The Autonomous Ocean Sampling Network (AOSN) represents an integration of multiple observing platforms, data assimilation schemes and high-resolution numerical models of the coastal ocean. A triply-nested version of the Regional Ocean Modeling System (ROMS) is used to produce 3-day ensemble forecasts of the coastal ocean, for the AOSN-II observing period in the Monterey Bay area (August 2003). The image illustrates a coastal upwelling event during this period.

An ensemble of representations of surface wind stress (based on a single 3-km resolution COAMPS model forecast) is designed to force the ocean model ensemble, via spatial and temporal shifts of surface wind features. The sensitivity of the ocean model response to wind stress perturbations will be presented, as will the ability of the Ensemble Transform Kalman Filter adaptive observing strategy to predict the reduction in ROMS forecast error variance due to deployments of Autonomous Underwater Vehicles during AOSN-II.