26 March 2009

Tangjiashan – images of a potential disaster that was averted

Posted by Dave Petley

This is the second of my series of photographic reviews of the earthquake affected area in Sichuan Province. The other sets are as follows:
Part 1: Beichuan town
Part 2 (this part): The Tangjiashan landslide
Part 3: Hanwang town
Part 4: The Mianyuanhe area
Part 5: The Xingyiu area

On Sunday I was lucky enough to be allowed to visit the landslide site at Tangjiashan, thanks again to my friends from Chengdu University of Technology. To remind you, Tangjiashan was the most hazardous of the 40 or so valley-blocking landslides triggered by the earthquake. Over a period of about a month a team battled heroically to drain it – ultimately succeeding. I blogged about these efforts in detail back in May and June.

Since the earthquake the dam site has been closed, so I was exceptionally fortunate to be allowed to go up there. Access is via a track along the river bed. This will inevitably be lost in the rainy season, which starts in May, so I suspect that few other people will make it up there before the autumn.

A good reference point for this is this pair of NASA ASTER images of the Tangjiashan site and the river down to Beichuan. The image on the left is before the earthquake and on the right is an image from after the event (whilst the dam was still intact. I have annotated on the image pair the location of Beichuan and of the Tangjiashan landslide (as always, click on the image for a bigger version in a new window):

The basic geography of the situation should be clear from the above. The Tangjiashan landslide had blocked the river and a lake was forming behind. Downstream are two smaller landslides that had also blocked the river, and then downstream again was the now ruined town of Beichuan, which still had thousands of dead victims trapped under the rubble.

The landslide scar is clearly shown in the image above. Unfortunately I struggled to get a good image of it as it was very hazy and we were looking into the sun, and am not very adept at improving images with photoshop, so the below is the best that I could manage:

The landslide occurred on a slope that is structurally rather complex, but at least in places it is clear that the landslide slipped on natural joints or bedding planes inclined out of the slope. If these planes are persistent then there is a risk of a further failure. This needs investigation. The image below shows a set of these planes located in the central part of the slope:

The landslide body is large. This image gives an idea of the scale. It was taken from the top of the landslide mass looking down into the channel. If you look very carefully you will see that there is a full-sized, green-blue back hoe excavator working in the channel in the centre of the image:

Similarly, this image was taken from the upstream side of the dam looking down the channel onto the landslide. Again, if you look very carefully you can see two people standing on top of the landslide body:

And here is an image looking upstream onto the landslide body. For scale, note the road up the mass and, if you look very carefully, you can just see a building on top:

When the landslide came down the slope it hit the opposite valley wall. A zoom in on the above image shows a very sharp contact between the valley wall and the landslide body. There is no evidence of there having been an air blast:

Of course after the channel was constructed, half of the dam eroded away. However, there is still a small lake on the upstream side, terminated by a fairly steep bar:

Again, if you look carefully at the above you will see that there is a back hoe in the channel trying to widen it – the scale of this feature becomes clear once you get your eye in to this digger. The channel itself is still quite steep but the bed is mantled with some large blocks of rock:


However, the narrowness of the channel is causing real concern given that the rainy season is only two months away. As the image below shows, the contractors are working very hard indeed at trying to widen the channel – this is a massive effort:


Downstream of the dam the channel is quite wide. The flood plain deposits left by the flood are clear to see. Note the multiple slope failures on the valley walls and the debris flow deposits from the September 2008 rains in the valley mouths (there is a large fan at the far end of the valley floor that must post-date the flood from Tangjiashan). There is a huge volume of sediment waiting to be transported in the tributary valleys and gullies:

The road up the valley to the dam is quite clear on this image – clearly this will not survive the wet season, so the contractors are trying to open / rebuild the old valley side road (this road can be seen on the “before” satellite image):

However, as the above image shows, in opening the road the team are undercutting the slope and leaving it unsupported. Unless some support is emplaced I doubt that this will survive the wet season.

The flood from Tangjiashan swept downstream, impacting the small hydro plant shown on the satellite images above. The central sluice gates of the dam were swept away, although the ones on the edge of the channel survived:

Finally the flood passed through Beichuan itself, fortunately without causing too much damage. This bridge shows the force of the water. It is difficult to imagine what the flood would have been like if the dam had overtopped naturally:

So, that is the current state of affairs at Tangjiashan. As ever, your comments are very welcome. I hope that these images are helpful and interesting.

This is the second of my series of photographic reviews of the earthquake affected area in Sichuan Province. The other sets are as follows:
Part 1: Beichuan town
Part 2 (this part): The Tangjiashan landslide
Part 3: Hanwang town
Part 4: The Mianyuanhe area
Part 5: The Xingyiu area