Embed Size (px)
Citation preview
SW CRE
LW CRE
Net CRE
Is the Net CRE Constrained to be Uniform over the
Tropical Warm Pools?
Casey Wall* Dennis Hartmann
Joel Norris
*Scripps Institution of Oceanography
SW CRE
LW CRE
Net CRE
CRE (Wm-2)
-80 -40 0 40 80
Cloud Radiative Effects — JJA Mean
data: CERES
Why is net CRE about the same in
convective and non-convective
regions?
If the “cloud shading feedback” is strong, then R1 ≈ R2.
Hypothesis (Hartmann et al., 2001)
Warm Cold
R1 R2
top-of-atmospherenet radiation
~1500 km
Climate Model Experiments
Model Configuration• Community Atmosphere Model v. 4 (CAM4) • global RCE • non-rotating aquaplanet with uniform insolation • 5 m slab ocean • daily output
Is this model useful for testing the hypothesis?• conserves energy • resolves motions with scale ~1500 km • simulates cloud-shading feedback
Climate Model Experiments
Experimental Design
Control(cloud shading feedback on)
save SW↓ at ocean surface for every time step
randomly permute SW↓ in time dimension
Random SW(cloud shading feedback off)
model sees randomized SW↓ at ocean surface
If the hypothesis is correct, then we will see:Control
(feedback on)Random SW(feedback off)
SST gradients small largeenergy transport small large
CRE contrast small large
Hypothesis Testing
SST
Anom
aly
(°C
)
7
5
3
1
-1
-3
-5
-7
SST GradientsControl
(feedback on)
Random SW(feedback off)
ascent
subsidenceControl
SST Anomaly (°C)
-8 -6 -4 -2 0 2 4 6 8
-8 -6 -4 -2 0 2 4 6 8
(°C
-1)
0
0.05
0.1
0.15
0.2
0
0.05
0.1
0.15
0.2
(°C
-1)
Random SW
SST Gradients
(feedback on)
(feedback off)
ascent
subsidenceControl
SST Anomaly (°C)
-8 -6 -4 -2 0 2 4 6 8
-8 -6 -4 -2 0 2 4 6 8
(°C
-1)
0
0.05
0.1
0.15
0.2
0
0.05
0.1
0.15
0.2
(°C
-1)
Random SW
SST Gradients
(feedback on)
(feedback off)
If the hypothesis is correct, then we will see:
Control Random SWSST gradients small large
energy transport small largeCRE contrast small large
energytendency
surfaceheatflux
columnradiativeheating
atmosphericenergy
transport
Horizontal Energy Transport
@H
@t= SH + LH +RLW +RSW �D
h = cpT + Lq + gz
H =
Z pt
p0
h
⇣� dp
g
⌘
moist static energy
column moiststatic energy
column moiststatic energy
budget
Net Energy Export from Ascent Regions (Wm-2)10080604020
0
0.02
0.04
0.06
((Wm
-2)-1
)
Horizontal Energy Transport
0.08
ControlRandom
SW
Net Energy Export from Ascent Regions (Wm-2)10080604020
0
0.02
0.04
0.06
((Wm
-2)-1
)
Horizontal Energy Transport
0.08 Uniform SST
ControlRandom
SW
Net Energy Export from Ascent Regions (Wm-2)10080604020
0
0.02
0.04
0.06
((Wm
-2)-1
)
Horizontal Energy Transport
0.08 Uniform SST
ControlRandom
SW
If the hypothesis is correct, then we will see:
Control Random SWSST gradients small large
energy transport small largeCRE contrast small large
Cloud Radiative Effects
Δ = 8
Δ = 29
-106
SW CRE(Wm-2)
LW CRE(Wm-2)
Net CRE(Wm-2)
-151
71
88
-35
-63
-43
-48
Control
Random SW
Regions ofUpward Motion
Regions ofDownward Motion
SW CRE(Wm-2)
LW CRE(Wm-2)
Net CRE(Wm-2)
16 -27
-3414
(feedback on)
(feedback off)
Cloud Radiative Effects
Δ = 8
Δ = 29
-106
SW CRE(Wm-2)
LW CRE(Wm-2)
Net CRE(Wm-2)
-151
71
88
-35
-63
-43
-48
Control
Random SW
Regions ofUpward Motion
Regions ofDownward Motion
SW CRE(Wm-2)
LW CRE(Wm-2)
Net CRE(Wm-2)
16 -27
-3414
(feedback on)
(feedback off)
If the hypothesis is correct, then we will see:Control Random SW
SST gradients small largeenergy transport small large
CRE contrast small large
Conclusion: Cloud shading feedback causes net CRE to be more uniform.
