Montserrat Volcano Observatory, Montserrat, West Indies

Scientific Report 33
August 1996

This report covers the period 1 to 31 August, 1996. It was written in February 1997, and the author was not in Montserrat during August; therefore the detail contained in the report is not as great as in normal MVO scientific reports. Much of this report is based on the MVO daily reports and other data files.


The month began at a point directly following a major dome collapse on 31 July. The level of activity in terms of rockfalls and pyroclastic flows was low, but the lava extrusion rate into the amphitheatre created by the previous collapse was inferred to be high. However, rapid growth of the dome quickly replaced material excavated during the period of heightened pyroclastic flow activity and a pattern of near-continuous rockfalls and pyroclastic flows re-commenced by 4 August. Another major collapse took place on 11 August, followed by further extrusion and finally another collapse on 21 August. Throughout the month the dome grew towards the north-east. Rockfalls any pyroclastic flow continued to travel down the Tar River valley on the north side of Castle Peak, reaching as far as the new delta during the two major collapse events. Ash clouds produced by this activity caused ashfalls in areas to the west of the volcano. On some occasions light and variable winds caused ash to be deposited in the central parts of the island. During the major collapse events, once pyroclastic flows reached the sea, ash fell as wet accretionary lapilli and were accompanied by frequent lightning strikes.

Visibility was frequently hampered by low cloud cover but there were a number of exceptionally clear periods when it was possible to get excellent views and conduct dome volume measurements.

Visual Observations

At the beginning of month, the spoon-shaped collapse structure produced by pyroclastic flows of 29 and 31 July was rapidly refilling. Observation during clear weather on 2 August showed that the previous collapse scar was virtually full of fresh, dark lava. The new feature had a whaleback form with several large blocks. Vigorous steaming with small amounts of ash was emitted from the new excavation and this corresponded with periods of high amplitude harmonic tremor. Steam emission was observed from most of the dome, from some areas of the Tar River valley and from areas where recent pyroclastic flows had entered the sea. Most of the rockfalls and small pyroclastic flows originated from the north-east flank of the dome and did not produce many ash clouds. From 2 to 4 August the new extrusion grew rapidly upward and eastward, forming a bulging and steep mass on the dome complex NE flank.

Frequent rockfalls and small pyroclastic flows re-commenced as of 4 August. Ash clouds were observed up to 10,000 ft on 4 August generated by pyroclastic flows produced by collapses from over-steepened parts of the dome. Some parts of the new extrusion began to appear over-steepened and unstable by 6 August. Rockfall activity was concentrated in an area just north of Castle Peak. No major changes were observed during good visibility on 7 August, and the visual impression was that the extrusion rate had decreased markedly since 2-4 August. The new dome now growing in the collapse scar had almost filled this structure and some areas continued to look over-steepened and unstable. Rockfalls were again concentrated from an area north of Castle Peak. These, along with small pyroclastic flows into the upper Tar River valley, produced ash clouds to the west over upper Gages.

During the late evening on 8 August a few pyroclastic flows were produced in the Tar River valley which travelled as far as the Tar River Soufriere and generated ash clouds up to 7,000 ft above sea level. Ash drifted to the west with the prevailing winds and caused small ashfalls in upper Gages, Amersham and Plymouth. A few episodes of vigorous, near-continuous ashy steaming were observed on 10 August during a short period of clearing. Near continuous rockfall and pyroclastic flow activity in the Tar River valley during the morning of 10 August produced ash clouds up to 5,000 ft and caused light ashfalls in Plymouth.

Another series of near-continuous rockfalls and pyroclastic flows in the Tar River valley caused by collapse of parts of the eastern flank of the dome began at 7:00 am 11 August. The first flow reached the sea at about 9:00 am. Near-continuous pyroclastic flow activity continued throughout the day, although prior to 17:00 the flows had runouts of only 800 m or less. The pyroclastic flow activity escalated to a maximum level of activity at around 17:00, and continued for about an hour. During this vigorous period, several flows reached the sea at Tar River and produced significant ashfalls in Plymouth, Richmond Hill, Fox's Bay, St George's Hill and Cork Hill. The largest ash cloud during this period was associated with lightening and thunder; it occurred at 5:17 pm and was estimated to reach 30,000 ft.

