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

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

Thursday, 25 February 2010

Coastal erosion as art

The Guardian has a slightly bizarre article today. It is entitled "The art of watching your house fall into the sea". It tells the story of artist Kane Cunningham, who has bought a bungalow on the cliff edge at Knipe Point near to Scarborough. This site has been featured here before as recent landsliding has threatened to destroy a number of houses. He says in the article:

"I'm going to turn the moment my studio collapses into an art work: I've set up cameras to film it, and I've commissioned music and poetry to celebrate it. Both our houses punch a hole in what we think of as the value of ­property, and remind us of our moral and ethical responsibility to nature... When my studio disappears, I'll have no sense of loss – that will be its beautiful final act. The sooner it goes, for me, the better."

He has a website about this here. Quite bizarre, but it could be very interesting to watch.

Wednesday, 24 February 2010

Overview photo of the Bandung landslide

The Jakarta Post has an overview image of the Bandung landslide:

Framlingham College Presentation on the typhoon Morakot disaster in Taiwan

Last night I gave a public lecture on the Typhoon Morakot disaster in Taiwan. The talk may be viewed and downloaded here:

Massive landslide in West Java, Indonesia

A large landslide occurred yesterday in the Dewata tea plantation area in Pasir Jambu, Southern Bandung in West Java, Indonesia, burying the accommodation for temporary plantation workers. It is estimated that 70 people were buried by the slide. This image, from AP, appears to show the head of the landslide, which looks to be a large but comparatively shallow flow type failure in regolith:

The toe of the slide, including some of the impacted buildings, is shown in this AP image:

News reports suggest that to date 7 bodies have been recovered from the debris, with rescue operations continuing amongst heavy rainfall, which must place the workers at considerable personal risk.

Java consistently appears in the landslide database with very high fatal landslide occurrence. Tea plantations are also areas of high landslide hazard.

Friday, 19 February 2010

A landslide dam in Haiti caused by the earthquake

Thanks to Lynn Highland for the heads-up on this one. The French language website has an article describing a valley blocking landslide triggered by the Haiti earthquake. The translation says:

"The dam is large and there are still several tens of metres before the water passes over it. The problem is that it has hardly rained since the earthquake of January 12. When the rainy season triggers, a few weeks at most, the water will very quickly here...The dam is located in the mountains about ten miles upstream of Grand Goave, a city wedged between the sea and a semi-circle of mountains that lies ahead."

This is a Google Earth image of Grand Goave after the earthquake:

Thursday, 18 February 2010

Impressive roadside landslide in Pomona, California

Thanks to reader CConkle for the heads-up on this one. An impressive slide happened this morning by the side of Freeway 10 near to Pamona in California. The slide has completely blocked the road (pictures from the LA Times):

It is uite interesting that the weather does not indicate an obvious trigger for what appears to be a failure on a reprofiled slope. However, California has suffered recent exceptional rainfall in the last few weeks, so it may be progressive failure is the key factor here.

Most recent reports suggest that the slope is still moving and the road is likely to be closed for a week.

Tuesday, 16 February 2010

An update on the Italian landslide(s)

The large landslide in southern Italy that was graphically caught on camera (see the video in my earlier post) has generated a great deal of interest - indeed this site has had its busiest ever day. Details of the slide are still quite sketchy, but this is what we know so far:

First, the media are confusing two different slides in the same general area of Calabria - in fact there were about 100 altogether, but two are directly affecting towns. The one in the video occurred on the outskirts of Maierato, which is this town:

It appears that the slope had been moving for some days, and there are some indications that it was in distress well before this. The slide appears to have affected the area shown on this image, although I am waiting for clarification of this:

Interestingly, as Jonas van Rutte has pointed out to me, the roads near the headscarp appear to have been recently repaired, judging by the Google Street View images here. No-one was killed or injured in this landslide, although 2300 people have been evacuated from their homes.

