Morphing ensemble Kalman filters

Abstract

A new type of ensemble filter is proposed, which combines an ensemble Kalman\nfilter (EnKF) with the ideas of morphing and registration from image\nprocessing. This results in filters suitable for nonlinear problems whose\nsolutions exhibit moving coherent features, such as thin interfaces in wildfire\nmodeling. The ensemble members are represented as the composition of one common\nstate with a spatial transformation, called registration mapping, plus a\nresidual. A fully automatic registration method is used that requires only\ngridded data, so the features in the model state do not need to be identified\nby the user. The morphing EnKF operates on a transformed state consisting of\nthe registration mapping and the residual. Essentially, the morphing EnKF uses\nintermediate states obtained by morphing instead of linear combinations of the\nstates.\n