Montserrat Volcano Observatory, Montserrat, West Indies

Scientific Report 27
05 June 1996


Activity at the Soufriere Hills Volcano from 30 May until 5 June showed continued dome growth and associated rockfalls from the flanks of the dome. Two of the rockfalls on 31 May were large enough to generate moderate-sized pyroclastic flows in the Tar River Valley, the first almost reaching the sea. There was a swarm of the small repetitive hybrid earthquakes which lasted from 30 May until 3 June. This was probably associated with a change in the active areas of the dome at about the same time; rockfall activity declined on the east, northeast and north flanks and increased on the south and west flanks of the dome.

Visual observations

Viewing conditions were excellent at the start of the week. There was vigorous steaming from many parts of the dome on 30 May and rockfall activity was concentrated on the north and northeastern flanks. Rockfalls occurred throughout the morning of 31 May, mainly from the north and east flanks of the dome. Large volumes of unstable hot material were noted high on the dome, especially on the eastern flank. There was also some new growth and steaming from the southeastern flank and some new material was seen avalanching down into the southern moat.

There were two pyroclastic flows on 31 May at 14:04 and 15:35. Both of these originated from the eastern flank of the dome and followed a well-developed channel into the Upper Tar River Valley. The first, larger, flow travelled to within 300 m of the sea, whilst the second reached well past the Tar River Soufriere. Both flows generated ash clouds which reached an altitude of 2 to 3 km and were blown to the northwest, depositing ash in the Cork Hill and Richmond Hill areas. Later that afternoon, the first large rockfall from the southern moat was witnessed from the air, with material moving less than a hundred metres out of the crater in the valley to the south of Castle Peak.

Viewing conditions were generally poor during the rest of the week, although there were partial views of the dome on most days. There was a small pyroclastic flow at 06:00 on 1 June into the upper Tar River Valley which produced an ash column which reached an altitude of 2 km. Rockfall activity on 1 June was concentrated on the northeastern, eastern and southern flanks of the dome. On 2 June, a small "whaleback"-type feature was observed on the eastern flank of the dome, just to the north of Castle Peak.

On 4 June, it was clear that activity had decreased on the eastern side of the dome with few rockfalls being observed. The main areas of dome growth appeared to be on the south and western parts of the dome; both rockfalls and gas emissions were seen in these areas. No prominent spines were seen on the top of the dome, which was flatter and rounder than it had been in the past. One small, but prominent spine had appeared on the northeastern flank of the dome.


Seismic activity at the Soufriere Hills Volcano during the week was dominated by variable amounts of small- to moderate-sized rockfall events and a swarm of the small repetitive hybrid events associated with dome growth. The table below lists the daily counts of the different types of events recorded - volcano-tectonic, long-period, hybrid, and rockfalls. It also includes an estimate of the amount of the broadband seismic tremor seen on stations close to the crater.

Earthquake types 30 May - 5 June 1996

Date          VT        LP        Hybrid    RF        Tremor Amount

30/05         1         5         0         17        Low to intermediate
31/05         1         14        96        97        Intermediate to high
01/06         1         0         307       116       Intermediate to high
02/06         0         0         132       83        Intermediate to high
03/06         0         1         19        32        Intermediate to high
04/06         0         5         18        51        Low to intermediate
05/06         0         17        8         57        Low to intermediate
Seismic events associated with rockfalls on the flanks of the dome occurred throughout the week. There was a noticeable increase in rockfall activity associated with the pyroclastic flows on 31 May. This heightened level of activity continued until 2 June.

A swarm of the small repetitive hybrid seismic events started at about 17:15 on 30 May with the events occuring at a rate of about 1 per minute. This gradually increased to about 2 events per minute by about 13:30 on 31 May. The rate then gradually decreased on 1 June to one event every one or two minutes. The hybrid events continued to occur at this rate until the early morning of 3 June, when the activity stopped. Throughout the period of activity, the hybrid events were of variable amplitude. In the past, swarms of hybrid events have been associated with changes in the dome growth rate or the position of the actively growing areas. This swarm may well have been associated with the observed change in activity from the north and east to the southern and western flanks of the dome.

The broadband tremor recorded on stations close to the crater was again present, in varying amounts, throughout the week. There was slightly more tremor between 31 May and 2 June, which may have been associated with the increase in rockfall activity. The tremor was mainly intermittent, but there were distinct near-continuous episodes from 07:50 to 11:05 on 30 May, from 02:42 to 04:49 on 3 June and from 23:10 on 3 June to 04:00 on 4 June. There was some sustained tremor-like signal on the Gages seismic station between 00:05 and 00:39 on 5 June, with very high amplitude from 00:15 to 00:23. Since it was recorded on only one station, it must have been generated by something local to that station, most likely a small mudflow in upper Fort Ghaut. The nature of the seismic signal, building slowly to a peak and then slowly decreasing again, is similar to previous signals from mudflows. There was muddy water running in Fort Ghaut in Plymouth at the time of the signal, which is consistent with a small mudflow higher up.

