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.

Monday, 30 November 2009

Slumps caused by thawing ground on Mars and Earth

The Planetary Geomorphology Working Group of the International Association of Geomorphologists has a rather nice article online comparing the landforms caused by slumping during thawing of the ground with similar features that have been seen on Mars. The article is available here:

On Earth, thaw slumps occur in permafrost areas like Alaska. This is an oblique aerial image of these features, taken from the site above:

The Natural Resources Canada has quite a nice (although I suspect somewhat old) cartoon to illustrate how these landslides work:

Essentially, the thaw of permafrost (ground that is usually frozen) allows the weak materials to fail and flow. This typically exposes a new face of frozen ground that, if the temperatures are high enough again, thaws and flows. Thus, over time, the back scarp of the landslide moves back into the hill - i.e. it retrogresses. This is what one looks like in a vertical view on Google Earth:

The features observed (by satellite obviously) on Mars have much the same morphology as retrogressive thaw slides on Earth:

The arrowed features are interpreted as pingos, which are only found in permafrost areas on Earth. Thus, the arcuate features upslope of the pingos are interpreted as the arcuate backscarp of the landslides. The existence of these features of course implies that at some point the ground thawed and there was liquid water present at some point in the past.

Sunday, 29 November 2009

Landslides in art

I thought that it would be interesting to run a series of occasional posts on landslides in art. The first is, rather predictably, called "Landslide" by Francie Lyshak, the source of which is here:

Landslide, by Francie Lyshak, Oil on linen, 50 x 37, 1990

The accompanying text, by Joe Vojkto, says:

"The cosmic site-gag of Francie Lyshak’s painting, Landslide, hits like a ton of bricks and opens up old scars, where pieces of ourselves have been removed, to shine an unforgiving light inside. Viewing this painting for the first time is a sobering act, resonating with aftershocks of emotional recognition that can drag your thoughts unwillingly back through creaky episodes of your own personal history that you never wanted to inhabit the first time around. Landslide unapologetically assaults, and in refusing to be simply beautiful, its stones us numb with the sour truth that beauty is a beast. It's Golgotha, the place of skulls, the scene of the crime for which we are all supposed to be paying. There are no crosses at the summit of the rubble, but the source of light makes it obvious that some heaven has delivered this cold load. Landslide remains in the memory, almost the quintessential symbol of life at this late hour in the rocky history of our species, a tarot card to represent the stasis of despair, the metastasis of betrayal."

So now you know (no, I don't understand it either)!

Saturday, 28 November 2009

The link between rainfall intensity and global temperature

The aftermath of a landslide in Taiwan caused by very heavy rainfall

One of the most interesting aspects of the global landslide database that we maintain at Durham is the way in which it has highlighted the importance of rainfall intensity in the triggering of fatal landslides. Generally speaking, to kill people a landslide needs to move quickly rapid, and rapid landslides appear to be primarily (but note not always) triggered by intense rainfall events (indeed in the reports the term "cloudburst" often crops up). So, a key component of trying to understand the impacts of human-induced global climate change on landslides is the likely nature of changes in rainfall intensity, rather than that of rainfall total. Put another way, it is possible that the average annual rainfall for an area might decrease but the occurrence of landslides increase if the rainfall arrives in more intense bursts.

There is of course a certain intuitive logic in the idea that rainfall intensity might increase with temperature. Warmer air is able to hold more moisture (as anyone who has been in the subtropics in the summer will know only too well!) and of course increased temperatures also drive greater convection, responsible for thunderstorm rainfall. Of course this is a very simplistic way to look at a highly complex system, so it is not enough to rely upon this chain of logical thought. However, until now there have been surprisingly few studies to actually quantify whether there is a relationship between global temperature and precipitation intensity, which has meant that for landslides understanding the likely impact of climate change has been quite difficult.

However, an important and rather useful paper examining exactly this issue has sneaked under the radar in the last few months. The paper, by Liu et al (2009) (see reference below), was published in Geophysical Research Letters a couple of months ago. The paper uses data from the Global Precipitation Climatology Project (GPCP). These data can be accessed online here (so no claims that climate scientists don't publish their data, please!) The dataset provides daily rainfall totals for 2.5 x 2.5 degree grid squares across the globe, extending back almost 50 years. Liu et al. (2009) looked at the data from 1979 to 2007, comparing precipitation density with global temperature in this time period.

Their results are both unsurprising and surprising. The unsurprising part is that they found that the occurrence of the most intense precipitation events does increase with temperature. The surprising part is the magnitude of the change - they found that a 1 degree Kelvin (Centigrade) increase in global temperature causes a 94% increase in the most intense rainfall events, with a decrease in the moderate to light rainfall events. Indeed the median rainfall increased from 4.3 mm day−1 to 18 mm day−1, which is a surprisingly high shift as well.

So why is this important in the context of landslides? Well, I think that there are probably two key implications:

1. It has long been speculated that anthropogenic warming will lead to an increase in landslides, but with little real quantitative evidence to confirm or deny this. The demonstration that higher global temperatures does lead to increased precipitation intensity starts to put some meat on the bones of this idea. Furthermore, if it is possible to directly link rainfall intensity to landslide occurrence (and there is some evidence both from my own work and from that of others that this may be possible), then it should be possible to start to examine the likely increase in landslides as warming proceeds.
2. The current global climate models assume a much lower increase overall in precipitation intensity with increasing temperature than Liu et al. (2009) suggest. Indeed most of the models assume about a 7% increase per degree Kelvin (Centigrade) warming. For the most intense precipitation events this means that the models predict about a 9% increase, which is an order of magnitude lower Liu et al. (2009) found. This suggests that the rainfall projections that are derived from the models are probably overly-conservative, and possibly very much so, which is a concern. If so, then forecasts of landslide occurrence that are derived from these models are likely to under-estimate the true impact.

