Agnieszka Mrowiec - giss.nasa.gov
Transcript of Agnieszka Mrowiec - giss.nasa.gov
isentropic analysis of convective motionsAgnieszka Mrowiec
Olivier Pauluis, Ann Fridlind, Andy Ackerman
*Image from “Upper atmosphere in motion” - C.O. Hines, 1974
• case study and model
• convective systems
• gravity waves
• method
• results
• new applications
24 January -MCS moved over NW Top End Overnight from squalline previous afternoon/evening;
96mm to 9am at Darwin airport;
24 January -MCS moved over NW Top End Overnight from squalline previous afternoon/evening;
96mm to 9am at Darwin airport;
The TWP-ICE CRM Intercomparison Specification and First ResultsAnn Fridlind ([email protected]), Andrew Ackerman ([email protected]), Adrian Hill ([email protected]),Jon Petch ([email protected]), Paul Field ([email protected]), Greg McFarquhar ([email protected]),Shaocheng Xie ([email protected]), and Minghua Zhang ([email protected])
TWP-ICE CRM intercomparison specification http://www.giss.nasa.gov/∼fridlind/twp-ice/info
! active and suppressed monsoon conditions only! 18 January - 3 February 2006 (16 days)! horizontal domain size = 176 x 176 km! vertical domain size ≥ 24 km! periodic boundary conditions (idealized marine)! sea surface temperature = 29◦C, albedo = 0.07! interactive surface and radiative fluxes! idealized ozone profile from observations (Figure 2)! idealized aerosol profile from observations (Figure 3)! domain-mean large-scale forcings from observations! apply forcings at full strength below 15 km! nudge to observations with a 6-h time scale! two simulations
! baseline: nudge to observations only above 15 km! sensitivity: nudge to observations above 0.5 km
Figure 1. TWP-ICE domainSource: Xie
100 101 102 103 104 105
Ozone, ppb
0
5
10
15
20
25
30
35
Altit
ude,
km
1/20 5 UTC1/22 5 UTC1/23 5 UTC
0 3.5•1018 7.0•1018
Ozone, #/cm2
0
5
10
15
20
25
30
35
Altit
ude,
km
OM
I = 2
40 D
U =
6.5E
18
1.8E
18
4.7E18
Figure 2: Idealized ozone profileData: Allen, OMI
20060120 - 20060127
0.01 0.10 1.00 10.00Diameter [um]
10-6
10-4
10-2
100
102
104
106
dN/d
logD
p [c
m-3]
0.5 km 1.5 km 2.5 km 3.5 km 5.5 km 6.5 km 7.5 km 8.5 km 9.5 km10.5 km11.5 km12.5 km13.5 km14.5 km
Figure 3: Idealized aerosol profileData: Allen, Gallagher, Williams
! preliminary results due 1 March! final results due 1 July
Figure 4: Baseline simulation at 22.375 (event B)Simulation: Fridlind/Ackerman
Where do simulations and measurements disagree? http://www.giss.nasa.gov/∼fridlind/twp-ice/data
! approach! sample models appropriately! match spatial resolution! poor statistics? use envelope! pursue all measurements
! examples (Figure 5)! precipitation
! models agree within ∼25%! measurements may be adequate
to constrain simulations! ice water path
! models disagree by ∼50%! measurements appear inadequate
to constrain simulations! liquid water path
! models disagree by ∼10%! measurements appear inadequate
to constrain simulations
20 22 24 26 28Julian Day
02468
10
Prec
ipita
tion
Rate
(mm
/h) + Forcing data (mean=0.78)
C-Pol radar (mean=0.79)DHARMA (mean=0.80)ISUCRM_2D (mean=1.00)UKMO (mean=0.87)
Event AEvent B
Event C
20 22 24 26 28Julian Day
02468
10
Ice
Wat
er P
ath
(kg/
m2 )
+ Liu retrieval (mean=0.26)ARM COMBRET (mean=0.08)DHARMA (mean=0.52)ISUCRM_2D (mean=0.34)UKMO (mean=0.41)
20 22 24 26 28Julian Day
012345
Liqu
id W
ater
Pat
h(k
g/m
2 )
+ MWR retrieval (mean=0.07)ARM COMBRET (mean=0.10)DHARMA (mean=0.53)ISUCRM_2D (mean=0.47)
20 22 24 26 28Julian Day
01020304050
Max
imum
Ver
tical
Win
d Sp
eed
(m/s
)
DHARMA (mean=14.7)UKMO (mean=17.0)
Figure 5. Baseline simulations, measurementsSimulations: Fridlind/Ackerman, Park/Wu, Hill/PetchMeasurements: Xie/May, Liu, Turner, Comstock/McFarlane/Mather
Does ice sublimation hydrate the upper troposphere?
