Three landslide-induced railway incidents in a single day

News today of three different incidents from around the world:
1. New Zealand
NZ Herald reports that a milk train hit a landslide in Manawatu Gorge.  Fortunately there were no injuries, or even a need to cry over spilt milk...



2. India
Bangalore Mirror reports that three coaches of the Thiruvananthapuram-Mangalore Express train were buried by a landslide in a narrow cutting at Mulunthurthy.  Three people were injured, plus there were some minor injuries during the evacuation.

3. Canada
Various Canadian newspapers report that there was a serious landslide-induced derailment of a freight train at St-Lazare in western Quebec.  The accident trapped the injured train crew in their cab, requiring that they were rescued by firefighters.  Heavy rainfall was reported to be the trigger.  The images of the site, from the Montreal Gazette, are impressive:

 

 Landslides represent  an important risk to railways in upland areas, and where extensive earthworks have been used.  Railway companies expend huge amounts of resource mitigating the threat, usually with success.  Incidents are quite rare, but previous examples described on this site include:
May 2010: 19 killed when a train struck a landslide in China;
April 2010: Nine people killed when a train was struck by a landslide in northern Italy;
July 2009: Four people killed when a train struck a landslide from a cutting in China;
Dec 2008: A train was struck by a flowslide formed from power station ash in Tennessee, USA;

Remarkable NASA imagery of catastrophic flooding around Manchhar Lake in Pakistan

NASA have provided the most dramatic evidence yet of the catastrophic floods that are occurring around Manchhar (Manchar) Lake in Pakistan (see my post yesterday on this issue).  This image was captured by the ALI instrument on 18th September:

Compare this with an image taken of the same area just three days earlier:

Or the Google Earth image of the same area, which was collected in May this year:

Geo-TV, an independent TV station in Pakistan, reports that some of the breaches that are causing these floods are now 1 km wide, whilst NTDTV notes that the number of people displaced by this most recent component of the floods may be as high as 250,000. 


The breaches have allowed the water level in the main body of the lake to fall slightly, bringing some relief elsewhere, but with new areas continuing to experience flooding that is not cause for much overall cheer.

Pakistan flood update

The Pakistan flood crisis drags slowly on, and probably has at least a month to go before the waters fully recede.  Although the news coverage of the event is now little more than a drumbeat in the background, huge numbers of people are still being affected for the first time by this event.  The current crisis is focused on Mancchar Lake, which has been the destination of the waters that have travelled down the "ghost" parallel water course to the west of the main Indus channel, as shown very clearly by this NASA image:

The key iprblem at Manchhar has been that the waterways that allow the lake to drain are inadequate for the task, which has allowed the lake level to rise, causing new flood damage.  The reasons for this are clear from the following Google Earth image, taken before the flood:

In recent years the lake has suffered from inadequate inflows and serious salinity problems, to the extent that last year water was diverted into the lake from the Indus to improve water quality.  This of course shows that the need to have adequate drainage to allow large inflows to be released has not been a pressing issue of late. Thus, the two channels lining the lake to the Indus are too small to deal with the inflow:

In order to drain the excess water in the lake, the levees have been breached in eight places (I am unsure whether these breaches are natural or artificial), but the water level is reportedly still rising, and according to the UN 100,000 people have been displaced in 215 villages.   

Meanwhile the World Food Program have produced a report, available online here, that documents the scale of the impacts of the floods to date.  There are a number of useful aspects of this report, not least the following map that shows the extent of the floods:

Note that the data notes that there is no data for the northern regions, which of course were also very badly affected.  The upshot is that the statistics in the report under-represent the true impacts.  Nonetheless the statistics are eye-watering:

• 14.1 million people directly affected
• 392,786 damaged houses.
• 728,192 destroyed houses
• 7,600 destroyed schools
• 436 health facilities damaged or destroyed
• Overall production loss of sugar cane, paddy and cotton is estimated to be 13.3 million tonnes
• 2 million hectares of standing crops were either lost or damaged.
• 1.2 million head of livestock (excluding poultry) lost
• 14 million livestock are at risk due to fodder shortages and heightened risk of disease.

