Ensemble Kalman filter

Abstract

Meteorological models are used to predict the weather, study atmospheric processes, and provide input to decision makers on the consequences of increased greenhouse gas emissions to the Earth's climate. Predictions of future atmospheric states are accomplished as an initial value or marching problem, where the initial atmospheric state is specified and the variables are advanced in time using numerical techniques. The challenge of data assimilation in meteorology is to estimate, based on a set of limited observations of varying types, the complete three-dimensional atmospheric state at a given time to provide an initial value for a meteorological model.