A surge’s dangerous offspring: secondary pyroclastic flows

T H Druitt, E S Calder, P D Cole, R P Hoblitt, S C Loughlin, G E Norton, L J Richie, R S J Sparks and B Voight, Small-volume, highly mobile pyroclastic flows formed by rapid sedimentation from pyroclastic surges at the Soufrière Hills Volcano, Montserrat: an important volcanic hazard. Geological Society of London, Memoirs (2002) 21: 263-279.

Because volcanology is a relatively new branch of science, phenomena that have not previously been observed are still regularly reported. During the years since 1995 the erupting Soufrière Hills Volcano (SHV) has served up two special surprises to the scientists observing it. Both of these matter to Montserrat residents, and not just the scientists, because potentially very dangerous things have already happened when they were totally unexpected (fortunately without casualties in these two cases).

One of these phenomena was described and discussed in another section of this website titled “How pyroclastic surges sometimes ‘bounce’ landwards when they reach the sea”. The second phenomenon appeared as soon as the surge ash clouds cleared along the north flanks of the volcano on 25 June 1997. The surges that devastated Farrell’s Plain and Streatham “escaped” from a series of pyroclastic flows (PFs) that travelled down Mosquito Ghaut and Paradise River to the area around Spanish Point and W. H. Bramble airport (Fig. 1). As the dense block-and-ash flows in Mosquito Ghaut curved eastwards within that valley system, the associated surges charged straight ahead over Farrell’s Plain. This is a relatively flat watershed between river valleys draining east (Mosquito/Paradise) and west (Dyer’s/Belham).

At that time the “textbook” model of how such a surge would behave was as follows:

1. 1.  The surge flows fast down steep slopes.
2. 2.  The surge slows and spreads out as it reaches Farrell’s Plain.
3. 3.  Grains and larger pieces of volcanic material, carried along when the surge moves very fast, are progressively dropped from its base when it slows. This process coats     the ground with a thin ash deposit and reduces the overall density of the surge.
4. 4.  The hot surge now has a low density and rises as a dusty cloud.

The implication was that anyone lucky enough to be standing just beyond where the surge reached would see it stop short of them and then loft harmlessly. As it happened, one of the eyewitness survivors on 25 June was ideally placed to see what actually happened. This person, a farmer (eyewitness 9 in the 2002 report of Loughlin and colleagues), was heading along the road from Streatham to Windy Hill when he looked back and saw; “A small drift of mainly ash came onto the road…. We never heard any sound…. When it hit the road it went straight down the road (westward towards Dyer’s – at right angles to the movement of the surge cloud!)…. The road was boiling and the ash moved around the bends on the road like a vehicle”.

When an MVO field team could safely study the deposits, they found that similar “boiling ash flows” had moved westwards off the Farrell’s Plain-Streatham plateau and amalgamated in the head of Dyer’s River valley. The resulting granular hot ash flow moved down the gently inclined Dyer’s and Belham River valleys, including around their tight meanders, and finally stopped 3 km downstream, in a gorge about 50 m below the small town of Cork Hill where ~200 children were in their school at the time. Much the same phenomenon happened again in the same area on 10 December 2009. Fortunately this is now within a forbidden zone.

The volcanic ash that dropped from the surge cloud and covered Farrell’s Plain was 5-20 cm thick and had two distinct sub-divisions of similar thickness. The lower layer was dust-poor and mostly sandy in grain-size, with a smaller coarser gravelly fraction that even included a few lava chunks several cm in size. The upper layer lacked gravel fragments and instead included more fine sand and dust. These two layers could also be found on the floor of Dyer’s/Belham River valley, where they formed a deposit 0.1-1 m thick that was plastered up the valley sides to a height of 6 m (slumping down afterwards in places). The vegetation above this deposit was scorched up to 10-15 m above the valley floor – small wonder, as the deposit was still at 350-410^oC internally several days after the eruption. The scorching and also a thin layer of tell-tale ash beneath the singed vegetation showed that the denser dust-poor sand-plus-gravel volcanic ash flow that travelled down the valley floor had its own accompanying surge cloud above. Both must have moved quite slowly because they failed to break trees and branches, despite scorching them strongly. Their maximum speed was estimated to have been 10 meters per second.

The sequence of events on 25 June revealed by the various deposits was as follows:

    1.  As a result of progressive collapse of part of the SHV summit lava dome, several pulses of block-and-ash pyroclastic flow travelled down Mosquito Ghaut, Paradise River and spread out on the less-sloping land below. These PFs and their deposits comprised lava boulders (up to 5 m or more in diameter) that were swept downhill within abundant volcanic ash.

    2.  The surges that detached from these PFs and covered Farrell’s Plain and Streatham moved fast enough at first to carry sand-gravel-sized ash and even some lavas chunks (up to 10 cm).

    3.  As the surges slowed they rapidly dropped their load of lava sand-pebbles and most of these simply settled on the land below.

    4.  Something weird happened to the debris falling out of the surge cloud as it slowed above the western side of Farrell’s Plain. Instead of hitting the ground and staying there, the sand-gravel-sized ash (and even some of the larger lava chunks) became an extremely mobile hot grain-flow that decanted like a liquid into the Dyer’s/Belham Valley and trundled slowly but inexorably down it for 3 km. Fig. 4 summarises what the investigators found in Dyer’s River valley.

