SCIENTIFIC AND HAZARDS
ASSESSMENT OF THE SOUFRIERE HILLS VOLCANO MONTSERRAT
Addendum to the September 2002 Report
29 November 2002
MONTSERRAT
VOLCANO OBSERVATORY
Executive summary
i. The September 2002 Report of the Montserrat Volcano Risk Assessment Panel evaluated the status of the Soufrière Hills Volcano and its attendant hazards and risks at the time. It was submitted to the Government of Montserrat and to the UK Government by 4th October 2002. Shortly afterwards, a significant change in the locus and trend of growth of the dome was observed by the Montserrat Volcano Observatory team, with a switch in extrusion direction to the northwest.
ii. Given the continuing growth of a very large dome and that the direction of growth could give rise to collapses capable of generating pyroclastic flows which might reach the lower Belham Valley, and given that areas bordering the lower Belham Valley were still occupied, the Government of Montserrat Volcano Executive Group requested an urgent update of the September risk assessment. This Addendum records the outcome of an immediate re-appraisal of the hazard and risk estimates, focussing on the increased exposure of people living in and near the Belham Valley area. The update is based on further consideration of the detailed technical analyses of dome collapse scenarios and pyroclastic flow runouts that were undertaken for the March 2002 and September 2002 risk assessment meetings.
iii. On the basis of developments at the volcano and observations of activity and changes in the dome, the Risk Assessment Panel estimate the expected individual risk exposure for persons living full-time in Population Zone 4 should be revised upwards from 1 in 300 annualised probability of being killed (as assessed in early September) to about 1 in 100 - that is, the risk level is in the HIGH category on the CMO's Scale.
iv. In situations where the general public is exposed to a major hazard, the Precautionary Principle is often applied by policy makers, necessitating a quantitative statement of the influence of scientific uncertainty on any related risk assessment. In the present case, a check at the +1 standard deviation level in the spread of scientific opinion indicates the estimated individual risk for persons on the margins of the lower Belham Valley would equate to an annualised exposure of 1 in 44 on this basis, placing the level of risk well within the HIGH category.
v. As noted in earlier assessments, the risk is strongly dependent upon which areas are populated at the time of any volcanic activity. A reduction in societal exposure, and minimising of individual risk exposure locally, can be achieved by incremental evacuation of neighbourhoods in or bordering the area most threatened.
vi. The current situation at the volcano might prevail for months without a major collapse, in any direction. As long as the present mass of new dome material remains perched above the NW flanks of the volcano, the danger of a very substantial and energetic collapse entering the Belham Valley will persist.
vii. The RAP judges, therefore, that the available scientific evidence and its interpretation substantiate the case that areas in or close to the Belham Valley are exposed to a substantial volcanic threat, given the existing conditions at the volcano.
Contents
Introduction
1. One of the regular, six-monthly meetings of the Montserrat Volcano Risk Assessment Panel (RAP, comprising Montserrat Volcano Observatory scientific staff and external advisers), at which the status of the Soufrière Hills Volcano is evaluated, took place on Montserrat from 3rd – 4th September 2002. A press release, giving the preliminary findings of that RAP assessment, was released by the Government of Montserrat Volcano Executive Group (VEG) on 11th September 2002 (Appendix 1), and copies of the full report of the meeting were submitted to the Government of Montserrat (GoM) and to the UK Government (HMG) by 4th October 2002.
2. The report from the September Risk Assessment considered the risk if a switch of dome growth were to occur to the north. Much of this risk assessment meeting was devoted to assessing the likelihood of collapses of the dome under these circumstances, and on the size of collapse necessary for pyroclastic flows to reach the lower parts of the Belham Valley. The Report concluded that, if a switch were to occur, then the areas along and at the margins of the Belham Valley would become areas of high risk. In these circumstances, the occurrence of rockfalls or pyroclastic flows entering the upper reaches of ghauts feeding into the Belham watercourse would confirm potentially unstable conditions in this sector and be an indicator of a direct and active threat.
3. Shortly after the Report was finalised, a change in the way the dome was growing took place, and the extrusion of magma did indeed switch into a northwest direction. This change was observed and confirmed by the Montserrat Volcano Observatory (MVO) team on 26th September.
