SEARCH

SEARCH BY CITATION

Keywords:

  • Montserrat

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data
  5. 3. Interpretation
  6. 4. Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

[1] The SEA-CALIPSO experiment was designed to image structure related to active volcanism beneath the island of Montserrat in the Caribbean. As part of that experiment, over 200 Texan recorders with 5 Hz geophones were deployed in 3 linear arrays at a nominal spacing of 100m, primarily to record an airgun source towed offshore around the island. Because the recorders were operating in continuous mode for three days, a number of shallow microearthquakes under the active summit of Soufriere Hills Volcano (SHV) were also recorded. 20 events were sufficiently well recorded and located that they could be used to identify and map reflections from deep subsurface structure. Here we report on the processing of those recordings as multichannel CMP reflection sources, with emphasis on careful statics correction and coherency enhancement. The resulting reflection images indicate subhorizontal layering at depths between 6 and 19km which we interpret as sills.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data
  5. 3. Interpretation
  6. 4. Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

[2] Reflections from crustal discontinuities identified on microearthquake seismograms have provided critical evidence of magma at depth at a limited number of sites around the world. Sanford and Long [1965] first reported on a midcrustal shear wave reflector beneath the Rio Grande Rift near Socorro, New Mexico, that was subsequently interpreted as an extensive magma sill [e.g., Sanford et al., 1973, 1977] . The distribution of such reflecting points suggest a lateral extent of at least 3400 km2 for the magma body [Balch et al., 1997]. Microearthquake recordings have been used in Japan to identify anomalous S-wave reflectors, likewise interpreted as magma bodies, beneath several volcanic complexes, typically at depths less than 20 km [e.g., Hasegawa et al., 1991; Inamori et al., 1992; Iidaka et al., 1993]. James and Savage [1990] used recordings at a single seismic station of multiple microearthquake sources to identify crustal reflectors at depths of 10–11 km and 13–14 km (Moho?) beneath the Big Island of Hawaii. In that study, 29 events were source migrated to remove the effects of varying focal depths [James and Savage, 1990].

[3] Anomalously strong reflections, or bright spots, have also been identified from deep seismic reflection surveys using controlled surface sources. Most of these bright spots have been interpreted as due to magma, with the large amplitudes attributed to liquid material in juxtaposition with solid country rock. Examples of such magma bright spots include the aforementioned Socorro Bright Spot [Brown et al., 1979], the Death Valley Bright Spot [de Voogd et al., 1986] and the Tibetan Bright Spots [Brown et al., 1996]. Amplitude anomalies attributable to deep magmatism have also been recognized on teleseismic converter imagery (receiver functions) at an increasing number of locations around the world [e.g., Sheetz and Schlue, 1992; Chmielowski et al., 1999]. The prominent expression of magma as a reflection/conversion on these relatively high resolution seismic recordings was one inspiration for the SEA-CALIPSO experiment.

[4] The SEA-CALIPSO experiment is described in the companion articles in this issue. The focus of this paper are recordings made by the dense reflection spreads consisting of over 200 Texan recorders equipped with 5Hz geophones. These deployments constituted three lines (Figure 1), two effectively radiating NW and N from SHV and the third providing fancoverage for sources on and SE of SHV. These arrays were designed in part to undershoot SHV with airgun sources on the RSS James Cook (e.g., L. D. Brown et al., manuscript in preparation, 2010). Although continuous, as opposed to windowed, recording was primarily imposed by the regular nature of the airgun source (one shot every 60 seconds), such recording also allowed the recording of natural sources, in particular a number of microearthquakes that occurred near the summit of SHV. These earthquake recordings, processed using a selected subset of traditional multichannel reflection techniques, provide the most substantive indications of crustal structure near SHV.

image

Figure 1. (left) Map of Montserrat showing the locations of the Texan seismic arrays (triangles), along with the best located microearthquakes used in this study (stars). The CDP reflection points corresponding to the Belham Valley recordings of a typical event are shown as circles. (right) Schematic cross-section illustrating depths of the sources relative to the recording spread, together with a resulting image (source gather).

Download figure to PowerPoint

2. Data

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data
  5. 3. Interpretation
  6. 4. Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

[5] The microearthquakes were recorded from Dec 17, 2007 through Dec 20, 2007. A total of twenty local events were identified from the continuous data recording from these three days and correlated with events located by Miller (personal communication, 2009) from the areal seismic network. Locations were made with HypoEllipse [Lahr, 1999] by using a multi-layer 1-D approximation of the SEA-CALIPSO velocity model developed by Shalev et al. [2010]. The epicenters were centered near the summit of SHV at relatively shallow depth, thus providing near vertical reflection coverage for depth points relatively close to SHV (Figure 1). The microearthquakes were treated in the same manner as borehole shots in a conventional controlled source profile. Although these events were recorded along all three profiles, attention here is focused on those associated with the Belham Valley line, as they sample most closely to the SHV (Figure 1).

