This blog provides a commentary on landslide events occurring worldwide, including the landslides themselves, latest research, and conferences and meetings. The blog is written on a personal basis by Dave Petley, who is the Wilson Professor of Hazard and Risk in the Department of Geography at Durham University in the United Kingdom.

This blog is a personal project that does not seek to represent Durham University.

Thursday, 30 April 2009

British Geological Survey landslide web page

The British Geological Survey (BGS) appear to have been quite active of late in overhauling their web pages, reflecting what seems to be an improved ethos of delivery of information to end users. The landslide pages seem to have been subjected to this upgrade of late, which has led to the creation of a really good set of case study pages, backed up with photos, reports and maps. These pages are available below - they are a little slow on my system, but that might be my wireless router:

http://www.bgs.ac.uk/science/landUseAndDevelopment/landslides/caseStudies.html

A good example of this resource is the 1993 Holbeck Hall landslide (see BGS page here), which includes this image:


Perhaps most excitingly, parts of the UK national landslide database are going to be made available online for free through the new Geoindex tool. A Beta version of this tool is already available - it looks to be a really exciting tool.

Great stuff!

Friday, 24 April 2009

EGU presentation on landslide fatalities in 2008

I thought that it would be helpful to make available my Powerpoint file for the presentation that I am giving on the occurrence of landslide fatalities in 2008. This presentation details the fatalities that I recorded on the database over the course of the year, providing maps and tables of some of the data. I hope that you will find it useful.

The file should be visible below:


Uploaded on authorSTREAM by Dr_Dave

And you should be able to download the file from the link above or from the following site:

http://www.authorstream.com/User-Presentations/Dr_Dave/

Wednesday, 22 April 2009

EGU Day 3

I am just going to write up one session from today's meeting as it contained the most interesting talks that I heard. This was an annual session on landslides induced by volcanoes and earthquakes, the first three talks of which focussed on the Wenchuan event. Regular readers will know that I am very interested indeed in this event (see here for example).

So first up was Ed Harp and two colleagues from USGS. They provided a pretty general overview of the landslides triggered by the earthquake, but supplemented with some very nice satellite imagery of key sites. He highlighted the death toll associated with the earthquake-induced landslides (20,000+) and the huge number of dams that needed clearing (he quoted 33 that required mitigation). The tour included Beichuan, Tangjiashan, etc. I guess there wasn't much new or scientifically-challenging here, but it was a good start. The final part of the talk highlighted the collaboration between USGS and the China Geological Survey, which is going to allow transfer of techniques for seismic landslide hazard analysis, data collation to test the USGS PAGER model and quantification of the sediment flux. In questioning Ed said that the planned seismic hazard analysis tool is Newmark Displacement. I do wonder whether this is the right tool in this part of China - it is probably appropriate for the initiation of the slides, but most of the failures that I saw have lower sections that are very complex, with massive entrainment of slope debris and colluvium. This is where the people and infrastructure are, so it seems to me that without substantial modification Newmark is going to be quite problematic.

Second up was Gorum and his colleagues from ITC. The poor chap had the misfortune of giving his first international conference talk to a packed house about a set of landslides that he had not visited! In that context he did very well indeed. He gave a slightly broader overview, making use of some of the data collected by the Chengdu University of Technology from their mapping. In particular, he highlighted that the landslide are focused close to the fault trace (NB this focus is not on the epicentral region), with many slides on the lower gorges. Using satellite imagery they have mapped an initial 11,308 landslides, compared with 1,638 before the earthquake. They also noted that they had mapped 256 valley blocking slides, with the highest concentration being on the fault trace. They are now working with the Chinese to understand the landslide distribution (very challenging) and to undertake a multi-hazard analysis. I cannot quite see how the latter will be done - the presenter seemed to imply that they will use the existing landslide distribution to drive a modelling exercise. I hope that this is not the case as the seismically-induced landslides will not give a good indication as to where rainfall-triggered slides will occur in a post-earthquake landscape.

