Attabad - situation report (Updated 26th June 2010)

Commentary and analysis accompanying this data is provided on Dave's Landslide Blog.

Situation as per 25th June:
Lake depth: 116.7 metres
Change in lake depth in last 24 hours: +0.13 metres (negative number = fall in lake level)
Spillway discharge: 352 cubic metres per second
Estimated seepage: 6 cubic metres per second
Estimated total outflow: 358 cubic metres per second
Estimated total inflow: 419 cubic metres per second

Lake level graphs:
Lake level and rate of filling from time of landslide:

Lake level since 21st May 2010:

Flow through the spillway:

Slowing evolution of the spillway at Attabad

Once again I am grateful for Focus Humanitarian Assistance for sending the latest photos of the spillway at Attabad.  The rate of change new seems to be slowing as flow is being controlled by the large boulders at the head of the channel.  This pair show the downslope side of the channel - the left hand image was taken on 7th June and the right hand image a day later:

Allowing for the slight change in camera angle, there is little indication of major change between the two images.  The large boulder on the left side of the upper part of the channel bank has slipped into the water on the right hand image, presumably indicating that channel widening is continuing.  Of course there may well be some continued erosion of the bed that would is not visible here.

In the upper channel there is comparably little evidence of change also.  The inset image is from 7th June, the main image a day later:

It may be that a more detailed inspection on the ground would reveal a more dynamic system.

It is my intention to stop duplicating posts on the two sites from today, with the Hunza Monitoring Blog being used once again to present data on the state of the dam and the lake, and Dave's Landslide Blog to provide a commentary.  On DLB I will also start covering other landslide events again.

Latest image of the spillway at Attabad

The wonderful people at Focus have provided an image of the state of the spillway at Attabad that is very instructive.  This was taken today:

The current state of play is I think as follows.  Flow appears to be constrained by two blockages but, as a commenter has noted, one is a rock spur out of the channel.  The blockage appears to have formed a small waterfall / rapid.  Downstream of the blockage the channel has widened dramatically - and indeed a section of the track has now been lost.

To me this suggests that the crisis may be far from being over.  In particular, the loss of this barrier could cause a rapid increase in flow rate that could be highly erosive.  It is however important to stress that interpreting the state of play from photos is difficult.

Attabad - an increasingly difficult hazard to manage

Apologies to all who emailed and left messages over the weekend requesting updates. 



FWO data collected at 8 am on Saturday suggests that the rate of flow was 124.6 cumecs, whilst the Pamir Times reported yesterday that it was 131.7 cumecs.  If so, the discharge time graph looks like this:

Thus, the discharge is apparently stabilising.  Various media reports from the site itself suggest that downward erosion of the channel has now reduced, but that some lateral erosion is still occurring.  I have not seen any images to confirm this though.  The lake has stopped rising for now.


It is hard from here to assess the current position with regards to the dam, or to forecast what will happen next.  However, from the start I have held the view that managing this hazard would be somewhat challenging if stable flow became established.  The chronic hazard has not gone away, although the acute hazard may have reduced.  As there is still a vast quantity of water stored in the landscape, the dam remains vulnerable to a series of processes, including:

1. An unexpected increase in erosion rate;
2. Erosion during flood events (the discharge of the Hunza will rise substantially in the next few weeks);
3. A further landslide on the banks of the lake, which could trigger a wave;
4. A seismic event.
   

Thus, NDMA have a huge challenge ahead in deciding when to allow people to move back into the high hazard area.  They will also need to decide whether to initiate erosion through the use of a controlled blast.  This is not an unenviable decision to make.  Inevitably, there are some exerting pressure to initiate a breach to drain the lake, whilst others are happy with the status quo.



Meanwhile, of course those on the upstream side remain isolated, with a huge lake impeding access.  The Karakoram Highway remains closed indefinitely, although the boat service has been resumed.  During summer floods, when inflow may briefly exceed outflow, the lake level could rise again by a small amount, unless progressive erosion of the spillway serves to lower the lake level.  Thus, overall, many challenges remain at Attabad.; it will be interesting to see how NDMA responds to them.

