Studying glacial and mountain hazards is necessary as the world seems to face major climate changes, which may increase the frequency of catastrophic events, e.g., debris flows. The latter processes are well understood, but changes in their frequency and relationships to climatic variability are not. In the present study, aiming at imaging an entire debris-flow system from the release area to the depositional area, correlations of lake sediments to specific slope deposits on land are looked for by using a combined approach including geology, sediment-core analysis and geophysical surveys (ERT, GPR and seismic). Our overall goal is to create a chronology of debris-flow events from lake Leirvatnet in order to be able to compare debris-flow records with existing paleo-climatic data and evaluate the climatic impact on debris-flow frequency in a long-term perspective. Following the mapping and thickness estimate of the sediment cover on the lake bottom obtained from GPR, we are evaluating the spatial continuity of reflectors associated with sections of high debris-flow activity to relate them to debris-flow deposits on land (imaged by ERT, GPR and seismic), having a complete Holocene sedimentological history from the lake. Actual results indicate a large temporal variability in debris flows.

Numerous glacier lakes have formed in recent decades due to worldwide glacier retreat induced by climate change. These lakes, dammed by glaciers and moraine ridges, are hazardous because of potential glacial lake outburst flows (GLOF). The GLOF probability is increasing in the Russian Central Caucasus, like at the Bashkara glacier which has been extensively studied, but detailed information about the ground is missing. A pilot geophysical campaign carried out during summer 2009 tested GPR and resistivity profiling at this site, using towed-systems to facilitate acquisition. The GPR measurements were successful with penetration depth down to 70 m on icy ground, though the acquisition was difficult due to rough ground terrain. The results show that GPR measurements would greatly improve the knowledge of the internal structure of that complex zone, thus helping for hazard assessments, but more field work is needed, including CMP measurements. The resistivity measurements were not that successful, the towed system requiring repeating each profile with increasing offset, the progression on the ground being heavy. Only the very first meters of the ground were retrieved, i.e, not really providing useful information. Results and experience gathered in 2009 are now analysed to plan another campaign summer 2011.

A debris flow occurred on 8 May 2004, in Fjǽrland, Western Norway, due to a Glacial Lake Outburst Flood and a natural terminal moraine failure. The site was investigated in 2004 and 2005, using pre- and post-flow aerial photos, airborne laser scanning, and extensive field work investigations, resulting in a good understanding of the mechanics of the debris flow, with quantification of the entrainment and determination of the final volume involved. However, though the moraine had a clear weak point, with lower elevation and erosion due to overflowing in the melting season, the sudden rupture of the moraine still needs to be explained. As moraines often contain an ice core, a possible cause could be the melting of the ice, inducing a progressive weakening of the structure. Geophysical investigations were therefore carried out in September 2006, including seismic refraction, GPR and resistivity. All methods worked well, but none revealed the presence of ice, though the depth to bedrock was determined. On the contrary, the moraine appeared to be highly saturated in water, especially in one area, away from the actual breach and corresponding to observed water seepage at the foot of the moraine. To estimate future hazard, water circulation through the moraine should be monitored over time.

On June 20th, 1996, a submarine land failure in a fjord near the village of Finneidfjord (Northern Norway) developed into a retrogressive quick-clay slide encroaching 100-150 m inland, resulting in the loss of life and significant material damage. Reports in the aftermath of the slide concluded that build up of excess pore pressure may have been the main reason for the slide, though other causes are invoked too. Several field campaigns in the fjord since 1997 consisted of sediment coring, VHR seismic surveys, and long term in-situ instrumentation. To complete offshore data with land data, several geophysical techniques were tested onshore in July 2007. Seismic, GPR and resistivity (method of choice in quick clay) were tried at two sites (one intact and one close to the 1996 slide scar), and all three techniques proved to be useful for a better definition of the underground. Despite clay-prone sites, a 50 MHz GPR antenna was used with success on one site because of the unknown presence of a thick layer of morainic material. Though standard refraction seismic profiles were acquired for P-wave tomography, surface-wave processing tests on the same data gave promising results.

B-17 LOCAL IMAGING APPROACH AND APPLICATIONS Summary 1 Local imaging obtained using a local plane-wavenumber approach of Generalized Diffraction Tomography is very similar to Synthetic Aperture Radar imaging in spotlight-mode acquisition. A local image of the scattering structures around any point in a model is calculated using ray-based Green’s functions computed at that point only. The cost of the local imaging is very small compared to classic imaging which needs grids of Green’s functions. The signal-processing approach inherent to this FFT-based local imaging allows also a very powerful and flexible imaging with full control of aliasing and noise filtering. The