Aerosol, Interhemispheric Gradient, and Climate Sensitivity Ching-Yee Chang Department of Geography University of California Berkeley Lawrence Livermore.

Aerosol, Interhemispheric Gradient, and Climate Sensitivity

Ching-Yee ChangDepartment of Geography

University of California Berkeley

Lawrence Livermore National Lab Seminar April 27, 2011

Collaborators: John Chiang (UC Berkeley)

Michael Wehner (Lawrence Berkeley Lab)

Sulfate Aerosols and Climate

Difference in SAT caused by sulfate aerosol indirect effect (Rotstayn & Lohmann 2002)

Direct forcing from anthropogenic sulfate forcing (Kiehl & Briegleb 1993)

IPCC AR4

Large uncertainty in the aerosol forcing

Kiehl 2007

IPCC AR4

Outline

• Sulfate aerosol control of Tropical Atlantic climate over the 20th century (Chang et al. 2011, in press for Journal of Climate)

• Projection of the Interhemispheric Gradient in 21st century

• Interhemispheric Gradient and Transient Climate Response

Sulfate aerosol control of Tropical Atlantic climate over the 20th century

Atlantic interhemispheric SST gradient over the 20th century

Interhemispheric SST index = South – North box

21-yr running mean on annual mean data

Hadley SST, ERSST, Kaplan SST

HADISSTERSST KAPLAN SST

Ensemble Empirical Mode Decomposition (EEMD)

Modes from Ensemble Empirical Mode Decomposition (EEMD) analysis:

Multidecadal (mode 4) and trend (mode 5)

HADISSTERSST KAPLAN SST

The interhemispheric SST gradient and the meridional position of ITCZ

Mode 1 of an MCA of SSTA and 10m winds, and regression of the SSTA mode 1 time series on precipitation (Chiang and Vimont, 2004)

Equatorial meridional winds

ICOADS Winds

Smith et al. 2010: Reconstructed global precip

CRU TS 2.1 land precip.

south-north

June-July-August precipitation

CMIP3 models simulation of Atlantic Interhemispheric Gradient in 20th century climate experiment

Variance explained by this EOF: 49%

• Most of the projection coefficient are positive (Most models have a upward trend in the SSTA

gradient indices)

Projection of EOF1 onto each run

1st EOF of AITG indices from 71 model runs

South Hemisphere warming more than North HemisphereThis trend mitigates after 1980

T-test value=4.41 p-value = 0.00001 (assuming 71 d.o.f. for the 20th

century runs, and 44 for the preindustrial)

Mean of 1900-1982 trend in SSTA gradient significantly different from preindustrial run

Unit: 0.1K/100yr

Regression of SSTA onto model-index EOF1

Hadley SSTA

Stronger warming in the South Atlantic

Model ensemble averaged SSTA

Southward shift of ITCZ

Regression of Precip. Anomaly onto model-index EOF1

Model ensemble averaged Precip. anomalyCRU Precip. anomaly

Attribution the cause of the trend

Trend behavior appears in model ensemble mean

Most likely to be externally forced Single forcing runs

Results from single forcing runs

CCSM3 (2 members)PCM1 (4 members)GISS modelE (1 member)s)

SST gradient index

ITCZ index

most resembles the 1st EOF of the indices of the 20C expt.

CCSM3 sulfate aerosol emission forcing data

CCSM3 simulated sulfate aerosol optical depth

EOF1 from different subsets of models

EOF1 from AIE models capture the turn of the trend better AIE models simulate the AITG trend closer to observation

AIE models: Models with both Aerosol Direct and Indirect Effect

No-AIE models: Models with only aerosol direct, but no Indirect Effect

X-axis unit: 0.1K/100yr

Modeled SSTA and Precip.A regression on model-index EOF1

• Warming asymmetry and ITCZ southward shift stronger in AIE models

Models with Aerosol Indirect Effect

Models without Aerosol Indirect Effect

Summary I Interhemispheric gradient of Atlantic SST

found to have an upward trend before 1980, indicating stronger warming in the South Atlantic and southward shift of ITCZ

A similar positive trend is detected in the IPCC models.