Attabad: spillway flow update of 08:30 this morning

There is a mass of confusing and contradictory information about the state of flow in the Attabad spillway, but the FWO data appears to be consistent at the moment.  The latest data, again via Dr Sadiq,  is that flow was 3700 cusecs (105 cumecs) at 08:00 local time this morning.  This makes the graph look like this:

This suggests that we are still seeing an incremental increase in flow.  However, interestingly, the rate of inflow and the rate of outflow now appear to be balanced, such that the lake level should stop rising (for a while anyway).  If this happens, and spillway flow continues to increase, we will know that erosion is developing.

Photos of the development of the situation at Attabad

Focus have kindly provided a new set of images of the situation at Attabad with respect to the spillway.  I think that these images were taken yesterday.

Most importantly, here is an overview image of the spillway:

It is helpful to compare this with the image of the spillway from 1st June (right hand image):

It is clear that the situation has developed considerably over the last few days.  The flow along the spillway has increased greatly, and the lower part of the channel has widened and deepened.  Flow appears to still be controlled by the large boulder in the middle of the channel, although the lowering of the channel downstream will be steadily undercutting this.The development of the spring on the channel edge is also interesting.


A closer view of the spillway clearly shows how widening is occurring:

The flow is undercutting the banks, which are then progressively failing.  Note also just how much scour is occurring even though the bed of the channel is boulder-strewn - it is clear that at the moment the boulders are not armouring the channel.

A view from the downstream side of the channel is fairly dramatic:

This gives the impression that there is a huge volume of water flowing.  However, a look downstream towards the old landslide deposit at Salmanabad (that is the big pile of debris upper right) shows that this is not really the case:

It is clear that the flow is still rather modest, despite the huge amount of erosion that the water has achieved.

Attabad: a different set of discharge data

Thanks to Aftab Sadiq of the School if Civil and Environmental Engineering at NUST for passing on a revised set of spillway discharge data.  This set was collected by the FWO.  It is almost identical to that published here previously, with one important exception.  The graph of discharge against time is as follows (correct to 17:00 yesterday, local time):

The difference between this and the data that I have published previously is shown in this graph:

The key difference is for the data of 2nd June, when the FWO recorded a rather higher discharge than did the earlier dataset.  I noted yesterday that the linear trend in the earlier dataset was somewhat odd - the FWO data does not show this.  The difference is minor on the one hand, but important on another.  If we take the original dataset, the apparent sudden stepped increased in discharge yesterday appeared to suggest a major change in process - i.e. that erosion had initiated.  The FWO data suggests that the process is unchanged - i.e. we are continuing to see an increase in flow associated just with the increasing lake level.  Given that as far as I am aware no breach has developed, my inclination is that the FWO data may have captured the processes on 2nd June better, but this is uncertain at the moment.


It will be very interesting to see what today's data might bring.  As it appears that discharge and inflow are now coming close to each other we might expect to see the discharge plateauing.  However, some channel erosion will continue, so the situation is far from resolved.

Discussions with the team that dealt with Tangjiashan showed  that one of their key concerns was that overtopping would be followed by equilibrium conditions, leaving the hazard in place and creating a very difficult management task downstream.  At Hunza we have a repeat of this situation.  In the end at Tangjiashan they resolved this by blasting the boulders that controlled the flow in order to initiate a breach event. Given that they do not communicate with me at all, I have no idea what the plans of NDMA might be.


Finally, there is a new NASA ASTER image, collected on 1st June, of the site available:

However, thanks to G.M. Shah of AKU for pointing out that there are some errors in the labeling on this series of images.  The key issue is that the location of Shishkot is incorrect:

I will post again when I have further information.  And finally, from a personal perspective, thanks to all who wished me luck in my move - it is complete.  I am now getting used to living on my own for the first time in 16 years!

Attabad appears to have shifted to the next stage

In brief, the discharge at Attabad appears to have dramatically increased in the last few hours, as the graph below shows:

I am trying to find out more information, but it appears to have shifted to a new phase of activity. Inflows and outflow may be approaching a balance. Downstream communities need to be very alert now.

More as soon as I get it.

Attabad - so why hasn't it breached yet?

Several people have emailed to ask why the lake has yet to breach.  Well, I would like to emphasise two things - first, that every landslide is different, so predicting the behaviour of any particular slide is effectively impossible.  Second, there is a peculiarity about the spillway that appears to be delaying the breach event.