Some long-period earthquakes were recorded during the week, but the numbers were not significant. Three VT earthquakes were recorded. Those on 30 and 31 May were located at .75 and 1.5 km beneath English's Crater. That on 1 June was located at a depth of 2 km to the southeast of the crater.

Ground Deformation

The poor viewing conditions during the week made it difficult to take EDM measurements. However, lines in the eastern triangle were occupied on 30 and 31 May and on 2, 3 and 4 June. The lines to Castle Peak show movements consistent with the recent increase in the rate of shortening. Analysis of recent data indicates that this change occurred in the second half of April. The lines from both Long Ground and Whites show a consistent shortening of about 1 mm per day between the beginning of December and the end of April. From 1 May until 4 June, they show an average daily shortening of 2.4 mm from Long Ground and 2.1 mm from Whites.

There was one apparent departure from this rate. The Long Ground to Castle Peak line shortened by 12 mm between 30 and 31 May. The Whites to Castle Peak line shortened by 6 mm at the same time and by 9 mm over the next two days. When both lines were remeasured on 4 June, the rates were back to normal.

The western triangle of the EDM network was measured on 4 June. The results continue to show no consistent changes in line lengths.

Dry-tilt levelling lines at Brodericks were measured on 04 June for the first time since 5 January. Analysis of the data showed there may have been some changes. However the errors on the new measurements were quite large and may have accounted for the observed changes.

No GPS measurments were made because of continuing problems with the equipment.

Gas Measurements

SO2 flux measurements were made during the week using a COSPEC correlation spectrometer which was mounted in a car and driven between Cork Hill and St. Patricks under the gas plume. The results of the measurements for the week are given in the table below, as the mean and standard deviation of all the runs made each day. The results show some variability but are consistently low, with a mean value of 169 tonnes per day.

COSPEC measurements of SO2 flux

Date     Number    Mean SO2 flux       Standard
         of runs   (tonnes / day)      deviation
30/05         2         224            58
31/05         2         88             72
01/06         3         52             13
02/06         6         193            37
03/06         3         192            13
04/06         4         240            65
05/06         7         194            55

Gas concentration from FTIR measurements (BDL = below detectable level)

Date Location Path length (m) HCl (ppb) SO2 (ppb) SO2:HCL ratio 15/05 Kinsale shore road 150 - 500 34 BDL 17/05 Upper Amersham 200 546 <480 <0.88 21/05 Parsons Estate 200 51 14 0.27 22/05 Amersham-St Georges 1800 54 11 0.20 24/05 Plymouth docks 275 35 80 2.28 24/05 Foxes Bay cliff rd 380 BDL BDL 24/05 Foxes Bay main rd 400 BDL BDL 28/05 Amersham 325 40 BDL 30/05 Amersham 350 52 27 0.52 02/06 Amersham 350 34 18 0.53

Some gas measurements using Fourier Transform Infra-Red Spectroscopy (FTIR) were conducted during the week. It has not proved possible to measure the gas concentrations in the plume using either the active or passive techniques. The active technique -using an artificially generated infrared light source- has been used to measure gas concentrations at ground level. All the measurements made to date are given in the table above. The errors in these measurements are typically 50% or more. The errors for the measurements made in Plymouth Docks on 24 May are larger, about 100%.

The analyses show that on the lower slopes of the volcano, excluding the results from upper Amersham, the ambient concentrations of both HCl and SO2 are well below 100 ppb. These levels pose no significant short-term health hazards. The long-term hazards do not apply, since noone should be living in these areas, but the hazard is probably no different from living in a moderate-sized city.

The SO2:HCl ratio is generally well below 1.0. The only exception to this is on 24 May, when the measurement errors were large. If we assume that the gas ratios measured at ground level are the same as those in the plume, then the low ratios would indicate a moderately degassed magma chamber.

Additional measurements

Gas measurements were made at the soufrieres around the volcano.

Staff changes
Lutchman Pollard, Seismic Research Unit, Trinidad

Desmond Supersad, Seismic Research Unit, Trinidad
Wilkie Balgobin, Seismic Research Unit, Trinidad

Jean Battaglia, Guadaloupe Volcano Observatory

Montserrat Volcano Observatory