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.

Friday, 30 January 2009

Name the Landslide number 1

OK, a bit of fun for a Friday afternoon. The challenge is to provide via the comments section for this page, the name of the famous landslide shown on the Google Earth image below. The winner must also provide a two sentence (maximum) description of the big movement event:

Whoever gets it right gets to choose the next landslide for us to try to name. I have turned message moderating off to allow this to work. This one is not easy, but there are clues in the picture.

Some interesting landslide sites

I thought that it was high time to post a summary of some interesting landslide sites:

Aowanda landslide
There is an interesting youtube video of a smallish progressive rockslide occurring at Aowanda in Taiwan here. It should also be visible below:

Not quite as dramatic as some others, but useful nonetheless.

Five disastrous landslides that changed the Canadian Landscape
There is quite a nice non-scientific piece here that describes in non-technical language five large and significant historic landslides in Canada over the last century or so. Good images!

Rockfall blog
There is an interesting new blog here that looks at measures that can be used to prevent rockfalls. Although it is in Italian, Google Language Tools do a pretty good job of translating the text. I guess one should be a little careful of the fact that it is promoting a company, but there is some nice stuff there.

The largest known landslide?
National Geographic has a nice story here that speculates that a large deposit on Mars might be the debris from a landslide that would have been the size of the United States. Working on Mars must be great because no-one can check your interpretation with field data.

Flashflood footage
It is a fine line between a debris-rich flash flood and a debris flow, so I thought that I would highlight this youtube video of a flash flood in Australia. The first minute isn't terribly exciting, but stick with it! It should be visible below:

Extraordinary video of the Baldwin Hills dam disaster
There is an excellent video of the collapse of the Baldwin Hills dam in 1963 on youtube here. This is a great illustration of the dangers associated with the formation and collapse of landslide dams. Again, it should be visible below:

Thursday, 29 January 2009

Was the Wenchuan (Sichuan) earthquake triggered by humans?

At recent edition of Science carries a really intriguing review piece that examines the triggers of the Wenchuan earthquake. Although earthquake triggering is really beyond the scope of this blog, the huge number of devastating landslides that the earthquake induced means that this is of interest. The article can be read online here.

The article has emerged from work undertaken at two different locations. Christian Klose of Columbia University presented a paper (abstract here) at the AGU Fall meeting in December in which he suggested that the earthquake could have been induced by "local mass imbalances" associated with the accumulation of "at least 320 million tonnes of water" in the upper Min Valley - a clear reference to the construction of the Zipingpu Dam, which is located close to the fault, and which was completed in December 2004. The Google Earth image below shows the dam and lake, although the quality is not great:

The idea is that the additional weight of the water acted in a manner that is very similar to the triggering of landslides on slopes - i.e. it acted to both increase shear stress and to decrease the normal effective stress, rendering the fault more likely to fail (slide). This is apparently backed up by a paper by Lei Xinglin, of the China Earthquake Administration in Beijing and GSK/AIST, which was published in the Chinese journal "Geology and Seismology" last month. That paper is also available online here, but it is in Chinese, so I cannot really tell what it is saying in detail. According to the Science article, the paper apparently also suggests that the reservoir may have been a factor, without coming to a firm conclusion. The diagrams in the paper are worth a look, and have English captions. The abstract is as follows:

So how likely is this? To be honest my gut reaction is not very likely at all, but this certainly needs further work. Klose in his abstract says that:
"Shear stresses increased by >1kPa on the Beichuan fault at the nucleation point in about 20km depth. Normal stresses decreased by <-4kPa and weakened the fault strength. Pore pressure increases might have additionally destabilized the fault locally due to pore pressure diffusion. This effect, however, might be minor in 20km depth, because of low lateral fracture connectivity and permeability between the area of water accumulation and the Beichuan fault."

