Abstract Marine turbidite records have been used to infer palaeoseismicity and estimate recurrence intervals for large (>M w 7) earthquakes along the Cascadia Subduction Zone. Conventional models propose that upper slope failures are funneled into submarine canyons and develop into turbidity flows that are routed down-canyon to deep-water channel and fan systems. However, the sources and pathways of these turbidity flows are poorly constrained, leading to uncertainties in the connections between ground shaking, slope failure and deep-water turbidites. We examine the spatial distribution of submarine landslides along the southern Cascadia margin to identify source regions for slope failures that may have developed into turbidity flows. Using multibeam bathymetry, sparker multichannel seismic and chirp sub-bottom data, we observe relatively few canyon head slope failures and limited evidence of large landslides on the upper and middle slope. Most of the submarine canyons are draped with sediment infill in the upper reaches and do not appear to be active sediment conduits during the recent sea-level highstand. In contrast, there is evidence of extensive mass wasting of the lower slope and non-channelized downslope flows. Contrary to previous studies, we propose that failures along the lower slope are the primary sources for deep-sea seismoturbidites in southern Cascadia.

Abstract The Queen Charlotte Fault defines the Pacific–North America transform plate boundary in western Canada and southeastern Alaska for c. 900 km. The entire length of the fault is submerged along a continental margin dominated by Quaternary glacial processes, yet the geomorphology along the margin has never been systematically examined due to the absence of high-resolution seafloor mapping data. Hence the geological processes that influence the distribution, character and timing of mass transport events and their associated hazards remain poorly understood. Here we develop a classification of the first-order shape of the continental shelf, slope and rise to examine potential relationships between form and process dominance. We found that the margin can be split into six geomorphic groups that vary smoothly from north to south between two basic end-members. The northernmost group (west of Chichagof Island, Alaska) is characterized by concave-upwards slope profiles, gentle slope gradients (<6°) and relatively low along-strike variance, all features characteristic of sediment-dominated siliciclastic margins. Dendritic submarine canyon/channel networks and retrogressive failure complexes along relatively gentle slope gradients are observed throughout the region, suggesting that high rates of Quaternary sediment delivery and accumulation played a fundamental part in mass transport processes. Individual failures range in area from 0.02 to 70 km 2 and display scarp heights between 10 and 250 m. Transpression along the Queen Charlotte Fault increases southwards and the slope physiography is thus progressively more influenced by regional-scale tectonic deformation. The southernmost group (west of Haida Gwaii, British Columbia) defines the tectonically dominated end-member: the continental slope is characterized by steep gradients (>20°) along the flanks of broad, margin-parallel ridges and valleys. Mass transport features in the tectonically dominated areas are mostly observed along steep escarpments and the larger slides (up to 10 km 2 ) appear to be failures of consolidated material along the flanks of tectonic features. Overall, these observations highlight the role of first-order margin physiography on the distribution and type of submarine landslides expected to occur in particular morphological settings. The sediment-dominated end-member allows for the accumulation of under-consolidated Quaternary sediments and shows larger, more frequent slides; the rugged physiography of the tectonically dominated end-member leads to sediment bypass and the collapse of uplifted tectonic features. The maximum and average dimensions of slides are an order of magnitude smaller than those of slides observed along other (passive) glaciated margins. We propose that the general patterns observed in slide distribution are caused by the interplay between tectonic activity (long- and short-term) and sediment delivery. The recurrence (<100 years) of M > 7 earthquakes along the Queen Charlotte Fault may generate small, but frequent, failures of under-consolidated Quaternary sediments within the sediment-dominated regions. By contrast, the tectonically dominated regions are characterized by the bypass of Quaternary sediments to the continental rise and the less frequent collapse of steep, uplifted and consolidated sediments.

Abstract The Santa Cruz Basin (SCB) is one of several fault-bounded basins within the California Continental Borderland that has drawn interest over the years for its role in the tectonic evolution of the region, but also because it contains a record of a variety of modes of sedimentary mass transport (i.e., open slope vs. canyon-confined systems). Here, we present a suite of new high-resolution marine geophysical data that demonstrate the extent and significance of the SCB submarine landslide complex in terms of late Miocene to present basin evolution and regional geohazard assessment. The new data reveal that submarine landslides cover an area of ~160 km 2 along the eastern flank of the Santa Rosa–Cortes Ridge and have emplaced a minimum of 9 to 16 km 3 of mass transport deposits along the floor of the SCB during the Quaternary. The failures occur along an onlapping wedge of Pliocene sediment that was uplifted and tilted during the later stages of basin development. The uplifted and steepened Pliocene strata were preconditioned for failure so that parts of the section failed episodically throughout the Quaternary—most likely during large earthquakes. Once failed, the material initially translated as a block glide along a defined failure surface. As transport continued several kilometers across a steep section of the lower slope, the material separated into distinctive proximal and distal components. The failed masses mobilized into debris flows that show evidence for dynamic separation into less and more mobile components that disturbed and eroded underlying stratigraphy in areas most proximal to the source area. The most highly mobilized components and those with the lowest viscosity and yield strength produced flows that blanket the underlying stratigraphy along the distal reaches of deposition. The estimated volumes of individual landslides within the complex (0.1–2.6 km 3 ), the runout distance measured from the headwalls (>20 km), and evidence for relatively high velocity during initial mobilization all suggest that slides in the SCB may have been tsunamigenic. Because many slopes in the California Continental Borderland are either sediment starved or have experienced sediment bypass during the Quaternary, we propose that uplift and rotation of Pliocene deposits are important preconditioning factors for slope failure that need to be systematically evaluated as potential tsunami initiators.