Introduction

Invisible clouds in this study mean super-thin clouds which cannotbe detected by MODIS but are classified as clouds by CALIPSO.

These sub-visual clouds may exist globally and may have effects onEarth-atmosphere radiation budget and remote sensing of aerosols.

In this study, 12-month (Jan 1 – Dec 31, 2007) CERES, MODIS, CALIPSO, and AIRS measurements are analyzed for these clouds.

Wenbo SunScience Systems and Applications, Inc.

Mail Stop 420, NASA Langley Research Center, Hampton, VA23693, USAwenbo.sun-1@nasa.gov

Study Invisible Clouds for Glory Aerosol Product

Glory Science Team Meeting, August 10-12, 2011, New York

Total attenuated backscatter at 532nm from CALIPSO lidar

20km CERES FOV

1km MODIS Pixel

333m CALIPSOLidar Shot

. Cloud coverage percentage is calculated using along-CALIPSO-track CALIPSO and MODIS data.

. Radiation energy budget effect of invisible clouds is estimated on CERES FOVs of MODIS clear and CALIPSO cloudy.

Method and Data

CCCM data – CERES, CALIPSO, MODIS, and MOA

AIRS data – L3 daily 1°x1° gridded standard retrieval product V5

CCCM data

MODIS-derived 12-month clear percentage of CERES FOVs

CALIPSO-derived cloudy percentage in MODIS-clear CERES FOVs

-180 -120 -60 0 60 120 180180

150

120

90

60

30

0

Longitude (deg)

Co

latitu

de

(d

eg)

0

1

2

3

4

5

6

7

8

9

10

-180 -120 -60 0 60 120 180180

150

120

90

60

30

0

Longitude (deg)

Co

latitu

de

(d

eg

)

0

1

2

3

4

5

6

7

8

9

10

-180 -120 -60 0 60 120 180180

150

120

90

60

30

0

Longitude (deg)

Co

latitu

de

(d

eg

)

0.0

2.0

4.0

6.0

8.0

10

12

14

16

18

20

-180 -120 -60 0 60 120 180180

150

120

90

60

30

0

Longitude (deg)

Co

latitu

de

(d

eg

)

0

2

4

6

8

10

12

14

16

18

20

Daytime Purely Clear

12-month CERES FOVs Sampling Distribution

Daytime Invisibly Cloudy

Nighttime Purely Clear Nighttime Invisibly Cloudy

Zonal and altitude distribution of invisible cloud occurrence frequency (in the unit of CERES FOV number) for daytime (left panel) and nighttime (right panel) ocean

Zonal and altitude distribution of invisible cloud

0 30 60 90 120 150 1802

4

6

8

10

12

14

16

18

Zonal M

ean H

eig

ht of H

ighest L

ayers

(km

)

Colatitude (deg)

One-year zonal mean of invisible cloud heights for nighttime oceans

The extent of Hadley cell is a metric of climate change.

Invisible clouds provide a novel way for satellite remote sensing of Hadley cell.

Invisible clouds correlate with general circulations

Regular ice clouds do not well correlate with general circulation cells

Sun and Lin (2011) ACPD

Instantaneous CERES SW flux is converted to diurnal 24-hour mean value by using previously made lookup tables from CERES TRMM processing-orbit data(Loeb & Manalo-Smith 2005).

Invisibly thin clouds have ~2.5 Wm-2 diurnal mean SWcooling effect.

Daytime Invisible Clouds’ Radiation Effect

0 30 60 90 120 150 18075

80

85

90

95

Colatitude (deg)

Zon

al M

ean

Inta

ntan

eous

SW

Flu

x (W

m-2)

Comparison of CERES outgoing LW flux for clear (filled circle) and invisibly cloudy (open circle) cases

Nighttime Clear Sky and Invisible Cloud Radiation

The CERES LW flux difference between clear and invisibly cloudy FOVs could be a result of water vapor absorption. This makes the quantification of the invisible clouds’ effect on LW radiation difficult.

Humidity and temperature difference between clear and invisibly cloudy environment

0 30 60 90 120 150 1800

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

AIR

S W

ate

r V

ap

or

Mix

ing

Ra

tio

(g

/kg

)

Colatitude (deg)

1000mb 925mb 850mb 700mb 600mb 500mb 400mb 300mb 250mb 200mb 150mb 100 mb

cloudy

clear

Daytime zonal mean instantaneous column water vapor amount from AMSR-E (filled circle) and AIRS (open circle) for clear (black and red) and invisibly cloudy (blue and green) ocean

Daytime zonal mean instantaneous temperature profiles from AIRS for clear (thin curve) and invisibly cloudy (thick curve) ocean

Comparison of modeled outgoing LW flux for clear (filled circle) and invisibly cloudy (open circle) cases using atmospheric profiles of clear CERES FOVs.

