Coupled Model System for Southeast Florida: Wave Model Validation Using Radar and In Situ Observations.

摘要

This document describes the wave model validation efforts for a Florida Straits test case for the fully coupled Coastal Ocean/Atmosphere Mesoscale Prediction System (COAMPS) model during spring of 2005. The main focus of this work is validation of the waves portion of the COAMPS model; however, brief point analysis of winds and tides are also included. The 'ground truth' for waves validation includes in situ data (ADCP and buoy) and high frequency Wellen Radar (WERA HF) data collected by the University of Miami. The wave height comparison to in situ data is generally good, though some wave events are overpredicted by the model. It is found that the accuracy of predictions of swell events is highly sensitive to the swell direction, and further that the range of directions from which swells are able to pass through islands and shoals to arrive at the in situ data region is highly sensitive to whether currents are included in the forcing. In the comparison to radar, the time series comparisons are mostly good, and the wave model with surface current forcing included shows a modestly improved accuracy compared to a model without currents. With regard to the ability of the models to capture the spatial variability evident in the radar, this depends on whether it is variability in the azimuthal or range position relative to the radar transmitter. In the context of the azimuthal position, it is suspected that the error statistics are unfortunately more reflective of the relative quality of the radar than the model skill. In the context of the range position, the situation is more favorable: this direction approximates the cross-shore direction, and variation of model and radar are in qualitative agreement in this case. The wave model with surface current forcing has higher correlation with the radar than does the model without currents.