[6] One challenge using natural sources is the inherent uncertainty in the location and origin times of the sources. To reduce the influence of inaccuracies in depth, especially for stacking purposes, the events were separated into three different groups based upon reported precision in the location. For events that had a horizontal location error less than one kilometer, the reported location was accepted for processing purposes. The seven events in this category (A) are shown explicitly in Figure 1. Events with horizontal location errors greater than one kilometer but less than 2.5 kilometers were grouped together as category B. An average location was used for processing these seven events. Events with horizontal location errors greater than 2.5 kilometers were assigned to category C. The six events in this category were visually inspected for consistency with the other earthquake recordings, but were not used for joint processing (e.g., stacking) due the relatively large location uncertainties.

[7] The raw earthquake gathers (e.g., Figure 2) all show clear indications of organized energy that cannot be attributed to direct P, S or surface wave energy, but rather suggest moveout consistent with reflected arrivals. However, individual reflections are difficult to trace undisrupted across the array, which we suspected was due to relative static shifts associated with the overlying crust. While conventional seismic processing may draw upon an extensive library of signal enhancement tools, the relatively limited nature of the data here and the uncertainties with source location mentioned above, led us to focus a few basic procedures. These procedures are illustrated in (Figure 2). Starting with the raw data, elevation statics were applied to correct for changes in topography along the seismic lines, then the data were bandpass filtered from 1 to 8 Hz. To further improve reflection coherence, we applied a form of refraction statics. First P wave arrivals were aligned to near -horizontal using linear moveout (LMO) corrections. Deviations of the first arrival time from horizontal were manually picked and used to apply a static shift to force alignment of the first arrival. The LMO correction was subsequently removed, hopefully with increased lateral alignment of reflections as well as first arrivals. A normal moveout correction (NMO) was then applied using an average velocity of 5km/s from 0s to 10s to image reflection geometry at depth. Several additional coherency enhancement techniques were used to further increase the visibility of reflected energy. These included FX-deconvolution, trace mixing, and FK-filtering. Various combinations of these routines were applied to find the combination that best improved the visual coherency of the reflectors; the most effective appeared to be application of FX-decon twice followed by a single pass trace mix (Figure 2). Although more quantitative measures of improved coherence were considered, we felt that they were essentially circular with the enhancement techniques involved and less suited to discrimination of reflections from coherent noise than the interpreter's eye.

image

Figure 2. Example microearthquake gather illustrating the processing steps used to enhance possible deep reflections. (a) Raw data, (b) data with bandpass filter and elevation statics, (c) alignment using first arrivals and linear moveout, (d) display with NMO and (e) NMO, FX-decon and trace mix (applied twice).

Download figure to PowerPoint

[8] As cited above, seven of the recorded microearthquakes provided locations with particularly low error estimates. However, due to technical issues with the recording equipment on the Belham Valley line for a portion of the survey, not all of the data for all seven events were recovered. In spite of this loss, the processed gathers for all seven events (Figure 3) show strong similarities, e.g. subhorizontal reflectivity, despite being recorded for different earthquake sources. It is unclear whether one can defend a reflection-for-reflection correlation between the various gathers. However such a detailed correspondence should not be expected given the fact that the reflecting points in the subsurface differ for each event, not to mention possible variations in arrival time due to location uncertainties. Moreover, as is often the case with conventional reflection data, a degree of interpreter subjectivity is inherent: weaker events may be judged more significant than strong events when considered in context. Here we simply assert that the overall similarity in reflectivity argues that geological layering at a common depth, rather than noise, is being imaged. Note that the microearthquake gathers in Figure 3 are displayed in depth using a provisional average upper crustal velocity of 5 km/s.

image

Figure 3. Source gathers for the 7 best located earthquakes, as recorded along the Belham Valley line. Differences in the number of traces in each gather are due to failure of a subset to the array to operate during this time period.

