Moist Thermodynamics of the Madden-Julian Oscillation in a Cloud-Resolving Simulation

The moist thermodynamic processes that determine the time scale and energy of the Madden-Julian oscillation (MJO) are investigated using moisture and eddy available potential energy budget analyses on a cloud-resolving simulation. Two MJO episodes observed during the winter of 2007/08 are realistically simulated. During the inactive phase, moisture supplied by meridional moisture convergence and boundary layer diffusion generates shallow and congestus clouds that moisten the lower troposphere while horizontal mixing tends to dry it. As the lower troposphere is moistened, it becomes a source of moisture for the subsequent deep convection during the MJO active phase. As the active phase ends, the lower troposphere dries out primarily by condensation and horizontal divergence that dominates over the moisture supply by vertical transport. In the simulation, the characteristic time scales of convective vertical transport, mixing, and condensation of moisture in the midtroposphere are estimated to be about 2 days, 4 days, and 20 h respectively. The small differences among these time scales result in an effective time scale of MJO moistening of about 25 days, half the period of the simulated MJO. Furthermore, various cloud types have a destabilizing or damping effect on the amplitude of MJO temperature signals, depending on their characteristic latent heating profile and its temporal covariance with the temperature. The results are used to identify possible sources of the difficulties in simulating MJO in low-resolution models that rely on cumulus parameterizations.