1U.S. Geological Survey.
P-wave arrivals from 298 events were used to determine the three-dimensional velocity structure in the region of Mount Pinatubo. The inversion results show several well-resolved regions of low-velocity (5 to 10 percent) which are inferred to be locations of magma bodies. There is a large low-velocity region between depths of 6 and 11 kilometers that is estimated to have a volume of 40 to 90 cubic kilometers. An extension of this low-velocity region reaches toward the surface under Mount Negron, which could represent magma that supplied eruptions at that vent. Another shallow low-velocity zone is located below the northwest flank of Mount Pinatubo, under the small caldera that formed during the June 15 eruption. If all of these low-velocity regions are interpreted as magma bodies, we estimated a total volume of 60 to 125 cubic kilometers for the magma system under Mount Pinatubo and Mount Negron, which is one of the largest active magma chambers. The three-dimensional velocity structure was also used to relocate all of the earthquakes recorded from May 5 to August 16, 1991. Compared to the locations done with a one-dimensional model, the hypocenters tend to be more tightly clustered and show more clearly the linear trends in both the preeruption and posteruption seismicity.
Note to readers: Figures open in separate windows. To return to the text, close the figure's window or bring the text window to the front.
The seismic activity associated with the 1991 eruptive activity of Mount Pinatubo provided a data set of many closely spaced earthquakes that can be used, through seismic tomography, to determine the three-dimensional P-wave velocity structure of the volcano. These results provide some of the few observations about the structure at depth under this volcano and are of particular interest for inferring the location and size of magma bodies that may have been the source of the large eruption of June 15, 1991. Three-dimensional velocity studies at other volcanoes (Iyer, 1992) have identified low-velocity zones that have been interpreted as magma bodies, such as Kilauea, Hawaii (Thurber, 1983), Mount St. Helens, Washington (Lees, 1992), Phlegrean Fields, Italy (Aster and Myer, 1988), Medicine Lake, California (Evans and Zucca, 1988), Newberry Volcano, Oregon (Achauer and others, 1988). Most of the magma bodies identified in these papers have been relatively small, with dimensions of only a few kilometers. The size of the 1991 Mount Pinatubo eruption of about 3.7-5.3 km3 of magma (W.E. Scott and others, this volume) suggests there is a large source that may be imaged with the tomographic methods. Eruptions are thought to expel only a small percentage of the magma chamber volume, so the feature for which we are searching may have a volume on the order of 100 km3. Some other studies have searched for large magma chambers under calderas at Long Valley, California (Sanders, 1984, Romero and others, 1993) and Yellowstone Caldera, Wyoming (Clawson and others, 1989), but there is still relatively little direct evidence for active magma reservoirs of 100 km3 in size.
After determining the three-dimensional velocity structure, this velocity model can be used to relocate the earthquakes more accurately compared to the hypocenters calculated by using one-dimensional structures. The goal of this paper is to search for low-velocity regions that might be the magma source for the 1991 eruption and to investigate locations of the low-velocity regions in relation with the accurately relocated seismicity. This information provides valuable insights into the "plumbing" structure of the volcano and the processes that were occurring before and during the eruption.
This study used a set of 298 small earthquakes (M 2.0-4.0) recorded on the short-period network at Mount Pinatubo (Lockhart and others, this volume, Murray and others, this volume). There were 70 events from the preeruption period of May 16 to June 12 and 228 events from the posteruption period of June 29 to August 16 that were chosen to be relatively well-distributed in space around the volcano (fig. 1). The earthquakes were located in an irregular volume that extended from the summit to distances of 10 to 20 km and ranged in depth from the surface to 20 km. The data for the tomographic inversion were the P-wave arrival times that were carefully picked from the digital seismograms with estimated uncertainties of 0.02 to 0.1 s. S-wave arrivals, especially those picked from horizontal components at PPO (Patal Pinto) and PIE (Pinatubo east), were used to help the location of the earthquakes, but there were an insufficient number to be used for an S-wave velocity inversion. All but two of the seven sites that recorded the arrivals of the preeruption events were completely destroyed by the eruption so, there is a different configuration of seven stations for the posteruption events (fig. 1). The two stations that were common to both time periods are CAB (Clark Air Base) and BUG (Sitio Buag).