These stress changes seem to be rather small to have a significant effect. To put this into context, average atmospheric pressure at sea level is 101.325 kPa, so the changes that he is describing are a fraction of this. It is hard to believe that this has had a significant effect at 20 km depth, although the argument seems to be that the fault was so close to rupture that this tipped it over the edge, assisted by pore pressure diffusion. The depth of the earthquake nucleation point also renders this a little dubious, but Klose argues that in fact most of the energy release occurred close to the surface. In addition the fault appears to be rather further from the dam than the article suggests (the Science article says 500 m, but this is surely not the case (I think it is about 30 km to the epicentre for example). Note also that this reservoir is really not very large (c. 8 km long)

It will be very interesting to see how this plays out over the next few months. I remain very doubtful indeed but am certainly open to being persuaded if the science stacks up properly. Of course, if it is the case that the dam was a factor then this would have profound implications legally, socially and in terms of future development, especially given that HEP is coming back into fashion as non-renewable sources of energy decline.

Sunday, 25 January 2009

Two interesting recent landslides

Two interesting landslides to report in the last few days.

1. Landslide in Iztapalapa, Mexico City
On Thursday morning a 50 cubic metre landslide in the Itzapalapa, a poor suburb of Mexico City slid onto a house at the toe of the slope, killing two people. Chinagate has published a rather dramatic picture of the site:

The interesting thing about this image is the very large (apparently 5 m high) retaining wall that had been built between the slope and the houses. Clearly the wall has failed under the weight of the landslide (not surprising given the size of those boulders actually), but I wonder who built the wall (surely not the people living in these houses - it is far too big)?

2. Landslide-induced wave on Lake Roosevelt, Washington State, USA
The second is a report from the Spokane arm of Lake Roosevelt near to Spokane in Washington State. The website Spokesman Review is carrying a report that says:

"Property owners in the Spokane arm of Lake Roosevelt were swamped by a huge wave caused by a landslide last Friday. The 17-acre landslide on Jan. 16 near Mill Canyon – northeast of Davenport and downstream from Little Falls Dam – slid into the water, unleashing a wave that reached 30 feet above the full-pool mark across the lake into Breezy Bay, according to Adam Kelsey, chief ranger for the Lake Roosevelt National Recreation Area. Private docks and vessels were destroyed or damaged for at least 1.5 miles downstream from the site."

They also have an image of the site:

It is reasonable to assume that the displaced mass on the far side of the lake is the landslide as the Google Earth image does not show a slide at this site:

It is interesting to note that the section of slope that has failed shows clear signs of toe erosion and some interesting linear features on the Google Earth imagery:

Interestingly, in the years after Lake Roosevelt was filled there were a series of landslides. The Emergency Management Division of Washington Military Department has a document here that describes a series of events as the lake was filled and emptied:

"1944 to 1953 – Massive landslides generated a number of inland tsunamis in Lake Roosevelt in Eastern Washington:
• April 8, 1944 - A four to five million cubic yard landslide from Reed Terrace generated a 30-foot wave, 5,000 feet away on the opposite shore of the lake about 98 miles above Grand Coulee Dam.
• July 27, 1949 - A two to three million cubic yard landslide near the mouth of Hawk Creek created a 65-foot wave that crossed the lake about 35 miles above Grand Coulee Dam; people 20 miles away observed the wave.
• February 23, 1951 – A 100,000 to 200,000 cubic yard landslide just north of Kettle Falls created a wave that picked up logs at the Harter Lumber Company Mill and flung them through the mill 10 feet above lake level.
• April 10 – 13, 1952 – A 15 million cubic yard landslide three miles below the Kettle Falls Bridge created a 65-foot wave that struck the opposite shore of the lake. People observed some waves six miles up the lake.
• October 13, 1952 – A landslide 98 miles upstream of Grand Coulee Dam created a wave that broke tugboats and barges loose from their moorings at the Lafferty Transportation Company six miles away. It also swept logs and other debris over a large area above lake level.
• February 1953 – A series of landslides about 100 miles upstream from Grand Coulee Dam generated a number of waves that crossed the lake and hit the opposite shore 16 feet above lake level. On average, observed waves crossed
the 5,000-foot wide lake in about 90 seconds.
• April – August 1953 – Landslides originating in Reed Terrace caused waves in the lake at least 11 different times. The largest wave to hit the opposite shore was 65 feet high and observed six miles away. Velocity of one of the series of waves was about 45 miles per hour.