The second slide occurred at San Fratello in Sicily, which is this town:

This slide appears to have damaged the centre of the town, leaving 1500 people homeless. The level of damage is high (image from Corriere della Sera):

Watch this! Extraordinary landslide video - Maierato, Italy

See update post here

The BBC has an extraordinary video of a landslide in Maierato in Southern Italy. More later, but for now the link is here:

UPDATE: Youtube has a longer version of the video below:

See update post here

Saturday, 13 February 2010

An analysis of fatal landslides in the Asia-Pacific region for 2006 to 2008

In my last post I published two maps of fatal landslide occurrence in the years 2006 to 2008 inclusive, based upon my long term fatal landslide database. In this post I focus on the Asia Pacific region. This analysis does not include seismically-induced landslides, most notably the Wenchuan (Sichuan) earthquake, which triggered a large number of slides, killing over 20,000 people.

The basic statistics of the data are in the table below - as per usual you can get a better version of the table and figures by clicking on the image:

In total, I recorded 796 fatal landslides (note that in this context I use landslides generically to include all non-avalanche mass movements, including rockfalls) in which 9941 people were killed. In terms of fatalities, the Philippines ranks highest (Fig. 1), although India is the country with the most fatal landslides (Fig. 2). Of course if the Wenchuan earthquake were included China would be the top of both lists

Fig. 1: The number of recorded fatalities organised by country

Fig. 2: The number of recorded fatal landslides organised by country

The seasonality of landslide occurrence varies greatly within this area. In South Asia there is a very strong influence from the S. Asian monsoon, which is very apparent in the monthly data (note that for these three graphs I have used the same y-axis scales so that they are directly comparable):

Fig. 3: Monthly recorded fatal landslide occurrence (line graph) and loss of life (bar graph) for South Asia

In East Asia there is also a very strong seasonal signal, but note that here there is a more distinct peak in July and a decline thereafter (Fig. 4):

Fig. 4: Monthly recorded fatal landslide occurrence (line graph) and loss of life (bar graph) for East Asia

In South-East Asia there is no strong seasonal signal - this is unsurprising in an area that is mostly tropical. Two distinct peaks do occur, one in February and one in November.Fig. 5: Monthly recorded fatal landslide occurrence (line graph) and loss of life (bar graph) for South-East Asia

As with yesterday's post, I have no problem with this information being used elsewhere, but please reference the figures and information as follows:

Petley, D.N. 2010. An analysis of fatal landslides in the Asia-Pacific region for 2006 to 2008. Dave's Landslide Blog URL:

Friday, 12 February 2010

Maps of global fatal landslides

NB there is an alaysis of the data in thiese maps in this post:

I have recently worked up two new maps of the distribution of fatal landslides, based upon the fatal landslide database that I maintain. This first map shows the global distribution of non-seismic landslides for the period 2006-2008. Each dot represents a single fatal landslide. The colours indicate the year:

Red = 2008
Blue = 2007
Green = 2006

The background image is the ETOPO digital elevation model, with the darker colours indicating higher ground (click on the image for a better view in a new window):
There are a whole host of interesting aspects of this map. First, note the clustering of the landslides in areas that are tectonically-active (e.g. Western S. America, the Himalayas, Indonesia, etc). Second, note how the patterns change from year to year (e.g. take a look at the south coast of China - this is related to typhoon landfalls). Finally, note that the vast majority of recorded fatal landslides occur in Asia. This is shown better by the second map, for Asia only:

Please feel free to use the information and figures on websites, reports, theses etc, but please reference this as follows:

Petley, D.N. 2010. An analysis of fatal landslides in the Asia-Pacific region for 2006 to 2008. Dave's Landslide Blog URL:

Wednesday, 10 February 2010

Continuing works on the Attabad landslide

The National Disaster Management Authority in Pakistan is posting daily updates on the situation with the Attabad landslide on its website here. The reports appear with a lag of about 4 days, but nonetheless provide a helpful insight into the rate of progress. As of 6th Feb a total of 26,000 cubic metres of debris had been moved to create the spillway. Meanwhile the lake level is increasing at 60 cm per day - it will be interesting to see if this rate decreases with the recent weather.

