chien wang massachusetts institute of technology

Chien WangMassachusetts Institute of Technology

A Close Look at the Aerosol-Cloud Interaction in Tropical Deep Convection

Why Does Aerosol Matter to Clouds

Saturation requirement to form new particles through homogeneous homo-molecular nucleation in the atmosphere: S > 3.5

Note: Typically, in-cloud S < 1.01

Lowering the required S1) Mixed vapor of 2 (binary) or 3 (ternary) species – hetero-molecular

homogeneous nucleation: mainly aerosol nucleation2) Existing surface – heterogeneous nucleation on insoluble (with small

contact angle) or soluble aerosols (ion factor as well)

Heterogeneous nucleation of droplets and ice crystals

Four ice nucleation modes:

Heterogeneous deposition

Immerse

Condensation-freezing

Contact

H2O(g)IN Ice

crystal

cloud droplet

Water droplet nucleation:Hygroscopic aerosols acting as nuclei. Note that it is existing aerosol NOT molecular collision efficiency that determines the nucleation rate

water vapor

CCN IN

clouddroplet

icecrystal

raindrop

snowflake

graupel

hailstone

nucleation

condensation/evaporation

riming/freezing

melting

precipitation

collection/coagulation/conversion

Aerosol and Cloud microphysical processes

Scavenging:nucleation & impaction

Production:evaporation (recycling)

Aerosol-Cloud Interaction

Aerosols Clouds

Aerosol Indirect Effect

Cloud Indirect Effect(?!)

Radiation

and Its Impact on Radiation

A Three-Dimensional Cloud-Resolving Model

Radiation:Radiation:-four-stream -four-stream

including ice cloud including ice cloud

Radiation:Radiation:-four-stream -four-stream

including ice cloud including ice cloud

Cloud Properties:Cloud Properties:winds, T, P, Qv, lightningwinds, T, P, Qv, lightning7 Hydrometeors (Q & N) 7 Hydrometeors (Q & N)

40+ microphysical 40+ microphysical conversionsconversions

Cloud Properties:Cloud Properties:winds, T, P, Qv, lightningwinds, T, P, Qv, lightning7 Hydrometeors (Q & N) 7 Hydrometeors (Q & N)

40+ microphysical 40+ microphysical conversionsconversions

Chemistry:Chemistry:Species: 25g+16c,r+7iSpecies: 25g+16c,r+7i

Reactions:Reactions:35g + 21eq + 32aq + 7h35g + 21eq + 32aq + 7h

Chemistry:Chemistry:Species: 25g+16c,r+7iSpecies: 25g+16c,r+7i

Reactions:Reactions:35g + 21eq + 32aq + 7h35g + 21eq + 32aq + 7h

Environment:Environment:large-scale forcingslarge-scale forcings

and input fluxesand input fluxes

Environment:Environment:large-scale forcingslarge-scale forcings

and input fluxesand input fluxes

Aerosols:Aerosols:N of CCN, INN of CCN, IN

or Multiple mode or Multiple mode multi-moment modelmulti-moment model

Aerosols:Aerosols:N of CCN, INN of CCN, IN

or Multiple mode or Multiple mode multi-moment modelmulti-moment model

11 22

1a 2

5 4a

4b

3b6a

1b

3a6b

References: Wang and Chang, 1993; Wang et al., 1995; Wang and Prinn, 2000; Wang 2005; Ekman et al., 2004; 2006

How does tropical deep convection respond to the 1) increase of CCN concentration; 2) change of aerosol chemical composition; and 3) modified aerosol properties at different altitudes?

What are the chemical and physical consequences of aerosol effect on convection?

Research Issues

Model and Simulations

CEPEX March 8 soundings; 200 100 50 grids with 2 2 0.5 km resolution; 4 hours simulation; supporting runs with 1.0 – 0.25 km horizontal and 250 – 50 m vertical resolutions.

Prognostic CCN (hygroscopic, Aitken or accumulation mode) and IN (water-insoluble); activation of CCN: N = CSk

No “external sources” 90 runs with 30 initial concentrations of CCN, CCN0 from 100 to

5500/cc with a increment of 200/cc; also 50 and 6000/cc; different autoconversion.