! simulated budget terms! actual moisture tendency! large-scale horizontal advection! large-scale subsidence/ascent! vertical advection and mixing! exchange with hydrometeors! nudging! budget residual
! approach! time periods: 6-hour, 24-hour! first consider all elevations! then focus on tropopause layer
Baseline
20.75 Ordinal 2006
-10 -5 0 5 106-Hour Mean Tendency (K/d)
0
5
10
15
20
Altit
ude
(km
)
(dQv/dt)act(dQv/dt)ls(dQv/dt)sub(dQv/dt)adv(dQv/dt)micro(dQv/dt)nudge(dQv/dt)err
Baseline
22.75 Ordinal 2006
-10 -5 0 5 106-Hour Mean Tendency (K/d)
0
5
10
15
20
Altit
ude
(km
)
(dQv/dt)act(dQv/dt)ls(dQv/dt)sub(dQv/dt)adv(dQv/dt)micro(dQv/dt)nudge(dQv/dt)err
Baseline
30.75 Ordinal 2006
-10 -5 0 5 106-Hour Mean Tendency (K/d)
0
5
10
15
20
Altit
ude
(km
)
(dQv/dt)act(dQv/dt)ls(dQv/dt)sub(dQv/dt)adv(dQv/dt)micro(dQv/dt)nudge(dQv/dt)err
Figure 6: 6-hour water vapor budgets at all elevations! ice is a net sink for water vapor above the melting layer! net hydration by hydrometeors occurs primarily below 5 km! water vapor lifted by convection deposits to hydrometeors
Baseline
20.75 Ordinal 2006
-0.02 -0.01 0.00 0.01 0.026-Hour Mean Tendency (K/d)
14
15
16
17
18
19
20
Altit
ude
(km
)
(dQv/dt)act(dQv/dt)ls(dQv/dt)sub(dQv/dt)adv(dQv/dt)micro(dQv/dt)nudge(dQv/dt)err
Baseline
22.75 Ordinal 2006
-0.02 -0.01 0.00 0.01 0.026-Hour Mean Tendency (K/d)
14
15
16
17
18
19
20
Altit
ude
(km
)
(dQv/dt)act(dQv/dt)ls(dQv/dt)sub(dQv/dt)adv(dQv/dt)micro(dQv/dt)nudge(dQv/dt)err
Baseline
30.75 Ordinal 2006
-0.02 -0.01 0.00 0.01 0.026-Hour Mean Tendency (K/d)
14
15
16
17
18
19
20
Altit
ude
(km
)
(dQv/dt)act(dQv/dt)ls(dQv/dt)sub(dQv/dt)adv(dQv/dt)micro(dQv/dt)nudge(dQv/dt)err
Figure 7: 6-hour water vapor budgets near the tropopause! hydrometeors are a mean water vapor sink at the tropopause! large amounts of water vapor cycle through the TTL! only one 6-hour period with moistening above 14 km (right)
Baseline
20.50 Ordinal 2006
-0.2 -0.1 0.0 0.1 0.224-Hour Mean Tendency (K/d)
14
15
16
17
18
19
20
Altit
ude
(km
)
(dQv/dt)act(dQv/dt)ls(dQv/dt)sub(dQv/dt)adv(dQv/dt)micro(dQv/dt)nudge(dQv/dt)err
Baseline
23.25 Ordinal 2006
-0.2 -0.1 0.0 0.1 0.224-Hour Mean Tendency (K/d)
14
15
16
17
18
19
20
Altit
ude
(km
)
(dQv/dt)act(dQv/dt)ls(dQv/dt)sub(dQv/dt)adv(dQv/dt)micro(dQv/dt)nudge(dQv/dt)err
Baseline
24.50 Ordinal 2006
-0.2 -0.1 0.0 0.1 0.224-Hour Mean Tendency (K/d)
14
15
16
17
18
19
20
Altit
ude
(km
)
(dQv/dt)act(dQv/dt)ls(dQv/dt)sub(dQv/dt)adv(dQv/dt)micro(dQv/dt)nudge(dQv/dt)err
Figure 8: 24-hour water vapor budgets near the tropopause! ice is a net sink during monsoon events A, B, and C! moistening does not occur if RHI is already high! uncertainty in RHI measurements will need to be considered
