Montserrat Volcano Observatory, Montserrat, West Indies

Scientific Report 74
27 September 1997

This report covers the 14 day period from 0:00 on 14 September to 00:00 (local time) on 28 September, 1997.


Activity on the volcano has been extremely high during the reporting period. During the first week, rockfall and pyroclastic flow activity was concentrated in Tuitt's Ghaut. The flows typically had runout distances of around 3km and filled in the upper reaches of the ghaut. Seismic activity was dominated by pyroclastic flow and rockfall signals, although a hybrid swarm preceded the 14 September flow, and there were a large number of long-period events. A major collapse of the dome occurred during the morning of 21 September generating pyroclastic flows and surges over the north-eastern flank of the volcano. Most of the area between Trant's Yard Estate and White's Yard has been impacted and the flows reached the coast in many places. The volcano has been in a phase of explosive activity since 22 September with vulcanian explosions occurring every several hours. The explosions produced vertical eruption columns and collapsing fountains of material which fed pyroclastic flows. The flows were pumice-rich and were shed into all the ghauts around the volcano. Most of and northern and western Montserrat has received pumice and ash fall.

Visual Observations

Activity was very high during the first seven days of the reporting period with moderate-sized pyroclastic flows occurring down Tuitt's Ghaut and over Farrell's plain. Rapid accumulation of material occurred in these two areas and the upper reaches of Mosquito Ghaut were so full that material spilled across into Tuitt's Ghaut building a small debris fan. The largest individual pyroclastic flow in Tuitt's Ghaut occurred on the morning of 17 September, depositing material 200m beyond the Tuitt's / Mosquito Ghaut confluence. Continued degradation of the southeast side of the dome has occurred generating rockfall debris and small pyroclastic flows in Tar River.

At 03:54 on 21 September a large collapse took place from the NE flank of the dome. Pyroclastic flow activity occurred down Tuitt's Ghaut devastating the area from Trant's Yard to White's Yard. The airport terminal building has been destroyed as have most properties in Spanish Point, Bramble and Tuitt's. Bethel village has been buried. Small fans were built into the sea at the mouths of Farm River and White's Ghaut.

Most deposition from the flows occurred around Trant's and Farm Estate Yard greatly extending the 25 June fan. A lobe detached from the main Tuitt's Ghaut flow, 500m south of the confluence with Mosquito Ghaut and spread ENE across Bramble village. The deposit was similar to that of the Bethel lobe emplaced on 25 June, consisting of a thin ashy, unit with abundant, large (metre to several-meter sized) blocks. Material also spilled from Tuitt's Ghaut into White's Ghaut at a low point in the valley wall. This generated pyroclastic flows in White's Ghaut which reached the sea, and extensive surges which travelled over Tuitt's Estate, Bethel Estate and White's Yard. Volumetrically the collapse is estimated to have generated around 8 million m3 of deposits - considerably larger than that of 25 June.

Following the collapse, a series of vulcanian explosions began at 00:55 on 22 September. There have been 15 explosions to the end of the reporting period occurring at intervals of between 7h:14min to 17h:11min (average 9h:44min). The explosions have been similar in many ways to those of early August beginning with the abrupt generation of a dark-grey explosion cloud comprised of radiating spears on ballistic trajectories. This rises quickly to 600-1000m above the dome. This then partly develops into a convecting column and partly forms a collapsing fountain around the plume. During this time a rapidly-building roar can be heard and ballistics could be seen falling to the ground up to 1.5km from the dome. Pyroclastic flows were generated during most explosions and travelled down all the six major ghauts on the volcano (Tar River, Tuitt's, Mosquito, Tyre's, Gages and White River). The explosion columns rose to heights of up to 7600m. Vigorous pulses of ash venting were observed for up to 1hour after the explosions and low rumbling sounds were heard for up to 30minutes after each explosion. At night time the explosions appeared as a huge fire-ball above the volcano which would expand laterally showering incandescent ballistics over Farrell's plain, Gages Mountain and Chances Peak. Incandescent surges lasting a few seconds were seen on several occasions travelling over Farrell's plain and down Tuitt's Ghaut. Observers at JackBoy Hill saw extreme venting of gas from the fumaroles on the eastern flank of the dome at the onset of one explosion.