Of course it is important to consider these impacts in the context of the pre-existing situation in Pakistan.  This map, also from the WFP report, shows the state of food security in most of Pakistan before the floods.  Much of northern and western Pakistan were considered to be extremely food insecure. The central swathe of Pakistan most seriously affected by the floods was in a better (though in most cases not ideal) food security state:

The effect of these flood will be to plunge a large part of Pakistan into a state of extreme food insecurity through the coming winter, at a time when houses, infrastructure and health facilities are seriously compromised. 

Tangjiashan again - and a possible new Chinese flowslide

Back in 2008 I dedicated a great deal of space on this blog to the extraordinary efforts by the Chinese Army to draining the landslide lake at Tangjiashan, just above the town of Beichuan, which was created by the May 2008 Wenchuan Earthquake. 

 These efforts were ultimately successful, but in my visit to the site in Spring 2009 it was clear that a threat remained at the site in the form of another block of material that was showing signs of deformation. Over the last few days this area has received very high levels of rainfall.  Yesterday, Xinhua reported that a 300,000 cubic metre block has detached from the scarp above the barrier, and blocked the river to a depth of 10 metres:

"More than 6,200 residents were relocated Tuesday as torrential rains pounded Beichuan Qiang Autonomous County in southwest China's Sichuan Province, local authorities said Tuesday...Some 300,000 cubic meters of debris brought by the landslide caused a dam, blocking the lake's outlet. The dam's lowest point is 10 meters higher than the present water level, the statement said.  Debris still continue to come down from the hills, and if there were more rains, then the lake level would further rise, threatening the lives of people in nearby townships.The rains had disrupted the normal life of 58,000 local residents in the county, causing huge economic losses, the statement said."

China is well-versed in dealing with these hazards, but given the magnitude of the destruction in the Beichuan area, such events must cause great heart-ache.

It is clear that the elevated level of landslide activity in the aftermath of the landslide is a major issue.  I am travelling to Chengdu on Sunday, so will see whether I can ascertain more information about these issues.

Meanwhile, Bloomberg reports a probable flowslide failure in a tin mine in Guangdong yesterday:

Zijin Mining Group Co. said a dam built to hold tin-mining waste collapsed in China’s Guangdong province following torrential rain, less than three months after one of its copper mines leaked toxic waste into a river.About 60 centimeters (24 inches) of rain from Typhoon Fanapi and mud and rock slides triggered the accident at the company’s Yinyan tin mine at about 10 a.m. local time today, Shanghang, Fujian province-based Zijin said in a statement. 

China has been impacted by a series of these events in recent years, including one that caused multiple fatalities in 2008.  There appears to be a strong need to improve the safety of these facilities before another major accident occurs.

Gifts and Perils of Landslides

Ken Hewitt has written a wonderful article for Scientific American entitled "Gifts and Perils of Landslides", in which he examines the inter-relationship between the development of society and the occurrence of landslides in the Upper Indus valleys.  Ken is the guru of high mountain landslides in Pakistan, having spent many field seasons mapping rock avalanche deposits in the remote upper valleys of the Hindu Kush.  The article is available online at the following link:

http://www.americanscientist.org/issues/feature/2010/5/gifts-and-perils-of-landslides

His key point is that these giant landslides create both destruction and benefits to humans in this very rugged topography, the latter because they create terrain that is fertile (e.g. on lake beds formed behind landslide dams) and less steep. 

The piece will be accompanied in due course by a slideshow, which will be online here.  This is not yet available.

