Monitoring activities and disaster mitigation responses at the Soufriere volcano, Montserrat, Caribbean 19951998.

Keith Rowley
Montserrat Volcano Observatory, P.O. Box 16, Port of Spain, Trinidad, West Indies.
Tel. : +1-868-633-1653, Fax: +1-868-633-1653,
E-mail: krowley@trinidad.net

The 80 square kilometre island of Montserrat in the northeastern Caribbean is among the smallest of the volcanic islands of the Lesser Antilles chain. The 1000 m Soufriere volcano dominates the southern half of the small island. The current eruption of 199598 is the first since the island was colonised in the 16th Century. However, during this same period the volcano has had, at least, three observed seismic disturbances which threatened but did not culminate in any eruption. The last eruption prior to this one, as determined from C14 dating, took place approximately 350 years ago when the island was virtually uninhabited.

Prior to the 1995 re-awakening of the volcano the last period of dormancy was interrupted several times by seismic disturbances which are generally interpreted as precursors to eruptive activity. The latest period of elevated seismic activity was first observed in 1992 and by July of 1995 the volcano burst into phreatic eruption which eventually produced a cold ash cloud which enveloped the capital town of Plymouth which is located on the lower southwestern flanks of the volcano.

The early phreatic phase was terminated by a dangerous magmatic phase of rapid to intermittent dome growth which was punctuated by explosions and dome collapse events which generated a variety of hazards in the form of super-heated pyroclastic flows, hot surges, directed blast and debris avalanche as well as sporadic occurrences of air-fall deposits of cristobalite-bearing volcanic ash.

Soufriere, Montserrat, along with other known potentially active Caribbean volcanoes, has been under continuous surveillance for decades through the deployment of a low density seismic network system which is monitored from the University of the West Indies in Trinidad. Continuous evaluation of the instrumental data and other scientific observations enabled a team of international scientists to forecast the relative course of the eruption thereby providing the authorities with critical information which was used to assess the relative exposure of the population resulting in a series of phased withdrawals as the eruption progressed from a mildly phreatic phase in early July 1995 to violently Vulcanian in 1997. During 1995 the vast majority of the population of Montserrat and virtually all of its economic life and administration was located in close proximity to the flow paths of the several hazards associated with explosive Antillean volcanism. An orderly and timely evacuation of hazardous areas determined through the monitoring processes contributed to the safety of a population which was always juxtaposed to an erupting volcano on a very small island.

For the period July 1995 to mid-1998 the integrity of the hazardous zones which were established and refined along the way remained remarkably intact, since all identifiable, forecasted hazards were confined within the projected areas. There were some casualties during a June 1997 eruption. This development indicates that whereas long-term monitoring systems could lead to notification of the imminent onset of hazardous conditions and the declaration of evacuated zones could result in dramatic shift in populations to places of significantly reduced risk or of safety, the final outcome in disaster reduction is heavily influenced by the response which prevails on the day of any given eruption. The Montserrat experience points to two levels of responses which are required if maximum reduction of the effects of this type of natural disaster is to be achieved. It identifies the need for a long-term physical planning strategy which permanently acknowledges the presence and potential of the volcano and another level of interventionist policies to be effected once the threatened eruption