Bayesian filtering for stochastic dynamical systems via Markov chain Monte Carlo

Stochastic dynamical systems have been increasingly used in natural sciences. Data assimilation, which can effectively combine observation data and theoretical models, improves the applicability of dynamical models. In this study, a statistical data assimilation method, Bayesian filtering, is presented. Its performance is examined with a dynamical model of aquatic ecosystem. It is found that the new method can give a satisfactory state estimate and be applied to general dynamical model in biological and environmental sciences.