Changes of the Hadley Circulation Under Global Warming and Their Linkage

with Climate Sensitivity and Hydrological Sensitivity

CFMIP Meeting, Tokyo, Japan, September 25-28, 2017

Hui Su1

Jonathan H. Jiang1, J. David Neelin2, Janice Shen1, Chengxing Zhai1, Qing Yue1, Zhien Wang3, Lei Huang4, Yong-Sang Choi5,

Graeme L. Stephens1, Yuk L. Yung6

1Jet Propulsion Laboratory, California Institute of Technology2University of California, Los Angeles

3University of Wyoming4Joint Institute for Regional Earth System Science and Engineering, UCLA

5Ewha Womans University6California Institute of Technology

Large Spread in Climate Sensitivity and Hydrological Sensitivity

dP/d

T s (%/K

)

0

0.5

1

1.5

2

2.5

3

3.5

CSIRO_acc

ess1

.0

UKMO_had

gem2-a

CCCMA_can

am4

MPI_esm

-lr

BCC_csm

1.1

MIROC_esm

MPI_esm

-mr

CSIRO_acc

ess1

.3

GFDL_cm3

IPSL_cm5b

-lr

CNRM_cm5

NCC_nores

m1-m

MIROC_miro

c5

INM_cm4

BCC_csm

1.1m

NCAR_ccs

m4

IPSL_cm5a

-lr

IPSL_cm5a

-mr

GISS_e2-r

MRI_cgcm

3

Temperature Mediated Precipitation Change per Unit Surface Warming

Temperature-mediated dP/dTs: 2%/K − 3%/K

Equilibrium Climate Sensitivity: 2 K −4.5K

2.6%/K

Changes of the Hadley Circulation, Clouds and Precipitation Under Global Warming

∆ = 2074-2098 in “RCP4.5” – 1980-2004 in “historical run”

Multi-model-mean from 15 CMIP5 coupled models

“Wet Get Wetter,Dry Get Drier”

(Su et al., JGR, 2014)

Quantifying the Model Differences in Circulation and Relation with Cloud Radiative Effect Changes

• Differences in the Hadley Circulation are highly correlated with the inter-model spread in net CRE.

The explained variance by the 1st

EOF is 57%

(Su et al., JGR, 2014)

Normalized Response

Smallest Circu. Change

Largest Circu. Change

Lower Tropospheric Mixing Index

• Stronger circulation change is associated with greater boundary layer drying

Largest Circu. Change

Smallest Circu. Change

Circulation Change and Lower Tropospheric Mixing

• Stronger circulation change is associated with stronger boundary layer drying

Circulation, Cloud Feedback and Climate Sensitivity

(Su et al., JGR, 2014)

Circulation, High Cloud and Precipitation Changes

Tightening of Hadley Ascent Links to High Cloud Cover Reduction

Interannual Centennial

(a) Interannual

-20 -10 0 10 20Latitude (o)

-20

-10

0

10

20

-20

-10

0

10

20

dCF/dTsdω/dTsω250 climdP/dTs

dCF/

dTs (

% K

−1);

dP/d

T s x4

(mm

day

−1K

−1)

dω/dT

s (hPa day−1K

−1); ω250 clim

atology (hPa day−1)

W WS

THATHA

(b) Centennial

-20 -10 0 10 20Latitude (o)

-4

-2

0

2

4

-4

-2

0

2

4

dCF/dTsdω/dTsω250 climdP/dTs

dCF/

dTs (

% K

−1);

dP/d

T s x4

(mm

day

−1K

−1)

dω/dT

s (hPa day−1K

−1); ω250 /10 clim

atology (hPa day−1)

W WS

THATHA

a BCC_csm1.1b BCC_csm1.1mc CCCMA_canam4d CNRM_cm5e CSIRO_access1.0f CSIRO_access1.3g CSIRO_mk3.6h GFDL_cm3i GFDL_esm2gj GISS_e2rk INM_cm4l IPSL_cm5a-lrm IPSL_cm5a-mrn IPSL_cm5b-lro MIROC_esmp MIROC_miroc5q MPI_esm-lrr MPI_esm-mrs MRI_cgcm3t NCAR_cam5u NCAR_ccsm4v NCC_noresm1-mw UKMO_hadgem2-a

(c) Interannual

-4 -3 -2 -1 0 1dFω/dTs (% K−1)

-2.5

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

dCF/

dTs(%

K−1

)

a

b

c

e

f

g

h

j

k

lm

n

o pqr

st u

v

w

correlation = 0.56

(d) Centennial

-1.0 -0.5 0.0 0.5dFω/dTs (% K−1)

-1.5

-1.0

-0.5

0.0

0.5

1.0

dCF/

dTs (

% K

−1)

ab

ce

f

g

h

ij

lm

n

o

pq

r

s

tu

v w

correlation = 0.65

dFω/dTs: the fractional change of tropical ascending area per unit surface warmingdCF/dTs: the tropical-mean high cloud fraction change per unit surface warming

Su et al. (2017, Nature Comm.)