Thin, pumiceous pyroclastic flow deposits were generated by the explosions. Although not observed as clearly as some of the August events, they were clearly generated by fountain collapse as a veneer of pumice was left over the area between Gages and Peak 'C', over Chances Peak, over the Galways area and over much of the dome's surface. The time of their generation varied however from just a few seconds after the onset of the explosion to more than 30 seconds. They had maximum runout distances of around 4.5km in Tuitt's Ghaut, reached the sea in Tar River, 2.25km in Tyre's Ghaut, 4km in White River, 2.6km down Gages Valley and 2.5km down Farrell's plain. In unconfined areas such as the Tar River Fan, the Farrell's Plain and the Trant's/Farm Fan the flows left long, thin tongue-like deposits and more rarely spread out into thin sheets. Flows confined to the narrow ghauts appear thicker, narrower and have steeper termini.

Fallout from the explosions has been varied due to the variable wind directions. Most of northern and western Montserrat has suffered moderate ash fall and fallout of pumice fragments. 2cm pumice fragments have been recorded at the Observatory on Mango Hill and at Little Bay, 3cm at Cudjoehead and Olveston and 5cm fragments on Davy Hill. A 6cm pumice was reported in Woodlands.

The scar developed above Tuitt's Ghaut extends 300m back into the dome, where it leads into an explosion crater of about 300m diameter. There is a low wall of tephra separating the crater and the scar, and the depth of the crater below the level of the pumice wall is estimated at around 150m. Much of the dome's surface and the upper area of Chances Peak are blanketed by fresh pumice deposits. Two of the three peaks on the northern crater wall (Peaks 'B' and 'C') have been severely eroded and just poke out of the dome talus and pumice deposits.


From the 14th to the 20th of September the seismicity was dominated by rockfall and pyroclastic flow signals. There were few earthquakes in this period, although the pyroclastic flow on the 14th September was preceded by a hybrid swarm (table 2) which included some large events. These hybrids had amplitudes on the broadband Windy Hill station as great as any recorded since the installation of the broadband network in October 1996. There were also an unusually high number of long period earthquakes throughout this period - which tended to occur before, rather than after, pyroclastic flows. On the 16th of September a bang was heard that correlated with a long period earthquake.

The night of 20th / 21st of September showed a marked change in seismicity. There was a hybrid swarm before the big pyroclastic flow at 3:54 am but the hybrids did not decrease very much after the flow and a high level of hybrid activity continued for the rest of the day. Then at 00:55 on the 22nd September came the first in a sequence of explosive events. These explosions occurred at intervals of 6 to 10 hours for the remainder of the reporting period (table 3). As in August each explosion was recorded on the seismometers as a pyroclastic flow signal preceded by an approximately 1 Hz signal of varying amplitude relative to the pyroclastic flow amplitude. This long period energy continued throughout the pyroclastic flow signal and continued afterwards at low amplitude as tremor. A few of these explosions were preceded by hybrid swarms (see table 2) but others had very little seismicity leading up to them. When a swarm did occur it tended to continue for a short time after the explosion. All of the explosions were followed by a period of tremor lasting from 20 minutes to 3 hours. This tremor was less harmonic than in August but instead had two or three well defined spectral peaks. As in August there was good visual confirmation that the tremor was correlated with ash venting from the crater.

Volcano-tectonic earthquakes from throughout this period were located at between 2 and 4 km below the top of the dome.

Table 1: Earthquake types

These earthquake counts are of events that triggered the broadband network's event recording system between 16:00 and 16:00 each day (local time).

Date		VT	Hybrid	LP	Dome RF	LPRF*	HYRF*
14 Sep 97	4	2	5	65	0	0
15 Sep 97	10	31	12	95	0	0
16 Sep 97	13	3	20	47	0	0
17 Sep 97	12	0	9	56	0	0
18 Sep 97	0	14	11	63	0	0
19 Sep 97	0	0	36	32	6	0
20 Sep 97	9	10	14	22	0	0
21 Sep 97	0	109	6	50	1	0
22 Sep 97	5	82	10	24	1	0
23 Sep 97	9	33	0	8	0	0
24 Sep 97	0	88	5	11	0	0
25 Sep 97	5	47	17	20	0	0
26 Sep 97	1	50	0	17	0	0
27 Sep 97	0	30	3	28	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.