Figure 1. Map view of the Mount Pinatubo area showing the grid of node points (crosses) and earthquakes (small circles) used for the inversion. Stations of the preeruption network are shown by triangles, and stations of the post-eruption network are shown by squares. One station of the posteruption network is off the figure to the north. The topographic contour interval is 500 m. The west-to-east cross section shows all the hypocenters on the map and the depth distribution of the node points.
Having the two different station configurations for the preeruption and posteruption periods is an advantage to the tomographic study because it provides a more varied distribution of source-station ray paths. The preeruption seismicity located within several kilometers of Mount Pinatubo and recorded with nearby seismic stations can resolve the velocity in the shallow region close to the summit. The posteruption seismicity was spread over a much larger area and depth and therefore enabled us to estimate the velocity structure in a much larger volume around the volcano. However there were no stations close to the summit for the posteruption period to resolve the shallow summit structure. The two data sets are complementary and, when combined, allow a velocity inversion of both the shallow region near the summit and also the larger surrounding volume. One potential problem is that the velocity structure may have changed significantly during the June 15 eruption. If this is the case, combining these two data sets may be mixing data from two different velocity structures. This problem will cause errors in the shallow portion of the velocity structure, but the deeper portions of the model should still accurately reflect the velocities of the posteruption period. The preeruption earthquakes were all shallow and close to the summit, so ray paths to the stations do not travel at very large depths through the model and therefore should not affect the results in that region.
A computer program originally written by Thurber (1981, 1993) and modified by Eberhart-Phillips (1993) was used to invert the arrival times for the three-dimensional velocity structure. This program can simultaneously solve for the velocity structure and the hypocentral locations by using an iterative damped least-squares inversion. The velocity structure is parameterized by a three-dimensional grid of node points. The velocity structure is derived by minimizing the traveltime residuals rij for event i from station j with respect the earthquake hypocentral parameters (ti, xi, yi, zi) and the n node point velocities (vn).
For the calculation of traveltimes through the structure, linear velocity gradients are used between the node points, and rays are traced by using an approximate three-dimensional algorithm that produces curved nonplanar ray paths (Um and Thurber, 1987). This method results in a final velocity model with smooth velocity variations within the volume rather than sharp velocity contrasts.
In this study we started with a one-dimensional velocity structure (table 1) and first inverted the data to find the best one-dimensional model that fits the data (table 1). This model already reflects some of the anomalous velocity structure near the volcano, with a low-velocity zone at 7 km depth. The one-dimensional model was used as our starting velocity structure for the three-dimensional inversion. The node points for the three-dimensional inversion were spaced at 1- and 3-km intervals horizontally and at depths of 0, 1, 4, 7, 10, and 15 km, as shown in figure 1. There are relatively more ray paths from the distribution of earthquakes and stations that cross in the central region of the grid near the volcano summit, so the closer spacing of the node points allows finer resolution of the velocity structure in the summit area.
Table 1. Starting velocity model (left) and best one-dimensional model (right) from inversion.
|
Depth |
P-wave velocity |
P-wave velocity |
|---|---|---|
|
0.0 |
3.30 |
3.89 |
|
1.0 |
4.00 |
4.59 |
|
4.0 |
5.00 |
5.72 |
|
7.0 |
5.70 |
5.44 |
|
10.0 |
6.00 |
5.48 |
|
15.0 |
6.70 |
5.70 |
For the velocity inversion, P-wave data were used with weighted arrivals, depending on the uncertainty of the phase pick. One weakness of this data set is that there are relatively few arrivals for each earthquake. There can be a maximum of only seven, and sometimes only five or six stations recorded an earthquake. If we solve simultaneously for the velocity structure and the earthquake locations, there are relatively few data points compared to the number of unknowns, and the inversion can be unstable. For this reason, we used an iterative procedure of fixing the earthquake locations and solving for the velocity structure; then we fixed the derived velocity structure and relocated the earthquakes. The procedure minimized residuals (equation 1), alternately holding fixed the partial derivatives with respect to the hypocentral parameters and then holding fixed the partials with respect to the velocities. In the location step we included the S-wave arrivals, which helped to constrain the depths of the earthquakes. We used a constant ratio of the P- to S-wave velocity (vp/vs = 1.8). Three iterations of this process caused the inversion to converge to a solution.
Figure 2. Depth slices of the three-dimensional velocity structure derived in this study. The colors show velocity differences from the average velocity at a given depth. Hotter/colder colors correspond to slower/faster velocities. Lines on the 1-km depth slice show the positions of the cross sections in figure 3.
