S U M M A R YWe present a 3-D joint inversion framework for seismic, magnetotelluric (MT) and scalar and tensorial gravity data. Using large-scale optimization methods, parallel forward solvers and a flexible implementation in terms of model parametrization allows us to investigate different coupling approaches for the various physical parameters involved in the joint inversion. Here we compare two different coupling approaches, direct parameter coupling where we calculate conductivities and densities from seismic slownesses and cross-gradient coupling, where each model cell has an independent value for each physical property and a structural similarity is enforced through a term in the objective function.For both types of approaches we see an improvement of the inversion results over single inversions when the inverted data sets are generated from compatible models. As expected the direct coupling approach results in a stronger interaction between the data sets and in this case better results compared to the cross-gradient coupling. In contrast, when the inverted MT data is generated from a model that violates the parameter relationship in some regions but conforms with the cross-gradient assumptions, we obtain good results with the cross-gradient approach, while the direct coupling approach results in spurious features. This makes the cross-gradient approach the first choice for regions were a direct relationship between the physical parameters is unclear.

[1] Risks associated with unstable rocky slopes are growing as a result of climate change and rapid expansions of human habitats and critical infrastructure in mountainous regions. To improve our understanding of mountain slope instability, we developed a microseismic monitoring system that operates autonomously in remote areas afflicted by harsh weather. Our microseismic system comprising 12 three-component geophones was deployed across $60,000 m 2 of rugged crystalline terrain above a huge (30 million m 3 ) recent rockfall in the Swiss Alps. During its 31-month lifetime, signals from 223 microearthquakes with approximate moment magnitudes ranging from À2 to 0 were recorded. Determining the hypocenters was challenging for several reasons: (1) P wave velocities were highly heterogeneous, varying abruptly from <1.5 km/s to >3.8 km/s. (2) First-break picks were either inaccurate or lacking for some microearthquakes. (3) There were no reliable S wave picks. (4) Numerous microearthquakes occurred just outside the network boundaries. These issues were addressed by using a three-dimensional (3-D) P wave velocity model of the mountain slope determined from refraction tomography in a nonlinear inversion for hypocenter parameters and their probability density functions. Recordings from geophones at different altitudes and in boreholes constrained microearthquake depth estimates. Most microearthquakes were concentrated within 50-100 m of the surface in two zones, one that followed the recent rockslide scarp and one that spanned the volume of highest fracture zone/fault density. These two active zones delineated a mass of rock that according to geodetic measurements has moved toward the scarp at 1-2 cm/yr.

No abstract

As transport routes and population centers in mountainous areas expand, risks associated with rockfalls and rockslides grow at an alarming rate. As a consequence, there is an urgent need to delineate mountain slopes susceptible to catastrophic collapse in a safe and noninvasive manner. For this purpose, we have developed a 3D tomographic seismic refraction technique and applied it to an unstable alpine mountain slope, a significant segment of which is moving at 0.01-0.02 m/year toward the adjacent valley floor. First arrivals recorded across an extensive region of the exposed gneissic rock mass have extraordinarily low apparent velocities at short ͑0.2 m͒ to long ͑Ͼ100 m͒ shot-receiver offsets. Inversion of the first-arrival traveltimes produces a 3D tomogram that reveals the presence of a huge volume of very-lowquality rock with ultralow to very low P-wave velocities of 500-2700 m/s. These values are astonishingly low compared to the average horizontal P-wave velocity of 5400 m/s determined from laboratory analyses of intact rocks collected at the investigation site. The extremely low field velocities likely result from the ubiquitous presence of dry cracks, fracture zones, and faults on a wide variety of scales. They extend to more than 35 m depth over a 200ϫ 150-m area that encompasses the mobile segment of the mountain slope, which is transected by a number of actively opening fracture zones and faults, and a large part of the adjacent stationary slope. Although hazards related to the mobile segment have been recognized since the last major rockslides affected the mountain in 1991, those related to the adjacent low-quality stationary rock mass have not.