Increase of the satellite retrieved aerosol optical depth due to the enhanced illumination optical depth due to the enhanced illumination of cloud-free areas in the vicinity of clouds
Alexander Marshak (GSFC)
Thanks to: Guoyong Wen (UMBC)L i R (GSFC)Lorraine Remer (GSFC)
Jim Coakley (OSU)
March 26, 2007 Alexander Marshak 1
How do clouds affect aerosol retrieval?aerosol retrieval?
clouds are complex and “satellite analysis may be affected clouds are complex and satellite analysis may be affected by potential cloud artifacts” (Kaufman and Koren, 2006);
Both• cloud contamination (sub-pixel clouds)• cloud adjacency effect (a clear pixel with in the vicinity of clouds)cloud adjacency effect (a clear pixel with in the vicinity of clouds)may significantly overestimate AOT.
But they have different effects on the retrieved AOT: But they have different effects on the retrieved AOT: while cloud contamination increases “coarse” mode, cloud adjacency effect increases “fine” mode.
March 26, 2007 Alexander Marshak 2
Aerosol-cloud photon interaction
0
Thick clouds0
Thin clouds
500
1000ow
1000
row
1500
2000
ro
50
60
0 20 40 60 80
0.373
0.383
0.392
0.402
.47u
m)
Aerosol Optical Thickness (0.47um)
0 1000 2000
2000
col
0.0 0.2 0.4 0.6ASTER_B3N_BL2100x2490_bin
0 500 1000 1500 2000col
0.000 0.125 0.250 0.375 0.500 0.625ASTER B3N TL2100x2490 bin20
30
40
50
Y (
km)
0.335
0.344
0.354
0.364
0.373
osol
Opt
ical
Thi
ckne
ss (
0.4
March 26, 2007 Alexander Marshak 3
ASTER B3N TL2100x2490 bin
0 20 40 60 80X (km)
0
10
0.306
0.316
0.325 Aer
os
Aerosol-cloud radiative interactionThi k l d 15Thin clouds, <τ>=7 Thick clouds, <τ>=15
14
0 2 4 6 8 10 12 14
45.2
50.0
Cloud Optical Depth
14
0 2 4 6 8 10 12 14
45.3
50.0
Cloud Optical Depth
6
8
10
12
Y (
km)
20.9
25.8
30.6
35.5
40.3
Clo
ud O
ptic
al D
epth
6
8
10
12
Y (
km)
21.8
26.5
31.2
35.9
40.6
Clo
ud O
ptic
al D
epth
0 2 4 6 8 10 12 14X (km)
0
2
4
1.5
6.4
11.2
16.1
0 2 4 6 8 10 12 14X (km)
0
2
4
3.0
7.7
12.4
17.1
AOT0.66=0.1
enhancement: Δρ= ρ3D-ρ1D
March 26, 2007 Alexander Marshak 4
Δρ ~ 0.004Δτ ~0.04 or ~40%
Δρ ~ 0.014Δτ ~0.14 or ~140%
from Wen et al. 2007
Contributors to the cloud enhancement
0.05
h t
• Rayleigh scattering• Aerosols
0.04
ecta
nce
AOT=0.1
enhancement: Δρ= ρ3D-ρ1D
• Aerosols• Surface reflectance 0.03
ent o
f Ref
lect
AOT=0.5
AOT=1.0
0.01
0.02
nhan
cem
en
For dark surfaces, the
0.00
0.01Enh
,enhancement only weakly
depends on the AOT
March 26, 2007 Alexander Marshak 5
0.0 0.1 0.2 0.3 0.4Surface Albedo
0.00
If Rayleigh scattering is the main source for the enhancement in the source for the enhancement in the
vicinity of clouds then
we retrieve larger AOT and, namely, and, namely,
the “fine” mode fraction
However, it is not clear yet how much the observed aerosol grows (in the vicinity of clouds) comes fromgrows ( n the v c n ty of clouds) comes from
• the 3D cloud effects in the retrievalsl ph si s ( h midifi ti n)
March 26, 2007 Alexander Marshak 6
• real physics (e.g., humidification)
Enhancement of radiance near cloudsthe twilight zoneg
Cumulus clouds over Atlantics
March 26, 2007 Alexander Marshak 7
Fig. 2 from Koren et al., GRL (in press)
How to (realistically) remove the 3D l d ff t l ?cloud effect on aerosols?
The enhancement is defined as the difference between the two radiances:
• one is reflected from a broken cloud field with the scattering Rayleigh layer above it• and one is reflected from the same broken cloud field but with the Rayleigh l h i ti ti b t tt ilayer having extinction but no scattering
Rayleigh layerRayleigh layer
March 26, 2007 Alexander Marshak 8
Broken cloud layer
Enhancement (θ0=60o, θv=0o)
0.10.47 μm
0.03
0.0350.65 μm
0.06
0.08
cem
ent
0.02
0.025
emen
tAc = 1.0Ac = 1.0
0.02
0.04
enha
nc
0.01
0.015
enha
nce
Ac = 0.5
A = 0 3
Ac = 0.5
A = 0 3
00 10 20 30 40 50
average cloud optical thickness
0
0.005
0 10 20 30 40 50
average cloud optical thickness
Ac = 0.3 Ac = 0.3
average cloud optical thickness average cloud optical thickness
The enhancement vs <τ> for three Ac=1.0, 0.5 and 0.3; Ac=1 corresponds to the pp approximation. Dots (correspond to the thin and thick clouds) and dash lines (correspond to the two Brazilian scenes) are from the
March 26, 2007 Alexander Marshak 9
and dash lines (correspond to the two Brazilian scenes) are from the Wen et al. (2007) MC calculations.
Angstrom exponent
0.47 μm vs 0.65 μm 0.65 μm vs 0.84 μm
p ll t d p ll t d
2
nt 2
2.5
nt
very polluted very polluted
1
1.5
stro
m e
xpon
e
1.5
2
trom
exp
onen
pollutedpolluted
0.5
Ang
s
0.5
1
Ang
sclean
clean
00 10 20 30 40 50
average cloud optical depth
00 10 20 30 40 50
average cloud optical depth
March 26, 2007 Alexander Marshak 10
Three cases: clean, polluted and very polluted.
Conclusions
• 3D cloud enhancement only weakly depends on AOT; • 3D cloud enhancement only weakly depends on AOT; molecular scatt. is the key source for the enhancement;
h h h f f f l • The enhancement increases the fraction of fine aerosol mode;
• Retrieved AOT can be corrected for the 3D radiative effects;
March 26, 2007 Alexander Marshak 11
Airborne aerosol observations in the i i it f l dvicinity of clouds
March 26, 2007 Alexander Marshak 12Courtesy of Jens Redemann
Stochastic model of a broken cloud field
Clouds follow the Poisson distr. and are defined by• average optical depth, <τ>
l d f ti A• cloud fraction, Ac• aspect ratio, AR = hor./vert.
March 26, 2007 Alexander Marshak 13
AR = 2 Ac = 0.3 AR = 1