Does dehydration occur in overshooting convection?Baseline
20.25 Ordinal 2006
-0.02 -0.01 0.00 0.01 0.026-Hour Mean Tendency (K/d)
14
15
16
17
18
19
20
Altit
ude
(km
)
(dQv/dt)act(dQv/dt)ls(dQv/dt)sub(dQv/dt)adv(dQv/dt)micro(dQv/dt)nudge(dQv/dt)err
Baseline
23.00 Ordinal 2006
-0.02 -0.01 0.00 0.01 0.026-Hour Mean Tendency (K/d)
14
15
16
17
18
19
20
Altit
ude
(km
)
(dQv/dt)act(dQv/dt)ls(dQv/dt)sub(dQv/dt)adv(dQv/dt)micro(dQv/dt)nudge(dQv/dt)err
Baseline
23.75 Ordinal 2006
-0.02 -0.01 0.00 0.01 0.026-Hour Mean Tendency (K/d)
14
15
16
17
18
19
20
Altit
ude
(km
)
(dQv/dt)act(dQv/dt)ls(dQv/dt)sub(dQv/dt)adv(dQv/dt)micro(dQv/dt)nudge(dQv/dt)err
Figure 9: 6-hour vapor vapor budgets near the tropopause! peak activity of events A, B, and C are shown! only the strongest event dehydrates above the tropopause! results may be sensitive to simulated ice fall speeds
Baseline
20.50 Ordinal 2006
-0.02 -0.01 0.00 0.01 0.0224-Hour Mean Tendency (K/d)
14
15
16
17
18
19
20
Altit
ude
(km
)
(dQt/dt)act(dQt/dt)ls(dQt/dt)sub(dQt/dt)adv(dQt/dt)sed(dQt/dt)nudge(dQt/dt)err
Sensitivity
20.50 Ordinal 2006
-0.02 -0.01 0.00 0.01 0.0224-Hour Mean Tendency (K/d)
14
15
16
17
18
19
20
Altit
ude
(km
)
Baseline
23.25 Ordinal 2006
-0.02 -0.01 0.00 0.01 0.0224-Hour Mean Tendency (K/d)
14
15
16
17
18
19
20
Altit
ude
(km
)
(dQt/dt)act(dQt/dt)ls(dQt/dt)sub(dQt/dt)adv(dQt/dt)sed(dQt/dt)nudge(dQt/dt)err
Sensitivity
23.25 Ordinal 2006
-0.02 -0.01 0.00 0.01 0.0224-Hour Mean Tendency (K/d)
14
15
16
17
18
19
20
Altit
ude
(km
)
Baseline
24.50 Ordinal 2006
-0.02 -0.01 0.00 0.01 0.0224-Hour Mean Tendency (K/d)
14
15
16
17
18
19
20
Altit
ude
(km
)
(dQt/dt)act(dQt/dt)ls(dQt/dt)sub(dQt/dt)adv(dQt/dt)sed(dQt/dt)nudge(dQt/dt)err
Sensitivity
24.50 Ordinal 2006
-0.02 -0.01 0.00 0.01 0.0224-Hour Mean Tendency (K/d)
14
15
16
17
18
19
20
Altit
ude
(km
)
Figure 10: 24-hour total water budgets near the tropopause! integrations over events A, B, and C are shown (baseline and sensitivity results for each event)! convective events result in drying in all except the baseline simulation for event B! but water vapor tendency is often small relative to other terms, including budget residual
Next steps! refine model-data comparisons! time series analysis of budgetsAcknowledgments! NASA Radiation Sciences! NASA Advanced Supercomputing! DOE ARM Program! DOE ACRF data archive! TWP-ICE science team! ACTIVE science team
• Case: TWP ICE 2006, active monsoon event of January 23 - 24 2006
• Intercomparison: Fridlind et al. 2010 Fridlind et al. 2012
TWP ICE - Tropical Warm Pool International Cloud Experiment was a multi-agency project centered in Darwin, Australia. A detailed description of experiment conditions, general climate, and measurements can be found in May et al. (2008).