The final talk from Wenchuan was by Chigira and his colleagues from Kyoto, with substantial co-authorship from China. This was the best judges of the three, provising a nice summary of the key points issues, well-illustrated using good images. They highlighted the role of dissolution in raising landslide susceptibility - the point being that dissolving limestone beds creates voids that allows drainage of groundwater, reducing susceptibility to rainfall induced slides, but creating point-to-point contacts that increases susceptibility to earthquake induced sliding. This is a nice point. He concluded by looking at some of the very largest slides, concluding that the geomorphology before the earthquake showed depressions and dips on the big slides that indicated that they were potentially unstable. Thus, the biggest slides were considered to be predictable. I am not sure that I agree with the latter point completely (unless all slopes with these features failed, which I don't think is the case), the observations about the morphology are well-made.

The penultimate talk upon which I will comment was by Niels Hovious from Cambridge, with co-authors from Taiwan and elsewhere. Niels used the 1999 Chi-Chi earthquake in Taiwan to examine the distribution of landslides that are generated, and then to look at the production of sediment. First, Niels argues that the landslides closely map onto the distribution of ground shaking, with which I agree, but then argued that the highest landslide density occurs around the epicentre. This may well be true for Chi-Chi, but it was not for one of jis other examples (Northridge) and it was not true for Pakistan, where the highest densities are at the fault rupture. Most importantly, it is also not true for Wenchuan, where again the highest densities lie close to the surface expression of the fault rupture and not around the epicentre. Niels then showed that the density of landslides increased remarkably in the aftermath of the Chi-Chi earthquake - in the Chenyoulan ctachment that they studied the number of landslides before the earthquake was 8123, with a further 3,800 being triggered in the event. However, in the nearly ten years since a further 48,370 landslides have been triggered. However, Niels showed that the sediment concentration in the rivers is now close to base level again, suggesting that the earthquake's impacts are now reducing. This is good stuff but, given that the landscape is affected by typhoons that are exceptionally extreme events one wonders how applicable it is to other places. It seems to me that they need to work in China! An important aside is that this talk does highlight the importance of being prepared for massive sediment production in China.

The final talk that I shall briefly mention is that of Merri and his colleagues from Italy, who are using a finite difference model (FLAC3D to simulate the impacts of magma intrusion on the stability of the Stromboli volcanic edifice. The presentation was quite nice, but the model seems a little flawed. First, it assumes that the volcano is geologically homogenous - volcanoes certainly are not, and given that deposits are layed down in slope parallel layers, this heterogeneity can be a big factor in slope instability. Second, the model appears to ignore pore pressure affects (I asked whether pore pressures are being modelled - the presenter ducked the issue by talking about over-stress. Given that injections of hot material cause increases in pore pressure for certain, and these may well be very important in understanding slope stability - this is a substantial omission.

Comments welcome as ever, especially from the presenters and other attendees. Do feel free to comment if you disagree with what I have written. Finally, apologies for typos, spelling mistakes, etc. The spell check function isn't working and I don't have time to check.

Tuesday, 21 April 2009

EGU Day 2

The landslide elements of Day 2 at EGU were split between two sessions on landslide forecasting and two on landslide risk. The latter is of comparatively little interest to me, and I had a load of work to do and meetings to attend, so I only attended the morning sessions.

In terms of landslide forecasting, there was some pretty good stuff presented. I would say that some speakers need to think a little more about their audience - using endless meteorological acronyms might work well if you are talking to weather specialists, but when the audience is mostly composed of landslide geologists it is a surefire way to lose the focus of your audience.

The first talk that caught my eye was that of Brunetti and colleagues from Italy, who used datasets culled from the literature to look at the statistical properties of landslide volumes. This sounds pretty terminal, but actually it is interesting as the distribution of landslide sizes in any given area follows a very specific relationship - a so-called power law. The paper examined the caharctersitics of this power law for a range of landslides, finding that there were consistent patterns that appeared to be determined by the mechanism of failure - i.e. soil slides formed a group, irrespective of location, that was distinctly different from the group associated with rockfalls. This felt like a substantial step forward - power lay relationships have been around for a while but we have struggled to understand what this tells us. Detailed studies like this will help greatly.