A 3-D CRM Study(Wang 2005a&b; JGR)

0 1000 2000 3000 4000 5000 60001

10

100

1000

Cloud Droplet

Clo

ud D

ropl

et C

once

ntra

tion

(1/c

m3 )

Initial CCN Concentration (1/cm3)

(a)

0 1000 2000 3000 4000 5000 60000.0

0.3

0.6

0.9

1.2

1.5

Ice

Cry

sta

l Co

nce

ntr

atio

n (

1/c

m3 )


(c)

Ice Crystal

The Response of Cloud Particle Number

Concentrations to the Increase of Initial CCN

Concentration

Note: All runs use the same initial IN profile.

Effective Radius of Hydrometeors vs. Initial CCN Concentration

0 1000 2000 3000 4000 5000 60000

5

10

15

20

25

30

35

Cloud Droplet

Effe

ctiv

e R

adiu

s (

m)


(a)

0 1000 2000 3000 4000 5000 6000200

300

400

500

600

Raindrop

Effe

ctiv

e R

adiu

s (

m)


(b)

0 1000 2000 3000 4000 5000 600040

50

60

70

80

90

100

110

120

Ice Crystal

Effe

ctiv

e S

ize

(m

)


(c)

0 1000 2000 3000 4000 5000 6000

200

300

400

500

600

Effe

ctiv

e R

adiu

s (

m)


Graupel

(d)

Total Precipitation vs.

Initial CCN Concentration

Maximum Coverage of Cloud vs.

Initial CCN Concentration

0 1000 2000 3000 4000 5000 600010

20

30

40

50

60

70

Max

imum

Clo

ud C

over

(%

of d

omai

n ar

ea)


Berry scheme Kessler scheme NOAC

0 1000 2000 3000 4000 5000 60000.0

0.1

0.2

0.3

0.4

0.5

Tot

al D

omai

n P

reci

p (m

m)



(a)

0 1000 2000 3000 4000 5000 600020

30

40

50

60

Est

ima

ted

"P

reci

p E

ffeci

en

cy"

(%)



0 1000 2000 3000 4000 5000 6000-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Pre

cip

(m

m)


qvf

lux

(mm

)

0 1000 2000 3000 4000 5000 6000

0

5

10

15

20

(b)


Co

lum

n L

oa

din

g o

f Co

nd

en

sed

Wa

ter

(g/m

2)


Budget of Water:

• Supply and consumption of water vapor

increases with CCN0;• precipitation efficiency

varies little with CCN0

3100/cc 3100/cc

700/cc

100/cc

700/cc

100/cc

60 min.

60 min.

60 min.

90 min.

90 min.

90 min.

Surfaces of Updraft 5m/s (Brown) or Downdraft 2 m/s (Blue)

Correlations of microphysical conversions and precipitation:

The importance of riming

0 20 40 60 80 1000

20

40

60

80

100

120

140

160

0 20 40 60 80 1000

20

40

60

80

100

120

140

160

0 20 40 60 80 1000

20

40

60

80

100

120

140

160

0 20 40 60 80 1000

20

40

60

80

100

120

140

160

0 20 40 60 80 1000

20

40

60

80

100

120

140

160

Rel

ativ

e C

hang

es o

f Mic

roph

ysic

al C

onve

rsio

ns

Relative Change of Total Precipitation

migr

crim rrim frg depi

0 1000 2000 3000 4000 5000 6000

50

60

70

80

90

100

Con

tribu

tion

of Ic

e-P

hase

Pro

cess

es to

R

ain

Pro

duct

ion

(%)


Berry scheme mass number Kessler scheme mass numberNOAC runs mass; number = 100

Budget of rain:The dominant role of

Ice-phase microphysics

Cloud-Area-MeanCloud Shortwave

Forcingvs. Initial CCN Concentration

Domain-MeanCloud Shortwave

Forcingvs. Initial CCN Concentration

0 1000 2000 3000 4000 5000 6000-450

-400

-350

-300

-250

-200

Clo

ud-m

ean

Forc

ing

(W/m

2)