The results of the velocity inversion are shown in figure 2 as horizontal slices through the model at depths of1, 4, 7, 10, and 15 km. For the final velocity model and hypocenters, the root-mean-square fit of the predicted traveltimes to the observed traveltimes for the 296 events was 0.069 s, compared to 0.178 s for the best one-dimensional model. The colors reflect velocity differences from the average velocity at each depth, with slower/faster velocities shown by hotter/colder colors. The velocities generally range +-10% from the average value for each depth. The colors have been smoothed to show graduated velocity changes throughout the model, although very strong contrasts between adjacent nodes are reflected in the blockiness of some parts of the images. The inversion solves for discrete velocities at the nodes and assumes a linear gradient between the nodes for calculation of the raytracing and the velocity partial derivatives for the inversion. Therefore, the smoothed color images are representative of the smooth velocity variations that were determined in the model.
The pattern of velocity variations tends to be more complicated for the shallower depth slices at 1 and 4 km, which have closely spaced regions of fast and slow velocities. This might be expected in the vicinity of a volcano where the shallow crustal structure could have a complexity caused by magma intrusions, lava flows, ash deposits, and variations in degree of fracturing. Also, parts of the shallow velocity structure away from the summit are not well resolved, and the strong variations could be reflecting instabilities in the inversion. This is particularly true for the strong perturbations along the edges of the depth slices. At greater depth the model has a simpler image, which may be partly due to more homogeneous structure and partly to better stability of the model in a region where there are more crossing ray paths. One conspicuous feature in the 7- and 10-km depth slices is the strong low-velocity region south of the summit area. This region is 2 to 3 km across in the east-west direction and 4 to 5 km in the north-south direction, with velocities that are 5 to 10% slower than the surrounding material. At 15 km and greater depth, there is little resolution of the model because of the small number of crossing ray paths.
In the top portion of figure 3, the velocity model is plotted in a south-to-north cross section through the summit area. The black dots show the locations of the preeruption seismicity from May 7 to June 11. This image clearly shows the large low-velocity region south of the summit that was seen in the 7- and 10-km depth slices of figure 2. The low-velocity body is 4 to 5 km across at depths between 6 and 11 km, with an extension from the southern edge that reaches toward the surface. Another low-velocity region, possibly connected to the deeper one, is seen at shallow depth (1 to 3 km) below the northern flank of the old Pinatubo summit. The 1-km depth slice in figure 2 shows that this area is actually on the northwest flank of Mount Pinatubo, under the small caldera that formed during the June 15 eruption.
Figure 3. Cross sections of the velocity perturbations from the average velocity at a given depth. Top, South-to-north cross section of the velocity structure through the summit area. The black dots are all the preeruption seismicity recorded from May 7 to June 11. Bottom, West-to-east cross section of the velocity structure through an area 2 km south of the summit. The black dots are the posteruption seismicity from June 29 to August 16 in the region 3 km north and south of the summit.
The bottom portion of figure 3 is a west-to-east cross section through the strong low-velocity zone 2 km south of the summit. The black dots are the posteruption earthquake locations. The low-velocity region falls within the outward trending limbs of the seismicity starting from about a depth of about 7 km and extending to the bottom of the model. There are relatively few earthquakes located within the low-velocity region between the two outward-trending limbs of seismicity.
We estimated the volume of the large low-velocity region between 6 and 11 km in depth to be 40 to 90 km3 and the extension toward Mount Negron as 15 to 25 km3. The smaller low-velocity region at shallow depth under the northwest flank of Mount Pinatubo was estimated to be 6 to 9 km3. The sum of these three volumes gives a total of 60 to 125 km3.
