Assessing the Lagrangian Predictive Ability of Navy Ocean Models
DELAWARE UNIV NEWARK
Pagination or Media Count:
We have developed a variety of Lagrangian analysis tools with prior ONR support and we have studied the role of submesoscale and mesoscale dynamics in ocean transport by applying these tools to archived ocean model velocities. In many of the regions weve studied, Lagrangian analysis of model forecasts reveals complex, slowly evolving Lagrangian coherent structures LCS that define mixing boundaries in the flow. Increasing model spatial resolution allows more small-scale variability in these mixing boundary structures to emerge. Since LCS maps are constructed from tens of thousands of modeled trajectories, their usefulness depends entirely on the Lagrangian forecast skill of the underlying ocean models. Our objective is to assess the Lagrangian forecast skill of operational Navy ocean models in different geographic regions. Since it is extremely difficult to benchmark LCS maps with observations thousands of observed drifters would be needed, we focus on a more practical objective quantifying trajectory forecast skill over one forecast cycle typically 72 hours by comparing predicted and observed trajectories. We are also interested in exploring the range of Lagrangian forecast skill among all members of an ocean model ensemble, since this indicates the impact of model Eulerian uncertainties on the quality of trajectory forecasts.
- Physical and Dynamic Oceanography