This impressive rate of progress appears to be possible because the area that they are excavating is very fine-grained, as this video shows:

Meanwhile, the ever helpful Pamir Times reports that the river blockage is starting to have substantive impacts downstream due to the reduced flow levels. These impacts are clearly most acute directly below the dam, where the river is now dry, but extend right down to the Tarbela dam, close to Islamabad.

Tuesday, 9 February 2010

Problems at the Hattian Bala landslide dam in Pakistan

The largest landslide in the 2005 Kashmir earthquake occurred at Hattian, 25 km or so from Muzaffarabad:

My colleagues and I wrote a paper on this landslide a few years ago (you can download the pdf here). It was notable for the fact that it blocked the valley - in the aftermath the Pakistan Army built a spillway to ensure that uncontrolled overtopping did not occur. The landslide has been stable since, although many of us have warned that this area is far from out of the woods, with the first really intense precipitation event being a key test.

In the last few days this area has suffered a very heavy rain and snowfall event - as in fact has a vast area of the Himalayan Arc - indeed I have never seen a 1 day TRMM landslide potential map like this - it is quite bizarre:

Late tonight, media sources in Pakistan are suggesting that the lake has broken its banks, flooding 23 houses. So far the dam is holding, but clearly this event will ask substantial questions of it. Hopefully, the spillway engineers have got this right!

Elsewhere in the Muzaffarabad area, heavy rain is reported to have triggered many landslides. I have not yet seen any reports of the situation at Attabad, but this is not likely to be helping.

Sunday, 7 February 2010

Mudslides occur in the areas affected by the 2009 Station Fire in Los Angeles

The Station Fire was a huge forest fire that occurred on the northern edge of Los Angeles in late 2009:It was finally brought under control on 16th October 2009, by which time it had burnt a huge area, thought to be over 160,000 acres (image from NASA):

In the aftermath of the fire considerable concern has been expressed about the potential for damaging mudflows in the burnt area, magnified by the current El Nino conditions, which typically bring very heavy rainfall to California at this time of the year.

Early on Saturday, unexpected heavy rainfall triggered a series of mudflows from the burnt area that affected the northern part of Oceanview Boulevard in La Canada Flintridge, causing extensive damage to properties but fortunately no fatalities. The most seriously affected area appears to be Manistree Drive, which is shown on this Google Earth image:

This is apparently the location of this image, from the LA Times:

At least 43 houses were damaged, together with 25 cars (see image below from AP). A dozen houses have reportedly suffered major structural damage.

The Attabad landslide - caught between the devil and the deep blue sea

News from the landslide site at Attabad is increasingly grim, despite the ongoing work to create a spillway. The Pamir Times has an excellent article here outlining the impact of the rapidly expanding lake. This lake is now 11 km long and has inundated 460,000 square metres of land, 11 houses and 3.5 km of the Karakoram highway.

However, the spillway that is under construction is a further 79 m above the current lake level, meaning that at its maximum point the lake is expected to inundate 7.55 million square metres of land, including 187 houses, the homes of 1,736 people, and 25 km of the Karakoram Highway. The next problem is expected to occur at the bridge between Shashkat and Gulmit, which the lake has recently reached (see the image below, from the Pamir Times):

To give you an idea of the size of the lake - this Google Earth image shows the location of the landslide dam and the bridge (click on the image for a better view in a new window:

The next crisis will arise when the bridge is inundated as at this point the communities along the Karakoram Highway between the bridge and the dam will be isolated. The Pamir Times article suggests that there are 3000 people in this area, which is probably correct based on the images:

All of this emphasises the magnitude of the problems in this area, which are receiving scant attention from the outside world. When the water reaches the top of the dam it will have flooded a very large area. If the dam remains intact, a huge reconstruction programme will be needed for the Karakoram Highway, and a huge number of people will be displaced. This image shows the location of Hussaini village, which is where the article estimates that the head of the lake will be when the spillway is reached:

On the other hand, if the dam is eroded away then an immense flood will travel down the Hunza River, endangering a very number of people. The 1856 flood caused damage down as far as Attock, which is 360 km away as the crow flies! Unfortunately, the possibility of a sudden collapse of the dam cannot be ruled out as the lake fills.