Hypothesis
Summary
Why is net CRE nearly uniform over the tropical warm pools?Question
Findings Modeling support for hypothesis
OpenQuestions
• Is this theory a complete explanation? • Implications for cloud-climate feedbacks?
Cloud Albedo
Sea Surface Temperature Circulation
This is a key mechanism for obtaining uniform net CRE over the warm pools [Hartmann et al., 2001].
Hypothesis
Summary
Why is net CRE nearly uniform over the tropical warm pools?Question
Findings Modeling support for hypothesis
Cloud Albedo
Sea Surface Temperature Circulation
This is a key mechanism for obtaining uniform net CRE over the warm pools [Hartmann et al., 2001].
ReferenceWall, Hartmann, & Norris (in press) Is the Net Cloud Radiative Effect Constrained to be Uniform over the Tropical Warm Pools? Geophysical Research Letters
Extra Slides
SW CREsubsidence (Wm-2) LW CREsubsidence (Wm-2) Net CREsubsidence (Wm-2)
SW C
REas
cent
(Wm
-2)
LW C
REas
cent
(Wm
-2)
Net
CRE
asce
nt (W
m-2
)
-130 -10-110 -90 -70 -30
-10
-50
-70
-90
-110
-130 0
20
40
60
80
100
120
0 20 40 60 80 100 120 -80 -60 -40 -20 0 20 40-80
-60
-40
-20
0
20
40
IPSL-CM5A-LR
MIROC5
MRI-CGCM3
-30
-50
CMIP5 — SW CRE CMIP5 — LW CRE CMIP5 — Net CRE
A Note on Model Tuning: CRE in CMIP5 Models
average over days with upward motion at 500 hPa
data: CMIP5 fully-coupled
historical runs (1979-2005)
average over days with downward
motion at 500 hPa
-30 -15 0 15 30-60
-50
-40
-30
-20
-10
0
10
20
30
40
Hea
t Flu
x An
omal
y (W
m-2
)
Control
LHSHSWLW
-30 -15 0 15 30Time (days)
-1
0
1
2
3
SST
anom
aly
(C)
-30 -15 0 15 30-60
-50
-40
-30
-20
-10
0
10
20
30
40
Hea
t Flu
x An
omal
y (W
m-2
)
Random SW
-30 -15 0 15 30Time (days)
-1
0
1
2
3
SST
anom
aly
(C)
-30 -15 0 15 30-60
-50
-40
-30
-20
-10
0
10
20
30
40
Hea
t Flu
x An
omal
y (W
m-2
)
Constant Surface SW
-30 -15 0 15 30Time (days)
-1
0
1
2
3
SST
anom
aly
(C)
-30 -15 0 15 30-60
-50
-40
-30
-20
-10
0
10
20
30
40
Hea
t Flu
x An
omal
y (W
m-2
)
Control
LHSHSWLW
-30 -15 0 15 30Time (days)
-1
0
1
2
3
SST
anom
aly
(C)
-30 -15 0 15 30-60
-50
-40
-30
-20
-10
0
10
20
30
40
Hea
t Flu
x An
omal
y (W
m-2
)
Random SW
-30 -15 0 15 30Time (days)
-1
0
1
2
3
SST
anom
aly
(C)
-30 -15 0 15 30-60
-50
-40
-30
-20
-10
0
10
20
30
40
Hea
t Flu
x An
omal
y (W
m-2
)
Constant Surface SW
-30 -15 0 15 30Time (days)
-1
0
1
2
3
SST
anom
aly
(C)
-30 -15 0 15 30-60
-50
-40
-30
-20
-10
0
10
20
30
40
Hea
t Flu
x An
omal
y (W
m-2
)
Control
LHSHSWLW
-30 -15 0 15 30Time (days)
-1
0
1
2
3
SST
anom
aly
(C)
-30 -15 0 15 30-60
-50
-40
-30
-20
-10
0
10
20
30
40
Hea
t Flu
x An
omal
y (W
m-2
)
Random SW
-30 -15 0 15 30Time (days)
-1
0
1
2
3
SST
anom
aly
(C)
-30 -15 0 15 30-60
-50
-40
-30
-20
-10
0
10
20
30
40
Hea
t Flu
x An
omal
y (W
m-2
)
Constant Surface SW