This trend is likely due the north-south disparity in anthropogenic sulfate aerosol emissions

• 3 different scenarios are examined • A1B, A2, B1• Both Atlantic and Pacific sectors

Projection of the Interhemispheric Gradient in 21st century

Indices are defined as south box minus north box

Global mean of various anthropogenic forcing agents in future scenarios

IPCC AR4 WG1, Fig.10.26

95-year (2004~2098) trend statisticsPacific Atlantic

A1B

A2

B1

Most models project downward trend of the Pacific index in the 21th century in these 3 future scenarios => North Pacific warming stronger than South

Less conclusive results on the projection of the Atlantic index trend

X-axis unit: 0.1K/100yr

A1BAtmos. Sulfate burdenunit: 10e-6 kg/m2

Pacific

Atlantic

High-lat index (35~60)Tropical index (5~35)

Stronger change in the interhemispheric gradient of sulfate aerosol forcing across the equator in the Pacific sector

(From miub_echo model)

It’s projected that most of the decrease of sulfate aerosol mainly comes from Asia

More aerosol emission from Tropical Atlantic than from North Atlantic

Atmospheric Sulfate burden

A1b A2 B1

All three scenarios have stronger change in sulfate aerosol forcing across the equator in the Pacific sector

1%/yr to 2xCO2 experiment

• However, similar change of Pacific gradient is found in 1%/yr to double CO2 experiment, but with weaker magnitudes

Comparison: 1%/yr to 2xCO2 and A1B experiments

-0.8 -0.6 -0.4

-0.199999999999999 0 0.2 0.402468

1012

1%/yr to 2xCO2

Tropical Pacific SAT Gradient change (C )

Freq

uenc

y

mean = -0.15 C

-0.8 -0.6 -0.4

-0.199999999999999 0 0.2

0.3999999999999980

4

8

12

A1B

Tropical Pacific SAT Gradient change (C )Fr

eque

ncy

mean = -0.26 C

(Yr60~Yr80) – (Yr1~Yr20) (2079~2098) – (2005~2024)

Most models project negative trend in the Pacific interhemispheric gradient – the rate of the warming in the north Pacific speeds up at the end of 21st century

Possibly related to the decrease of the aerosols in the north Pacific in the future, but GHG forcing or other factors may also contribute

Summary II

Interhemispheric Gradient and Climate sensitivity

Trend Statistic from different subsets of models

AIE models simulate the AITG trend closer to observation

AIE models: Models with both Aerosol Direct and Indirect Effect

No-AIE models: Models with only aerosol direct, but no Indirect Effect

X-axis unit: 0.1K/100yr

Kiehl 2007:• Total forcing inversely correlated

to climate sensitivity• Large uncertainty in the aerosol

forcing

Equilibrium Climate sensitivity and Transient Climate Response

Equilibrium Climate Sensitivity

IPCC AR4 Table 8.2

Transient Climate Response

300 ppm

600 ppmCO2 concentration

+ Slab ocean

AGCM

T0

T’

CO2 concentration

AGCM + OGCM300 ppm

600 ppm

T0

T’

Climate sensitivity and Transient Climate Response

IPCC AR4 Table 8.2

1.5 2 2.5 3 3.5 4 4.5 50

0.5

1

1.5

2

2.5

3

Equilibrium Climate Sensitivity (C )Tr

ansi

ent C

limat

e Re

spon

se (C

)

Atlantic SSTA Grad. Trend v.s. Climate Sensitivity

• There seems to be a linear relationship between the gradient trend and the Transient Climate Response (TCR) among most of the models

1.5 2 2.5 3 3.5 4 4.5 5-1

0

1

2

3

4

5

Equilibrium Climate Sensitivity (C)

Atl

antic

SST

A g

rad.

Tre

nd

(0.1

C/10

0yr)

1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8-1

0

1

2

3

4

5

Transient Climate Response (C)

Atl

antic

SST

A G

rad.

Tre

nd

(0.1

C/10

0yea

r)

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10

0.5

1

1.5

2

2.5

3

3.5

4giss_model_e_hgiss_model_e_ripsl_cm4miroc3_2_hiresmiroc3_2_medresmpi_echam5ukmo_hadcm3ukmo_hadgem1miub_echo_gcccma_cgcm3_1 cnrm_cm3csiro_mk3_0 gfdl_cm2_0 gfdl_cm2_1 inmcm3_0mri_cgcm2_3_2a ncar_ccsm3_0 ncar_pcm1 Obs

Transient Climate Response (C )

Trop

. Atla

ntic

SST

Grad

ient

Tre

nd (0

.1C/

100y

r)

Regional Transient Climate response

Regional Transient Climate Responses (TCR) in the Tropical Atlantic regions are similar in the North and South

0 0.5 1 1.5 2 2.50

0.5

1

1.5

2

2.51%/yr to 2xCO2

North Trop. Atlantic Regional TCR (C )

Sout

h Tr

op. A

tlanti

c Re

gion

al T

CR (C

)

1 1.5 2 2.5 30

0.5

1

1.5

2

2.51%/yr to 2xCO2

N.Trop.Atl.SAT.change

S.Trop.Atl.SAT.change

Transient Climate Response (C )

Regi

onal

TCR

(C )

Roughly a linear relationship between regional TCR and global TCR

A linear relationship between TCR and Interhemispheric Gradient Trend

0 0.5 1 1.5 2 2.5 30

0.51

1.52

2.53

3.54

4.55

f(x) = 2.823566 x − 2.46865046666666R² = 0.848919004915189

20C expt.

