Data assimilation in ocean models

Publisher Website
Google Scholar
Abstract
This review covers recent advances in applying data assimilation techniques to problems in physical oceanography. The introduction and appendices provide the non-specialist reader with background in ocean circulation and observing methods. The 4D variational assimilation approach is covered in depth showing how model - data misfits can be minimized using Lagrange multipliers in an unconstrained variational problem. Applications to modelling tropical Pacific temperatures, sea surface heights and circulation in the North Atlantic, and tomographic (sound travel time) data are all presented. The use of variational methods for deriving average climatological circulation patterns is also described as well as applications to error growth during numerical forecasting. The paper then focuses on three important topics in physical oceanography, the evolution of ocean mesoscale eddies in middle latitudes, the development of El Nino events in the tropical Pacific, and the evolution of ocean surface waves. Recent improvements in data acquisition and modelling in these areas mean that data assimilation is practical and is providing new insights and forecasting capabilities to varying degrees. Results using a variety of assimilation techniques are presented, concluding with a forward look to a time when routine forecasting of ocean developments will be possible with important implications ranging from understanding climate change to fishing and pollution control.

This publication has 84 references indexed in Scilit: