An examination of four-dimensional data-assimilation techniques for numerical weather prediction

Abstract

Four-dimensional data-assimilation methods, along with the most commonly used objective analysis and initialization techniques, are examined from a historical perspective. Operational techniques, including intermittent data assimilation and Newtonian nudging, and next-generation methods (Kalman-Bucy filtering and the adjoint method) are briefly described. Several methods are compared, with primary emphasis being placed on recent papers dealing with the operational assimilation techniques. Ongoing and future research is outlined, and some important implications of this research are discussed. -Authors