Of course, this is only one study, and it should also be noted that the most intense rainfall events are usually associated with tropical areas and with those in the path of hurricanes and in particular typhoons. There is a great deal more work to do on this topic, but the initial results provide real cause for concern.

Liu, S., Fu, C., Shiu, C., Chen, J., & Wu, F. (2009). Temperature dependence of global precipitation extremes Geophysical Research Letters, 36 (17) DOI: 10.1029/2009GL040218

Thursday, 26 November 2009

The damage caused by landslides during earthquakes

Below is the presentation file of the keynote lecture that I gave today at the Chilean Geological Congress in Santiago. I have removed a few of the figures as they have not yet been published.

You should be able to download and to view the file below:

The powerpoint file is hosted on Authorstream, which also holds many more of my presentations here:

The Copenhagen Diagnosis

As a rule on this blog I try not to drift too far off-topic. Occasionally I will track a hurricane or a typhoon if there is the potential for them to trigger lots of slides. I never stray into the political domain on this blog - and I hope that I never will.

So, today for the probably the first time, I am going to drift into a different field. I am going to remain strictly within the science, but I am going to highlight a key document that is now available. Ahead of the Copenhagen climate discussions a group of key climatologists have released "The Copenhagen Diagnosis", which is an update to the most recent IPCC report on climate change.

Now, over the last week or so there has been some extraordinary garbage on both the internet and in the mainstream media in relation to the stolen emails and files from the University of East Anglia. I am appalled at the way in which comments within those emails have been manipulated and misrepresented, even by some who should know better, to undermine climate science. Meanwhile the physics of the atmosphere continues as before, and our knowledge of how it is behaving is unchanged, and the picture is not a good one. The strength of the Climate Diagnosis report is the fact that it relies upon good quality observational data. And lets be clear about what the data show:
  • Greenhouse gas emissions are increasing;
  • The climate is continuing to warm (and despite the guff that the denialists spout, tenperatures have increased over the last decade);
  • Sea level is rising at the top end of the previous estimates;
  • Arctice sea ice decline has accelerated
So why have I drifted into featuring climate science on this blog? Well, for two reasons. First, the denialists continue to peddle the myth that there is no consensus. This is not my experience, and so I think it is time for the silent majority of environmental scientists to stand up and be counted. Second, the CRU / Hadley Centre email hack was clearly designed to derail the science of climate ahead of Copenhagen. This is an outrageous attack on science and scientists that we should all oppose.

The document can be downloaded here. Please read it and try to understand what it is saying. This is so very, very important.

Wednesday, 25 November 2009

A follow up to the Tennessee landslide video

A couple of weeks ago I highlighted a new landslide video from Highway 64 in Tennessee, USA. Thanks to the excellent GeoPracNet for highlighting that some aerial imagery is now available from News Channel 9 of the slide:

Photo by Dennis Hutcheson

The sliding plane was clearly an existing discontinuity sub-parallel to the slope, with movement perhaps being allowed by the cutting of the slope to create the road bench.

Tuesday, 24 November 2009

More El Salvador lahar satellite imagery

In an earlier post today I highlighted the availability of satellite imagery covering the El Salvador lahar disaster, triggered by Hurricane Ida earlier this month. Whilst sitting in a conference session this morning it occurred to me that the Disaster Charter, which provides satellite images for disaster relief and recovery operation, was also triggered by this event. So I had a quick look at the web page for this triggering of the charter, and sure enough there are some good materials there.

Perhaps the best image is this interpretation of the lahar tracks, produce using the Taiwanese Formosat2 satellite (famous for its images of Beichuan and Tangjiashan after the Wenchuan Earthquake) - do click on the image to see this properly, it is worth it!:

There is also a more detailed look at the town of Guadeloupe using the same satellite, showing before and after images of the town:

Viewed in conjunction with the NASA images in my earlier post, this is a very useful resource.

Satellite imagery of the El Salvador lahars

NASA has released before and after satellite images of the lahars (volcanic debris flows) triggered by Hurricane Ida in El Salvador earlier this month. The final toll of the landslides remains unclear, over 60 is the widely reported statistic.

The images are false colour composites collected by the ASTER instrument. These images differentiate between vegetation, which appears as red colours, and bare ground, which appear as a blue-grey colour. Such images are ideal for delineating new landslides, which usually strip away vegetation to leave bare soil.

This is the before image of the area affected by the landslides:

As usual, click on the image for a better view in a new window. Here is the image after the hurricanes passed through:

The lahar tracks are pretty clear. Note that they start high on the volcano, mostly with small, translational landslides. These slides then enter the channel, entraining (eroding and incorporating) material along the channel. The towns are quite clear on the images - the debris flows have hit the settlements in several locations, leaving this type of damage (Image from here):

When the landslides reach the plains below the volcano they spread out and stop, leaving a large area covered in debris. This is also clear on the image.

Monday, 23 November 2009

The ten greatest landslide papers?

Whilst pushing my five year old daughter on the swing the other day (a task that leaves plenty of time to think!), I was pondering upon the greatest landslide papers of all time. I thought that it would be interesting to compile a list and invite suggestions of alternatives. So, here is my list, in no particular order:

1. Terzaghi on the principle of effective stress
In 1936 Karl von Terzaghi laid the foundation for modern soil mechanics, and the basis of our understanding of how landslides move, by stating the principle of effective stress. The recognition that pore water pressures control the frictional resistance of slopes remains the most important concept in understanding landslide behaviour.
Reference: Terzaghi, K., 1936. "The Shear Resistance of Saturated Soils". Proceedings of the first International Conference on Soil Mechanics and Foundation Engineering, Cambridge, U. S. A., 1, 54-56.

