28 May 2015
The 12th International Symposium on Landslides (12 ISL) will be held on 12-19 June 2016 at the Royal Continental Hotel in Naples, Italy. This is a symposium jointly organized by FedIGS, ISSMGE, ISRM, IGS, JTC1. The previous symposia were as follows:
- 11th in 2012 in Banff, Canada (this is the meeting at which I gave the opening keynote lecture)
- 10th in 2008 in Xi’an China;
- 9th in 2004 in Rio de Janeiro in Brasil
The organisers of the meeting have selected “Experience, Theory, Practice” as the key theme.
- Weather-induced landslides
- Earthquake-induced landslides
- Volcanic activity and landslides
- Man-induced (I assume that this means human-induced!) landslides
The Scientific Committee is encouraging authors to submit abstracts dealing with one or more of the following topics:
- Soil and rock properties
- Investigation and slope monitoring
- Mechanisms and mechanical aspects
- Analysis and modeling
- Risk analysis, assessment and management
- Control works
My view is that the International Symposium on Landslides remains the most important landslide meeting in the calendar. I have attended the last two, both of which have been both scientifically fascinating and hugely enjoyable. I strongly recommend it!
27 May 2015
Wind River Canyon in Wyoming
Unusually heavy rainfall in Wyoming on Sunday triggered a series of landslides that has led to the closure, probably until Wednesday, of the US20 highway between Thermopolis and Shoshoni as well as the BNSF railway line.
The best coverage of the event is on the County 10 website, which has a splendid gallery of images of the landslides. This image for example is both beautiful and technically interesting:
These are not simple landslides by any means. In all three cases they involves water that has been channelised on the steep slopes, from where it has collected debris, and which has triggered the detachment of boulders from the bluffs. The flows have transitioned into true debris flows on the fan below the bluffs. Note that in most cases the flows have eroded sediment from the fan in the steeper portion, adding to their volume. At the point at which the gradient of the slope reduces, just above the railway line, the flows have stopped eroding and have started to deposit, with the larger boulders being left higher on the slope and finer material further down. This is classic debris flow behaviour.
Other failures show slightly different behaviour. This one for example appears to have been erosive all the way to the river, probably because these was no reduction in slope gradient all the way down to the floor of Wind River Canyon:-
Google Earth gives a better understanding of the nature of these landslides. In fact above the bluffs each of these gulleys has quite a large catchment of more erodible material:
These upper catchments are likely to be the source of much of the sediment that forms the debris fans at the slope toe and that comprises the solid part of the debris flows.
26 May 2015
Kali Gandaki: an interesting landslide dam and breach
Early on Sunday a landslide occurred at Ranche in Myagdi District of Nepal, blocking the large Kali Gandaki river. The area affected by the landslide was affected by the earthquake of exactly a month ago, but interestingly there appears to have been no trigger. In this case the response of the authorities appears to have been exemplary, with downstream settlements, most notably the town of Beni, being evacuated and an assessment team being flown by helicopter to the site. The dam breached about 15 hours later, releasing a significant flood that caused damage but no loss of life.
The best source of information throughout was the twitter account of @gocoolchhetri, even though it is in Nepali (and Google translate does not do a good job on Nepali to English). But from the start he tweeted a series of photographs of the blockage on the Kali Gandaki, including this one of the landslide dam itself:
Note the large amounts of dust from ongoing landslide activity. This one showed the lake as it filled:
This one showed the top pf the landslide dam as the water level rose. Note the fine grained material mantling the top of the blockage:
And the dam from above:
And then the overtopping initiating:
And then developing rapidly:
Ekantipur has an article today on the aftermath of the landslide, which states that:
A 2km long and 100 metre deep lake was formed after the landslide dammed the river. The site is about 9km north of Beni, the district headquarters of Myagdi. Massive rocks started falling past Saturday midnight, damaging 27 houses in Bhagawati and Ramche VDCs. Bhimsen KC, a local of the landslide-buried Baisari, said the dam is about 40 metres high from the riverbed. He believes there is no possibility of an outburst of the lake as the dam is about 100 metres wide.
And it includes this image of the aftermath of the landslide that blocked the Kali Gandaki:
This interesting event serves once again to highlight the challenges that face Nepal in the months ahead.
21 May 2015
A landslide in Guiyang, China
A landslide yesterday morning in the Yunyan District of Guiyang, the capital city of Guizhou demolished completely an apartment building that was home to 114 people. Of these people, eight bodies have been recovered whilst eight remain missing. The authorities report that their mobile phone signals have been traced to the rubble.
It is clear that the building collapse at Guiyang was triggered by a landslide, caused by heavy rainfall. Some of the images show the collapsed block with the landslide behind. Both of these images are from Xinhua:
Judging by the images it appears that the building was connected to another, which fortunately did not collapse.
Such events are rare, but not unknown. The two best known examples are:
- The 1972 Po Shan Road landslide in Hong Kong, which killed 67 people. I featured the animation of this landslide in a previous post.
