SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data Acquisition and Processing
  5. 3. Analysis of Velocity, Vp/Vs, and VpVs Logs
  6. 4. Conclusion
  7. Acknowledgments
  8. References

[1] A vertical seismic profiling experiment has been conducted at the DOE well LVF 51–20 at the resurgent dome at Long Valley caldera, California. The current paper is intended to report the velocity results derived from this experiment to seismically constrain the area around the resurgent dome. We report the results of velocity logs, Vp/Vs ratios, and VpVs products for the depth interval from 880–1980 m. The Bishop tuff can be divided into three zones (top-to-bottom) of intermediate velocities (Vp ≈ 4.0 km/s, Vs ≈ 2.3 km/s), followed by higher velocities (Vp = 4.7 km/s, Vs = 2.9 km/s) in the lower-central part, and a decrease back to the lower values at the bottom of the sequence. The analysis of VpVs data suggested a decrease in porosity in the lower-central section of the tuff as a possible explanation, which was supported by direct density measurements from a nearby location.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data Acquisition and Processing
  5. 3. Analysis of Velocity, Vp/Vs, and VpVs Logs
  6. 4. Conclusion
  7. Acknowledgments
  8. References

[2] The Long Valley caldera (LVC) is a Quaternary volcanic structure, located at the intersection of the Sierra Nevada frontal fault escarpment and the western margin of the Basin and Range tectonic province. A map view of the LVC region is presented in Figure 1.

image

Figure 1. Map of Long Valley region adopted from USGS. The location of LVF 51–20 is indicated by the circle inside the resurgent dome.

Download figure to PowerPoint

[3] In 1980, strong evidence of renewed magma movement became apparent following a series of large magnitude earthquake sequences accompanied by rapid surface uplift. While several geophysical studies suggest the existence of a magma chamber beneath the resurgent dome [i.e., Hill, 1976], other investigations do not support the presence of a sizable magma body at shallow to mid-crustal depth beneath LVC [Kissling, 1988]. The latter study relates the encountered lower mid-crustal velocities to hydrothermal alteration and/or extensive fractures.

[4] Most of the seismological studies were performed as surface refraction experiments [Hill et al., 1985], facing the difficulties of two-way wave propagation through the highly attenuative non-welded tuff and interpretations of the reflected seismic energy with little knowledge about the actual subsurface velocities.

[5] In an attempt to determine a better velocity model, among other goals, the Lawrence Berkeley National Laboratory (LBNL) conducted a vertical seismic profiling (VSP) experiment within LVC under the Department of Energy's Basic Energy Science and Continental Scientific Drilling program. The experiment was conducted at the Long Valley Federal 51–20 well (LVF 51–20, also referred to as LVEW) after completion of the second drilling phase and is the only deep VSP performed at LVC to date. The location of the well lies within the center of the resurgent dome (Figure 1). The borehole measurements have the advantage of improved signal-to-noise ratio, because of the quieter environment and because of one-way wave propagation through the highly attenuating surface layers. The current paper is intended to report the results of the velocity analysis, including velocity ratio (Vp/Vs) and velocity product (VpVs), to provide local velocity constrains within the resurgent dome for future seismic studies of LVC.

2. Data Acquisition and Processing

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data Acquisition and Processing
  5. 3. Analysis of Velocity, Vp/Vs, and VpVs Logs
  6. 4. Conclusion
  7. Acknowledgments
  8. References

[6] The seismic source of the VSP experiment was a P- vibroseis truck, located 165 m from the well, exciting sweeps 12 seconds long with frequency content of 10 Hz to 58 Hz. The data were acquired with a sample rate of 2 ms between 550 m and 2072 m depth at 15 m intervals. Shallower data acquisition was prevented by multiple borehole casing, which degraded data quality. The receiver was a three-component, high-temperature, hydraulic wall-locking borehole seismometer. The corner frequency of the geophones was 14 Hz with a flat response up to 500 Hz.

[7] Data processing consisted of noise editing, trace stacking to improve the signal-to-noise ratio, and correlating the traces with the source sweep. Strong P-to-S-wave conversions in the shallow subsurface produced sufficient S-wave arrivals on the seismograms, such that P- and S-wave data could be analyzed.

