- 0 -

The Montserrat Volcano Observatory: its evolution, organization, rôle and activities.

W. P. Aspinall1,2, S.C. Loughlin1,3, F.V. Michael4, A.D. Miller1,5 G.E. Norton1,6, K.C. Rowley1,7,

R. S. J. Sparks1,8, S. R. Young1.

Montserrat Volcano Observatory

1 Montserrat Volcano Observatory, Mongo Hill, Montserrat, W.I. 2 Aspinall & Associates, 5 Woodside Close, Beaconsfield, Bucks, HP9 1JQ, UK. 3 British Geological Survey, Murchison House, West Mains Road, Edinburgh, EH9 3LA, UK. 4 Emergency Department, Govt. of Montserrat, St John’s, Montserrat. 5 GEOWALKS, 24 Argyle Place, Edinburgh EH9 1JJ, Scotland. 6 British Geological Survey, Keyworth, Nottingham, NG12 5GG, UK. 7 LANDATA Ltd, Trinidad & Tobago. 8 Department of Earth Sciences, University of Bristol, Bristol, BS8 1RJ, UK.

Published in slightly modified form as: Aspinall, W. P., S. C. Loughlin, et al. (2002). The Montserrat Volcano Observatory; its evolution, organization, role and activities In: The eruption of Soufriere Hills Volcano, Montserrat from 1995 to 1999 (eds. Druitt, T. H. and Kokelaar, B.P.); The Geological Society of London, Memoir 21: 71-91.

- 1 -

Abstract

The Montserrat Volcano Observatory (MVO) is a statutory body of the Government of

Montserrat, and is the agency responsible for volcano monitoring operations on the island of

Montserrat, in the West Indies. It was formed shortly after the first phreatic explosions from the

Soufrière Hills Volcano occurred on 18 July 1995, and evolved from a hastily created, loose

interim entity to a fully established volcano monitoring operation. Participating scientific teams

have been drawn mainly from the Seismic Research Unit of the University of the West Indies,

the U.S. Geological Survey, the British Geological Survey and various US and UK universities.

Valuable support has come from the University of Puerto Rico and the Institut de Physique du

Globe de Paris through its Observatories in Guadeloupe and Martinique. Despite its hurried

inception, the MVO has been able to provide timely, high quality hazard advice to the civil

authorities and has maintained an exceptional record of all scientific aspects of the eruption. Its

public education and information efforts have been extensive, and there have been unusually

high levels of interaction between scientists and the civil authorities, and between scientists and

the public, both within Montserrat and outside in the wider world. The experience of setting up

and running the MVO under difficult conditions has exemplified the advantages of teamwork

and flexibility within monitoring operations, and the benefits of openness and clarity in public

interactions.

1. Introduction

The first historic eruption of the Soufrière Hills Volcano (SHV), on the island of Montserrat

(Figure 1), began on 18 July 1995 with phreatic explosions from within the central crater of this

dome complex volcano. This style of activity lasted until November 1995, and was followed by

an extended period of dome growth; this second phase of the eruption came to an abrupt end in

early March 1998 (Young et al., 1998a; Norton et al., this volume). There then followed a

twenty-month interval in which no significant new magma was emplaced in the crater and only

mild activity involving ash venting, explosions, generally low levels of seismicity, and several

collapses of the dome, took place. In mid-November 1999, however, a new lava dome began to

grow in the big void that had been left in the remnants of the 1995-1998 dome by collapses in

1998 and 1999. This new dome increased steadily in size until it, too, collapsed almost entirely

- 2 -

down the Tar River valley on 20 March 2000. After this incident, dome growth recommenced

with increased vigour and has continued to the time of writing (April 2000). Magma production

rates have largely determined eruptive style with high production rates being associated with

major dome collapses and explosions. This eruption, protracted over a period nearly five years,

has caused tremendous disruption to the lives of all residents on the island and has necessitated

the continued presence on island of a full-time scientific team and a dedicated monitoring

facility.

For the past five years, the Montserrat Volcano Observatory (MVO) is the agency that has been

responsible for continuous monitoring of the Soufrière Hills Volcano (SHV). The inception,

evolution and operation of the MVO occurred against a background of complex administrative

arrangements, which characterised Montserrat as a British Dependent Territory (now an

Overseas Territory). The administration of the island involves both a Governor, who represents

the Queen and the British Government, and a locally elected Government, led by a Chief

Minister. There is limited self-government, with a legislature and civil service, but Britain has

responsibility for defence, internal security and foreign relations. There is an Executive Council

consisting of the Governor, the Chief Minister, the Attorney-General, the Financial Secretary and

three more ministers, and an eleven-member Legislative Council, of which seven are elected by

universal adult suffrage (from age 18), two are nominated, and two are appointed ex-officio. The

British Government Department with administrative responsibility for Montserrat is the Foreign

and Commonwealth Office. However, this department has virtually no financial resources with

which to respond to a natural disaster, and the actual financial and logistic responsibility for

Montserrat in an emergency devolved to the Overseas Development Administration (now called

the Department for International Development - DFID). Another complication in relation to the

Foreign Office is that Montserrat affairs have been dealt with internally both by the West Indies

and Atlantic Department (WIAD), in London, and by the Dependent Territories Regional

Secretariat (DTRS), based in Barbados. The intricate relationships between all these ministries,

departments and organisations, and their respective rôles in the crisis, were not always clear and,

as a result, were confusing to many of the scientists responding to the emergency.

- 3 -

Within the Eastern Caribbean, the regional agency for earthquake and volcano surveillance, the

Seismic Research Unit (SRU) of the University of the West Indies, had been actively monitoring

Montserrat for several decades, with permanently installed seismographs and occasional field

visits. Indeed, the history of SRU’s involvement in Montserrat goes back to, and beyond, the last

significant volcano-seismic crisis there in 1966 (Shepherd et al., 1971). Their experience of

handling volcanic crises in the region includes the eruptions of the Soufriere of St. Vincent in

1971-72 (Aspinall et al., 1972) and 1979 (Shepherd et al., 1979), as well as numerous other

volcano-seismic swarms and related studies. Between January 1992 and July 1995, seismic

stations on Montserrat and the surrounding islands recorded more than fifteen episodic swarms of

small earthquakes. These originated at depths of 10-15 km and had epicentres defining a northeast

trend off the East Coast of Montserrat (Ambeh et al., 1998). As the swarms recurred, the SRU

upgraded and augmented its seismic network on Montserrat, and made field measurements of dry

tilt, and observations around the volcano to watch for changes in fumarolic activity. They were

assisted by scientists from the Guadeloupe Volcano Observatory, who undertook annual sampling

visits to the main fumaroles and thermal spring areas for physiochemical monitoring (Hammouya

et al., 1998). These latter efforts, however, did not reveal any significant changes in fumarole

activity or chemistry indicative of an impending eruption. When activity escalated dramatically

on Montserrat in July 1995, an operational base was immediately established on the island by

SRU and, from this foundation, the MVO evolved into a fully functional local organisation

staffed by a multinational, multi-institutional team, utilising a variety of geophysical, geological

and geochemical techniques to monitor the volcano. Funding for these activities was provided

mainly by the Government of Montserrat (GoM) and by the British Government. The initial

absence of suitably trained local staff and the longevity of the eruption necessitated a major

turnover of staff at the MVO: over 100 individual scientists and technical staff have worked

there during the past five years. At the height of the crisis, a staff complement of up to 8 UK

scientists (generally appointed through the British Geological Survey), three SRU scientists and

technical staff could be mustered, eventually supplemented by seven Montserratian technical and

administrative staff. In addition, staff and scientists from the US Geological Survey (USGS),

and UK and US universities have been involved in MVO operations from time to time, and

generous support has been received also from the Institut de Physique du Globe de Paris, through

- 4 -

staff from its Volcano Observatories in Guadeloupe and Martinique, and from the University of

Puerto Rico. From the resident Montserrat population itself, several volunteers have provided

invaluable assistance with observations, routine support work and night duty.

The establishment of a smoothly functioning scientific team during a crisis with staff drawn from

various nationalities and organisations is inherently difficult. The stressful situation in

Montserrat combined with a complex and unsatisfactory overall management structure and

funding deficiencies to create a number of significant problems. Notwithstanding these

difficulties, however, a great deal of valuable scientific work has been done by the MVO,

particularly in the documentation of the eruption. There has been rapid publication of accounts

of activity and scientific results in various media (e.g. Ahmad, 1996; Robertson and Wadge

1996; Montserrat Volcano Observatory Team, 1997; Young et al., 1997; Ambeh et al., 1998).

A collection of papers focusing on the first 2 years of the eruption was published in a two-part

special section of Geophysical Research Letters (Aspinall et al., 1998a; Young et al., 1998b),

and overviews of the scientific advances appear here (Sparks and Young, this volume), and

elsewhere (Robertson et al., 2000). There have been extensive efforts in public education, and

the MVO has become a primary source of information for the local administrative authorities and

the public, and a major attraction for visitors to the island. More importantly, the continued

operation of a volcano monitoring facility on Montserrat is essential to the present and future

occupation of the island, and recent administrative changes to the status of the MVO have now

placed its managerial and financial structure on a stable basis.

There are many valuable lessons to be learned from the experience of setting up and running the

MVO; important insights and issues of a detailed scientific nature are treated in the companion

papers, or elsewhere (see references, below). This paper summarises some of the unique features

of the new institution and outlines essential elements in its evolution, structure and organization.

However, myriad other factors and influences shaped events and outcomes in Montserrat, and

these cannot all be covered in this one paper: a book by Pattullo (2000) and a comprehensive

report by Clay et al. (2000) each document many aspects of the broader picture.

- 5 -

2. Evolution of the MVO

The genesis of a permanent volcano observatory on Montserrat can be traced back to the periods

of increased seismic activity that began in Montserrat in the early 1990’s, first recognised by the

SRU from their regional monitoring activities, which prompted strengthening of the network on

island. Then, immediately following the first phreatic explosions from the Soufrière Hills

Volcano on 18 July 1995, the SRU established an operational base on Montserrat to provide

direct scientific advice on the state of the volcano to local authorities. Acting on the advice of

SRU, the Government of Montserrat (GoM) invited scientists from the USGS Volcano Disaster

Assistance Program (VDAP) and the Guadeloupe Volcano Observatory to join the SRU team on

island, to assist with their monitoring activities. The SRU and USGS scientists, together with

one UK scientist and two student volunteers recruited in early 1995 from the local secondary

school, formed the embryonic observatory. A temporary facility was established near the

government headquarters in Plymouth in July 1995 (Figure 2) and was first called the “Soufrière

Hills Volcano Observatory”. The physical location of the Observatory moved (Figure 1) from

Plymouth to Vue Pointe Hotel in August 1995 (Figure 3), and then to a rented villa in Old

Towne in October 1995 (Figure 4), at which time the name changed to the Montserrat Volcano

Observatory.

The USGS VDAP Team spent six weeks on Montserrat from late July 1995, and reinforced the

pre-existing seismic network with the establishment of new short-period seismograph stations.

In addition, three electronic tiltmeters were installed and a programme of sulphur dioxide

monitoring was started, using a Correlation Spectrometer (COSPEC). An automated seismic

data acquisition and processing system was also installed, and this eventually replaced the pre-

existing Soufrière System software that had been the basis of the data acquisition for the SRU

network.

At these early stages of the crisis (i.e. from August to November 1995), the Foreign and

Commonwealth Office of the British Government commissioned short visits to Montserrat by

individual UK consultants with experience of volcanic activity in the Caribbean. The missions

were contracted and funded through the Department for International Development (then called

- 6 -

ODA - the Overseas Development Administration). These scientists functioned essentially as

advisors to the Governor and to the British Government, but also assisted MVO with routine

monitoring work. In October 1995, the first of a series of short-term contracts from ODA

supported the direct involvement of staff from the British Geological Survey (BGS), to work

with the incumbent monitoring team. Later contracts with the BGS supported staff and UK

university scientists and students (whose numbers increased from 2 in January 1996 to 7 in July

1996), and other specialists from time to time. The local complement of staff at the MVO also

increased in October 1995 with the secondment of several Montserratian civil servants from

other government departments.

The day-to-day running of the MVO was managed by a Chief Scientist (CS), who was

responsible for co-ordinating the scientific work and for reporting to the Government of

Montserrat and to the Governor. During the first year of its operation, the Head of the SRU

fulfilled this rôle with various members of SRU staff acting on his behalf when he was off island.

By this time, the need had been recognised for a sizeable scientific, technical and administrative

staff to monitor the volcano 24 hours a day. This staff was drawn from various institutions and,

by 1996, typically consisted of eight local staff, two to three staff from the SRU and six to eight

staff drawn largely from the UK. The US Geological Survey, the Institut de Physique du Globe

de Paris, Brown University and the University of Puerto Rico also provided significant

assistance. The UK contingent was drawn both from the British Geological Survey and the UK

Universities with strong research expertise in volcanology (principally from the universities of

Bristol, Lancaster, the Open University and Cambridge). However, priorities for scientific work

and responsibilities for giving advice between all these different institutions and individuals were

initially far from clear and, at times, relationships became very strained.

