MJO Modulation of Lightning in
Mesoscale Convective Systems
Katrina S. Virts and
Robert A. Houze, Jr.
University of Washington
Seminar, Pacific Northwest National Laboratory, Richland, WA, 4 June 2014
Mesoscale Convective Systems (MCSs)
Radar echoes showing
the precipitation
in the 3 MCSs
Stratiform
Precipitation
Convective
Precipitation
Madden-Julian Oscillation
Intraseasonal time scales (~30-80
days)
Enhanced convection develops
over equatorial Indian Ocean
Eastward propagation
Associated circulation anomalies
Image courtesy Madden and Julian (1972)
MJO modulation of cloud population
Field campaigns (TOGA COARE, DYNAMO/AMIE)
Satellite observations
– Passive sensors
• “Superclusters” (Nakazawa 1988)
• MJO “modulates cloud clusters of all sizes, but larger clusters
are proportionately more affected than smaller clusters”
(Mapes & Houze 1993)
• MJO “associated with weaker or stronger mesoscale
organization of deep convection” (Tromeur & Rossow 2010)
MJO modulation of cloud population
Satellite observations (continued)
– TRMM
• Shallow cumulus and congestus prior to onset of deep
convection (Benedict & Randall 2007)
• “The precipitating cloud population of the Madden-Julian
Oscillation over the Indian and western Pacific Oceans”
(Barnes and Houze 2013)
– CloudSat
• “A familiar evolution of cloud type predominance” (Riley et al.
2011)
• “Shallow and congestus clouds in advance of the [MJO] peak,
deep clouds near the peak, and upper level anvils after the
peak” (Del Genio et al. 2012)
– Other A-Train satellites (Yuan and Houze 2013)
MJO modulation of cloud population
(Barnes and Houze 2013)
Echo types:
– Isolated shallow echoes (ISEs) — echo tops at least 1 km
below freezing level
– Deep convective cores (DCCs) — radar echo ≥ 30 dBZ
up to at least 8 km
– Wide convective cores (WCCs) — radar echo ≥ 30 dBZ
covering at least 800 km2
– Broad stratiform regions (BSRs) — stratiform echo covering
at least 50,000 km2
Indian Ocean
NW Western Pacific SE Western Pacific
Image courtesy Barnes and Houze (2013)
MJO modulation of
lightning
Out of phase with rain
(Morita et al. 2006)
Image courtesy Morita et al. (2006)
MJO active
MJO inactive
Image courtesy Kodama et al. (2006)
MJO modulation of lightning
Break period (phases 8-1-2) minus
active period (phases 4-5-6)
Image courtesy Virts et al. (2013)
Out of phase with rain
(Morita et al. 2006)
Suppressed over large
islands during active
period (Kodama et al.
2006)
Modulation of diurnal cycle
(Virts et al. 2013)
MJO modulation of lightning
Break period (phases 8-1-2) minus
active period (phases 4-5-6)
Image courtesy Virts et al. (2013)
Out of phase with rain
(Morita et al. 2006)
Suppressed over large
islands during active
period (Kodama et al.
2006)
Modulation of diurnal cycle
(Virts et al. 2013)
What about individual
convective clouds?
Identifying MCSs using A-Train data
MODIS 10.8 m
brightness
temperature
AMSR-E rain rate
Years included:
2007-2010
Details in Yuan and Houze 2010
260K
Separated
HCS
Details in Yuan and Houze 2010
260K
Closed
contour
Separated
HCS
Details in Yuan and Houze 2010
260K
Closed
contour
Separated
HCS
“HCS”
Details in Yuan and Houze 2010
260K
Heavy Rain
Closed
Rain
contour
Separated
HCS
“HCS”
Details in Yuan and Houze 2010
“Separated”
Separated
HCSMCS
active
260K
Heavy Rain
Closed
Rain
contour
“HCS”
“Connected”
active MCS
Details in Yuan and Houze 2010
World-Wide Lightning Location Network
(WWLLN)
Global network of 70+ sensors
Monitors very low frequency waves
Lightning strokes located to within 5 km and a few s
Preferentially detects cloud-to-ground lightning
World-Wide Lightning Location Network
(WWLLN)
Lightning in one-hour window
– Separate coordinate system for each MCS, centered on
largest raining core
– Lightning in cloudy grid boxes (lightning density)
Indian
Ocean
% CMCSs
MCS lightning density
29.5
Maritime
Continent
17.6
Western
Pacific
30.0
SPCZ
29.3
Indian
Ocean
Maritime
Continent
Western
Pacific
SPCZ
% CMCSs
29.5
17.6
30.0
29.3
MCS lightning density
2.9
26.5
2.5
7.6
CMCSs most frequent with
peak precip.
