Forecasts of MJO during DYNAMO in a coupled tropical channel model, Part I: Impact of parameterization schemes

A coupled tropical channel model is developed to forecast the Madden-Julian Oscillation (MJO) event during the Dynamics of the MJO (DYNAMO) campaign period. Eight sets of 30-day hindcast experiments are performed using different combinations of cumulus parameterization and planetary boundary layer (PBL) schemes. The results indicate that a well-simulated mean state is a prerequisite for MJO forecast, and the model's ability to reproduce the mean state is sensitive to the parameterization schemes. The experiment with Tiedtke (TDK) cumulus parameterization and University of Washington (UW) PBL schemes can successfully predict the MJO event. Over the equatorial Indian Ocean, this experiment has a good representation of zonal sea surface temperature (SST) gradient and low-level westerly wind, favouring a reasonable atmospheric circulation. As a result of boundary-layer convergence in the eastern Indian Ocean, more water vapour is gathered in the low-middle troposphere by updrafts, contributing to the PBL moistening ahead of the MJO convection and the eastward-propagation of MJO precipitation. Further analysis reveals that the high cloud is accompanied by strong convection. A better simulation in tropical high cloud fraction helps trap the cloud radiation heating to warm middle-upper troposphere, which promotes the MJO propagation. Our results suggest that the selection of proper parameterization schemes with a better-simulated mean state is crucial to predicting the MJO event in the tropical channel model.

Automated summary

A coupled tropical channel model is developed to forecast the Madden-Julian Oscillation (MJO) event. The results indicate that a well-simulated mean state and proper parameterization schemes are crucial to predicting the MJO event.