Particle Filtering and Gaussian Mixtures―On a Localized Mixture Coefficients Particle Filter (LMCPF) for Global NWP

In a global numerical weather prediction (NWP) modeling framework we study the implementation of Gaussian uncertainty individual particles into assimilation step localized adaptive particle filter (LAPF). We obtain local representation prior distribution as mixture basis functions. step, calculates weight coefficients and new locations. It can be viewed combination LAPF version filter, i.e., Localized Mixture Coefficients Particle Filter (LMCPF).