had decades of experience in volcanic crises throughout the world, and most VHP personnel had invaluable hands-on participation in VDAP. The miniaturized instruments, using sensor-on-a-chip technology developed in partnership with the Department of Energy, began monitoring strain signals associated with the pulse-like rise of magma within the conduit system. The VDAP team was prepared to handle not only the voluminous influx of monitoring data, but also the increasingly aggressive media attention that the volcanic awakening had created.

By this time, the USGS had issued its first low-level eruption forecast. The time and the magnitude were still unclear, but the potential for an eruption grew increasingly more likely. The projections were updated periodically as more data became available, in much the way that weather forecasts were done at the end of the twentieth century. Seismic, strain, and geochemical data were available in near real time over the Internet. Because of extensive outreach and education, the

public was generally able to manage the information flow and make reasonable decisions. Many residents living within the highest hazard zones decided to leave the area. Disaster preparedness planning was well advanced. The Washington State Emergency Management Agency task force completed a tabletop simulation of the Mount Rainier eruption emergency plan. Modifications to the response plan were made based on lessons learned from the exercise.

Eleven months after the first indications of ground deformation and five months after the onset of increased seismic activity, the first signs of volcanic gases were recorded. Miniaturized gas sensors placed on broad fracture systems on the sides of the volcano picked up emissions of CO₂ and helium. In addition to noting high concentrations of these gases, the in situ sensors were able to measure their isotopic compositions, which showed a clear magmatic signature. At the same time, space-based sensors, developed in partnership with NASA, including LIDAR, with the capability of monitoring gas compositions, revealed very low amounts of gas being released from the volcano. This was particularly alarming, because the estimates of magma volume, from geodetic, strain, gravity, and seismic tomography indicated a large migrating body. The absence of gas emissions suggested that the magma was not degassing at the surface, and therefore was building in pressure.

By this time the governor of the State of Washington had ordered the evacuation of the volcanic hazard zones. This was based on information from the VHP personnel estimating a high probability of an explosive eruption of Mount Rainier during the months of June or July. Although air traffic was routine, an official notice to aircraft flight personnel from the FAA and the Volcanic Ash Advisory Center was released that described the potential for an eruption to inject large amounts of ash into the atmosphere. Existing contingency plans for rerouting air traffic were implemented. Shortly thereafter, the president of the United States declared a state of emergency for Washington State. This made federal resources available to aid the state and its affected residents. The governor and president acknowledged the tremendous coordination of scientific information by the VHP, not only to public safety officials, but also to the affected population. This coordination minimized confusion about what to believe and whom to believe. On May 15, 2010, the governor of the State of Washington announced the completion of the evacuation of hazard zones identified by VHP personnel.

On May 18 the eruption process accelerated exponentially. Until this time the effects had been subtle. Indeed, none of the many signs tracked by USGS scientists would have been detected without sensitive instrumentation, developed over the last 10 years in collaboration with numerous university and government colleagues. Beginning at 1:00 p.m., an intense swarm of earthquakes started, with hypocenters shallowing markedly over the next two hours. Harmonic tremor amplitude increased dramatically. Automated event location algorithms identified zones of intense fracturing ahead of a rising magma body. Strain-measuring instruments and GPS receivers recorded motions of meters in a matter of hours as the dike rose toward the surface. At this point, a short-term forecast of high probability of an explosive eruption was issued. Air traffic was diverted away from the region and critical facilities went into automated shutdown procedures. The Air National Guard was called in to deploy a widely dispersed network of relatively inexpensive, case-hardened, biodegradable microsensors capable of detecting and relaying, via satellite, ambient temperature, pressure, humidity, and geochemical conditions to recording stations tens of kilometers away. This technology, developed for use on battlefields, had recently been partially declassified for application to natural hazard emergencies.

The sequence was unusual in the rapid acceleration toward the climactic eruption. At 3:00 p.m. the north flank of the volcano exploded in a directed blast. This was followed shortly by an eruption cloud rising to 30 km in the atmosphere. Seismic and acoustic infrasound networks combined with space-based optical, radar, and thermal sensors, operated by NOAA, NWS, NASA, and nuclear treaty monitoring networks, rapidly detected the onset of the eruption, thereby broadcasting an instantaneous notification around the world. A combination of these data and the ambient information provided by the widely dispersed microsensors was used to rapidly quantify the size and explosivity of the eruption. Numerical models of ash dispersal combined with accurate models of wind direction and strength, operated in collaboration with the NWS and the FAA, provided timely warning of ash hazard to aircraft and other affected entities.

As it rose, part of the volcanic ash column became unstable and collapsed, producing pyroclastic flows that raced down the Puyallup and Carbon River valleys. Within hours, debris flows inundated other areas low on the flank of the volcano. Some of these debris flows were

diverted by engineered structures, built low on the flanks of the volcano for this purpose. Property damage and loss of life in these areas were limited. Elsewhere, engineered systems failed to contain the debris flows. Residential areas and business parks in the lower Nisqually River were devastated when the Alder Dam failed in response to a lahar pulse entering the reservoir around 5:00 p.m. Fortunately, these events were anticipated by hazard assessments and, as a result, by the public. Even in these areas, loss of life was minimal because of prompt evacuation by communities well aware of the risks. Property damage, however, exceeded several billion dollars.

As the eruption progressed through the night, VHP staff focused on forecasting rates of ash accumulation in nearby communities, continued lahar hazards, and the probable duration of the eruption. All of this information was crucial in the following days to organize a safe and rapid response to the disaster.

This page intentionally left blank.