4. On 7th October, an extraordinary meeting of the VEG was held, and MVO were asked if the RAP could undertake a prompt assessment of the risks to the lower Belham Valley area, given the altered (but not unanticipated) circumstances. This appraisal was undertaken between 7th and 9th October 2002, utilising email and telephone conferencing, and was by a structured opinion elicitation by email. It involved mainly the external members of the Panel who had met a month earlier in Montserrat (listed in Appendix 2), working closely with the Director, MVO and his staff.
5. A brief statement on the changed situation and the general implications for risk levels in and around the lower Belham Valley was prepared by the external members of the RAP on 9th October 2002 (Appendix 3), and passed to the Government of Montserrat by MVO.
6. The present document summarises briefly the key observations made at the time by the MVO, and the results of the updating of the hazard and risk estimates undertaken by the external members of the RAP that underpinned the statement released on 9th October. It thus forms a prompt addendum to the main September 2002 Report.
7. For readers unfamiliar with basic terminology and concepts in volcanology, a glossary can be found in Appendix 4 of the September 2002 Report, and a more complete introduction to the major hazards can be found in the report “Assessments of the Status of the Soufrière Hills volcano, Montserrat and its hazards”, first issued on 18th December 1997. Further detailed descriptions of the various hazards scenarios considered both in the main report and in this addendum are given in Appendix 6 of the March 2002 report: “Scientific and hazards assessment of the Soufrière Hills Volcano, Montserrat”, issued by MVO. For completeness, however, the Chief Medical Officers Risk Scale is again reproduced here, in Appendix 4.
8. Whilst the Risk Assessment Panel has once more followed, as closely as possible, working procedures within the Code of Practice for Scientific Advisory Committees published by the Office of Science and Technology (December 2001), it should be noted that there are limitations[1] to these risk appraisals (see also Appendix 5 which must be read as part of the whole document).
Scientific observations in September and early October 2002
9. A chronological summary of events at the volcano in the period 19th September – 3rd October 2002 and a record of MVO’s observations in that period are provided in Appendix 6 (the information presented there has been extracted from an MVO internal document in draft form, prepared after a moderate size collapse of the dome had occurred on 29th September 2002).
10. On 26th September, a hybrid earthquake swarm occurred in the volcano. Visual sightings by MVO indicated that the ‘northern lobe’ (Fig. 1), which had developed in mid-August and then stagnated in early September, had become reactivated by a switch in direction of dome growth towards the northwest. This change produced semi-continuous pyroclastic flow activity, with small to moderate flows coming off the northern face of the lobe and funnelling into Tuitt’s and White’s Ghauts. Whilst these were not particularly large flows, and reached only a few kilometers down the ghauts, they were very persistent. At one or two stages in this period of elevated activity, the Windy Hill camera indicated small surges occurring near the uppermost part of Tyre’s Ghaut, but nothing was detected entering the main ghaut itself. Small surges and rockfalls also covered the apex of the fan above Farrell’s Plain (i.e. the headwater area where Mosquito and Tuitt's ghauts originally began before they were buried during the present eruption). There was also some pyroclastic flow activity down Gages Valley.
11. An observation flight on 28th September confirmed that the direction of growth had switched to the northwest, with a large whaleback-shaped extrusion lobe directed NW towards the ‘NW Buttress’. A big central spine, at the back of the lobe, could be seen also.
12. A moderate size collapse of the dome occurred on 29th September, generating flows down the northeast slopes of the volcano, which reached the sea near Spanish Pointe. This left a slippage scar in the talus slope above and around the ‘NW buttress’; however, that scar was quickly refilled over the next few days.
13. After the collapse, visibility was restricted for several days. On 6th October, the MVO team were again able to inspect the dome clearly. The western lobe, having now become well-established, was continuing to extrude outwards in a northwest direction, building a significant apron of talus (see Figure 1). By this time, the lobe was described by observers as ‘enormous’ and ‘steep sided’. In these circumstances, it was judged by MVO that the potential existed for a sudden collapse of a substantial volume of material, sufficient to generate a moderate to large pyroclastic flow down to the lower Belham. It was considered this could take place without there necessarily being a gradual build-up, in the form of smaller, preliminary flows, especially as the region was in the middle of its hurricane season: recent previous major collapses of the dome had occurred during periods of intense rainfall (see the September 2002 Report).