-30 -15 0 15 30Time (days)
-1
0
1
2
3
SST
anom
aly
(C)
-30 -15 0 15 30-60
-50
-40
-30
-20
-10
0
10
20
30
40
Hea
t Flu
x An
omal
y (W
m-2
)
Control
LHSHSWLW
-30 -15 0 15 30Time (days)
-1
0
1
2
3
SST
anom
aly
(C)
-30 -15 0 15 30-60
-50
-40
-30
-20
-10
0
10
20
30
40
Hea
t Flu
x An
omal
y (W
m-2
)
Random SW
-30 -15 0 15 30Time (days)
-1
0
1
2
3
SST
anom
aly
(C)
-30 -15 0 15 30-60
-50
-40
-30
-20
-10
0
10
20
30
40
Hea
t Flu
x An
omal
y (W
m-2
)
Constant Surface SW
-30 -15 0 15 30Time (days)
-1
0
1
2
3
SST
anom
aly
(C)
-30 -15 0 15 30-60
-50
-40
-30
-20
-10
0
10
20
30
40
Hea
t Flu
x An
omal
y (W
m-2
)
Control
LHSHSWLW
-30 -15 0 15 30Time (days)
-1
0
1
2
3
SST
anom
aly
(C)
-30 -15 0 15 30-60
-50
-40
-30
-20
-10
0
10
20
30
40
Hea
t Flu
x An
omal
y (W
m-2
)
Random SW
-30 -15 0 15 30Time (days)
-1
0
1
2
3
SST
anom
aly
(C)
-30 -15 0 15 30-60
-50
-40
-30
-20
-10
0
10
20
30
40
Hea
t Flu
x An
omal
y (W
m-2
)
Constant Surface SW
-30 -15 0 15 30Time (days)
-1
0
1
2
3
SST
anom
aly
(C)
-30 -15 0 15 30-60
-50
-40
-30
-20
-10
0
10
20
30
40
Hea
t Flu
x An
omal
y (W
m-2
)
Control
LHSHSWLW
-30 -15 0 15 30Time (days)
-1
0
1
2
3
SST
anom
aly
(C)
-30 -15 0 15 30-60
-50
-40
-30
-20
-10
0
10
20
30
40
Hea
t Flu
x An
omal
y (W
m-2
)
Random SW
-30 -15 0 15 30Time (days)
-1
0
1
2
3
SST
anom
aly
(C)
-30 -15 0 15 30-60
-50
-40
-30
-20
-10
0
10
20
30
40
Hea
t Flu
x An
omal
y (W
m-2
)
Constant Surface SW
-30 -15 0 15 30Time (days)
-1
0
1
2
3
SST
anom
aly
(C)
SWnet
latentheat LWnet
sensibleheat
-30 -15 0 15 30 -30 -15 0 15 30 -30 -15 0 15 30
-30 -15 0 15 30-30 -15 0 15 30-30 -15 0 15 30-1
0
1
2
3
-1
0
1
2
3
-1
0
1
2
3
-60
-40
-20
0
20
40
-60
-40
-20
0
20
40
-60
-40
-20
0
20
40Control Random SW Uniform SW
Time Lag (Days) Time Lag (Days) Time Lag (Days)
SST
Anom
aly
(°C)
Surfa
ce H
eat F
lux
Anom
aly
(Wm
-2)
a b c
d e f
Surface Heat Budget
-100 -75 -50 -25 0EOF1 - omega (hPa/day)/(stdev PC1)
0
5
10
15H
eigh
t (km
)controlrandom SW
-50 -25 0 25 50EOF2 - omega (hPa/day)/(stdev PC2)
0
5
10
15
Hei
ght (
km)
-100 -75 -50 -25 0EOF1 - omega (hPa/day)/(stdev PC1)
0
5
10
15H
eigh
t (km
)controlrandom SW
-50 -25 0 25 50EOF2 - omega (hPa/day)/(stdev PC2)
0
5
10
15
Hei
ght (
km)
Hei
ght (
km)
𝜔 EOF1 (hPa day-1 𝜎1-1) 𝜔 EOF2 (hPa day-1 𝜎2-1)
RandomSW
Control
-100 -75 -50 -25 0 -50 -25 0 25 500
5
10
15
0
5
10
15ExplainedVariance:
68%85%
ExplainedVariance:
23%9%
Modes of Vertical Motion in Tropic World
0 300 600 900 1200 1500 180014
30
18
22
26
Time (days)
SST
(°C)
Control
Random SW
Global-mean SST in Tropic World Simulations
SW CREsubsidence (Wm-2) LW CREsubsidence (Wm-2) Net CREsubsidence (Wm-2)
SW C
REas
cent
(Wm
-2)
LW C
REas
cent
(Wm
-2)
Net
CRE
asce
nt (W
m-2
)
-130 -10-110 -90 -70 -30
-10
-50
-70
-90
-110
-130 0
20
40
60
80
100
120