High Cloud Shrinkage Leads to Enhanced Longwave Radiative Cooling

Interannual Centennial

a BCC_csm1.1b BCC_csm1.1mc CCCMA_canam4d CNRM_cm5e CSIRO_access1.0f CSIRO_access1.3g CSIRO_mk3.6h GFDL_cm3i GFDL_esm2gj GISS_e2rk INM_cm4l IPSL_cm5a-lrm IPSL_cm5a-mrn IPSL_cm5b-lro MIROC_esmp MIROC_miroc5q MPI_esm-lrr MPI_esm-mrs MRI_cgcm3t NCAR_cam5u NCAR_ccsm4v NCC_noresm1-mw UKMO_hadgem2-a

(a) Interannual

-2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.00.0dCF/dTs (% K−1)

0

1

2

3

4

5

dOLR

/dT s

(W m

−2K

−1)

a b

c

de

f

g

hj

k

lm

n

op q

rt

uvw

correlation = −0.77

T+AAquaTerraISCCP

CER

ES

(b) Centennial

-1.0 -0.5 0.0 0.5 1.0dCF/dTs (% K−1)

-2

-1

0

1

2

3

4

dOLR

/dT s

(W m

−2K

−1)

a

bc

e

fg h

ij

l

mn o

p

qr

s

t

uvw

correlation = −0.71

Su et al. (2017, Nature Comm.)

Observational Constraint on Hydrological Sensitivity

• Unconstrained hydrological sensitivity: 2.1%/K to 3.2%/K• Observation-constrained hydrological sensitivity: 2.6%/K to 2.9%/K

a BCC_csm1.1b BCC_csm1.1mc CCCMA_canam4d CNRM_cm5e CSIRO_access1.0f CSIRO_access1.3g CSIRO_mk3.6h GFDL_cm3i GFDL_esm2gj GISS_e2rk INM_cm4l IPSL_cm5a-lrm IPSL_cm5a-mrn IPSL_cm5b-lro MIROC_esmp MIROC_miroc5q MPI_esm-lrr MPI_esm-mrs MRI_cgcm3t NCAR_cam5u NCAR_ccsm4v NCC_noresm1-mw UKMO_hadgem2-a

(a)

0 2 4 6Interannual Tropical dOLR/dTs (W m−2K−1)

0

1

2

3

Inte

rann

ual G

loba

l Lvd

P/dT

s (W

m−2

K−1

)

a

b

c

d

e fg

h

jkl

m n

opq

r

s

t

u

vw

correlation = 0.68

0 2 4 60

1

2

3

(b)

0 1 2 3Interannual Global LvdP/dTs (W m−2K−1)

1.5

2.0

2.5

3.0

Tem

pera

ture

Med

iate

d G

loba

l Lvd

P/dT

s (W

m−2

K−1

)

a

b

c

d

e

fh

j

k

lm

no

p

qr

s

t uv

w

correlation = 0.64

OBS.

0 1 2 31.5

2.0

2.5

3.0

Su et al. (2017, Nature Comm.)

Interaction between Circulation, Clouds and Precipitation

Narrowing of ITCZ

Tightening of Hadley ascent

Expansion of dry and clear area

Decrease of tropical high

clouds

Enhanced atmospheric

longwave cooling

Reduced cloud longwave warming effect

Intensification of hydrological

cycle

Higher dP/dTs ; Wet Get Wetter and Narrower

Greater subtropical boundary

layer drying

Stronger Hadley

Circulation Change

Reduced shortwave

cloud cooling effect

Greater decrease of subtropical low clouds

High climate sensitivity

Conclusions• Changes of the Hadley Circulation exhibit latitudinally alternating

weakening and strengthening structures.

• Stronger Hadley Circulation change is associated with greater boundary layer drying, which may contribute to the greater decrease of subtropical low cloud fraction and thus higher climate sensitivity.

• The tightening of the ascending branch of the Hadley Circulation is strongly coupled with the decrease of tropical-mean high cloud fraction, which contributes primarily to the inter-model spread in the rates of longwave radiative loss to space and therefore the spread in global-mean precipitation change per unit surface warming.

• Understanding the interactions between circulation, clouds, and precipitation is central to reducing uncertainties of climate sensitivity and hydrological sensitivity.

Atmospheric Energy Constraint on Precipitation

DeAngelis et al. (2015, Nature)dOLR/dTs (W/m2/K)

0.5 1 1.5 2 2.5 3

dP/d

T s (W/m

2 /K)

1.8

1.9

2

2.1

2.2

2.3

2.4

2.5

2.6

2.7

2.8

Corr = 0.67

Model Precipitation Change vs OLR Change per Unit Surface Warming

Implications for Extreme Precipitation Centennial

-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4Tropical dFω/dTs (% K−1)

1.0

1.5

2.0

2.5

3.0

3.5

L vdP

/dTs

(W m

−2K

−1)

ab

c

d

e

f

g

h

ijk

lm

n opq

r

st

uv w

Global LvdP/dTscorrelation = − 0.65

a

b

c

d

efg

h

i

jk

lm

no

p

q

r

st

u

v

w

Wet-area LvdPwet /dTscorrelation = − 0.52

a BCC_csm1.1b BCC_csm1.1mc CCCMA_canam4d CNRM_cm5e CSIRO_access1.0f CSIRO_access1.3g CSIRO_mk3.6h GFDL_cm3i GFDL_esm2gj GISS_e2rk INM_cm4l IPSL_cm5a-lrm IPSL_cm5a-mrn IPSL_cm5b-lro MIROC_esmp MIROC_miroc5q MPI_esm-lrr MPI_esm-mrs MRI_cgcm3t NCAR_cam5u NCAR_ccsm4v NCC_noresm1-mw UKMO_hadgem2-a

Su et al. (2017, Nature Comm.)

Changes of the Hadley Circulation, Clouds and Cloud Radiative Effects in the RCP4.5

∆ = 2074-2098 in “RCP4.5” –1980-2004 in “historical run”

Multi-model-mean from 15 CMIP5 coupled models

“The Wet Get Wetter,The Dry Get Drier”

(Su et al., JGR, 2014)

Normalized CRE Changes

Largest Circu. Change

SmallestCircu. Change

Net CRE

SW CRE

LW CRE