Thursday, 22 January 2009

Whitby - the old and the new

As someone interested in earth processes, one of the great powers of the internet is the ability to access quickly images of places in which I am interested. Over the last few years my colleagues and I have been studying rockfall processes at the town of Whitby in N. Yorkshire, which is in NE. England. I thought as a bit of fun I would compare two images of the cliffs in the town. The first, assessed from here, was taken in the 1900's - i.e. a century ago:

The second, accessed from here, was taken in the last few years:

There are several things of interest here. It is important to note that the two images were not taken from quite the same angle -the 2000's image was taken further from the sea than was the 1900's one, though there is not much difference. However, there is quite a lot of interest here. First, note the very different beaches between the two images - in the 1900's image there is very little beach visible, and indeed boats are moored by the small pier. Now the beach has encroached into this area, presumably through natural accretion though I guess it could have been artificially stocked.

More interesting, the profile of the cliff has changed dramatically. The cliff a century ago was far less steep, with a large, protruding toe. In the recent image most of this toe has been eroded away, leaving a far steeper profile. Interestingly though the profile of the top of the slope is little changed, except being a little steeper. Presumably as the toe continues to erode, and thus the slope to steepen, the rate of erosion at the top will increase, or a large failure will occur that will cause the cliff top to erode back a few metres. Fortunately, it is a long way to the abbey and church, and there is little else at the cliff top to worry about unduly.

Tuesday, 20 January 2009

Earthquakes, landslide disasters and the collapse of civilisations

There is a very interesting paper being published next week in the Proceedings of the National Academy of Sciences by Dan Sandweiss, Ruth Shady, Mike Moseley, Dave Keefer and Charles Ortloff. This paper, which is reported in a long press release from the University of Florida, combines archaeology, geology and geomorphology to look at the collapse of the Supe civilisation in Peru, which survived for a couple of millennia in the coastal areas of North Peru before it appears to have abruptly collapsed and disappeared about 2000 years ago. The civilisation is intriguing because they built large and complex pyramids, some over 30 m high, as shown in the Google Earth image below. To be able to build such structures the society must have been complex and organised, so why did it disappear so abruptly?

The paper suggests that about 3,600 years ago the area was struck by a massive earthquake, an idea that is not unreasonable given that we know that this a highly seismically active area. The earthquake caused widespread destruction to the towns, but given the length of time that the civilisation survived in a seismically-active area, this would not have been the first time that this happened in all probability. The key factor was the triggering by the earthquake of massive landslides on the adjacent hills, which moved large amounts of debris into the valleys (as happened in the Wenchuan earthquake last year). This debris was then mobilised by floods triggered by heavy rainfall (again, as happened recently in Sichuan), probably associated with strong El Nino events, which bring heavy rain to this part of the world. This mobile debris would have been a hazard in itself, but more importantly it washed out to sea, whereupon it was redeposited on the coast to form a feature called the Medio Mundo, which is a large coastal bar (see below - click for a better image in a new window):

The Medio Mundo sealed off coastal bays, and thus eliminated a major source of food. Furthermore, sand from the ridge blew inland on the prevailing winds, swamping farms and communities. The huge resultant sand deposits, with the wind sculpted features, can be seen on the image above. Within a few generations the civilisation collapsed, never to recover in that form.

All-in-all a very interesting story that shows how combinations of events can lead to catastrophes. There are lessons to be learnt by modern societies from this!

Sunday, 18 January 2009

Landslide in Lombok, Indonesia

The Jakarta Globe is today reporting a quite large landslide at Buwun Mas village in West Nusa Tenggara Province. This slide is reported to have killed at least four people, with a further 11 potential victims believed to be buried in the debris. The report is quite interesting as it says that:

The area, [Rustam Pakaya, the head of the Ministry of Health’s Crisis Center] said, is used for illegal gold mining. “We are still searching for the missing people,” Rustam said. “We don’t have any information yet on who these victims are. We don’t know if they are miners or not.”