[8] A numerical rotation of the geophone components was applied to maximize the depth coherence of P- and S-waves [Daley et al., 1988]. The rotated data are presented in Figure 2, separated into radial and transverse components. The quality of the shallowest traces is adversely effected by scatter and ringing in the waveforms, caused by problems with the cement bond of the well casing. Therefore, the following velocity analysis will be performed over a depth interval from 880–1980 m and 980–1980 m for P- and S-waves, respectively.

image

Figure 2. Seismic data after correlation and rotation into the wavefront coordinate system. In the new coordinate system, the components correspond to radial and transverse components. The upper 16 traces show strong reverberations in the wave forms caused by problems with the cement bond of the well casing.

Download figure to PowerPoint

3. Analysis of Velocity, Vp/Vs, and VpVs Logs

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data Acquisition and Processing
  5. 3. Analysis of Velocity, Vp/Vs, and VpVs Logs
  6. 4. Conclusion
  7. Acknowledgments
  8. References

[9] The data processing was followed by the determination of the first arrival times for the direct P- and S-waves. Assuming straight ray geometry between the source and receiver positions in the well, interval velocities for every receiver location were computed. The arrival times were determinated by cross-correlation technique yielding subsample accuracy. This was achieved by first cross-correlating the traces one sample at a time, followed by taking their ratio in the frequency domain and determining the slope of the resulting phase spectrum at low frequencies [Nadeau et al., 1994]. The slope was then added to the shift determined in the time domain to gain subsample accuracy. Because some scatter remained in the resulting velocities estimates, they were subsequently averaged over a 10-point running window of 150 m length that was comparable to the dominant compressional- and shear wavelength of λp = 160 m and λs = 125 m, respectively. The estimated uncertainty of the arrival times is ±1 ms for P-waves and ±3 ms for S-waves over the 10-point average. These uncertainties were subsequently converted to upper and lower bounds for the velocity estimates.

3.1. P- and S-Wave Velocities

[10] The velocity-depth estimates are shown in the left panel of Figure 3. Within the Bishop tuff, the velocity profiles can be subdivided into three sections as indicated by the dashed horizontal lines. The upper and lower zones reveal intermediate velocities (Vp = 3.5–4.5 km/s, Vs = 2.0–2.7 km/s), while the central section reveals higher velocity values (Vp = 4.7 km/s, Vs = 2.9 km/s). At the bottom of the tuff, the transition to metavolcanics and the underlying metasediments is reflected by an increase in P- and S-wave velocities to Vp ≥ 5.0 km/s and Vs ≥ 3.0 km/s, respectively. Previous seismic studies in this area produced similar results. [Romero et al., 1993] conducted Vp and Vs tomography studies using local earthquake travel time data, and reported, for the area around the current VSP well, velocity estimates of Vp = 2.7 km/s at 0 km depth and Vp = 5.4 km/s at 2 km depth. Interpolation between these estimates would yield value of Vp = 4.05 km/s at 1 km depth. Similarly, the reported S-wave velocity estimates are Vs = 1.7 km/s at 0 km depth, Vs = 3.0 km/s at 2 km depth, and an interpolated value of Vs = 2.35 km/s at 1 km depth. These values are well in accordance with the velocity estimates reported in the present study. In addition, [Hill et al., 1985] conducted seismic P-wave refraction studies across the resurgent dome and reported velocity estimates of Vp = 3.9–4.1 km/s at 1 km depth and Vp = 4.9–5.2 km/s at 2 km depth, which again agrees with the present findings. The depth distribution of alteration minerals, presented in the right column in Figure 3, shows an even distribution of minerals throughout the tuff section, which precludes the mineralogy as the leading cause for the velocity variations. The transition to metavolcanic rocks at the base of tuff, however, is characterized by an increase in both P- and S-wave velocity, which is caused by a change in mineralization, primarily the presence of Chlorite and Epidote, with velocities of Vp = 6.75–7.1 km/s for the former and Vp = 6.43 km/s and Vs = 4.2 km/s for the latter [Carmichael, 1982]. The velocity distribution within the Bishop tuff is caused by the depositional and cooling history of tuffs, which produces significant variations in density and porosity, as we will address in a later section.

