Eastward Propagating Upper-Tropospheric Divergence and Temperature Anomalies and the Timing of Large-Scale Organized Convective Development during AMIE

Description

The AMF2 at Gan Island was located within a large region of the central Indian Ocean where large-scale organized convection associated with a Madden-Julian Oscillation (MJO) event often first develops before propagating to the east. Sounding profiles of zonal wind (u), temperature (T), and humidity (q) captured the development and passage of three such events during the AMIE campaign. Satellite-based views of humidity and precipitation show that the evolution of convection and humidity observed via rawinsonde is representative of the evolution of convection and humidity throughout the central equatorial Indian Ocean. The changes observed in each are also depicted in ERA reanalysis, and this allows for its use to characterize existence of signals in u, T, q, and relative humidity prior to large-scale convective onset. Clear, eastward propagting signals in u at 150 hPa correspond with velocity potential anomalies that exist prior to large-scale convective development, which occurs when a negative velocity potential anomaly (divergent anomaly) approaches from the west and passes over the central Indian Ocean. Positive T and q anomalies are observed between 500 and 200 hPa where convection occurs, and they follow convection as it propagates eastward. Positive T anomalies at the same levels appear to propagate into the region from the west prior to convective onset, but q anomalies do not predate convection. This lends support to Powell and Houze (2013), who conclude that upper-tropospheric humidity anomalies associated with convective MJO events are likely caused by some combination of cloud moistening and moisture advection caused by the anomalous circulation associated with the MJO. Early regional modeling results using the Weather Research and Forecasting model are shown. Using a spatial resolution of 8 km, the sensitivity of convection to ~30 day variability in T, u, q, and sea surface temperature (SST) is tested. The resolution is not enough to explicitly resolve convection, but it can effectively model the presence of large-scale organized convection. When using ERA reanalysis as forcing and updating SSTs, WRF can effectively simulate the presence of widespread organized convection for the October and November events observed at Gan. Modeling experiments in which the 20-40 day variability in the above variables is removed from the boundary conditions show the sensitivity in the model of convection to the intraseasonal variability of each parameter.