Using the three-dimensional velocity structure derived in this study, we recalculated hypocenters for all of the available data from May through August 16. To show the improvement in the hypocenters, we also relocated the earthquakes using the best one-dimensional structure (table 1). Figure 4 compares the posteruption locations determined using the one- and three-dimensional velocity structures for the preeruption period. Figure 5 shows the same for the posteruption period. For the three-dimensional structure, the hypocenters are more tightly clustered and appear to form more distinct spatial patterns, particularly in the cross sections. Dipping trends of earthquakes in both the preeruption and posteruption locations are more apparent for the locations when using the three-dimensional structure. The sharper images of earthquake distributions produced by the three-dimensional velocity model suggest that the locations are more accurate. The most significant difference was for the cluster of the earthquakes northwest of the summit area in the posteruption data. When the three-dimensional structure is used, locations of these earthquakes form a clear trend that dips outward from the summit area, matching a similar trend of earthquakes that dips outward on the east side of the summit (west-east cross sections in fig. 4). When the one-dimensional structure is used, the westward dipping trend is much more diffuse and harder to recognize. This area, where the largest difference between the locations calculated in the one- and three-dimensional structures exist, is also the region where the station coverage is the worst. Since most of the stations are to the east, the majority of ray paths from the earthquakes pass through the region of extreme velocity heterogeneity. This demonstrates that in an area where significant lateral variations in velocities can affect the traveltimes, the problems of locating earthquakes using a one-dimensional velocity structure are magnified when there is poor azimuthal station coverage.
The plots of earthquake locations shown and discussed in this volume by Harlow and others (this volume) and Mori, White, and others (this volume) use hypocenters located using the three-dimensional velocity structure.
Figure 4. Comparison of the all preeruption seismicity recorded from May 5 to June 11 located with a one-dimensional velocity structure (top) and a three-dimensional velocity structure (bottom). The cross-section shows all the events projected on a plane oriented N.45°W. Triangles are stations of the preeruption network.
Figure 5. Comparison of all the posteruption seismicity recorded from June 29 through August 16 located with a one-dimensional velocity structure (top) and a three-dimensional velocity structure (bottom). The cross-section is oriented west to east and shows events within a 10-km-wide strip through the summit region. Triangles are stations of the posteruption network.
We have attributed importance to the low-velocity regions seen in the cross sections of figure 3, and in order to ensure that these areas are well resolved by the velocity inversion, we examined the model resolution matrix from the inversion. The spread of values around the jth diagonal elements is an indication of the degree of resolution for the velocity value of the corresponding node. This spread (Sj) can be estimated by using the expression of Michelini and McEvilly (1991),
where skj are elements of the resolution matrix weighted by their distance (Dij) from the node point. Small values of the width indicate that the values surrounding the eigenvalue are relatively small and the velocity is well resolved. Higher values represent poorer resolution and indicate that there may be smearing of the velocity values among adjacent nodes. These values are shown in figure 6 for the velocity node points used to construct the cross sections of figure 3. The smaller spreads are shown by darker colors, representing better resolved velocities. The white areas are regions where few or no ray paths crossed, so there is little information about the velocity.
Figure 6. Estimates of the resolution for velocity values shown in figure 3, as measured by the spread in values about the diagonal elements of the model resolution matrix.
Figure 6 shows that the low-velocity regions are located in regions that are well resolved by the model, so the large size of the low-velocity region is not a result of smearing a small anomaly over a large volume.The model is weakest in the deeper portions, especially east and south of Mount Pinatubo, where there were few earthquakes to provide information. The lack of resolution in the deeper regions directly under the low-velocities areas leaves open the possibility that any inferred magma body could extend to greater depth and have larger volumes than we estimated.
Another uncertainty in the velocity model calculation is due to the coarseness of the grid (1 to 3 km). Small-scale features such as pipes or dikes that are a few hundred meters across cannot be resolved with the present data, because there are relatively few stations. Ideally, seismic monitoring on volcanoes would include sufficient station density not only for locating earthquakes but also for making three-dimensional velocity studies that can resolve the location of the magma. More S-wave data from three-component stations would be useful for incorporation into the inversion procedure and for identifying S-wave shadows (i.e. Matumoto, 1971) to confirm the existence of the large magma bodies.
The images in figures 3 and 4 show some strong low-velocity regions, particularly near and beneath the summit region. It is common to interpret low-velocity zones in volcanic areas as magma bodies; however, they could also be porous units of incompetent rock or unconsolidated sediments (Achauer and others, 1988; Evans and Zucca, 1988; Romero and others, 1993) or areas dominated by fluid saturation (Romero and others, 1993; O'Connell and Johnson, 1991). For the deeper low-velocity regions, we prefer the magmatic interpretation because it seems unlikely that unconsolidated material would be present at the lithostatic pressures associated with depths of 6 to 11 km. Vera and others (1990) observed a large low-velocity zone at 5 to 9 km in depth under the East Pacific Rise that they infer is not a magma chamber because the seismic amplitudes indicate that it behaves as a solid. Similar arguments might be made for Pinatubo, where S waves are sometimes observed for ray paths that cross the large low-velocity regions. Alternatively, these observations may indicate that the large magma bodies are not single reservoirs of magma but instead a region of dikes and sills intruded into competent rock. If this is the case, the volumes of the low-velocity bodies estimated above would be upper limits on the volume of magma.
