This talk will present experiences and lessons learned during the development of a regional 3-dimensional ocean nowcast/forecast system. The model is based on the Regional Ocean Modeling System (ROMS) with multi-domain nested configurations. A multi-scale 3-dimensional variational (3DVAR) data assimilation scheme is used to assimilate both in situ (e.g., gliders) and remotely sensed data from both satellite and land-based platforms (e.g., high-frequency radars). For the ROMS assimilation system, the spatial resolution for the most inner domain is typically 1 km, comparable to the highest resolution satellite sea surface temperature and HF radar current observations. An ensemble modeling methodology has been tested to estimate the forecast uncertainties. Preliminary results coupling the regional circulation model with biogeochemistry/ ecosystem/fishery models will be presented. Several demonstration cases using ROMS forecasts in support of decision-making by the U.S. Coast Guard during Search and Rescue operations will be discussed.
