The coastal forecast systems sought by Navy planners are ones that retain their robustness regardless of the region of application (Blain and Preller, 2007), are targeted to local areas of interest, and whose products are able to address the full suite of Navy parameters that impact coastal operations including acoustic communication, mine hunting, and underwater vehicle operations, to name a few. For this type of forecast system, a better understanding of systematic errors is sought in order to optimize construction of the coastal forecast system to minimize error and maximize robustness.

Assessments of the fidelity of a coastal forecast system typically involve comparisons between forecast products and concurrent in-situ observations. A bulk error measure for each physical parameter such as water level, temperature or currents is computed and reported. This error strictly indicates the deviation of a forecast parameter from the observed value at a specific location and window of time, e.g., Ko et al. 2008, Allard et al. 2008. What it does not indicate are the sources of error that lead to the mismatch between modeled and observed parameters. As coastal forecast systems advance to consider the implementation of sophisticated forms of data assimilation, e.g. Kurapov et al. 2002, Li et al. 2008, quantification of the error sources is increasingly important. An understanding of the error source can not only lead to more effective data assimilation strategies but also better configuration of the coastal forecast system itself, which will ultimately result in improved forecast products.

Two-day forecasts of three-dimensional baroclinic currents produced over a 10-day period from 04-14 June 2010 at the mouth of Chesapeake Bay are systematically examined via comparisons to real-time observations. These currents are computed by a limited domain coastal forecast system that utilizes unstructured grids within its hydrodynamic core. Comparisons of the predicted currents to available observations do not merely quantify the data fit at a particular point in space and time but aim to determine the source of any forecast error. In this context, both spatial and temporal resolution of the forcing is examined for its impact on the forecast products. Forms of the forcing include density initialization and boundary specification derived from a regional model, the prescription of tides at offshore boundaries, and applied surface winds.

Interplay between the temporal frequency of the regional model boundary forcing and the application of external tides to the coastal model impacts the tidal characteristics of the coastal current, even contributing a small phase error. Frequencies of at least 3-hr are needed to resolve the tidal signal within the regional model, otherwise, externally applied tides from a database are needed to capture the tidal variability. Spatial resolution of the regional model (3-km vs 1-km) does not impact skill of the current prediction. Tidal response of the system indicates excellent representation of the dominant semi-diurnal tide for water level and currents. However, the primary diurnal tide is amplified unrealistically with the application of coarse 27-km winds. Higher resolution winds reduce current forecast error with the exception of wind originating from the SSW, SSE, and E. These winds run shore parallel and are subject to strong interaction with the shoreline that is poorly represented even by the 3-km wind fields. The vertical distribution of currents are also well-predicted by the coastal model. The analyses conclude that spatial and temporal resolution of the wind forcing including areas close to the shoreline are the most critical component for accurate current forecasts. Additionally it is demonstrated that wind resolution plays a large role in establishing realistic thermal and density structures in upwelling prone regions.