New pictures of Attabad

Many thanks to David Butz, who is a Professor of Human Geography at Brock University in Canada, for sending to me a set of photographs of the situation at Attabad.  David was there about eleven days ago, when the boat service was still running (it has now ceased due to a lack of fuel).  David has kindly allowed me to post some pictures here - note however that he retains ownership and copyright of them.

The dam from from downstream (click on "read more" to get the remainder of the post):

A close up of the downstream part of the spillway.  Note the very large boulder:

The upper part of the spillway, the point at which the flow rate increases:

The landslide mass from the upstream side:

The towering landscape and the heavily overloaded boats that ply the lake:

Many of the displaced people are still located in very hazardous locations:

And to finish, the best image that I have seen to date that is able to show the size of the landslide.  Remember that the lake is 120 metres deep:

Attabad lake level is rising again. It is clear that temperature is the key control at present.

The lake level at Attabad is rising again - the latest NDMA report suggests that the level has increased by 19 inches (41 cm) in the last 24 hours:

It is now clear that the lake level is controlled primarily by local environmental conditions (which in turn are changing the inflow), rather than the spillway properties.  Compare the following two graphs.  The first is the lake level through July to date:

The second is the Pakmet hydrograph for the same period for Besham, which is much further down the Hunza valley:

The similarity in patterns is not because Attabad is controlling the flow at Besham (the Attabad discharge is just a fraction of that at Besham) - the similar patterns are because the flow at Besham is being controlled by air temperature (i.e. the rate of melt on the snowfields and icecaps), as is the flow, and hence the lake level, at Attabad.

Regular commenter BeforeGoreKneel kindly pointed out that temperature graphs are available online here too.  The temperature graph for Gilgit is as follows:

The role that temperature is playing in controlling this system is clear.

Meanwhile, it is still not clear as to whether the Frontier Works Organisation (army) has started their operations to lower the lake. 

Attabad lake level stabilises again

The latest report from the NDMA suggests that the lake level at Attabad has stabilised once again, matching the stabilisation of flow in the Indus catchment:

There is no news as yet as to when blasting to drain the lake will start.

Large drop in the lake level at Attabad

The latest NDMA update (released yesterday) reports a further very substantial drop in the level of the lake.  The report is that the level of the lake fell by 28 inches (71 cm) in the 24 hour period to 9 am (local time) on 23rd July.  This means that the lake level graph looks like this:

The cause of this reduction appears to be a dramatic decrease in inflow, caused by the substantial drop in temperature in recent days.  Regular reader and commenter BeforeGorekneel pointed out that hydrographs for some of the major Indus flows are available online in real neartime here.  The data supports the view that the river flows have decreased markedly in the last few days:

Flow through the spillway is declining slowly in consequence, but as of yesterday remained quite high.  This is my calculation of spillway flow, which is the NDMA data for Ganish Bridge less the seepage and downstream component:

Finally, reader and commenter Torsten has provided a comparison of images from 14th June and the somewhat grainy Pamir Times image of 18th July.  This rather helpfully helps to show the magnitude of the volume loss on the downstream side of the dam:

The lines are intended to assist in comparing the two images.  The amount of volume loss is remarkable.

Finally the Pamir Times reports that the government is "all set" to begin blasting in order to lower the lake level.  I wish them well.

Strange goings on at Attabad

The NDMA daily report for today (22nd July) reports a dramatic decrease in lake level - 19 inches (48 cm) in the last 24 hours.  This is by far the largest fall recorded to date, producing a lake level the looks like this:

This is quite surprising.  Assuming that the surface area of the lake is 1202 hectares (in fact it is probably a little larger in reality), this implies that about 5.8 million cubic metres of water has left the lake in the last 24 hours, representing a rate of about 67 cubic metres per second.  However, the reported spillway discharge has not increased substantially - discharge at Ganish Bridge was 20,800 cubic feet per second yesterday and 20,837 cubic feet per second today. 

So what is going on?  Well it could be a measurement error, perhaps of the discharge.  Alternatively it could be that the inflow has dramatically reduced, maybe because the weather has cooled.