Meanwhile, the Pamir Times has an image of restarted works on the Attabad landslide, the aim of which is to widen and then, I understand, to deepen the spillway:

Landslides in Art Part 7 - Jennifer Williams

Regular readers will know that I occasionally highlight a piece of art featuring landslides.  This is usually a piece of visual art (such as this and this), but occasionally it is a song or even an installation.  This time I thought I'd feature a piece of art by the American artist Jennifer Williams, who paints environmentally-orientated pieces using acrylic on birch panels.  Jennifer works mostly in the Pacific Northwest, where both the landscape and the sky is large and colourful.  This is reflected in her portfolio of work, which resonates strongly through its depictions of the landscape of hazards. 


I'd like to highlight here two pieces.  First there is a painting called lahar, created as part of a portfolio encompassing the natural regeneration of the Mount St Helens area:

Second, and for me more strikingly, is a piece entitled simply "Landslide", painted in 2008, and visible through the Waterworks Gallery website:

To me this captures the chaos and ruin of the catastrophic earthflows we see in the Himalayas and other high mountain environments during heavy rainfall.  Amazing stuff.

Earthquake damage in Christchurch - an ironic billboard

This billboard, which predates the Darfield earthquake, was attached to a building in the Central Business District of Christchurch in New Zealand:

Do you think this is what they had in mind when they described the "open plan office with balcony access" and "cool funky office environment"?:

Images of the Darfield (Canterbury) earthquake fault rupture

Yesterday I was exceptionally fortunate to be able to spend the morning looking at the surface expression of the fault responsible for the Darfield earthquake, which hit the Canterbury plains area of New Zealand a fortnight ago.  I was kindly guided around by Russ Van Dissen of GNS Science, and the visit was organised by Chris Massey, also of GNS - so many thanks to them.  This GNS map shows the surface expression of the fault across the Canterbury Plain:

These pictures will work from east to west.

The east-most expression of the fault occurs close to the town of Rolleston, where a railway track crossed the fault.  The rails have been repaired, but the kink in the formerly straight railway track is fairly obvious.

Just around the corner a road crosses the fault as well.  Here the damage, again to the formerly straight road, is unrepaired as the fault is expressed as a broad deformation zone accommodating about 60 cm of movement:

If you are struggling to see the feature, remember that the road stretching into the distance (not the bit at the very bottom of the picture) was straight before the earthquake.  You should be able to see that from the middle of the image the road has been shifted to the right.  In this area there is little evident vertical change, bearing in mind that the surface was not perfectly flat before the earthquake.

A few kilometres to the west there is another formerly-straight road crossing the fault.  Here we see a little more lateral movement - actually about a metre.  Again, there is no vertical movement.  The movement is evident in this image - look at the alignment of the edge of the road:

Lets now jump a few more kilometres to the west, where things start to get really interesting!  Another formerly straight road is our indicator of strain - but now it is becoming difficult to believe that this happened in a single 30 second event (I can assure you it did):

Note here the highway people have resurfaced the road, so the original cracks are no longer visible.  There is a ditch running down the side of the road that also shows the deformation rather clearly:

The movement here is about 3 metres or so.  Again there is little evidence of vertical deformation.

The farmer at this site very kindly allowed us to enter his field (please do not do this without permission), where the array of cracks, and associated deformation, is astonishing.  GNS have an aerial view of the field here; below is the ground view:

The movement of the fault is evident in the power lines that cross the fault here.  The movement of the fault has put the cables under tension, meaning that the insulators (the pieces that join the cable to the pylon) are no longer hanging vertically:

A few more kilometres to the west and we are into the maximum displacement area.  Here we see over 3 metres of horizontal deformation and about 1.5 metre vertically.  This is the view from the hanging wall side - the drop down onto the footwall, and the lateral motion should be evident:

The picture below was taken on the footwall side looking back towards the fault - note the horizontal motion (the road has been patched up) and the vertical change.  Remember that this was a straight road across a flat plain before the earthquake:

The maximum deformation is recorded a little to the west again, where a road is offset by almost its entire width, indicating movement of about 4 metres.  There is some vertical deformation too, but this appears to be more of a ground roll than a true vertical movement of the hanging wall block:

Our last site is at the western end of the fault trace.  Here the surface expression of the fault is reducing, leaving a small step in the road that is quite hard to see.  The best evidence is from the power cables, some of which are now very tight (those on the upper right), whilst others are very slack (those linking to the pole on the far right).  Note how the tension cable supporting the pole on the left has also gone loose - this would have been taut before the earthquake:

It is at this point, almost perfectly on the fault, that the highest accelerations were measured.

