During this reporting period the volcano was extremely active with major pyroclastic flows to the north and west devastating the villages of Harris, Streathams, Windy Hill, Riley's Yard, Bethel, Bramble and Farms. The flows reached to within 50 m of the airport on the east. Tragically 10 people are now confirmed dead and 9 people are missing presumed dead as a result of the activity on 25 June.
The volcano's most destructive act to date struck during this reporting period. On 25 June a substantial portion of the massive lava dome nested in the crater of Soufriere Hills collapsed, generating a series of pyroclastic flows that swept across the northern flanks of the volcano to the sea. The flows and associated surge deposits destroyed 150 homes, claimed 10 lives and left 9 people unaccounted for.
The days leading up to the event of 25 June were relatively quiet. Activity between 21 and 23 June was limited to a few rockfalls and minor pyroclastic flows from points high up on the east and north-east face of the dome. The flows generally travelled no further than the base of the dome in the Tar River Valley in the east and came to rest in the upper reaches of Tuitts Ghaut in the north-east. The largest flow ran down the Tar River Valley to within 100 m of the delta (~2.0 km long). In the evenings, glow and incandescent rock falls were observed on the north-east face of the dome. A brief daylight glimpse of the dome on 23 June revealed a blocky summit profile punctuated by three prominent spines, and extensive ash and steam emissions from the summit, particularly in the area above Mosquito Ghaut.
On 24 June the activity shifted northward slightly and a series of minor rockfalls and pyroclastic flows (1 km in length travelled down Mosquito Ghaut. Rockfalls continued on the east face of the dome throughout the day.
Between 06:00 and 08:00 on 25 June there were a near-continuous series of rockfalls and small pyroclastic flows that fed into Mosquito Ghaut. Sporadic rockfalls continued to occur on the east face of the dome. The activity subsided as the morning progressed. Later in the day, at ~12:45, a dilute steam and ash cloud billowed upwards from the summit of the dome. At ~13:00 a dense, dark ash column shot up from the north flank of the dome above Mosquito, reaching 45,000 feet in a matter of minutes. In rapid succession three major pyroclastic flows pulsed down Mosquito Ghaut. Pyroclastic surge clouds spread laterally over the ridges near the head of Mosquito Ghaut as the flows shot past MVO reference peaks "B" and "C". The flows fed into the Paradise River and spread laterally through Trants Village and Spanish Point as they entered the lowlands below the town of Bramble. A distinct lobe of pyroclastic flow material spilled out of the ghaut and travelled into Bethel. This lobe deposited large blocks over the area around Bethel. The upslope surges fanned westward to the head of Dyers Valley, northward to Streatham, and eastward to Paradise Estate. Near the town of Gun Hill, the surge was funnelled into the Dyers Valley and travelled west along the Belham River as far as Cork Hill.
The pyroclastic flows and surges of 25 June were the largest and most destructive ever produced during this eruption. The intensity of the event exceeded that of the 17 September 1996 explosion. An estimated 4 to 5 million cubic metres of lava dome were unloaded during the event creating a spoon-shaped scar with a steep back wall on the lower flanks of the dome above Mosquito Ghaut. The resulting deposits covered ~ 4 square kilometres to depths as great as 30 m in places. A detailed account of the 25 June 1997 event can be found in MVO Special Report 3.
The days immediately following the event were relatively quiet. The most notable events were two small steam explosions on the morning of 27 June, which showered 0.5 cm-sized rock fragments over the western flank of the volcano, in the areas of Lovers Lane, Dagenham, Richmond Hill, and Foxes Bay.
Pyroclastic flow activity picked up again on the 28th. Between the hours of 10:24 and 11:42, moderate collapses on both the north and north-west flanks of the dome sent 5 pyroclastic flows down Mosquito Ghaut ((1 km long) and 2 flows down Gages Valley/Fort Ghaut into central Plymouth (( 3.5 km long). Significant pyroclastic flow activity continued through 5 July in Mosquito Ghaut, Gages Valley/Fort Ghaut, and to a lesser extent, in Tuitts Ghaut and over Galways Wall. A listing of the flows observed (there may have been others) over this very active period is given below:
Many of the flows between 28 June and 5 July started with a resounding "BOOM" and a vertical ash column that shot upwards of 35,000 feet. On occasion, the event ended with large ash emissions and was not followed up by pyroclastic flow activity. Ash column rise rates of 9-17 m/s were measured (unfortunately, there are no measurements on the most robust of the columns occurring over this period).