Either way it is a surprising observation. 

Finally, I somehow missed until today this low resolution image published in the Pamir Times a few days ago of the dam:

I am astonished at how much downstream erosion there has been since I last saw an image of the site.


Comments and thoughts welcome please.

The Government has reportedly decided to use blasting to lower the lake at Attabad by 80 metres

The Associated Press Pakistan has a report of a somewhat important development in the Attabad landslide dam story this evening.  The report is of a high level meeting, chaired by the President of Pakistan, which has decided to lower the level of the lake to 30 metres (it is currently 111 metres, implying a drop of about 80 metres from the current level.  This will be achieved through controlled blasting.  Meanwhile a proper ferry service will be set up on the lake (at last).  Finally, "it was also decided to undertake feasibility study for building by-pass as a permanent solution to the problem of landslides in the future."

Interestingly, this decision follows a visit to China by the President.  The Karakoram Highway is of great strategic importance to China.  It is good to see that China has offered technical assistance, given that they have the greatest experience in operations of this type.

No timetable has been provided for the works.

Meanwhile, the latest NDMA reports suggest that the lake level has fallen by about 40 cm in the last 48 hours, possibly reflecting lower inflow rates.  This means that the lake level and spillway discharge graphs are now as follows:

Latest Attabad lake level information

I have now extracted the lake level data from the latest NDMA daily report.  The latest graph on lake level above the spillway floor at the time of overtopping is as follows:

I haven't updated the long term lake level graph for a long time, so here it is:

Finally, the flow through the spillway is also now increasing, even though the lake level is going up:

This increased lake level must be causing problems upstream, although there are few media reports at present.  The weather forecast for the next two days is for warm weather, so expect inflow to continue to be high.  Based on the analysis by David Archer, the mean date for peak flow is 29th July - i.e. about ten days from now.  Given the cold winter it may occur late this year, but within a month the inflow should be starting to reduce.

The Attabad lake is now at its highest level so far

Today's NDMA report paints a surprising picture of the Attabad barrier lake. It reports that the water level  has risen by 36 inches (91 cm) over the weekend, meaning that it is now at its highest ever level.  I will try to post the graphs of lake level and discharge in the next three hours or so.

Attabad - lake level is high but steady

The latest NDMA reports show that the lake level at Attabad is at its highest level to date, but is holding steady.  As before, I have used the reports and the NDMA data to reconstruct the lake level data, to produce a graph that is consistent with NDMA for the last two weeks:

Meanwhile, it appears that the discharge through the spillway has increased again, once more using the NDMA data:

This allows us to look at the discharge - lake level graph again:

The most recent three values are indicated by open diamonds.  There are some suggestions that operations are underway again to widen the spillway, which may explain the increased discharge, but I have no way of verifying this suggestion.

Finally, the press are also reporting that the IDPs will be allowed to return to their villages in early August, when the river discharge has peaked.  Fortunately the enhanced level of landslide activity from the Attabad scar that was observed last week has now reduced.

Attabad - the lake level has apparently risen again

Interesting developments at Attabad, as various press reports suggest that the lake level increased substantially once again yesterday.  The Pamir Times for example reports that the lake level went up by 11 inches (28 cm) yesterday; this magnitude of increase is reported elsewhere too.  I have used the data and the graph in the NDMA daily reports to build a dataset of the lake level with time:

If these reports of an 28 cm increase are true then the lake level is probably very close to the previous maximum level.  Any further increases could of course lead to more losses upstream, and increase the potential energy stored behind the dam.  This increase in lake level is interesting though because at the same time the reported outflows have declined from those of a week or so ago.  The estimated spillway flow graph is as follows.  I have assumed that the seepage rate has remained constant, and there is a considerable uncertainty in these media reported values:

The resulting relationship between the lake depth and the spillway flow is shown below.  For clarity I have plotted the five most recent points as open diamonds:

It is not clear to me as to why the spillway discharge has declined even though the lake level has increased.  There are a number of possible explanations, but it t is difficult to speculate with so little information and no images, and so at this stage I will not do so.