Regular readers (of which there are now several thousand each day I think - thanks to you all) may remember that a few days ago I noted the presence of a large boulder in the channel, right at the crown of the spillway:

 Well it is this boulder that is preventing the scour from propagating upstream and that is, in effect, protecting the upper part of the channel against further erosion, as this image shows.  The boulder is circled:

The lower part of the channel is continuing to erode with some power, and the channel is widening.  The red lines show substantial cracks suggesting further potential widening.  The steepest part of the channel is continuing to erode backwards, such that sooner or later the boulder will start to be undercut.

Meanwhile, there appears to be an increasing level of frustration about the very confused spillway discharge statistics emerging from NDMA.   This is understandable.  If we take the last four daily reports we get:
29th May 18:00: 150 cusecs (cubic feet per second) from the spillway, 350 total discharge

30th May 18:00: 900 cusecs total discharge

31st May 18:00: 900 cusecs total discharge

1st June: 18:00: 1050 cusecs discharge at Ganesh Bridge.



This is a quite bizarre mixture of measurements, but we should all understand that the NDMA team dealing with this are under huge pressure too. 

Meanwhile, the more reliable measurements of inflow suggest that the rate is continuing to increase linearly with time, probably because of the choke provided by that boulder.  However, I am very surprised by just how linear this increase is proving to be, so please treat this with some caution:

The lake level upstream continues to rise, increasing both the damage and the amount of water in the lake.  In the last 24 hours the lake level rose by 75 cm, perhaps suggesting that the rate of inflow is further increasing. 


Finally, several people have emailed to ask why I have not covered the multiple landslide accidents around the world in the last few days, including the multiple slides in Central America (and that amazing sinkhole), the landslides in China, the non-fatal slide in New Zealand, etc.  I will try to catch up on all of those things as and when this event ends - please accept my apologies.  You will understand that I have to maintain my day job whilst doing this, plus yesterday and today I am moving house!

Attabad - flow continues to increase with time

The Pamir Times reports that as of lunchtime today the flow through the spillway had increased to about 850 cubic feet per second - i.e. 24 cubic metres per second, giving a discharge - time graph that looks like this:

For the last 72 hours or so the increase in discharge with time appears to have been linear, which is interesting.  This suggests that erosion of the saddle has yet to initiate fully.  Fortunately so far the in situ (i.e. not excavated) lacustrine silt and clay has been more resistant to erosion than some had speculated, although the disturbed (excavated and dumped) material has proven to be very erodible.  However, the inflow remains more than three times greater than the outflow (seepage plus spillway flow), and retrogressive erosion will in due course steepen the channel, which will increase the erosive potential.


In the meantime the level of the lake is continuing to rise at 50 cm or so per day, with the upshot that further flooding is occurring upstream.  The Pamir Times has a tragic (but beautifully shot) photo piece (plus here) on these continued impacts.  The level of impact up there is all too clear:

Attabad - continued retrogression. Close to the next stage?

The latest images from Attabad suggest that the retrogressive erosion in the channel continues to accelerate, and that this is now by far the most likely failure mechanism.  A breach now looks to me to be inevitable.  This image, taken this afternoon, shows how the spillway has developed:

Note the people on the road for scale. 

Compare it with yesterday's image:

The waterfall has now smoothed out, probably because the water is now eroding the landslide mass rather than the excavated material from the spillway, to create a set of rapids that are clearly eroding back up the channel.  The head of the rapids are close to the saddle.  The key point is probably the location where the channel become notably steeper - this is where scour accelerates.  One this point is reaches and passes the saddle, the rate of flow will start to increase and we might well see the breach developing.  Unless the top of the rapids is being impeded by a large boulder or similar, this will probably develop quite quickly.


It remains hard to know how quickly this will develop once the saddle is reached, but downstream communities need to be prepared for a rapid breach.

Attabad - further retrogressive erosion

The latest photos from the site, taken late yesterday, show that the retrogressive erosion of the channel is continuing to develop rapidly.  This image was taken at about 5 pm yesterday:

The lower part of the channel has widened and, more importantly, the steep section has moved up channel considerably.  Bear in mind that this image is more than 15 hours old now.  The best illustration of the way that this is developing is with this sequence of three images, which are taken about a day apart, starting on 29th May on the left and ending with 31st May on the right:

It seems to me that release of the water in the lake is now highly likely in the next few days.  The most likely mode of failure looks to be this retrogressive erosion, which could lead to a rapid release event, but the initiation of erosion through downcutting cannot be ruled out.  Downstream communities need to be prepared.


I have received no update on the state of flow in the channel - I will post again when some information becomes available.