B1312
Title: Statistical emulation to quantify uncertainties in tsunami modelling using high performance computing
Authors: Serge Guillas - University College London (United Kingdom) [presenting]
Abstract: Solutions to the investigation of uncertainties in tsunami impacts in three settings are presented. First, we consider landslides as a source of tsunamis from the Indus Canyon in the Western Indian Ocean. We employ statistical emulation, i.e. surrogate modelling, to efficiently quantify uncertainties associated with slump-generated tsunamis at the slopes of the canyon. We demonstrate that the emulator-based approach is an important tool for probabilistic hazard analysis since it can generate thousands of tsunami scenarios in few seconds, compared to days of computations on High Performance Computing facilities for a single run of the tsunami solver. We then examine future tsunami hazard from the Makran subduction zone in the Western Indian Ocean. We capture these phenomena in high resolution (down to 10m) using carefully constructed unstructured meshes for the port of Karachi. An emulator approximates the functional relationship between inputs and outputs maximum velocity and free surface elevation. Finally, we create emulators that respect the nature of time series outputs. We introduce here a novel statistical emulation of the input-output dependence of these computer models: functional registration and Functional Principal Components techniques improve the predictions of the emulator. We apply this approach to the high resolution tsunami wave propagation and coastal inundation for the Cascadia region in the Pacific Northwest.
