Uncertainty Quantification of geophysical models

In this talk, we show various strategies for the calibration and emulation of simulators having uncertain inputs and internal parameters, with applications to CFD, tsunami wave modelling, simulated earthquakes catalogues and climate models. We first illustrate fully Bayesian calibration and emulation techniques on the k-epsilon turbulence CFD model. For the landslide-generated tsunami model, a fast surrogate is provided by the outer product emulator. It can enable either real-time warnings according to uncertain speed, position and shape of the landslide, or full uncertainty quantification for hazard assessment. We also show some new realistic simulation and hazard assessment of earthquake-generated tsunami risk in Cascadia (Western Canada and USA), using VOLNA. VOLNA is a solver of nonlinear shallow water equations on unstructured meshes that is now accelerated on the GPU system Emerald.
As for the emulation of earthquake catalogues, stochasticity of the outputs is taken into account trough the distribution of inter-events times, in the framework of non-extensive statistical physics. Finally, enhancements in terms of sequential design of experiments are shown for the climate model CESM, where points in the design are chosen according to novel adaptive strategies.