-150

-120

-90

-60

-30

0 Hr2 mean Hr3 mean Hr4 mean

Dom

ain-

mea

n Fo

rcin

g (W

/m2)

Solar TOA

Radiative Effects

3.0

3.2

3.4

3.6

3.8

4.0

4.2

Wa

ter

Vap

or

Co

nce

ntr

ati

on

(m

g/m

3)

12-16 km

10.6

10.8

11.0

11.2

11.4

11.6

0 1000 2000 3000 4000 5000 6000590

600

610

620

630

640

650

Wa

ter

Vap

or

Co

nce

ntr

ati

on

(m

g/m

3)


0 1000 2000 3000 4000 5000 6000820

840

860

880

900

920

6-12 km

Cloud-Area Mean


Ambient Mean

Water vapor redistribution by the modeled convection

Efficiency of vertical transport

of gaseous species from the lower to upper troposphere

0 1000 2000 3000 4000 5000 60000.04

0.06

0.08

CH2O


0.61

0.62

0.63

0.64

0.65

SO2

0.64

0.65

0.66

0.67

0.68

Rat

io o

f 4t

h H

our

Mea

n C

loud

Mol

e F

ract

ions

in 1

2-16

kman

d In

itial

Mea

n A

mbi

ent

Mol

e F

ract

ions

in 0

-6km

CO

0 1000 2000 3000 4000 5000 60000.00

0.10

0.20

0.30

0.40


(S

O2)

0.60

0.65

0.70

0.75

(C

H2O

)

0.86

0.87

0.88

0.89

0.90

(H

2O2)

BL

UTconv

Y

YY

)1(1

UTBL

convBL

XX

XX

Scavenging efficiency of a fast soluble gas Y (Cohan et al., 1999):

Here, the dilution factor of a species X:

CO is used as the reference species in the modeled case with β = 82-88%.

Influence of the Modeled Cloud on Gaseous Chemistry:Total condensed water (0.1 g/kg surfaces in yellow color)

and NO2(g)/NO(g) ratio (2.0 surfaces in blue color), all from CCN0 = 100/cm3 run

0 1000 2000 3000 4000 5000 6000

0.1

0.2

0 - 6 km


0.5

0.6

6 - 12 km

Clo

ud-A

mbi

ent R

atio

of O

H C

once

ntra

tion

(Hou

r 4)

1.0

1.1

1.2

1.3

12 - 16 km

Redistribution of OH Radicals

Influence of the Modeled Cloud on Heterogeneous Chemistry O3(s) 2 pptm (blue), CH3OOH(s) 2 pptm (green),

and HNO3(s) 1 pptm (brown); all from CCN0=100/cm3 run.

Summary

CCN CDNC CD re

LWC

W

FreezingLevel & Q

WarmQR

QiCld area

Precip QI

Subl.

TotalQR

Soluble

Tracers

Likely changing color for continental cases

Radiativeforcing

UTreactions

LTphotolysis

Important Points:

Many properties of modeled tropical deep convection DOES NOT respond MONOTONICALLY to the change in CCN0.

Aerosol effect could be more substantial in clean environment (low CCN0 cases).

Dynamics AND microphysics are equally important in determining the response of convective cloud to CCN0.

Ice-phase microphysics plays an important role in precipitation formation and development of modeled cloud.

Some conclusions drawn from this study perhaps can be only applied to the specific cloud type.

Importance of Including Prognostic Aerosol Properties in the Model:Results of a Size-Resolving Aerosol Model (Ekman et al., 2004)

Note: Observed. max. value (>7nm) in anvil: 2.5·104 Modeled max. value (>6nm) in anvil: 5.5 ·104

Al ti

tude

(km

)

20

10

050 250Horizontal distance (km)

177.5 316.2 562.5 1000.0

Aitken mode (6nm<d<30nm)

Concentration (100 cm-3)100.0

0.1 1.0 10.0 100.0 1000.0

Acc. mode aerosol (d>30nm)

20

10

0Al ti

tude

(km

)

50 250Horizontal distance (km)

Concentration (cm-3)

chien wang massachusetts institute of technology

Documents