It should be said that other, unfortunately non-local, reports are rather more certain about the mining issue:

A landslide at a gold mine killed four workers and left 11 missing on the Indonesian island of Lombok, a Health Ministry official said Sunday."

Fortunately, this is an area with excellent Google Earth imagery, so a quick look at the site of the landslide shows the following (click on the image for a better view in a new window):

The area around the village appears to be quote densely forested, but note that the Google Earth data covers the junction between two epochs (periods) of imagery, which you can see from the colour change across the image above. I have shown this below:

Note how to the west of the line (north is approximately at the top of the image) the forest cover is quite intact, but to east it is mainly denuded. The junction is along the line that I have marked above. This presumably means that the imagery to the west is older and was taken when the forest was mostly intact, whilst to the east the images are more recent. In the meantime there has been extensive deforestation. Unfortunately, the consequences are all to clear to see - to the east of the line the ground is visibly eroded. Just a couple of kilometres from the village there is an area of extensive recent shallow landslides:

A perspective view shows that these are breaking out all over the deforested landscape:

Given that this is an area of illegal mining and extensive deforestation, the occurrence of destructive landslides should not be a surprise. The level of vulnerability here is indicated by the fact that the Tropical Rainfall Measuring Mission have not detected unusually high levels of rainfall in this area over the last few days.

The combination of deforestation and mining in tropical environments is really bad news from a landslide perspective.

Thursday, 15 January 2009

Intriguing long runout landslide in Death Valley

Thanks to reader Gregory T. Farrand who brought my attention to a very intriguing feature located in Death Valley in California. This is a feature that was first identified and mapped as a long runout landslide by Michael W. Hart. Greg and Michael, together with Brian Olson and Phil Shaller, are currently studying this slide, which they are terming the "Eureka Valley Landslide". The source rocks for the landslide are Cambrian marine sediments, mostly dolomite (dolostone). The slide is partially buried by Holocene alluvial fan deposits.

The landslide shows up really well on Google Earth. First, a vertical view with north almost at the top of the image:

Next an oblique view looking eastward:

Michael and Gregory are writing a paper on the slide, which will be very interesting indeed when it comes out. In the meantime, a few observations:
  • Identifying this is a great spot. Finding features like this is far from easy;
  • In fact, our understanding of landslides in very arid environments is rather poor, so such observations are definitely interesting;
  • The slide has appeared to have moved in such a way that it has formed a narrow tongue-like deposit. Such deposits are rare but certainly not unknown. Incidentally, the runout distance across the fan is about 1.7 km by my reckoning, although from the source area it may be twice this;
  • There is that interesting feature NNW of the toe of the slide deposit. I wonder if this is just a result of the landslide preventing the debris fan channels from actively depositing in this area, perhaps allowing wind erosion, or is it landslide material that has perhaps eroded and then been deposited;
  • The type of landslide of which this most reminded me is those seen in imagery from Mars. The image below was collected by the NASA Mars Global Surveyor Mars Orbitor Camera in 2004. The size of the landslide is very similar to that in Death Valley. I guess in some ways the conditions are similar (i.e. very arid), which is interesting in itself.

I hope that Gregory and Michael will let me know when the paper is published - I'll note on the blog when this is the case.

Guatemala landslide - new satellite image

The Advanced Land Imager instrument of the NASA Earth Observation satellite EO-1 has recently collected an image of the Guatemalan landslide, which they have made available online. The image is here. I have cropped it to show just the slide - it is quite useful:

Tuesday, 13 January 2009

Costa Rica Earthquake - the aftermath

I thought a brief follow up email to my previous posts about the Costa Rica earthquake would be useful now that we are five days post the event. So what do we now know?

Well, first the final death toll is likely to be about 40, most of whom were killed by landslides. Officially, as of today, 19 bodies have been recovered and 21 more are buried under landslide debris. It is possible that there is a small number of addition casualties under collapsed slopes. Landslides continue to occur during aftershocks and there is growing concern about the potential for future slides (see image below). Initial estimates of the damage are $100 million. In a country with a GDP of $26 billion, this is a huge amount (in comparison, the GDP of the UK is $2.13 trillion, meaning that the equivalent loss in terms of GDP would be $8.2 billion). Estimates are that the earthquake destroyed 518 houses, triggered the loss of 16 miles of road and collapsed eight bridges. Of course there will also be a substantial hit to tourism, which is vital in this area.