In part at least, the tensions over scientific obligations and priorities can be traced to the very

difficult question of what level of response was appropriate at the time, given the uncertainty

surrounding the future course of the crisis. Only two decades earlier, in 1976, a volcanic crisis

on the neighbouring island of Guadeloupe, with many volcanological similarities to conditions in

Montserrat in 1995, had failed to develop into a significant, life-threatening eruption after a

- 7 -

major evacuation of the population had been ordered. The circumstances surrounding the

decision to instigate the evacuation became the focus of intense debate among volcanologists

(see Tazieff, 1980, and preceding contributions in the same journal referenced therein; also

Fiske, 1984), and the management of the emergency was the subject of a wide-ranging

international commission of enquiry, sponsored by the French government. One major factor

that was considered in analyses of the Guadeloupe episode was the enormous modern economic

cost of an unfulfilled warning of dangerous eruption, notwithstanding the haunting political

spectre of that earlier Caribbean volcanic catastrophe, the 1902 eruption of Mt. Pelée in

Martinique. Against this backdrop, it is understandable that in 1995 some in the Eastern

Caribbean, particularly those in Montserrat itself, were reluctant to initiate urgent precautionary

measures until the extent of the threat could be better ascertained. They would be very conscious

of the potential for a devastating impact on the economy of their small island, which had only

just been re-established after the ravages of Hurricane Hugo, in 1989. On the other hand, the

USGS VDAP scientists, some of whom had witnessed the catastrophic consequences of the

eruptions of Mt. St. Helens and Pinatubo, took a more pessimistic view of the potential outlook.

In these circumstances, higher political directorate, in the form of the UK government, had to be

mindful of its ultimate responsibility for the welfare of the people of Montserrat. Steeped in

recent public disasters and health scares, British politicians and civil servants were probably

much more risk averse than their counterparts in the local government, and in the position to

assert their judgements and authority more forcefully.

This political dichotomy generated dissonances within the scientific group. Contention arose as

differences developed as to the perceived level of risk and, with it, difficulties relating to attitude,

personality, and management styles. These were exacerbated in the first weeks of the crisis by

the disparities in funding and resourcing that were initially available to, or deployed by, the

various groups within the team. It was not conducive to team spirit and collaboration that the

regional scientists were obliged to make use of separate, cheaper accommodation, away from

others, and had to endure restrictive transport arrangements and so on; this promoted a “them-

and-us” atmosphere, and reduced opportunities for the informal scientific dialogue (and banter)

that is so invaluable in crisis circumstances. A simple instance of a contentious issue was the

- 8 -

accumulating cost of frequent and extended international phone calls from the nascent

observatory facility, taken for granted as a necessity by many of the metropolitan scientists, but

incurring a major expense for the host government in local terms. It is unfortunate that, at this

early stage, it was not possible to forge a unified approach to these issues within the whole MVO

team, even though a large measure of concurrence existed and the British involvement by now

included several former staff members of SRU, all familiar with the constraints faced by that

organisation. Thus, while the regional team from SRU were initially disinclined to widen the

scope of monitoring (and scientific advice) much beyond that which could be sustained by or

through the local government, some other scientists were less restrained about pressing forward

recommendations for strengthening the scientific efforts (which, in effect, implied seeking direct

financial support from the British government).

As things stood then, it is perhaps not surprising that differing perceptions of the balance of risk

and economic imperative overlapped into another area where substantial difficulties always arise

in a volcanic crisis: the articulation of scientific uncertainty in the forecasting of future activity.

Where the public is exposed to a given hazard, natural or otherwise, the lack of sufficient

knowledge to quantify the attendant risk(s) with useful accuracy impinges upon a government’s

ability to comprehend, and justify, the extent to which steps should be taken to eliminate or

reduce that risk In the first few weeks of the Montserrat crisis there was perhaps, at times, some

unwarranted scientific dogmatism about what might or might not happen at the volcano,

especially in terms of it turning magmatic and explosive. The confounding effects of these

diverging, categorical stances were then compounded for a short while by an overall diminution

in communication between scientists and the various civil authorities. The result was a dip in the

confidence of the authorities in the MVO team and, with it, some loss of public credibility; this

was not fully restored until later, when a consensual approach was achieved. Thus a situation

developed, as the crisis started to drag on and deepen slowly, in which there was increasing

pressure from the British government, in particular, for the wider involvement of more scientists

and specialists in their deliberations. This reflected both the natural instinct of politicians to

canvass as many opinions as possible, and new UK government policy guidelines (Office of

Science and Technology, 1997, 1998; The Royal Society, 1999; see also Curnow, 1999) which

- 9 -

enjoin politicians and civil servants to organise and engage wide-ranging scientific consultation

in any science-related decision-making process, a precept emerging from recent high profile

health risk crises such as BSE. This pressure to extend the scope of the scientific effort in

Montserrat started to manifest itself by an increasing presence of British personnel at MVO and

this, in turn, put additional strain on SRU’s efforts to manage all the scientific elements of the

crisis. In short, SRU was under-resourced for a sustaining what was turning out to be a

burgeoning and prolonged crisis rôle on one island while it still had many other regional

commitments to fulfil and, as an institution, it initially showed some reluctance to seek additional

technical and scientific assistance once the VDAP team had left.

The effects of the background political realities in London led, later on, to further strains in

relationships between the SRU staff from Trinidad and staff from the UK. A major contributory

factor was the multiplicity of reporting lines for scientific advice in the early stages of the

emergency - at times, there were three alternative routes of communication to the UK

government: through the Governor and his office on Montserrat; through the Foreign Office,

and through the Department for International Development. This caused SRU scientists to feel

that their rôle was being undermined or their advice countered by separate, independent opinion

being expressed in the UK. Another factor was that the UK government placed all contracts for

scientific monitoring work in Montserrat through BGS as their principal source of geological

advice. The involvement at any level in this process of SRU/UWI, the main regional

organisation which, up to that time, had held full responsibility for volcano monitoring in

Montserrat, was limited. In part, this one-sided arrangement served to sour inter-institutional

relationships, both at administrative and scientific levels. In addition, during the first two years

of the eruption, ODA/DFID itself had virtually no direct contact with the SRU concerning

monitoring in Montserrat or the needs of the MVO. Matters were not helped by changes in

SRU’s status within the University of the West Indies: in 1996, the Seismic Research Unit,

having been a self-administered regional Research Unit for all its existence, was subsumed into

the bigger Physics Department, where teaching and theoretical science are the mainstays. Thus,

at the administrative level there was a shift to a directorate with other concerns, little experience

of the demands and ramifications of mounting a major crisis response, and apparently limited

- 10 -

appreciation of the opportunities for reinforcing the group’s work by international exposure and

external funding that such a prominent eruption affords. The change impinged on the

deployment in the field of individual SRU scientists and technical staff and, together with some

residual internal difficulties that were legacies from the start of the eruption, significantly

curtailed the group’s former responsiveness, resilience and flexibility. With all these internal

and external factors taken together, it is not surprising that SRU, a small team labouring under

severe constraints in difficult conditions, started to lose some momentum in its lead rôle at

MVO. Having said all this, however, the underlying root cause for most of the difficulties faced

by the scientific team, jointly and as members of separate institutions, stemmed directly from the

political duality that existed in Montserrat.

Recognising the unsatisfactory nature of the arrangements at MVO, an “audit” of the situation

was commissioned by the Governor in January 1996. Its main purpose was to consider medium-

term needs (i.e. the following 3 years) and the longer-term development of the MVO and

associated scientific programmes. The report concluded that the eruption could be long-lived

and that there was a need to establish and staff a well-equipped and effective monitoring facility.

The report recommended the installation of new seismological equipment; the training of local

staff and the construction of a purpose built permanent observatory building. The report also

recommended the establishment of a management board and greater co-ordination in handling

the UK contribution to the monitoring effort. While the Audit Report was duly completed in

January 1996 and its recommendations apparently accepted, most of these (save for the

procurement and installation of a new state-of-the-art seismic network) proved to have an

inordinately long gestation period, or were not implemented at all.

During the course of 1996, there was a shift in the rôle and responsibilities of the main agencies

involved in, and responsible for, the management of the MVO. At the start of the eruption, the

MVO was locally administered through the Development Unit within the Chief Minister’s

Office. In August 1996, responsibility for the MVO along with all other aspects of emergency

operations was transferred to the Governor’s Office. An Administrative Manager was employed

to assist with the co-ordination of non-scientific aspects of the MVO work, and to manage local

- 11 -

staff. In September 1996, a decision was taken for the MVO to be jointly managed by the BGS

and SRU with the position of Chief Scientist alternating between staff drawn from the SRU and

the UK. Each Chief Scientist took on a tour of duty, which lasted between four to six weeks,

typically requiring a commitment to fourteen or more hours per day, seven days a week, and

disturbed nights when the volcano was very active. This rotation arrangement sought to reduce

the stress associated with work at an erupting volcano and to ensure that the individuals who

undertook this important rôle functioned at optimal efficiency in the conditions, without

suffering undue fatigue. The switches in leadership caused some inconsistency in management,

unavoidable discontinuity whenever there was a change of Chief Scientist, and difficulties in

developing long-term strategic policies for the MVO. However, these shortcomings were

ameliorated by posting to the team selected scientists who undertook longer stints of duty as

Deputy Chief Scientist (DCS); the DCS took on a significant proportion of the routine

administrative workload of the Chief Scientist in Post, and provided much needed continuity in

management (but still acted fully as a member of the scientific team). As the crisis unfolded, the

direct load of responsibility on the Chief Scientist was also relieved by the presence at the

observatory of nominated Senior Scientists, experienced volcanologists who could assist with the

tasks of interpreting observations and data for day-to-day hazard assessment and risk mitigation.

With the establishment of a Chief Scientists’ Committee in early 1997, an attempt was made to

develop policies that were more consistent. This committee met three times between 1997 and

1998 and, in between, there were frequent contacts with the Chief Scientist in Post via the

Internet so that assessments and views coming from the MVO had the understanding and support

of all the other scientists who filled the CS rôle from time-to-time.

In November 1996, the BGS submitted a proposal to the Department for International

Development (DFID), to rationalise the management of all scientific input to the MVO that was

funded directly by the British Government. The project covered a two-year period beginning

April 1997 and provided for the overseas input to the MVO to be managed by the BGS, through

whom all UK funding to the MVO would be channelled. For a variety of administrative reasons,

the project was not fully approved until mid-1997, and then there were further delays in the

signing of a contract between UWI and the UK Natural Environment Research Council (NERC)

- 12 -

for the involvement of the SRU, and this was not completed until early 1998. The very slow

speed of completing this contract led to further difficulties between the two principal institutions

involved in monitoring the volcano.

With the inception of the two-year contract, the formal reporting line to the UK government then

became that from BGS to DFID. This involved a combination of meetings between DFID staff

and weekly reports from the Observatory to DFID on the state of the volcano. DFID also

received scientific feedback from the Foreign Office through the informal input of scientific

advice or indirectly via the Governor’s Office. These multiple reporting channels and complex

inter-relationships continued to lead to occasional misunderstandings and confusion during the

crisis. The on-island reporting mechanism – bi-weekly briefings by the Chief Scientist in Post

(see Section 5.1) – remained the most important direct interface to the authorities and remained

largely unaffected by these administrative changes.

By late 1997, funding for the MVO was derived jointly from the Government of Montserrat

(local and recurrent expenses) and from the British Government through DFID (foreign scientific

support and major expenditure on monitoring equipment). The GoM supported the local staff,

MVO’s local budget and the direct participation of the SRU during the period July 1995 to April

1997. Various short-term contracts between DFID, the BGS and individual consultants

supported the UK and other non-SRU overseas scientists up until April 1997. The new 2-year

DFID contract, which began in April 1997, provided the means for continued involvement of all

overseas scientists at the MVO, including those seconded from the SRU. Initially, the authorities

had been slow to appreciate the significance of the longevity of the eruption and the crisis, and to

understand the implications for scientific requirements for monitoring an active volcano. This

had led to a number of short-term administrative responses and the early development of the

MVO in a civil service setting that, in turn, allowed little flexibility in spending, thereby

frustrating rapid reaction to changing circumstances at the volcano or to equipment breakdowns,

for instance. The commitment by DFID to a two-year period of stable funding for the

monitoring activities improved arrangements. However, the scientific team was still unable to

respond as rapidly to breakdown of equipment or a need to acquire new equipment as they would

- 13 -

have liked since the BGS, as manager of a publicly-funded project, was also bound by restrictive

procedures that allowed little room for local flexibility.