Thursday, 4 February 2010

The Rivermist subdivision in San Antonio, Texas is on the move again - and it is a rotational retaining wall failure

The woes of the families displaced by the slope failure at the Rivermist Subdivision in San Antonio, Texas have been exacerbated by a new phase of movement. AP Texas News reports that heavy rain has triggered additional cracking of the wall, and the movement of some debris. However, the houses at the top of the slope are apparently not moving.

I have been wondering why they are so confident that the wall won't fail. In an earlier post I suggested that this is not a simple retaining wall failure, given the toe scarp - a couple of subsequent commenters agreed. These aerial images, from My SA news show that this is the case:

The key part of the image is actually at the bottom of the first image above, although once you have your eye in you can see these features in the other image too. This is an annotated version of that first image:

This landslide is clearly not a simple retaining wall failure. It is a rotational slip in the slope that has caused the retaining wall to fail. It is not surprising that they do not expect the wall to collapse during these rains as the wall is being rotated to lean backwards by the failure.

Of course, as per the comment below, ensuring that the retaining wall would not be affected by this type of rotational failure should always be a key part of the design process for such structures. I suspect that this will become a textbook example of the problems that can affect retaining walls.

Tuesday, 2 February 2010

The causes of the Shiaolin landslide disaster in Taiwan

The Shiaolin landslide disaster in Taiwan, which occurred during typhoon Morakot last August, has been the source of huge controversy. To recap, the landslide, which occurred during an exceptional rainfall event, wiped out Shiaolin village, killing about 500 people. The controversy centred on two key aspects - first, the perceived very slow response of the government to the disaster and second the possibility that tunneling associated with the Zengwen Reservoir project may have been a contributing factor to the slope failure. In response the Executive Yuan of Taiwan commissioned an investigation from the Public Construction Commission, which released its final report yesterday.

The report is of course in Mandarin, but very helpfully there is a powerpoint file available that summarises the findings and provides some illustrations of the key issues. This powerpoint file is available here (warning it is a large file in Powerpoint in pptx format):

The report is available here:

The key finding of the report is in my view correct - this is that the tunnel project was not the cause of this landslide disaster - they factor was the exceptional rainfall experienced in this event. The powerpoint file provides a dramatic illustrations of the magnitude and intensity of this rainfall:

Click on the image for a better view in a new window. The map on the left is the recorded rainfall for the storm, the table on the right is the total rainfall for a number of stations in the worse affected area. Note that the precipitation totals are extreme in every sense of the word - c.2500 mm (2.5 metres of rain) is the equivalent of three years total rainfall for the temperate area in which I live. This is the largest rainfall event ever recorded in Taiwan, and probably the most intense rainfall event worldwide for half a century.

The report shows that disturbance associated with the tunnel is not sufficient to be a factor in the landslide - a conclusion that I support. Instead, they show that the slope underwent a dip slope failure that led to a massive rockslide that destroyed the village. The report suggests that the landslide had a maximum depth of about 86 m and a volume of 2.5 million cubic metres. From what I can tell the slide itself was a wedge failure with a dip-slope defining part of the wedge.

There is only one aspect of the report that continues to cause concern. This is the interpretation of the mechanism of failure. This slide shows a long profile of the landslide site, which shows bedding parallel to the slope right down to the river (section A-A'):

This just doesn't seem to accord with what Chris Massey and I observed on site at the toe of the slope:

This picture is taken from the north end of the toe of the slope looking upstream - note the bedding on the far side of the valley - this is near vertical.

This picture was taken at the site of the old bridge abutment at the toe of the slide (the concrete in the middle of the image is this abutment I think) - again, note the very steeply dipping rocks at this point.

The final point to make is that the Shiaolin landslide was of course not the only failure to occur in the area during Morakot. Mapping of this region has identified 880 landslides covering an area of 2058 hectares (20.88 square kilometres).

There can be no doubt that Morakot was an extraordinary event.