During the period 12-15 August there were frequent but not continuous rockfalls concentrated on the east flank of the dome, and a few small pyroclastic flows in the Tar River valley which reached as far as the Soufriere and Tar River estate house. On 14 August fresh blocky lava was observed being extruded into the head of the collapse cleft. Also that day a jetting ash-laden plume was observed rising from the cleft area, and was not directly associated with any rockfall or pyroclastic flow activity. Ash clouds were often generated which drifted to the west with the prevailing winds.

Over the period 15-17 August, the cleft formed by the previous collapse became almost filled with debris and new lava. Most of the east face of the infilling material was a steep, near planar surface of coherent lava. An eroded gully channelling material from the upper flanks of the east dome had developed by 17 August. The eastern flank seemed unstable with most rockfalls occurring from this area. The largest rockfalls were associated with small ash clouds the highest reaching 6,000 ft above sea level. From the 18 August rockfalls and pyroclastic flow activity escalated again, in terms of frequency and runout, and were focused on this channel eroded just north of Castle Peak on the eastern flank of the volcano. These produced ash clouds which drifted to the west depositing ash on areas on the western flank of the volcano. The eroded gully gradually became wider and deeper as rockfalls and pyroclastic flows continued to be focused in this area.

Three episodes of near-continuous rockfalls and pyroclastic flow activity on 19 August combined with a change in wind direction to cause light ashfalls in central Montserrat. However ashfalls were heaviest in Amersham, Parsons, Plymouth, Richmond Hill and Kinsale. Satellite imagery and aircraft reports indicate that a belt of very light ash extended from Montserrat to close to Puerto Rico as a result of ash clouds produced during the day. The flow runouts during active periods were around 800 m but a few longer flows extended to about 1.5 km.

Pyroclastic flow activity continued at a heightened level through 20 August. The longest pyroclastic flow resulting from activity on 20 August reached the new delta but did not enter the sea. Several of these flows generated ash clouds up to 5,000 to 12,000 ft and ashfalls to the west and central parts of the island. Sometime shortly after 10 am on 21 August continuous pyroclastic flow activity began again, culminating in flows first reaching the delta at about 11 am. Pulsing flows continued to move around 2 km down the Tar River valley until 13:45. Flows on 21 August reached the new delta but generated ash clouds only up to 8,000 ft.

Good visibility on 22 August showed another cleft carved in the NE flank of the dome, in the same location as the previous collapses. At this time a new, low extrusion was again visible at the head of the cleft. After 2 days of poor weather, good views on 25 August showed that the collapse scar was almost refilled. Once again there was vigorous ashy gas emissions from the top part of the collapse structure. Several other areas on the dome were also steaming. Rockfalls were occurring from the east flank of the dome and were channelled through the eroded gully which continued to be filled. The remnant of the new dome which was extruded in the void left by the 29 and 31 July collapse appeared to be very unstable. Visual observations during variable conditions on 26 August indicated that the gully to the north of Castle Peak was filled. Near-continuous rockfalls of a very small size were observed from the eastern side of the dome. On 27 August there were several episodes of vigorous steam and/or ashy steam emission from the top of the collapse structure. Throughout this period several areas of the dome continued to look very unstable and rockfalls and small pyroclastic flows were frequently generated in the Tar River valley.

The dome appeared essentially unchanged on 28 August with less vigorous steam emission from several areas of the dome. Rockfall activity continued to occur from the east flank and several areas of the dome looks very unstable.

Brief views obtained on 30 August suggested that two new features (possibly spines), had been extruded in the central part on the eastern side of the dome. Light ashfall was associated with some of the larger rockfalls and or small pyroclastic flows seen to the south and south east. Ash clouds reached above 3500 feet.


Near-continuous periodic bursts of rockfalls and pyroclastic flow signals were the most dominant feature of seismicity in August. A few volcano-tectonic and hybrid earthquake swarms occurred during the latter part of the month. Long period events remained at background levels and a few regional earthquakes were recorded by the network. Flash flooding in Fort Ghaut also generated a seismic signal on several days during the month.

Table 1: Earthquake types: 1 to 31 August 1996
These earthquake counts are of events that triggered the short-period seismic network event recording system between 0000 and 2400 each day.