Formation of a Management Board for the MVO, which had been recommended by the January

1996 Audit Report, was finally completed in October 1996. However, the first meeting of this

board did not occur until 7 March 1997. At that meeting, the MVO was declared an entity of the

Government of Montserrat; a mission statement was approved, and decisions taken on the

management structure, staffing policy, funding and long-term development of the MVO. As was

the case with the Audit Report few, if any, of the Board’s recommendations were fully

implemented. The Board itself had no direct rôle in the management of the MVO and did not

devise any clear mechanisms for ensuring that its decisions were implemented, and there was no

absolute line of authority to link the Board to the MVO operations. In addition, the changes in

administrative and funding arrangements that were put in place during late 1996 to mid-1997

were paralleled by a shift, as noted above, in the relative rôles of the SRU and BGS, the two

main scientific organisations involved within the MVO. The BGS now managed the

involvement of non-SRU overseas scientists at the MVO (who were in the majority), and

effective co-ordination of nearly all aspects of the scientific work passed to that organisation.

Outside the observatory, the escalating volcanic activity in the middle and latter parts of 1997

caused another relocation of the MVO base after nearly two years in Old Towne (Figure 5), to its

present location at Mongo Hill (Figures 1 & 6). The requirement for a permanent MVO had by

then been well established. However, recurring problems regarding management and financing,

and the longevity of the eruption itself, pointed to the need for a wide-ranging governmental

review to ensure the survival of the MVO beyond the eruption. This led to a Joint

Comprehensive Review (JCR) of the MVO, which was commissioned by DFID in February

1998. The main purpose of the review was to clearly define the objectives for, and outputs

required from the MVO, and to formulate plans on how this could be achieved cost-effectively in

the short-, medium- and longer term. The review was intended to formulate the basis for future

funding and management of the MVO initially for the next three-year period. The final report

was delivered in mid-1998, and contained recommendations on all aspects of the MVO’s

- 14 -

operations. The JCR recommended that the MVO should become a statutory body of the GoM,

managed by a two-tiered board system. A Main Board would meet annually to determine policy

while an Operational Board (with wider representation), would meet quarterly and be responsible

for implementation of policy and accountability. A Director with an appropriate scientific

background was to be recruited as soon as possible and arrangements made for continued

provision of specialist advice, whenever needed. It was recommended that all funding to the

MVO should be arranged through a single channel. Provision of a forward observation post,

which would give the MVO team a clear view of the volcano, was seen as a critical operational

issue.

On 15 July 1998, the second meeting of the Management Board of the MVO was convened.

This meeting had become somewhat overdue and attempted to deal with several important

issues. However, most of the discussion centred on the results of the Joint Comprehensive

Review, and several on-going matters were left outstanding. Despite this, a number of key

decisions were taken at this meeting. The recommendation of the JCR to establish the MVO as a

statutory body was endorsed; the composition of both boards was discussed, and agreed upon. It

was decided that an Interim Director, with responsibility for overall management of the MVO,

would be appointed as soon as possible. In addition, separate subcommittees were established to

make specific recommendations on the future rationalisation of the two seismological networks

operated by the MVO and on the recruitment of a permanent director, respectively.

Unlike previous occasions, follow-up action was actually taken on all the main decisions of the

Board. An Interim Director of the MVO was appointed (who held post from October 1998 to

June 1999), members of the Operational Board were selected, and a draft bill was formulated for

the incorporation of the MVO as a statutory body of the Government of Montserrat. This bill

subsequently passed into law in August 1999. The appointment of the Interim Director allowed

administrative action to be taken forward in a number of areas. The Operational Board held

several meetings and made recommendations for consideration and endorsement by the Main

Board on the establishment of a permanent observatory building, monitoring strategy, funding

and a number of policy issues. As a result, the MVO now has a full-time Director who currently

- 15 -

supervises seven local technical and administrative personnel and one contracted seismologist.

They are assisted as needed in the short-term by one or more professional geoscientists, called

down through the BGS/DFID arrangements. Funding is now provided by the British

Government through budgetary aid to the Government of Montserrat, and includes provision for

the construction of a permanent observatory.

The wider social, political and economic contexts of the Montserrat eruption crisis and its

management, within which the evolution, vicissitudes and re-structuring of the MVO are but a

small part, have been exhaustively evaluated by Clay et al. (2000), on behalf of DFID.

3. Monitoring

Dedicated monitoring of the Soufrière Hills volcano remains the main objective of the MVO.

This has involved application of a variety of techniques guided by the need to acquire sufficient data

to provide reasoned scientific advice to civilian authorities. The strategy has been flexible with new

techniques being introduced to address specific concerns about the hazards posed to the population

at risk. Routine monitoring revolves around three main activity areas: visual observation of the

volcano, operation of the seismological network, and ground deformation measurements (Figures

8 & 9). There is an ongoing programme of dome morphology and volume assessment, and

deposit mapping. Gas monitoring is done through direct and remote sampling. Environmental

monitoring of groundwater, rainwater, ash and atmospheric geochemistry has been undertaken

since 1995. Other techniques used include: petrological and geochemical analysis of the erupted

products; gravity surveys and rock strength measurements. Video and still photographs are taken

whenever conditions dictate and activity permits. A summary of the main areas of scientific

activity at the MVO is given in Table 1; in respect of scientific results, an overview can be

found in Sparks and Young (this volume).

Visual observations have been made from helicopter and from various observation points around

the volcano since eruptive activity began in July 1995. These observations have been vital to the

monitoring operations and have been supplemented with basic photography using small and

medium format still cameras and video photography. The data collected have allowed the

- 16 -

detailed and accurate documentation of the evolution of volcanic activity, of changes in dome

morphology and dome volume.

Two seismic monitoring networks have been used in seismological monitoring at the MVO. A

short-period (SP) network using analogue telemetry has been in place since July 1995, including

some stations that had been occupied by the SRU before the crisis took hold. The network

consists mainly of vertical 1 Hz instruments, although two 3-component instruments have been

deployed. Throughout the eruption at least five and usually eight or more stations have been

continuously transmitting by VHF radio telemetry or telephone lines to a PC-SEIS data-

acquisition system (Lee, 1989) installed at the MVO by the USGS VDAP Team. Events are

inspected routinely, classified by type (Miller et al., 1998) and locations computed using the

HYPO71PC program (Lee and Valdes, 1989). The seismic signals are monitored 24 hours a day

at the Observatory. Signals from four of the stations are written to paper drum recorders to give a

real-time view of the seismic activity. A network of five 3-component Guralp CMG-40T

broadband seismometers was installed in October 1996 and operated in conjunction with the

short-period network (Figure 8). Three single component Integra LA100/F seismometers were

also included in the network. The data are transmitted to the observatory as 24-bit digital signals

by UHF radio telemetry where it is acquired by SEISAN data acquisition system. Further details

of these networks along with the data collected during the first two years of operation can be

found in Aspinall et al. (1998b), Miller et al. (1998) and Neuberg et al. (1998).

Ground deformation monitoring began with the use of a Wild NA2 precision level and 3-metre

invar levelling rods at three dry tilt stations located on the western flank of the volcano (Figure

9). Concern over the stability of the Gages crater wall overlooking Plymouth led to the

introduction of a Leica TC1100 Total Station. This was used with single or triple prism targets

to monitor possible movement on the Gages crater wall and Castle Peak dome. Six target sites

were initially established in late September 1995. The network was radically modified between

November 1995 and January 1996 when further instrument and target sites were added (Jackson

et al 1998). Destruction of reflectors during the progressive encroachments of the volcanic

activity on to the surrounding flanks undermined the effectiveness of this network which, by mid

- 17 -

1998, was effectively abandoned when the single remaining reflector on the northern flank was

lost. A new network was established in and around Long Ground in February 1999 to monitor

possible movement on the eastern side of the volcano, and the line from Chance’s Steps to

Amersham was re-occupied in June 1999 after a two-year period of inaccessibility.

The ground deformation programme was reinforced from April 1996 by measurements using the

portable Leica System 300, dual-frequency, differential GPS system (Figure 7). This allowed

resolution of changes in line lengths between points on the volcano into vertical and horizontal

components. The equipment was used in the Rapid Static mode and data processed using the

Leica SKI software system (Shepherd et al., 1998). In addition to the MVO GPS system, the

University of Puerto Rico (UPR) has maintained a GPS network at the Soufrière Hills Volcano

since July 1995 (Mattioli et al., 1998). At present, this network consists of three permanent GPS

stations, which continuously transmit data to the MVO, and to the UPR via the Internet. The

MVO established a permanent GPS station at Harris in April 1998. Closer collaboration with the

UPR GPS program from February 1998 allowed the use of data collected by the UPR network in

processing data by MVO GPS stations. During early 1998 and again in March 1999, the MVO

GPS network was upgraded with new receivers and another permanent station established at

South Soufrière Hills. A further MVO permanent GPS station was installed at Spring Estate in

June 1999. Details on the results obtained from the GPS program at the Soufrière Hills volcano

can be found in Shepherd et al (1998), and Mattioli et al (1998).

Electronic tiltmeters have been used intermittently since July 1995 when three stations were

deployed at Spring Estate, Amersham and Long Ground. The stations used high-gain tilt sensors

with bubble-type biaxial platform tiltmeters and digital telemetry designed and built by the

USGS Cascades Volcano Observatory (Murray et al., 1996). Two instruments were moved to

Chance’s Peak in December 1996 because of increased concern over the possibility of sector

collapse, and others were placed at Hermitage and Gages Mountain in 1998 (Figure 9).

However, apart from the stations on Chance’s Peak, there have been no perceptible variations or

trends in the data collected from these tiltmeters (Voight et al., 1998). But the Chance’s Peak

stations were important for discerning cyclic patterns of conduit and dome pressurisation and

- 18 -

these variations were used, together with seismicity changes, to provide accurate short-term

forecasts of activity during June to August 1997 (Voight et al., 1998; 1999).

Estimates of dome volume have been made since January 1996 using compass and Abney level

surveys, supplemented by photographs and theodolite measurements (Watts et al., this volume).

Repeated observations and photos from fixed positions allowed estimates of dome volume to be

made from trigonometry and triangulation, and from simple geometric measurements between

successive photographic images (Sparks et al., 1998). In August 1996, a GPS kinematic survey

method using mobile equipment and laser range-finding binoculars was developed and used to

determine dome volume. During late 1998, a computer program for such GPS data, based on an

approach similar to the photographic method, was developed and used for estimating dome

volumes (Herd et al., 1998).

Measurements of the volume, distribution and character of pyroclastic deposits have also been

integral in understanding the eruption. Regular calculations of deposit volumes using the laser

range-finding binoculars and kinematic GPS method together with isopach mapping of ashfall

(Bonadonna et al., this volume) and simple estimates of deposit thickness have provided

information on total volume of magma extruded (Sparks et al., 1998). Changes in topography of

pathways for pyroclastic flows were monitored giving direct input to hazard assessments, such as

the numerical modelling of runouts (Wadge et al., 1998). Detailed mapping, sampling and

temperature measurements of the new deposits (Cole et al., 1998; Calder et al., 1999; Druitt et

al., this volume; Loughlin et al., this volume; Komorowski et al., 2000) provided critical data

for assessing the likely runouts and areal extents of different types of phenomena. Coupled with

studies of the geology of the Soufrière Hills (Roobol and Smith, 1998; Harford et al., this

volume), this resulted in more accurate assessments of the likely range of future behaviour of the

volcano. Petrological work (Devine et al., 1998a,b; Barclay et al., 1998; Murphy et al., 1998;

2000) has given indications of the state of the magma chamber and the ascent rate of the lava.

Gas monitoring has involved a variety of techniques throughout the course of the eruption. The

environmental impact of sulphurous gases has been monitored by filter packs (Allen, 1996), SO2

- 19 -

diffusion tubes and measurements of acidity levels in rainfall and surface water (Norton, 1997).

Remote sampling of gas concentrations has been undertaken using a Mini-COSPEC Correlation

Spectrometer (Young et al., 1998c) and open-path Fourier Transform Infra-Red spectroscopy

(Oppenheimer et al., 1998); these measurements helped elucidate cyclic emission patterns and

deeper processes. Direct sampling of dome gases was done on one occasion and sampling of the

associated hydrothermal system (Hammouya et al., 1998; Boudon et al., 1998) was carried out

routinely during the early stages of the eruption until sample sites became too dangerous or

covered by volcanic deposits.

