Montserrat Volcano Observatory, Montserrat, West Indies

Scientific Report 71
16 August 1997

This report covers the 14 day period from 00:00 on 2 August to 24:00 (local time) on 16 August.


The activity has been characterised by tilt cycles, hybrid earthquake swarms, pyroclastic flows and vulcanian explosions. This has been one of the most intense periods of activity at the Soufriere Hills Volcano during the current eruption which began on July 18, 1995.

Visual Observations

Activity at the beginning of August was characterised by abundant rockfalls. Eruptions became increasingly explosive and the block and ash flows more voluminous. Two tilt cycles and associated hybrid swarms on 2 August each culminated in block and ash flows down Gages. The same occurred the following day, but the flows were larger. The dome was observed at 14.00 on the 3rd, when an active face of large blocks and spines was present high above Gages wall. From 18.00 to 20.30 at the peak of the af ternoon tilt cycle on the 3rd, a succession of pyroclastic flows travelled down Gages. Then at 18.10 a single large flow descended the length of Fort Ghaut to the sea, causing extensive damage in Plymouth, where fires burned late into the night along the north and south banks of Fort Ghaut.

The first explosive activity occurred on 4 August. Eruption of a pyroclastic flow at 06.30 was accompanied by a loud rumbling, followed by fallout of lithic and crystal lapilli up to 5mm diameter at MVO. A second explosion at 16.43 sent a dark grey jet of ash inclined at about 60o high above the dome. Moments later, pyroclastic flows swept 3.5km down Tuitts Ghaut, 3.5km down the Tar River Valley to the sea, 4km down Fort Ghaut to the sea, and an unknown distance down the White River. The resulting plume rose to 5km and fragments of dome rock and dense pumice as large as 1.5cm diameter fell at MVO. The flow that entered Plymouth started new fires.

A second explosion took place in the early morning on 5 August near the end of a 2.5-hour hybrid earthquake swarm. The plume from this eruption rose to 6-10km and drifted to the west. The lapilli fallout that began at MVO five minutes later consisted of fragments of dome rock and dense pumice.

Between the afternoon of 5 August and the morning of 12 August , there were ten Vulcanian explosions. The first eight occurred regularly every 10 to 12 hours during or immediately after hybrid swarms and generated pumice flows by column collapse. The last two occurred on the 11 and 12 August. These appeared slightly weaker than the others and were separated by hybrid swarms without explosive activity.

Each explosion began with the rapid rise of one or more dark vertical plumes. A boom was heard, followed by an intermittent roar which lasted up to several minutes. Jets of dark material with large ballistic bombs were thrown symmetrically outwards is high, curving arcs. Shortly after, pyroclastic flows formed and travelled between 3 to 5km down all the major ghauts around the dome (Gages, White River, Tuitt's, Mosquito, Tar River and Tyre's). Those to the west and east reached the sea down the Fort Ghaut and the Tar River valley respectively. The flows were formed by the fallback of material onto the slopes surrounding the dome. The flows travelled very fast initially, but decelerated beyond about 2km from the dome and were observed several minutes after the eruption onset moving very slowly as thin, creeping carpets near their distal ends.

Each eruption column rose rapidly to a height of between 9 and 12km, where it spread out to form an umbrella cloud. Pumice clasts as large as 2cm fell at St. Johns. Ash clouds rose swiftly from the pumice flows and merged with the main plume. After several minutes, the eruption settled into a phase of pulsing, less intense discharge, generating a low, bent-over plume of ash that was distributed to the west and WNW by low-level winds. The plume slowly decreased in height over about an hour until the eruption ceased.

Ballistic impact craters from the explosions extend out to 1.6km from the centre of the dome. Blocks as large as 1.2m of fresh dome rock occur on the Farrells plain, at Hermitage, and on the NW slopes of the South Soufriere Hills. Pumice flows from the explosions travelled down six drainage systems around the dome. A striking feature is the similarity of runout distances in different directions, showing that column collapse was essentially symmetrical.

The explosions occurred from a large circular crater excavated in the summit of the dome. The crater persisted through the last major explosion of the morning of 8 August, and was seen from MVO at 17.00 on the 9th. When seen again at midday on the 10th a small new dome was visible above the western crater lip. This apparently survived the final explosion (12 August) and over the next three days grew into a large spine 950m high at 10.00 on 13 August. By 14th August the crater was nearly filled by new lava.


The whole of this period was characterised by intense hybrid earthquake swarms occurring at regular intervals. For the first 10 days the interval between swarms was between 10 and 12 hours. On August 2nd and 3rd the swarms culminated in continuous tremor and pyroclastic flows. The explosive activity, beginning on August 4th, was preceded in almost all cases by hybrid earthquake swarms of varying intensity. The intensity and duration of swarm generally correlated with the intensity of explosion. Swarms typically started 1-4 hours before the explosions. As the swarm developed the frequency (number of events per unit time) increased. The individual events themselves had highly repeatable spectra. Explosions and pyroclastic flows continued to follow the hybrid swarms for the next 4 days.

