Comparison of a mid-shelf wave hindcast to ADCP-measured directional spectra and their transformation to shallow water

In conducting a cross-shelf wave transformation experiment off the Atlantic coast of north Florida, a unique opportunity was exploited in which an Acoustic Doppler Current Profiler (ADCP) instrument was installed 30 km offshore at the exact location of one of the archive-nodes of a WAM-like wave hindcast model (OWI3G). A second ADCP was installed 550 m from shore. Approximately 53 days of directional wave spectra collected with the two ADCPs are used to (a) locally test the reliability of a subsequent update of the hindcast, (b) document the loss in energy as the waves crossed the broad, relatively shallow continental shelf between the two instruments, (c) test the ability of the SWAN (Gen2) nearshore wave transformation model to replicate the measurements taken in shallow water when driven by the offshore ADCP spectra, and (d) reassess the spectral transformation results when the offshore hindcast is used as input. In addition to direct comparison of the time series of frequency spectra and the directional distribution of energy, typical spectral parameters are each subjected to standard error tests. Results indicate that the offshore hindcast performs well in replicating significant wave height, fairly well for mean period, but not as reliably for peak period. Directional spreading in deeper water is generally well-represented, although vector mean direction is not, and is believed due to the proximity of the coast to the hindcast node. The nearshore model requires an order-of-magnitude reduction in bed roughness from its default value before agreement in wave energy at the nearshore ADCP can be achieved. Outcomes of the error tests for the hindcast-driven versus the ADCP-driven nearshore results (after roughness calibration) are quite similar, but nevertheless indicate that transformed wave period, wave direction, and directional spreading require improvement.