4. Outreach

The MVO has taken many positive steps to help educate and inform the public of Montserrat on

the hazards posed by their volcano (Table 2). Throughout its existence, the MVO has

maintained a commitment to an open policy as regards information dissemination and, as far as

is practicable in a crisis, an open-door policy for interaction with the public. This policy attitude

evolved directly from public concerns about the openness of MVO in the early stages of the crisis

when there were differences of opinion about whether the volcano would erupt. Then, it was

perceived by some that the scientists and civil authorities were jointly withholding information and

bad news, and by others, later, that some scientists were working to a hidden political agenda. (As a

British territory, there is no freedom of information act enshrined in law in Montserrat). To

minimise these concerns, the civil authorities endorsed the open information policy and actively

supported the scientists in their public outreach.

Thus, for four years the team has engaged in an extensive programme of public information and

communication, in partnership with civil officials, welfare and health professionals, schools,

churches and the media. Many of these public education exercises were co-ordinated by the

Government Information Service and the GoM Emergency Department in response to specific,

identified requirements, and the public outreach efforts have resulted in a highly informed

population. However, it would be unrealistic to pretend that even this intensive programme of

education and dialogue solved all problems: the long-lived character of the eruption and its

crisis has caused particular difficulties.

- 20 -

During July-September 1995, daily reports and ad hoc scientific briefings were given by the

Chief Scientist on local radio and television. From early 1996 to mid-1998, morning and

evening reports (which were circulated to key government agencies and other interested

organisations by fax) were read out on the local radio station. During periods of heightened

activity, additional lunchtime radio updates were, and still are, presented live by MVO staff.

These enable specific topics to be explained and promote improved immediate public

understanding of the latest volcanic situation. From early 1997, one or more of the MVO senior

scientists (and usually the Chief Scientist in Post) has been interviewed on the radio most weeks

on a Friday, for half an hour or more. Special radio call-in programmes, allowing the public to

interact directly with the scientists and local technical staff, were aired frequently and were

effective in clarifying both issues of uncertainty for the public and important topics of concern.

Articles on the volcano have been regularly contributed to the main local newspaper. In

addition, the Montserrat National Trust and the MVO published jointly a monthly magazine called

SeismiCity News, which provided further explanation of the science, and contained sections aimed

specifically at children because the latter were found to be quite influential opinion-formers within

families. This magazine was instituted in early 1997, following a period in December 1996 when

escalating volcanic activity, rapid and frequent changes in alert level, and general public unease

raised concerns about MVO’s credibility (see Section 5.2, below). SeismiCity News was sponsored

by DFID and circulated to the public free of cost. Other public information bulletins and

information sheets have been prepared in collaboration with the Emergency Department and

circulated to the public. From time to time, individual MVO staff have given more formal public

lectures, which are always broadcast live on national radio, and these were invariably followed by

extensive periods of intense questioning from the audience.

Any organization or group on Montserrat that wanted information or advice from MVO scientists

was able to obtain it. Informal talks and tours of the Observatory were given to groups and

individuals on request. Scientists made visits to schools, various workers’ groups, religious

groups and commercial enterprises, whenever possible or appropriate. From April 1996 to June

1997, when Plymouth was evacuated but the essential services still maintained operations there, a

- 21 -

great deal of effort was invested in making certain that the workers involved were fully aware of the

risks that they faced. Visits were made to these work places and talks given with lengthy question

periods. When some residents of Spanish Point on the north-eastern flank of the volcano refused to

move away from danger, MVO scientists visited their homes and presented information about the

risks that they were running. Bulletin boards were placed in the main occupied town of Salem and

at the airport, with weekly updates to the displays being prepared by two school students, helped by

observatory staff.

All these public education and information efforts have been fully supported by the Government of

Montserrat, the UK Government, and other agencies, and there have been strong and positive

comments on their benefits from all sections of the Montserratian community (including those

abroad). Of course, there have also been times when negative comments were received from the

public. Some were made on occasions when the public became aware of internal difficulties

within the MVO and usually concerned the independence, impartiality, cultural sensitivity, or

otherwise, of the contributions of some members of the team (because of weight of numbers and

the political backdrop, these criticisms tended to be aimed at British scientists). More generally,

however, adverse comments were forthcoming during periods when there appeared to be

differences in perception between the MVO and the public regarding the level of hazards posed

by the volcano, and what mitigation was most appropriate. At those times when strong eruptive

activity waned and visible activity was low for weeks on end, it seemed many of the public

quickly felt that the volcano was doing nothing at all and posed little or no threat, despite the fact

that monitoring instruments continued to indicate that the eruption had not ended. These periods

were often followed by episodes when the volcano escalated to highly dangerous activity often

only in a matter of hours. Keeping the population sufficiently alert to the dangers during such

lulls in activity was a major difficulty and, on some occasions, led to a few vocal people

expressing public disagreement with MVO advice as to the level of risk involved. The situation

was not helped by the fact that, in some instances, although the scientific team had correctly

identified the incipient dangers, these did not always materialise promptly. In the case of the

major 1997 Boxing Day collapse and surge that devastated the south of the island, scientific

anxiety had been elevated for over a year before the event actually happened.

- 22 -

There were a few other times when the credibility of the scientists seemed to be questioned in the

public mind: specific instances might be when individual episodes of heightened activity were

not “predicted”, as in the case the 17 September 1996 major magmatic explosion (although the

general likelihood of such an event occurring at some time had been repeatedly expressed, and

the scientists had invariably indicated that specific predictions of this nature were not possible).

However, throughout the crisis, the majority of the general public were both very supportive of

the work of the Observatory scientists and appreciative of their efforts and, overall, confidence in

the scientists and their assessments generally improved as forecasts of what might happen indeed

came to pass, and as the team got better at both anticipating the behaviour of the volcano and

communicating their insights and understanding.

One novel aspect of the MVO’s outreach has been the extensive use of the Internet as a conduit for

information dissemination. This medium has been used to communicate with the wider community

of scientists, non-resident Montserratians who were otherwise unable to obtain up-to-date

information on the state of the volcano, and volcano enthusiasts generally. This was achieved

through co-operation with the Michigan Technological University where a Webpage containing

MVO reports and other information was maintained for most of the eruption. However, this new

medium of communication also threw up an example of how additional difficulties for relationships

between scientists, and between scientists and politicians, can be jeopardised if care is not taken.

One group unwisely put more information and interpretation than MVO wished to issue into a

report on their Website that was openly available to the surfing public and, in the process, used

some infelicitous phrasing for expressing their views. This, of course, both provoked further public

claims of information suppression on island and, at the same time, irritated and offended some

colleagues. Eventually, in September 1999, the MVO was able to institute its own Webpage, which

can be found online at: <http://www.mvomrat.com>.

The Internet has also been used by MVO scientists to obtain valuable comments and to maintain

contact with professional colleagues throughout the world. This has enabled access to a wider

variety and depth of scientific thought than would have otherwise been available to the on-island

- 23 -

monitoring effort alone. As one example, in October 1996 there were fears that a sector collapse

could generate a large volcanic landslide and associated tsunami. Contact with numerical modelling

specialists in France and the UK via the Internet enabled sophisticated calculations to be done and

thus allay concerns as to the potential size of any tsunami generated in this way (Heinrich et al.,

1999). Thus, in principle and in practice, the Internet allows an observatory to consult very rapidly

with the world’s leading experts. For the group of MVO Chief Scientists, communication by email

was also used effectively to keep each informed of developments on Montserrat between rotations.

Perhaps one shortcoming of the public education programme organised by the MVO has been

the absence of a well-defined structure and lack of a formal mechanism for gauging the

effectiveness of the outreach. This has been due mainly to the lack, until September 1999, of a

suitably trained permanent member of staff to co-ordinate such activities. Apart from the regular

daily public scientific reports, many of the methods used to disseminate information to the public

have not been sustained through the course of the eruption. Often, they were either a one-off

response to a particular concern within the scientific team regarding public awareness and safety,

or resulted from the (short-term) presence at MVO of personnel who had specific interests in

public education. Greater support for such outreach activities became possible with the

recruitment to the MVO of a full-time Information Officer.

One of the few occasions on which comments were solicited by MVO from the public occurred

immediately following the tragic events of 25 June 1997 (Loughlin, Aspinall et al., this volume;

Loughlin, Baxter et al., this volume). This consisted of a series of interviews with people who

were in the Exclusion Zone at the time of the event, and those involved in rescuing survivors.

While the information gathered was intended to provide documentary evidence on details of the

eruption itself, the responses also allowed MVO to investigate how well people felt they had

been informed of the dangers associated with this volcano. The results of the interviews suggest

that the MVO’s efforts at public education had been largely effective in informing the public of

the risks, and showed that individuals’ decisions regarding entry into high-risk areas on the

volcano were complex, and could not be attributed to any single causative factor. People in

Montserrat appeared willing to take high risks if the perceived return or benefit to them or the

- 24 -

community was considered of sufficient importance (e.g. to continue farming), or if the

alternatives were repugnant (e.g. living in shelters).

In the wider context of scientific advancement, the MVO has encouraged collaboration with

external scientists who wish to undertake research at the Soufrière Hills volcano. Attempts have

been made to ensure that such collaboration is done within a defined framework and following

established protocols. A set of guidelines was drafted in 1997 to serve as operational procedures

for researchers working at the MVO, and for those wishing to collaborate with the MVO. Due to

number of problems, mainly relating to the management structure at the MVO, these protocols

have not always been strictly adhered to or consistently implemented. However, most scientists

interested in working in Montserrat have attempted to follow the guidelines provided and have

established contact with the Chief Scientist or, latterly, the Director prior to their arrival on the

island.

5. Crisis management

The crisis management of the Montserrat volcanic eruption has been characterised by extremely

close interactions between the scientific team, civilian authorities and public. In addition, a

custom-made Montserrat Alert System was drafted which was specific to the conditions on the

island. This has been supplemented by volcanic hazard mapping, vulnerability assessment, risk

maps, formal and informal reports, scientific briefings, public meetings, news media interviews

and articles. During the first few months of the eruption, a framework was set up to facilitate

interaction among scientists and also between the scientists and the administrative authorities.

An existing volcanic hazard and risk assessment prepared originally by Wadge and Isaacs (1987;

1988) was initially adopted, and then adjusted, as the nature of the eruption became clearer. The

MVO has subsequently produced a series of hazard maps for the authorities and worked closely

with civilian authorities to produce risk maps with zones indicating the level of danger and also

whether occupation is advised or allowed. These maps are published in the local newspaper and are

widely circulated. Scientific meetings of the MVO team provided the basis for a consensus

opinion on the state of activity, which was presented at government briefings by the Chief

Scientist.

- 25 -

5.1 Government Briefings

Since the onset of eruptive activity at the SHV in July 1995, the MVO have provided regular

briefings to local authorities on the state of the volcano. Very close and frequent

communications were developed between the scientists at the Observatory and the political

directorate and other civil authorities. The key officials became thoroughly familiar with both

the scientific understanding of the volcano and the implications of scientific uncertainty, so that

the most appropriate decisions could be made for civil response and mitigation.

From July to September 1995, the Chief Scientist, often accompanied by another senior team

member, provided daily briefings to the crisis management teams. During this period, numerous

informal briefings were given to the civil authorities during their frequent visits to the MVO,

which was then located next to government headquarters. As the pattern of events became more

certain and dome growth began, the briefings became more structured and were given at least

weekly to the crisis management team. This consisted of the Governor and the Chief Minister,

senior civil servants and aid agency personnel. In addition, the Governor and Chief Minister

attended at least one of two internal scientific meetings held weekly at the MVO to discuss the

state of the volcano. This continued up to the middle of 1997 when a system of bi-weekly

briefings was initiated, which remains in effect to the present. The first of these is usually given

on a Monday to members of the Volcanic Executive Group, the highest decision-making body

involved in disaster management on the island, comprising the Governor, the Chief Minister,

senior ministers of government, the Chief of Police, the Chief Scientist and the Head of the

Emergency Department. The second briefing is given on the Friday to a larger group called the

Volcanic Management Support Group which consists of senior civil servants and key sector

agencies. The briefings are usually given by the Chief Scientist (now the Director) or Deputy

Chief Scientist and consist of a synopsis of activity for the week along with an up-to-date

assessment of what may be expected from activity during the ensuing week. Advice is also

given on various aspects of crisis management, from the scientific perspective.

- 26 -

As noted earlier in Section 2, one of the peculiarities of Montserrat was the existence of a

complex administrative and political system of government due to the (then) status of the island

as a Dependent Territory of the United Kingdom. There were two major ministries of the UK

Government (DFID and the Foreign and Commonwealth Office) involved in the UK response to

the crisis, a Governor appointed by the UK Government and a local democratically elected

government. Inevitably these arrangements at times led to a lack of clarity about where, and in

what form, essential scientific information should be provided. Thus, sometimes more than one

line of reporting was instigated with civil servants in London, with the Governor and his Office

on Montserrat, or with the political leaders of Montserrat, each wanting information and advice.

