Numerical Models, Observing Systems, and Data Assimilation for Prediction of Ocean Mesoscale Eddies

Abstract

Mesoscale eddies pervade the global ocean, characterized by a typical horizontal scale of approximately 100 km and a timescale on the order of a month. Forecasting these eddies is the primary objective in predicting the ocean’s “weather” over periods ranging from days to a month. This study provides a historical perspective on milestones in the evolution of successful mesoscale eddy-forecasting systems over the past half-century. Advances in eddy-resolving numerical models, observing systems, and, in particular, data assimilation (DA) algorithms have led to success in forecasting mesoscale eddies. Mesoscale eddies arise from baroclinic flow instabilities, making their forecast highly sensitive to initial conditions. A forecasting model must be appropriately initialized to generate subsequent forecasts successfully. DA integrates various observations into the model forecast, producing optimal estimates of the ocean state to initialize numerical models. An effective combination of observations from the Argo float observing network and a constellation of altimetry satellites is crucial for accurate estimates of the ocean state at the mesoscale through DA, ensuring success in forecasting mesoscale eddies. The temporal and spatial scales of the ocean state at different depths are greatly different. These ocean state characteristics pose multiple challenges in altimetry DA. To tackle these challenges, multiscale DA (MSDA) algorithms have been suggested, formulated, and implemented. Different strategies for the implementation of MSDA are discussed. The most pressing needs for further model development and enhancing mesoscale DA are outlined.