The next talk was a slightly odd one, by Peter Lehmann and his co-author on self-organised criticality. This relates to the power law issue above, but here the starting point appeared to be that avalanches in a sand pile also show power law behaviour and that this is associated with self-organised criticality. Ergo, landslides occur because of self-organised criticality, which means that the characteristics associated with SOC can be used to look at precursors to slope failure. This step may be problematic because the SOC displayed by sand piles is associated with frictional systems, whereas landslides are generally cohesive. Therefore I remain to be convinced that tje same precursors will occur in SOC for natural slopes. Clearly there is more work to do here, so I will watch with interest. This is one of those things that could be brilliant or it could be very esoteric.

Matthias Jakob went next, talking about the design of a debris flow warning system for N. Vancouver in Canada. His starting point was that given the risk of debris flows some sort of warning system is needed, but the cost of a full blown deterministic system is too high given the area covered. So, using 30 years of very high quality data, they had looked at understanding the relationship between debris flows and rainfall. Interestingly, they have rejected the standard intensity - duration relationship, instead undertaking a discriminant function analysis on the data to find that the three key factors are:
  • long term antecedent rainfall (fills up the groundwater stores)
  • medium term antecedent rainfall (tops up groundwater)
  • short term rainfall intensity (triggers failure)
The upshot was an equation that allows warnings to be issued. Three alert levels will be used (no debris flows, debris flow watch, debris flow warning). The authors thought that on average five warnings will be issued each year, of which two on average will generate debris flows. The system has worked well in trials this year, but it will be interesting to see how the community reacts to so many warnings.

A rather peculiar presentation was given by Thiebes and his colleagues from the Department of Geography at Vienna. The talk was well delivered and the topic was both interesting and scientifically valid. So what was odd? Well, the team are part of the ILEWS consortium that is trying to develop early warning systems for landslides. To do so they have instrumented a landslide in the Swabian Alb area. The aim is to use the instruments to drive an online data collection and analysis tool that incorporates end-user driven modelling via the CHASM code. This is great - and I fully support such initiatives. The odd thing is that the landslide that they are instrumenting has moved 1 cm in the last 2.5 years - this hardly sounds like a slide that needs an early warning system! This is a shame as there are so many slides around that do need such a system.

Serval Miller from Chester University presented a very detailed analysis of a landslide susceptibility mapping exercise that he had undertaken in Jamaica. Interestingly, they had tested a range of GIS based techniques, concluding that a Bayesian Model provided the best results. Second best was a straight forward layer combination model. This was quite interesting, but with this and other presentations on landslide susceptibility analysis I do end up wondering whether the time spent would be better used to provide a geomorphological map that indicated where landslides are considered likely. I wonder whether this would really be any less accurate?

Finally, I would like to note the work of Devoli and her colleagues, who have been trying, with some success, to implement a landslide risk reduction programme in Nicaragua in the aftermath of Hurricane Mitch a decade ago. Although the in-country team is small (three geologists), a huge amount appears to have been achieved through SINAPRED, the national emergency commission. This presentation highlighted two web resources that are well worth a look:

Georiesgos-ca.info, which provides georisk information across all of Central America (in Spanish)
http://mapserver.ineter.gob.ni/website/mapas/Estudios/viewer.htm, which is a mapping server that provides access to hazard maps for Nicaragua.

Monday, 20 April 2009

Updated: European Geosciences Union Day 1

Updated to include the afternoon sessions
This week is the annual European Geosciences Union assembly in Vienna. This is the biggest annual landslide meeting - there are >300 landslide related papers this time around - and since I am the scientific secretary for the landslide session I cannot allow the opportunity to comment on what I see to pass. So here are my thoughts on Day 1. My intention is not to comment on everything that I see, but instead on those that I found interesting.

Morning sessions
The first, unfortunately poorly attended, session focused on landslides associated with loess deposits. Dr Meng and colleagues from China presented a very well illustrated overview of landslides on the loess plateau. There were some remarkable statistics - for example, in the 20th Century over 60,000 people were killed by loess landslides in China, whilst in Gansu province alone in the decade between 1875and 1985 there were over 1000 disastrous landslides, killing in total >2000 people! Interestingly, the key factor determining the spatial occurrence of landslides was the neotectonic activity - areas of active uplift have far more slides than those that are subsiding, presumably because of undercutting and oversteepening during incision. The presence of sinkholes was also shown to be rather important.

