Magma flow instability and cyclic activity at Soufriere Hills volcano, Montserrat, British West Indies

Science

Volumn 283, Issue 5405, 1999, Pages 1138-1142

  Voight, B ^a,b   Sparks, R S J ^c   Miller, A D ^c   Stewart, R C ^c   Hoblitt, R P ^c   Clarke, A ^c   Ewart, J ^c   Aspinall, W P ^c   Baptie, B ^c   Calder, E S ^c   Cole, P ^c   Druitt, T H ^c   Hartford, C ^c   Herd, R A ^c   Jackson, P ^c   Lejeune, A M ^c   Lockhart, A B ^c   Loughlin, S C ^c   Luckett, R ^c   Lynch, L ^c   Norton, G E ^c   Robertson, R ^c   Watson, I M ^c   Watts, R ^c   Young, S R ^c   more..

^a PENNSYLVANIA STATE UNIVERSITY   (United States)

^b CASCADES VOLCANO OBSERVATORY   (United States)

^c NONE

Author keywords

[No Author keywords available]

Indexed keywords

CYCLICITY; MAGMATISM; VOLCANO;

ARTICLE; ENVIRONMENTAL IMPACT ASSESSMENT; ENVIRONMENTAL MONITORING; GEOGRAPHY; GEOLOGY; MONTSERRAT; PRIORITY JOURNAL; VOLCANO;

MONTSERRAT;

EID: 0033582738     PISSN: 00368075     EISSN: None     Source Type: Journal    
DOI: 10.1126/science.283.5405.1138     Document Type: Article

References (31)