Meanwhile the Pamir Times also reports that there were 13 large rockfall events yesterday, some of which created waves on the lake.  Secondly, a rather strange report in The News says that "...construction of road leading to Attabad lake is underway. Deputy Commissioner Hunza Nagar Zafar Waqar Taj told Geo News that the construction of road will be completed within a week followed by installation of a floating bridge through which five big trucks will be pass at a time. It will make possible transportation of goods in huge quantity between Pakistan and China and swift delivery of relief items to upper Hunza Gojal affectees."  That is an intriguing possibility, involving a major development over a very short period of time.  I cannot quite work out what is meant by this!

Increased landslide occurrence at Attabad

Various media reports suggest that the rate of landslide activity has dramatically increased today on the slopes above the Attabad lake.  For example, SAMAA is reporting that "Series of landsliding still continues in intervals in Ataabad, Hunza, causing difficutlties for the boat service, SAMAA reported Friday.  Process of landsliding has started in intervals since last night in Hunza, Ataabad. Boat service however being continuing."


A key concern remains the potential for a large failure to create a wave that triggers a rapid breach event.  The ability for mass movements to create waves is best illustrated by this video of an ice collapse event:





This is a very small collapse event compared with a large rockfall.  Large collapses are often preceded by increasing event rates (i.e. in this case increasing rates of rockfall activity).  The best (and absolutely fantastic) example of this is a large ice collapse event captured in Argentina.  Do watch this if you can!



This is not to say that such an event is coming in Hunza, or indeed that a wave would be large enough to trigger a breach, but the need for caution is clear.

Attabad - what can NDMA do now?

The situation facing NDMA at Attabad remains very challenging.  They continue to need to manage both an understandably restive displaced population and a dam that has reached some kind of dynamic equilibrium, with no major changes appearing to be occurring.  In addition, the optimum solution for the upstream population (draining of the lake) is at odds with the probable wishes of the downstream population (stabilise the lake in place such that there is no damaging flood).  Throw into the mix a complex political environment, and the situation becomes very difficult indeed.

So what are the options?  As far as I can tell the possibilities are as follows:
1. Maintain the status quo until after the peak flood has passed.  
One option is to do nothing except to monitor the lake.  Given the apparent stability of the dam, this option has attractions, although of course it leaves the displaced population in a state of limbo.  It is unclear as to how long the IDPs both upstream and downstream can continue to live in their current difficult state - the ongoing protest on the dam site suggests that things may be quite volatile now.  In addition, the risk from the dam is undiminished, but with large numbers of people out of harm's way this is being managed.

2. Allow the population to return, protected by a warning system
A second option is to set up a full warning system and to allow the population on the downstream side to start to return to their land.  There are risks associated with this of course - and these would need to be fully explained to the vulnerable people - but perhaps the local people would view those risks as being tolerable compared with the challenges of their current daily lives.

3. Widen the spillway further
NDMA could attempt to further widen the spillway, although it is now clear that widening operations to date have not caused a substantial drop in lake level.  A wider spillway may increase the capacity to withstand the peak flood, and might avoid further rises in lake level.


4. Deepen the spillway
If the aim is to resolve the situation then it is necessary to deepen the spillway, presumably using explosives delivered via military hardware.  This was the approach used at Tangjiashan to initiate the breach:

A similar approach was used elsewhere in the Wenshan earthquake area - for example this is the Shinangou landslide dam being demolished:

The size of the blasts needed to initiate a breach in China was very large, as the images show. Needless to say the use of explosives in this way carries very considerable risks - especially in an area as geologically unstable as Attabad - and therefore should only be undertaken with great care.  The caution that NDMA are showing with this approach is well-advised.  In my view such an approach may well be too risky, but I am sure that the authorities are weighing up the pros and cons fully. 

Thus, at this stage the viable options remain really very limited indeed.  The situation remains one of waiting to see what happens, frustrating though this is for the local people.

Options during the winter?
If the dam survives the summer floods then the flow rate will decrease substantially during the winter.  This might be a time to start trying to induce a controlled lowering of the lake or to start to armour and protect the spillway.  Perhaps siphons and pumps could be used to bring the lake level down to the point that it is possible to work on the spillway, which could then be deepened or engineered.  This would require a substantial investment in time and money, but may well be viable.  However, this is a long way off at the moment.