I hope this gives a useful overview of the fault.  I am happy to answer any questions, though Geonet and GNS Science are the experts. 

Comments welcome

Zhouqu landslide - NASA satellite image

NASA have somehow managed to find a gap in the clouds to collect a satellite image of the Zhouqu landslide site.  Readers at the time of the landslide last month will remember that I was wondering about the source of the slides.  Although this image does not give a full answer, it gets us a part of the way there:

A zoom into the affected area yields this:

The slides (note the smaller one to the right of the main event) clearly originated as a series of smaller slips on the walls and headwaters of the upstream catchments.  These appear to have coalesced to form the main events.  This is consistent with processes observed elsewhere during periods of exceptional rainfall.

Meanwhile, I am now in Christchurch in New Zealand, site of the earthquake of 12 days ago.  Every 30 minutes or so my chair shakes as an aftershock rattles the area, which is a very surreal experience.

Road failure video from Thailand

Thanks to my former student Kurtis Garbutt for the heads up on this one.  The video below shows the progressive development of a road in Thailand.  I suspect that this initiated as a culvert failure, with multiple collapses due to undercutting:



The final frames of the video show an overview of the site.  Check out where the people are standing!

As an aside, Kurtis also runs a very interesting blog on natural hazards - take a look.  It is a very useful resource with a scope that goes way beyond my site.

Pakistan floods - the extraordinary duration of the elevated water levels

The floods in Pakistan may have faded from the headlines in Europe, but unfortunately the impact continues, even though the rainfall events that caused them occurred more than a month ago.  The most dramatic illustration of this is a set of satellite images collected by NASA using the MODIS instrument.

This is an image of the area around Sukkur taken on 7th July, before the rainfall event that initiated the disaster:

By way of comparison, this image was taken on 7th September, showing the same area:

Two things to note.  First the extensive flooding along the line of the main part of the Indus remains.  This may well be an indication that water is draining very slowly from behind the broken levees, which are now serving to keep water in despite their failure to keep the water out in the initial event.  Such problems with broken levees are common.  Second, note the parallel course of the Indus that has been created, where the flooding is much more intense.  I assume that this is the river occupying one of its old drainage paths.  The problem here may be that the downstream end of this huge (check out the scale bar on the imagery) new watercourse is Mancchar Lake.  This is an artificial lake, created in the 1930s, with a range of existing environmental problems.  The lake appears to be connected to the main Indus River by just two small canals (see Google Earth image below):

This limited drainage path is may cause the floods to drain very slowly, prolonging the agony for those affected.  The flood waters must now be highly polluted, increasing the potential for health problems for the affected population.

Ruapehu lahar information

One of the many highlights of the splendid IAEG Congress in Auckland this week was a talk by GNS geologist Chris Massey on the 18th March 2007 lahar at Mount Ruapehu on North Island.  The lahar occurred as a result of the failure of a tephra wall holding back the crater lake at the summit, and is shown by this NASA image:

The potential for a lahar had been anticipated and the site was intensively monitored with real time instruments such as water level sensors and geophones; with  two web cams; and with periodic surveys using a terrestrial laser scanner.  An emergency plan was in place and worked well.  The need for caution was undeniable - on 24th December 1954 a lahar from the same site demolished a railway bridge at Tangiwai, killing 151 people on a train that tried to cross a bridge that had been destroyed by the lahar.