In some instances the communications that are essential with multiple reporting lines were

inadequate and led to difficulties. The situation was exemplified early in the eruption by the

appointment of independent consultants to the FCO and DFID at the same time as there was an

on-going reporting line in Montserrat, originating from the embryonic MVO. This problem was

soon recognised and strenuous efforts were made to establish a single line of communication to

the authorities. However, the problem of how to communicate uniform and coherent scientific

advice to several interested arms of the two governments, in parallel in two separate countries in

different time zones, remained a significant complication throughout the eruption. One of the

major difficulties was that, even when a single channel was active, the content of urgent

scientific advice easily became garbled in being passed from one uninitiated civil servant to

another, or even distorted when an ulterior political agenda was in play.

The introduction of six-monthly risk assessments from August 1997 onwards, formally prepared

by a group of senior MVO scientists, improved the situation. These reports were delivered to all

interested parties simultaneously and the principle was established that the Chief Scientist in Post

(more recently, the Director) should be the channel for all immediate scientific advice. The

problems that were encountered emphasise the importance to scientists of establishing clear lines

of communication in a volcanic crisis, although this can prove far from straightforward if, as

noted above, there is duality of political accountability, or any of the main political or

administrative actors seek to manipulate arrangements for their own ends.

- 27 -

5.2 Alert Scheme

From the onset of the eruption of the Soufrière Hills volcano, the Montserrat Alert Scheme has

served as a useful tool for guiding the management and response actions to varying levels of

volcanic activity. The scheme has undergone several iterations since the initiation of eruptive

activity on 18 July 1995. The first system used was a generic one patterned on that from the

UNDRO handbook on Volcanic Emergency Management (Table 3). It was not specifically

tailored to the peculiarities of the environment on Montserrat and proved inadequate to guide

mitigation efforts as the eruption slowly evolved.

The explosive eruption that occurred in mid-September 1996 was a significant departure from

the extrusive style of eruption that had occurred up to that date. Consequently, it was necessary

to revise the existing volcanic alert scheme to incorporate new information pertaining to the

explosive hazard potential of the volcano and the corresponding impact on vulnerability and

response. This resulted in the first draft of an Alert Scheme specifically for Montserrat (Table

4), a framework that provided a guide for actions to be taken by civil authorities within defined

risk zones around the volcano at specified alert levels determined by the MVO Team (Figure 10).

It underwent a number of revisions and alterations as activity escalated. During the volcanic

events of early December 1996, when the Galway’s segment of English’s Crater appeared to be

under severe stress, differences in interpretation of these emergency procedures emerged

between scientists at the MVO, the civil administrators and the public. The alert procedures

came under closer scrutiny in mid-December 1996 when, upon raising the alert level in response

to increased activity at the volcano, only a small percentage of the residents in a particular

community that was deemed vulnerable heeded the official notice to evacuate. The increased

activity that triggered the raising of the level on this occasion had all but died down by the time

the public announcement of evacuation was made.

Once volcanic activity had encroached on most of the built-up areas and vital services around the

volcano in the tragic events of 25th June 1997, the Alert Scheme could no longer be retained as a

microzonation management tool. The approach was then simplified to a system whereby the

entire island was divided into a ‘northern zone, a ‘central’ zone and an ‘unsafe’ zone, with severe

- 28 -

restrictions being placed on entry into the unsafe or Exclusion Zone (Figure 11). This system

was considered applicable for as long as the eruption continued in the same vein and, with minor

administrative changes, has remained in effect ever since. The boundaries of each zone were

modified slightly as activity at the volcano decreased during 1998 (Figure 12), and were used to

determine the zone from which people remain evacuated.

5.3 Expert Elicitation

In an attempt to provide good advice to the decision-makers in the form of scientific consensus, a

formalised procedure for eliciting expert judgements was adopted by the MVO in August 1995.

This became necessary due to the need to provide urgent daily advice to the authorities at a stage

when the nature and magnitude of future volcanic activity was most uncertain, and time for repeated

and protracted scientific debates amongst a large group was just not available. Arriving at an agreed

position each day by committee took an increasing amount of time and this became frustrating both

for the scientists, who needed to press on with observations and measurements, and for the

authorities who found it preferable to have a rapid and definitive answer provided as promptly as

possible.

The formalized procedure which was introduced is based on the "Classical Model" for structured

expert judgement (Cooke, 1991), following the suggestion first made at the International

Symposium on Large Explosive Eruptions in Rome in 1993 to consider using this approach for

producing a collective scientific opinion in a volcanic crisis (Aspinall and Woo, 1994). The

method performs weighted combinations of expert judgments, where weights are determined by

'calibration' and 'information' performance on questions for which the true values are known, or

become known a posteriori. However, in application, the procedure had to be adapted to the

needs of real time crisis management. In open meetings, where some preferred to hold their

counsel, others were more forceful in giving their opinions so, from the outset, it was felt better

to focus the procedure on quantifying the informativeness of each individual's views, rather than

rely too heavily on a hurried and questionable calibration score. (The problems of calibrating the

expertise of a group of volcanologists are non-trivial at the best of times, let alone with an

eruption going on outside the window!). This emphasis on informativeness (or degree of

- 29 -

conservatism) meant there was an implicit assumption that roughly equal expertise attached to

each member of the team, a reasonable supposition for the initial scientific group assembled in

Montserrat. In fact, when the concepts of this approach were being introduced to the

administrators and scientists it was novel to most, and several pressed for the scheme to be

administered in such a way that no single participant was ever zero-weighted: all views would be

used with some weight in the decision process. However, in order to bring an element of

calibration into play, a preliminary set of five suitable seed questions was hastily prepared.

While this set was aimed primarily at measuring the individual's informativeness factor, it was

also used to provide a basic measure of how well each person might make quick judgements on

issues related to safety and hazard mitigation in an emergency. The latter element of the exercise

was undertaken, with general agreement, as an exploratory trial of the method in application in a

live crisis. Therefore, the emphasis in the seed questions was placed on judgement of generic

hazard-related variables and factors - such as percentage casualties from different sorts of

volcanic action; durations of eruptions, and so on - rather than on the science of volcanology

itself. Devising a broad range of fair technical questions to encompass geologists of all kinds,

geodesists, geochemists and seismologists, for full execution of the procedure, was impracticable

under the circumstances.

As the emergency progressed month by month, three shortcomings with the adopted approach

emerged. First, having once used the initial set of seed questions for calibration, for uniformity

these had to be repeated for scientists subsequently joining or replacing others in the team; over

the course of three year’s activity, a total of more than sixty individual scientists and technical

specialists participated in the different elicitations, although there were generally only between

five and twenty present at any one time (to everyone's credit, the calibration facts were not

leaked to new arrivals at the observatory). Secondly, an experienced technical facilitator was not

always available in Montserrat to supervise the elicitations, and thirdly, there was a fluctuating

mandate to undertake formal elicitations as conditions and levels of anxiety varied.

Notwithstanding these slight drawbacks, the approach was very successful in handling key on-

going issues, such as regular assessments of the scientific team's "comfort" with the current alert

level.

- 30 -

There were divided opinions amongst the scientific group as regards the utility and soundness of

the method, based as it is on concepts of subjective probability which are new to many, and, with

an experienced technical facilitator not always available at the MVO, the technique was not used

continuously. However, the majority of scientists supported the approach. After being

discontinued in late 1995, the methodology was again re-introduced after the 17 September 1996

explosive eruption. It was then used more widely in the decision making process, being integrated

into the alert scheme and used on a weekly basis to assist the Chief Scientist in Post to determine the

appropriate level of alert. The method continued in operation in this context until the alert scheme

was simplified in July 1997. Its use for assessing the appropriate alert level gave continuity to the

decision making process, and provided a traceable record through time of the views of the scientific

team (see, for example, Figure 13). Further details on the theory and potential application of this

technique to volcanic crises are given in Aspinall and Woo (1994), and a brief account of its use at

the MVO is contained in Aspinall and Cooke (1998).

5.4 Risk and Hazard Assessments

With the change in eruptive style that occurred in 1997, a more structured approach to hazard and

risk assessment was invoked. The first of what became regular reviews of the state of the volcano

and its activity was undertaken in December 1997, and involved several of the senior scientists who

had been involved in work at the MVO. These assessments provided a considered and

comprehensive review of the most recent activity and arrived at a prognosis for the immediate (6

months) and more distant future (> 5 years).

The various possible eruptive scenarios, which these reviews developed for the volcano on the

basis of available evidence, were usually summarised in the form of event probability trees, for

presentation to decision-makers. The formalised elicitation methodology mentioned above in

Section 5.3 has been used to provide consensus probability value distributions for input onto the

branches of the trees (see Figure 14), in conjunction with traditional decision-conferencing and

numerical modelling. In turn, these probability distributions are used as inputs to Monte Carlo

simulations of population casualty estimates, as part of the overall on-going assessment of

- 31 -

volcanic hazard and risk that the MVO undertakes for guiding civil mitigation decisions. As the

eruption progressed, and its behavioural patterns changed, these event trees became increasingly

more specific in terms of threats to particular areas, and as population zone occupancies changed,

the casualty impact simulations also grew in complexity and detail (see Figure 15, for examples

of F-n exposure curves for Montserrat, where the F-n curve is a risk assessment concept

representing the exceedance probability of frequency of a cumulative number of casualties). It

was essential that the varied opinions from a diverse and changing group of experts were

incorporated into these models in a structured way in order to maintain balance and consistency

in the outcomes, and the approach to this problem was one more example of how creative hazard

management techniques have been developed within the MVO during the crisis.

The risk assessments have been updated at six-monthly intervals (or as circumstances demanded)

and, in the later stages of the crisis, the results have formed the main basis for planning by the

civilian authorities. They have provided outlooks which, as events proceeded, were not

overturned by any major surprises in the behaviour of the volcano.

5.5 Acceptable Risk

In Montserrat, the administrative authorities held responsibility for specifying the level of

acceptable risk throughout the volcanic crisis. From the onset of the eruption, their stated policy

was to set the tolerability threshold at a low level, and to maintain it there. Whenever the risk to

an area near the volcano was deemed to be approaching life-threatening levels, preventative

evacuation of such areas was undertaken. Having accepted accountability for determining the

threshold of risk acceptability, the government also assumed the responsibility for providing

relief assistance to the population that was evacuated from the endangered areas. Bans were

placed on the re-entry to such areas until the danger had subsided, and in order to apply them,

police checkpoints were established at key entry points. The purpose of these checkpoints was to

limit access to authorised essential personnel only, thereby restricting the number of entries to a

level manageable in accordance with the emergency and risk management plans and capability.

However, because of a number of social and political factors, these exclusions were not totally

enforceable.

- 32 -

Furthermore, it is inevitable that there is also a small, residual risk involved in living and

working just outside the limits of any marginal exclusion zone. Thus, by remaining close to, or

by making unauthorised entry to the exclusion zone, many individuals in Montserrat have been

able to set their own level of acceptable risk to a certain extent. The actual threshold that

individuals assigned to acceptable risk varied throughout the crisis. At the start of the crisis,

when public knowledge of the volcano was not as informed as later and uncertainty about the

future was greatest, many persons, having heard stories of the 1902 Martinique disaster, accepted

only a low risk and heeded evacuation recommendations. As the eruption progressed and both

knowledge of the hazards and experience of this particular volcano increased, risk thresholds

appeared to change, and many individuals seemed willing to accept higher levels of personal

risk.

6 Summing up

A new institution, the Montserrat Volcano Observatory, has been established on Montserrat to

monitor the Soufrière Hills volcano. While the MVO owes its existence to the eruption that

started in July 1995, it was not created from a standing start: it is founded on the Seismic

Research Unit’s long experience and knowledge of crises in the islands as regional agency for

earthquakes and volcanic activity the Eastern Caribbean, and on the strength of UK volcanology,

supported by the scientific project management experience of BGS. However, in the present

case, even such resources as did exist prior to the eruption were not been fully utilised in good

time for mitigation (e.g. the overlooked volcanic hazard assessment in the 1987 Wadge and

Isaacs report). Moreover, the nature of the present crisis in Montserrat changed from one that

was initially expected to be short-lived, with high levels of public sensitisation-to-threat, to one

that became unusually protracted, with consequent difficulties in sustaining both public anxiety

and official caution. As a result, fresh new challenges were continually arising for the scientific

team, both in respect of the behaviour of the volcano itself, and in terms of necessary interactions

with administrators and the public. Under the circumstances, developing new capabilities for

hazard monitoring and instituting innovative mitigation measures within a besieged community

in a prolonged emergency, with a frequently changing cast of participants, was fraught with

- 33 -

difficulties and pitfalls. It is to the credit of all involved that the performance of the MVO in

terms of the quality and timeliness of its scientific advice was as good as it turned out to be.