A rare viewing window on the morning of the 29 June showed a large collapse scar visible on the northern flanks above Mosquito Ghaut. This scar extended back as far as the central point of the dome with a classic spoon-shape and steep back-wall. The uppermost ca. 100 m of the pre-June 25 dome had been removed and the eastern and north-eastern flanks of the dome appeared remarkably unscathed with all the pyroclastic flows being funnelled down Mosquito Ghaut from the collapse area. A relatively small active growth area was visible in the central/rear floor of the scar and was generating rockfalls that overspilled into Mosquito Ghaut. An observation flight on the 1 July showed that the growth area within the scar had increased considerably with ca. 65 % of the void being filled particularly in the central and rear portions of the scar. The central part of the growth area had several small spines on the top and large rockfalls generated from the base of this area were commonly spilling down the chute that had developed at the head of Mosquito Ghaut. Both the visual observations of the dome and frequent pyroclastic flow activity seemed to confirm the relatively high extrusion rate that is believed to have predominated during this period.
The 22 of June showed a marked change in volcanic seismicity. The previous 20 hour cycles of rockfall activity were replaced by 8 hour cycles of hybrid swarms. This change in activity coincided with a small swarm of volcano-tectonic events. Hybrid swarms with large numbers of events were recorded until the large pyroclastic flow of 25 June. After 25 June the 8 hour cycles continued, but at the end of the reporting period most cycles corresponded to a hybrid swarm of relatively few events changing to tremor before a pyroclastic flow. Throughout the two weeks rockfall activity continued, but not with any particular periodicity.
Table 1: Earthquake types
These earthquake counts are of events that triggered the broadband network's event recording system between 0000 and 2400 each day (local time).
Date VT Hybrid LP Dome RF LPRF* HYRF* 22 June 97 9 96 18 180 12 7 23 June 97 3 452 5 84 2 2 24 June 97 0 421 1 60 0 1 25 June 97 0 130 0 48 0 1 26 June 97 0 55 3 38 1 1 27 June 97 0 106 0 40 0 6 28 June 97 0 152 0 72 0 4 29 June 97 0 117 0 60 0 2 30 June 97 0 82 0 54 0 3 01 July 97 0 43 0 29 0 2 02 July 97 0 73 2 77 1 0 03 July 97 0 4 7 63 3 0 04 July 97 0 14 1 26 1 0 05 July 97 0 86 1 65 0 0
*LPRF: LP earthquake followed by rockfall signal. HYRF: Hybrid earthquake followed by rockfall signal. The LPs, hybrids and rockfalls in these signals are also counted in their respective columns.
GPS occupations of the volcano's BIGNET (Harris, Whites, Long Ground, Dagenham and Brodericks) and EASTNET (Harris, Whites, Long Ground, Windy Hill) were completed on 22 and 24/25 June respectively. Both networks were cut short by one line due to high seismic activity. 1 cm shortening on the line Harris to Windy Hill was noted. The movement of direction of the Long Ground station appears to have changed abruptly. Through mid 1996 to end April 1997 the station moved 3 cm to the east. Five measurements to the station since April indicate an abrupt movement of 3.4 cm to the north-east which has produced a 2.5 cm shortening on the line to Harris. Future measurements of these network will be severely hampered due to problems of accessibility and the necessity for full helicopter support during each occupation.
A new GPS network has been set up covering the north-western flank of the volcano that is still accessible by road. The 4 sites on this network (LEESNET) are Lees Yard (on the northern flanks of Gages Mountain), M40 EDM (on the Waterworks Estate), St. George's (the south-west flanks of St. George's Hill) and M27 (in Old Towne). The sites at Garibaldi and a pin on the MVO pool deck will also be used for theodolite measurements during clear periods using the pole on the tip of St. George's Hill as the point of reference.