Fluidisation

The key technical term at this point is “fluidisation”. This process is widely used in industry and converts a mass of solid grains – like a pile of sand – into a fluid state where they flow just like a liquid. Fluidisation is best explained in the opposite direction from what happened on Farrell’s Plain. Let’s start with some dry sand in a container and pump air very slowly into its base. Nothing visible happens because the air trickles smoothly through the network of gaps between the sand grains and escapes out from the surface of the sand. Now (if you must!) try the opposite and pump air very fast through the sand. The air lifts up all the sand grains and they fly around turbulently, as in a surge cloud. Now (after cleaning up the mess!) return to the very-slow-air-flow experiment and gradually increase the air flow rate through the sand. At a certain flow rate each sand grain is surrounded by enough air to be able to shuffle around a bit but not to fly freely. This is the fluidised state. The surface of the sand rises in its container, because the sand with air passing through it occupies more space than the sand alone, and its surface appears to “boil”. If you place a coin on top of such a fluidised layer of sand, it will sink – as though into a liquid. Needless to say, if you can imitate Nature and organise a situation where a mass of unconfined fluidised sand is on a slope, it will flow downhill – like water in a river.

All these words may be brought to life by looking at 4 videos on YouTube:

http://uk.youtube.com/watch?v=3a7qNU9PRc8&feature=related

This is the “classic” laboratory experiment. Now you can see why eyewitness 9 talked about volcanic ash “boiling” and moving “around the bends in the road like a vehicle” (behaving like a liquid).

http://uk.youtube.com/watch?v=0HQbIoT4QKg&feature=related

This shows violent fluidisation used in industry for reacting, drying etc.

http://uk.youtube.com/watch?v=uLhblc48NIA&feature=related

This is just for fun but emphasises how the sand becomes a type of liquid. Do not try this at home without permission!

http://uk.youtube.com/watch?v=Ixu4BTksTeQ&feature=related

The temperature in this industrial fluidisation furnace is 840^oC, so that the particles are red hot. This is what scientists deduce to have happened when molten viscous (sticky) lava has been erupted as PFs (often extremely large – vastly bigger than those on Montserrat) and the fragments have stayed hot enough when they stopped moving to weld together, under their own weight, and become a solid hard glassy rock type called “ignimbrite”.

Now we can think about how pyroclastic flows form in Nature. The two types seen during the SHV eruption have been:

    1.  Collapse of a dome, sending a violent cascade of volcanic rubble down the steep slopes of the volcano. Some are just rock falls. Others build up enough internal gas pressure from various sources (exploding gas-rich pieces of lava; overridden, trapped and heated air escaping upwards; trapped moisture flashing into steam and doing likewise) to establish fluidisation of the debris and birth of a pyroclastic flow.

    2.  Ash, blocks etc cascading down from a vertical explosion often use the same gas generation and trapping methods to become pumice-rich PFs, slithering down the volcano flanks. This happened on a large scale most recently on 11 February 2010.

The events of 25 June 1997 in the Dyer’s/Belham Valley showed that a small-scale version of exactly the same process can happen as flying debris settles from a surge cloud.

At first the highly mobile grain-flow down Dyer’s/Belham Valley on 25 June 1997 was a one-off phenomenon that could not be studied in detail because the deposits lay downhill from a dangerously unstable lava dome. Fortunately (for science) the same thing happened again on 26 December 1997, when a wholesale collapse of part of the upper SW sector of the volcano unroofed the buried parts of the lava dome. The latter immediately exploded sideways and sent an enormous and particularly violent pyroclastic surge across that sector of the island, sweeping the village of St Patrick’s and the surrounding area wholesale into the sea. Afterwards the deposits of a series of these distinctive, dust-poor, sand-and-gravel-grainsize secondary pyroclastic flows (carrying lava chunks up to a few cm in size) were found in several gullies. The most remarkable was in Dry Ghaut – again at right angles to the direction of the surge! This Dry Ghaut deposit was studied in great detail by Druitt and his colleagues and confirmed everything they had seen of the 25 June 1997 example.

Could “secondary” pyroclastic flows cause problems for Montserrat residents in the future?

The research outlined above on the two recorded instances of secondary surge-derived PFs on Montserrat show clearly that these strange fluid “things” form while flying sand-and-gravel-sized volcanic ash is settling to the ground beneath a big pyroclastic surge. As the fragments pass from turbulent flight to a stable layer of ash on the ground, they go through a fluidised state and if there is a suitable route downhill for them to flow (a gully or road for instance) they become a secondary PF. Druitt and his colleagues estimated that the accumulation rate of the ash was at least several millimetres per second.

As the upper Belham Valley is already evacuated long-term, the only threat from pyroclastic surges to island residents will be if one engulfs the lower Belham Valley and its flanks. The probability of such an event was discussed at the end of section 4 (surge damage to buildings) of this website. This probability remains small but non-zero at the time of writing (according to MVO/SAC estimates).

A potentially dangerous situation for secondary PFs would occur if a very large surge, probably generated by a lateral explosion of the SHV lava dome to the NW, was to invade the area between the Belham and Nantes rivers. If such a surge generated secondary PFs as it dropped its suspended ash load, these might wander in unpredictable directions down gullies and perhaps also along sloping roads and drives.

The best way to avoid such a horrendous situation is obviously to evacuate the area fast, whenever MVO/SAC and their government colleagues judge this to be necessary. This is an irritating prospect but far less irritating than being engulfed by the double whammy of a surge and its secondary PFs.