The Impact of Ocean Model Limitations on Observing System Simulation Experiments

An ocean OSSE system requires models that accurately reproduce oceanic phenomena of interest so that viable nature runs can be generated. Ocean OSSEs are in their infancy compared to the atmosphere, with ocean models tending to reproduce many oceanic phenomena with significant errors and biases that will limit the usefulness of nature runs. Our OSSE system development efforts are focusing on different oceanic phenomena that highlight different sets of strengths and weaknesses in present-day ocean models. One effort focuses on observing system design over the Gulf of Mexico and the adjacent coastal region to the north, emphasizing ocean “weather” systems such as cyclones, anticyclones, boundary currents, fronts, and coastal jets. This effort is akin to atmospheric OSSEs that emphasize enhanced observation of synoptic weather systems with the goal of improving weather forecasts. In this case, nature run generation will benefit from the fact that ocean models can now reproduce ocean weather systems with statistically realistic structure and variability. The other two projects are more challenging. One emphasizes ocean climate processes, specifically the design of an observing system to monitor the Atlantic Meridional Overturning Circulation. Different models display large differences in their representation of the strength of the AMOC and its associated heat flux and fluid pathways. The other project involves observing system design with the goal of improving ocean model initialization in coupled hurricane forecast models. In this case, a key ocean model limitation is the vertical mixing parameterization. Vertical profiles of temperature, salinity, and density, along with vertically-integrated ocean heat content, are sensitive to this parameterization and differ substantially among ocean models. Strategies to ameliorate these challenges will be presented.