Jack Lausa, 4 August 2008


I will ask one question at this time of yourself and your readers: Sometime ago there was mention of levels of water changing in wells close to the volcano in apparent relation to the actions apparent at the volcano's surface. How does this phenomenon play out in relation to the movement of magma and possible explosive events? Would regular reports of this level help in assessments?



Bob Thompson, 11 August 2008


This is a fascinating topic. There are plenty of published studies about how the changing stress fields within the crust around a zone that is about to rupture, and thus generate an earthquake, causes fissures in the rocks to either seal or open and hence cause changes in the water levels within wells nearby. When new magma is pumped into the upper parts of a volcano, it raises the pressure within any reservoirs of magma already there. In turn this process squeezes the crustal rocks around the volcano, causing them to bulge upwards and changing the openness of rock fractures. In turn you might then expect changes in the water table measured in wells around the volcano, as these stresses interfere with the circulation of local groundwater.


There’s a bit of published research about cases where groundwater has shown strange behaviour before eruptions. Apparently water poured out of fissures at Mt Pelée just before the 1902 eruption. Water levels rose in wells around My Mayon just before a Mt Mayon, Philippines, eruption. The best known case is Mt Usu, Japan, where water both fell and then rose in surrounding wells during the months before the March 2000 eruption. These are all the published examples quoted by Steve Sparks in his excellent 2003 review of that material (R S J Sparks, Forecasting volcanic eruptions. Earth and Planetary Science Letters (2003): 210, 1-15).


Despite these examples, both Sparks, in his review, and Chris Newhall (USGS), in another excellent 2004 review “Promise and pitfalls of eruption forecasting” (,1,Can eruptions be forecast?), a  PowerPoint presentation, only draw attention to such results and do not dwell on them as major potential forecasting tools. Newhall gives a cartoon that shows the main methods of monitoring impending eruptions.



Although groundwater is mentioned (hydrology) it is the other methods that both he and Sparks concentrate upon.  The focus by MVO staff is on seismicity, ground deformation (especially bulging) and escaping gases (especially sulphur as sulphur dioxide, SO2). They also use remote sensing data and imagery whenever they can get the material. These methods are producing so many successes that there is probably not much impetus for MVO to use precious staff time checking other ideas.


I would imagine that someone visiting or temporarily attached to MVO in the past may well have researched water levels in Montserrat wells during eruptions but I cannot find any record of such studies. Of course the ratio of scientific studies to published ones is always low! It would be, to my mind, an excellent idea to approach MVO and see whether or not this possible relationship has been studied on Montserrat during the current long eruption. Then, if such research is not ongoing, it would be an excellent idea for interested residents to organise such a study themselves. These days we are all much too inclined to leave even observational scientific research to paid professionals (let alone experiments). Up to the nineteenth century is was normal for such research (including many amazing efforts) to be done entirely by amateurs.


The water levels data in wells are presumably something that observers could measure or get hold of without difficulty. It might take quite a long time (maybe several years) before the correlations between groundwater levels and eruptions did, or did not, emerge clearly. The tricky bit would be handling the raw data in such a way as to spot the water level changes related to impending volcanic activity, if they were partly masked by changes happening for other reasons, such as seasonal rainfall. This NASA diagram shows what a fiendishly complicated place the groundwater beneath an ocean island like Montserrat can be.



By the way, the Newhall PowerPoint presentation is one of many accessible from a 2004 conference on Volcanic Ash and Aviation Safety ( Some readers might find other presentations very interesting.



Rod Stewart, Acting Director of MVO, 15 August 2008


MVO and its collaborators work on a lot of areas that are not widely reported, including the groundwater on Montserrat. Montserrat’s water supply comes from natural springs, not wells. The flow of the springs is monitored regularly by Montserrat Water Authority staff. There are some variations in the data that might be related to the eruption but so far there is nothing that could potentially be used as a warning for impending volcanic activity.



Chris Newhall, Earth Observatory of Singapore, 23 October 2008


Your answer is fine, though we can go a bit beyond anecdotes.  I'm attaching a short review paper on the topic that a student and I wrote a few years back in the Journal of the Geological Society of the Philippines.   Basically, volcanoes are like water-saturated sponges and they respond to volumetric strain (compression, dilatation) induced by intrusions.   Broadly, the systematics of response follow Mogi predictions, as in the diagram that I reproduced in that little paper, but in detail there are at least as many exceptions as there are examples, so it's not really reliable as a monitoring method.   Much depends on the hydraulic conductivity and degree of confinement of an aquifer.  I'd be happy to send a pdf copy of this paper to anyone who asks me at




UK School student, 17 October 2008


How much damage was there after the eruption?