case study:mesoscale convective system
(MCS)
simulation
DHARMA - Distributed Hydrodynamic Aerosol and Radiative Modeling Application(Stevens et al., 2002; Ackerman et al., 2003) cloud resolving model
Detailed description of model configuration may be found e.g. in Fridlind et al. 2012 or Mrowiec et al. 2012
• domain 176 km 176 km
• horizontal resolution 900 m
• stretched vertical grid 100 - 250 m
• periodic lateral boundary conditions
• model domain height of 24 km
• initial conditions derived from the mean observed profiles of potential temperature
and water vapor mixing ratio
• uniform sea surface temperature of 29C, experimental domain idealized as marine
• surface albedo fixed at 0.07
• large-scale forcings were based on observations (Xie et al., 2010)
• two-moment microphysics based on the 5-class scheme of Morrison et al. (2009)
• trimodal aerosol profile derived from observations (Fridlind et al., 2010)
structure of a thunderstormS
Building blocks from the point of view of convective parameterization:
• Convective updrafts• Convective downdrafts• Stratiform updrafts• Stratiform downdrafts
February 5th, 2008, Picture taken from the International Space Station
Frederick and Schumacher, MWR 2008
• INTENSITY: any point with reflectivity higher than 40 dBZ is called convective
• PEAKEDNESS: any point that exceeds the average reflectivity taken over surrounding area (radius of11 km) by reflectivity difference shown by the line below are also convective
• SURROUNDINGS: any point within a radius dependent on the mean background reflectivity from convective center is also included in convective area
• remaining precipitating areas are called stratiform
convective / stratiform partitioning
Ze ~ 10 log (Z(1mm)-1∫C D6 N(D) dD)where C is a constant, D is droplet diameter and N is a droplet size distribution
Steiner (1995) algorithm based on simulated radar reflectivity expressed in decibel (dBZ) and typically is normalized by the echo of a 1 mm diameter droplet:
3km
6km
what rain radar can see
• partitioning done at altitude of 3km
• we have additional criterium of reflectivity greater than 5 dBZ at 6 km to eliminate shallow convection
• everything else we call a non-precipitating region
convective / stratiform partitioning
C
S
S
S
C
C
C
S
resulting cloud mask from a simulation at some time step during the peak of the eventconvective regions (C) stratiform regions (S)non-precipitating regions (white)
Warm colors w>0Cool colors w<0
updrafts downdrafts
simulated MCS: updrafts / downdrafts
everything looks great...., right?
convective motions generate gravity waves...