14.
These were the conditions in the first
week of October 2002 that gave rise to the VEG request for an immediate
re-appraisal of related hazard probabilities and risk levels in respect of the
Belham Valley area.
Volcanic hazards threatening the Belham Valley
15.
Since the last major dome collapse,
which occurred on 29th July 2001, more than 60 million cubic metres
of lava have been extruded on to the mountain.
By early September 2002, the total dome volume stood at 175 million
cubic metres - the largest it has been. The height
of the active part of the dome had reached an altitude of more than 1050 metres
above sea level and, above about 925 metres altitude, the active part of the
dome was no longer confined by pre-existing topography.
16. As at 6th September, the volume of new material in the most active part of the dome above the altitude of topographic barriers was estimated to be about 3 million cubic metres, and as much as 10 million cubic metres of material might be readily available for collapse if remnant material also became involved in a failure. While a collapse of this size would not be expected to channel all its material through the deep northwest gully into Tyre’s Ghaut, sufficient could go that way to generate a pyroclastic flow capable of affecting areas outside the limits of the Exclusion Zone (then in force).
17. However, the sector for a dangerous directed collapse of the dome of this kind is usually quite restricted: the combined probability of (a) having a large collapse in that direction, and (b) this generating a sufficiently large pyroclastic flow, had been estimated by the RAP in September at only 4% in the following six months, given the circumstances that existed at the time.
18. But, as noted in the September 2002 Report, during the eruption there have been significant changes in the topography of the northern flanks of the volcano: these changes are important because they make it more likely than before that pyroclastic flows can move into the Belham Valley, and thus threaten areas which were at the margins and just outside the Exclusion Zone (as it was then). The topographic changes, and the pyroclastic flow models that were used in the September 2002 assessment, were described in detail in Appendix 6 of that report.
19. The Panel had, in that assessment, considered the distance that pyroclastic flows would travel if they moved down the northern flanks and into the Belham Valley. Observations and theory show that the distance a pyroclastic flow travels down a volcano is related to the volume of the dome collapse, to the elevation of its source and to the topography on the upper flanks of the volcano. These insights are used as the basis for modelling flow run-outs, and could be calibrated for the Soufrière Hills volcano by data from previous pyroclastic flows on Montserrat. The group concluded that a collapse of 3 million cubic metres or greater, taking place to the northwest, was likely to reach the golf course in the lower Belham Valley, and that larger collapses were likely to reach the sea. Taking all the factors together, a hazard line had been defined by the RAP for the northern side of the Belham Valley, which would separate areas likely to be affected by pyroclastic flows and surges from areas at lower risk (see Fig. 3 of the September 2002 Report).
20. By early October 2002, when even more material had accumulated in the western lobe (see Fig. 1) and small avalanches were entering the top of Tyre’s Ghaut, there was increased concern for the direct threat of such pyroclastic flows and their associated surge clouds in the Belham Valley. Thus, the situation required that the September 2002 risk assessment be brought up-to-date for the Belham Valley flow scenario in particular.
Re-appraisal of the volcanic hazards and risks to the lower Belham Valley
21. As noted above, circumstances on the volcano changed significantly in late September 2002. This addendum report revises the earlier calculations of exposure risk for identified threats to people in and near the Belham Valley area by making adjustments to the analysis based on a re-appraisal of conditions and changes on the dome. The approach and methodology is again that described in the December 1997 Report, validated by the UK Government’s Chief Scientific Adviser’s consultative group.
22. For consistency, the geographic population zones considered in this risk re-assessment are the same as those shown on Fig. 1 of the September 2002 Report, and population numbers are assumed unchanged from that assessment.
23. In order to be compatible with what had been done before and, in particular, to test the impact of any changes of perceived probability on the risk levels determined in their September 2002 Report, the RAP has re-assessed the relevant probabilities of occurrence of the ‘initiating events’ which might give rise to pyroclastic flows and surges in the Belham Valley (the full framework of scenarios used in the risk modelling for Montserrat is described in Appendix 6 of the MVO March 2002 Scientific and Hazards Assessment Report). This entailed a re-elicitation of key event probabilities and related conditional probabilities, on the same terms as before (i.e. for a six-month outlook).
