Search for a project

Predicting the formation and lifetime of landslide dams triggered by earthquakes

Dr Bill Murphy (SEE), Dr Vern Manville (SEE), Chris Massey (GNS Science, Lower Hutt, New Zealand)

Project partner(s): GNS Science, New Zealand (CASE)

Contact email: w.murphy@leeds.ac.uk

Large landslides are an important secondary effect of strong earthquakes. The formation of relatively short-lived landslide dams in river valleys constitutes a significant cascading hazard due to potential breakout floods often resulting in urgent engineering remediation and potential evacuation. The formation of multiple landslide dams after the 2005 (M=7.6) Kashmir Earthquake created long term difficulties. Evidence of distress in the Attabad landslide dam prompted intervention amid fears of collapse of the landslide mass. While the presence of valley-blocking landslides is evident in high mountains, such phenomena also occur in areas of more modest relief. The formation of Lake Waikaremoana in New Zealand is a clear example of such a slope failure. Further evidence of the threat posed was highlighted by the formation of multiple landslide dams triggered by the Kaikoura (M=7.8) earthquake (for example, see figure 1). It is clear therefore that the cascading hazard associated with the formation, lifetime and potential failure of such dams is problematic.

Figure 1 The Hapuku River landslide dam (5/12/2016). Seepage is developing from the face of the dam and side streams are forming ponds on the debris. (Photo: B Lyndsel, GNS Science)

There are numerous areas of uncertainty in understanding these landslide related features. The investigation of these have tended to focus on the dam itself with limited geotechnical input to inform decision making. However, there are three significant problems to be considered.  The first is the stability of the slopes above river systems whilst the second relates to the properties of the landslide mass formed. The third relates to the rate of overland flow, catchment hydrology, and any associated seepage. The aim of this research is to advance our understanding of the first two components of this problem: namely, how to evaluate the likelihood of a big enough landslide being triggered to form a landslide dam and then how to evaluate the properties of the dam itself to allow for a prediction of both failure type and occurrence. There are two significant challenges to predicting the formation and lifespan of landslide dams. These are predicting the size of the causative landslide and also predicting the permeability of the resulting slide mass and how quickly seepage erosion will occur. While it is possible to evaluate the stability of slopes during shaking, evaluating the geometry of the slide mass remains challenging as the distribution of vibration with depth is poorly constrained. In addition, given the dynamics of the stress state and the material properties in natural slopes it is unclear whether current models for ambient stress conditions adequately predict the geometry of  landslides caused by earthquakes. In order to constrain this we will:

  • Data will be collected on the landslides triggered by the Kaikoura Earthquake to examine the geometry of the triggered by the earthquake by comparison to the geometry predicted from available stability analysis methods;
  • Rock mass properties will be investigated to evaluate whether an effective medium model gives a reasonable approximation of mobilised strength and landslide geometry and under what conditions;
  • Block size, block shape and landslide size will be examined in order to compare the source of the landslide dam and the volume of slide mass and how this can best be modelled;
  • Predictive models of hydraulic conductivity of the slide mass will be established to identify the role of seepage and erosion in triggering collapse. The evolution of these systems will be compared with numerical simulations made using FLAC to model the loss of fines contributing strength/permeability changes.

We will produce a predictive model that is based on the geotechnical properties of the ground to evaluate the likelihood of dam formation during earthquakes and estimate the probability of collapse given the properties of the dam.

The successful applicant will be expected to spend a minimum of 3 months working at GNS Science in Lower Hutt, New Zealand and will undertake fieldwork in the northern part of South Island, New Zealand.

Related undergraduate subjects:

  • Civil engineering
  • Earth science
  • Geological science
  • Geology
  • Geophysical science
  • Geophysics
  • Geoscience