Warm colors w>0Cool colors w<0
updrafts downdrafts
buoyant oscillations(gravity waves)
overshoot
Is there an effective way to exclude these oscillations when analyzing the properties of drafts in the sub-regions?
Isentropic Analysis of Convective Motions
A similar approach can be taken to analyze convective motions.
Here, we will replace both horizontal coordinates (x and y) by the
equivalent potential temperature.
In practice, we introduce an isentropic averaging:
We then compute the mean upward isentropic mass flux
and the isentopic streamfunction:
First, we define our isentropic surfaces as surfaces of constant equivalent potential temperature, defined following Emanuel (94) textbook.
Isentropic coordinates have been widely used in studies of the global circulation to separate large-scale planetary overturning from synoptic scale eddies. Here, we are using it to analyze convective scale overturning.
2. Methodology59
a. Isentropic averaging60
Averaging circulation on isentropic surfaces provides a more direct description of the La-61
grangian flow. There are several ways in which the isentropic surfaces can be defined. Here we62
base the calculation on the equivalent potential temperature defined as ✓e
= T
⇧RH�Rvqv
C exp�Lvqv
CT
�,63
where C is given by C = cp
+ cl
qt
following ? textbook.64
Condensation and precipitation have little impact on ✓e
, thus the change in it in the free65
troposphere is mainly caused by the radiative cooling at the cloud top. Because of deep66
convection, the average temperature profile in the tropics is close to moist adiabat (Xu and67
Emanuel 1989), which makes ✓e
a good candidate for an invariant of the flow.68
discuss not including freezing and fusion.69
Averaging on isentropic coordinate captures the mean convective transport of air from70
the surface (where the evaporation and sensible heat are the sources of ✓e
) to the cloud top71
(where radiative cooling plays role of the sink of ✓e
).72
Eulerian mass flux can be defined as73
M(x, y, z, t) = ⇢o
(z)(w(x, y, z, t)� w(z, t)), (1)
where ⇢o
(z) is the mean atmospheric density profile, w(x, y, z, t) is the vertical wind compo-74
nent and w(z, t) is the horizontal mean of the vertical wind. This mass flux is spatially and75
temporally averaged to give a bulk number for a profile of a mass transport76
M(z) =1
L2T
ZZ Z⇢o
(z)(w(x, y, z, t)� w(z, t))dx dy dt, (2)
but it can also be defined in the sub-regions of the circulation, such as upward or downward77
mass fluxes:78
M+(z, t) =
1
L2
ZZM(x, y, z, t)H(M > 0)dx dy, (3)
M�(z, t) =
1
L2
ZZM(x, y, z, t)H(w < 0)dx dy, (4)
4
where
2. Methodology59
a. Isentropic averaging60
Averaging circulation on isentropic surfaces provides a more direct description of the La-61
grangian flow. There are several ways in which the isentropic surfaces can be defined. Here we62
base the calculation on the equivalent potential temperature defined as ✓e
= T
⇧RH�Rvqv
C exp�Lvqv
CT
�,63
where C is given by C = cp
+ cl
qt
following ? textbook.64
Condensation and precipitation have little impact on ✓e
, thus the change in it in the free65
troposphere is mainly caused by the radiative cooling at the cloud top. Because of deep66
convection, the average temperature profile in the tropics is close to moist adiabat (Xu and67
Emanuel 1989), which makes ✓e
a good candidate for an invariant of the flow.68
discuss not including freezing and fusion.69
Averaging on isentropic coordinate captures the mean convective transport of air from70
the surface (where the evaporation and sensible heat are the sources of ✓e
) to the cloud top71
(where radiative cooling plays role of the sink of ✓e
).72
Eulerian mass flux can be defined as73
M(x, y, z, t) = ⇢o
(z)(w(x, y, z, t)� w(z, t)), (1)
where ⇢o
(z) is the mean atmospheric density profile, w(x, y, z, t) is the vertical wind compo-74
nent and w(z, t) is the horizontal mean of the vertical wind. This mass flux is spatially and75
temporally averaged to give a bulk number for a profile of a mass transport76
M(z) =1
L2T
ZZ Z⇢o
(z)(w(x, y, z, t)� w(z, t))dx dy dt, (2)
but it can also be defined in the sub-regions of the circulation, such as upward or downward77
mass fluxes:78
M+(z, t) =
1
L2
ZZM(x, y, z, t)H(M > 0)dx dy, (3)
M�(z, t) =
1
L2
ZZM(x, y, z, t)H(w < 0)dx dy, (4)
4
We introduce isentropic averaging, where we replace both horizontal coordinates (x, y) by Θe. In practice the properties of air parcels are averaged over finite size bins of equivalent potential temperature.