Many valuable lessons have been learned about the processes of interfacing with decision-

makers, key actors and stakeholders, and the public, some of which are now being transferred

through regional disaster preparedness organisations to other islands, e.g. Dominica, where a

recent volcano-seismic crisis has raised public concern. For long periods in the Montserrat

crisis, the micromanagement of hazards on a small island with an erupting volcano and a

population intent on remaining there pushed the society and its politicians to use to the limit such

guidance as the scientists could provide: the implications of this experience, and how it might

redefine the scientific rôle in volcanic crises, are worthy of more detailed appraisal.

The establishment of a local entity with responsibility for monitoring a single volcano, operating

at arm’s-length from the regional agency, represents an unusual departure from conventional

ways of approaching the problem. Individual observatory buildings and dedicated local teams

have been set up before in the region, e.g. on St. Vincent, and on the French islands of

Martinique and Guadeloupe, but these all operate under the aegis of bigger organisations.

Exceptional circumstances in Montserrat (dual political hierarchy; protracted, slowly escalating

eruption) were major contributory factors, together with a multitude of other influences, in

moving towards the setting up of this separate body, the MVO. The nature of this particular

crisis required prolonged staffing and an extended range of skills, and thus demanded more input

than the SRU could provide from its own, limited resources. This raises the fundamental

question (which will probably recur in future at other small volcanic island states in the region)

as to whether an alternative scheme for resourcing the monitoring effort through the SRU could

have been implemented. It is too facile to argue that there were traditional colonial political

pressures making sure that management control remained in the hands of the home/donor

country, and that this is the basic explanation for the way things developed as they did. In the

earliest stages of the crisis, attempts were made to configure a viable way forward for SRU as

lead team in Montserrat, but a combination of singular personalities, unique institutional

conditions and characteristics, and many other internal and external factors, were sufficient to

divert the evolution of MVO into other channels, at least for the time being. The SRU

- 34 -

experience in Montserrat also underlines yet again the clichéd but difficult challenge that

volcanologists need to be able to convince governments to make long-term investments in hazard

mitigation measures and related resources so that good quality, cost effective solutions are

available when disaster strikes.

In the present case, the solution to the issue of staffing a monitoring team for dealing with an on-

going crisis was found by combining, on a rotating basis, a corps of young scientists and Ph.D.

students with a kernel of experienced volcanologists from various different institutions. While

this has been beneficial overall, and has been invaluable as a training regime for the young

scientists, it does depart from the way volcanic eruptions have been handled in the region in the

past, and at many volcanoes elsewhere in the world. This course of action has not been without

drawbacks: for the Chief Scientist or Director in post, additional people also mean extra

management overheads and resources (vehicles, radios, hot-suits, etc), and increased anxiety

about safety of personnel. The ambitions, goals and experience of some may be more oriented to

academic research and collecting data for that purpose, than to hazard assessment and mitigation;

and the circumstances of a major volcanic crisis are not always ideal for fostering a student’s

best interests or for helping with inter-personal relationships that can become intense and

exaggerated in crisis conditions. Not least, it is debatable what effect witnessing a bad

catastrophe, such as the fatalities of 25 June 1997, might have on a young, impressionable

individual, especially when that individual may, unduly, feel some responsibility as a member of

the monitoring team. That being said, the Soufrière Hills eruption crisis has clearly been a major

stimulus for many young scientists including, most importantly, several young Montserratians.

In recent discussions, the two governments and the various agencies involved in Montserrat have

been seeking to map out the long-term direction of the MVO and its eventual integration into

regional monitoring. The resurgence of dome growth in November 1999 may delay these

initiatives, and more immediate attention is being focused on drawing up a monitoring strategy

for coping with yet another major episode of persistent eruptive activity. However, it is foreseen

that, in due course, the legacy of resources, knowledge and experience gained at the MVO during

this eruption will pass locally into the custodianship of trained Montserratians, and will enhance

- 35 -

the future management of volcanic crises through a strengthened regional agency. In the

meantime, with the eruption on-going, the new institution will have the opportunity to play an

important rôle in advancing volcano monitoring and volcanology.

Acknowledgements

The people of Montserrat, so badly affected by the disaster in their island, have given wholehearted support to MVO, and their

encouragement and expressions of confidence have been greatly appreciated by all the scientists involved. The MVO itself has

benefited from the hard work of many individuals who have all contributed to its development and, in particular, the efforts of

local staff and individuals, who selflessly devoted their time and energies when families were under threat and homes being lost

to the volcano. Scientists and technical staff from the Seismic Research Unit of UWI, the British Geological Survey, the USGS

Volcano Disaster Assistance Program, the French Volcano Observatories in Martinique and Guadeloupe, the University of Puerto

Rico, Brown University, Lancaster University, Bristol University, the Open University, Cambridge University and numerous

other institutions have all contributed to the development of the MVO, and hence the content of this paper. The authors wish to

single out and acknowledge the wholehearted dedication, commitment and unique contributions made in this respect by their

colleagues from SRU: Lloyd Lynch and Richie Robertson, two of the former co-Chief Scientists at MVO. Prof. C. Newhall and

Dr. J-C. Komorowski provided a range of constructive and challenging comments as referees. Funding for the MVO has been

provided by the Government of Montserrat and by the British Government through its Department for International

Development. For relevant authors, this paper is published by permission of the Director, British Geological Survey (NERC).

References

Ahmad, R. (ed.) (1996). Abstracts from The Second Caribbean Conference on Natural Hazards and Disasters: Science, Hazards And Hazard Management, held 9-12 October 1996, Kingston Jamaica. Publication No. 1, Unit for Disaster Studies, Dept. Geography and Geology, University of the West Indies, Jamaica, pp26-41.

Allen, A.G. (1996). The influence of local volcanic emissions on the atmospheric environment of Montserrat, West Indies. Abstract, Applied Geoscience 96, Warwick, U.K.

Ambeh, W.B., L.L. Lynch, J. Chen, J. and R.E.A. Robertson (1998). Contemporary seismicity of Montserrat, West Indies (Abstract). In: Transactions of the 3rd Geological Conference of the Geological Society of Trinidad and Tobago and the 14th Caribbean Geological Conference (eds. W. Ali, A. Paul & V. Young On), Vol. 1, pp. 1-2.

Aspinall, W.P., H. Sigurdsson and J.B. Shepherd (1972). Eruption of Soufriere Volcano on St. Vincent Island, 1971 - 1972. Science, Vol. 181, pp. 117 - 124.

Aspinall, W.P., L.L. Lynch, R.E.A. Robertson, K.C. Rowley, R.S.J. Sparks, B. Voight and S.R. Young (1998a). The Soufrière Hills Eruption, Montserrat, British West Indies: Introduction to Special Section, Part 1. Geophys. Res. Lett., Vol. 25, No. 18, pp. 3387.

Aspinall, W.P., A.D. Miller, L.L. Lynch, J.L. Latchman, R.C. Stewart, R.A. White and J.A. Power (1998b). Soufrière Hills eruption, Montserrat, 1995-1997: volcanic earthquake locations and fault plane solutions. Geophys. Res. Lett., Vol. 25, No. 18, pp. 3397-3400.

Aspinall, W.P. and R.M. Cooke (1998). Expert judgement and the Montserrat Volcano eruption. In: Proceedings of the 4th International Conference on Probabilistic Safety Assessment and Management PSAM4, September 13th -18th, 1998, New York City, USA (eds. Ali Mosleh and Robert A. Bari), Vol.3, pp. 2113-2118.

Aspinall, W.P. and G. Woo (1994). An impartial decision-making procedure using expert judgement to assess volcanic hazards. Atti dei Convegni Lincei, 112, 211-220.

Barclay, J., M.J. Rutherford, M.R. Carroll, M.D. Murphy, J.D. Devine, J. Gardner and R. S. J Sparks (1998). Experimental phase equilibria constraints on preeruptive storage conditions of the Soufrière Hills magma. Geophys. Res. Lett., Vol. 25, No. 18, pp. 3437-3650.

Bonadonna, C., E.S. Calder, C. Choux, A-M. Lejeune, S.C. Loughlin, M.D. Murphy, G.E. Norton, D. Pyle, G. Ryan, S. Young and R.S.J. Sparks (2000). Tephra fall deposits in the 1995-1998 eruption of the Soufrière Hills Volcano, Montserrat. This volume.

- 36 -

Boudon, G., B. Villemant, J-C. Komorowski, P. Ildefonse and M.P. Semet (1998). The hydrothermal system at Soufrière Hills Volcano, Montserrat (West Indies): characterisation and role in the ongoing eruption. Geophys. Res. Lett., Vol. 25, No. 19, pp. 3693-3696.

Calder E. S., P.D. Cole, W.B. Dade, T.H. Druitt, R.P. Hoblitt, H.E. Huppert, L. Ritchie, R.S.J. Sparks and S.R. Young (1999). Mobility of pyroclastic flows and surges at the Soufrière Hills Volcano, Montserrat. Geophys. Res. Lett., Vol. 26: No. 5, 537-540.

Clay, E., C. Barrow, C. Benson, J. Dempster, P. Kokelaar, N. Pillai and J. Seaman (2000). An Evaluation of HMG’s Response to the Montserrat Volcanic Emergency. London: Department for International Development Evaluation Report EV635, 2 volumes.

Cole, P.D., E.S. Calder, T.H. Druitt, R. Hoblitt, R. Robertson, R.S.J. Sparks and S.R. Young (1998). Pyroclastic flows generated by gravitational instability of the 1996-1997 lava dome of Soufrière Hills. Geophys. Res. Lett., Vol. 25, No. 18, pp. 3425-3428.

Cooke, R.M. (1991). Experts in Uncertainty. Oxford Univ. Press, 321pp. Curnow, R.N. (1999). Unfathomable nature and Government policy (the address of the President of the Royal Statistical

Society). The Statistician, J. Roy. Statist.Soc., Series D, Vol. 48 Part 4, 463-476. Devine, J.D., M.D. Murphy, M.J. Rutherford, J. Barclay, R.S.J. Sparks, M.R. Carroll, S.R. Young and J.E. Gardner (1998a).

Petrologic evidence for pre-eruptive pressure-temperature conditions, and recent reheating, of andesitic magma erupting at the Soufrière Hills Volcano, Montserrat, West Indies. Geophys. Res. Lett., Vol. 25, No. 19, pp. 3669-3672.

Devine, J.D. M.J. Rutherford and J.E. Gardner (1998b). Petrologic determination of ascent rates for the 1995-1997 Soufrière Hills Volcano andesitic magma. Geophys. Res. Lett., Vol. 25, No. 19, pp. 3673-3676.

Druitt, T.H., P.D. Cole, L. Ritchie, R.S.J. Sparks, E. Calder, R.P. Hoblitt, S.C. Loughlin, G.E. Norton, B. Voight and R. Watts (2000). Surge-derived pyroclastic flows (Montserrat, Antilles): a new volcanic hazard. This volume.

Fiske, R.S. (1984). Volcanologists, journalists, and the public: a tale of two crises in the Eastern Caribbean. In: Explosive Volcanism: Inception, Evolution and Hazards. Nat. Res. Counc. (U.S.) Geophysics Study, National Academy Press, Washington D.C., pp. 170 –176.

Hammouya, G., P. Allard, P. Jean-Baptiste, F. Parello, M.P. Semet and S.R. Young (1998). Pre- and syn-eruptive geochemistry of volcanic gases from Soufrière Hills of Montserrat, West Indies. Geophys. Res. Lett., Vol. 25, No. 19, pp. 3685-3688.

Harford C, M. Pringle, R.S.J. Sparks and S.R. Young (2000). Evolution of Montserrat using 40Ar/39Ar geochronology. This volume.

Heinrich, P., S. Guiborg, A. Mangeney and R. Roche (1999). Numerical modeling of a landslide-generated tsunami following a potential explosion of the Montserrat volcano. Phys. Chem. Earth (A), 24(2): 163-168.

Herd, R., E. Calder, J. Barclay, R.S.J. Sparks, M. James, P. Cole, C. Harford, R. Watts, G. Norton, S. Young and R. Robertson (1998). Methods used to survey the lava dome and associated products at the Soufrière Hills Volcano, Montserrat. MVO Open File Report 98/16.

Jackson, P., J.B. Shepherd, R.E.A. Robertson and G. Skerritt (1998). Ground deformation studies at Soufrière Hills volcano, Montserrat I: Electronic distance meter studies. Geophys. Res. Lett., Vol. 25, No. 18, pp. 3409-3412.

Komorowski, J-C., G. Norton, G. Boudon, B. Voight, R. Watts, S. Young, S. Sparks, R. Herd and C. Walker (2000). Field-relationships and texture of the striking debris-avalanche deposit from the 1997 flank-collapse of Soufrière Hills Volcano, Montserrat. Geoscience 2000, abstract volume, p157.

Lee, W.H.K. (1989). PC-SEIS: a toolbox for seismic data acquisition, processing, and analysis. IASPEI Software Library, Vol. I (ed. W.H.K. Lee), 284p.

