Modified convection scheme in CAM to improve MJO predictability

We assess the predictability of the Madden-Julian Oscillation (MJO) event during the DYNAMO observational campaign period using the Community Atmosphere Model with a modified closure assumption in the Zhang-McFarlane convection scheme. This model has previously been shown to depict some major MJO features realistically. DYNAMO observations are used extensively to force, constrain and evaluate the model simulation of MJO initiation. ECMWF global reanalysis data through the period of DYNAMO is used for relaxation and sensitivity experiments to improve the convection scheme even further and the MJO concomitantly.

The MJO initiation phase, among all the MJO phases, has been previously shown to have the least predictability in all the current statistical and dynamical models present globally. Our improved model is used to understand the observed MJO initiation during the period and eventually improve the MJO simulation and predictability especially during the MJO initiation phase. Results from predictability experiments and factors that helped improve the predictability of the MJO will be presented.