Variability of Oceanic Mesoscale Convective System Vertical Structures Observed by CloudSat in Indo-Pacific Regions Associated with the Madden-Julian Oscillation

Vertical structures of mesoscale convective systems (MCSs) during the Madden-Julian-Oscillation (MJO) are investigated using 2006-2011 CloudSat radar measurements for Indo-Pacific oceanic areas. In active phases of the MJO, MCSs with larger sizes occur relatively more frequently. MCSs in connected from also occur more frequently compared to MCSs in separated form. The frequency of occurrence of connected MCSs peaks in the onset phase, a phase earlier than separated MCSs.

Compared to separated MCSs, in all MJO phases connected MCSs in all sizes have weaker reflectivity above 8 km in their deep precipitating portions as well as thick anvil clouds closely linked to their parenting rain areas, suggesting more “stratiform” physics associated with them. Coincident data from TRMM Precipitation Radar measurements consistently show that over tropical oceans the “stratiform” rain fractions in connected MCSs is generally higher than separated MCSs. All MCSs show stronger/weaker reflectivity above 8 km in their deep precipitating portions before/after the onset phase of the MJO compared to the onset phase, indicating an overall shift towards more “stratiform” rain physics after the onset phase of the MJO.

Connected MCSs in all phases have systematically lower anvil productivity (amounts of anvil clouds relative to areas of their rain cores) compared to separated MCSs. In active phases of the MJO, while deep convective systems occur more frequently, MCSs with larger sizes tend to have lower anvil productivity compared to suppressed phases. The reduced anvil productivity associated with connected MCSs as well as large MCSs in active phases (ie. more chances to connect with other deep convective systems) suggest that the “merging” among different MCSs or between MCSs and other deep convective systems might be the reason. SMCSs and CMCSs together produce relatively the least anvil clouds in the onset phase, resulting from both the increased weight of the frequency of occurrence of connected MCSs (40%) and the overall reduced anvil productivity associated with large MCSs.

Thus an overall picture of the variability of MCSs after the onset phase of the MJO is a shift toward more “convective” organization because separated MCSs maximize after the onset, while their internal structures appear more “stratiform” because internally MCSs have weaker reflectivity above 8km. The corresponding changes in the anvil productivity show strong coherence with the organization (category & merging etc.) of MCSs.

            Further investigations on environmental conditions show that connected MCSs coincide with a more humid middle troposphere spatially, even at the same places a few days before they occur. The middle-tropospheric moistening both of the domain-mean and in regimes closely associated with deep convection peaks in the onset phase. Moistening of the free troposphere around MCSs shows relatively stronger moistening/drying below the 700 hPa before/after the onset phase compared to domain-mean averages. Lower-topped clouds are found to occur most frequently around connected MCSs and in active phases, consistent with the presence of a moister free troposphere. Coexistence of these phenomena suggests that the role of middle troposphere moisture in the formation of connected MCSs needs to be better understood.

========================================================

Figures:

Upper block: The cloud coverage (in percentage) of high cloud systems excluding MCSs (green solid), separated MCSs (blue dash) and connected MCSs (blue solid) in 8 MJO phases. The left y-axis is for separated MCSs and connected MCSs. 

Middle block: Differences of true CFADs of the deep precipitating areas of connected MCSs and separated MCSs for Small (a), Medium (b) and Large (c) systems. Black contours show True CFADs of Small (a), Medium (b) and Large (c) separated MCSs. Contour intervals are 0.025. To produce true CFADs, every layer of each CFAD for each phase is normalized so that the total frequency of occurrence at each layer of each CFAD is one. Then the difference is taken by subtracting one composite from another for each phase. Black and white markers ‘x’ mark the area where the difference has passed the 95% significance test of the 8-sample t-test. 

Bottom block: Differences of specific humidity as functions of phases for the IPO region. Reading from top row to bottom row: separated MCSs, connected MCSs and Large separated MCSs, respectively. Reading from left column to right column: day 0, -1, -2 and -3, respectively. The difference of the specific humidity is defined as the difference between the mean specific humidity profiles of associated systems and that of areas out side high cloud systems.