24. Three pairs of scenario probabilities are crucial in this context, representing three classes of collapse scale:
i. The probability of 0.3x ref collapse (3 x 106 cu m, in any direction in next 6 months);
ii. The conditional probability of flows/surges to NW (Belham), given a 0.3x ref collapse occurs;
iii. The probability of 1x ref collapse (10 x 106 cu m, in any direction in next 6 months);
iv. The conditional probability of flows/surges to NW (Belham), given a 1x ref collapse occurs;
v. The probability of 3x ref collapse (30 x 106 cu m, in any direction in next 6 months);
vi. The conditional probability of flows/surges to NW (Belham), given a 3x ref collapse occurs.
25. Five sets of updated probabilities were received from members of the RAP for these six questions, which were circulated by email. The opinions are pooled, as before, using the EXCALIBR structured elicitation procedure, and the resulting values entered into the Monte Carlo simulation model to determine the annualised individual risk exposure (IRPA) for persons in the Population Zone 4 (that is, Lower Friths-Happy Hill-Old Towne South, see Fig. 1 of the September 2002 Report).
26. With these revised probabilities, the revised IRPA for Population Zone 4 goes up, from an assessed level in September 2002 of 1 in 300 annualised probability of being killed, to about 1 in 100 - in other words just into the HIGH category on the CMO's Scale. Thus, working on the same basis and with the same assumptions as those used in the assessment of early September, there is judged to be a three-fold increase in the estimated risk exposure along the margins of the lower Belham Valley.
27. In deriving this risk estimate by simulation techniques, there is uncertainty in the result arising from imprecisions in the modelling procedure itself. Multiple repeat runs of the simulation program, using a different random number generator seed for each trial, confirm the expected IRPA for Population Zone 4 (i.e. 1 in 100), with a standard deviation of ±4 on this central value. This equivalent to a ±3 standard deviation range of from 1 in 88 to 1 in 112 - in other words, there is very little chance that the result falls outside this range, because of numerical modelling effects.
28. Variations in the outcome of a risk simulation because of limitations to scientific understanding are a different matter, however. In many situations involving the estimation of public risk exposure the Precautionary Principle is applied, often necessitating quantitative scrutiny of the influence on a risk assessment of scientific uncertainty. In the present case, the latter issue can be engaged by examining the outcome of adopting the +1 standard deviation points in the spreads of the synthesised probabilities derived from the opinions of the RAP group, rather than the central values; the higher probabilities are then used as single inputs to the simulation model to provide a measure of the extent to which variation in the elicited opinions can influence results. With this constraint in place, the IRPA for the Zone 4 Population area computes to an exposure of 1 in 44, i.e. unequivocally in the HIGH category of the CMO’s scale. (The numerical simulation modelling uncertainties in this case provide a ±3 standard deviation range of exposure odds from 1 in 40 to 1 in 48.)
29. Where significant scientific uncertainty exists, some proponents argue that the +1 standard deviation level is the measure of risk to rely upon (or an even higher level) in ‘risk-informed’ decision-making, rather than the expected value, as given in para 25. Whichever yardstick is adopted, in the present case the current assessment is that the volcanic risk to individuals in the areas of concern must be regarded as HIGH.
30. As noted in earlier assessment reports, the overall societal risk from volcanic activity is strongly dependent upon which areas are populated at the time of any activity. Moving north from the Belham Valley, there is a steep decline in risk to residents. If it is required to reduce total overall societal exposure, or minimise local individual risk exposure, this can be achieved by incremental evacuation of areas bordering the Exclusion Zone. Early warning and other mitigation steps could also reduce exposure.
31. For any strong volcanic activity, such as that being considered here, there is also a concomitant threat of people being injured (as opposed to killed outright). In respect of injuries, the related risk probabilities are probably significantly greater than those given above for fatalities, but this issue can be judged more reliably by medical and emergency specialists.
32. The area along the northern side of the Belham River that is assessed as being at high hazard from pyroclastic flows and surge clouds in the present circumstances is shown on the map in Fig. 3 of the September 2002 Report: the edge of this area is a line extending from the mouth of the ghaut at Lime Kiln Bay southeastwards along the ghaut to the roundabout in Old Towne; and from there eastwards along Logwood Drive to Olveston House; then continuing eastwards to the Happy Hill/Friths main road junction; and from there south-eastwards to the Waterworks Estate).