The event has been written up in a paper (Massey et al. 2010), and there is a spectacular set of images of the event captured by the web cam available here (NB it took me a while to get my eye into these images).   

An interesting aspect of this event is that one can examine just how good the natural hazard science community is at assessing hazard.  There is a New Zealand Civil Defence report, written in 2002, about the threat of a lahar at Ruapehu online here. 


Reference
Massey, C., Manville, V., Hancox, G., Keys, H., Lawrence, C., and McSaveney, M. 2010.  Out-burst flood (lahar) triggered by retrogressive landsliding, 18 March 2007 at Mt Ruapehu, New Zealand—a successful early warning.  Landslides 7 (3),303-315, DOI: 10.1007/s10346-009-0180-5.

Italian debris flood video

Thanks to Ivan Montanari for highlighting this one.  Youtube has an extraordinary video of a debris flood that appears to have occurred yesterday at Atrani on the Amalfi Coast in Italy:



The video makes a slowish start, so keep watching.  This is the first debris flow of cars that I have seen on video...!  This is a Google Earth perspective image of Atrani:

I guess it is not hard to see how very heavy rainfall, or another hydrological event in the catchment, can lead to such an event.

A round up of intersting events and documents in the world of landslides and slopes

1. Please vote for a slope monitoring project
One of the recurrent themes of the IAEH Congress is the growing importance of slope monitoring techniques to provide warning of the development of failure.  One of the most exciting projects is the ALARMS project, led by Prof. Neil Dixon at Loughborough University, which is developing techniques that use acoustic emissions to monitor slope movements.  The project has been entered into an Enterprise competition at Loughborough University, which will be decided by a free vote.  So, please can you visit this page: http://www.lboro.ac.uk/business/enterprise-awards/ and vote for the Alarms project.  It would be a very worthy winner of the prize.


2. Manual on Engineering Geological Practice in Hong Kong
On behalf of the Hong Kong Geotechnical Engineering Office, AECOM have produced a new manual on engineering geological practice in Hong Kong.  Unsurprisingly, it has a strong slopes and landslide component, and it is relevant in areas outside of Hong Kong itself.  This can be downloaded as a pdf for free here.  I recommend it - I think that it will become a very useful guide and reference text.

3. 11th Symposium on Landslides (ISL)
The big four yearly slopes event is the ISL, which was last held in Xian, China.  The next event is in June 2012, and will be held in Banff, Alberta Canada.  This should be a fantastic event in a wonderful location.  The meeting now has a website, and the call for abstracts has been released, with submissions due by 15th April 2011.  Get writing!

4. Slope Stability 2011
In September 2011, Canada will also host a conference on rock slope stability, with a strong emphasis on the mining industry.  This event, which will be held in Vancouver, will be a great opportunity for academics and practitioners to interact.  The site has a website, and again a call for abstracts is out, with a deadline of 7th January 2011.

IAEG presentation and paper on Attabad

Today I presented an invited paper at the 11th Congress of the IAEG in Auckland, New Zealand on the topic of landslide hazards along the Himalayan Arc.  This paper starts by looking at global and regional landslide hazard before presenting the state of play at Attabad.  As usual, I have uploaded the PowerPoint file to Authorstream, such that you should be able to both view and download it below:

10_09 IAEG petley v2

See more presentations by Dr_Dave | Upload your own PowerPoint presentations

The written paper has been published in the conference proceedings.  I do not have a pdf of the actual printed version but I do have one of the final manuscript.  I am working on putting that online, so watch this space.

A pdf of the paper itself is available here:
http://www.mediafire.com/file/w34v38zb4uvzcj4/10_06%20Petley%20invited%20final.pdf

Note that this paper contains a map of global landslide losses in terms of fatalities and some data both for global losses and for those in the Himalaya. 