Download figure to PowerPoint

[9] Before addressing the interpretation of the apparent reflectivity on these gathers, we must consider the potential artifacts that are inherent in such heavy use of algorithms such as FX-decon and trace mixing. This combination can be expected to emphasize sub-horizontal energy versus dipping energy, and linear events versus curved events (e.g., diffractors). There are certainly indications of diffractions from presumptively small bodies, such as intrusions, on some of the unprocessed shot gathers (e.g. arrow in Figure 2c). A more serious concern is that the heavy use of these multichannel filters is generating spurious coherence out of background noise or very weak arrivals (direct or multiple) from the airgun sources. To assess these possibilities, we processed background gathers consisting of data from the minute preceding and the minute following the 60 sec interval beginning with the origin time of a sample earthquake. These noise gathers were then processed with exactly the same sequences and parameters used for the earthquake record they bracket. These gathers contain a similar signal from the airgun that would appear on the earthquake record, but for our purposes the gathers would be considered noise. These noise gathers were then processed in the exact same way as the corresponding earthquake using the FX-decon and the trace mix, and compared in true amplitude format with the earthquake gathers. As shown in the auxiliary material, these noise gathers do not replicate the key features of the microearthquake gathers and thus we are confident that the coherent energy evident in Figures 24 originates from the microearthquake sources.

image

Figure 4. CDP stack for the Belham Valley line. (left) No coherency filter. (middle) FX-decon with two iterations of trace mixing. (right) For comparison, an individual earthquake gather with identical processing. Arrows indicated reflections on the CDP stack that seem to correlate with events on the gathers.

Download figure to PowerPoint

[10] Most of the processing focused on signal enhancement within individual source gathers (Figures 2 and 3), in part due to concerns over source location uncertainties but also because the reflection points for each earthquake scatter laterally from each other because the sources themselves are not co-linear with the recording spread. However, we did attempt to stack the Belham Valley recordings. Because of the lateral dispersal of the common reflecting (mid) points, we restricted our stack to those events whose common reflections points were close to co-linear. The resulting low fold (up to 5) stack is shown in Figure 4, with and without coherency enhancement. Given the uncertainties cited above, it is not surprising that the stack shows little, if any, improvement relative to the individual shot gathers. This exercise is perhaps more useful as a feasibility demonstration that CMP methods can be applied to microearthquake recordings, while emphasizing the need for more sources and better source locations.

3. Interpretation

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data
  5. 3. Interpretation
  6. 4. Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

[11] The individual earthquake gathers from the Belham Valley line (Figure 3) show consistent subhorizontal energy between 6 and 19km depth beneath the NW flank of Soufriere Hills Volcano. On the Belham Valley CDP stack (Figure 4), the stronger reflectors are at similar depths to the reflectors seen on the individual gathers. As is always the case with 2D surveys, some of this energy may be arriving from off line sources (e.g., sideswipe). However without 3D coverage, this issue cannot be usefully addressed here. These amplitudes, which need considerable enhancement just to be visible, provide little support for their interpretation as fluid bodies at depth One could argue that intrinsic attenuation in the crust near SHV might render even a bright spot reflection as a relatively weak arrival, but this is mere speculation. Given the quality of the unprocessed data, attempts to identify the polarity of these reflectors have been unfruitful. We are thus left with the conjecture that these reflectors represent either buried volcanic layering and/or later sills intruded into the crustal edifice. We prefer the sill interpretation, primarily because they would be more likely to provide the needed impedance contrasts to generate detectable reflections from depth. Receiver functions (W. I. Sevilla et al., manuscript in preparation, 2009) suggest that the Moho lies near 30 km, much deeper than any of the prominent reflections indicated by these reflection images.

4. Conclusions

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data
  5. 3. Interpretation
  6. 4. Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

[12] The presence of sill-like features in the upper crust beneath an active volcano is not surprising. The primary value of this study is its demonstration that relatively high resolution reflection imaging of crustal structure is feasible using microearthquake sources. In regions near active volcanoes, like SHV, placement of controlled sources near the summit, as required to obtain near vertical illumination of subvolcanic structure, is problematic at best. Microearthquakes offer an alternative in such situations, requiring only dense array recording in continuous mode over long periods of time (compared, at least, to a typical controlled source survey). As we have shown, the data can be processed with conventional reflection techniques. A major attraction of using natural sources is the inherent potential for time lapse imaging of subvolcanic features without the prohibitive cost of controlled sources for multiple surveys. The efficacy of this approach, however, is dependent upon the quality of relative event locations and the number and distribution of such sources in space.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data
  5. 3. Interpretation
  6. 4. Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

