Mesoscale Convective Systems in the
Initiation of the MJO
Jian Yuan and Robert A. Houze
University of Washington
CloudSat/CALIPSO Science Team Meeting
Montreal, Quebec, Canada, 16 June
The Madden-Julian
Oscillation (MJO):
• Play important roles in
weather and climate
• Current prediction skill,
especially for the initial
phase of MJO is very
limited
• Cumulus parameterizations
in GCMs is the primary
limiting factor in MJO
simulation and prediction.
(Zhang et al. 2010, DYNAMO)
(Courtesy of US CLIVAR MJO Working Group)
MJO initiation processes
Feedbacks between:
• Clouds
• Radiative heating
Fundamental
• Convectionprocesses
related
to MCSs that are
• Precipitation
crucial
to understand MJO:
• Ocean
a key
to understanding
•arethe
diabatic
heating
thestructure
MJO.
• convective sensitivity to
environmental moisture
• cloud microphysics
• convective organization
Courtesy of Zhang et al. 2009, DYNAMO
After Stephens et al. 2004, “Humidistat Feedback”
MCSs including both raining and anvil
components are identified using A-Train
instruments
Yuan and Houze 2010
MODIS TB11 + AMSR-E (Yuan and Houze 2010)
combined to find“cold centers” & “raining areas”
Locate 1st
closed
contour
Use 260 K
threshold
Use 1 mm/h
threshold for
rain rate
Associate
pixels with
nearest
cold center
Use 6 mm/h
threshold for
heavy rain
MCS Criteria (Yuan and Houze, 2010)
Systems whose largest raining cores have
• Area > 2000 km2
• Min TB11 ≤ 220 K
Must have one dominant core
• with intense cells, and
• accounting for >70% rain area
56% all tropical rain
MCSs are further
divided to two
groups :
1. Separated (40
% rain fall)
1. Connected
(>=3 MCSs
share the
same rain
feature, 16%
rain fall)
• Separated MCS: Frequently found over all convective
zones, especially continents
• Connected MCS: more organized convection, primarily
found over warm ocean area
MODIS/AMSR-E/CloudSat
identifies MCSs
obtains the global distribution of
MCSs
investigates variability of MCSs in
MJO
(EIO:-15-15oN;75-100oE; Composite of 8 phases; Wheeler
and Hendon 2004)
More Connected MCSs observed in MJO
active phases
OLR
Deeper MCSs observed in
pre-onset, initial and active phases
Low level Θe likely determines the Tb_min
(“hot tower” hypothesis)
Phase 1-3
Climatology of EIO:
•T150 hp ≈ 205 K
•Θe150 hp≈ 352.6 K
Phase 5-7
Moisture effects need to be better understood
Deeper
MCSs
Less MCSs; More MCSs;
more
less
organized
organized
Summary and Conclusions
 A-Train instruments make it possible to identify
MCSs (raining + anvil components) globally
 MJO pre-onset phase  active phases over EIO:
 Deeper MCSs & Warmer low level Θe (both)
 Less  More MCSs
 Relatively Less  More organized MCSs
 Drier  Moister middle troposphere
The moisture effect needs to be better
understood.
End
MJO activities viewed in OLR
MCSs Over the Whole Tropics: oceanic conditions
favor larger systems
Smallest 25% (<12,000 km2)
Largest 25% (>40,000 km2)
“Superclusters”