Role of water vapor and convection-circulation decoupling in MJO simulations by a tropical channel model

A tropical channel model failed to simulate an observed MJO event. As a means to find reasons for the failure, model errors in water vapor (q(v)) and other fields were corrected to different extents by nudging towards a global reanalysis product. An MJO metric was developed to quantify the realism of a simulated MJO event in the amplitude and zonal propagation speed of its precipitation and zonal wind at 850 hPa (U850). Results show that correcting errors in q(v) alone is necessary and sufficient for a realistic simulation of the MJO event. Correcting errors in other fields is insufficient by itself and unnecessary with water vapor correction. For a realistic MJO simulation, error corrections are needed for q(v) perturbations of the planetary scales (zonal wave number 1-3) together with its zonal mean in both the lower and middle troposphere. Correcting errors in q(v) at upper levels, planetary-scale perturbations of any single zonal wave number, synoptic-scale perturbations, or temporal/zonal means alone is not sufficient for a realistic MJO simulation. Amplitudes of MJO precipitation and U850 are significantly correlated, but their propagation speeds are not. MJO signals in U850 are easier than those in precipitation to improve by correcting errors in q(v). In the absence of MJO intrinsic dynamics, inserted MJO signals in q(v) in a simulation may create a phantom MJO event in precipitation but not in the circulation. This study confirms the existing knowledge on the importance of qv to MJO simulations, sheds new lights on the role of the q(v) spatial structure, and presents evidence for possible precipitation-circulation decoupling in MJO simulations.