Note that the typeset format is slightly different from that of the actual published paper, but the content is the same.  The reference to the paper is:
Petley, D.N., Rosser, N.J., Karim, D., Wali, S., Ali, N., Nasab, N. and Shaban, K. 2010. Non-seismic landslide hazards along the Himlayan Arc.  In: Williams, A.L., Pinches, G.M., Chin, C.Y., McMorran, T.J. and Massey, C.I. (eds) Geologically Active.  CRC Press, London, pp. 143-154.

Updated: The New Zealand earthquake

Bizarrely, I am currently sitting in the QANTAS lounge at Sydney Airport waiting for a flight to New Zealand, where the IAEG Congress starts on Monday.  The theme of the conference is "Geologically Active"...

So, what do we know about the earthquake so far.  The best source of information is the Geonet site - http://www.geonet.org.nz/ - which puts seismic data online in real time.  They are reporting that the earthquake occurred at 4:35 am local time 30 km west of Christchurch at a depth of (now updated to) 10 km.  The reported magnitude is (now updated to) 7.1 - USGS is reporting 7.0, but this is probably a moment magnitude.  This is the isoseismal map from Geonet:

This suggests that some damage in the Christchurch area is likely, which seems to be confirmed by the initial news reports.  Hopefully the timing of the earthquake, and the high level of preparation for earthquakes in New Zealand, will mean that casualties will be light and damage manageable.  Perhaps the most interesting data at this stage is the map of locations in which people have reported they felt the earthquake.  This looks like this at the time of writing.  The dark orange dots represent an intensity sufficiently large to cause significant damage:

GEONET shake map

	There are no reports of landslides as yet, but given the proximity of the Southern Alps some landslides are likely:
	More later.

August fatal landslide data

The end of the month marks the point at which I post the accumulated totals for fatal landslides, based upon the global fatal landslide database that I maintain.  The raw figures for August are stark.  The total number of fatal landslides recorded in the month was 71, resulting in 2,740 fatalities.  This total is of course dominated by the Zhouqu landslide in China, which is reported to have resulted in 1,765 deaths.  (Click on read more to access the remainder of this post).


These totals continue the exceptional trend of landslide occurrence in 2010.  The average number of recorded fatality-inducing landslides in August in the period 2002-2009 is 46.7, resulting in an average of 348.3 deaths.

This graph illustrates the magnitude of the anomaly for 2010 to date.  This shows the cumulative total number of recorded fatality-inducing landslides for 2003 (which had the lowest total of any year on the database), 2009 (the previous record year), 2010 to the end of August, and the average for 2003-2009.  There is no doubt that 2010 is exceptional:

It is interesting that these totals have been recorded in a year that to date has had exceptionally low levels of tropical cyclone activity - indeed Northern Hemisphere tropical cyclone energy release remains at the lowest level for 30 years.   When they occur, landfalling NW Pacific tropical cyclones (typhoons) have a major impact upon landslide occurrence.  It is notable in particular that the Philippines and the Caribbean islands (especially Haiti) have been spared substantial tropical cyclone events to date this year.  Given that there are strong signs that tropical cyclone activity is currently increasing in both the N. Atlantic and the Pacific, the next month will be very interesting.

Of course, the root cause of the exceptional occurrence of recorded landslides in August has been the intensity and volume of rainfall associated with the summer monsoon, primarily in China and Pakistan, and to a lesser extent Nepal.  The monsoon should start to weaken during September, but more landslides are likely in the interim given the near-saturated state of the ground.

Another large landslide in China

China has today suffered another large landslide event.  Xinhua reports that a slide happened at 10:20 pm on Wednesday in the village of Wama in Longyang District, Yunnan, leaving 8 people dead and 40 missing.  Xinhua have this image of the slide:

As far as I can tell, this is the location:

Unfortunately the Google Earth image for this area is poor, but there is a report on a small hydro-electric scheme at this village available online here.  I should add that I am not suggesting that the HEP scheme was responsible for the slide.  This article on recent permission for logging in the area is also interesting.