2. Keefer on earthquakes caused by landslides
Dave Keefer's 1984 paper on landslides triggered by earthquakes was a remarkable analysis. Taking a huge dataset of earthquake-triggered landslide inventories, Keefer demonstrated both the simplicity and the complexity of the relationship between numbers and areas of landslides and the earthquake parameters. The approach taken in this paper has been much reproduced since; remarkably it the core conclusions are essentially unaltered.
Reference: Keefer, D.K., 1984, Landslides caused by earthquakes. Geological Society of America Bulletin, 95, 406-421.

3. Hutchinson and Bhandari on undrained loading
Undrained loading is one of those concepts that makes you go "of sourse" when it is explained to you. The idea is that the sudden application of a load (such as for example a rockfall from a cliff onto an mudslide below) drives a dramatic increase in pore pressures, which can't dissipate quickly. This reduces the effective normal stress, allowing the slope to move. It explains landslides in many settings - a crucial step forward.
Reference: Hutchinson, J. N. & Bhandari, R. K. 1972. Undrained loading, a fundamental mechanism of mudflows and other mass movements. Geotechnique 21, 353-358

4. Caine on thresholds associated with landslide initiation

Almost all of the working large area landslide warning systems are based on this paper. Caine set out to understand the rainfall thresholds at which landslides occur, recognising that it is a combination of medium term low intensity rainfall (to get the ground wet) and short duration rainfall (to get initiate movement) that is the key. This is literally the paper that launched a thousand studies - and more appear every year. There is even a website dedicated to this type of work:
Reference: Caine, N., 1980. The rainfall intensity-duration control of shallow landslides and debris flows. Geografiska Annaler, 62A, 23-27.

5. Skempton on the principle of residual strength
Residual strength is a key idea within landslide science. Skempton's work was prompted by the difficulties of explaining why low gradient slopes were failing. The idea that materials have a residual strength that is less than the peak strength is key - slopes that have in previous times been reduced to this lower strength will fail much more easily when disturbed. This concept really made Skempton's career, and he was ultimately knighted for his scientific contributions.
Reference: Skempton, A.W. 1964. Long term stability of clay slopes. Geotechnique, 14, 77-101 (the fourth Rankine lecture).

6. Carson and Petley on the existence of threshold slopes
In case you are wondering, this Petley is not me, but shall we say that when I was a child I spent a great deal of time with him (and continue to do so, but more rarely now). In many ways the dataset in this paper is not comprehensive, but the ideas that it introduced are fundamental. The paper suggested that in any environment a slope will relax to a constant gradient that reflects its stength and the conditions to which it is exposed. The idea was spot on, and is just being rediscovered by geophysicists!
Reference: Carson, M.A., and Petley, D.J., 1970, Existence of threshold hillslopes in denudation of
landscape. Transactions of the Institute of British Geographers, 49, 71–95.

7. Carrara on the evaluation of landslide hazardLandslide hazard assessment is big business these days. Carrara's paper was not the very first to do this, but it was a pioneering study in terms of using the rapidly evolving technologies that were becoming available. The myriad of factor and suchlike based studies that have followed over the last 26 years all owe a great deal to this study.
Reference: Carrara, A. 1983. Multivariate models for landslide hazard evaluation. Journal of the International Association for Mathematical Geology, 15, 403-426.

8. Hoek and Bray on rock slope engineering
This is the only item on my top ten that is a book rather than a paper. The stability of rockslopes is primarily controlled by the properties, orientations and interactions of discontinuities rather than by the intact material strength. This book brought together for the first time the principles of rockslope stability and design. It has been revised and reprinted on numerous occasions, and remains the bible of rock slope engineers.
Reference: Hoek, E. and Bray, J. 1974. Rock slope engineering. Institution of Mining and Metallurgy, 309 pp.

9. Bjerrum on progressive failure
The principles and processes of progressive failure remain poorly understood, but Bjerrum's 1967 examination of this key topic remains the salient work in this area. Bjerrum proposed a model for the weakening of slopes with time as a crack grows through the base. The concepts remain fresh and valid today - there is surely a great opportunity for someone to take this issue and produce the definitive follow-up on how this process actually operates.
Reference: Bjerrum L. 1967. Progressive failure in slopes of overconsolidated plastic clay and clay shales. Journal of the Soil Mechanics and Foundations Division, ASCE, 93, 1-49.

10. Bishop and Wesley on soil testing
Without laboratory testing of laboratory materials slope engineering would still be in the dark ages. The key research machine in every geotechnical laboratory is the triaxial cell, and research purposes the stress path cell is the most important tool. This paper described just such a machine, opening the way to our fundamental understanding of soil behaviour.
Reference: Bishop, A.W. and Wesley, L.D. 1975. A hydraulic triaxial apparatus for controlled stress path testing. Geotechnique 25, 657-670.

I'm sure that you don't agree with me, so please tell me why I am wrong, and suggest alternatives. I am worried that there is nothing since 1984 - surely there must have been great papers in the intervening 25 years. What have I missed?

Sunday, 22 November 2009

Landslide impacts in Turkey

The last few days have seen very heavy rain in Turkey, with a number of landslides. ITN has a short clip of home footage showing a large buikding collapsing as it is hit by a landslide:

The distress shown by the man in the foreground - presumably an owner or resident - illustrates graphically the impact of these events on people.