Lee, W.H.K. and C.M. Valdes (1989). HYPO71PC. IASPEI Software Library, Vol. I (ed. W.H.K. Lee), pp. 203-236. Loughlin, S.C., W.P. Aspinall, A. Clarke, P.D. Cole, R.A. Herd, J.L. Latchman, A-M. Lejeune, R. Luckett, M. Mangan, A.D.

Miller, L. Pollard, L. Rodriquez, K.C. Rowley, R.S.J. Sparks, I.M. Watson, R. Watts, R. and MVO staff. (2000). Pyroclastic flows generated by the June 25 1997 dome collapse, Soufrière Hills volcano, Montserrat. This volume.

Loughlin, S.C., P.J. Baxter, C.L. Harford, B. Darroux, A.D. Miller and W.P. Aspinall (2000). Soufrière Hills volcano, Montserrat, WI: summary of eyewitness accounts of the 25th June 1997 pyroclastic flows and implications for disaster mitigation. This volume.

Mattioli, G.S., T.H. Dixon, F. Farina, E.S. Howell, P.E. Jansma and A.L. Smith (1998). GPS measurement of surface deformation around Soufrière Hills Volcano, Montserrat from October 1995 to July 1996. Geophys. Res. Lett., Vol. 25, No. 18, pp. 3417-3420.

Miller, A.D., R.C. Stewart, R.A. White, R. Luckett, B.J. Baptie, W.P. Aspinall, J.L. Latchman, L.L. Lynch and B. Voight (1998). Seismicity associated with dome growth and collapse at the Soufrière Hills Volcano, Montserrat. Geophys. Res. Lett., Vol. 25, No. 18, pp. 3401-3404.

Montserrat Volcano Observatory Team (1997). The Ongoing Eruption in Montserrat. Science, Vol. 276, pp. 371-372 Murphy, M.D., R. S. J Sparks, J Barclay, M R. Carroll, A-M. Lejeune, T S. Brewer, R. Macdonald, S. Black and S. Young

(1998). The role of magma mixing in triggering the current eruption at the Soufrière Hills Volcano, Montserrat, West Indies Geophys. Res. Lett., Vol. 25, No. 18, pp. 3433-3436.

Murphy, M.D., R.S.J. Sparks, J. Barclay, M.J. Carroll and T.S.Brewer (2000). Remobilization of andesite magma by intrusion of mafic magma at the Soufrière Hills volcano, Montserrat, West Indies. Journal of Petrology, Vol. 41, pp. 21-42.

- 37 -

Murray, T.L., J.W. Ewert, A.B. Lockhart and R.G. LaHusen (1996). The integrated mobile volcano-monitoring system used by the Volcano Disaster Assistance Program (VDAP). In: Monitoring and Mitigation of Volcano Hazards (eds R. Scarpa and R.I. Tilling), Springer, Berlin pp. 315-362.

Neuberg, J., B. Baptie, R. Luckett and R. Stewart (1998). Results from the broadband seismic network on Montserrat. Geophys. Res. Lett., Vol. 25, No. 19, pp. 3661-3664.

Norton, G.E. (1997). Summary of the environmental effects of volcano emissions, Soufrière Hills volcano, Montserrat. MVO Open File Report, 97/15.

Norton, G.E., W. P. Aspinall, B. Baptie, C. Bonadonna, P.W. Francis, R.A. Herd, C. Harford, R.P. Hoblitt, A.D. Jolly, S.C. Loughlin, R. Luckett, L. Ritchie, K.C. Rowley, R. S. J. Sparks, R. Watts, S. R. Young (2000). Residual volcanic activity at the Soufrière Hills Volcano Montserrat: 1 March 1998 – 3 November 1999. This volume.

Office of Science and Technology (1997). The Use of Scientific Advice in Policy Making. London: Department of Trade and Industry, March.

Office of Science and Technology (1998). The Use of Scientific Advice in Policy Making: Implementation of the Guidelines. London: Department of Trade and Industry, July.

Oppenheimer, C., P. Francis and A.J.H. Maciejewski (1998). Spectroscopic observation of HCl degassing from Soufrière Hills Volcano, Montserrat. Geophys. Res. Lett., Vol. 25, No. 19, pp. 3681-3684.

Pattullo, Polly (2000). Fire from the Mountain. Constable, London. 217pp. Robertson R.E.A., Aspinall W.P., Herd R.A., Norton G.E., Sparks R.S.J. and Young S.R. (2000) The 1995-1998 eruption of the

Soufrière Hills volcano, Montserrat, WI. Phil. Trans. Roy. Soc. London, Series A, Vol. 358, No. 1770, 1619-1637. Robertson, R.E.A. and G. Wadge (1996). Overview of the Soufrière Hills eruption 1995-96. In: The Soufrière Hills Eruption,

Montserrat (ed. G. Wadge), Geological Society of London, pp. 9-11. Roobol, M.J. and A.L. Smith (1998). Pyroclastic stratigraphy of the Soufrière Hills Volcano, Montserrat: Implications for the

present eruption. Geophys. Res. Lett., Vol. 25, No. 18, pp. 3393-3396. Shepherd, J.B., J.F. Tomblin and D.A. Woo (1971). Volcano-seismic crisis in Montserrat, West Indies, 1966-67. Bull.

Volcanol., Vol. 35, No. 1, pp. 143-163. Shepherd J.B., W.P. Aspinall, K.C. Rowley, J. Pereira, H. Sigurdsson, R.S. Fiske and J.F. Tomblin (1979). The eruption of

Soufriere Volcano, St. Vincent April - June 1979. Nature, Vol. 282, 24-28. Shepherd, J.B., R.A. Herd, P. Jackson and R. Watts (1998). Ground deformation measurements at the Soufrière Hills Volcano,

Montserrat II: Rapid static GPS measurements, June 1996-June 1997. Geophys. Res. Lett., Vol. 25, No. 18, pp. 3413-3416. Sparks, R.S.J., S.R. Young, J. Barclay, E.S. Calder, P. Cole, B. Darroux, M.A. Davies, T.H. Druitt, C. Harford, R. Herd, M.

James, A.M. Lejeune, S. Loughlin, G. Norton, G. Skerritt, M.V. Stasuik, N.S. Stevens, J. Toothill, G. Wadge and R. Watts (1998). Magma production and growth of the lava dome of the Soufrière Hills Volcano, Montserrat, West Indies: November 1995 to December 1997. Geophys. Res. Lett., Vol. 25, No. 18, pp. 3421-3424.

Sparks, R.S.J and S.R. Young (2000). Overview of the 1995-1999 eruption of the Soufriere Hills Volcano, Montserrat. This volume.

Tazieff, H. (1980). Letter to the Editor. J. Volc. Geotherm. Res., Vol. 8, pp.3-6. The Royal Society (1999). Science and Society: a response to the inquiry by the House of Lords Science and Technology

Select Committee. London: Statement 12/99, 4pp. Voight, B., R.P. Hoblitt, A.B. Clarke, A.B. Lockhart, A.D. Miller, L. Lynch and J. McMahon (1998). Remarkable cyclic ground

deformation monitored in real-time on Montserrat, and its use in eruption forecasting. Geophys. Res. Lett., Vol. 25, No. 18, pp. 3405-3408.

Voight, B., R.S.J. Sparks, A.D. Miller, R.C. Stewart, R.P. Hoblitt, A. Clarke, J. Ewart, W.P. Aspinall, B. Baptie, E.S. Calder, P. Cole, T.H. Druitt, C. Harford, R.A. Herd, P. Jackson, A-M. Lejeune, A.B. Lockhart, S.C. Loughlin, R. Luckett, L. Lynch, G.E. Norton, R. Robertson, I.M. Watson, R. Watts and S.R. Young (1999). Magma flow instability and cyclic activity at Soufrière Hills Volcano, Montserrat, B.W.I. Science, Vol. 283, p. 1138-1142.

Wadge, G. and M.C. Isaacs (1987). Volcanic Hazards from Soufriere Hills Volcano, Montserrat, West Indies. Report to the Government of Montserrat and the Pan Caribbean Disaster Preparedness and Prevention Project. [Reading: Dept. of Geography, Univ. of Reading] UK: 34pp.

Wadge, G. and M.C. Isaacs (1988). Mapping the volcanic hazards from the Soufrière Hills Volcano, Montserrat, West Indies using an image processor. J. Geol. Soc. London, Vol. 145, pp. 541-551.

Wadge, G., P. Jackson, S.M. Bower, A.W. Woods and E. Calder (1998). Computer simulations of pyroclastic flows from dome collapse. Geophys. Res. Lett., Vol. 25, No. 19, pp. 3677-3680.

Watts, R. B, R.S.J. Sparks and S. Young (2000). Growth patterns and emplacement of the andesite lava dome at the Soufrière Hills volcano, Montserrat. This volume.

Young, S.R., P.W. Francis, J. Barclay, T.J. Casadevall, C.A. Gardner, B. Darroux, M.A. Davies, P. Delmelle, G.E. Norton, A.J.H. Maciejewski, C.M.M. Oppenheimer, J. Stix and I.M. Watson (1998c). Monitoring SO2 emission at the Soufrière Hills volcano: Implications for changes in eruptive conditions. Geophys. Res. Lett., Vol. 25, No. 19, pp. 3681-3684.

Young, S.R., R.S. Sparks, W.P. Aspinall, L. Lynch, A.D. Miller, R.E.A. Robertson and J.B. Shepherd (1998a). Overview of the eruption of Soufrière Hills volcano, Montserrat, 18 July 1995 to December 1997. Geophys. Res. Letts., Vol. 25, No. 18, p. 3389-3392.

- 38 -

Young, S., S. Sparks, R. Robertson, L. Lynch and W. Aspinall (1997). Eruption of Soufrière Hills Volcano in Montserrat Continues. EOS, Transactions, American Geophysical Union, Vol. 78, No. 38, pp. 404, 408-409.

Young, S.R., B. Voight, K. Rowley, R.S.J. Sparks, R.E.A. Robertson, L.L. Lynch and W.P. Aspinall (1998b). The Soufrière Hills Eruption, Montserrat, British West Indies: Introduction to Special Section, Part 2. Geophys. Res. Lett., Vol. 25, No. 18, pp. 3387

- 39 -

Figure Captions

Figure 1: Map of Montserrat showing the location of the volcano, main place names used in the

text and the migration of the MVO to the north of the island during the period to 1995-1999.

The dashed lines are main roads around the island.

Figure 2: The original Soufrière Hills Volcano Observatory location in government buildings in

Plymouth. Dan Miller, John Power, Cynthia Gardner (all USGS) and Lloyd Lynch (SRU) watch

and wait, August 1995. (photo © W.P. Aspinall)

Figure 3: Temporary accommodation for the observatory team at the Vue Pointe Hotel,

September 1995. Rick Hoblitt (USGS), Lloyd Lynch and Richie Robertson (SRU), and Levar

Cabey (Montserrat Secondary School) work in luxury. (photo © W.P. Aspinall)

Figure 4: View of the eruption cloud on 25th June 1997, from the balcony of MVO(S) at Eiffel

House, Old Towne. (photo © W.P. Aspinall)

Figure 5: Pulling out of MVO(S), September 1997, due to the threat of fountain collapse

pyroclastic flows from vulcanian explosions. Simon Young and Billy Darroux, with Leroy Luke

driving. (photo S.C. Loughlin © BGS/NERC)

Figure 6: Moving into the Mongo Hill building that became MVO(N), September 1997. The

site is on the northern side of the Centre Hills, near St. John’s; view looking north. (photo S.C.

Loughlin © BGS/NERC)

Figure 7: MVO field team, Rick Herd and Tappy Syers, having uncovered the marker pin from

under new surge deposits, are setting up the Tar River GPS station for overnight measurement,

January 1999. The ruins of the Tar River estate House are in the background, and the view is

looking NE out to sea from the volcano. (photo © R.E.A. Robertson)

- 40 -

Figure 8: Map showing the original network positions of the five broadband and three single

component seismometers that comprised the ‘MVO Broadband Network’. Most of the stations

were located at the same sites as pre-existing stations of the short-period network. The combined

networks operated together from October 1996, with some subsequent modifications due to

damage from volcanic activity.

Figure 9: Network of stations used for precise levelling (dry tilt) monuments, electronic tiltmeter

installations, and Electronic Distance Meter (EDM) measurements at the SHV. Occupation of

dry tilt stations was discontinued in 1996, and eruption damage to electronic tiltmeters and EDM

targets effectively reduced these networks to a single EDM line by 1998.

Figure 10: Volcanic Risk Management Map from 6th June 1997, drafted by the MVO in

consultation with Montserrat government officials and disaster managers. This and the

following maps were based on hazard maps produced for the main types of volcanic events that

were expected from the SHV, and were used along with population distribution and

infrastructure maps to delineate risk boundaries.

