An adjoint method for the assimilation of statistical characteristics into eddy-resolving ocean models

The work investigates perspectives of the parameter estimation problem with the adjoint method in eddy resolving models. Sensitivity to initial conditions resulting from the chaotic nature of this type of models limits the direct application of the adjoint method to the forecast range. Beyond this range an increasing number of secondary minima of the cost function emerges and prevents convergence of this method. In the framework of the Lorenz model it is shown that averaged quantities are suitable for describing invariant properties of the attractor and that secondary minima are for this type of data transformed into stochastic deviations. An adjoint method suitable for the assimilation of statistical characteristics of data and applicable on time scales longer than the forecast range is presented. The approach assumes a larger predictability for planetary scales which are here defined by spatial and temporal averaging. The adjoint to a prognostic model for statistical moments is invented for calculating cost function gradients that ignore the fine structure resulting from secondary minima. Coarse resolution versions of eddy resolving models are used for this purpose. Identical twin experiments were performed with a quasigeostrophic model to evaluate the performance and limitations of this approach for improving models by estimating parameters. The wind stress curl is estimated from simulated mean stream function and sea surface height variability. Even very simple closure schemes for second order moments are shown to give reasonable estimations of gradients that perform efficiently in minimizing cost functions. (orig.)Available from TIB Hannover: RN 3292(310) / FIZ - Fachinformationszzentrum Karlsruhe / TIB - Technische InformationsbibliothekSIGLEDEGerman