Mamyrova and her colleagues presented a paper on the investigation of the mechanics of two loess landslides in Kyrgyzstan, a place about which we here far too little from a landslide perspective. She presented some rather nice repeat pass Quickbird imagery to show the evolution of the landslides, demonstrating in particular that one of the slopes showed clear tension crack development before the main failure event. Perhaps most interestingly, the initiation of the main failure event occurred in a wet, but not exceptionally wet, year (1994). A month or so earlier there was a magnitude 3.9 earthquake just 17 km from the landslide. The authors speculated as to whether this might have played a role. This project is clearly in its early stages, but the combination of an under-reported area, a very strong research team and an aspiration to combine FLAC (finite difference) and PFC (discrete particle) modelling makes it one to watch for sure.

Later in the morning, in a far better attended session (I ended up sitting on the floor!), there was a very nice presentation by Dewez and his colleagues frim BRGM in France on the use of terrestrial laser scanning to look at coastal cliff hazards in chalk. In many ways this mirrored the work that my colleagues and I do on cliffs in NE. England and, unsurprisingly, the results were similar. Over about 1 km of cliff, and scanning from eight stations, the survey picked up 8567 individual rockfall events. The most interesting thing for me was the investigation that they had done of the role of large vs small events in erosion on the cliffs, showing that over this time a single large event contributed 85% of the volume change, and that 99.8% of the volume change comes from rockfalls that are >1 cubic metre. This is quite a different result from our observations in layered sedimentary rocks in N. Yorkshire, emphasising the role that geology plays. It should be noted though that the comparatively short duration of this study (ours is >twice as long) may mean that the results are biased by the large events that happened to occur. The authors recognised that there is a need to extend the time-base - I do really hope that this is possible as the development of long datasets is crucial if we are to understand rockfalls properly.

Travelletti and his colleagues, also from France, presented a study of the use of terrestrial laser scanning for the monitoring of a large mudslide - in this case the Super Sauze slide. In my view the application of TLS for monitoring this type of slide is too rarely presented. In this case they have used the system to look at the evolution of the steep source area and to examine the movement of the toe of the flow. Both were pretty convincing, but I was particularly impressed with the work that they had done tracing the movement of boulders to generate displacement fields. In some cases they saw 16 m of movement in a single summer, which is far greater than the c.0.25 m error in the technique at this location. This is looking like a mature technique now that really justifies more extensive use.

Afternoon
The afternoon landslide sessions focused upon geophysical techniques, many of which are a little beyond my area of expertise. I will comment on a couple of presentations though. A general point to note is that, as the chair of the session pointed out, geophysical approaches to landslide analysis have moved from being rather esoteric and obscure to generating quite a lot of interest in a very short period. This is primarily because these techniques are increasingly good at discriminating between different types of wet sediment and debris, which means that they are useful. However, I do note that the presentations that I heard were on the whole about the geophysics, and not about the landslide, so these methods are not really in the mainstream yet.

There were two presentations on the Trieve landslide complex in France, the first by Kneiss and the second by Renalier. The former focused upon the use of seismic noise to map out the subsurface geology of the slope. In particular, they used a technique that they called the H/V method to delineate the soft sediment – bedrock boundary. The story was pretty nice in that the 3D topography that they showed explains the dynamics of the movement, although it does seem to be based on a slightly simplistic view of the mechanics of the slides. Personally I would like to see some verification of the 3d model (perhaps a couple of extra boreholes to see if they can predict where the drill will meet rockhead), but overall it is a pretty nice piece of work. The second was by Florence Renalier, who was a little nervous I suspect, but did a good job (well done if she reads this). The intriguing result here was that they showed that shear wave velocity is inversely correlated with displacement rate for the landslide that they studied (in the same complex as the previous paper). That seems to me to be a pretty fundamental result, but it was not really dwelt upon. In questions the team said that they think that the shear wave velocity is related to microcrack density. I would like to hear a lot more about this to be honest.