Can Attabad fail?

A number of commenters here have suggested that the structure of the dam is such that it cannot now fail due to the presence of the large boulders.  So I thought I'd write a short post exploring whether this is actually the case.  I believe that it can still fail, and indeed that eventually it probably will, but I have no idea as to the likely time-frame.  The reasons that I believe that a failure is still highly possible are as follows:


1. This is a landscape littered with large landslide scars
In the Attabad area alone Shroder (1998) identified six previous very large landslides:

"Multiple overlapping rockslides have repeatedly thundered into the Hunza River near Atabad, several with serious and destructive consequences. The six slope failures of interest here are: (1) older Serat rockslide; (2) younger Seratrockslide; (3) Ghammessar slope failure and lake; (4) Ghammessar breakout flood and retrogressive slump failure; (5) 1962 Ghammessar rockslide and lake; and (6) 1991 Sulmanabad rockfall."

Both upstream and downstream this pattern is repeated - there are literally hundreds of large rockslide scars in this landscape, many of which will have blocked the valley in a similar way to Attabad.

So how many landslide dams are left?  There are fragments and remains of them in many locations, but there are very few intact landslide dams.  This suggests that most such valley blockages eventually fail, although not necessarily rapidly.  There are no real grounds at this stage to assume that the dam at Attabad is exceptional.

2. We are still some way from peak flow
The data from David Archer that I presented in an earlier post suggests that we are probably three weeks or so from the peak flow, which may well be 30-50% greater than at present.  Whilst there are grounds for optimism that the structure may survive such flows, it is far from certain that this will be the case.

3. Landslides into the lake are a real threat
Landslides continue to occur on the walls of the valley.  A large slide still has the potential to create a wave that could trigger a rapid collapse.  This threat has not diminished.  We believe that this was the failure mode for the 1858 landslide dam just downstream at Salmanabad.  There is a need for proper assessment of this threat before one could sound the all-clear.  I hope that NDMA are on the case.

4. The dam is still losing volume
Images of the downstream area of the river show that the water is still carrying a substantial amount of sediment, as this Pamir Times image from a week or so ago shows:

The loss of volume implies that the dam is weakening with time, but it is not clear how fast or where.  Nonetheless, until this ceases the potential for failure remains.

5. This is a river with GLOFs
GLOFs are glacial lake outburst floods, which are flash floods created by the collapse of lakes dammed by glaciers or moraines high in the mountains.  GLOFs create short duration, very large magnitude floods.  The Hunza suffers GLOFs on a regular basis.  Such an event would lead to a greatly increased flow rate over the spillway, threatening its stability.


6. Earthquakes
This is an area of high seismic hazard.  A substantial earthquake would threaten the dam in a number of ways.  First, the dam itself could undergo slope failure and collapse.  Second, the earthquake could create a seiche (standing waves) in the lake that could overtop the dam, inducing failure.  Third, an earthquake could trigger further slope failures into the lake, causing waves.  The likelihood of an earthquake is low, but the consequences could be very serious.

Please do not believe that the boulders rule out the possibility of the release of the lake.  This is not the case, despite their size.  If the flow velocity and volume is sufficiently high then this dam can still fail.  Unfortunately it is impossible to say when and how this might occur, or how rapidly such an event might develop.  The chances of a very rapid failure are comparatively low, but are not negligible by any means

Meanwhile, NDMA reported that the lake level rose 2 inches (5 cm) yesterday, although it does appear to have fallen the day before (the link on the NDMA website to the report yesterday is dead).  Local people are reportedly protesting about government action and the limited amount of compensation by camping close to the dam.

Unfortunately though the flow of information from the site remains very limited.


Reference

Shroder J.F. 1998. Slope failure and denudation in the western Himalaya, Geomorphology, 26 (1-3), 81-105.  DOI: 10.1016/S0169-555X(98)00052-X.

