New images of Beichuan and its lakes

CCTV has posted the following image of the lake at Beichuan:

Meanwhile Reuters have published this image, captioned: "An aerial view of the roofs of houses in earthquake-hit Beichuan, Sichuan province May 20, 2008. The houses, which were originally built next to a reservoir, were uprooted and floated downriver after water levels started to rise." I suspect that by reservoir they mean river?


AP has this image of the landslides in Beichuan, captioned "The ruins of collapsed buildings lay among those still standing following last week's earthquake, in Beichuan, in China's southwest Sichuan province Thursday, May 22, 2008."

Beichuan landslide-dammed lake - new imagery

NASA has now released new images collected by the FORMOSAT-2 satellite of one of the lakes forming near to Beichuan in Sichuan behind the landslide dam shown in this post (click on the image for a better view).

The first image shows a reach of the river upstream of the dam after the earthquake (on May 14th) but before the lake really started to form. The second image shows the lake a day later. Note that now the lake has started to fill, threatening the two villages built on the river banks. The river has overtopped its banks and is starting to inundate the farm land. In the third image, collected on 19th May, the lake is well formed. River flow has more or less stopped and a large pool has formed. Both villages and the farmland have been inundated. The only sign of them is a large amount of debris floating on the surface. This may well be all that is left of people's houses.

Note that the volume of water behind the dam is large and that the lake is filling quickly. Although the authorities appear to be responding quickly and appropriately, this situation remains abolsutely critical. One can only hope that the reports in the Times yesterday, which said the following, are incorrect:

One plan is to build up earth barriers halfway across the Beichuan river, at several points downstream from the reservoir. These barriers would divert the water, slowing its flow and reducing the danger to communities in its path. Once these were ready, some or all of the water could be released over the next day or two, possibly by blasting away part of the landslide.

Such an approach would be unconventional at the very least. Blasting the landslides, rather than building a spillway, is not an experiment that I would recommend. The idea of controlling the resultant flow with barriers is at best optimistic.

The problem of cracks in mountains after earthquakes

Xinhua is today reporting the following:

Nearly 9,000 people in a quake zone were evacuated on Tuesday for fear that huge cracks on a mountain could lead to further disasters. Many crevices, measuring up to 1,500 meters long, 250 m high and 50 centimeters wide, have been spotted on the Shiziliang Mountain in the Qingchuan County seat, Guangyuan City, since Sunday, threatening about 50,000 people and quake-relief soldiers. Part of the mountain, distorted in the 8.0-magnitude earthquake and many aftershocks over the past week, has sunk about 1 meter and caused many road works to cave in. A house at the foot of the mountain was damaged by a falling boulder. Any new aftershock or heavy rainfall in the area could trigger serious landslides and cause casualties, according to the quake relief headquarters at the site. The local government on Tuesday launched an emergency evacuation of 9,000 residents near the mountain. The area has been cordoned off and is under round-the-clock monitoring.

Reading this gave me a very strong sense of deja vu. One of the greatest challenges in Pakistan after the 2005 earthquake was trying to interpret the meaning of similar crack networks, which we found all over the hillsides. The following two images show what these looked like.

A cracked slope near to Muzaffarabad in Kashmir after the 2005 earthquake (click for a bigger image). Cracked slopes like this occurred throughout the earthquake zone. To date very few have led to a slope failure, but we remain very concerned by them.

A further cracked slope near to Muzaffarabad in Kashmir after the 2005 earthquake (click for a bigger image). Note that in this case the cracks are very close to a steep cliff. Nonetheless, this slope has still not failed.

In Pakistan we assumed that these were incipient landslides and worked with the authorities to move people off the worst affected areas before the monsoon. We also put instruments to measure four of the most notable sites. However, this was only an assumption. As it turned out, almost none of these slopes have to date failed. We don't know why. Are they in fact not landslides at all - if so, what are they? If they are landslides, are they in fact sufficiently stable that they can only move in a large earthquake? Or are they incipient landslides, but they have not failed yet as the rainfall associated with the two monsoons (2006 and 2007) since the earthquake has not been intense in this area (which is the case)? I must admit that I err on the latter, and worry greatly about what will happen in Pakistan in the next really strong monsoon.