The next paper that caught my eye was by Clara Levy and her colleagues, also from Grenoble (are they trying to corner the market in research level landslide geophysics?). This was a study of the seismic precursors to a rockfall that occurred from chalk cliffs in 2002. By luck or good design (or both) the team had a seismic network in place before the 2000 cubic metre failure occurred. It was collecting data at 30 kHz, which must be a mind boggling dataset, but what they captured was 200 seismic events in the 2 hours before the collapse. They attribute this to breaking of the remaining rock bridges (which seems reasonable). They have tried to model this effect, but have chosen to do so with a Mohr-Coulomb failure model, which is perhaps not ideal. Nonetheless, the model does show how the failure propagates through the rock bridges, with those at the toe of the slope being the last to go. The model is I think not really faithful to reality, but it is an interesting piece of work.

Finally, Arnhardt and his colleagues from Aachen decided to break up the French – Italian axis by presenting a paper looking at the use of low cost sensor networks for monitoring landslides. This is highly worthy work, but I seem to have heard numerous similar presentations, all of which come to nothing. The problem is that the low cost sensors in question are usually designed for use in nice, controlled, clean, stable environments (like in cars, high tech production lines and hospitals). Landslides are the exact opposite – wet, cold (or hot), dirty and mobile, which means that the sensors that work so well in the lab really don’t function well. This is the same issue that Pen Hadow and his friends are having in their misjudged nightmare in the high latitudes. I guess someone will crack this problem in the end – and the team from Aachen seem sufficiently serious to be the ones to do it – but I do hope that they realise just how hard it will be. I really do wish them luck and will watch their progress with interest.

Overall it has been an excellent day (thanks to all the speakers), with some great science presented. Tomorrow we have two sessions on landslide forecasting and two on landslide risk. I can’t wait!

Friday, 17 April 2009

The rising cost of landslides in the Three Gorges dam area


One of the great concerns about the Three Gorges Dam in China has always been the potential for large-scale landslides. As the project nears completion, inevitably very over-budget, the level of the lake is rising and the cost of these landslides in becoming apparent. Today AFP and Xinhua have runs stories that raise concerns still further. These stories report that since September there have been 166 landslide and debris flow events on the banks of the reservoir, forcing the relocation of 28,600 people. The economic losses associated with these slides are approximately US$79 million.

Unfortunately these events have occurred before the onset of this year's rainy season and before the planned further increase in the water level later this year.

Large landslides in Peru and Kyrgyzstan, the Afghanistan earthquake plus heavy rain expected in the Wenchuan area

Each year in mid-April we move into the global "landslide season", when the development of the Northern Hemisphere summer, and the associated weather patterns elsewhere, means that the number of landslides starts to increase dramatically. This is all too clear from the range of landslide events in the last few days, plus the threat of heavy rainfall in the earthquake affected areas of China:

1. Major landslide in Peru
A range of news agencies (for example AFP and CRI) are reporting that there was another large landslide in Peru, again in La Libertad Province (the second major landslide this week in that province). Although details are sketchy, this time the landslide appears to have been very large (one report suggests 1 km long), hitting two villages (Chamanacucho and Aricapampa). Reports suggest that about 30 people were killed. The coordinates of Aricapampa are (-7.80583, -77.7172), which yields the following Google Earth images:


It appears that the landslide is still active, which is hampering the recovery operation substantially.

2. Major landslide in Kyrgyzstan
According to RIAN there was also a large landslide in Kyrgyzstan yesterday. The landslide appears to have hit Raikomol village Jalalabad province in in south Kyrgyzstan, killing 16 people and a large number of cattle. All of the victims, 11 of whom are apparently children, have been recovered. The ENG24 website has posted this rather grainy, but very helpful, image of the slide:

The reports suggest that it is about 300 m long. It appears to be a massive earthflow. Unfortunately the source zone is not in the image - I would be very interested to see how and where this started.

