Multi-scale Dynamics of MJO Convective Onset during DYNAMO

Radar and rawinsonde observations collected during DYNAMO, supplemented with space-borne precipitation radar from the now-defunct TRMM platform, indicate that large-scale convective onset of the MJO events observed during DYNAMO, marked by the fairly sudden development of widespread deep convection and mesoscale convective systems over the Indian Ocean, occurred in a two-step process. Moistening of the troposphere by moderately deep cumulonimbus clouds before MJO onset, which has long been suspected to play a role in MJO onset, is the second step of that process. TRMM observations collected over the central equatorial Indian Ocean during DYNAMO depict a rapid increase in the areal coverage of moderately deep cumulonimbi about 7 days prior to MJO onset. Moreover, sounding budget analysis on 1º grid spacing derived from the rawinsonde network in place during DYNAMO depicts a negative moisture sink (Yanai’s Q2 term, which represents drying by sub-grid scale processes) between about 600–800 hPa during the week-long “transition” period when moderately deep clouds are prevalent. Both results confirm that clouds directly moisten the troposphere, supporting the later development of mesoscale systems that are present during an active MJO.

The first and more often overlooked step concerns how the moderately deep convection became prevalent in the first place. As a circumnavigating equatorial Kelvin wave enters the Indian Ocean, it supports a reduction in tropospheric subsidence of about 0.002 m s^-1. While very small, if the anomalous reduction of subsidence is integrated over about a week, it supports cooling below 500 hPa, thus reducing convective inhibition above the boundary layer and allowing more boundary layer cumuli to develop into moderately deep elements.

WRF model simulations of two DYNAMO MJO cases replicate moistening by moderately deep clouds during a transition period. In both the model and observations, the transition period is followed by MJO onset and preceded by a convective regime dominated by boundary layer cumuli. WRF model simulations can be employed to investigate important cloud moistening processes and to quantify changes in buoyancy experienced by cloud updrafts during suppressed, transition, or deep convective periods. The WRF model simulations support several conclusions or hypotheses: 1) Upscale growth of convective elements into mesoscale systems occurs after relative humidity in the lower half of the troposphere reaches a critical threshold of roughly 60–70%. 2) Simulated isolated cumulonimbi moisten their local environment by evaporation on the edges of clouds and horizontal advection of that moisture to within about 5 km of the cloud. The aggregate moistening of many cloud elements over a large region is the primary contributor to large-scale moistening during transition periods. 3) Reduction of large-scale subsidence in the model prior to the start of transition periods cools the 850–700 hPa layer by about 0.2–0.4K. The seemingly small change in temperature has a critical impact on the mean buoyancy of updrafts in this climatologically neutral to stable layer and allows transition periods, and thus the MJO onset process, to begin.