Identifying the smaller low-velocity zone under the northwest flank of Mount Pinatubo as a magma body is more uncertain, because there could easily be unconsolidated sediments with low seismic velocities at this shallow depth. However, the formation of the small caldera (2.5 km diameter) during the June 15 eruption is consistent with the inference that there was a significant magma body here prior to the eruption. The caldera (2.5 km3 collapse volume, W.E. Scott and others, this volume) is centered about 1 km northwest of the old summit, almost directly over the shallow low-velocity zone on the northwest flank of Mount Pinatubo (see 1-km depth slice of fig. 2). This feature could have formed as a collapse in response to volume loss in the shallow magma body.
The pattern of seismicity is also suggestive that the low-velocity regions are magma bodies. Because the located seismicity consists mostly of high-frequency volcano-tectonic events, these events should have hypocenters in the areas of competent material. As the large magma bodies inflate or deflate, there would be significant volumetric strains surrounding the magma bodies, and one might expect most of the seismicity to occur in the regions of hard rock close to the magma bodies. The top portion of figure 3 shows that the preeruption seismicity occurred in two areas adjacent to the shallow low-velocity region under the northwest flank. The posteruption seismicity, shown in the bottom portion of figure 3, consists of outward-dipping limbs surrounding the strong low-velocity zone that is inferred to be the large magma chamber at depth. In both cross sections of figure 3, the earthquakes tend to occur in the areas of higher (blue) velocity, possibly indicating regions of competent rock.
Petrologic data indicate that most of the erupted material came from depths of 7 to 11 km below the surface (Rutherford and Devine, this volume; Pallister and others, this volume; Fournelle and others, this volume). This depth coincides with the large low-velocity region and suggests that the main source of the magma is this volume beneath the ridge that connects Mount Negron and Mount Pinatubo. This could be the main magma reservoir, with a volume of 40 to 90 km3, that supplies eruptions at Mount Pinatubo. The red Y-shaped pattern in the south-north cross section of figure 3 could be interpreted as a large magma reservoir at depth connected by two branches to volcanic vents at Mount Pinatubo and Mount Negron. If this system supplied the magma for past eruptions at both Mount Pinatubo and Mount Negron, it would have to be a long-enduring feature since the only available date for an eruption at Negron is 1.3 to 1.2 Ma (Delfin and others, this volume). The magma would have to travel a relatively long distance to reach the vents, but there is evidence from other volcanoes of large "plumbing" systems that can connect reservoirs and vents over long distances. One example is the 1912 eruption of Mount Katmai, Alaska, where the main eruption vent was 10 km from the summit collapse (Hildreth, 1983). Another example is the simultaneous eruptions at two vents 6 km apart at Rabaul caldera, Papua New Guinea, in 1937 (Fisher, 1939) and 1994.
The large volume of the inferred magma system (total of 60 to 125 km3) under Mount Pinatubo and Mount Negron suggests there is a potential of future eruptions being larger than the 1991 activity. Large magma reservoirs of this size may be responsible for caldera-forming eruptions, and we speculate that an eruption at Pinatubo could be large enough to reduce the volume of the magma body significantly. The ridge between Pinatubo and Negron could lose support from beneath and result in collapse around a ring fault to form a large caldera. There is a hint in the posteruption seismicity of a developing ring fault system surrounding the large magma reservoir (Mori, White, and others, this volume). Geological data, however, indicate that there is a recent trend of diminishing sizes in the Pinatubo eruptions (Newhall and others, this volume). Whether this system evolves into a large caldera event or continues with decreasingly smaller eruptions, the large size of the inferred magma reservoir indicates that eruptive activity will probably extend significantly into the future.
We inverted the P-wave arrival times for earthquakes around Mount Pinatubo to produce a three-dimensional velocity structure in the region around the volcano. The velocity structure enabled better locations for the earthquakes recorded during the preeruption and posteruption periods. Hypocenters calculated with the improved velocity model show more clearly dipping trends of earthquakes for both the preeruption and posteruption seismicity.
