Stochastic parametrization: An alternative to inflation in ensemble Kalman filters

We investigate the application of a stochastic dynamical model in ensemble Kalman filter methods. Ensemble Kalman filters are very popular in data assimilation because of their ability to handle the filtering of high-dimensional systems with reasonably small ensembles (especially when they are accompanied with so-called localization techniques). The stochastic framework presented here relies on location uncertainty principles that model the effects of the model errors on the large-scale flow components. The experiments carried out on the surface quasi-geostrophic model with the localized square-root filter demonstrate two significant improvements compared with the deterministic framework. First, as the uncertainty is a priori built into the model through the stochastic parametrization, there is no need for ad hoc variance inflation or perturbation of the initial condition. Second, it yields better mean-square-error results than the deterministic ones.