Saturday, 21 November 2009

A very large ancient rockslide in Chile

I am en route to Santiago in Chile to attend the Chilean Geological Congress, the organisers of which kindly invited me to give one of the keynote lectures (on Friday). I thought therefore that I would point out that Chile has an extraordinary set of very large rock avalanches. Earlier this year, Antinao and Gosse (2009) published an interesting review of a set in the Chilean Cordillera Principal. I do not intend to publish a full review of the paper here, but thought I would highlight just one of the slides, called Yeso - Meson Alto. This is a very large landslide, as the Google Earth image below shows:

I have eyeballed in the main landslide deposit (note that in places this is covered with more recent fluvial (river) deposits, so this is very approximate. I have used the supplementary information from Antinao and Gosse (2009) as a guide as this contains a map of the landslide, and another large slide downstream (which I have included below my indicative lines on the Google Earth image:

The statistics for this landslide provided by Antinao and Gosse (2009) are impressive:
Volume: 4.5 cubic kilometres
Original surface area: 22.7 square kilometres
Length (travel distance): 7.6 km

The landslide is considered to have been translational as the deposit retains the original stratigraphy. Interestingly, it is thought to have occurred in the Holocene (i.e. in the last 12,000 years).

Antinao, J., & Gosse, J. (2009). Large rockslides in the Southern Central Andes of Chile (32–34.5°S): Tectonic control and significance for Quaternary landscape evolution Geomorphology, 104 (3-4), 117-133 DOI: 10.1016/j.geomorph.2008.08.008

Friday, 20 November 2009

Extreme rainfall in NW England and SW Scotland

The Northwest of England and Southwest of Scotrland have suffered an extreme rainfall event over the last few days that has brought extensive flooding. The Met Office in the UK are suggesting that this might represent a UK rainfall record (we will see), but certainly the impacts are extraordinary (this is Cockermouth in Cumbria; there is also extensive flood damage in Dumfries and Galloway in Scotland):

Image from Skynews

Image from BBC news.

The UK Met Office have provided this graphic of the rainfall distribution:

The images above are from Cumbria, which is the area highlighted as having the highest rainfall.

Unfortunately, there is more rainfall on its way. This is the BBC Weather forecast for this area at noon tomorrow:

Fortunately, this rainfall event should not be on a par with the last 24 hours (the Met Office are forecasting up to 40 mm over higher ground), when >170 mm of rain fell in some areas of Cumbria, but it certainly won't help the clean up.

There are a few reports of landslides, such as:
  • Train services disrupted because of a landslide between Carlisle and Penrith;.
  • A landslide in Rossendale in Lancashire on Wednesday that damaged a packaging factory:

Thursday, 19 November 2009

Surely the most over-the-top headline of the year!

Various daily newspapers in the UK are reporting on the Pennan landslide situation (more of which below). However first prize in the hysterical over-reaction category (the British tabloids pride themselves in their ability to out-do each other in this class) goes to the Mirror for this headline:

Unbelievable! Meanwhile, back in the real world, the residents of the village have now been advised to leave the village for a week whilst the landslide is remediated. The plan appears to be to remove the unstable section above the road. The BBC has a good short video of the site here.

Wednesday, 18 November 2009

Tailings dam failures and the price of commodities

A couple of weeks ago the University of Alberta hosted a conference entitled "Tailings and Mine Waste 2009". Impressively, all of the presentations from the meeting have been posted online on an FTP site here. Most of the presentations focus upon technical aspects of tailings facility design, but there is a very interesting talk from Michael Davis and Todd Martin from AMEC online as a pdf here. This presentation examines the relationship between the occurrence of tailings dam failures and the economic cycle of commodities. I should say upfront that I find parts of the presentation rather uncomfortable (especially the back-slapping aspects of the last slide regarding the oil sands industry), but the core point of the presentation is certainly thought provoking.

In the presentation they note that between 1968 and August 2009 there were 143 documented tailings dam failures worldwide. However, the occurrence of these failures appears to be cyclic with time, with peaks in the periods 1976-8, 1984-6, 1990-2, 1998-2000 and 2008-now. They compared these peaks with the cyclicity of the global copper and gold prices. The key part of the presentation is a table that compared the timing of the peaks:

The authors' conclusion is that there is a relationship between the peak in commodities prices and the occurrence of tailings dam failures, with a lag between the two of about two years. I must admit that I am a little unconvinced by the statistics of this analysis (I would like to see a proper regression analysis to see whether this link is statistically valid - and to be fair the authors recognise that this is not a scientifically-rigorous analysis), but the central point is one that is certainly very thought provoking. Increased commodity prices drive increased exploitation. The relationship between the peak in prices an the peak in accidents is ascribed by the authors to:
  1. The rush to mine quickly means that design and construction standards may be low;
  2. Rapid turn-over of key staff as new (presumably lucrative) opportunities arise during the boom;
  3. The boom drives the development of resources in areas that are known to be difficult;
  4. after the boom there are pressures to cut costs as commodity prices decline;
  5. The boom drives the use of inappropriate designs imported from other locations;
  6. There may be a lack of independent review, presumably to avoid the time delays and costs associated with this.
The implications of all this are sobering. The most recent boom ended in 2008, and the average lag is 25 months for copper and 29 months for gold. This should mean that the peak in tailings dam failures should be expected 12 to 30 months from now.

Hat-tip to Jack Caldwell's excellent I think Mining blog for highlighting this paper.