The velocity model was also used to infer the physical structure under the volcano, particularly in identifying low-velocity regions that are interpreted to be locations of magma bodies. A large region between depths of 6 and 11 km under the ridge that connects Mount Pinatubo and Mount Negron may be the magma reservoir (40 to 90 km3) that has supplied eruptions at these two vents. Another low-velocity region extends from the reservoir upward toward Mount Negron with a volume of 15 to 25 km3. A small (6 to 9 km3), shallow low-velocity region on the northwest flank of Mount Pinatubo is located under the crater that formed during the June 15 eruption. Volume loss in this shallow magma body may have initiated the collapse that formed the small summit caldera. We have inferred a large system of magma bodies under the Mount Pinatubo-Mount Negron complex with total volume estimated to be 60 to 125 km3. The deeper portions of the velocity model are poorly resolved, and the volume could be larger if the magma body extends below 11 km. This large volume of magma may be typical of volcanoes that produce eruptions of the size that occurred at Mount Pinatubo in 1991 and indicates there is the potential for future eruptions to be comparable or larger.
Data used in this study were the result of hard work by J.A. Power, T.L. Murray, F. Fischer, P. Okubo, R.A. White, L. Bautista, E.T. Endo, J.W. Ewert, A.B. Lockhart, J. Lockwood, J.N. Marso, A. Miklius, E.P. Laguerta, B. Bautista, C.G. Newhall, E. Ramos, E.W. Wolfe, and others. R.P. Hoblitt and J.S. Pallister provided many useful discussions. We thank U.S. Air Force personnel at Clark Air Base for their support. Helpful comments on the manuscript were provided by A. Pelayo, J. Lees, and C.G. Newhall.
Achauer, U., Evans, J.R., and Stauber, D.A., 1988, High-resolution seismic tomography of compressional wave velocity structure at Newberry Volcano, Oregon Cascade Range: Journal of Geophysical Research, v. 93, p. 10135-10147.
Aster, R.C., and Myer, R.P., 1988, Three-dimensional velocity structure and hypocenter distribution in the Campi Flegrei caldera, Italy: Tectonophysics, v. 149, p. 195-218.
Clawson, S.R., Smith, R.B., and Benz, H.M., 1989, P wave attenuation of the Yellowstone caldera from three-dimensional inversion of spectral decay using explosion source seismic data: Journal of Geophysical Research, v. 94, p. 7205-7222.
Delfin, F.G., Jr., Villarosa, H.G., Layugan, D.B., Clemente, V.C., Candelaria, M.R., Ruaya, J.R., this volume, Geothermal exploration of the pre-1991 Mount Pinatubo hydrothermal system.
Eberhart-Phillips, D., 1993, Local earthquake tomography: Earthquake source regions, in Iyer, H.M., and Hirahara, K., eds., Seismic tomography: Theory and practice: London, Chapman and Hall, p. 613-643.
Evans, J.R. and Zucca, J.J., 1988, Active high-resolution seismic tomography of compressional wave velocity and attenuation structure at Medicine Lake Volcano, Northern California Cascade Range: Journal of Geophysical Research, v. 93, p. 15016-15036.
Fisher, N.H., 1939, Geology and vulcanology of Blanche Bay, and the surrounding area, New Britain: Territory of New Guinea Geological Bulletin, v. 1, 68 p.
Fournelle, J, Carmody, R., and Daag, A.S., this volume, Anhydrite-bearing pumices from the June 15, 1991, eruption of Mount Pinatubo: Geochemistry, mineralogy, and petrology.
Harlow, D.H., Power, J.A., Laguerta, E.P., Ambubuyog, G., White, R.A., and Hoblitt, R.P., this volume, Precursory seismicity and forecasting of the June 15, 1991, eruption of Mount Pinatubo.
Hildreth, W., 1983, The compositionally zoned eruption of 1912 in the Valley of Ten Thousand Smokes, Katmai National Park, Alaska: Journal of Volcanology and Geothermal Research, v. 18, p. 1487-1494.
Iyer, H.M., 1992, Seismological detection and delineation of magma chambers: Present status with emphasis on the western USA, in Gasparini, P, Scarpa, R., and Aki, K., eds., Volcanic seismology: Berlin, Springer-Verlag, p. 299-338.