Transient Climate Response (C )

Atl

antic

SST

A G

rad.

Tre

nd

(0.1

C/10

0yea

r)

•Similar regional TCRs, in the Tropical Atlantic region across the equator

•Roughly a linear relationship between regional TCR and global TCR

•Models with higher TCR are models

with stronger aerosol forcing, due to the constraint of the 20C global mean SAT change

Stronger aerosol forcing with larger TCR => stronger SST gradient

If we also constrain the models with observed interhemispheric gradient change?

Summary III• A linear relationship between the modeled Atlantic

SST Interhemispheric Gradient and Transient Climate Response for most of the models

• This relationship can be explained by the uncertainty of the aerosol forcings among the models

• Further confirms that the important role of aerosol on the change of the Interhemispheric Gradient

• Constraint on the simulation of Interhemispheric Gradient change (or trend) may be a way to confine the uncertainty of models’ climate sensitivity

Thank you for your attention

Comparison of SAT grad. change for 20C and 1%to2xCO2

1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8

-0.1

-5.55111512312578E-17

0.1

0.2

0.3

0.4

0.520C experiment

Transient Climate Response (C )

SAT G

rad.

Cha

nge

(C )

1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8

-0.5-0.4-0.3-0.2-0.1

00.10.20.30.40.5

1% to 2xCO2

Transient Climate Response (C )SA

T G

rad.

cha

nge

(C )

Comparison of SAT grad. change for 20C and 1%to2xCO2

1 1.5 2 2.5 3

-0.5

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

1%to2xCO220c

TCR (C )

SAT

Gra

d. C

hang

e (C

)

Climate sensitivity and total anthropogenic forcing

}{

}{

222

111

aerosolGHG

aerosolGHG

FFT

FFT

FT

0;0 aerosoli

GHGi FFIn general,

20 Century temperature change = climate sensitivity × Radiative Forcing

aerosolaerosol FF 21

}{

}{

22

11

aerosolGHG

aerosolGHG

FFT

FFT

}{

}{

22

11

aerosolGHG

aerosolGHG

FFT

FFT

Smaller total anthropogenic forcing, larger climate sensitivity

Larger total anthropogenic forcing,smaller climate sensitivity

Let

Consistent with Kiehl 2007

Regional TCR and interhemispheric gradient}{ aerosol

iGHG

i FFT

South – North Gradient change, G△

}{ Sii

Niiii SulbSulbG

Gj

Sj

SGi

Si

Sul

SulW

Sul

Sul

aerosoliSNi

GiiSNii

FWW

SulbWWG

)(

)(

TWWFWWG SNGHG

SNii )()(Linear relationship btw.TCR and interhemispheric gradient change

})()({ Ni

Siii AeroFAeroFG

Si

Si

GHGSi

Si

Si

Si

si AeroFTAeroFGHGFT )(})()({ ,

Ni

Ni

GHGNi

Ni

Ni

Ni

Ni AeroFTAeroFGHGFT )(})()({ ,

-0.8 -0.6 -0.4

-0.199999999999999 0 0.2 0.402468

A1B

Tropical Pacific SAT Gradient change (C )Fr

eque

ncy

mean = -0.3

Model internal averaged Gradient change

-0.8 -0.6 -0.4

-0.199999999999999 0 0.2 0.402468

1% to 2xCO2

Tropical Pacific Gradient change (C )

Freq

uenc

y

mean = -0.15

-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.50

1

2

3

4

5

6

1%/yr to 2xCO2

Trop. Atlantic SAT change

Freq

uenc

y

Pacific and AtlanticInterhemispheric Gradient

HADISSTERSST(NOAA)KAPLAN SST

Pacific

Atlantic

20C experimentPacific Atlantic

49%51%

EOF1 from all models

Projection of EOF1 on each run

A1B

B1

A2

Atlantic Interhemispheric SST Gradient

Pacific Interhemispheric SST Gradient

A1B

B1

A2