- The 1993 Highland Towers landslide in Taman Hillview, Kuala Lumpur, which killed 48 people
20 May 2015
Timing the landslide season in Nepal
In a previous post I highlighted the threat posed by landslides in the SW monsoon in Nepal. A key question now is the timing – i.e. when will the rainy season (i.e. the landslide season) start? The IMD has a tracking map for the advance of the monsoon – at the time of writing the monsoon front is lying to the SW of India in the Andaman Sea, and the advance looks to be about normal for the time of year:
Based upon the timings on this map the rains might be expected to reach the eastern part of Nepal on about 5th June, and to have covered the entire county by about 20th June. These are average figures of course, so there may be considerable variability. But of course the arrival of the monsoon front is not necessarily the start of the landslide season. Landslides can be triggered by pre-monsoon rainfall (especially convective events), and the heavy monsoon rainfall may not be associated with the arrival of the monsoon front. So I have been looking at my Nepal landslide database. In the graph below I have plotted by year for the period from 2004 to 2014 inclusive the cumulative number of fatal landslides in Nepal:-
Today (20th May) is day number 140, so it is clear that the main landslide season has yet to start. In most years the number of fatal landslides starts to increase at around about day 160 (about 8th June), with the main focus starting from about day 180 (28th June). So the main threat will start in about three weeks from now, and will intensify towards the end of next month. The landslide season typically ends at about day 280, i.e. somewhere around early October. Note that there is a great deal of variation between years both in terms of timing and the number of events.
I have highlighted two curves – 2006 and 2009. These were both years that represented the start of El Nino events – a medium-sized event in 2009-10 and a weak event in 2006-7. Interestingly both years saw a later than average start to the landslide season. We are currently in weak El Nino conditions. I am unsure as to whether this is significant.
19 May 2015
Salgar, Antioquia, Colombia
Early yesterday morning, a landslide struck the small town of Salgar in Antioquia, Colombia. Latest reports (in Spanish) are that at least 63 people were killed. At least 37 people have been injured and an unknown number are missing.
The event appears to have been a debris flow, triggered by heavy rainfall, that tore through the small town in the floor of the valley. Although the exact location of this event is not yet clear, the Salgar area is characterised by villages lying in the floor of valleys in quite steep upland areas:
Images from the scene – there is a good gallery on the El Tiempo website – suggest that the flow may have been rapid and catastrophic:-
18 May 2015
Landslides in Nepal and the SW Monsoon
As Nepal moves out of the emergency phase of the earthquake disaster and starts to transition into the recovery and reconstruction phases, attention is now turning to the potential impacts of the SW (summer) monsoon period. In particular, there is now real concern that the heavy rainfall that the monsoon brings may reactivate many landslides, and trigger a host of new failures as well. Based on experience from elsewhere, these concerns are valid.
Not entirely coincidentally, back in 2007 I wrote what is perhaps the definitive paper on the triggering of landslides by the monsoon in Nepal, based on a database of landslides that we constructed as part of the DfID Landslide Risk Analysis project. This paper, Petley et al. (2007), is available for download for free from the Natural Hazards journal website.
In the paper we demonstrated three key aspects of landslides in Nepal:
1. The SW Monsoon is the dominant factor:
This graph from my paper shows the monthly occurrence of landslides that cause loss of life in Nepal (the data is for 1978-2005 inclusive):
The trend is clear – landslides occur almost solely in the summer period, with low incidence in May, a few more in June (the data show that these are mostly in the last week of the month, and then increasing dramatically in July and remaining high for three months. The threat in the months ahead is clear.
2. The earthquake affected area is extremely landslide-prone:
Map A below shows the distribution of landslides that caused loss of life over the same period, plotted in terms of number of landslides per unit area:
Map B is the distribution of rainfall. It is clear that landslides are most serious in a swathe of Hill Districts in Central Nepal. This starts to look deeply worrying when compared with a map of the areas most seriously affected by the earthquake:
The coincidence between the area of intense landslides and the area affected by the earthquake is deeply worrying.
3. Nepal landslides are at their worst when the SW Monsoon across the whole region is weak.
This may seem like a perverse relationship, but we found that the worst years for landslides in Nepal occurred in years when the SW Monsoon for the whole region (as indicated by SASMI – an index of monsoon intensity) was low:
The reason that the data split into two groups is that the later values (with far higher losses from landslides) occurred after the start of the spate of road building in the mountains of Nepal, after which the occurrence of landslides increased dramatically. So, we might expect the landslide problem in Nepal to be particularly serious this year if the SW Monsoon across the whole region is particularly weak.
Petley, D.N., Hearn, G.J., Hart, A., Rosser, N.J., Dunning, S.A., Oven, K., and Mitchell, W.A. 2007. Trends in landslide occurrence in Nepal. Natural Hazards 43:23–44. Download the PDF from here (no cost).
17 May 2015
A massive ice – rock avalanche in Akto County, Xinjiang, China?
