ORE/CIMAR Seminar: Nearshore Wave Model for Assessment of Coastal Flooding

May 11, 3:30pm - 4:30pm
Mānoa Campus, Holmes Hall 247 & Zoom (see description for meeting ID and passcode)

Quantitative assessments of wave-driven run-up and inundation scenarios have been of high interest for coastal residents, engineers, emergency managers, and scientists. Computational approaches of coastal flooding from waves only have been widely based on empirical formulations but nowadays make increasingly use of numerical models. Due to the complexity of the problem, numerical computations of wave run-up are mostly based on solutions of the free surface elevation and velocity components of individual nearshore waves. Obviously, a phase-resolving computation requires high grid resolution and, thus, is computationally much more demanding than a solution from a spectral model. Consequently and from an operational perspective, it is crucial to optimize phase-resolving models with the focus on cost-effective reliable run-up assessments. We present the ongoing development of a new numerical model for nearshore wave transformation and run-up/inundation. Our work focuses on improving both the accuracy and speed of the numerical solution. The new model is built around a lean numerical structure combined with a GPU implementation that enables real-time run-up predictions on commodity hardware over substantial segments of the coastline. We also employ a nested domain approach, which provides a targeted grid refinement and therefore achieves high accuracy solutions locally at a low computational cost. This fundamental work has led to further insights into the key elements, which are crucial for accurate wave run-up computations. We have investigated several numerical schemes with respect to their impact on preserving the underlying properties of the governing equations such as the ability to form shocks and to describe the moving boundary along with the wet/dry interface. We have also assessed the diffusivity of several conventional numerical schemes that can unfavorably alter the wave properties and subsequently the run-up solution. We further highlight the benefits of the nested grid approach with respect to low-frequency and high-frequency contributions to wave run-up. Finally, we will demonstrate the computational efficiency of GPU processing over conventional CPU parallelization techniques. This seminar will be given by Fatima-Zahra Mihami, PhD student at Université de Pau et des Pays de l’Adour, France and Dr. Volker Roeber, E2S Chair HPC-Waves. Zoom Meeting ID: 960 4654 5799, Passcode: OREseminar


Event Sponsor
Ocean and Resources Engineering, Mānoa Campus

More Information
(808) 956-7572, https://www.soest.hawaii.edu/ore/event/seminar_220511/

Share by email