image

Figure 3. Velocity, velocity-ratio, and velocity-product logs for the DOE exploratory well LVF 51–20. The upper and lower bounds are indicated by the thin dashed lines. The dashed horizontal lines are intended to divide the velocity logs into sections of similar velocities (see text). The stratigraphic and mineralogic columns, reproduced from McConnell et al. [1995], are based on data from core analyses and televiewer information. Qer: Postcaldera volcanics (rhyolite, vesicular pumice, welded tuff); Qbt: Bishop tuff; Qer Int.: dike intrusions (rhyolite); Mb: mixed breccia; Mmv: Mesozoic metavolcanics.

Download figure to PowerPoint

3.2. Vp/Vs-Ratio

[11] The interpretation of seismic velocity logs is non-unique, because many rocks with different physical states have similar seismic velocities. Factors such as fracturing, pore pressure, fluid saturation, and partial melt affect seismic velocities. Therefore, it is useful to consider velocity ratios to differentiate between the different physical effects in the medium. Moos and Zoback [1983] show that fluid saturated fractures decrease both Vp- and Vs-velocities, while it increases the Vp/Vs-ratio. Dvorkin et al. [1999] demonstrated that for water saturated rocks Vp/Vs values increase with increasing pore pressure (i.e., differential pressure decreases), while for dry rocks Vp/Vs values decrease with increasing pore pressure. Sanders et al. [1995] investigate Vp/Vs-ratios from seismological observations at LVC and conclude that high values suggest the existence of high temperatures and perhaps partial melt or an increase in fracture concentration in saturated rocks. Low Vp/Vs-ratios, in contrast, are consistent with supercritical fluids and the presence of gas phases.

[12] The Vp/Vs-ratio, derived from the velocity measurements, is shown in the central panel in Figure 3. The average value over the depth interval of 1.69 ± 0.09 is slightly lower than the crustal average of 1.73. The upper and middle zones of the Bishop tuff from 900–1520 m show only small variations in Vp/Vs-ratio, while the P- and S-wave velocities increase simultaneously. In the lower section of the tuff (1520–1750 m), however, the Vp/Vs-ratio increases, which is caused by the stronger decrease in S-wave velocity relative to the P-wave velocity over the same depth interval. Partial melt is unlikely the cause for the high ratio, because the temperature profile in the borehole, shown in Figure 4, does not reveal an increased temperature gradient for this depth interval. Furthermore, the shaded region of the temperature profile, representing the depth interval of the velocity estimates, shows an almost constant gradient of 30°C/1000 m throughout the Bishop tuff, while the values at the bottom of the profile indicate a slightly higher gradient within the basement rocks. In contrast, the temperature increases at 550 and 750 m depth are likely associated with zones of lateral hydrothermal fluid flow delineated by Sorey et al. [1991]. In the absence of temperature anomalies around the well, the likely cause for the Vp/Vs high is a fluid saturated zone of high porosity or fracture concentration. Similar observations were reported by Lees and Wu [2000] for fluid saturated production zones at the Coso geothermal field, California.

image

Figure 4. Temperature log of LVF 51–20. The grey shaded area indicates the depth range of the current velocity analysis.

Download figure to PowerPoint

3.3. VpVs-Product

[13] The VpVs-product of seismic velocities has been used to determine porosity in volcanic settings, where it was observed that VpVs decreases with increasing porosity [Lees and Wu, 2000]. The estimation of VpVs-values is particularly useful in geothermal areas, where the understanding of fluid flow depends more on porosity than lithological settings. In the following section we attempt to relate VpVs to the porosity profile of the Bishop tuff.

