Role of Model Configurations in MJO Simulations

Most coupled models still have issues to reproduce the MJO signal realistically, especially in terms of its amplitude and eastward propagation. The subtle differences in cumulus convection schemes, planetary boundary layer (PBL) processes, and lateral boundary conditions may all play a role in the simulation of MJO. Previous studies have suggested the importance of humidity effects to MJO simulations, which include the mean state, sensitivity of precipitation to humidity, vertical structures of diabatic heating, and zonal asymmetry in tropospheric humidity. Using a channel atmosphere-ocean coupled model with a resolution of 30km over tropical region, we demonstrated that MJO event can be better simulated only by improving parameterization schemes of different processes all together as a package. The combination of errors from cumulus convection scheme, PBL, and microphysics can be identified to be the main sources of the dry biases in precipitation and column-integrated water vapor over Indian Ocean and Maritime Continent (MC), which makes the simulated MJO unrealistic. The influences of PBL convergence can penetrate into the upper troposphere and those by cumulus convection into the boundary layer, which change diabatic process and zonal-vertical distribution of moisture associated with MJO event. This conjecture is supported by a series of sensitivity tests. These sensitivity tests also demonstrate that the simulated MJO initiation and propagation depend critically on the performance of simulated precipitation and column-integrated water vapor over MC.