In terms of landslides, the emergency management office has indicated that the earthquake triggered more than 246 landslides. Repair of roads damaged by the landslides is expected to cost $15 million (see image below).

Sunday, 11 January 2009

Costa Rica earthquake - landslides have proven to be a major threat

News continues to flow in regarding the impact of the earthquake in Costa Rica. It is increasingly evident that landslides have probably been the major problem, which is unsurprising given the topography and soils. The landslides have clearly caused substantial loss of life and are also hampering rescue operations.

The Washington Post is reporting, quite rightly, that there are now concerns about the stability of the slopes in the affected area in the next wet season. The Reuters image below shows seriously deformed ground above the crown of a seismically-triggered landslide at San Rafael de Vara Blanca:

I found the diagram below showing rainfall in San Jose, which is just a few tens of kilometres away:

Fortunately the period January to April is pretty dry, but problems will certainly start in May. Action will be needed to address this as the level of hazard will be high in many areas. At the moment the biggest problems seem to be around the village of Cinchona, on the flank of the volcano, were a family restaurant was buried (see AP images below of Cinchona):

There are certainly lots of other landslides too, as this AP image shows:

Finally, the Washington Post is carrying a slightly intriguing report:

"A wave caused by the earthquake damaged the Cariblanco hydroelectric plant and it will take at least a year to repair, local media quote a senior official as saying."

A bit of digging shows that AP are reporting that "Officials reported meanwhile that the country's leading Cariblanco hydroelectric plant will be out of operation for about a year, after its generators were buried by tons of mud from flooding and landslides unleashed by the earthquake."

According to this site, Cariblanco appears to consist of two 40 MW turbines. It came into operation in 2007 after construction costs of US$170 million. This would appear to be a very expensive set of landslides for such a new project! Unfortunately, our experience elsewhere (e.g. Taiwan) is that sediment delivery increases in the rainy season as further landslides activate, so the operators of this site would be well advised to get some landslide advice.

Friday, 9 January 2009

Future British seasonal precipitation extremes - implications for landslides

ResearchBlogging.orgOne of the great questions of the age is of course the ways in which climate change will affect the weather patterns that we are likely to see in the future. In the case of landslides the key issue is the ways in which precipitation patterns will alter, especially the most intensive rainfall events that are responsible for many of the most damaging landslides. One of the most significant steps forward over the last few years has been the ability of global climate models to handle these extreme events, meaning that at last we are starting to develop some capability.

This week an important paper has been published by Fowler and Ekstrom (2009), which seeks to look at the likely changes to very intense rainfall events in the UK. Helen Fowler, is based just up the road from me at Newcastle (the city with the chronically under-performing football team), and her co-author have used modelling ensembles to examine how UK precipitation regimes are likely to change in the time period 2070 to 2100 under the SRES A2 emissions scenario, which is currently effectively our best estimate as to how carbon dioxide emissions will change with time (Fig. 1).

Fig 1: SRES Emissions Scenarios. A2, as used in this study, is shown in Fig. (b). Source:

Ensemble modelling looks at the results of a series of different climate models to examine the range of outputs. Each model operates in a slightly different way, meaning that there will always be a range of results. Therefore, papers presenting ensemble model outcomes always present a range. One of the key issues of interest is whether there is some consistency between them. In this study. Of course the results of such modelling runs are highly complex - in this paper the authors have looked at the 1 day and 10 day precipitation events with a current return period of 25 years. The 1 day event can be thought of as the impact of an intense storm; the 10 day probably simulates a series of low pressure systems tracking across the country, as has happened several times in the last few of years. In landslide terms the 1 day storms might trigger the catastrophic debris flow and sallow failure events, whilst the 10 day events might trigger deeper seated and large slope failures.