0 20 40 60 80 100 120 -80 -60 -40 -20 0 20 40-80
-60
-40
-20
0
20
40
IPSL-CM5A-LR
MIROC5
MRI-CGCM3
-30
-50
CMIP5 — SW CRE CMIP5 — LW CRE CMIP5 — Net CRE
A Note on Model Tuning: CRE in CMIP5 Models
average over days with upward motion at 500 hPa
data: CMIP5 fully-coupled
historical runs (1979-2005)
average over days with downward
motion at 500 hPa
Climate Model Experiments
Model Details• CAM4 global RCE • RCEMIP standard protocol • aquaplanet
• 5 m slab ocean • daily output
-15 -10 -5 0 5 10 15-120
-80
-40
0
wba
r (hP
a/da
y)
140
160
180
200
220
240
SW d
own
surfa
ce (W
m-2
)
-15 -10 -5 0 5 10 15Time (days)
-0.4
-0.2
0
0.2
0.4
SST
anom
aly
(C)
-15 -10 -5 0 5 10 15-120
-80
-40
0
wba
r (hP
a/da
y)
140
160
180
200
220
240
SW d
own
surfa
ce (W
m-2
)
-15 -10 -5 0 5 10 15Time (days)
-0.4
-0.2
0
0.2
0.4
SST
anom
aly
(C)
-15 -10 -5 0 5 10 15
-15 -10 -5 0 5 10 15
(hPa
day
-1)
ω̄
0
-40
-80
-120
Surfa
ce S
W↓
(Wm
-2)
240220200180160140
ω̄Surface
SW↓
-0.4
-0.2
0
0.2
0.4
SSTAnomaly
SST
Anom
aly
(°C)
Time Lag (days)
Cloud Shading Feedback in CAM4
ΔSST ≈ 0.6 °Ccomparable to observations
[Wall et al. 2018]
Other
Upper-Ocean Temperature in the West Pacific Warm Pool
Diurnal warming events of western equatorial Pacific warm pool 1067
Diurnal heating in the COARE domain at different wind speed conditions I , I I I I
6-
E8 5 g-10- n
12-
14-
16-
18- A
A: 9 profiles during 7-m/s wind speed (shifted -0.5%)
B: 10 profiles during 2.5-m/s wind speed
C: 5 profiles during calm weather
I3 C
20’ I I I I I I 28 29 30 31 32 33 34
Temperature (“C) Fig. 5. Vertical temperature profiles in near-surface layer of the ocean in the TOGA COARE
domain obtained by free-rising profiler at different wind speeds in the afternoon.
In Fig. lOa, the surface temperature calculated using the transilient model is compared with the bucket-thermometer and dry air temperature data. The SST calculated using the cool-skin model, including the solar radiation absorption in the upper millimeters of the ocean (Soloviev and Schluessel, 1996) is also shown. According to Fig. lOa, the daily variation in temperature in the atmosphere is even larger than it is in the ocean. Hoeber (1969) previously observed a similar effect in the equatorial region. It occurs because the moist equatorial atmosphere absorbs solar radiation directly during the daytime and is again cooled during the nighttime (see discussion in Kraus and Businger, 1994, p. 170).
Also remarkable is the diurnal variation of the relative humidity (Fig. lob). It decreases with the diurnal SST increase, thus enhancing the latent heat flux. This is a manifestation of the negative feedback mechanism stabilizing the SST (Greenhut, 1978; Gautier, 1978; Lukas, 1990a; Kraus and Businger, 1994).