Tuesday, 17 November 2009

Threat of another landslide at Pennan, the site of the film "Local Hero"

The film Local Hero was released in 1983, having been filmed in the village of Pennan, near Aberdeen in NE Scotland. In my view it is one of the best films of the 1980's - understated, beautiful and genuinely funny, with a very strong environmental message. The village at Pennan has become something of an attraction as a result but, located so far from the beaten track, it has never been a real mecca for visitors. The attraction of the village to the film crew is clear from this photograph (from here):

However, look carefully at the picture above and you will probably see the problem that the village faces. Unfortunately the slope behind the village is unstable - in 2007 the village suffered a mudslide that flowed into homes, the pub and the village hall, causing considerable damage. The pub remained closed until last month, when a new landlord took over. Finally things started to look up for the village again.

Thus, it was with great sadness tonight that I read that the residents of the village have been told to prepare for another potential evacuation as another tension crack has opened up on the cliffs above the village. BBC News reports that this crack is 25 m long. Fortunately, this time the potential landslide does not appear to threaten the houses directly, but the road is threatened. The Google Earth image below shows just how critical this road is to the village:

The UK has been suffering heavy rain over the last few days, which is of course the likely cause of this problem. Tomorrow is forecast to be drier in this area, but further rain is forecast for Thursday and, in particular, Friday. It will be a nervous few days for the people of Pennan. I wish them well.

Images of the Shanxi loess landslide

Xinhua has now published two images of yesterday's loess landslide in Shanxi, which killed 23 people, including eight young children, in Zhangjiaju Village, in Zhangzishan Township of Zhongyang County:
The reports are that this was a "loess avalanche" with a surface area of about 20,000 cubic metres. An investigation into the cause is ongoing.

Now that is what I call a boulder!

Thanks to Dr Kurt Douglas at the University of New South Wales for drawing my attention to this. The Australian Courier Mail has a story today about a family that survived a near miss from a 1500 tonne boulder that rolled from the hill above their house in southern Queensland on Friday. The rock detached from the slope above their house and rolled to within 50 m of their property at 3 am. All this sounds quite unremarkable until you see the pictures of the boulder:

And the source:

Monday, 16 November 2009

Mining related loess landslide in China kills at least 12

See update here

The Chinese language version of the Xinhua website is reporting that a landslide occurred this morning in Zhangzishan Township, Zhongyang County, Luliang in Shanxi province, burying five houses. Reports suggest that about 25 people were buried - of these two have been rescued alive whilst 12 bodies have been recovered so far. Xinhua has images of the scene:

You may well have noticed the slightly odd aspect of this slide - which is that there is snow on the ground. Landslides under these conditions are not that common unless there is rapid snow melt occurring (which is unlikely at this time of year). However, the proper explanation probably lies in the fact that the houses that were hit were the temporary camp from a coal mine. This material that makes up the slide does not look like loess not coal waste, but the landscape does look like it has been excavated and disturbed, so I suspect that this is another tragic accident in a long history of mining accidents in China.

Typically, the Google Earth image of the area is low resolution in the area of interest (the reported location is at the bottom of the image):

However, zooming into the loess area at the top of the slide suggests that open pit coal mines are active in this area:

See update here

Sunday, 15 November 2009

Filling at the Three Gorges Dam suspended due to landslides?

Probe International image of a landslide

The Three Gorges Dam in China is one of mankind's greatest achievements or a folly on a grand scale, depending upon your point of view. Whichever is correct, there is little doubt that its construction is an extraordinary engineering achievement. The final aim of the project was due to be achieved this month as the lake is finally elevated to its maximum level. However, filling of the reservoir has suddenly been suspended four metres short of the maximum level, and there is a great deal of speculation as to why this should be the case.

AP is reporting upon an interesting difference between the official story and that being reported locally. The official story is that filling has had to be suspended due to low river flows caused by a drought downstream in Hunan and Jiangxi, downstream of the dam. However, the AP article speculates that at least in part this cessation of filling may be associated with increasing concerns about the threat of landslides on the gorge walls as the final phase of filling is completed.

So, what is the real situation? Well, there is little doubt that there is a serious drought in southern China at the moment. For example, the China Meteorological Agency reported on 30th October:

"China's Hunan and Jiangxi provinces, known for their rivers and lakes, are often portrayed as lands abundant with rice and fish. However, even they have not been spared by a drought that has plagued a wide swathe of south and east China since August, usually part of the rainy season. The drought has damaged crops and left 2 million people suffering water shortages and thousands of boats grounded on the shallow rivers or cracked riverbeds...Hunan received 60 percent less rain than normal in August and September. Jiangxi received 60 percent less in September and 90 percent in October.The drought was worsened by unusually high temperatures. The average temperature in Jiangxi was 2.5 degrees Celsius higher than average, and the highest since 1963. The Xiangjiang and Ganjiang rivers, two major tributaries of the Yangtze River running through the provinces, are reporting record low water levels. The water flow in the Xiangjiang River is down to about 500 cubic meters per second, compared to 1,200 cubic meters per second, which is the average for the time of year."

However, on 6th November the Three Gorges authority reportedly defended the continuing plan to fill the reservoir:
"China Three Gorges Corporation defended the ongoing plan to raise water level in reservoir to 175 meters during dry season, saying its water storage will help prepare for possible worse drought in coming months...Beginning Sunday morning, the Three Gorges project had increased its water flow to downstream to over 9,000 cubic meters per second, about 38 percent more than originally planned, to relieve the drought situation downstream...Demands for the Three Gorges Dam to discharge more reservoir water were on the increase, forcing the project to slow down its water-raising pace to 175 meters, a target that was rescheduled to early November from the end of October. The water level of the Three Gorges reservoir stood at 170.47 meters at around 11:00 a.m. Sunday. Zheng Shouren, Chief Engineer of the Yangtze River Water Resources Committee, said those blaming the drought in Hunan and Jiangxi solely on the Three Gorges project got a partial picture."