3. The Afghanistan earthquakes
The two moderately-sized (USGS Mw=5.5 and 5.1) but shallow (USGS depth = 5.7 and 3,.2 km) earthquakes in Afghanistan this morning appear to have caused damage in at least some villages, with about 20 reported fatalities at the moment. A Google Earth image of the area affected suggests that it really is a very remote zone:

Given the remoteness of the area and the rugged terrain the number of reported fatalities might well rise during the day. Earthquakes of this size would not normally cause much damage, but the early indications are that these two events really are exceptionally shallow. I would anticipate that there will have been at least some landslides in the upland areas, but probably over quite a limited area.

4. Heavy rain forecast for the earthquake affected areas of China
Xinhua is forecasting that heavy rain will hit the areas affected by the Wenchuan earthquake over the next few days. Up to 100 mm is expected to fall. This will be the first heavy rainfall of this years rainy season. Given the amount of mobile sediment on the hillsides, and the occurrence of debris flows in heavy rainfall last September, some further problems might be expected if this heavy rainfall does occur:

Tuesday, 14 April 2009

Updated: Large landslide in Peru

Various Peruvian news agencies, such as Adonde.com, are carrying reports of a large rainfall-triggered landslide on Saturday night / Sunday morning in Peru. This slide, which appears to have occurred in the town of Retamas in Parcoy District of Pataz Province, has reportedly buried a number of houses, killing 13 people. The following perspective Google Earth image of the location suggests that landslides may well be a substantial problem in this area (the location is -8.0197, -77.4783 if you want to take a look for yourself):


Helpfully, the following image of Retamas is available on Panaramio at this page:


I guess it is not hard to understand how a landslide can kill 13 people in this landscape. As an aside, given that this is an area of high seismic hazard, those ridge line dwellings look very poorly located.

Update: the Latin American Herald Tribune has an article with further information here. The aricle also has an image:


The slide appears to have occurred on a steep slope above the town. The article notes that the town is “is situated on the side of a mountain and whose residents mine gold and silver...The informal mining operations in the town have affected the stability of the soil in the area", which explains the large scarps on the Google Earth image.

A round-up of some recent landslide events

I thought that it was time to provide a round-up of some landslide events from around the world:

1. A narrow escape in Australia
Thanks to Remke van Dam for bringing this one to my attention. This Australian family had a pretty lucky escape last week when a pair of large rocks struck their car (image from here):


2. A road-blocking rockfall in Alaska
Thanks to John Fritz for this one. In Alaska, a pretty large rockfall has blocked the Whittier tunnel (see news report here):


The landslide occurred on Saturday; the road is expected to be blocked until Wednesday.

3. A fatal landslide in Idaho
In contrast to the first item above, a very unlucky couple were killed by a landslide in Idaho last week. According to this report, the couple's 4x4 (SUV) was hit by a tree that had toppled onto the road as a result of a small landslide near to Orofino. As the image below (from here) shows, the tree hit the cab, killing the occupants:

Monday, 6 April 2009

Cave rockfall in Shropshire, England

The British newspapers are reporting that overnight a rockfall occurred in Hermitage Cave near to Bridgnorth in Shropshire, England. This is news because at the time there was a group of teenagers camping in the cave. Sadly, one of them was killed and another was badly injured. These caves are old dwellings cut into the Bridgnorth Sandstone (see image below from here):

The caves used to be much larger but rockfalls over time have progressively reduced their extent. There are reports that the ill-fated group lit a fire in the cave. I must stress that this is unconfirmed. Sometimes however falls like this are triggered by campfires lit in the cave, which induce thermal stresses in the ceiling that trigger the collapse. I am not sure if this is the case here - it could well be that the group were just very unlucky indeed - but fires in caves with low ceilings can be a quite significant hazard. Sadly, it is very unlikely that the group would have known about that particular danger.