The Chinese new agency Xinhua has some rather vague reports this morning of a massive “glacier collapse” in Akto County, Xinjiang, China. The reports in the English version are somewhat hard to interpret:
A huge body of glacier has collapsed in China’s Xinjiang, burying more than 100 cattle, crushing dozens of houses and damaging 1,000 hectares of grassland, local government said on Saturday. No casualties has been reported in the Kongur Tiube glacier slide that occurred in the Akto County of the Kizilsu Kirghiz Autonomous Prefecture. About 60 families were evacuated after the 20-kilometer-long and one- kilometer-wide glacier devoured their land.
But a catastrophic collapse that generated a slide 20 km long and one kilometre wide is likely to be an ice – rock avalanche, or perhaps a more conventional landslide. IFENG has a more detailed report in Chinese, suggesting that the volume of the collapse might be 500 million cubic metres, and that the deposit is 20 – 50 m think. If so this is a truly enormous event. This report includes an image, but it is far from clear what this represents:
The IFENG report also includes a video, but again it is unclear what this is showing. It does include a sequence from which this still is taken:
This could be an ice-rock avalanche deposit perhaps?
This is the Google Earth image of the area affected of Akto County by this landslide:-
This area is in the far northwest of China, near to the border with Tajikistan and Kyrgyzstan, so it is very unlikely that this event is associated with the Nepal earthquake sequence. Does anyone have any further information on this landslide? It appears to be a very unusual and interesting event.
16 May 2015
The second Nepal earthquake
The M=7.3 second Nepal earthquake on Tuesday released only about 13% of the energy of the first event. but the location and shallow depth meant that it still had the potential to generate landslides, albeit over a smaller area. The region affected by the second Nepal earthquake had already been affected by landslides triggered by the Gorkha earthquake, and this is one of the most landslide-prone regions of Nepal.
Digital Globe continues to release imagery of the earthquake-affected area on their online platform, for which we are extremely grateful. They have now collected some images of the areas affected by the second Nepal earthquake, most notably along the Araniko Highway, which is the main link between Nepal and Tibet. This image was collected on 13th May – a day after the second Nepal earthquake. The image quality is not great simply because there is so much dust in the air – and indeed the ground and buildings seem to be draped in a layer of dust, all generated by the landslides. But once you get your eye is you will see that there is massive landsliding across this area – indeed far more than was the case in the first earthquake:
This is what the same area looked like before the earthquake (from Google Earth). It is almost unrecognisable:
The yellow markers are the larger landslides mapped by the Durham / BGS team after the Gorkha (25th April) earthquake. This is what the same area looked like after that event:
The level of landsliding has clearly dramatically increased, and the road and associated infrastructure is very badly damaged. The landslides triggered in the second Nepal earthquake are characteristic of coseismic slides – they seem to be mostly shallow, disrupted rockslides, many of which originate from the ridge tops or other breaks of slope. The reasons why the second Nepal earthquake has generated so many landslides compared with the first event are not clear. This will need further study in the weeks ahead.
There can be little doubt that this area is going to be extremely difficult in heavy monsoon rainfall. The monsoon is a little more than a month away. In that context, the decision to refuse permission for the three Royal Air Force Chinook helicopters to operate in Nepal looks utterly absurd. Having witnessed first hand the amazing role that these helicopters played in the aftermath of the 2005 Kashmir earthquake, where the damage was so much worse, the refusal of the Nepal Government to allow these machines to operate is absolutely crass. In particular, the ability of these helicopter to carry large volumes of material in underslung loads was crucial in Pakistan. It is just crazy to stop that happening in Nepal.
12 May 2015
Sorkjosen (Sørkjosen) underwater landslide
Away from Nepal for a moment, a very interesting landslide occurred at the town of Sorkjosen (Sørkjosen) in Finnmark, Norway on Sunday. This appears to be the location as shown on Bing Maps:
There is a good archive of images of the aftermath of the landslide on the NRK.no website (the article is in Norwegian):-
The landslide appears to have taken out an area of fill on the edge of the fjord and a large part of the jetty, plus some buildings, though fortunately no houses. If you compare the photographs with the Big Maps image the size of the Sorkjosen landslide becomes apparent. The timing of the landslide, in the early hours, meant that none of the buildings were occupied. The landslide has also effectively cut Norway in half as the highway through the town – the E6 road – has been closed. The diversion route is 700 km long via Finland (see this translation of the Norwegian article).
Interestingly, the images suggest that close to the crown of the landslide there was a large ongoing construction project on the highway, including the construction of a new tunnel (as seen in the image above). There is some speculation in the various articles that the landslide might have been associated in some ways with these works. Indeed, this article suggests that local fishermen have been concerned about this site for some time because of the dumping of spoil from the tunnel. The concern now will be to ensure that the landslide is not likely to retrogress and remove the road itself. In the longer term, it will be interesting to find out the causes of the landslide, and the mobility of the slipped mass.
This landslide is quite reminiscent of the Nord-Statland landslide, also on Norway in January of last year.