Eugene C. Morgan

Shallow Landslides and Their Dynamics in Coastal and Deepwater Environments, Norway

In this manuscript, we present the first results of integrated slope stability studies to investigate smaller-scale mass movement processes in different physiographic settings of Norway. These include coastal areas (Sorfjord, Finneidfjord), and pristine open ocean settings in intermediate (Vesteralen) and deep waters (Lofoten) on the Norwegian margin. Triggers, pre-conditioning factors and sedimentary processes associated with these landslides are currently not well constrained.

Finneidfjord: a Field Laboratory for Integrated Submarine Slope Stability Assessments and Characterization of Landslide-Prone Sediments: A Review

Sorfjord outside the village of Finneidfjord has a history of landsliding throughout the Holocene. The 1996 landslide – the focus of this study – has many characteristics typical of submarine landslides (well-developed slip plane, outrunner blocks, peripheral thrusting and lateral spreading). Due to its sheltered and accessible location, Finneidfjord has become a natural laboratory for testing high-resolution and multidisciplinary techniques to improve our understanding of landslide development. This study integrates multiple sediment cores, swath-bathymetry surveys, single- and multi-channel 2D seismic data (Topas, boomer, sparker, airgun), very-high-resolution 3D chirp seismics, ocean-bottom seismometer as well as free fall and traditional cone penetration testing (CPTU). The cores have been subjected to both geological and geotechnical laboratory analyses. Of particular interest is the correlation of the regional slip plane as a high-amplitude package of reflections in the geophysical data with the results of the sediment and in situ measurements. Comparison of 3D traces with synthetic seismograms based on multi-sensor core logs show that the most prominent slip plane lies within a thin clay unit sandwiching a sand seam. The slip plane is difficult to identify from CPTU data alone. The top part of this composite unit has in places been eroded under the 1996 mass-transport deposit (MTD). This composite unit’s formation is associated with turbidite deposits from terrestrial quick clay landslides and possibly river floods in the catchment of the fjord. While the MTD is extensively deformed, different flow facies are identified within the landslide body revealing a complex, multi-phase failure. The seismic data were also used to infer physical properties (mean grain size, gas saturation from P-wave attenuation). Interestingly, shallow gas adjacent to the landslide appears not to have played a role in the landslide development. Fjordbed stability is strongly influenced by shallow subsurface structure, with geotechnical properties and lateral continuity of stratified beds acting as primary controls on slide plane depth and failure mechanisms. This study can well form a template for near-shore areas prone to landsliding. Currently, a long-term pore pressure monitoring programme is in progress, after the installation of several piezometers close to the depths of the slip plane close to the shoreline in September 2012.

Evaluating Gas-Generated Pore Pressure with Seismic Reflection Data in a Landslide-Prone Area: An Example from Finneidfjord, Norway

On the 20th of June, 1996, a multi-phase landslide that initiated under water and retrogressed onto land ultimately killed four people, destroyed several houses, and undermined a major highway in Finneidfjord, Norway, an area with a known history of landsliding in the Holocene. Geological and environmental conditions inherent to the 1996 slide include excess fluid/gas pressure (particularly in gas-bearing sediment). In this study, we quantify pore pressures within the free gas accumulation at very shallow sub-surface depths using seismic reflection data. The gas front (a few meters below the seabed) produces a strong, polarity-reversed reflection, dramatically attenuating sub-surface reflections. On x-ray images of cores collected from the 5 km2 large gas zone, gas appears as vesicular spots. We use a previously published method incorporating continuous wavelet transforms to quantify attenuation produced by gas-bearing sediment. Taking the output from this method, and knowing or assuming values for other physical parameters, we invert for in situ pressure and equivalent thickness of the free gas layer. We compare our results to pressure data collected from a single piezometer penetrating the gas front. This analysis demonstrates the utility of seismic reflection data in analyzing the dominant parameter in submarine slope stability (i.e. excess pore pressure), which could be useful in assessing geohazards in similar geological environments.