[14] The VpVs-product is displayed in the right panel in Figure 3. The trend of VpVs increases in the top half of the depth section, peaking in the lower-central part, before it decreases again towards the base of the tuff sequence. The increase below 1650 m is caused by higher velocity values in the underlying basement rocks that are mapped in the tuff by the length of the smoothing window (150 m). The trend of the VpVs-values suggests higher porosity values at the top and base of the tuff, while the center appears more dense. To investigate this possibility a density profile of the Bishop tuff was plotted in the left column of Figure 5. The density data was measured on exposed sections of the Bishop tuff in Owens Gorge outside the boundary of LVC about 20 km south-east of VSP well [Ragan and Sheridan, 1972]. Therefore, the thickness of the tuff in Figure 5 is reduced relative to its thickness at the location of the VSP well near the center of the eruption within LVC. Nevertheless, the cooling history of the tuff is very similar in both locations, such that the general trend of the data in Figure 5 can be used as an analog for the tuff layer in the vicinity of the VSP well. In order to estimate porosity from the density profile, we use the bulk density ρb of a fully water saturated rock, with porosity ϕ, matrix density ρm, and water density ρw as given by

  • equation image

Using this equation the proposity ϕ can be expressed as

  • equation image

Based on this equation, the density values in Figure 5 can be converted to porosity using the density of water and the matrix as ρw = 1.0 g/cm3 and ρm = 2.5 g/cm3, respectively. The value for the matrix density was chosen to represent the grain density of tuff. The resulting porosity profile in Figure 5 mimics the reverse trend of the VpVs in Figure 3. The lower-central section of the tuff sequence shows the lowest porosity with values of ≈5%. The density/porosity profile was created by different cooling regimes within the tuff. The top and bottom cooled fastest after the initial eruption, such that welding could not take effect as efficiently as in the central section of the tuff layer, which created a more compact layer. The reason for the displacement of the low porosity/high density zone below the center of the tuff flow may be explained by faster cooling rates from the top [Ragan and Sheridan, 1972].

image

Figure 5. Density and porosity depth profile for the Bishop tuff at Owens Gorge, California. The density profile was measured by [Ragan and Sheridan, 1972], while the porosity profile was calculated using equation (2) and the density profile (see text).

Download figure to PowerPoint

[15] The increase in VpVs in Figure 3 correlates well with the presence of the metavolcanics at the bottom of the profile, suggesting an additional reduction in porosity relative to the Bishop tuff.

4. Conclusion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data Acquisition and Processing
  5. 3. Analysis of Velocity, Vp/Vs, and VpVs Logs
  6. 4. Conclusion
  7. Acknowledgments
  8. References

[16] The intention of this paper was to report the velocity estimates obtained during the VSP experiment at LVF 51–20, to better constrain the velocity model in the vicinity of the resurgent dome for future seismic investigations. The acquired data yielded P- and S-wave velocity results between 880–1980 m and 980–1980 m, respectively. The determination of Vp/Vs and VpVs from the velocities helped to constrain the interpretation of the data. Nevertheless, a more thorough investigation of the results could quantify the interpretation of the data in the vicinity of the well by reproducing the observed seismic data through numerical modeling using all available physical parameters. However, such investigations are beyond the intended scope of this paper.

[17] The available temperature log did not indicate the presence of an anomalous temperature zone in the vicinity of the well between 900 and 2000 m depth. The combination of velocity- and VpVs-logs suggested a region of decreased porosity throughout the lower-central densely-welded part of the Bishop tuff with higher porosity estimates above and below. These findings correlate well with density measurements of the tuff from locations along the boundary of LVC. The lower section of the Bishop tuff may be additionally fractured as indicated by an increase in Vp/Vs. The metavolcanics at the base of the survey are represented by an increase in both P- and S-wave velocities and a further reduction in density. The current analysis indicated a degree of heterogeneity throughout the Bishop tuff, similar to that of ash flows in other areas, which warrants further studies of the physical parameters to obtain a better understanding of the eruption processes and the depositional history of LVC.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data Acquisition and Processing
  5. 3. Analysis of Velocity, Vp/Vs, and VpVs Logs
  6. 4. Conclusion
  7. Acknowledgments
  8. References

[18] This research was supported by the Director, Office of Energy Research, Division of Basic Energy Sciences, Engineering, and Geosciences, of the U.S. Department of Energy under contract DE-AC03-76SF00098.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data Acquisition and Processing
  5. 3. Analysis of Velocity, Vp/Vs, and VpVs Logs
  6. 4. Conclusion
  7. Acknowledgments
  8. References