The explosion signal consisted of three discrete phases (1) long period precursor (2) high amplitude pyroclastic flow signal (3) lower amplitude harmonic tremor. The precursor phase was a very monochromatic 1 Hz signal that started 10-20 seconds before the signal caused by the pyroclastic flows. When the pyroclastic flow signal died down, approximately 6 minutes later, the 1 Hz signal could be seen to be continuing as low amplitude harmonic tremor. This harmonic tremor lasted up to 80 minutes after the explosion and was probably linked to "puffs" of ash venting observed at the dome.

From the 9th onwards the swarms of hybrids continued but often ended without any external activity - although small pyroclastic flows did occur after some swarms. The swarms culminated in intense episodes of high amplitude tremor. The spectral characteristics of the tremor episodes were almost identical to the spectra of individual hybrid events. On the 13th the cycle changed slightly with swarms and tremor occurring every 8 hours. The peak level of tremor increased throughout this period.

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*
 2 Aug 97	0	155	10	45	4	3
 3 Aug 97	0	122	16	89	2	0
 4 Aug 97	0	136	3	50	1	0
 5 Aug 97	2	78	0	10	0	0
 6 Aug 97	0	99	7	20	1	0
 7 Aug 97	0	132	15	36	2	0
 8 Aug 97	0	137	9	9	0	0
 9 Aug 97	0	158	14	4	0	0
10 Aug 97	0	244	8	8	0	0
11 Aug 97	0	287	0	21	0	0
12 Aug 97	0	285	2	26	0	1
13 Aug 97	0	818	0	4	0	0
14 Aug 97	0	417	1	26	0	0
15 Aug 97	0	218	0	26	0	0
16 Aug 97	0	405	0	17	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.

Ground Deformation

An occupation of the GPS network EASTNET (Harris, Whites, Long Ground and Windy Hill) was made on 12 August. The occupation confirmed the shortening observed on the lines Harris-Whites and Harris-Long Ground. Total shortening on the two lines is around 3cm. The positions of the sites indicate a northwards movement of Whites and Long Ground.

Further measurements have been made to the recently established target site at Lees Yard from instrument sites at the Water Works Estate, MVO and Garibaldi Hill. Fluctuations in the slant distances between Garibaldi Hill and the Water Works estate and the target site at Lee's Yard appear to have disappeared from the recent measurements. The source of these fluctuations is still unclear and attempts are being made to find the reasons for such changes. The recent measurements indicate that there is a very slow shortening trend on the slant distances to the Lee's Yard target. It is hoped to increase the frequency of measurements to expand the database for these lines.

Dome Volume Measurements

A series of GPS-laser binocular surveys were carried out by helicopter on August 13, 14 and 15. The objective was to obtain a shap-shot of the deposit volumes in the ghauts and a dome volume. In this period surveys of the Tar River Valley, the Tar River Fan, Mosquito ghaut and Tuitts ghaut were completed. Volumes of deposits for all the ghauts and the dome are required before an extrusion rate can be given, but current deposit volumes (not DRE) for the four measured areas are in million m3:

Location		million cubic meters
Tar River Valley	13.55
Tar River Fan		15.72
Mosquito Ghaut		8.88
Tuitts Ghaut		1.68

(The end of Mosquito and Tuitts ghaut are taken at the point where they converge SE of Harris. The eastern limit of the Tar River Valley is taken as the original zero topographic contour).

Two main areas of deposition are present in the Tar River Valley. A lower ponded area below the second break in slope contains 4.16 million m3 of block and ash flow deposits. The area stretches 700m to the east to the fan and is around 200m wide. Maximum deposit thickness is around 45m in the original channel of Hot River. The bulk of the Tar River deposits are above the main break in slope immediately east of the dome and comprise rockfall deposits and shorter run-out pyroclastic flows.

The Tar River Fan volume was last measured on 7 February. Since then there has been additional volume contributed from flows in mid-February and throughout most of May. Although many flows made it onto the fan in this time, few reached the sea and the coastline trace was essentially unchanged. The explosive events of early August covered around 80% of the fan area with a veneer of thin, lobate pumice flows. The pumice flows are typically 0.5-1m in thickness.

The upper reaches of Mosquito ghaut filled in rapidly after the 25 June pyroclastic flow activity, due to the accumulation of deposits from rock falls and small pyroclastic flows. The western wall of the ghaut is less than 5m high in two stretches along the Farrell's fields and will be filled completely by the next large flows down the ghaut. This will allow material to spill out of the ghaut at a high level and spread across the Farrell's plain towards the upper reaches of the Belham valley. Deposits in the upper reaches of the ghaut are over 40m in thickness and thin to around 20m in the area of Paradise turning. Immediately south of Harris the deposits are 10-15m thick and reach 25m at the confluence with Tuitts ghaut.