(Pauluis and Mrowiec 2012)
constant equivalent potential temperature
NO
YES
at each z level, at each time step all particles that fall into a given ϴe bin (0.5 K)
In the process of this conditional averaging the fast oscillatory motions are filtered out from the slower thermodynamically driven circulation.
A similar approach can be taken to analyze convective motions.
Here, we will replace both horizontal coordinates (x and y) by the
equivalent potential temperature.
In practice, we introduce an isentropic averaging:
We then compute the mean upward isentropic mass flux
and the isentopic streamfunction:The mean isentropic mass flux may be defined as: FM=
A similar approach can be taken to analyze convective motions.
Here, we will replace both horizontal coordinates (x and y) by the
equivalent potential temperature.
In practice, we introduce an isentropic averaging:
We then compute the mean upward isentropic mass flux
and the isentopic streamfunction:
and the isentropic streamfunction:
This method allows to follow parcels with similar thermodynamic properties even when their trajectories are quite complicated. Reversible buoyant oscillations happen fast, their properties will not change between upward and downward motions, therefore they should get averaged out. Main updrafts are moist and have high equivalent potential temperature and downdrafts are drier and colder and will get nicely separated in this analysis.
isentropic analysis of convective motions
Isentropic analysis provides a more direct Lagrangian description of a flow, but in an averaged sense. The spatial information is lost.
isentropic mass flux
330 335 340 345 350 355 360O
e (K)
5
10
15
Altit
ude
(km
)
-0.020
-0.016
-0.012
-0.008
-0.004
0.000
0.004
0.008
0.012
0.016
0.020
Convective (C)
330 335 340 345 350 355 360O
e (K)
5
10
15
Altit
ude
(km
)
-0.020
-0.016
-0.012
-0.008
-0.004
0.000
0.004
0.008
0.012
0.016
0.020
Stratiform (S)
330 335 340 345 350 355 360O
e (K)
5
10
15
Altit
ude
(km
)
-0.020
-0.016
-0.012
-0.008
-0.004
0.000
0.004
0.008
0.012
0.016
0.020
Non-precipitating (NP)
330 335 340 345 350 355 360O
e (K)
5
10
15
Altit
ude
(km
)
-0.020
-0.016
-0.012
-0.008
-0.004
0.000
0.004
0.008
0.012
0.016
0.020
Total
majority of the upward mass flux in C
strongest C fluxes initiated at about 5km
weak S ascent
no ascent in NP region
S downdrafts more mass flux than C downdrafts
a significant overturning in the stratiform boundary layer - shallow convection after all?
