Conundrum on magmatic reservoir of Soufriere Hills volcano, Montserrat: enigmatic evidence and the case for a vertically-elongated reservoir
The Soufriere Hills Volcano(SHV) has been heavily investigated 1995-present but its magma storage structure remains poorly constrained and some critical evidence conflicts. The reservoir top is >5km (~130MPa) based on phenocryst assemblages, melt inclusion data on volatiles, and the deepest locations of volcano-tectonic earthquakes occurring near the conduit. Early GPS data were interpreted to suggest a reservoir depth of ~5km coupled to a deforming dike, but post-1997 data supported a deeper source, >9km, assuming a spherical reservoir geometry. The cumulative volume of the eruption (~0.9km3) and its chemical and petrological consistency over 13 years suggests that the shallow andesite magma source is voluminous, many km3. The response of CALIPSO borehole strainmeters to a major dome collapse suggested a volatile-saturated magma body of several km3 with a top 5-7km depth. Some crystal phases suggest incorporation >10km, and mixed lavas require a deep supply of mafic magma. Earthquakes are too shallow to enable passive source tomography or S-wave shadows of the magma reservoir. Teleseismic events are being studied but probably cannot resolve reservoir details. The SEA-CALIPSO active source experiment (Dec 2007) aimed to image the lithosphere and, if possible, region of magma storage, but velocity tomography reveals a difficulty in 3D imaging >5km (where magma storage is) due to control of ray curvature by velocity structure. Above this level, the several volcanic centers show higher speed perturbations with respect to the initial velocity model. Travel times from four OBSs and four land stations on a SE-NW line through SHV were inverted to obtain a 2D seismic section, revealing a body with high average velocity underneath the island, about 10km wide, from the surface to 8km depth. These results can be explained by crystallized intrusions, with the currently active magma storage region(s) contained inside this body but masked at the seismic resolution achieved. Thus the strongest constraints on reservoir details may be GPS imaging, with the most complete data sets (10 sites) post-2003. We examine the merits of a stratified single vertically-elongated crustal reservoir with top below 5km and centroid near 9 or 10km, supplied by a deep factory for hydrous andesites fractionated from mafic magma (with minor magma mixing) near the crust-mantle boundary. This model is an alternative to some others with simpler geometry that have been proposed. Deep reflections that might be Moho are revealed at ~10s (~30km) in SEA-CALIPSO shot gathers. The vertical reservoir elongation is supported by some GPS data (e.g., 2003- 05), which display a central dimple of vertical displacement. The upper part of the reservoir contains crystal- rich, highly viscous, gas-saturated, compressible, rhyolitic melt (similar to erupted lava), overlying a lower part of crystal-poor less-viscous hydrous andesite. The model is consistent with both mineralogical constraints indicating shallow storage of erupted lava, GPS indications of a deeper mean-pressure source when the full reservoir expands or contracts, and magma budget anomalies. The presence of an exsolved gas phase greatly increases magma compressibility and suggests that most of the magma volume transferred into or out of the reservoir is accommodated by compression or decompression of stored reservoir magma.
AGU Fall Meeting Abstracts
- Pub Date:
- December 2008
- 7270 Tomography (6982;
- 8145 Physics of magma and magma bodies;
- 8439 Physics and chemistry of magma bodies