So how likely is it that landslides are the issue? Well, slope instability on the banks of the Three Gorges reservoir has long been a fear (see for example this article from 1999), and there have been many well-documented landslides already. Back in 2007 I wrote an article in China Dialogue expressing concerns about the potential for landslides on the banks of the lake. The concern has certainly increased in recent months, and last week the Chinese investigative magazine Caijing quoted official reports that stated concerns about the likelihood of landslides. AP reports that on 16th October new instability was noted above the township of Quchi, with a tension crack 400 m long being observed.

It is unsurprising that increased movement will be observed as the reservoir rises, although of course as this is the dry season the slopes should be more stable now than they will be in the late summer. However, concerns about instability are of course well-founded, so it will be interesting to watch this story develop.

Wednesday, 11 November 2009

A surprising spell of landslides

In most years in which I have been collecting landslide fatality statistics (since 2002) November has been a month with quite low landslide occurrence. Not this year though - in fact the last few days have been more like the summer monsoon season, with fatal landslides occurring in a wide range of places. Examples include:
  • El Salvador on 6th November, in which the volcanic lahars triggered by Hurricane Ida appear to have killed over 100 people, with a few (very speculative) reports of rather more;
  • South Sulawesi in Indonesia, when 13 people were killed in a slide Palopo;
  • Tamil Nadu on 9th and 10th November, when a number of landslides killed at least 30 people;
  • Casamicciola on Ischia in Italy on 9th November, when a 15 year old girl was killed by a slide that carried her out to sea in a bus;
  • Bukittinggi in West Sumatra on 10th November, when two people were killed;
  • Wassa-Akropong in Ghana on 10th November, when up to 30 people were killed in a landslide in an illegal mine;
  • Goha village in Tanzania on 10th November, when 20 people were killed in a landslide.
In addition there have been reports of non-fatal landslides in the USA, Serbia, Vietnam, etc. I cannot tell at the moment whether this is just coincidence or something more significant, perhaps due to the El Nino conditions. With Cyclone Phyan currently bringing heavy rainfall to central western India, an area that is landslide-prone, there could be further events in the next few days.

One of the best landslide videos so far!

Many of the landslide videos that I have featured on here have been quite low quality. The one featured below is unusually good. I have updated this to provide the (better) Youtube version:

The slide apparently occurred at Polk County in Tennessee, USA, on 10th November 2009. The slide appears to be a rockslide on a discontinuity dipping parallel to the slope. An interesting element of this is the clear signs of precursory activity - note the blocks on the road and the small slide that happens just before the main slide.

Tuesday, 10 November 2009

Reported landslide accident in SW India

AP is reporting that a landslide in S India has killed 39 people. The landslide occurred in the Ooty and Coonoor region of Tamil Nadu states, hitting 300 mud huts. The report quotes a "state flood control official" as saying that 14 bodies were recovered on Monday and a further 25 today.

Tamil Nadu is in the very south of India. The TRMM landslide warning system has this region as being likely or very likely to suffer landslides:

Monday, 9 November 2009

El Salvador landslide disaster

The nature of the landslide disaster in El Salvador caused by Hurricane Ida is now becoming a little clearer. The largest event appears to have happened in the town of Verapaz, where it appears that a debris flow hit and destroyed part of the town. The reported death toll is 16 people, with a further 47 thought to be missing.

There are two intriguing aspects to this landslide. The first is that the maps seem to show that Verapaz is some distance from the San Vicente volcano (also known as Chichontepec volcano) from which the landslide (probably actually technically a lahar) came, as the Google Earth image below shows:

So, I wonder whether the landslide actually affected this town - in which case it is very large indeed- or whether it was one of the smaller communities nearer to the volcano.

Second, this area was affected by an earthquake on 13 February 2001, which killed over 300 people. has an interesting comparison image:

This is clearly not the same street.

Acciording to this abstract, In that earthquake two landslides were triggered on San Vicente volcano (the volcano shown above). Part of the abstract states (note my emphasis):

"Large landslides dammed two major rivers, the Río El Desagüe and Río Jiboa, and two large slides occurred on Volcán San Vicente. The ±1.5-million-m3 Río El Desagüe landslide temporarily dammed the river and formed a shallow, 1.5-km-long lake, but the dam has been overtopped and is stable. The ±12-million-m3 Río Jiboa landslide blocked about 700 m of the valley with debris composed of poorly consolidated tephra, and the upstream lake could potentially have been as deep as 60 m and about 2 km long. A 20-m-deep spillway was excavated to decrease the maximum lake volume and reduce the possible catastrophic failure of the unstable landslide dam. On the northern flank of Volcán San Vicente, ±250,000 m3 of loose landslide debris filled the upper part of Quebrada El Blanco; remobilization of this material in debris flows could inundate part of a downstream village. On the volcano's northwest flank, ±200,000 m3 of lithified andesite blocks slid in the upper part of Quebrada Del Muerto, but this material will not likely remobilize and threaten downstream settlements.."

Verapaz is to the north of the volcano.