New EO-1 ALI satellite image of Attabad

NASA have released a new satellite image of the Attabad landslide and lake, collected by the Advanced Land Imager (ALI) instrument on the EO-1 satellite on Saturday 2nd July.  The images are excellent, showing both the lake and the dam and spillway with only limited cloud cover (note that north is to the right on these images):

The box indicates the following close-up image:

They also link to this site, which is appreciated.  Meanwhile, the NDMA site suggests that the lake level has dropped by about 3.5 inches (about 9 cm).

Attabad - no substantial changes in the lake level reported by NDMA

Thanks to various commenters on the Attabad threads for their incredibly helpful and interesting contributions - this is something that I really value - and also for seeking out the latest information on the situation now that my usual feed has dried up.  NDMA posted a report dated 5th July (but helpfully they got the file name wrong  - thanks to Divalent for tracking down the right one here) that provides a latest update on the lake level.  The graph looks like this (the y-axis is lake level above the spillway base at the time of overtopping, measured in inches):

This seems to suggest that the lake level has now stabilised again.  The temporary drop noted on about 2nd July may be an error.  The report itself suggests that the lake level had increased by 2 inches (5 cm) in the period since 6 pm the day before, although changes of this magnitude may be no more than the effect of the wind.  Thus, it appears that the situation is one of (temporary) equilibrium again.

Meanwhile a new video of the spillway, apparently dated 6th July (i.e. today), has appeared:



This appears to show a steep cascade of water that is apparently clean, suggesting that the bed of the flow is formed of large boulders, although this is mere speculation of course. The lake level shows little change from earlier videos, confirming the NDMA figures.

Finally, we must not forget the continued huge toll that this crisis is inflicting on both upstream and downstream communities. The Pamir Times has a good gallery of Gulmit, near to the head of the lake here.  It is worth  a look:

Attabad - Analysis of NDMA data on lake level

Yesterday's daily update from NDMA contained a new, and very welcome addition: a graph showing lake level, starting 29th May (the date of overtopping).  The y-axis is slightly confusingly labeled on the sheet (the label having been added later by the looks of it), but is clearly the lake level (in inches - NB 1 inch = 2.54 cm) above the spillway base level at the time of overtopping:

There appears to be a datapoint every two days until the last four days (which have daily data).  This data suggests that the water level peaked on 24th-26th June, and has subsequently fallen somewhat, which is really good news.  This coincides with the cessation of reports about houses being flooded.

This of course allows one to check the data that I have been publishing against the NDMA data.  Below is the graph of my data up to the point that the feed ceased for me (the points are daily rather than two daily), with the lake level above the overtopping point in inches on the y-axis

There are some minor differences, as one would expect of two different datasets of the same natural  phenomenon (the major difference is that my data are slightly higher for the middle part of the dataset), but the similarity between the two graphs is striking, which is reassuring at every level.  The peak water levels recorded are almost identical (within +/-2% of each other).

I am still struggling to reconstruct the more recent changes in water level, so I have used a tool that extracts data from graphs to determine (probably +/-2%) the most recent lake level data from the NDMA plots.  I have then added this to my data below to generate a compound graph.  The open symbols are the NDMA data:

Spillway discharge has also substantially increased.  I have tried to continue to compile reported spillway flow figures, cross-checking these against the Ganesh Bridge data that NDMA reports.  Where the data are consistent I have included it; the resulting relationship between discharge and time looks like this:

Perhaps most interesting is the relationship between spillway discharge and lake depth:

There is a fascinating level of complexity in this plot that will deserve better analysis in due course (there are going to be some great papers from this example), but it is clear that the relationship between discharge and lake level is now changing, with continued increases in discharge even as lake level falls.  This suggests that the spillway efficiency in transporting water is increasing.  I wonder if this is because the spillway is now downcutting, or because widening (natural or artificial) is starting to have a real effect.  It will be worth watching this over the next few days.  It is a great shame that we have no images from the site at the moment.

All-in-all the situation on the ground is evolving, which means that we are still moving towards some sort of resolution.  Of course we expect to see inflow continuing to increase over the next few weeks.