4.4. Salinity effects
Salinity effects may substantially modify diurnal heating of the near-surface layer in the western equatorial Pacific warm pool. Freshwater influx due to precipitation influences the diurnal cycle by trapping heat near the surface. This has been demonstrated recently by modeling of the diurnal cycle with and without rain effects and comparison with TOGA COARE mooring time series data (Anderson et al., 1996). Note the salinity depression on 3
strongwind
med.wind
lightwind
-2 0 2 4
-4
-2
0
2
SW c
loud
feed
back
(W m
-2 K
-1)
-2 0 2 4LW cloud feedback (W m -2 K-1 )
-4
-2
0
2
-2 0 2 4
-4
-2
0
2
LW Cloud-Climate Feedback (Wm-2K-1)
SW C
loud
-Clim
ate
Feed
back
(Wm
-2K-
1 ) Free-Tropospheric Clouds Boundary-Layer Clouds
perfect cancellation
-2 0 2 4
-4
-2
0
2
SW c
loud
feed
back
(W m
-2 K
-1)
-2 0 2 4LW cloud feedback (W m -2 K-1 )
-4
-2
0
2
-2 0 2 4
-4
-2
0
2
-2 0 2 4
-4
-2
0
2
-4
-2
0
2
-2 0 2 4
Future Research: Implications for Climate Change Cloud Feedbacks over the Tropical Western Pacific from CMIP5 Models
Free-Tropospheric Clouds
Clo
ud F
eedb
ack
(Wm
-2K-
1 )
Total Amount Altitude OpticalDepth
Residual
-4
-2
0
2
4 LW Net SW
CMIP5 Cloud Feedbacks over the West Pacific Warm Pool
10 20 30 40 50 60 70 80LW CRE (Wm-2)
-100
-90
-80
-70
-60
-50
-40
-30
SW C
RE
(Wm
-2)
10 20 30 40 50 60 70 80-100
-90
-80
-70
-60
-50
-40
-30
LW CRE (Wm-2)
SW C
RE
(Wm
-2)
CMIP5CERES
net CRE =0 Wm -2
-20-40
Climatology of CRE over the West Pacific Warm Pool in CMIP5 Fully-Coupled Simulations
The Coincidence Hypothesis (Kiehl, 1994)
• The thick part of anvils determines the average CRE. • These clouds are very thick, so their albedo and
emissivity are fixed. • Therefore
1. SW CRE is determined only by insolation 2. LW CRE is determined only by the tropopause
temperature 3. Uniform net CRE is the result of a fortuitous
coincidence• Weak hypothesis: Life cycle of anvil clouds produces uniform net CRE
[Hartmann et al., 2018].
Hypothesis
Outline
Why is net CRE about the same in convective and non-convective regions over the tropical warm pools?
Question
Cloud Albedo
Sea Surface Temperature Circulation
This is a key mechanism for producing uniform net CRE over the warm pools [Hartmann et al., 2001].
ObjectiveTest hypothesis with idealized GCM experiments
Cloud shading feedback reduces SST gradient, which reduces lateral heat transport,
which reduces the contrast in net radiation.
If the feedback is strong, then R1 ≈ R2.
Warm Cold
R1 R2
top-of-atmospherenet radiation
Hypothesis (Hartmann et al., 2001)
Warm Cold
R1 R2
top-of-atmospherenet radiation
Hypothesis (Hartmann et al., 2001)
ObjectiveTest hypothesis with idealized GCM experiments
Climate Model Experiments
Model Details• CAM4 • global radiative-convective
equilibrium configuration • RCEMIP standard protocol
• aquaplanet • 5 m slab ocean • daily output
Is this model useful for testing the hypothesis?• conserves energy • resolves motions ~1500 km in scale • simulates cloud shading feedback;
SST fluctuations agree with observations [Wall et al., 2018]
Is the Net Cloud Radiative Effect Constrained to be Uniform over the Tropical Warm Pools?
Casey Wall*, Dennis Hartmann, Joel Norris*Scripps Institution of Oceanography
Climate Model Experiments
Model Configuration:CAM4 Global Radiative-Convective Equilibrium
• aquaplanet • no rotation • uniform insolation
• no aerosol effects • 5 m slab ocean • daily output
Experimental Design
Control(cloud shading feedback on)
save SW↓ at ocean surface for every time step
randomly permute SW↓ in time dimension
Random SW(cloud shading feedback off)
model sees randomized SW↓ at ocean surface
Hypothesis Testing
If the hypothesis is correct, then we will see:Control
(feedback on)Random SW(feedback off)
SST gradients small largeenergy transport small large
CRE contrast small large
If the hypothesis is incorrect, then we will see: Similar values in the two experiments
“cloud shading feedback”
Cloud Albedo
AtmosphericCirculation
Sea SurfaceTemperature
Hypothesis (Hartmann et al., 2001)