Characterization of the slope-destabilizing effects of gas-charged sediment via seismic surveys

Finneidfjord, Norway has a history of submarine slope failure and hosts areas of buried, gas-charged sediment. Using shipboard seismic survey data from Finneidfjord, we illustrate how novel seismic data processing techniques can yield estimates of geotechnical properties of this gas-charged sediment. This data processing involves two steps: 1) estimate the amount of seismic attenuation that a gas-bearing layer creates, and 2) fit a wave attenuation model to the observed attenuation. We accomplish the first step using a wavelet-based spectral ratio approach, and find the seismic quality factor (inversely related to attenuation) from the change in amplitude spectra across the gas layer. The novelty of this paper really comes in the second step, where we create a Bayesian hierarchical model that combines a wave attenuation model with a spatial random (Gaussian) process model. The advantage of this combination is that our estimates of the gas properties not only must conform to the observations of attenuation (quality factor), but must also smoothly vary over space, as one would expect of any natural process. Ultimately, we end up with posterior distributions of gas properties at points throughout our gas-bearing layer. We use these posterior distributions to probabilistically evaluate the role that gas plays in slope failures in Finneidfjord. This method has other applications in natural gas resource exploration and carbon sequestration.

Assessing the Probability of Occurrence of Earthquake-induced Landslides Offshore the U.S. East Coast: A First-order, Second Moment Approach

Submarine landslides pose a direct threat to offshore infrastructure, and an indirect threat to coastal communities via tsunami generation. Recent studies have investigated the potential role that submarine landslides play in causing tsunamis on the U.S. East Coast. This paper quantitatively assesses submarine landslide hazard offshore Long Island and New Jersey, as an example, but the method herein can be applied to the entire Atlantic margin. Using publicly available bathymetry, surficial sediment data, undrained shear strength values, and earthquake ground motion predictions, we map the conditional probability of slope failure over our entire study area. We calculate this probability using a first-order, second moment estimate of the variance of critical acceleration needed to overcome the resisting forces in the infinite slope stability analysis. We show that this first-order, second moment approximation serves as a convenient and computationally efficient way of assessing submarine landslide hazard over a broad region, while also accounting for the significant uncertainties in the slope stability parameters.

A Case Study of Alternative Site Response Explanatory Variables in Parkfield, California

The combination of densely-spaced strong-motion stations in Parkfield, California, and spectral analysis of surface waves (SASW) profiles provides an ideal dataset for assessing the accuracy of different site response explanatory variables. We judge accuracy in terms of spatial coverage and correlation with observations. The performance of the alternative models is period-dependent, but generally we observe that: (1) where a profile is available, the square-root-of-impedance method outperforms VS30 (average S-wave velocity to 30 m depth), and (2) where a profile is unavailable, the topographic-slope method outperforms surficial geology. The fundamental site frequency is a valuable site response explanatory variable, though less valuable than VS30. However, given the expense and difficulty of obtaining reliable estimates of VS30 and the relative ease with which the fundamental site frequency can be computed, the fundamental site frequency may prove to be a valuable site response explanatory variable for many applications.

Quantitative Seafloor Geomorphology and Offshore Geohazards

Seafloor geomorphology reflects dominant geologic and tectonic processes. By quantifying the scale of geomorphic features, we can gain insight into the mechanisms that drive the erosive processes. If erosive processes follow frequency-magnitude relationships that are related to the driving mechanisms as they do on land (e.g. big and infrequent storms will cause lots of large landslides whereas small, frequent rainfall events will cause smaller landslides), then it stands to reason that the largest erosive features (i.e. landslides) are caused by infrequent events. Quantifying the dimensional and spatial variability of large-scale erosive features is crucial in characterizing future mass-wasting events. Using the GEBCO global bathymetry data, we are classifying continental margins based on their morphology. Calculating semivariance of submarine topography establishes the scales of different erosive features (via the sill) and gives a measure of the distribution of these features (via the range). Margins dominated by gullies and closely-spaced canyons should generate semivariograms with relatively small sills and ranges. These erosive features suggest frequent, small-scale erosive events, likely triggered by frequent earthquakes (here, ‘frequent’ being anywhere from 1 per 100 years to 1 per 1000 years). On these margins, slope failure largely depends on the nature of earthquakes (e.g. segmentation and ‘slow’ vs. ‘fast’ events). In contrast, margins with large and well-preserved landslides should yield semivariograms with large sills and ranges, and are suggestive of a less frequent process, such as sea level change with interspersed intervals of relative quiescence. These geostatistical characteristics can then be applied towards assessing submarine landslide and tsunami risk to coastal areas across the globe.