Zhouqu, China, disaster - data capture, modelling and preliminary geohazard assessment

At 1am on Sunday the 8th of August a large flow slide took place in Zhouqu, a county town in SW Gansu Province, China. Current information suggests that the trigger was a rainfall event delivering a reported 97mm in less than 40 minutes (annual precipitation averages from 400 to 900mm). Antecedent rainfall may also have played a role. The devastating event caused the deaths of more than 1000 people (1,144 as of Friday 13th of August in Zhouqu alone with at least 600 people still missing). In the wider region, mudslides and debris flows continue to affect the local population with further fatalities in Longnan (south and downstream of Zhouqu) and Tianshui (to the northeast). The region is characterised by a seismically active mountainous terrain, an extension zone of the Tibetan uplift, and was affected by the Wenchuan earthquake (12/5/2008) that generated and re-activated a large number of landslides. At Zhouqu, the local erosion base is formed by the Bailong River at about 1300m asl. The surrounding peaks reach heights of about 3500 to 3750m. The catchment where the sediments originate is approximately 20km2 and is characterised by very steep slopes. Vegetation cover is sparse with only a small remnant of upland forest cover remaining. Following the disaster a lively debate has sprung up apportioning blame to deforestation, hydropower development and mining in the area. In order to learn lessons from this kind of events, instead of speculation, a scientific approach is required to collect data, verify their robustness and make impartial observations and develop reliable process models that will be essential to develop appropriate mitigation strategies for the future. The site was visited in November 2009 by Dijkstra and Gibson and discussions were held with the government of Zhouqu about the risk posed by large landslides in the area. It is vitally important to collect data soon. As with all landslide events, an understanding of local geological structure, drainage, morphology and geotechnical parameters are essential in building a robust model of landsliding (e.g. Geertsma et al. 2006). The key characteristics of landslides and landslide dams in this region are often removed rapidly by the very instability that causes them. It is inherently difficult to map out the potential for landslides to transfer into catastrophic viscous flows with complex and transient mechanical and rheological interactions. However, this is an essential component in any reliable assessment of the risks. These sort of assessments are rare because of the scarcity of events available for this type of research. When these occur we must make the most of it. Mostly, our understanding of these processes is derived from small-scale laboratory experiments, requiring upscaling that may affect the relevance of the observations. In this case, a unique field laboratory is waiting to be explored to gain a better understanding of the risks posed by geohazards to the population in this region.