First the model is run for the a control period (1961-1990) to check that they can realistically simulate observed conditions. They can. The models are then run to look at what would happen in the period between 2070 and 2100, and the results are then pooled using a fairly interesting approach. Well, the first thing to say is that the Global Climate Models (GCMs) do predict a much warmer climate - global mean temperatures are predicted to be 3.1 to 3.56 degrees warmer than at present. Interestingly though the occurrence of these intense rainfall events also greatly increases for three of the four seasons:
Winter: Increases in occurrence of extreme precipitation of 5 to 30%
Spring: Increases in occurrence of extreme precipitation of 10 to 25%
Summer: Very varied results, with some models suggesting decreases and other increases. More work is needed
Autumn (Fall): Increases in occurrence of extreme precipitation of 5 to 25%

A few of the models do predict larger (and smaller) increases - look at the paper for the full detail. Overall, the authors conclude that "Nevertheless, importantly for policy makers, the multi-model ensembles of change project increases in extreme precipitation for most UK regions in winter, spring and autumn. This change is physically consistent with warmer air in the future climate being able to hold more moisture. The use of multi-day extremes and return periods also showed that short-duration extreme precipitation is projected to increase more than longer-duration extreme precipitation, where the latter is associated with narrower uncertainty ranges."

The implications for landslides are stark. Increases on this level of the occurrence of extreme precipitation events will inevitably increase the occurrence of slope failures. Therefore, we should expect to see an increase in the occurrence of slope failures. Unfortunately, as landslides are triggered by just a small proportion of our existing rainstorm events, increases in this range are likely to have a disproportionate impact.

Of course the next thing to do will be to build the outputs of these models into slope stability models. This will be a fascinating exercise.

H. J. Fowler, M. Ekström (2009). Multi-model ensemble estimates of climate change impacts on UK seasonal precipitation extremes International Journal of Climatology DOI: 10.1002/joc.1827

Landslide hazard and the Guatemala rockslide

As commenters on my earlier threads have pointed out (thanks to them), the location of the landslide in Guatemala can now be pinpointed using a map produced by CONRED. This map comes from a very useful report, with some good images of the rescue and recovery operation, available here. I have reproduced the map below (click on the map for a better view):

Helpfully, this allows the location to be pinpointed on Google Earth (unfortunately the high resolution imagery starts just east of the landslide location - click on the image for a decent view):

Satellite imagery can be difficult to interpret until a 3D perspective is available. This is one of the key strengths of Google Earth. The situation becomes so much clearer when a perspective view is taken:

A magnified and annotated image below shows that the landslide occurred in a very clear bowl shaped feature that would cause any good geomorphologist to be very nervous in terms of slope stability. I have highlighted the boundary of the bowl and the location of the landslide:

Note that there are a number of other locations that look vulnerable to slope instability as well, such as the bowl to the east of the area that failed. Reports suggest that the authorities are being cautious and have arranged evacuations. This is probably prudent, but it is important that a proper hazard assessment is undertaken, and that the risks associated with landslides are then balanced against the (social) risks to the people associated with moving them from their homes and their land.

Magnitude Mw=6.1 earthquake in Costa Rica

At 19:21 UTC (13:21 local time) on 8th January a Magnitude Mw=6.1 earthquake struck Costa Rica. Although this is not a particularly large earthquake, it is certainly substantial enough to have triggered many landslides, especially given that it occurred very close to the surface (the USGS depth estimate is just 4.5 km). Several reports have come in already about fatalities associated with landslides. Confirmed are the deaths of two girls near the city of Alajuela, NW of San Jose, whilst less well confirmed reports suggest that a family of five may also have perished in a landslide.

The USGS has provided a shake map:
This is an area of known high levels of seismic hazard:
This area is probably particularly prone to landslides as the soils consist of heavy weathered soils and those derived from volcanic products, both of which are potentially unstable when disturbed. I have come across two very interesting resources. First, a map of damage reports (in Spanish) is being put together, based on information from Twitter. This should be visible below:

Ver mapa más grande

Second, there is an interesting Youtube video of TV footage of the aftermath, much of it shot from a helicopter. The first part of the footage shows I think attempts to recover the remains of the two girls killed in a landslide, whilst the middle section shows landslides triggered in the epicentral region. Based upon this it might well be that further fatalities become apparent when daylight reaches this area. The video should be visible below, or can be viewed here.