The presence of potentially dangerous landslides on the northern flank of the volcano is fairly clear:

Sunday, 8 November 2009

The role of landslides in coral reef destruction

Loss of coral reefs is a widely reported and serious problem, caused by a range of factors including changes in sea temperature and chemistry; pollution; fishing; development; and mining. Reefs are often also damaged by severe storms. It is is thus unsurprising to read this report of serious damage to the coral reefs at Orchid Island, to the south-east of Taiwan, as a result of Typhoon Morakot:

"Coral reefs off Taiwan will need up to 100 years to recover from Typhoon Morakot, which lashed the island in early August killing more than 600 people, a scientist said Tuesday...Some of the shallow-water coral reefs look as if they've been crushed by road rollers," said Chen Chaolun, a researcher at the institution. "They will need up to 100 years to recover."...The live coral coverage near Orchid Island has tumbled from 68 percent to less than 18 percent, he said."

However, the cause of the damage is a surprise:

"The reefs, popular with diving enthusiasts, were damaged by driftwood thrust into the sea by the typhoon and mudflows crashing into the ocean from coastal areas."

This driftwood, which accumulated on the beaches of Taiwan and even choked harbours in Japan, was released from the hillsides by the huge numbers of landslides triggered by the typhoon, and then carried out to sea by the rivers. The volume of driftwood was extraordinary, as this Guardian image shows:

I have not seen previous reports of landslide-released driftwood causing coral reef destruction, so this is an interesting new landslide impact.

Reports of a very large landslide tragedy at San Vicente volcano in El Salvador caused by Hurricane Ida?

It is becoming clear that Hurricane Ida has wreaked havoc in El Salvador over the last 48 hours, with many landslides. Over 100 people are known to have been killed, with the toll expected to rise over the next few days as more remote areas resume contacts. Intriguingly, the El Salvador news site has a report on the disaster here. It reports:

"Según los últimos reportes, en Verapaz, en las faldas del volcán de San Vicente, un alud ha cubierto un área de ocho kilómetros aproximadamente. El panorama es sombrío: Grandes rocas y árboles se observan por todos lados, por lo que la movilización es difícil. La tierra ha cubierto colonias enteras. Familiares y socorristas buscan desesperadamente a las víctimas. Testigos afirman que la tragedia es similar a la que ocurrió en Las Colinas, Santa Tecla, en 2001 cuando el país fue afectado por un terremoto."

This roughly translates as:

"According to recent reports, in Verapaz, in the foothills of the San Vicente volcano, a landslide has covered an area of eight kilometres. The outlook is grim: Large rocks and trees are found everywhere, so that mobilization is difficult. The earth has covered entire colonies. Relatives and rescuers desperately looking for victims. Witnesses say the tragedy is similar to what occurred in Las Colinas, Santa Tecla, in 2001 when the country was affected by an earthquake."

The Santa Tecla landslide killed about 600 people.

I have no way to validate or refute this report at this time, so let's hope that this report is erroneous. This is a Google Earth perspective view of the volcano in question, San Vicente:

I will post again when information becomes available.

Coastal erosion and climate change : Implications for the shoreline over the next century

On Friday I gave a public lecture at the Badbury Centre in North Yorkshire (NE England), looking at the potential impacts of climate change on coastal erosion. The presentation was focused on the field site that we maintain at Boulby, close to the lecture venue. In the presentation I start by looking at our landslide database and then go on to talk about climate change. The main part of the presentation is an examination of the likely impact of climate change, looking at the ways in which potential erosion rates have been determined, and then using our own very high resolution work to examine how this can be done properly. I hope that in the presentation I show that in this area at least the impacts of climate change on erosion are not as great as some might fear, and that proper scientific analysis allows one to develop proper models of future impacts.

I must stress here that this does not suggest that climate change is not a real and urgent problem - it is - and in the talk and in the questions afterwards I stressed the vital importance of a successful outcome to the Copenhagen talks.

The presentation is available via Authorstream below:

My personal Authorstream page, which contains a selection of my recent presentations, is available here:

Friday, 6 November 2009

New landslide video - houses destroyed in Brazil, November 2008

I have come across an impressive and dramatic landslide video, apparently from Brazil and taken in November 2008. The video quality is reasonably good:

Or you can go directly to the Liveleak web page.

Wednesday, 4 November 2009

The basic physics of landslides is the same everywhere

Whilst landslides come in all sorts of shapes and sizes, the fundamental physics that controls material deformation on slopes is the same everywhere. I was reminded of this truism this week when Andrew Giles pointed out a large rockslide that occurred a few days ago, blocking Interstate 40 at Pigeon Gorge in North Carolina, USA. The website has a nice report and images of the slide:

The images reminded me very strongly of the 2003 Bukit Lanjan landslide, thousands of kilometres away in Malaysia (image from CKC Malaysia):

Tuesday, 3 November 2009

The Willis Research Network - the world's most important hazard and risk collaboration?

Willis is a large insurance and reinsurance broker based in London. A key part of their primary business lies in arranging insurance for catastrophe risk - i.e. losses from mega-events such as an earthquake in Tokyo or San Francisco, a storm surge flood in London or a volcanic eruption near to Mexico City. Calculating the risks associated with these events is a challenging task, but of course the stakes are high as large events can induce catastrophic losses. Parts of the reinsurance industry was badly burnt by Hurricane Katrina for example. So, in order to calculate the risks and potential losses reinsurance companies use Catastrophe Models (usually called Cat Models), which are complex simulations of the impacts of large events. Getting these models to sensibly estimate potential losses is difficult - and of course requires a good knowledge of the science of the hazard in question.

A few years ago Willis approached one of my colleagues, Prof. Stuart Lane, and I to see if we would be interested in joining a research network that Willis would support. The idea was to bring together key parts of the reinsurance industry and top academic researchers on hazards and risk in order to improve the ways in which risk is modelled and handled in the insurance industry. And so the Willis Research Network was born. Initially the network consisted of seven very carefully chosen UK academic institutions - Durham, Bristol, Reading, City, Cambridge, Exeter and Imperial. Membership of the network is by invitation only, and active participation is ensured through the sponsorship by Willis of a research fellow in each institution.

