Particle filter is a data assimilation technique, which is based on Monte Carlo method. In this sense it is similar to EnKF or EnSR. The difference is that particle filter does not assume Gaussianity in the pdf of the model state. The distribution function of the model state is determined empirically from the particles (ensemble). At the analysis step, it uses the observation to assign different weight to each particle to determine a new density function: the closer a model output to the observation, the larger the weight. A new set of particles is the resampled from this density function. Different sampling strategies are proposed in the literature. Until now, only particle filter based on residual resampling is available in OpenDA.