Table 2: Swarms

Start			Stop		number of vts	number of hybrids
20:51 14/9/97		23:23 14/9/97		0		25
19:30 17/9/97		21:00 17/9/97		0		12
05:21 20/9/97		08:07 20/9/97		9		10
20:49 20/9/97		23:39 20/9/97		0		43
06:12 21/9/97		10:27 21/9/97		0		42
08:05 22/9/97		13:00 22/9/97		0		45
16:15 23/9/97		00:28 24/9/97		0		78
21:30 24/9/97		03:39 25/9/97		3		37
17:11 25/9/97		20:04 25/9/97		0		23

Table 3: Explosions

Explosion of	Duration of LP signal	Amplitude of 	Duration of PF	Amplitude of 
		before PF signal 	LP signal	signal (min)	PF signal	
		starts (sec)

00:55 22/9/97		12		16900		555		64400	
10:46 22/9/97		11		22000		540		83200	
20:42 22/9/97		10		12500		400		56800	
07:23 23/9/97		8 	  	8500		340		22800	
00:34 24/9/97		10		14600		370		21200	
10:54 24/9/97		10		12300		560		33200	
17:16 24/9/97		-				
03:55 25/9/97		15		28000		480		52200
11:09 25/9/97		9		9300		400		23000	
20:05 25/9/97		14		13900		350		34900
04:28 26/9/97		14		7400		300		20200
14:56 26/9/97		17		34600		330		25600
00:01 27/9/97		17		20300		290		31500
09:46 27/9/97		17		11100		310		23700
17:15 27/9/97		-				
All values are measured on the Windy Hill broadband station. Amplitudes are peak-to-peak in arbitrary units.

Ground Deformation

A GPS occupation of EASTNET was made on 20 September including sites at Harris, White's, Long Ground and Hermitage. The occupation confirmed the shortening that has been reported in the most recent scientific reports on the Harris-White's and Harris-Long Ground lines. The shortening is now around 3.5cm on these lines due to northward movement of White's and Long Ground. The movement does not at the moment appear to be accelerating. The Hermitage site had not been occupied since 21 May due to the activity on the NE side of the volcano. The line to Harris showed a further 1cm shortening (movement of Hermitage to the NE) consistent with its movement in May. The White's site was impacted the following morning by pyroclastic surges. This activity also damaged the UPR permanent GPS site at White's.

No EDM measurements have been made during this reporting period. Ash from pyroclastic flows and explosions has obscured the target at Lee's Yard and it is currently unsafe to enter this area to clean the reflector. An intensive EDM program had been planned to look at the details of deformation in the NE sector of the volcano as a follow-up to the reported movements picked up by GPS. The plans are being modified following the pyroclastic flow activity in this area on 21 September.

Volume Measurements

No dome or deposit volume measurements were made during this interval. Static photographs were obtained from the ground at Whites and from a hover position close to Windy Hill.

A series of accuracy tests were carried out with the GPS-laser binoculars from the helicopter. This was to assess their suitability for longer-distance work for dome mapping. Shots were taken with the binoculars to known points from a series of distances, azimuths and dips, to see how closely the coordinates of the fixed point could be approached. To date the working range has often been less than 200m, which is now considered too dangerous. Although the binoculars have a quoted precision of 1m in distance and 1? in dip and azimuth there is too much shake in the helicopter to make them useful at distances greater than 400m.

Environmental Monitoring

Dust Trak sampling has been carried out at several sites around the island showing comparatively high values in the northern and eastern area and high values in the Salem area. The central area also shows elevated values. Results are shown in Table 4.

Table4. PM10(concentrationinmg/m3)

