Operational ocean prediction (Invited Presentation)

The monitoring and forecasting of the global ocean environment requires proper representation of the physical processes in numeric models and the assimilation of observations to maintain the non-deterministic ocean field. Oceanographic monitoring and prediction in the deep ocean is very similar to the atmospheric problem. However, ocean features are much smaller in size and persist much longer. Thus, numerical models are more computationally intensive and can provide longer time forecasts in the deep ocean. Significant forecast skill has been demonstrated in the deep ocean out to 30 days. Satellite observations critical for assimilation into ocean models include measurements of sea surface height and sea surface temperature. In addition to the surface observations, correlation to subsurface structure is a vital component of the system. Present systems rely on continuous input from satellite altimeter and temperature sensors to maintain accurate estimates of the nondeterministic mesoscale field.

Nearer to the coast, tides, wind-driven effects, and surface waves dominate physical variations. To resolve the processes in numerical models, it is necessary to nest higher resolution systems into global models. Observations within these areas present challenges to resolve the space and time scales. The solution is expected to require a combination of in situ observations such as the Integrated Ocean Observing System (IOOS) and remote sensing. Several future satellite missions provide possibilities to begin to address these problems. In particular, the Wide Swath Ocean Altimeter can provide high resolution tide information. The Aquarius and Soil Moisture and Ocean Salinity SMOS) sensors could provide information on fresh water sources. Possibly the largest untapped contribution resides in use of hyperspectral observations of water properties that are linked to the ocean dynamical state. Presently the SeaWiFS, MODIS Aqua and MODIS Terra satellites provide optical property information that is very important to many operational applications. However, a predictive capability for these properties does not yet exist operationally. Combining this information into assimilation systems requires construction of models that represent not only the physics but also biogeochemical processes in the ocean.