The limit of predictability in the stratosphere versus troposphere of a coupled GCM

The limit of predictability in the stratosphere versus troposphere in a coupled GCM is investigated using the National Center for Environmental Prediction (NCEP) Climate Forecast System (CFS). A set of identical twin experiments is obtained by using the interactive ensemble coupling strategy, where a single oceanic general circulation model (MOM3 of GFDL) is coupled to the ensemble average of multiple realizations (in this case 6) of an atmospheric GCM (NCEP Global Forecast System, GFS). The atmospheric GCMs are initialized from slightly different initial conditions, but the ocean state is the same. In this strategy, the atmospheric realizations represent different possible atmospheric states that have equally chances of occurrence. We compare the normalized errors of zonal wind and temperature fields in the troposphere and stratosphere for various wave groups. The results show a large error growth in the stratosphere compared to troposphere. We also investigate the predictability of sudden stratospheric warming events. Preliminary results suggest that the predictability of sudden stratospheric events is low because small errors in individual fields lead to large errors in the EP flux.