Evolution of Humidity and Convection Prior to MJO Onset and Their Sensitivities to Upper-Tropospheric Equatorial Wave Dynamics

No proven dynamical explanation exists for how convection related to an instance of the Madden-Julian Oscillation (MJO) organizes over the Indian Ocean when no prior widespread convection was present. Among proposed hypotheses is the “discharge-recharge” hypothesis, which suggests that convective clouds and humidity feed back onto each other over a large area until the large-scale environment is conducive to widespread mesoscale convection. Another hypothesis involves circumnavigating anomalies of velocity potential (i.e. divergence) in the upper-troposphere, with the divergent anomaly aloft propagating over a region favoring development of organized convection. However, the mechanism(s) through which upper-tropospheric dynamics impact(s) convection in the lower-troposphere is unclear.

Results from the Dynamics of the Madden-Julian Oscillation (DYNAMO) and ARM Madden-Julian Oscillation Experiment (AMIE) show that during three cases of MJO onset over the central Indian Ocean (near Addu City, Maldives), the vertical build-up of tropospheric humidity and the increase in convective radar echo depth occur over just 3 to 7 days. This timescale is much shorter than that proposed by the “discharge-recharge” hypothesis. We show that compositing radar data from the three DYNAMO/AMIE events together yields a result that suggests that the buildup of humidity and convection occurs over a two-week long period, which would be more in line with the timescale proposed by “discharge-recharge”. However, analysis of any individual event does not support “discharge-recharge.” Because our result highlights the importance of examining the evolution of each unique MJO event separately, we explore the evolution of humidity and precipitating cloud echo prior to MJO events since 1999 using data from the Tropical Rainfall Measuring Mission (TRMM) precipitation radar and output from the ERA-Interim reanalysis.

Comparison of ERA-I and TRMM data with field observations shows that the evolution of convective echoes and humidity in a relatively small volume near the field site closely matches their evolution across the wider central equatorial Indian Ocean during DYNAMO/AMIE. We therefore use ERA-I data to expand the spatially limited rawinsonde dataset from the field program. When we examine ERA-I output globally, we see that 25 to 30 day variability observed in rawinsonde data in temperature above 500 hPa and zonal wind near 150 hPa is linked to eastward propagating upper-tropospheric anomalies of zonal wind and temperature, which itself is probably linked to equatorial wave activity. Convection becomes widespread and organized over the Indian Ocean only when the anomalies reach the area of MJO convective onset. In addition, results from high-resolution regional modeling over the Indian Ocean basin using the Weather Research and Forecasting (WRF) confirm and add further insight into the sensitivity of MJO convective development in model to upper-tropospheric wind, temperature, and humidity anomalies.