Modeled Invisible Clouds’ Radiation Effect

Comparison of CERES and modeled LW flux for clear CERES FOVs

Modeling LW flux for daytime and nighttime ocean using atmospheric profiles from MOA and AIRS dataset

Sun et al. (2011) JGR

daytime

nighttime

Invisible cloud effect

0 20 40 60 80 100 120 140 160 180

0.1

1

Invisibly CloudyClearDaytime Ocean

Ae

roso

l Op

tica

l De

pth

at 0

.55

mColatitude (deg)

Zonal mean MOD04 aerosol optical depth at 0.55 µm for daytime ocean

MOD04 cloudy; CALIPSO clear MOD04 clear; CALIPSO cloudy

3.93%

8.17%

76.55%

11.36%

MOD04 clear; CALIPSO clear MOD04 cloudy; CALIPSO cloudy

Statistics of 1km x 1km areas with matched and unmatched cloud masks from CALIPSO and MOD04

Effect of invisible clouds on MODIS aerosol product

~25% MOD04 aerosol product is contaminated by invisible clouds

Updated Work Outline

1. Use CALIPSO and MODIS data to study the distribution of invisible clouds;

2. Develop a polarized radiative transfer model and use RSP data to validate the model;

3. Use RSP data plus model to study the BPDFs of different surfaces;

4. Use model to study the sensitivity of the polarized reflectance at 1.37 µm to invisible clouds, aerosol, low clouds, and surface optical properties;

5. Develop the algorithm to use Glory 1.37-µm polarized radiance to retrieve the physical properties of invisible clouds;

6. Develop the algorithm to remove the invisible clouds’ effect from Glory radiances, so that the corrected radiances are suitable for retrieval of aerosol and cloud phase.

1. Adding-Doubling radiative transfer model: This can calculate full Stokes vector (I, Q, U, V). 2. Atmospheric profiles: Standard Atmosphere now.3. Spectral gas absorption: Line-by-Line and k-distribution plus ozone cross-section table.4. Molecular scattering: Rayleigh.5. Particulate absorption and scattering: Mie for water clouds (Gamma size distribution); FDTD for aerosols (lognormal size distribution with fine and coarse mode); FDTD plus GOM for ice clouds (lognormal or measured size distributions).6. Surface reflection model: Lambert surface for land now. Cox & Munk + foam for wind-roughened ocean.

The polarized radiative transfer model

0 10 20 30 40 50 60 70 80 900.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

RAZ = 180o

VZA (o)

Refl

ecta

nce

90 80 70 60 50 40 30 20 10 00.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Ocean Wind: 2.5 m/s 5.0 m/s 7.5 m/s10.0 m/s15.0 m/s

VZA (o)

WL = 670 nm

SZA = 30o

RAZ = 0o

Refl

ecta

nce

90 80 70 60 50 40 30 20 10 00.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Ocean Wind: 2.5 m/s 5.0 m/s 7.5 m/s10.0 m/s15.0 m/s

WL = 670 nm

SZA = 30o

RAZ = 0o

DO

P

VZA (o)0 10 20 30 40 50 60 70 80 90

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

RAZ = 180o

DO

P

VZA (o)

Sea-Salt AOT = 0.075

Sensitivity of clear ocean total reflectance and DOP to wind speed

0 30 60 90

90 60 30 00.4

0.5

0.6

0.7

0.8

0.9

Ice Cloud

Water Cloud

SZA = 30 degWL = 865 nmTo

tal R

eflec

tance

Loeb et al (2005)

CERES SW anisotropic factors in the principal plane

Water Clouds Ice Clouds

Model results have excellent agreement with CERES data in total radiance angular anisotropy, except the ice cloud specular reflection, since we assume purerandomly oriented ice crystalsin the model.

Conclusion

1. CALIPSO shows that ~50% of the clear sky are actually covered by invisible clouds.

2. These clouds have big impact on global radiation energy budget, but are not observed and studied sufficiently.

3. These clouds may introduce significant uncertainties into Glory aerosol data if not identified and properly removed from Glory product.

4. For these super-thin clouds, Glory APS is the most precise instrument to measure their physical properties.

5. We will use Glory and CALIPSO data to study the physical properties of invisible clouds. These properties will then be used to reduce the uncertainty in the Glory aerosol and cloud phase products.