1. S. R. Young et al., Geophys. Res. Lett. 25, 3389 (1998).
2. R. E. A. Robertson et al., ibid., p. 3429.
3. W. P. Aspinall et al., ibid., p. 3397.
4. J. Neuberg, B. Baptie, R. Luckett, R. C. Stewart, ibid., p. 3661.
5. A. D. Miller et al., ibid., p. 3401.
6. B. Voight et al., ibid., p. 3405.
7. P. Jackson, J. B. Shepherd, R. E. A. Robertson, G. Skerrit, ibid., p. 3409.
8. J. B. Shepherd, R. A. Herd, P. Jackson, R. Watts, ibid., p. 3413.
9. S. R. Young et al., ibid., p. 3681.
10. R. S. J. Sparks et al., ibid., p. 3421.
11. Types of seismicity are discussed by A. D. Miller (5) and B. A. Chouet [Nature 380, 309 (1996)]. Here we correlate cyclic tilt time series with triggered earthquakes, hybrid earthquakes, LP earthquakes, VT earthquakes, and seismic amplitude. The latter is recorded by real-time seismic amplitude measurement (RSAM), described by E. T. Endo and T. L. Murray [Bull. Volcanol. 53, 533 (1991)], to yield a simple quantitative measure of overall seismicity. RSAM measures the average absolute amplitude of seismic signals by using a sampling rate of about 60 samples per second. It does not discriminate between various event types, apart from spikes that reflect large pyroclastic flows. Triggered earthquakes are the number of earthquakes automatically recorded at a giver station during a 10-min period (data are presented in triggers per hour) (6) and are mainly a measure of hybrid earthquake swarms.
12. Types of seismicity are discussed by A. D. Miller (5) and B. A. Chouet [Nature 380, 309 (1996)]. Here we correlate cyclic tilt time series with triggered earthquakes, hybrid earthquakes, LP earthquakes, VT earthquakes, and seismic amplitude. The latter is recorded by real-time seismic amplitude measurement (RSAM), described by E. T. Endo and T. L. Murray [Bull. Volcanol. 53, 533 (1991)], to yield a simple quantitative measure of overall seismicity. RSAM measures the average absolute amplitude of seismic signals by using a sampling rate of about 60 samples per second. It does not discriminate between various event types, apart from spikes that reflect large pyroclastic flows. Triggered earthquakes are the number of earthquakes automatically recorded at a giver station during a 10-min period (data are presented in triggers per hour) (6) and are mainly a measure of hybrid earthquake swarms.
13. 2) is porphyritic (∼35 to 45 volume % crystals), with phenocrysts of plagioclase, hornblende, orthopyroxene, titanomagnetite, and minor ilmenite and apatite. The groundmass (<80 μm) is composed of the same mineral assemblage except that hornblende is replaced by clinopyroxene. The degree of crystallization of the poorly vesicular groundmass varies with eruptive style and magma ascent rate, indicating that significant groundmass crystallization occurs in a short time at shallow depths. Vapor-phase cristobalite occurs as vesicle and fracture fillings (to 15 to 20 weight %) from lavas that have resided within the dome for more than a few days; rapidly erupted and ejected lava lacks cristobalite. Likewise, hornblende is heavily oxidized in lavas with much dome residence time.
14. 2) is porphyritic (∼35 to 45 volume % crystals), with phenocrysts of plagioclase, hornblende, orthopyroxene, titanomagnetite, and minor ilmenite and apatite. The groundmass (<80 μm) is composed of the same mineral assemblage except that hornblende is replaced by clinopyroxene. The degree of crystallization of the poorly vesicular groundmass varies with eruptive style and magma ascent rate, indicating that significant groundmass crystallization occurs in a short time at shallow depths. Vapor-phase cristobalite occurs as vesicle and fracture fillings (to 15 to 20 weight %) from lavas that have resided within the dome for more than a few days; rapidly erupted and ejected lava lacks cristobalite. Likewise, hornblende is heavily oxidized in lavas with much dome residence time.
15. 11 Pa s.
16. Shear strength is estimated from the pressure supported near the base of a vertical rock column, assuming failure occurs in a shear mode.
17. R. S. J. Sparks, Earth Planet. Sci. Lett. 150, 177 (1997).
18. R. D. Mindlin and D. H. Cheng, J. Appl. Phys. 21, 926 (1950); K. Mogi, Bull. Earthquake Res. Inst. 36, 99 (1958).
19. R. D. Mindlin and D. H. Cheng, J. Appl. Phys. 21, 926 (1950); K. Mogi, Bull. Earthquake Res. Inst. 36, 99 (1958).
20. J. Walsh and R. Decker, J. Geophys. Res. 76, 3291 (1971).
21. H2O of 115 to 130 MPa, equivalent to a water-saturated magma chamber at 5- to 6-km depth.
22. edifice. Full resolution of the issue requires three-dimensional elastic-plastic modeling. Seismic threshold overpressures in the text are calculated by recalibrating the maximum amplitude pressure to 10 to 27 MPa, as deduced by R. E. A. Robertson et al. (2), and applying similar methods for the August explosions.
23. Shear resistance was calculated by assuming an extrusion overpressure of 10 to 27 MPa and setting it equal to shear strength times boundary area of plug. Plug depth was assumed as ∼270 m. Weight of extruded magma is accounted for by definition of overpressure.
24. M. V. Stasiuk, C. Jaupart, and R. S. J. Sparks [Earth. Planet. Sci. Lett. 114, 505 (1993)] estimated initial excess chamber pressures of 10 to 23 MPa for lava eruptions at Lonquimay, Chile, in 1989; Soufriere, St. Vincent, in 1979; and Paracutin in 1946 to 1952. During strong tilt cycles, the outlet pressure may be augmented by the average of uppermost-conduit pressure variations.
25. J. D. Devine, M. J. Rutherford, J. E. Gardner Geophys. Res. Lett. 25, 3673 (1998).
26. E. Rabinowicz, Proc. Phys. Soc. (London) 71, 668 (1958).
27. P. Okubo and J. H. Dieterich, in Earthquake Source Mechanics, S. Das, J. Boatwright, C. Scholz, Eds. (American Geophysical Union, Washington, DC, 1986), pp. 25-36; C. H. Scholz, The Mechanics of Earthquakes and Faulting. (Cambridge Univ. Press, Cambridge, 1990).
28. P. Okubo and J. H. Dieterich, in Earthquake Source Mechanics, S. Das, J. Boatwright, C. Scholz, Eds. (American Geophysical Union, Washington, DC, 1986), pp. 25-36; C. H. Scholz, The Mechanics of Earthquakes and Faulting. (Cambridge Univ. Press, Cambridge, 1990).
29. Feedback between pressurization and flow from the chamber result in decline of flux as magma degasses and stiffens, and gas pressure builds, followed by accelerated flux after pressure is relieved. For one set of constraints, the system may be stable and independent of velocity perturbation; for another, it may be stable under quasi-static loading but unstable if subjected to a large pressurization or velocity jump [J. R. Rice and A. L. Ruina, J. Appl. Mech. 105, 343 (1983); J. C. Gu, J. R. Rice, A. L. Ruina, S. T. Tse, J. Mech. Phys. Solids 32, 157 (1984)].
30. Feedback between pressurization and flow from the chamber result in decline of flux as magma degasses and stiffens, and gas pressure builds, followed by accelerated flux after pressure is relieved. For one set of constraints, the system may be stable and independent of velocity perturbation; for another, it may be stable under quasi-static loading but unstable if subjected to a large pressurization or velocity jump [J. R. Rice and A. L. Ruina, J. Appl. Mech. 105, 343 (1983); J. C. Gu, J. R. Rice, A. L. Ruina, S. T. Tse, J. Mech. Phys. Solids 32, 157 (1984)].
31. We thank our colleagues at Montserrat Volcano Observatory for assistance. Supported by the Department for International Development (UK), the British Geological Survey (BGS), the Seismic Research Unit of the University of the West Indies, and the U.S. Geological Survey (USGS). B.V. acknowledges support from NSF, BGS, and USGS, and R.S.J.S. from the Natural Environmental Research Council, and the Leverhulme Trust.