upper level descent in S and NP has double minima - due to initial subsidence and then main convective burst that mixes up and warms the troposphere causing descent to be closer to the environmental mean
isentropic mass flux time series
Downward total
23.2 23.4 23.6 23.8 24.0 24.2 24.4Time (h)
5
10
15
Altit
ude
(km
) -0.20 -0.17 -0.14 -0.11 -0.07 -0.04 -0.01
a) Upward total
23.2 23.4 23.6 23.8 24.0 24.2 24.4Time (h)
5
10
15
Altit
ude
(km
)
0.01 0.04 0.07 0.11 0.14 0.17 0.20
b)
Downward convective
23.2 23.4 23.6 23.8 24.0 24.2 24.4Time (h)
5
10
15
Altit
ude
(km
)
-0.20 -0.17 -0.14 -0.11 -0.07 -0.04 -0.01
c) Upward convective
23.2 23.4 23.6 23.8 24.0 24.2 24.4Time (h)
5
10
15
Altit
ude
(km
)
0.01 0.04 0.07 0.11 0.14 0.17 0.20
d)
Downward stratiform
23.2 23.4 23.6 23.8 24.0 24.2 24.4Time (h)
5
10
15
Altit
ude
(km
)
-0.20 -0.17 -0.14 -0.11 -0.07 -0.04 -0.01
e) Upward stratiform
23.2 23.4 23.6 23.8 24.0 24.2 24.4Time (h)
5
10
15
Altit
ude
(km
) 0.01 0.04 0.07 0.11 0.14 0.17 0.20
f)
Downward clear sky
23.2 23.4 23.6 23.8 24.0 24.2 24.4Time (h)
5
10
15
Altit
ude
(km
)
-0.20 -0.17 -0.14 -0.11 -0.07 -0.04 -0.01
g) Upward clear sky
23.2 23.4 23.6 23.8 24.0 24.2 24.4Time (h)
5
10
15
Altit
ude
(km
)
0.01 0.04 0.07 0.11 0.14 0.17 0.20
h) Non-precipitatingNon-precipitating
Time (day)Time (day)
Time (day) Time (day)
Time (day) Time (day)
Time (day) Time (day)
isentropic streamfunction
330 335 340 345 350 355 360O
e (K)
5
10
15Al
titud
e (k
m)
-0.060
-0.050
-0.040
-0.030
-0.019
-0.009
0.001
Warm, moist air rising, colder drier air sinking
Convective overshoot
Moistening of subsiding air
Radiative cooling
Large scale forcing explains the location of the maximum and increase in equivalent potential temperature in updrafts
warm, moist air rising,
colder dry air sinking
A little bit of
convective overshoot...
Entrainment weakens
updrafts at low level
Moistening of
subsiding air
plenty of shallow
convection
isentropic streamfunction
TWP-ICE RCE
330 335 340 345 350 355 360O
e (K)
5
10
15
Altit
ude
(km
)
-0.060
-0.050
-0.040
-0.030
-0.019
-0.009
0.001
Warm, moist air rising, colder dry air sinking
Convective overshoot
Moistening of subsiding air
Large scale forcing explains the location of the maximum and increase in equivalent potential temperature in updrafts
freezing
(Pauluis and Mrowiec 2012)
isentropic vs. eulerian mass flux
-0.04 -0.02 0.00 0.02 0.04Mass Flux (kg m-2 s-1)
0
5
10
15
Altit
ude
(km
)
a)
-0.04 -0.02 0.00 0.02 0.04Mass Flux (kg m-2 s-1)
0
5
10
15
Altit
ude
(km
)
b)
-0.04 -0.02 0.00 0.02 0.04Mass Flux (kg m-2 s-1)
0
5
10
15
Altit
ude
(km
)
c)
convective
stratiform
non-precipitating
isentropic
Eulerian
as expected the isentropic mass fluxes are smaller
the same amount of mass flux is removed from upward and downward components
stratiform updrafts are greatly reduced
non-precipitating region dominated by the large scale subsidence
isentropic vs. eulerian mass flux difference
Downward convective
23.2 23.4 23.6 23.8 24.0 24.2 24.4Time (h)
5
10
15
Altit
ude
(km
)
-0.10
-0.08
-0.07
-0.05
-0.04
-0.02
-0.01a) Upward convective
23.2 23.4 23.6 23.8 24.0 24.2 24.4Time (h)