Thursday, 8 January 2009

The Guatemala landslide - video of the failure in December

The Guatemalan landslide becomes more interesting by the day. It is now apparent that the first failure, which occurred on 14th December and killed five people (the remains of three of whom were not recovered) was also caught on video. This is available on Youtube. I have embedded it below, so should work:

If not, then it can be viewed here.

Guatemala landslide

As the Guatemala landslide now fades from media prominence, Reuters have released probably the best image to date of the landslide site:

It is now clear that this was a a large rockslide (not a mudslide as reported elsewhere).

AP have released a good image of the site:

Whilst this Reuters image shows the runout of the landslide quite well:

Finally, CONRED released this image of the site to show the ongoing rockfall activity that caused suspension of rescue operations:

Meanwhile the number of recovered victims has now reached 37. About 50 victims remain buried in the debris.

Finally, there are various reports that an evaluation of the site has determined that the area around the landslide is too dangerous for habitation. My suspcicion is that this is right - the landslide appears to me to have occurred in a much larger landslide scar. I will post again on this soon.

Wednesday, 7 January 2009

An update on landslides in Beichuan

At the time of the Wenchuan earthquake I posted on numerous occasions about the landslides around the town of Beichuan. One of the concerns has always been the likely behaviour of the landslides during heavy rainfall events.

As part of the forthcoming Sinorock 2009 conference in Hong Kong there will be a tour to the earthquake-affected area to look at and discuss the landslides. The organisers have put together a web page about this tour (see here). Interestingly, thy have posted three images of Beichuan:

First, before the earthquake:

Second after the earthquake:

And third, most importantly, after a very intense rainfall event on 24th September 2008:

Note the extensive movement of sediment, which has caused the river bed to agrade and bury many structures. This is very similar to what happened in parts of Taiwan after the 1999 Chi-Chi earthquake. Such behaviour represents a serious challenge for the authorities.

Another video of the Guatemala landslide

It now appears that there are two videos of the Guatemala landslide. The second one, which is not as clear as the original, has appeared on the Telediaro 3 web site. It can be viewed here:

Meanwhile, there remains considerable uncertainty as to the number of victims (we may never know for certain I suspect), whilst speculation is intensifying regarding the cause. Given the clear size of the landslide (see the video plus the image below from Prensa Libre), comments elsewhere that road widening was the primary trigger look to be wide of the improbable. Clearly there is no record of seismicity or of intense rainfall (everything looks dry in the video). The most likely trigger is therefore progressive failure, which is consistent with reports of increased activity in the fortnight before the failure, reports of the slope emitting noises and the fact that the road had been closed prior to collapse.

Finally, there is continued concerns that a second major failure might occur given the oversteepened nature of the upper part of the slope.

Tuesday, 6 January 2009

The Guatemala landslide was caught on video!

Unbelievably, it turns out that the Guatemalan landslide, which it appears killed at least 42 people, was caught on video. Geologist Marlon Contreras was recording at the edge of the site when the failure occurred. The video captures the movement, although mercifully little of the detail. The video can be viewed at this web page:

The failure that happens 1 minute and 50 seconds into the video defies belief! I urge you to take a look at the video.

Meanwhile, and quite rightly, rescue operations have been halted due to the dangers of continued falling rocks to the rescuers. Given that there is no possibility of any survivors this has to be the right decision, harsh though it undoubtedly is.

Finally, there is still a huge variation in the reported number of victims. Prensa Libre is reporting that 42 people were affected, but other news agencies suggest that it might be double this total, or even more.

The Tennessee flow slide accident - what went wrong?

The fallout from the recent flowslide accident at the Tennessee Valley Authority's (TVA) Kingston power plant in the USA on 22nd December, which I highlighted here, continue. There is an increasing sense that this was an unacceptable failure (it is hard to argue with that) and that it was lucky that there was not a loss of life in the accident.