A few years on, and the network is now an extraordinary entity. That initial group of seven has been joined by universities from the USA (e.g. Princeton and NCAR), Japan (e.g. Kyoto), Italy (e.g. Bologna), Germany (e.g. GFZ Potsdam) and Australia (e.g. SEA). I am pretty sure that it is now the largest and most dynamic academic-industrial hazard and risk network in the world, and it is generating some amazing research.

In the last couple of days the network announced associate membership of ten new institutions, including the British Geological Survey, the UK Met Office, The UK National Oceanography Centre, the Ordnance Survey and GNS Science in New Zealand. Research now spans natural perils, visualisation, social dislocation and financial management. I suspect that over the next few years this network will come to dominate research into catastrophic risk management. The publications section is well worth a look, not least because there are some very useful background presentations there from some of the world's top hazards researchers.

Sunday, 1 November 2009

Another fatal beach rockfall - this time in Spain

In August this year a rockfall on a beach in Portugal killed five people and injured four others. Unfortunately the newspapers in the UK are this evening reporting that a similar rockfall occurred today on a beach at Playa de los Gigantes in western Tenerife in Spain, killing a British woman and a local woman. The Telegraph has a report and picture (below) of the incident:

Coastal cliffs unfortunately are prone to erosion and thus to rockfalls. News reports suggest that access to this section of beach had been blocked due to an earlier event. Sadly, many people do not seem to appreciate that rockfalls can occur at any time , including in good weather.

A very surprising paper - movement of a landslide controlled by atmospheric tides

ResearchBlogging.orgJust occasionally a paper appears that makes me sit up with surprise, with a strong sense of "I did not see that one coming!". Just such a paper has appeared in pre-publication form in Nature Geoscience today - namely Schultz et al. (2009). This will undoubtedly represent the most surprising landslide paper of the year, and indeed for a few years in fact. I suspect that this paper will generate a fair amount of media interest in the next few days as well.

The subject of the paper is the Slumgullion landslide in SW. Colorado. This slide is a classic landslide site - it is large and moves almost continually. The slide is pretty obvious on Google Earth:

All the more so when viewed obliquely:

And there are a fair number of websites and online materials that describe it (see here and here for example).

This latest study is based on an intense monitoring campaign undertaken on the landslide using extension transducers. There is little doubt that the monitoring has been designed with great care by a team that is well-versed in such techniques. The team have used this movement data to examine the variation in movement through time. There is nothing terribly unusual about this approach except perhaps the level of precision with which the monitoring has been undertaken, but the results are very intriguing. The data suggest that the slide moves almost every day. This style of behaviour is quite unusual - most landslides are very episodic in terms of movement - but certainly not unique. However, examination the movement patterns on a daily basis suggested that the slide moved in distinct eight hour periods, essentially at night. During the day the landslide was to all intents and purposes stationary. This periodic movement of the slide was found to be correlated with the variations in air pressure associated with atmospheric tides - i.e. it appears that the movement of the slide is controlled by air pressure! Atmospheric tides are small variations in atmospheric pressure that result from solar heating of the air. The patterns of atmospheric tides are actually rather complex, but can be examined with a Fourier analysis.

The team have compared the Fourier analysis of the movement pattern of the landslide with that of the atmospheric tides - there is little doubt that they are remarkably similar. Thus the movement at least appears to be related to these atmospheric tidal cycles. The authors suggest that these low atmospheric tide conditions induce upward movement of water and air below the shear surface of the landslide, changing the "frictional stress" of the slide, allowing movement to occur. They have modelled the slide to explore this mechanism and present an analysis that, at least on first inspection, appears to be credible and to support their hypothesis.

Now, it must be stressed at this point that Slumgullion is not a typical landslide. In particular, the perennial slow landslide movement suggests that the factor of safety is essentially unity. Measurements suggest that the pore pressure on the shear surface varies very little on the landslide, meaning that the slide is potentially highly sensitive to effects such as that reported in Schulz et al. (2009). It is very unlikely that this effect is seen in anything other than a small proportion of slides that have exactly the right combination of stress state, thickness and material properties to allow this to occur, although I am willing to bet that a small number of other slides will be found to display this behaviour. However, as the authors of the paper hint, the most interesting aspect may well be the implications for landslides in areas that encounter conditions with very low atmospheric pressures. The key mechanism to deliver such conditions is the tropical cyclone, especially typhoons in East Asia and hurricanes in the Caribbean. Here, atmospheric pressures can drop very low (Typhoon Tip had a measured sea level air pressure of 870 hPa in 1970), as well as bringing high intensity rainfall of course. Thus, it may well be that the very high incidence of landslides under such conditions is not just the consequence of the rainfall but also the low atmospheric pressure. That is a very interesting observation indeed. Added to the recent observations that tropical cyclones can also trigger slow earthquakes, our understanding of the interactions between the atmosphere and the ground is developing quickly.

So, all in all this is a fascinating study for which the authors should be congratulated. It has added an intriguing new dimension to landslides, and will I suspect trigger a new wave of research projects. I should add here that for most landslides this effect is almost certainly very small (no-one living near a landslide should start worrying about atmospheric pressures - it is still rainfall and earthquakes that are the main issue), but for certain slides under certain conditions this may be an important effect.

Schulz, W.H., Kean, J.W and Wang, G. (2009). Landslide movement in southwest Colorado triggered by atmospheric tides Nature Geoscience : 10.1038/NGEO65