A permanent reflector has been installed near the GPS pin at Lees Yard and measurements to this reflector will be shot with the Total Station from the M40 EDM site on the Waterworks Estate. Measurements to this reflector will also be attempted using the Total Station from the Garibaldi site. The fixed reflector at Farrells was probably destroyed during the 25 June activity.
Prior to 22 June the Chances Peak tiltmeter showed a flattened pattern. At 05:30 on 22 June, a sharp increase in the rate of inflation occurred. Subsequent, sharp deflation at 06:30 was coincident with sustained pyroclastic flows which travelled approximately 1 km down the Tar River Valley. This event marked the beginning of a new pattern in the inflation-deflation cycles. The periodicity of the cycles shortened to 8 hours. and the amplitude increased to approximately 40 microradians. The change was accompanied by a short volcano-tectonic earthquake swarm which preceded the resumption of hybrid earthquake activity (see above). The number of hybrid earthquakes varied nearly exactly in phase with the inflation-deflation cycle, with the maximum number of hybrids occurring at the peak inflation.
Following the 25 June pyroclastic flow activity, the inflation-deflation cycle continued with the same period and amplitude which began 22 June until 5 July. The overall trend of the tiltmeter, prior to 25 June, showed a general inflation to the north or deflation to the south. Since 25 June, the trend shows a general deflation toward the centre of the dome.
No dome volume measurements were possible during this period due to poor visibility. The volume of deposits of the 25 June flows was estimated at 4.61 million cubic metres, with the total volume of deposits in the area affected estimated at 5.42 million cubic metres as of 25 June. This volume was obtained using data from a kinematic GPS survey of the area on 5 July for the more distal areas of the region unaffected by flows in the intervening period, along with rougher estimates of deposit thicknesses in the higher regions of the ghaut immediately following the 25 June event. Representative deposit thicknesses after the event were 5 m in Mosquito Ghaut, 14 m in Pea Ghaut and 6 m in the Farm's area. Estimates of the volume of the scar left by the 25 June eruption were difficult due to poor visibility, but a brief glimpse allowed a poorly constrained estimate of about 6.5 million cubic metres. There is a discrepancy between this volume and the pyroclastic flow volume, and the better constrained flow volume will be used in future calculations. The volume of the anomalously mobile flow down the Belham valley on 25 June is estimated to be around 90,000 cubic metres.
The total volume of deposits on the 5 July in the Mosquito, Paradise, Farms and Farrell's area from the kinematic GPS survey is estimated to be 9.24 million cubic metres, indicating accumulation of 3.82 million cubic metres since 25 June. The Gages valley was surveyed using the kinematic GPS technique on 7 July and a volume of 3 million cubic metres estimated.
Kinematic GPS surveys of the northern ghauts prior to the reporting period gave a volume of 0.42 million cubic metres for Tuitt's Ghaut on 9 June, and 0.81 million cubic metres for Mosquito ghaut on 20 June. These represent the first significant flows down each of these ghauts on 5 June and 17 June respectively.
Rain and trough water was collected from several sites around the volcano on 22 June and the results are shown in Table 2. The samples from Upper and Lower Amersham are extremely acidic and have high concentrations of total dissolved solids (TDS).
Table 2 Rainwater geochemistry, 22 June 1997
Location pH Cond. TDS Flrd Chlrd Slpht mS/cm g/l mg/l mg/l mg/l Weekes 6.34 0.090 0.044 0.30 14 nd Upper Amersham 3.05 2.080 1.030 1.10 450 >200 Upper Amersham 3.05 2.080 1.030 1.10 450 >200 Lower Amersham 3.22 1.060 0.530 1.20 218 79 Amersham cattle trough 6.53 0.487 0.243 1.15 72 33 Trials reservoir 7.51 0.916 0.459 0.35 106 33 Cattle trough, SS 7.76 0.285 0.143 0.10 41 nd MVO 6.50 0.093 0.046 0.25 16.6 nd Molyneaux 6.61 0.190 0.095 0.25 25.0 nd
nd: not detected
The miniCOSPEC was used to measure the sulphur dioxide flux from the volcano. Results are shown in Table 3. These values show a marked increase in flux before the 25 June. Since 25 June, no measurements have been possible along the roads along the central corridor or through Plymouth because of the extreme risk in these areas. However, experiments have been attempted by using the helicopter (29 June) and from static scanning methods at the observatory (5 July). Results from these experiments will be released at a later date.