This AP image is a good starting point as it is much clearer from this as to what has happened:

Perhaps the most useful resource though is this article from the, which I thoroughly recommend. It provides an excellent interpretation of what happened in the lead up to the accident. They have provided two aerial images of the site, before:

And after:

The gist of the article is as follows:
  1. The failure occurred on a 20 m high embankment constructed primarily from ash blocks that collected in the base of the plant's burners. The fly ash is deposited into the lagoon in a wet state;
  2. In November 2003 the lagoon suffered from a blow out (failure) caused by piping and seepage (i.e. water starts to flow through the ash, which then induces erosion. This then allows a pipe to form, which collects, channels and accelerates the flow, allowing erosion to intensify. This can rapidly create a weak zone);
  3. In December 2003 a mitigation scheme was proposed, consisting of "a series of trench drains at different levels on the dike, another drain at the base of the mound and a riprap (i.e. armoured) channel."
  4. This mitigation scheme was completed in October 2005
  5. In 2006 a further failure occurred, with a 1.5 x 1.5 m section failing through seepage, releasing some ash;
  6. This section was repaired and wells were installed behind the embankment, presumably to lower water pressure, monitored using 30 shallow piezometers;
  7. The article mentions that the embankment was inspected and passed in a report produced on 15th February 2008 and found to be stable, but that "plant operators were commended for mowing the landfill slopes. Trees too large to be mowed should be cut, the stumps removed, the area backfilled with soil and seeded". "They mentioned small trees being removed," said Tschantz, the dam safety consultant. "I'm wondering if trees had a role to play. You don't just pull those things out. The root channels have to be filled and compacted."
In December 2008 the site received 125 mm of rainfall (the average for December is 70 mm). Failure occurred on 22nd December.

Clearly it is far too early to identify exactly what has gone wrong or who, if anyone, is to blame. The key issues must be to clear up the mess and to learn the lessons, especially if there are other sites that are also vulnerable to failure. One suspects that the costs of this incident will be very high indeed - far more than the cost of constructing a safe embankment in the first place.

I do recommend the article (here) - it appears to be a good piece of journalism.

As an aside, it appears from the images that sections of the embankment have been moved intact (see image below - these can also be seen in the photo at the top of this post):

The sides of the embankment also appear to have failed (marked "lateral failure?" above). This suggests to me that pore pressures in the ash were very high indeed. I wonder if a factor here is the use of ash in the embankment as its comparatively low density might have made the embankment rather vulnerable to failure due to high buoyancy forces.

Monday, 5 January 2009

December 2008 landslide map

I am just in the process of putting together my annual collation of fatal landslide statistics for 2008, together with the map of all the landslides, but in the meantime here are the statistics and map for December.

So first the data. I recorded a total of 16 fatal landslides in December, causing 46 fatalities. This is well below the long term average, which 149.8 fatalities. The largest event was the Kainantu Mine landslide in Papua New Guinea on 4th December, which killed 10 people. I guess the highest profile event was the Bukit Antarabangsa landslide near to Kuala Lumpur, which killed five people on 6th December. This has really dominated the news in Malaysia since.

Here is the map. You will see that I have taken the "copyright" statement off to make it more usable. Please do acknowledge this blog if you use it in presentations etc, and please don't publish it without my permission. Click on the map for a better view:

As mentioned above I will provide a proper commentary for the 2008 dataset in due course (might take a week or two), but in the meantime here is the provisional final map for 2008. As before, click on the map for a better view:

Comments and/or corrections welcome. One of my tasks for this year is to get a better mapping program than Versamap (suggestions welcome)...

Guatemala landslide update: 33 killed and up to 70 people reported missing

A short update on the Guatemala landslide, which is looking like a dreadful tragedy.

Diaro de Centro America has published this dramatic image of the site:

Meanwhile Prensa Libre is reporting that the number of people killed is now known to be 33, with more people still missing. Telediaro suggests that the number missing may be as high as 70 people! Landslides are continuing to hamper the rescue operations.