Lees, J.M., 1992, The magma system of Mount St. Helens: Non-linear high-resolution P-wave tomography: Journal of Volcanology and Geothermal Research, v. 53, p. 103-116.
Lockhart, A.B., Marcial, S., Ambubuyog, G., Laguerta, E.P., and Power, J.A., this volume, Installation, operation, and technical specifications of the first Mount Pinatubo telemetered seismic network.
Matsumoto, T., 1971, Seismic body waves observed in the vicinity of Mt. Katmai, Alaska and evidence for the existence of molten chambers: Geological Society of America Bulletin, v. 82, p. 2905-2920.
Michelini, A., and McEvilly, T.V., 1991, Seismological studies at Parkfield: I. Simultaneous inversion for velocity structure and hypocenters using cubic B-splines parameterization: Bulletin of the Seismological Society of America, v. 81, p. 524-552.
Mori, J., White, R.A., Harlow, D.H., Okubo, P., Power, J.A., Hoblitt, R.P., Laguerta, E.P., Lanuza, L., and Bautista, B.C., this volume, Volcanic earthquakes following the 1991 climactic eruption of Mount Pinatubo: Strong seismicity during a waning eruption.
Murray, T.L., Power, J.A., March, G.D., and Marso, J.N., this volume, A PC-based realtime volcano-monitoring data-acquisition and analysis system.
Newhall, C.G., Daag, A.S., Delfin, F.G., Jr., Hoblitt, R.P., McGeehin, J., Pallister, J.S., Regalado, M.T.M., Rubin, M., Tamayo, R.A., Jr., Tubianosa, B., and Umbal, J.V., this volume, Eruptive history of Mount Pinatubo.
O'Connell, D.R.H. and Johnson, L.R., 1991, Progressive inversion for hypocenters and P wave and S wave velocity structure: Application to the Geysers, California, geothermal field: Journal of Geophysical Research, v. 96, p. 6223-6236.
Romero., A.E., Jr., McEvilly, T.V., Majer, E.L., and Michelini, A., 1993, Velocity structure of the Long Valley caldera from the inversion of local earthquake P and S travel times: Journal of Geophysical Research, v. 98, p. 19869-19879.
Rutherford, M.J., and Devine, J.D., this volume, Preeruption pressure-temperature conditions and volatiles in the 1991 dacitic magma of Mount Pinatubo.
Pallister, J.S., Hoblitt, R.P., Meeker, G.P., Knight, R.J., and Siems, D.F., this volume, Magma mixing at Mount Pinatubo: Petrographic and chemical evidence from the 1991 deposits.
Sanders, C., 1984, Location and configuration of magma bodies beneath Long Valley, California, determined from anomalous earthquake signals: Journal of Geophysical Research, v. 89, p. 8287-8302.
Scott, W.E., Hoblitt, R.P., Torres, R.C., Self, S, Martinez, M.L., and Nillos, T., Jr., this volume, Pyroclastic flows of the June 15, 1991, climactic eruption of Mount Pinatubo.
Thurber, C.H., 1981, Earth structure and earthquake locations in the Coyote Lake area, central California: Cambridge, Mass., Massachusetts Institute of Technology, Ph.D. thesis.
------1983, Seismic detection of the summit magma complex of Kilauea volcano, Hawaii: Science, v. 233, p. 165-167.
------1993, Local earthquake tomography: Velocities and Vp/Vs--theory, in Iyer, H.M., and Hirahara, K., eds., Seismic tomography: Theory and practice: London, Chapman and Hall, p. 563-583.
Um, J., and Thurber, C.H., 1987, A fast algorithm for two-point seismic ray tracing: Bulletin of the Seismological Society of America, v. 77, p. 972-986.
Vera, E.E., Mutter, J.C., Buhl, P., Orcutt, J.A., Harding, A.J., Kappus, M.E., Detrick, R.S., and Brocher, T.M., 1990, The structure of 0- to 0.2-m.y.-old oceanic crust at 9°N on the east Pacific Rise from expanded spread profiles: Journal of Geophysical Research, v. 95, 15529-15556.
PHIVOLCS | University of Washington Press | U.S.Geological Survey
This page is <https://pubs.usgs.gov/pinatubo/mori2/>
Contact: Chris Newhall
Last updated 06.11.99