Near-surface level water vapor leads precipitation by one pentad over the Indian Ocean and Western Pacific while upper level water vapor lags the peak in precipitation. The composite evolution of the MJO as tracked by specific humidity shows markedly different vertical structures as a function of longitude. There is a clear westward tilt with height of the moisture maximum associated with eastward propagating MJO disturbances, which disintegrates into a nearly vertical positive anomaly as it reaches the Western Pacific, and then becomes trapped to the lowest level in the Eastern Pacific. ISCCP cloud fraction is highly correlated with humidity, and also leads observed precipitation. However, the maxima in rainfall do not occur at the same longitude as the maxima in cloud amount, and there appears to be a shift in the dominant cloud type as the MJO systems propagate eastward.
Another important feature is a rectification of the MJO cycle onto the mean state of the atmosphere, with coherent non-zero average rainfall and moisture patterns occurring over the life cycle of strong events. Over the course of a whole winter season that contains above average MJO activity, there are also significant moistening and drying trends. While this latter effect has been checked for possible contamination from a few large ENSO events, such an artifact still needs to be more systematically ruled out.
We also examine the surface heat flux (from Jones et al. 1999) and moisture transport (calculated using TOVS and ECMWF winds) along with the above rainfall and three-dimensional moisture evolution in order to get an integrated view of the MJO's hydrological cycle. While there is a high correlation of these fluxes with the rainfall and water vapor, the interrelationship of all the hydrological quantities shows considerable spatial variability.
Given that even models with significant intraseasonal variability often fail to display many features listed above, this assessment will serve as a powerful yardstick for evaluating model performance with respect to the MJO. This work should provide some guidance for model development of intraseasonal simulation of the tropics.
Supplementary URL: http://terra.msrc.sunysb.edu/~dmyers/Revised.pdf