Mechanisms of poleward propagating, intraseasonal convective anomalies in a cloud-system resolving model William Boos & Zhiming Kuang Dept. of Earth & Planetary Sciences Harvard University October 16, 2009
Feb 24, 2016
Mechanisms of poleward propagating, intraseasonal convective anomalies in a
cloud-system resolving model
William Boos & Zhiming KuangDept. of Earth & Planetary Sciences
Harvard UniversityOctober 16, 2009
Outline• Background and observations
• Results from quasi-2D models with explicit convection
• Mechanisms of instability and propagation
Main message:For intraseasonal convective anomalies during boreal summer:• Poleward propagation occurs due to convectively-coupled beta-drift of a vorticity strip• Instability occurs due to moisture-radiation feedback
Borealsummer MJO lifecycle of TRMM precip
diagnostic from CLIVAR MJO working group, based on EOFs after Wheeler & Hendon (2004)
propagation has prominent poleward component
some events do exhibit poleward propagation without eastward propagation
Viewed as poleward migration of ITCZ
1.5 m/s
NOAA OLR anomalies, 80-100°E, summer 2001
Several events typically occur each boreal summer, modulating intensity of South Asian monsoon
History of axisymmetric model studies
• Land-atmosphere interactions (Webster & Chou 1980)
• Poleward gradient of convective instability (Gadgil & Srinivasan 1990)
• Dynamical coupling of anomalies to baroclinic mean state (Bellon & Sobel 2008, Jiang et al. 2004)
… but all of these studies use idealized parameterizations of moist convection, and mode characteristics depend on convective closure
Test in model with explicit convection
• System for Atmospheric Modeling (SAM, Khairoutdinov & Randall 2003)
• 1 km horizontal resolution• Beta-plane, 70°N – 70°S• 4 zonal grid points• Oceanic lower boundary
with prescribed SST
precipitation
Model with wider zonal dimension4 zonal grid points 32 zonal grid points
Precipitation snapshots when ITCZ is near 10N:
60
40
20
0
-20
-40
-60
latit
ude
60
40
20
0
-20
-40
-60
Old domain: 140° meridional x 4 km zonal
New domain: 140° meridional x 960 km zonal
For computational efficiency, use RAVE methodology of Kuang, Blossey & Bretherton (2005):
30 km horizontal resolution, RAVE factor 15
Similar results obtained for RAVE factors ranging from 1-15 at 30 km resolution, and for one standard run with 5 km resolution
x (km)x (km)
mm/day
0 500 960
Precipitation in wide-domain model
0.5 m/s
mm/day
Zonal meanvertical structure for wide domain
m/s
m/s
m/s
Composite 950 hPa vorticity
• Zonal mean vorticity satisfies necessary condition for barotropic instability
• Anomalies form closed cyclone for part of poleward migration, and zonal strip for remainder
• Suggestive of “ITCZ breakdown” (Ferreira & Schubert 1997)
zonal mean vorticity
compositerelative vorticity
latit
ude
Animation of two events
Poleward drift of vorticity patch/strip on β-plane… coupled to moist convection
latit
ude
x grid point
Shading: 930 hPa relative vorticity
Black contours: precipitation
Schematic: propagation mechanism
1. deep ascent creates (barotropically unstable) low-level vortex strip
3. Ekman pumping in vortex strip humidifies free-troposphere poleward of original deep ascent, shifting convection poleward
Convectively-coupled beta-drift of vortex strip
deep ascent
2. perturbed vortex strip migrates poleward
deep ascent
vorticity anomaly
xy
vorticity anomaly
yz
Test mechanism in dry model
• β-drift biases low-level convergence poleward of free-tropospheric heating
applied (constant) thermal forcing surface meridional wind
Surface wind in dry model
latit
ude
constant imposed heating
x grid point
Looks like unstable moisture mode
J/kg
MSE tendencies
composite moist static energy anomaly
Model tests of instability mechanismmm/dayfixed radiative cooling
fixed surface heat fluxes
control run
Precipitation Hovmollers:
Instability mechanism is non-unique
Dashed black lines denote latitude of peak moist static energy anomaly
Control run Run with fixed radiative cooling
Summary• Axisymmetric cloud permitting models fail to produce robust poleward
propagating, intraseasonal convective anomalies
• Meridional “bowling alley” domains O(1000 km) wide do produce such anomalies
– Suggested propagation mechanism:convectively-coupled beta-drift of vortex strip
– Anomalies destabilized by moisture-radiation feedback– Perhaps slowed and made more coherent by WISHE– Multiple instability mechanisms can operate, with structural changes
• Future work:– Behavior in wider domains– Validation of mechanism in simpler models
Additional slides
Wide domain permits high amplitude eddiesla
titud
e
x (105 m)
g/kgday 0 day 20 day 30 day 41 day 53
composite 930 hPa wind and humidity
Why does the wide domain make a difference? It’s the eddies…
J/kg
MSE tendencies
composite moist static energy anomaly
advective componentstotal & zonal mean advection
Propagation speed scaling
• Plots of precip and v wind for beta 0.75, 1, 2
Observed vertical structuredata: ERA-40 Reanalysis, composite of strong poleward events 1979-2002
latitude
pres
sure
(hPa
)
Note some similarties to eastward moving MJO
latitude
pres
sure
(hPa
)
Observed vertical structuredata: ERA-40 Reanalysis, composite of strong poleward events 1979-2002
Behavior depends on zonal width,not zonal d.o.f.
time (days)
latit
ude
latit
ude
5 km resolution with 32 zonal grid points
30 km resolution with 32 zonal grid points
OLR in wide domain model
Vertical structure for wide domain
(green line denotes position of peak precip signal used for compositing)
m/s
Turn off both WISHE & radiative feedbacksno WISHE or radiative feedbacks
control
time (days)
mm/dayPrecipitation:
MSE budget for run without WISHE or radiative feedbacks
moist static energy anomaly
latitude (degrees)
pres
sure
(hPa
)
moist static energy tendencies
W m
-2
“Convective downdraft instability”
J/kg