Probabilistic estimation of turbidity current risk for offshore developments under changing climate conditions

Recent advances in direct monitoring of submarine canyons activity suggest that metocean events can prime or ultimately trigger turbidity currents at sub-annual to centennial frequencies, challenging the paradigm that such highly destructive flows typically result from infrequent submarine landslides. Here we present a methodology aimed to link information on nearshore circulation and sediment resuspension events such as floods, tides and storms to the downslope evolution of turbidity flows, applied to cyclone-triggered turbidity currents affecting a deep-sea gas development offshore northern Mozambique. We evaluate seabed current velocities at the canyon heads derived from a 10,000 year long synthetic cyclone database consistent with expected future climate conditions. We couple the results for each event to a dedicated 1D turbidity current numerical model accounting for multiple grain-sizes, water entrainment and detrainment, overspill and spatial variations of the channel width and seabed gradient. The 1000's of individual modelled events allow a statistically consistent estimation of the return period of turbidity currents at different control points along the canyons, and provide an estimation of flow structure, density and velocity necessary to evaluate the risk for offshore developments. Results are validated by the analysis of core samples, seabed morphology and dedicated 3D modelling of critical events.