Table 3 Daily average SO2 flux using miniCOSPEC (tonnes/day)
Date tonnes/day 22 June 438 23 June 1157 24 June 1933
Ash was collected at a few points around the volcano after the 25 June eruption. However, many of the sites were destroyed or rendered inaccessible by the activity and so full coverage was not possible. After the small explosive event on 27 June, coarse lapilli (up to 10 mm) were collected from Dagenham and Richmond Hill.
The Risk Map
During June 1997, 4 successive risk maps were published in response to the increase in activity over the northern and western flanks of the volcano. With the advent of each map, the A-B zone (with no access) gradually increased in size to cover most of the south of the island. The risk assessment was clearly heading towards the situation where consolidation of risk zones and alert levels into a new risk map was required.
A new risk map was published on 4 Jul 1997 and was designed to simplify the previous system where there were 7 zones on the risk map and 6 alert levels. This meant that there were 42 different options for action depending on location and the state of the volcano. Now there are only 3 zones: the northern, central and exclusion zones, and only one alert level: "volcanic alert".
To decide where the boundaries between risk zones should lie, the extents of the pyroclastic flows and surges so far were marked on a map of Montserrat. These are shown on the risk map as black arrows. The maximum distance that pyroclastic flows and surges were expected to reach, given the current state of the volcano and with a similar type of activity, was estimated. There is now the potential for flows to reach much of the south of Montserrat, and so it was decided that an exclusion zone should be demarcated in these areas. This involved drawing a line across the centre of the island from Old Road Bay to north of the airport. This line is controlled primarily by the topography of the land and distance from the volcano.
North of the exclusion zone it is considered that the risk of pyroclastic flows and surges is low enough to allow people to live and work as normal. However, in the case of an increase in activity it is thought that people who live and work in the area directly north of the exclusion should be ready to move at short notice if necessary. Therefore a central zone has been designated in which people should be on increased alert. The further that you move away from the exclusion zone, the safer you get. Thus the northern boundary of the central zone is marked as a dotted line. In the event of an increase in alert level, citizens should move uphill and away from the Belham River Valley. If it is considered that an evacuation of the central zone is necessary, the wailing sirens will be sounded and maroons (explosive fireworks) will be fired.
So in summary, the new risk map is a simplification of the old risk and alert system, and it is hoped that it will be the current map for at least one month and possibly longer.
MVO staff changes
Dr. Keith Rowley (independent, British Geological Survey)
Dr. Maggie Mangan (Hawaii Volcano Observatory, US Geological Survey)
Dr. Gill Norton (British Geological Survey)
Mr. Richie Robertson (Seismic Research Unit, University of the West Indies, Trinidad)
Dr. Brian Baptie (British Geological Survey)
Ms. Chloe Harford (Bristol University)
Mr. Billy Darroux (MVO)
Dr. Rick Hoblitt (Cascades Volcano Observatory, US Geological Survey)
Dr. Ricky Herd (British Geological Survey)
Dr. Sue Loughlin (British Geological Survey)
Dr. Angus Miller (independent, British Geological Survey)
Dr. Paul Cole (Luton University)
Professor Steve Sparks (Bristol University)
Ms. Lizette Rodriguez (University of Puerto Rico)
Dr. Alan Smith (University of Puerto Rico)
Dr. Glen Mattioli (University of Puerto Rico)
Dr. John Shepherd (Lancaster University)
The current chief scientist is Richie Robertson (Seismic Research Unit, University of the West Indies, Trinidad).
6 Montserratian staff (Team Seismic: Pops Morris, David Silcott, Venus Bass, Thomas Christopher, Levar Cabey and Grace Lewis) went to the Seismic Research Unit of the University of the West Indies in Trinidad for a week to take part in a training course in seismology and ground deformation methods led by Dr. John Shepherd (Lancaster University). This was followed by a training week at MVO practising out their new skills under Dr. Shepherd's guidance.