TY - JOUR AB - This paper presents a novel approach for inverting a complex ocean model via a proper orthogonal decomposition. The inversion is achieved through the construction of an adjoint model and used to assimilate data in a similar manner to that used in weather forecasting. This is an incredibly important capability for an ocean model, however it is both complex to construct and also can be computationally expensive. The approach proposed here addresses both of these important issues by constructing an efficient and easy to compute adjoint directly from the reduced order model. The approach is demonstrated by inverting for initial conditions in an ocean gyre simulation. The methodology proposed here led directly to the award of a £1M EPSRC grant (EP/I00405X) to develop reduced order and adjoint models for coastal oceanography. Cited 11 times. AU - Fang,F AU - Pain,CC AU - Navon,IM AU - Piggott,MD AU - Gorman,GJ AU - Farrell,PE AU - Allison,PA AU - Goddard,AJH DO - 10.1002/d.1911 EP - 732 PY - 2009/// SP - 709 TI - A POD reduced-order 4D-Var adaptive mesh ocean modelling approach T2 - Int. J. Numer. Meth. Fluids UR - http://www3.interscience.wiley.com/journal/2861/home VL - 60 ER -