The pumice flow deposits of the August column collapse events were too thin to be individually surveyed using the kinematic GPS technique, although there volume was incorporated into the mid-August volume snapshot as already described. The volume of pumice flow deposits, estimated by mapping their areal extent and approximate thickness (>1-2 metres), is of the order of 1-2 million cubic meters. The estimated total volume of material carried in the Vulcanian eruption columns is also 1-2 million cubic meters, estimated from mass flux derived from column height and duration of the explosion signal on the broadband seismic network.

Environmental Monitoring

No rainwater, ash samples or gas samples were collected during this period. Environmental work has focused attention to the ambient concentrations of airborne particles to which the population living and working near the volcano on Montserrat is exposed. We have been using the Dust Trak aerosol monitor to measure PM10 values (Particulate Matter less than 10 microns). Dust Trak sampling started on the 10th August. Most parts of the non-evacuated areas were studied, with particular attention paid to busy places where people spend most of their time. Occupational and environmental measurements, both personal and static, have been carried out (the table shows only static measurements) showing that dusty work can increase values up to 7 times the background value. Values are also strongly linked to weather conditions (wind direction, rain, etc.) and to the volcanic activity (ash venting, visible fallout, etc.). Dust levels in Salem are generally above the Air Quality Standard (AQS: 0.05 mg/m3) accepted in the UK and calculations show that a person living and working in Salem can be exposed to values between 0.1 and 0.2 mg/m3 a day. On the other hand, measurements carried in areas north of Salem show values generally below the AQS, even in busy places.

PM10 (concentration in mg/m3

Location	10.Aug.97	11.Aug.97	12.Aug.97	13.Aug.97	14.Aug.97	15.Aug.97	16.Aug.97
BC		-		-		-		-		-		-		0.05
Bso		-		0.036		-		0.023		-		-		-
CB		-		-		-		-		-		-		0.037
CCSi		0.092		-		-		-		-		-		-
CCSo		0.131		-		-		-		-		-		-
CH		-		-		-		-		-		-		0.044
Friths		-		-		-		-		-		-		0.473
Fogarty		-		-		-		-		0.01		-		-
Gpi		0.044		-		-		-		-		-		-
Gpo		0.017		-		-		-		0.009		-		-
Hso		-		-		0.049		-		-		-		-
Lawyers		-		-		-		-		0.024		-		-
MSSi		0.11		-		-		-		0.063		-		-
MSSo		0.108		-		-		-		0.032		0.07		0.06
MVOi		a 0.193		-		-		-		-		bAV -		0.074
MVOo		-		a 0.065		c 0.065		c 0.054		dn 0.03 	bAV e 0.037	0.05
PCS		-		-		0.155		-		-		-		-
PS		-		0.734		-		-		-		0.138		-
RC		-		-		-		-		-		-		0.043
St. Peters	-		-		-		-		-		-		0.049
SCP		-		-		0.162		-		0.286		-		0.352
SSi		-		0.332		-		-		-		-		-
SSo		-		0.368		-		0.223		-		-		0.135
SSSo		-		-		0.068		-		-		-		-
SVo		-		-		a 0.051		0.051		-		-		-
WB		-		-		-		-		-		-		0.036

These values represent the average reading over the sample duration. The mean duration of sampling is usually between 15 to 30 minutes. Samples taken over longer duration are labelled as follows:

  1. 30 minutes to 1 hour
  2. 1-6 hours
  3. 6-12 hours
  4. 12-24 hours
  5. 24-36 hours
  6. 36-48 hours

- 	No samples taken
AV	Average between different samples taken on same day
n	sample taken at night
i	sample taken inside
o	sample taken outside

Key to locations:
BC - Brades Church (North)
BS - Brades Shelter (North)
CB - Carr's Bay (North)
CCS - Catholic Church Shelter (Salem)
CH - Cudjoe Head (North) 
GP - Gerald's Park Shelter (North)
HS - Hope Shelter (Salem)
MSS - Methodist School Shelter (Salem)
MVO - Montserrat Volcano Observatory (Old Towne)
PCS - Pentecostal Church Shelter (Salem)
PS - Police Station (Salem)
RC - Red Cross (Woodlands)
SCP - Salem Car Park
SS - Supermarket in Salem (Ram's Emdee)
SSS - Secondary School Shelter (Salem)
SV - Seascape Villa (Old Towne)
WB - Woodlands Beach. 

Staff Changes
Dr. Paul Jackson (Seismic Research Unit, UWI, Trinidad)

Dr Gill Norton (BGS)

Montserrat Volcano Observatory