[13] This research was funded by the Continental Dynamics Program of the National Science Foundation (grant EAR-060772). The authors would like to thank the entire SEA-CALIPSO research team for their contributions in the field and in discussion.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data
  5. 3. Interpretation
  6. 4. Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information
  • Balch, R., H. Hartse, A. Sanford, and K. Lin (1997), A new map of the geographic extent of the Socorro midcrustal magma body, Bull. Seismol. Soc. Am., 87, 174182.
  • Brown, L. D., P. A. Krumhansl, C. E. Chapin, A. R. Sanford, F. A. Cook, S. Kaufman, J. E. Oliver, and F. S. Schilt (1979), COCORP seismic reflection studies of the Rio Grande Rift, in Rio Grande Rift: Tectonics and Magmatism, edited by R. E. Riecker, pp. 169184, AGU, Washington, D. C.
  • Brown, L. D., W. Zhao, K. D. Nelson, M. Hauck, D. Alsdorf, A. Ross, M. Cogan, M. Clark, X. Liu, and J. Che (1996), Bright spots, structure, and magmatism in southern Tibet from INDEPTH seismic reflection profiling, Science, 274(5293), 16881690, doi:10.1126/science.274.5293.1688.
  • Chmielowski, J., G. Zandt, and C. Haberland (1999), The central Andean Altiplano-Puna magma body, Geophys. Res. Lett., 26(6), 783786, doi:10.1029/1999GL900078.
  • de Voogd, B., L. Serpa, L. Brown, E. Hauser, S. Kaufman, J. Oliver, B. W. Troxel, J. Willemin, and L. A. Wright (1986), Death Valley bright spot: A midcrustal magma body in the southern Great Basin, California? Geology, 14(1), 6467, doi:10.1130/0091-7613(1986)14<64:DVBSAM>2.0.CO;2.
  • Hasegawa, A., D. Zhao, S. Hori, A. Yamamoto, and S. Horiuchi (1991), Deep structure of the northeastern Japan arc and its relationship to seismic and volcanic activity, Nature, 352, 683689, doi:10.1038/352683a0.
  • Iidaka, T., K. Miura, and A. Ikami (1993), Evidence for the existence of a mid-crustal reflector in the Beppu-Shimabara Graben, Kyushu, Japan, Geophys. Res. Lett., 20(16), 16991702, doi:10.1029/92GL02800.
  • Inamori, T., S. Horiuchi, and A. Hasegawa (1992), Location of mid-crustal reflectors by a reflection method using aftershock waveform data in the focal area of the 1984 western Nagano Prefecture earthquake, J. Phys. Earth, 40(2), 379393.
  • James, D. E., and M. K. Savage (1990), A search for seismic reflections from the top of the oceanic crust beneath Hawaii, Bull. Seismol. Soc. Am., 80, 675701.
  • Lahr, J. C. (1999), HYPOELLIPSE: A computer program for determining local earthquake hypocentral parameters, magnitude, and first-motion pattern (Y2K compliant version), U.S. Geol. Surv. Open File Rep., 99–23.
  • Sanford, A. R., and L. T. Long (1965), Microearthquake crustal reflections, Socorro, New Mexico, Bull. Seismol. Soc. Am., 55, 579586.
  • Sanford, A. R., O. Alptekin, and T. R. Toppozada (1973), Use of reflection phases on microearthquake seismograms to map an unusual discontinuity beneath the Rio Grande Rift, Bull. Seismol. Soc. Am., 63, 20212034.
  • Sanford, A. R., R. P. Mott Jr., P. J. Shuleski, E. J. Rinehart, F. J. Caravella, R. M. Ward, and T. C. Wallace (1977), Geophysical evidence for a magma body in the crust in the vicinity of Socorro, New Mexico, in The Earth's Crust: Its Nature and Physical Properties, Geophys. Monogr. Ser., vol. 20, edited by J. G. Heacock et al., pp. 385403, AGU, Washington, D. C.
  • Shalev, E., et al. (2010), Three-dimensional seismic velocity tomography of Montserrat from the SEA-CALIPSO offshore/onshore experiment, Geophys. Res. Lett., 37, L00E17, doi:10.1029/2010GL042498.
  • Sheetz, K. E., and J. W. Schlue (1992), Inferences for the Socorro magma body from teleseismic receiver functions, Geophys. Res. Lett., 19(18), 18671870, doi:10.1029/92GL02137.

Supporting Information

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data
  5. 3. Interpretation
  6. 4. Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

Auxiliary material for this article contains a figure.

Auxiliary material files may require downloading to a local drive depending on platform, browser, configuration, and size. To open auxiliary materials in a browser, click on the label. To download, Right-click and select “Save Target As…” (PC) or CTRL-click and select “Download Link to Disk” (Mac).

See Plugins for a list of applications and supported file formats.

Additional file information is provided in the readme.txt.

FilenameFormatSizeDescription
grl26806-sup-0001-readme.txtplain text document2Kreadme.txt
grl26806-sup-0002-fs01.pdfPDF document1310KFigure S1. Two independent earthquake recordings along with their bracketing “noise” intervals.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.