Figure 11: Volcanic Risk Management Map for Montserrat as it stood in September 1997.

Microzonation was discontinued after 4th July 1997, when volcanic activity became more severe.

Figure 12: Volcanic Risk Management Map of Montserrat – revised in September 1998. The

boundaries of this map remained unchanged until April 1999 when some minor administrative

changes were implemented by the civil authorities in consultation with MVO.

Figure 13: Time-line chart of repeated appraisals of the volcanic Alert Level using expert

judgement elicitations in a weighted voting scheme, together with a record of the Alert Level in

force. This representation of scientific opinion, in which the current level setting is assessed

against thresholds to move up or down, provided decision-makers with both the central (optimal)

result of the poll and a measure of the spread of views, thus allowing them to select what level of

- 41 -

conservatism to accept. More often than not, the distributions of scientific opinion were skewed,

with longer tails to the more cautious end of the range.

Figure 14: A summary event probability tree for early stages of the Montserrat eruption,

showing elicited probability values attached to branches depicting different potential eruptive

scenarios. Such trees prompt the scientists to identify all possible scenarios and help

communicate their forecasts to public officials, summarizing on various time scales the hierarchy

of hazards against which mitigation measures can be planned. At MVO, the structured expert

elicitation methodology was used to update probabilities in a consistent manner, and associated

uncertainty distributions were, in turn, used as inputs to Monte Carlo risk simulations of

population impact (see Fig. 15).

Figure 15: Typical cumulative F-n risk curves from three risk assessments for Montserrat. The

curves show estimated exceedance probabilities for a given number of casualties occurring in the

next six months, based on Monte Carlo risk simulations combining identified volcanic hazard

scenarios with population numbers in given areas. Such estimates are useful to civil authorities

concerned with the tolerability of risk, and may be compared with societal risks from other

natural disasters, such as earthquake or hurricane. The examples shown here indicate how the

perceived risk levels dropped between April 1998 and February 1999 while the volcano was in

residual eruptive activity mode.

- 42 -

Table Captions

Table 1: Summary of the main representative areas of activity in which the MVO was involved

during the period October 1998 to April 2000. Specific operational tasks may change from time-

to-time, but it is anticipated that most of these MVO responsibilities will remain in being for the

foreseeable future.

Table 2: Summary of the main methods used by the MVO to disseminate information to the

public from August 1995 to April 2000. Most of these methods have become well established as

part of the MVO’s Outreach programme.

Table 3: The Montserrat Alert Scheme as it existed at the MVO in December 1995. This was

used as a guide when providing advice to the authorities about activity at the volcano. Although

no specific actions are listed here for the administrators, contingency plans drawn up by the

government’s Emergency Department were based on the identified alert levels.

Table 4: The Montserrat Alert System, March 1997. This was drafted with closer collaboration

between the disaster managers and the MVO and, as such, incorporated specific actions to be

taken by administrators and the general public. The Table had an accompanying risk map (e.g.

Figure 10) that illustrated the zones mentioned. These maps, along with a detailed listing of the

specific villages within each zone, were distributed to the public by the civil authorities.

- 43 -

Fig 1

- 44 -

Figs 2 & 3

- 45 -

Figs 4 & 5

- 46 -

Figs 6 & 7

- 47 -

Fig 8

- 48 -

Fig 9

NWhite's

LongGround

TarRiver

Castle Peak

Chance'sPeak

O'Garra's

Galway's

Windy Hill

Farrell'sRoad

Farrell's

St. George'sHill

Dagenham

LowerAmersham

GagesAmersham

Chance'sSteps

Eastern

Northern

Southern

Western

EDM Triangle

Active lava dome

Target & Instrument siteInstrument siteTarget site

2km

N

Castle Peak

O'Garra's

Windy Hill

Farrell's

St. George'sHill

AmershamBrodericks

2km

LowerWhites

UpperWhites

Upper

Single setup stationRadial line

Pre-1995 tilt meter site

Active lava dome

EDM networks

Dry & Electronictiltmeter networks

Electronic tiltmeters

GagesMtn

ChancesPeak

LongGround

Hermitage

- 49 -

Fig 10

Moving from Zones G to A represents an increasing risk,based on an evaluation of the volcanic hazard.

The status of each zone is dependent on the alert level.Potential hazards include pyroclastic flows, surges and ash fall.

Cork Hill

Salem

St Peter's

St John's

Harris Bethel

Plymouth

St Patrick's

Long Ground

Spanish Point

WH BrambleAirport

Lees

Kinsale

RichmondHill

A B

B

BC

D

D

E

F

G

C

C

AA

Little Bay

Fox's Bay

Soufriere HillsVolcano

- 50 -

Exclusion ZoneCentral Zone

Northern Zone

No admittance except for scientific monitoring and National Security Matters

Area with significantly lower risk, suitable for residential and commercialoccupation

Residential area only, all resident on heightened state of alert.All resident to have rapid means of exit 24 hours per dayHard hat area all residents to have hard hats and dust masks.

4 km

N

Long Ground

Spanish Point

Pelican Ghaut

Gages

St. Patrick's

Plymouth

Kinsale

Fox's BayRichmond

Hill

Cork Hill

Old Towne

Dyers

Molyneux

Lees

W.H. BrambleAirport

HarrisBethel

Trants

Streatham

St. John's

Salem

HillSt. George's

Soufriere HillsVolcano

Centre Hills

St. Peter's

Woodlands

Lawyers River

Nantes River North Olveston

Little Bay

Exclusion Zone

Central Zone

Northern Zone

Risk Map7 Sept 1997

Limit of high direct volcanic hazard

Fig 11

- 51 -

Fig 12

Risk MapSept 1998

Exclusion Zone: Bottom of Belham River valley and all areas south,

Limit of high direct volcanic hazard

4 km StPatrick's

Long Ground

SpanishPoint

Pelican Ghaut

Plymouth

Kinsale

Fox's Bay

RichmondHill

Cork Hill

Old Towne

Molyneux

W.H. BrambleAirport

Harris

St. John's

Salem

St. George'sHill

Soufriere HillsVolcano

Centre Hills

St. Peter's

Woodlands

Belham River

Little Bay Drummonds

Geralds

across Centre Hills to Pelican Ghaut

ExclusionZone

- 52 -

Fig 13

- 53 -

Montserrat volcano: risk assessment updates

0.0001

0.001

0.01

0.1

1.

1 10 100 1000

No. fatalities N

Prob

. of N

or m

ore

fata

litie

s in

6 m

onth

s

Feb 99 assmnt.Jul 98 assmnt.

Apr 98 assmnt.

Fig 14

- 54 -

Fig 15

- 55 -

Scientific Work Seismology • Processing of earthquakes & seismic signals on SP and BB networks • Analysis & archiving of seismic data • Manning of the Operations room

Ground Deformation • GPS measurements (once every 1-2 months) • EDM measurements (depending on visibility) • Electronic tilt (continuous - telemetered) • Theodolite measurements & photographs for dome volume calculations (depending on visibility) Geology • Deposits sampling (as activity permits) • Field mapping (as activity permits) General • Visual observations (daily) • Airfall ash & water sampling (as activity dictates) • Gas monitoring with SO2 tubes (every 2 weeks) and COSPEC (approx. 3 times per week) • Mudflow monitoring (as required) • Temperature measurements at pyroclastic flows (every two months) • Fumarole and hot springs sampling (2 –3 times/year, as conditions permit) • Video/photos of volcano (as conditions permit) Special Projects • Hazard mapping • Risk assessments and updates • Dedicated monitoring for authorised activities in Exclusion Zone • Preparation of public education/outreach material Electronics • Maintenance & repair of electronic equipment • Maintenance & operation of computer network • Video archive Outreach • Briefing of government officials (twice weekly & as activity dictates) • MVO Reports – weekly, (daily, monthly, special) • Press interviews • Weekly radio interviews • Public lectures; guided tours of MVO (as necessary; officially 1-3pm on weekdays) • Posters at MVO and elsewhere • Videos, and television & radio programs on the volcano & volcanism • Web pages • Collaboration with other scientists and groups

- 56 -

Management • Work programme & staffing etc. • Helicopter operations • Acquisition, repair and maintenance of monitoring equipment • Scientific and other professional visitors • Housing/accommodation • Advanced observations post • Customs • Collaboration with National Trust & Tourist Board on development of visitors centre

- 57 -

Method Start date End date MVO reports (daily – morning) August 1995 April 1998 MVO reports (daily – evening) August 1995 March 1999 MVO reports (weekly) April 1999 Continuing Scientific reports (weekly) November 1995 April 1998 Scientific reports (monthly) June 1998 April 1999 Scientific reports (quarterly) May 1999 Continuing Special volcanic activity reports September 1996 Continuing Lunch time updates February 1997 November 1997 Friday evening radio interview February 1997 Continuing Radio updates at times of high activity December 1996 Continuing “SeismiCity News” January 1997 March 1999 Government Information Service “Scientific Explanation” (weekly in the Montserrat Reporter)

1997 Continuing

Montserrat Reporter special articles 1996 Continuing Public posters in Salem and at airport December 1996 June 1997 Targeted meetings (e.g. to present risk map revisions) July 1995 Continuing Public lectures (occasional) August 1995 Continuing Visits to schools (occasional) November 1995 Continuing Visits by schools to MVO (occasional) November 1995 Continuing Visits to MVO by various sectors of the population (occasional) July 1995 Continuing Video nights (occasional) November 1995 Continuing Local Access TV (occasional) November 1995 Continuing Antigua Broadcasting Service broadcasts (occasional) November 1995 Continuing Off island visits/talks to Montserratian communities abroad November 1995 Continuing

- 58 -

Volcanic Activity Alert stage Actions by Administrators 1 Background seismicity with no new surface manifestation of volcanic activity

0 (WHITE)

Low-level local seismic activity, ground deformation and mild phreatic activity

1 (YELLOW)

Dome building in progress, periodic collapses generating rockfalls and occasional pyroclastic flows. Moderate level of seismic activity with no sudden changes.

2 (AMBER)

Change in style of activity anticipated within a few days. Pyroclastic flows common with associated light ash fall. High level of seismic activity.

3 (ORANGE)

Major dome collapse under way, with large pyroclastic flows and heavy ash fall. Explosive event possible if the activity continues.

4 (RED)

Ongoing large explosive eruption with heavy ash fall. 5 (PURPLE)

1 No specific actions for administrators were identified (c.f. Table 4).

- 59 -

Volcanic Activity Alert stage Actions by Administrators & General Public 1 Background seismicity with no new surface manifestation of volcanic activity

0 (WHITE)

All zones occupied. Review and update emergency plans on an ongoing basis

Low-level local seismic activity, ground deformation and mild phreatic activity

1 (YELLOW)

Maintain readiness of key personnel, systems and procedures. Keep stock of critical supplies Local evacuations may be necessary in Zone A. Zone B and C on standby

Dome building in progress, periodic collapses generating rockfalls and occasional pyroclastic flows. Moderate level of seismic activity with no sudden changes.

2 (AMBER)

Zone A - No access Zone B - Access limited to short visits by residents, officials and approved visitors with rapid means of exit Zone C - Daytime only visits for by residents, approved commercial activities and agriculture Zone D - Day and nighttime occupation by residents, high level of alert maintained Zone E, F - Full occupation of residents with national contingency plan for evacuation in readiness Zone G - Full occupation

Change in style of activity anticipated within a few days. Pyroclastic flows common with associated light ash fall. High level of seismic activity.

3 (ORANGE)

Zone A, B - No access Zone C - Access limited to short visits by residents and workers with means of rapid exit Zone D - Daytime occupation for essential services and agriculture, residents allowed access in daytime. Essential services operate with standby transport and evacuation plans in place Zone E - Full occupation with high level of alert maintained. Schools operate with standby transport Zone F - Full occupation by residents with contingency plan for evacuation. Warn of ashfalls in Zone E and F Zone G - Full occupation

Major dome collapse under way, with large pyroclastic flows and heavy ash fall. Explosive event possible if the activity continues.

4 (RED)

Montserrat Standing Operating Procedures for Red Alert in place All schools closed as required. People with special needs removed from Zone E & F Zone A, B, C, D - No access, rapid evacuation of all remaining persons Zone E - Rapid evacuation. Warn of potential for gravel, pumice and ash fall Zone F - Warn of potential for gravel, pumice and ash fall Zone G - Full occupation

Ongoing large explosive eruption with heavy ash falls.

5 (PURPLE)

Zone A, B, C, D, E - No access, rapid evacuation of all remaining persons Zone F - Initiate evacuation. Warn of potential for gravel, pumice and ash fall Zone G - Warn of potential for ash hazards

1 Recommended actions to minimise significant casualties. Zones as defined on Volcanic Risk Management Map.