The Explosion of March 2004 at Montserrat: Constraints From Borehole Strain Data

The dome-building volcano Soufriere Hills on Montserrat in the Carribean has been active since about 1995. By the end of January 2003, the CALIPSO group had established a small network of 4 borehole sites with the instrumentation including Sacks-Evertson dilatometers. Major dome collapse in July 2003, reported elsewhere, was accompanied by increased pressure in a deep (~5km) reservoir. In March 2004 a well observed and recorded explosion took place (Green and Neuberg, 2005). This event resulted in clear, coherent signals (peak amplitudes up to ~100 nanostrain) on the 3 strainmeters then being recorded. The amplitudes and polarities of the strain signals from this event cannot be satisfied by a single Mogi-like source: two closer sites at similar distances (but different azimuths) record similar amplitudes but opposite polarities while the most distant site has amplitude less than one-fifth that of the nearer sites. The data are well satisfied by a small shallow dike striking N55W, consistent with geologic fabric, together with corresponding loss of pressure in the deep reservoir. Seismic activity accompanying the explosion does not commence until some time after the onset of the strain event, i.e. after the dike begins to form. Later, the magma chamber is re-pressurized.