Frogs in saucepans, and cats and mice: key danger characteristics of the Montserrat eruption.




E Clay, C Barrow, C Benson, J Dempster, P Kokelaar, N Pillai, and J Seaman, An Evaluation of HMG’s Response to the Montserrat Volcanic Emergency. Department for International Development Evaluation Report EV635 (1999).


Only volume 1 of this two-volume report is available via the internet.



B P Kokelaar, Setting, chronology and consequences of the eruption of Soufrière Hills Volcano, Montserrat (1995-1999). Geological Society of London, Memoirs (2002) 21: 1-43.


The PDF version can be downloaded free but is a very large file (205Mb; many photographs).



S C Loughlin, P J Baxter, W P Aspinall, B Darroux, C L Harford and A D Miller, Eyewitness accounts of the 25 June 1997 pyroclastic flows and surges at Soufrière Hills Volcano, Montserrat, and implications for disaster mitigation. Geological Society of London, Memoirs (2002) 21: 211-230.



S C Loughlin, E S Calder, A Clarke, P D Cole, R Luckett, M T Mangan, D M Pyle, R S J Sparks, B Voight and R B Watts, Pyroclastic flows and surges generated by the 25 June 1997 dome collapse, Soufrière Hills Volcano, Montserrat. Geological Society of London, Memoirs (2002) 21: 191-209.




In this section I shall summarise the features of the Montserrat Soufrière Hills volcano (SHV) eruption that have made it particularly dangerous (both in the past and maybe also in the future) to nearby residents. Initially it is very difficult to see why an eruption that consists only of a rubbly lava dome atop the volcano, with the dome collapsing from time to time, and this process sometimes causing explosions triggered by expansion of volcanic gases under high pressure at depth, should be so much trouble to anyone living and/or working nearby. Surely it is just a matter of keeping an eye on the dome and making sure not to be too close when it collapses?


Within the last sentence lie the first two special dangers of the Montserrat eruption, as pointed out in the formal evaluation of the UK Government’s response to the early parts of the crisis by Clay and his colleagues in 1999. Kokelaar subsequently (2002) summarised this evaluation, with special emphasis on the actual volcanic events.



Danger 1: Reactive evacuation


There is a firm template in everyone’s mind as to what a major explosive volcanic eruption should be like. You could call it the “Hollywood eruption disaster movie template”: After centuries of peaceful dormancy, the volcano begins to stir; as volcanic activity rapidly escalates, the surrounding population wake up to the impending danger (prompted by the hero/heroine?); the volcano finally “blows” and everyone flees to safety, hopefully just in time; the eruption ends and people return to rebuild their lives.


As all Montserrat residents have discovered, lava dome eruptions are entirely different. A smallish heap of smoking rubble appears on top of the volcano and grows slowly. The reaction of both government and residents is “business as usual”. Only when a dome collapse and/or explosion are blatantly threatening to devastate an area, do people reluctantly move out of the way. As the dome enlarges, changes its site and gradually claims more and more of the land on the volcano slopes, residents very reluctantly retreat to safer places.


The problem with this piecemeal retreat is that it requires phenomenal concentration by scientists, government and residents to ensure that everyone takes a step backwards every time the dome takes a step forwards. Although it avoids expensive and demoralising fiascos, like the “false alarm” mass evacuation on Guadeloupe in 1976 (see Wikipedia), it is an immense strain on all involved – possibly for as long as a generation. The danger with reactive evacuation is that it relies on the whole complex “machine” of scientists, government and residents collaborating faultlessly for anything up to a lifetime. Any weakness in the humans and the volcano wins a round. This leads us to a human weakness that is the second danger.



Danger 2: The frog-in-saucepan phenomenon


There is a widely quoted adage that: “If you throw a frog into a saucepan of boiling water it will jump straight out, but if you put it in a saucepan of cold water on a very low heat then the frog will not realise that the water is slowly warming up and will boil to death”. Humans behave remarkably like frogs (Figs 1-2) when danger is developing in slow incremental steps. This is a universal human characteristic and Montserrat has been no exception, as pointed out by Kokelaar in 2002.



Fig. 1  Thanks to for permitting our use of this cartoon.

Fig. 2  A Montserrat mountain  frog at a safe distance from the volcano (Sarah Bradley). Tragically, since this 2007 photo was taken, this frog species has become extremely rare on both Montserrat and Dominica, ~200 km south. The cause is a lethal fungal infection and desperate attempts are being made by an international group of scientists to try to save it from total extinction.