33. The present situation at the volcano, with a large mass of dome material perched above the northwest flanks, might prevail for months without a major collapse occurring. The topography of the dome can change again, and switches in its direction of growth may take place. Nevertheless, until a major change takes place (e.g. a very large collapse down Tar River Valley), or growth ceases and the dome stagnates for a very long period of time (it will need many years to stabilise), the present configuration will remain ripe for a very substantial and energetic collapse, which could be triggered by an internal magma extrusion pulse or heavy rainfall. Such a collapse could involve dangerous flows and surges in the Belham Valley.
34.
The RAP judges, therefore, that the
available scientific evidence and its interpretation substantiate the case that
areas in or close to the Belham Valley are exposed to a substantial volcanic
threat, given the existing conditions at the volcano.
35.
As
noted previously (see RAP Statement made on 9 October 2002 – Appendix 3), the
RAP emphasises that this assessment is not predicting that a major collapse is
necessarily going to occur in the near future nor, if one does occur, that the
bulk of the material involved will inevitably go down the Belham Valley. Indeed the more likely outcome is that the
present threat, with the accompanying HIGH risk level in the Belham
Valley, will remain until conditions are changed by a
future event. A large collapse,
if it occurs, might be directed down the Tuitt's or Gages sectors, with only
marginal effects in the Belham Valley.
It is also possible that a rainfall-triggered collapse which was large
enough to remove the growing part of the dome in some other direction, for
instance down the Tar River Valley, could alleviate substantially the threat to
the Belham Valley.
Appendix 1 Volcano Executive Group press release - 11th September 2002
VOLCANO EXECUTIVE GROUP PRESS RELEASE
The Soufriere Hills Risk Assessment Panel met on 3 and 4 September. A Summary of their conclusions has already been issued to the Press.
Members of the Volcano Executive Group (VEG) have discussed these conclusions carefully with representatives of the Risk Assessment Panel. They noted the Panel’s estimate that the dome was now as large as it has ever been, and that while it was growing in a north easterly direction, as it was at present, any collapse would most likely occur in the Tar River area, as it did last year, or down Tuitt’s and White’s Ghauts towards the old airport.
They also noted the Panel’s conclusion, however, that if growth should switch back to the north or north west, as it did briefly a few weeks ago, areas along the margins of the Belham Valley could become at high risk within a very short time. The Scientists explained that this could lead to pyroclastic flows, possibly even reaching as far as the sea and accompanied by surge clouds which could affect areas along the margins of the Belham Valley.
The VEG concluded that if there were to be such a switch of growth to the north or north west and material started to spill into the upper reaches of the Belham Valley, measures to evacuate the areas around the Valley would need to be put in place at short notice.
In
consultation with the Risk Assessment Panel,
the VEG identified the areas that could be affected in this eventuality
as those being south of a line extending from
the mouth of the ghaut at Lime Kiln Bay southeastwards along the ghaut to the
roundabout in Old Towne; and from there eastwards along Logwood Drive to
Olveston House; then continuing eastwards to the Happy Hill/Friths main road
junction; and from there south-eastwards to the Waterworks Estate. The
Emergency Department are able to provide more advice to those wishing know
whether they may be affected.
It should be emphasised that this is not a warning that evacuation will be required. It is merely a precautionary measure to keep the population informed of risks which might occur if dome growth switches back to the north or north west. It should also be noted that the levels of risk associated with such a switch only apply while the dome is currently so large.
The VEG is satisfied that there is sufficient monitoring capability in place to ensure proper assessment of the dome during periods of cloudy weather when the summit region is not visible. The DTEZ also remains open for the time being.
Any enquiries about evacuation planning should be directed either to the Police or the Government Emergency Department.
