In-situ and satellite observations are used to improve the performance of a nowcast/forecast coastal ocean model. A simple and robust data-assimilation scheme (OI) is used to generate high-resolution cloud-free Sea Surface Temperature fields (through blending AVHRR and TMI SST) and improved coastal wind fields (through blending atmospheric model analysis with in-situ wind measurements). This procedure improves the coastal SST and wind field, which in turn improves the fidelity between the coastal ocean circulation model and in-situ temperature and currents. Comparisons between momentum analyses performed independently from the model and the data demonstrate the fidelity to be of a correct dynamical basis. Further improvements are achieved in experiments using in-situ air-sea flux measurements to drive the coastal ocean circulation model. We conclude that (1) the present coastal ocean model computation and parameterization schemes are reasonably good and that the primary limitation to ocean model performance lies with the boundary conditions; (2) the importance of emergent coastal ocean observing systems cannot be overemphasized. Along with their oceanographic variables these observing systems must have coverage sufficient to promote improvements in the coastal ocean surface flux fields.