5
10
15
Altit
ude
(km
)
0.01
0.02
0.04
0.05
0.07
0.08
0.10b)
Downward stratiform
23.2 23.4 23.6 23.8 24.0 24.2 24.4Time (h)
5
10
15
Altit
ude
(km
)
-0.10
-0.08
-0.07
-0.05
-0.04
-0.02
-0.01c) Upward stratiform
23.2 23.4 23.6 23.8 24.0 24.2 24.4Time (h)
5
10
15
Altit
ude
(km
)
0.01
0.02
0.04
0.05
0.07
0.08
0.10d)
Downward clear sky
23.2 23.4 23.6 23.8 24.0 24.2 24.4Time (h)
5
10
15
Altit
ude
(km
)
-0.10
-0.08
-0.07
-0.05
-0.04
-0.02
-0.01e) Upward clear sky
23.2 23.4 23.6 23.8 24.0 24.2 24.4Time (h)
5
10
15
Altit
ude
(km
)
0.01
0.02
0.04
0.05
0.07
0.08
0.10f)
Eulerian - isentropic time series
Time (day)
Time (day)
in convective region oscillations co-located with the main convective burst with the maxima near the surface, near the tropopause and at about 5km where the updrafts are initiated and where the forcing is strong.
in the stratiform region there are two major regions of wave activity: in the boundary layer and near the tropopause.
convective region downdraft to updraft mass flux ratio
0.00 0.02 0.04 0.06 0.08 0.10 0.12Mass Flux Up (kg m-2 s-1)
0.00
0.02
0.04
0.06
0.08
0.10
0.12
Mas
s Fl
ux D
own
(kg
m-2 s
-1)
(a=0.44, b=-0.00) (a=0.62, b=-0.00) Eulerian
Isentropic
|MD|MU = a
How much convective downdraft is there?
other parametersConvective
330 340 350 360 370Oe (K)
5
10
15
Altit
ude
(km
)
Stratiform
330 340 350 360 370Oe (K)
5
10
15
Altit
ude
(km
)
-2.00 -1.80
-1.60 -1.40
-1.20 -1.00
-0.80 -0.60
-0.40 -0.20
0.00 3.40
6.80 10.20
13.60 17.00
20.40 23.80
27.20 30.60
34.00Vertical Velocity (m s -1)
Convective
330 340 350 360 370O
e (K)
5
10
15
Altit
ude
(km
)Stratiform
330 340 350 360 370O
e (K)
5
10
15
Altit
ude
(km
)0 2 4 6 8 10 12 14 16 18 20 22
Total Water mixing ratio (g kg-1)
log10
330 340 350 360 370O
e (K)
5
10
15
Altit
ude
(km
)
-5.00
-4.00
-3.00
-2.00
-1.00
0.00
log10
330 340 350 360 370O
e (K)
5
10
15
Altit
ude
(km
)
-5.00
-4.00
-3.00
-2.00
-1.00
0.00
convective
stratiform
Potential Temperature [K] at the surface
−200 −150 −100 −50 0 50 100 150 200
−200
−150
−100
−50
0
50
100
150
200
298
300
302
304
306
308potential temperature at the surface
other applications: hurricane simulations
water vapor
mass flux
vertical velocity
Summary
• We introduce an isentropic analysis of convective systems by conditionally averaging the properties of the flow along the equivalent potential temperature lines
• We can study properties of thermodynamically similar air parcels and separate between warm, moist updrafts and cool drier downdrafts - which are fundamental aspects if moist convection
• We define isentropic streamfunction which depicts the convective overturning
• We determine the mean values of raising and subsiding parcels (vertical velocity, buoyancy, humidity and hydrometeors mixing ratios)
• We find that after the removal of reversible oscillations the ratio of downward to upward mass flux decreased from 0.6 to 0.4 - in this model, under monsoon conditions
• Isentropic analysis of convective motions can be used as a basis for comparison between cloud resolving models and cloud resolving model evaluation tool