Multi-scale calibration of a non-hydrostatic model for wave runup simulation

Modeling wave runup on beaches and structures is an important step toward accurate coastal flood prediction. Depth-averaged non-hydrostatic models such as XBNH (XBeach Non-Hydrostatic) are computationally costeffective yet accurate tools to numerically simulate random wave runup. XBNH includes a number of calibration parameters, which are determined on a case-study basis. This study aims to investigate the range of XBNH input parameters that results in accurate predictions regardless of the application scale, benefiting future applications of XBNH that lack the measurements needed for model calibration. The study performs calibration and sensitivity analysis to investigate the model response to the input parameters for applications with different spatial scales, including small- and large-scale laboratory wave runup experiments and a real-scale field experiment. A Grid search method is utilized to perform a multi-scale calibration of the input parameters used in the parametrization of the wave breaking process in XBNH. The effect of parameters in the JONSWAP spectrum used to generate the offshore boundary conditions is investigated. Results show that using a single combination of calibrated input parameters, the model is able to capture random wave runup on beach, beach-dune, and beach-seawall systems with various spatial scales. Moreover, wave spectrum parameters can highly influence the model performance, and calibration of these parameters makes the model more robust to replicate the actual water surface fluctuations.