Location	15/9/97	16/9/97	17/9/97	21/9/97	22/9/97	23/9/97	24/9/97	25/9/97	27/9/97	
Angelo's	-	-	-	-	-	-	-	-	-	
BC		-	-	-	-	-	-	-	-	-	
Bso		-	0.101	-	-	0.042	-	-	-	-	
CB		-	-	-	-	-	-	-	-	-	
GH		-	-	-	-	-	-	-	-	-	
Gpo		-	0.11	-	-	0.047	-	-	-	-	
JBH		-	0.101	-	-	-	-	-	-	-	
Mc		-	-	-	0.214	-	-	-	-	-
MVOi		-	-	b 0.097	-	-	-	-	-	-	
MVOo		b 0.074	b 0.135	b 0.298FAv 0.224	0.211	-	-	-	-	
MVOpd		-	-	-	-	-	-	-	-	-	
MVOn,0		-	-	-	-	-	-	-	F 0.031	c 0.022	
RBC		-	-	-	-	-	-	-	-	-	
RC		-	-	-	-	0.165	0.043	0.851	-	
SCP		0.197	-	-	-	-	-	-	-	
SPSh.I		-	-	-	-	-	-	-	-	-	
SPSh.o		-	0.114	-	-	-	0.164-P 0.	0.258	-	-	
Sso		-	0.16	0.12	-	-	-	-	-	-	-	
SSSo		-	0.126	-	-	-	0.056	0.315	0.057	-	
SSSi		-	-	-	-	-	0.277	-	-	
Fogarty		-	-	0.182	-	-	-	-	-	
Lawyers		-	-	1.025	-	-	-	-	-	
Lookout(i)	-	-	P 1.5	-	0.1	-	d 0.147	a 0.056	-	c 0.022	
Lookout(o)	-	-	-	-	-	a 0.032	-	F 0.372	-
MWAi		-	-	-	0.05	-	-	-	-	-
These values represent the average reading over the sample duration. The mean duration of a sample is usually between 15 and 30 minutes. Samples of longer duration are labelled with the following letters:
  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
i       sample taken inside
o       sample taken outside
F       during fallout
P       personal sample (during activity)

24 Hour averages:
<0.05-3 Low
0.05-0.1 mg m-3 Raised
0.1-0.3 mg m-3  Very High
> 0.3 mg m-3    Alert

Angelo's:	Supermarket (Cudjoehead)
BC:		Brades Church (North)
CB:		Carr's Bay (North)
GH:		Garibaldi Hill
GP:		Gerald's Park Shelters (North)
JBH:		Jack Boy Hill (eastern area)
Mc:		McChesney (Olveston)
MVO:		Montserrat Volcano Observatory (Old Towne)
MVOpd:		MVO pooldeck
RBC:		Royal bank of Canada (Olveston)
RC:		Red Cross (Woodland)
SPSh.:		St. Peter's Shelter (North)
SCP:		Salem Car Park
SS:		Supermarket in Salem (RAMS EMDEE)
SSS:		Secondary School Shelter (Salem)
MVOn:		(The new Volcano Observatory in the north on Mongo Hill)
Lookout:	(A villa in Olveston). 
MWA:		(The Montserrat Water Authority office in Woodlands).

The mini-COSPEC was deployed on 20 September with the use of the police launch. A series of traverses were made under the plume at different distances from the volcano. The average SO2 flux was found to be 600 tonnes per day. Unfortunately the launch broke down the following day and will be out of action for over a month.

Sulphur dioxide diffusion tubes were collected from four sites to the north of the volcano: MVO (south), Lawyers, Fogarthy and the MDF base at Geralds on 12th September. The diffusion tubes measure the average background level of SO2 gas during the exposure period. As in the previous sampling period, SO2 gas is not present in measurable quantities.

Table 5. SO2 concentrations (ppb) from diffusion tube sites. Recommended action level is 100 ppb.

Location	14-Aug-97 to 29-Aug-97		29-Aug-97 to 12-Sep-97
MVO(south)	0.0				0.0
Lawyers		0.0				0.0
Fogarthy	0.0				0.0
MDF base	0.0				0.6

Rainwater was collected on 21 September and showed low pH, one site also showed substantially elevated chloride content. See Table 6.

Table 6. Rainwater geochemistry 12th to 21st September

Location	pH	Cond	TDS	Flrd	Chlrd	Slpht
			mS/cm	g/l	mg/l	mg/l	mg/l
MVO(south)	3.01	1.655	0.829	>1.5	nd	nd
Lawyers		2.74	0.695	0.348	1.4	171	7
Fogarthy	2.95	3.03	1.51	nd	>650	57

Values in bold are outside WHO guidelines. Nd: not detected.

Ash is being collected regularly from 16 sites around the volcano and samples are being sent on a regular basis to the UK for analysis. Preliminary analyses suggest that the cristobalite content of the ash generated during the explosions is lower than that in ash generated during dome collapse events.

Staff Changes
Chloe Harford (University of Bristol, UK)
Rod Stewart (Independent, UK)

Dr Peter Baxter (University of Cambridge Clinical School, UK)
Dr Peter Francis (The Open University, Milton Keynes, UK)

Montserrat Volcano Observatory