In Sichuan, I would be willing to bet that close inspection will reveal that these cracks also occur very extensively across the landscape. Once again the same conundrum will occur. In the case described by Xinhua above the authorities are absolutely right to be cautious, especially if it is actually moving. Evacuation and monitoring appears to be the correct approach. I suspect that this drama will play out in many other places. In Kashmir essentially no resource was dedicated to finding out what these cracks actually meant. I hope very much that this is not repeated in Sichuan.

Beichuan landslides from Google Earth

The FORMOSAT-2 image of the Beichuan area is now available on Google Earth (see here). The image is interesting as it shows that the landslides were incredibly spatially variable - for example, in the Beichuan area (which is junction of the rivers on the middle right of the image - the Google label is a little way from the main part of the town) the occurrence of landslides is very high and the river is blocked in at least three places. Elsewhere, the occurrence of landslides is much lower.

FORMOSAT-2 image of the Beichuan area, from Google Earth. Click for a larger view.

This allows a first attempt at perspective view of the Beichuan area using the Google Earth rather basic terrain model. Note that the imagery and the model aren't quite aligned correctly, so the picture is a bit confused, but it is a good starting place. I have annotated the image below to indicate the major landslides and places where the river is blocked. As usual, click on the image to get a decent view (it will open in a new window).

Annotated FORMOSAT-2 image of the Beichuan area, using the Google Earth terrain model to give a perspective view. Click for a larger view.

Finally, as mentioned above elsewhere the density of landslides is lower. However, close inspection shows that there are a large number of landslides in the landscape. As the image below shows, these display the typical features of mass movements triggered by earthquakes, which is that they start right at the top of the slope near the ridge, then travel most or all of the way down the slope. This is because the maximum shaking occurs at the ridges due to a process known as topographic amplification. This is consistent with the observations of witnesses, who said things like (please excuse the link to the Sun): “The mountains just seemed to explode as if they’d been detonated with dynamite."

FORMOSAT-2 image of an area near Zicheng, using the Google Earth terrain model to give a perspective view. Note the landslides in the middle of the image start from the ridge and extend to the valley floor. Click for a larger view.

Sichuan - further images of giant landslides

Two further images have been sent to me of giant landslides from the Sichuan earthquake. At the moment I cannot find out where they are from, although they clearly both show the same giant landslide. Note that the valley is completely filled (which means that a lake if forming somewhere upslope) and the amazing flow structures in the landslide mass. I wonder if this is the same landslide as is featured in my earlier post.

Updated: Sichuan Earthquake aftermath - landslides are taking a terrible toll

In previous posts I have warned that landslides will wreak a terrible toll in the aftermath of the earthquake. Sadly, this appears to be all too true. Xinhua is today carrying this report:

More than 200 relief workers, engaged in repairing the broken roads in the quake-stricken Sichuan Province, had been buried in mud flows over the past three days, said Dai Dongchang, an official with the Ministry of Transport on Monday. Two construction machines and six vehicles were also buried in the mud flows, he said.

Given that the summer rains have not really started as yet, the problems that this area faces are all too clear.

Update: the report has now been changed to:
Five vehicles were buried over the weekend by mud flows as they attempted to leave quake-ravaged regions of Sichuan Province, and an estimated 158 people were known to have been on board, a transport official told Xinhua on Monday. The death toll was not confirmed as rescue work was still underway, said Director of Road Traffic Dai Dongchang with the Ministry of Transport in an interview. Dai earlier told the central government website (www.gov.cn) that more than 158 relief workers with the Sichuan transportation department were trapped by mud flows. The identities of the victims were yet to be confirmed, he said. The official confirmed two construction machines were buried by mud flows but the drivers escaped in time, with no injuries. The official admitted that mud flows and aftershocks following the deadly quake and blocked rivers swollen by heavy rain had hampered restoration of roads and rescue work in the quake zone.

Beichuan - what happens when a massive rockslide hits a town

I am not normally a fan of Chicken Noodle News (CNN), but they have an interesting report detailing the damage in Beichuan resulting from the impact of the rockslide. It is worth a view. I cannot find a way to link to the original report, but there is a good quality version of it here.

AP image captioned: Rescue workers search collapsed buildings in Beichuan, China's southwest Sichuan province, Sunday May 18, 2008.