Date		VT	LP	Hybrid	Dome RF
01 AUG 96	3	9    	43	34
02 AUG 96	6	0    	58	11
03 AUG 96	35	4    	136	27
04 AUG 96	13	0    	119	77
05 AUG 96	29	0    	126	24
06 AUG 96	15	0    	137	22
07 AUG 96	7	0    	129	42
08 AUG 96	41	2    	140	50
09 AUG 96	83	1    	92  	69
10 AUG 96	36	3    	29  	98
11 AUG 96	33	0    	85 	121
12 AUG 96	0	1    	50  	119
13 AUG 96	20	2    	57	119
14 AUG 96	21	2    	44	80
15 AUG 96	27	0    	35	65
16 AUG 96	1       6    	10	66
17 AUG 96	25	2    	14	69
18 AUG 96	14	1    	13	64
19 AUG 96	15	1    	24	99
20 AUG 96	21	1    	3	157
21 AUG 96	4       5    	39	73
22 AUG 96	6       7    	1	64
23 AUG 96	0	0    	5	43
24 AUG 96	3	0    	2	33
25 AUG 96	22	0    	10	45
26 AUG 96	6	0    	12	49
27 AUG 96	9	0    	2	18
28 AUG 96	0	17   	27	51
29 AUG 96	15	9    	20	29
30 AUG 96	64	3    	17	35
31 AUG 96	25	4    	16	61

Seismicity at the beginning of the month was characterised by episodes of sustained low to high amplitude, low-frequency, harmonic tremor which occurred at about three to four hour intervals. These episodes started with a series of small volcano-tectonic events followed by small, near-repetitive hybrid earthquakes, or with just small, near repetitive hybrids which increased in frequency and amplitude until they formed continuous tremor. After a period of sustained peak amplitude, the signal decayed first into smaller hybrids and then into background. These episodes of volcano-tectonic and/or repetitive hybrid and harmonic tremor lasted about 60-90 minutes and were felt to be associated with magma migration from shallow depths to surface or near-surface regions. On some occasions large steam plumes mixed with small quantities of ash were observed to be associated with times when the tremor level was very high. One possibility is that this behaviour is due to the interaction of magma with the volcano's ground water system. Another is that it represents vigorous ash laden fumaroles where magmatic gas is venting from the pressurized dome interior.

The seismicity pattern began to change during the second week in August with the duration of the tremor episodes becoming more variable and extending between less than one to more than two hours. In addition, the amplitude and rate of occurrence of the hybrids which preceded the tremor did not increase to the point where the events merged into continuous tremor but instead remained distinct events throughout the episodes.

At 17:20 on 8 August, a series of near-continuous small rockfalls from the eastern flank of the dome led to a few small pyroclastic flows into the Tar River Valley. This activity signalled a change in seismicity which thereafter became dominated by near-continuous rockfall and pyroclastic flow signals which generated several pyroclastic flows which reached the sea at the Tar River valley. This pattern continued up to 21 August, eventually becoming episodic in nature with up to 3 episodes of near-continuous rockfalls being experienced each day. Short bursts of volcano-tectonic earthquakes often preceded these periods of rockfalls and pyroclastic flow activity. Near continuous low-amplitude tremor accompanied rockfall activity on some occasions.

The level of seismic activity decreased from the 22 August onwards, although the style of activity remained essentially unchanged. Activity again picked up after 26 August but remained dominated by rockfall signals which often occurred during discrete periods. A few volcano-tectonic and hybrid earthquake swarms occurred on 30 August, but rockfall signals continued to be the most common events at the end of August.

Low amplitude broadband tremor recorded at the Gages seismic station were recorded at various times throughout the month. These were usually associated with flash flooding in Fort Ghaut or with increased steam emission from the crater.

A few regional earthquakes were also recorded during the month.

Ground Deformation

Measurement of ground deformation with the Leica Total Station firmware was carried out on the eastern, northern, western and southern networks during the month. The only lines which continue to show consistent changes were those on the Eastern Triangle to Castle Peak. Line lengths to Castle Peak from Whites and Long Ground decreased by 12-13 cm during the month giving an average daily shortening of about 4 mm per day during the month. Changes on the other networks were quite variable and do not define any long term trends. Increases in line lengths were observed on the radial lines from lower to upper Amersham line (1.3 cm) and from lower Amersham to Chances Peak steps (1 mm). Line lengths on the southern triangle from Galway's estate and O'Garro's to Chances Peak were also inconsistent; the Galway's-Chances Peak line increased by 3.2 cm while the O'Garro's-Chances Peak and O'Garro's-Galway's lines shortened by 4.7 mm. Line lengths on the western triangle to Farrells all shortened during the month. The largest amount of shortening on this triangle was recorded on the Windy Hill-Farrells line (15.9 mm); the St George's Hill-Farrells and St George's Hill-Windy Hill lines shortened by 6.9 and 7 mm respectively during the month.

