Convective interactions with the large-scale environment: Differences and similarities between the MJO and the Kelvin wave

The cloud population transition associated with the Madden-Julian Oscillation (MJO) life cycle transitions from predominantly shallow convective clouds in the onset stage to deep convective and then stratiform clouds in the mature stage. This progression is analogous to the lifecycle of a mesoscale convective system (MCS), also shown by other modes of tropical intraseasonal variability—the Kelvin wave and the inertio-gravity wave. There is, however, a growing view that the environmental moisture is intimately coupled to convection during the MJO, an aspect in which it differs from the Kelvin wave. We use observations from the DYNAMO/CINDY campaign to study how convection and its surrounding environment vary during the evolution of the MJO and the Kelvin wave, to compare and contrast the two modes of variability. A distinct difference, from a Eulerian point-of-view is that the moisture anomalies during the MJO are larger (about 3-4 times) than that of the Kelvin wave; this translates to a larger areal extent of stratiform convection during the MJO. We examine if the cloud-radiative feedbacks arising from the large area of cloud-cover could potentially aid in further moistening in the environment. One other difference is that the moisture variation during the MJO is asymmetric about peak rainfall—it has a gradual buildup and a rapid demise, in contrast to the Kelvin wave. This is because of strong vertical and horizontal advective drying that immediately follows peak MJO rainfall. We examine how the different terms in the moisture budget evolve during the MJO and the Kelvin wave, to look for signs of strong moisture-convection coupling in the MJO, and lack thereof in the Kelvin wave. For our analysis, we obtain latent heating profiles from Shared Mobile Atmospheric Research and Teaching Radar (SMART-R) and cloud radiative heating profiles from the Ka-ARM Zenith Radar (KAZR). Thermodynamical variables were provided by the quality-controlled sounding data from Gan Island; dynamical variables were provided the objective variational analysis of the large-scale forcing data, centered on Gan Island.