Evaluations of Global Wave Prediction at the Fleet Numerical Meteorology and Oceanography Center

Abstract

It is a major challenge to determine whether bias in operational global wave predictions is predominately due to the wave model itself (internal error) or due to errors in wind forcing (an external error). Another challenge is to characterize bias attributable to errors in wave model physics (e.g., input, dissipation, and nonlinear transfer). In this study, hindcasts and an evaluation methodology are constructed to address these challenges. The bias of the wave predictions is evaluated with consideration of the bias of four different wind forcing fields [two of which are supplemented with the NASA Quick Scatterometer (QuikSCAT) measurements]. It is found that the accuracy of the Fleet Numerical Meteorology and Oceanography Center?s operational global wind forcing has improved to the point where it is unlikely to be the primary source of error in the center?s global wave model (WAVEWATCH-III). The hindcast comparisons are specifically designed to minimize systematic errors from numerics and resolution. From these hindcasts, insight into the physics-related bias in the global wave model is possible: comparison to in situ wave data suggests an overall positive bias at northeast Pacific locations and an overall negative bias at northwest Atlantic locations. Comparison of frequency bands indicates a tendency by the model physics to overpredict energy at higher frequencies and underpredict energy at lower frequencies.