A first take on the Italy earthquake

You will probably be aware that an apparently quite destructive earthquake struck central Italy overnight. I thought I'd try to give a first take on the likely impact of this event. First, as ever, there is some pretty good information about the earthquake on the USGS Earthquake program website. They have provided the following helpful maps:

Earthquake location:
Shaking intensity:

Exposed population (PAGER):


At this point (c. 8:00 UT) the USGS is estimating that there are about 68,000 people living in areas that have suffered an earthquake intensity of VIII (severe shaking) or above, mostly in the town of L'Aquila. This is an area with a mix of old and new buildings built in a hilly area, as this image (from here) shows:

Although the earthquake is not huge (USGS estimates are Mw=6.3), the shallow depth (10 km) and fairly vulnerable buildings means that the impact could be quite substantial, albeit in a fairly limited area. Italy is well-prepared for earthquake response, which will help.

So, what of landslides? Well, an earthquake of this size should be capable of triggering a fair number of slides. A good starting point is the Keefer (1984) relationship between earthquake magnitude and area affected by landslides:

This gives an area affected by landslides as about 2000 km2 (give or take quite a lot, though). The area is certainly landslide prone, as this Google Earth perspective view shows (I have marked the epicentre location as per the USGS. The town in the foreground is L'Aquila):

Incidentally, the mountain in the background is Gran Sasso, which houses an important particle physics laboratory in a deep tunnel. I would be interested to know how the experiments have fared during the earthquake.

Sunday, 5 April 2009

The Casita landslide revisited

One of the most deadly hurricanes of modern times was Hurricane Mitch, which tracked across Central America in late October 1998. Many of the tens of thousands of victims were killed by landslides. Perhaps the most notable event was a lahar (a volcanic landslide) that swept down from near the summit of Casita volcano in Nicaragua, killing about 2500 people over the course of its 6 km path (and some more in the hyper-concentrated flow (debris rich flood) events that travelled a further 10 or so kilometres from the toe of the slide. Unfortunately, despite the magnitude of this event the amount of published literature about it has remained quite limited. It is therefore terrific that a paper has just been published by Graziola Devoli and her colleagues (Devoli et al. 2009) that seeks to summarise the published and unpublished reports about this remarkable landslide.

Wikipedia
has a very decent image of the upper track of the landslide, which gives a pretty good idea of the scale of this event:

Whilst MDA have a great overview image of the source, track and runout zone:

Complex landslides such as this are poorly understood. In particular, as in the landslides that I highlighted in Sichuan, the mechanisms of initiation and movement are quite intricate. Devoli et al. (2009) have used a range of geological, geotechnical and analytical techniques to get a better idea of what happened.

The landslide was triggered by very heavy rainfall - they suggest that about 750 mm (that's about a years worth for where I live) of rain fell in a little over 80 hours. Interestingly, they conclude that the landslide can be divided into three key phases:
  1. Failure started in a fractured and altered volcanic breccia in the northern area of the scarp which released a volume of about 260,000 cubic metres. The flow that developed from this failure swept downslope and entrained colluvium deposits at the toe of the slope in the southern part in less than about 40 seconds.
  2. The rapid removal of the colluvium on the slope triggered a second failure. This also originated in the scarp shown on the image above. In this phase about 640,000 cubic metres of volcanic breccia slipped over a unit of clay-rich pyroclastic deposits. It is unclear as to whether this flow joined the first one or occurred separately. Either way, blocks in this flow travelled 9 km or more downslope.
  3. The third and final stage consisted of a as a sudden debris / rock avalanche that originated in the uppermost section of what is how the landslide scar. This failure, with a volume of 690,000 cubic metres, appears to have occurred very soon after the first two events.
In the aftermath there has been considerable concern raised about the likelihood of failure of the entire flank of Casita Volcano (indeed, even I have published on this theme - see van Wyck de Vries et al. 2000). The paper concludes that under conditions of high groundwater instability of these flanks can occur, although as we understand these very large failures so poorly I would be cautious in the interpretation of this particular result. This is a slope that needs an active monitoring programme for sure.

References
Devoli, G., Cepeda, J. and Kerle, N. 2009. The 1998 Casita volcano flank failure revisited — New insights into geological setting and failure mechanisms. Engineering Geology, 105, 65-83.

van Wyk de Vries, B., Kerle, N., Petley, D., 2000. A sector collapse forming at Casita
volcano, Nicaragua. Geology 28, 167–170.