EXAMINING THE EFFECTS OF DIFFERING TIME SCALES IN COUPLED DATA ASSIMILATION USING THE LORENZ 96 DOUBLE-RING MODEL

The Lorenz (1996; hereafter L96) double-ring model is a toy dynamical system with coupled equations between a set of state components with fast-timescale variations and another with slower-timescale variations.  Given the challenge in ocean-atmosphere strongly coupled data assimilation of the differences in time scales, the L96 model can provide a useful testbed for examining issues related to these time scale differences.  Using a coupled state ensemble Kalman filter (EnKF) and synthetic observations related to both the slow and fast components, we will explore and discuss how EnKF filtering strategies may be adapted to provide reduced error coupled state assimilations.