Reducing systematic errors by empirically correcting model errors

Abstract

A methodology for the correction of systematic errors in a simplified atmospheric generalcirculationmodel is proposed. First, a method for estimating initial tendency model errors isdeveloped, based on a 4-dimensional variational assimilation of a long-analysed dataset ofobservations in a simple quasi-geostrophic baroclinic model. Then, a time variable potentialvorticity source term is added as a forcing to the same model, in order to parameterize subgridscaleprocesses and unrepresented physical phenomena. This forcing term consists in a (largescale) flow dependent parametrization of the initial tendency model error computed by thevariational assimilation. The flow dependency is given by an analogues technique which relieson the analysis dataset. Such empirical driving causes a substantial improvement of the modelclimatology, reducing its systematic error and improving its high frequency variability. Lowfrequencyvariability is also more realistic and the model shows a better reproduction of Euro-Atlantic weather regimes. A link between the large-scale flow and the model error is found onlyin the Euro-Atlantic sector, other mechanisms being probably the origin of model error in otherareas of the globe.