Bob Thompson, 18 October 2008


Since the eruption started in 1995 it has totally destroyed about two thirds of Montserrat.  Worse than that, the ruined part of the island included its capital town, Plymouth, the port and the only airport.  In 1995 there were about 15,000 people on Montserrat.  Now there are just under 5000.  Only 19 people were killed (in 1997) but, in such a small community, everyone knows everyone and so it was like losing members of their family.  The British Government (and also the EU) was rather reluctant at first to provide enough money to re-settle people in new houses in the north of the island. Therefore most people left and never returned.  For a while there was talk about forcing them all to leave but eventually the ones who stayed persuaded our government to let them try to keep the place running and gave money for some housing a etc and a small new airport.  They still lack a proper sea port and that is a major objective for the future.


If you want to imagine just how bad the eruption was for the Montserratians, think about a volcano totally destroying the same amount of England.  If the volcano was  roughly at Rugby, this would mean nothing left of England from the south coast up to north of Birmingham!  This shows what brave and determined people the Montserratians really are.



Constance Black and Salvatore Isola, February 2010


Is there any sort of connection between the Montserrat eruption and the massive Haiti earthquake on 12 January 2010?



Bob Thompson, 31 March 2010


This is another fascinating question because the two questioners asked about possible connections in "different directions". Salvatore wondered whether the Montserrat eruption caused the earthquake and Constance wondered if the earthquake had any impact on the Montserrat volcano. Although both events were/are spectacular at close range (and horrific in its effects on Haiti),  they were quite small on the scale of the forces that drive our planet. The Haiti earthquake was only news because it destroyed humans and their buildings etc in a densely crowded region. If it had happened somewhere free from humans, like the Atacama desert of Chile or the middle of the Pacific Ocean, we would have heard nothing about it, unless it triggered a tsunami.


The earthquake was caused by two of the Earth's relatively rigid so-called "tectonic plates" grinding together as they moved very slowly past each other. Look below or at Figure 7 of Topic 1 on this website and you will see that most of the Caribbean area is underlain by the Caribbean tectonic plate. In a simplified way, the Atlantic ocean floor of the South American plate dives (subducts) beneath it -- causing the Lesser Antilles volcanoes, including Montserrat -- and the North American plate slides sideways past it in the zone of islands that includes Haiti. Earthquakes at these two sorts of plate boundaries occur because the rocks of tectonic plates are extremely reluctant to move smoothly past each other and instead move in fits and starts, as the friction between them allows. Detailed maps of the tectonic situation in the Haiti and Montserrat region show that there is no direct connection between the local margin between two small tectonic plates that slipped to generate the earthquake and the Montserrat area.



How earthquakes concentrate along tectonic plate boundaries

Simplified diagram of tectonic plate boundaries around Hispaniola and Puerto Rico




Details of historic earthquakes in the Caribbean and Central America region

Historical large earthquakes around Hispaniola and Puerto Rico



I make no apology for including diagrams with French captions. Remember that the majority of Caribbean citizens living on islands with active volcanoes are French speaking. All the diagrams are taken from work by Eric Calais, Purdue University. The illustrations with French labelling come from a lecture he gave recently (early 2010) in Haiti. You will need to know a bit about seismology to understand the details of these diagrams fully (use Wikipedia to sort this out). The overall message from these diagrams is quite simple: the crust and underling lithosphere of the region around the islands of Hispaniola (includes Haiti) and Puerto Rico is chopped up into several small tectonic plates, but these are all well north of the subduction zone beneath the Lesser Antilles and therefore cannot influence events on it at all (and vice versa), except to the extent that extremely large earthquakes shake a very large area around their epicentres .


Is it therefore possible that a huge earthquake more than 1000 km away could shake the magma reservoir system of he Montserrat volcano sufficiently to trigger an eruptive episode? The only sensible answer to this question must be "maybe yes"! The Montserrat volcano has taken to exploding without detectable warning in the last couple of years. Without looking in any detail at the processes supposed to build up to such explosions, it is common sense that there must be a stage just before such an eruption where the slightest "prod" to the hair-trigger magmatic system below the volcano will be enough to initiate an explosion. Of course it's one thing to speculate on such a matter but quite another to prove it. There was a series of big explosions at the Montserrat volcano between early December 2009 and 11 February 2010 but none of these took place immediately after 12 January. Therefore I think that any connection between the earthquake and specific explosions at the volcano has to remain "non proven".


Update, late-November 2010.  A new scientific report on this topic has recently appeared in the important journal called just "Geology" (Earthquake-induced thermal anomalies at active volcanoes by Dario Delle Donne1, Andrew J.L. Harris, Maurizio Ripepe and Robert Wright). This group used satellite measurements to get evidence suggesting that large earthquakes in some areas may indeed trigger volcanic eruptions. Clearly there is much more research to do in this important subject.



Montserrat Volcano Observatory