SMCS timing varies,
reflects MJO stage
CMCSs experience greater
variability
MJO modulation of lightning in
Maritime Continent SMCSs
More frequent lightning, broader
lightning maximum during break period
Lifted Index (LI)
Measure of lower-tropospheric stability
LI = Te ( p) - Tp ( p)
Negative LI  parcel warmer than environment
Calculate using ERA-Interim fields
MCS environments more unstable during break
period
MJO modulation of
lightning density
Peak lightning at end of
break period
SPCZ: peak lightning at
beginning of break
period
Lower lightning density
in CMCSs
TRMM radar precipitation features (RPFs)
Contiguous areas with near-surface
rain rate > 0
Use features with maximum 30 dBZ height > 6 km
Size equivalent to smallest and largest 50% of
MCSs
Years included: 1998-2012
RPF data obtained from University of Utah
TRMM database. Details in Liu et al. 2008
TRMM radar precipitation features (RPFs)
Contiguous areas with near-surface
rain rate > 0
Use features with maximum 30 dBZ height > 6 km
Size equivalent to smallest and largest 50% of
MCSs
Years included: 1998-2012
convective rain volume
convective rain fraction =
convective + stratiform rain volume
RPF data obtained from University of Utah
TRMM database. Details in Liu et al. 2008
MJO modulation of
convective rain fraction
Peak at end of
break period
Varies strongly with
RPF size
MJO modulation of MCS characteristics
Isolated deep convection begins to aggregate
– Strong instability  strong updrafts  more lightning
– Dry mid/upper troposphere  smaller stratiform areas
MCSs become more numerous
– Stability increases  less lightning
– Increasingly extensive stratiform rain areas
MCSs increasingly more connected
– CMCS occurrence peaks with precipitation
MCSs decrease in number, size, connectedness
– Smaller stratiform areas  rain is more convective
– Increasing instability during break period  more lightning
MJO modulation of MCS characteristics
Isolated deep convection begins to aggregate
– Strong instability  strong updrafts  more lightning
– Dry mid/upper troposphere  smaller stratiform areas
MCSs become more numerous
– Stability increases  less lightning
– Increasingly extensive stratiform rain areas
MCSs increasingly more connected
– CMCS occurrence peaks with precipitation
MCSs decrease in number, size, connectedness
– Smaller stratiform areas  rain is more convective
– Increasing instability during break period  more lightning
MJO modulation of MCS characteristics
Isolated deep convection begins to aggregate
– Strong instability  strong updrafts  more lightning
– Dry mid/upper troposphere  smaller stratiform areas
MCSs become more numerous
– Stability increases  less lightning
– Increasingly extensive stratiform rain areas
MCSs increasingly more connected
– CMCS occurrence peaks with precipitation
MCSs decrease in number, size, connectedness
– Smaller stratiform areas  rain is more convective
– Increasing instability during break period  more lightning
MJO modulation of MCS characteristics
Isolated deep convection begins to aggregate
– Strong instability  strong updrafts  more lightning
– Dry mid/upper troposphere  smaller stratiform areas
MCSs become more numerous
– Stability increases  less lightning
– Increasingly extensive stratiform rain areas
MCSs increasingly more connected
– CMCS occurrence peaks with precipitation
MCSs decrease in number, size, connectedness
– Smaller stratiform areas  rain is more convective
– Increasing instability during break period  more lightning
MJO modulation of MCS characteristics
(simplified)
Few MCSs, mainly shallow or isolated deep convection
“Younger” MCSs with strong convection
“Older” MCSs with mature stratiform rain areas
Familiar…
Similar evolution in 2-4 day waves
during MJO active period
Image courtesy Zuluaga and Houze (2013)
Stretched building block model
(Mapes et al. 2006)
Convective clouds and MCSs “in different stages of a largescale wave have different durations of shallow convective,
deep convective, and stratiform anvil stages in their life
cycles,” such that evolution of mean characteristics of
convective clouds aligns with the evolution of individual
clouds.
Conclusions
MCSs over land contain more vigorous convection, more
lightning
MCSs over the ocean are more connected
Larger, more connected, and more numerous MCSs
during MJO active period
Peak lightning and convective rain fraction just prior to
active period (except over SPCZ)
Evolution of mean MCS characteristics aligns with MCS
lifecycle (stretched building block)
This work was funded by NASA (# NNX13AQ37G)
and the Department of Energy (#DE-SC0008452).