Lava dome and Tar River Surveys

Surveys of the dome, talus deposits in the Tar River valley and the new delta fan were conducted during periods of good visibility with the GPS-vector binocular helicopter surveys. Table 2 summarises the dome volume measurements.

Table 2: Measurements of dome volumes for August

Date		Volume (cubic metres)    
09-AUG-96	26.42
12-AUG-96 	23.70
16-AUG-96	25.87
17-AUG-96	27.32
25-AUG-96	27.86
04-SEP-96	27.68

The highest point on the dome was measured from Chances Peak on 16 August as 3161 feet. Measurements made on 9 and 12 August, bracketing a major collapse event, showed that the collapse represented the calving off of a lava dome flank about 130 m thick. The same measurements, and others done with the laser binoculars from the Tar River EDM site, showed that the collapse events and pyroclastic flows were highly erosive: near the dome, a maximum of 80 m thickness of Castle Peak dome rock had been removed beneath the collapsed flank, and in the Tar River valley local maxima of 5-10 m thicknesses of older rock had been scraped away.

Gas Measurements

The daily rate of sulphur dioxide flux was determined using a mini-cospec manufactured by Barringer Research and provided by a NERC emergency grant. The results are summarised in the table below.

Table 3: Measurement of sulphur dioxide flux with the Correlation Spectrometer for August.

Date		Sulphur dioxide flux (tonnes/day)
03 AUG 96	625  
09 AUG 96	326  
12 AUG 96	1197 
13 AUG 96	625  
25 AUG 96	254  
29 AUG 96	129  

Environmental Monitoring

Sulphur dioxide diffusion tubes were installed at locations in and around Plymouth. Results are shown in Table 4, and show values within acceptable levels. Another set of tubes were put at the same locations on 20 August, and will be collected in early September.

Table 4: Sulphur dioxide diffusion tube results for August 1996.
Levels are in ppb.

Location		Date
			22.07.96 to 20.8.96 
Upper Amersham		63.75
Lower Amersham		21.65
Police HQ, Plymouth	9.80 
Observatory		0.40 
Weekes			1.10 

Water samples taken at four sites around the volcano have been tested for pH, conductivity, total dissolved solids, sulphate, fluoride and chloride and show no significant changes. Results are shown in Table 4.

Table 5: Chemistry of rainwater from 4 sites around the volcano in August 1996
Units: conductivity (mS/cm), total dissolved solids (g/l), sulphate (mg/l), chloride (mg/l), fluoride (mg/l), nd (not detected)

Date Location			pH	Conductivity	TDS	Slphts	Chlrds	Flrds
					(ms/cm)		(g/l)	mg/l	mg/l	mg/l 
28.07 to 09.08 Upper Amersham	2.80	1.280		0.641	29	211	1.3  
09.08 to 20.08 Upper Amersham	3.14	2.960		1.470	nd	80	>1.5 
28.07 to 09.08 Parsons		2.88	0.932		0.467	33	144	1.15 
09.08 to 20.08 Parsons		3.63	0.349		0.174	14	74	1.3  
28.07 to 09.08 Plymouth		3.01	0.716		0.359	25	158	1.2  
09.08 to 20.08 Plymouth		3.24	1.107		0.653	10	38	1.45 
28.07 to 09.08 Weekes		4.05	1.790		0.897	>200	410	0    
09.08 to 20.08 Weekes		4.19	0.377		0.188	0	83	1.1  

Observatory Staff During August
William Ambeh, SRU
Anne-Marie Lejeune, Bristol University / BGS
Mark Stasiuk, Lancaster University / BGS
Chandradath Ramsingh, SRU
Rob Watts, BGS

Chloe Harford, Bristol University / BGS
Chris Kilburn, University of London / BGS
Simon Young, BGS

Sally Bower, BGS
Rod Stewart, BGS
John Shepherd, Lancaster University / BGS
Jane Toothill, Lancaster University
Glen Thomson, BGS

Clive Oppenheimer, Cambridge University
John Stix, University of Montreal

Montserrat Volcano Observatory