Attabad - limited information flows

Unfortunately, I am not currently receiving any updates on the situation at Attabad, so it is hard to post on what is going on.  The best that I can do at the moment is to report the various figures from the media, although it is far from clear as to whether these are reliable.

The last set of data that I received from the site was for 25th June.  At that time the sillway flow was reported to be as follows:
Spillway = 12,450 cusecs; Total outflow = 12,650 cusecs; Inflow = 14,810 cusecs

Media reported flows since have been (NB there are no indications as to the reliability of these figures, so I cannot vouch for them):

28th June: outflow = 15,000 cusecs; water level fallen 12 inches in two days (Pamir Times)
29th June: outflow = 16,000 cusecs; inflow = 14,000 - 15,000 cusecs; water level has fallen a "further" 13 inches (Hunza Times quoting the Chair of NDMA)
1st July: outflow - 16,554 cusecs; Water level increased 3 inches in previous 12 hours (Express Tribune)

I hope that the flow of information will improve soon, but can make no promises.

New images of spillway at Attabad

Unfortunately I cannot access images from Attabad at the moment, but the Pamir Times has this morning published a set that are well worth a look.  The images are here.  Interestingly, they also note that the water level has reduced by 13 inches (about 33 cm) over the last two days, which is a promising sign.  Unfortunately I cannot validate these figures at the moment.

The images seem to show that the key constraint remains the point of the large boulder as before, although the water has now eroded around this, leaving it stranded in the middle of the channel:

The channel is clearly widening on both sides through slope failures, increasing the capacity of the channel, but there is still remarkably little sign of down-cutting.  Meanwhile downstream, the dam is continuing to lose mass through erosion, a process that we must watch carefully:

This loss of volume is very slowly weakening the barrier.  There are no signs at the moment that this is a cause for real concern, but close monitoring is needed as the flow rate increases.

Attabad - situation as of 25th June,

NB I have updated this post, and changed the title.

Apologies for the lack of posts over the last few days - I have been in a meeting in Bermuda that has left little time to tend to this blog.  Meanwhile, although small scale blasting of the spillway has reportedly continued, as of yesterday the lake level at Attabad is continued to rise by 10 to 20 cm per day, with the effect of both further drowning houses, hotels and roads upstream, and increasing the volume of water in the lake.  Unfortunately I do not have any information about the blasting, or any pictures, but my analysis of the flow of the data is that the short term lake level graph now looks like this (correct to yesterday morning)

The lake level is now about 5.5 metres above the overtopping level, and at the moment there is little sign that this is reducing.  Thus, the long term lake level graph now looks like this:

Meanwhile, spillway discharge has  increased, driven primarily by a substantial rise in inflow:

I am intrigued to know whether this increased discharge has started to drive a new wave of erosion downstream.  Finally, this graph compares inflow and spillway flow over the last ten days or so:

It is clear that at the moment inflow and outflow are directly linked - i.e. that the increase in spillway flow is due at least in part to increasing inflow.  The grey line represents the point at which spillway flow equals inflow - i.e. the point at which the lake volume should stop increasing.  This does not account for seepage, but as this is now a tiny fraction of spillway flow this is not a major problem.  It is clear that recent increases in inflow were not being balanced by increasing outflow.

It should be expected that the inflow will continue to rise over the next few weeks, so this graph will be interesting to observe.  Hopefully the spillway operations will start to increase the outflow in a controlled manner in the next few days, such that the lake level starts to fall.

Blasting of boulders has begun at Attabad

Yesterday NDMA started the process of blasting the boulders in the channel at Attabad.  This image, from Focus, shows the state of the channel and the reconstructed access track:

One of the blasts yesterday was on the boulder at the entrance to the channel, visible on the left of the image above.  This image shows the crews drilling the blast holes:

Clearly this boulder is not the main issue in terms of the flow, but it may be that they are using this one, which is accessible, as a test.  Access to the other side of the channel is difficult, so the suggestion is that water cannon will be used to erode the softer material.  It will be interesting to see what they do today in terms of blasts.

Meanwhile the lake level rose by between 8 and 10 inches (20-25 cm) yesterday, suggesting that inflow has increased again.