SWAN vs. DG-WAVE: A Comparison of Numerical Wave Prediction Models
Loading...
Date
2013-12
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
The Ohio State University
Abstract
Established wave models are frequently used by government organizations such as NOAA (National Oceanic and Atmospheric Administration), the U.S. Army Corps of Engineers, FEMA (Federal Emergency Management Agency), and the Louisiana State University Hurricane Center. The Simulating Waves Near-shore (SWAN) model is a widely used spectral wave model, using the two-dimensional variance density spectrum to solve the action balance equation in five dimensions . SWAN solves its governing equation, the action balance equation, iteratively at discrete frequency and direction bins until some break-off criteria are met. This method performs well in application, but is computationally expensive and thus inefficient. This noted inefficiency leads to an interest in exploring the potential of a simpler wave model, the GLERL-Donelan wave model. This model utilizes a parametric approach to simplify the problem and reduce the number of dimensions, solving the momentum balance equation. A discontinuous-Galerkin method was selected as the approach to implementation and was then developed into a computer interface for the model, a Discontinious Galerkin-based Wave Prediction Model (DG-WAVE). The study presented herein is motivated by a desire to investigate a fair comparison of the accuracy and computational efficiency of both SWAN and DG-WAVE. We run the models in hind cast for several months of 2011 and present results for significant wave height shown in time series comparison with recorded buoy data for each month as well as average computational time for each model. Our results show that DG-WAVE and SWAN perform well when validating significant wave height against actual buoy data. Computational time results show that DG-WAVE is highly efficient. The DG-WAVE model is thus shown to be a robust and accurate wave model with the potential to be developed into a high powered and efficient computational tool for coastal water studies. With the development of a comprehensive model, domains such as the Great Lakes could be more efficiently studied with great accuracy and largely decreased cost.
Description
Keywords
wave model, discontinuous-Galerkin method, computational model