(signed) (signed)
HE The Governor Hon Chief Minister
Volcano Executive Group
Montserrat
11 September 2002
Appendix 2 List of participants
Scientific and hazard
assessment - Belham Valley risk update,
8th October 2002
External
Scientists
Prof. R.S.J. Sparks FRS Dept. of Earth Sciences, Bristol University, UK (chairman)
Dr. W.P. Aspinall Aspinall and Associates, UK
Dr. L.Neuberg Leeds University, UK
Prof. B. Voight Penn State University, USA (invited to provide information but not a formal contractor to RAP)
MVO Scientists
Dr. P.N. Dunkley Director
Dr. G. Thompson
Dr. R. Watts
Appendix 3 Statement from External Members of the
Montserrat Volcano Risk Assessment Panel
Hazard and Risk assessment (9th October 2002)
The members of the Risk Assessment Panel have been following developments on Montserrat over the last two or three weeks which have culminated in the critical situation that has now led to evacuation along the margins of the Belham Valley. The members of the RAP are fully in agreement with the scientific advice that has been provided by MVO over this period. The Director of the MVO requested on 8th October that members of the RAP evaluate the current situation on the volcano on the basis that the potential hazard from the volcano had reached the level at which the risk to the population close to the Exclusion Zone could have become HIGH on the Risk Scale of the Chief Medical Officer of the UK.
The current situation of a large and growing volume of lava above
the northern flanks of the volcano with a growth direction to the northwest
presents a definite and serious threat. This is the situation, which the last
three RAP reports identified as of greatest concern, and unfortunately has now
come about. Since the risk assessment 3 weeks ago the changes at the volcano
and elevated levels of activity have increased the chances that a major
collapse of the dome could take place with pyroclastic flows and surges moving
down the Belham valley, affecting populated areas
along the fringes of the valley.
During the last Risk assessment a number of risk model runs were done using elevated probabilities for directed collapse to the NW. These sensitivity models pushed the Individual Risk Exposure at the margins of the Belham Valley from MODERATE to HIGH. While the RAP has not yet had time to carry out a full analysis of the current hazards and risks there is agreement that the probabilities of such a collapse are now greater than the assumptions made in these sensitivity tests. Therefore it is certain that the level of risk is now HIGH.
The RAP agrees that the potential exists for a collapse large enough to affect the lower Belham valley, which may persist over the coming months. The RAP emphasises that this assessment is not predicting that a major collapse is necessarily going to occur in the near future nor, if one does occur, that the bulk will inevitably go down the Belham valley. Indeed the more likely outcome is that the threat with the accompanying HIGH risk level in the Belham valley will remain until alleviated by a future event. A large collapse, if it occurs, might be directed down Tuitt's, or Gages sectors with only marginal effects in the Belham valley. It is also possible that rainfall collapse down the Tar River Valley, which was large enough to remove the growing part of the dome, could lower risk dramatically.
W.P. Aspinall
J. Neuberg
R.S.J. Sparks
B.Voight
The information provided in this assessment is advisory and provides indications of the degree of risk that can be expected in the near future from the volcano in Montserrat. While Panel members believe that they have acted honestly and in good faith, they accept no responsibility or liability, jointly or severally, for any decisions or actions taken by HMG, GoM, or others, directly or indirectly resulting from, arising out of, or influenced by the information provided in this report.
Appendix 4 Chief Medical Officer’s Risk Scale
Negligible: an adverse event occurring at a frequency below one per million. This would be of little concern for ordinary living if the issue was an environmental one, or the consequence of a health care intervention. It should be noted, however, that this does not mean that the event is not important – it almost certainly will be to the individual – nor that it is not possible to reduce the risk even further. Other words which can be used in this context are ‘remote’ or ‘insignificant’. If the word ‘safe’ is to be used it must be seen to mean negligible, but should not import no, or zero, risk.
Minimal: a risk of an adverse event occurring in the range of between one in a million and one in 100,000, and that the conduct of normal life is not generally affected as long as reasonable precautions are taken. The possibility of a risk is thus clearly noted and could be described as ‘acceptable’ or ‘very small’. But what is acceptable to one individual may not be to another.
Very low: a risk of between one in 100,000 and one in 10,000, and thus begins to describe an event, or a consequence of a health care procedure, occurring more frequently.
Low: a risk of between one in 10,000 and one in 1,000. Once again this would fit into many clinical procedures and environmental hazards. Other words which might be used include ‘reasonable’, ‘tolerable’ and ‘small’. Many risks fall into this very broad category.