The first example is the situation that led to tragic fatalities in the Streatham Village area on 25 June 1997. In mid-May 1997 the summit lava dome began to grow on its NE side. Within a few days this led to rockfalls and small pyroclastic flows in Tuitt’s Ghaut, high above Bramble Airport. By mid-June the rockfalls and PFs resulting from dome collapses had switched to Mosquito Ghaut, above and beside the Streatham area, as described by Loughlin and her colleagues in a 2002 report. Nevertheless, a small number of farmers and others were “carrying on as normal”, right in the path of any pyroclastic surge that might escape from a PF in this ghaut and flow straight downhill towards them. Sure enough, earlier PFs partly filled the upper parts of the ghaut and reduced its capacity to channel future flows (Loughlin and colleagues, 2002). Eventually a larger-than-before collapse generated the lethal surge. A view of the situation taken on 16 June (Fig. 3) looks extraordinary with 20:20 hindsight (Fig. 4). What possessed people to remain so close to such violent volcanic activity?



Fig. 3  Dome collapse pyroclastic flow travelling down Mosquito Ghaut on 16 June 1997 (Copyright NERC). People continued working for 9 more days on the fields in the middle distance and around the houses in the foreground.

Fig. 4  The smooth white deposits and brown marginal zone of scorched vegetation mark the extent of the surge that detached from Mosquito Ghaut PFs on 25 June and swept through the still-populated area in the foreground of Fig. 3 (Copyright NERC).


It is sad but interesting to compare these photos with very similar ones taken during the 2010 eruption of Merapi, Indonesia,  where at least 330 people died in near-identical circumstances. It's extraordinarily difficult to persuade rural communities close to explosive volcanoes to evacuate in time.


This was the frog-in-saucepan phenomenon in action. The MVO scientists could see the danger clearly but nobody with the power to do so was prepared to remove the threatened residents forcibly to safety, and the economic and other circumstances of their lives encouraged them to stay. A probable further contributing factor was the cloud that caps the Soufrière Hills volcano most of the time (see Figs 3 and 4). Eyewitness accounts collected by Loughlin and colleagues following the 25 June tragedy (in a second 2002 report) show that people on the lower northern slopes of the volcano that day were less aware than they might have been about the dangerous behaviour of the summit lava dome because it was hidden in thick cloud.


Kokelaar (2002) dwells on the evacuation of Bramble Airport as his main example of frog-in-saucepan behaviour. Both the Montserrat residents and their two layers of government were determined to keep the airport open, unless closure was absolutely necessary. The result was an airport remaining in use until people there (fortuitously including the Governor) finally got the message from the volcano when a pyroclastic flow on 25 June reached to within 200 metres of the runway as the last aircraft was taking off (Kokelaar, 2002). Transpose this situation to one of the world’s major airports and you can see what the frog-in-saucepan phenomenon is all about. If you were to propose such a high level of risk to the management of, say, Heathrow Airport, you would surely be detained in a comfortable but secure place until your sanity returned! As things are in April 2010, the citizens of Montserrat, and countless other volcanic-ash-plagued places around the world, are having a huge belly laugh at the headless chickens in charge of northern European airspace during a trivial ashing event from Iceland! Most folk in the UK don't even realise that the daily brownish "dust" on their cars is ash!


I quote Kokelaar’s own suggestions as to how to deal with the Montserrat sort of situation; “Such problems deriving from progressive escalation of risks (as opposed to rapid onset) might be reduced if civil plans for disaster preparedness necessarily embed formal requirements for any emergency management team, comprising both scientists and officials, (1) to check regularly and deliberately, perhaps daily or even more frequently, for actual or potential progressive development towards dangerous or inappropriate activity, and (2) to moderate any such activity by default unless actions can be proved safe and/or appropriate to a high degree of probability. The slowly heated frog should be obliged to ask itself 'Is this comfortable?' and, if it is not quite certain, must jump out.”


As a subsequent postscript, I draw attention to the shocking state in 2010 of both earthquake and eruption forecasting in Italy. If the scientists involved fail to grow up and concentrate less on their personal grievances and more on the dangers they are well paid to think about, they may yet take their eyes off the dormant Albano volcano and literally still be bickering while Rome burns in the future. Montserrat has also had its fair share of such rivalry and recriminations in the past but things seem to be calmer and more consensual now.




Montserrat Volcano Observatory

Click here to reach page 2 of this topic