Aerosol retrievals from AERONET sun/sky radiometers: Overview of - inversion principles - aerosol retrieval products - advances and perspectives The Second International Conference of Aerosol Science and Global Change August, 18-21, 2009, Hangzhou, China O. Dubovik O. Dubovik 1,2 , A. Sinuyk , A. Sinuyk 2 , B.N. Holben B.N. Holben 2 and AERONET team 1 - University of Lille, CNRS, France 2 - NASA/GSFC, Greebelt, USA
29
Embed
Aerosol retrievals from AERONET sun/sky radiometers: Overview of - inversion principles - aerosol retrieval products - advances and perspectives Aerosol.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Aerosol retrievals from AERONET sun/sky radiometers: Overview of - inversion principles
- aerosol retrieval products - advances and perspectives
Aerosol retrievals from AERONET sun/sky radiometers: Overview of - inversion principles
- aerosol retrieval products - advances and perspectives
The Second International Conference of Aerosol Science and Global ChangeAugust, 18-21, 2009, Hangzhou, China
O. DubovikO. Dubovik1,2, A. Sinuyk, A. Sinuyk2, B.N. Holben B.N. Holben2 and AERONET team1 - University of Lille, CNRS, France
Multiple Scat:Multiple Scat: (scalar) Nakajima and Tanaka, 1988, or (polarized) Lenouble et al., JQSRT, 2007
ensemble of polydisperse randomly oriented spheroidsensemble of polydisperse randomly oriented spheroids(mixture of spherical and non-spherical aerosol components)
Accounting for multiple scattering effects
- cloud-free atmosphere;
- horizontal homogeneous atmosphere;
- assumed gaseous absorption and molecular scattering;
- vertically homogenous atmosphere (assumed profile of concentration !?)- bi-directional surface reflectance assumed from MODIS observations
- accounting for polarization effects !?!
ASSUMPTIONS in the retrievals:
AERONETAERONET model of aerosol model of aerosolspherical:spherical:
(Mishchenko et al., 1997)(Mishchenko et al., 1997)
Dubovik et al., 2006
Aerosol single particle scattering:
EACH AEROSOL PARTICLE
- sphere or spheroid (!!!);
- homogeneous;
- 1.33 ≤ n ≤ 1.6 (1.7- ???)
- 0.0005 (0 - ???) ≤ k ≤ 0.5
-n and k spectrally dependent (but smooth)
ASSUMPTIONS in the retrievals:
Aerosol particle size distribution
ASSUMPTIONS:
- dV/dlnr - volume size distribution of aerosol in total atmospheric column;
- size distribution is modeled using 22 triangle size bins (0.05 ≤ R ≤ 15 m);
- size distribution is smooth
0
0.05
0.1
0.15
0.2
0.25
0.3
0.1 1 10
Size Distribtuion Approximation
Particle Radius (m)
Vtotal
(r) = (i=1,...,22)
aiV
i(r)
0
0.05
0.1
0.15
0.2
0.25
0.3
0.1 1 10
Size Distribtuion
dV
/dln
(r)
(m3
/m2
)
Particle Radius (m)
Voriginal(r)
(Twomey 1977)
Trapezoidal approximation
Mixing of particle shapes
ASSUMPTIONS:
- dV/dlnr - volume size distribution is the same for both components;
- non-spherical - mixture of randomly oriented polydisperse spheroids;
- aspect ratio distribution N(is fixed to the retrieved by Dubovik et al. 2006
C Kspherical (
rmin
rmax
k;n;r )V(r )dr (1 C) K (k;n;r ,)
min
max
N()d
rmin
rmax
V(r )dr
retrieved
C + (1-C)
Aspect ratio distr.
0.1
1
10
100
0 40 80 120 160
Spheres Spheroids
Pha
se
Fun
ctio
n (0
.532
m
)
Scattering Angle (degrees)
spheroidspheroid kernels data basekernels data basefor for operational modeling !!!operational modeling !!!
Basic Model by Mishchenko et al. Basic Model by Mishchenko et al. 1997:1997:randomly oriented homogeneous spheroids () - size independent shape distribution
Flexible separation: minimum between: 0.194 and 0.576 m
0.45m
Integral parameters of dV/dlnR: t - total; f - fine ; c - coarseC(t,f,c) - Volume ConcentrationRv(t,f,c) - Mean Radius(t,f,c) - Standard DeviationReff (t,f,c) - Effective Radius
Retrieval accuracy and limitationsSensitivity tests by Dubovik et al. 2000
Real Part Imaginary PartSSA
≤ 0.05
80-100%0.05-0.07
≥ 0.02550%0.03
Size Distribution:
0.00
20.00
40.00
60.00
80.00
0.1 1 10
Err
ors
(%
)
Radius (m)
bias ∆ = ± 0.01Effective
Random errors Nonsphericitybiases
0
0.05
0.1
0.15
0.1 1 10
aureolefull almucantar
dV
/dln
R(
m3/
m2 )
Radius (m)
1.30
1.35
1.40
1.45
1.50
1.55
1.60
Wavelength (m)0.44 0.67 0.87 1.02
Real Part
wide angularcoverage
Error estimates:Error estimates:
New strategy: Errors are to be provided in each single retrievals for all retrieved parameters
Important Error Factors:- Aerosol Loading - Scattering Angle Range - Number of Angles (homogeneity)- Number of spectral channels- Aerosol Type
etc.
Rigorous ERRORS estimates:Rigorous ERRORS estimates: General caseGeneral case: : large number of unknownslarge number of unknowns and and
redundant measurementsredundant measurements
U - matrix of partial derivatives in the vicinity of solution
ˆ x i 2 ˆ x irandom 2 ˆ x i
bias 2
ˆ x
Above is valid: - in linear approximation
- for Normal Noise - strongly dependent on a priori constraints
AERONET estimated broad-band AERONET estimated broad-band fluxes in fluxes in solar spectrumsolar spectrum
Size distribution
FTOA and F
BOA
FTOA and F
BOA
Fbroadband F()dmin
max
Integrations details:min = 0.2 m, max = 4.0 m; more than 200 points of integration between;Aerosol: dV/dlnR - retrieved n() and k() are interpolated/extrapolated; from n(i) and k(i) retrieved;
Radiative transfer code uses 12 moments for P11()
Surface: Surface reflection is Lambertian; Values of surface refelctance are interpolated/ extrapolated from MODIS data valuesGases: Gaseous absorption is calculated using correlated k-distributions implemented by P. Dubuisson
Validation studies:Derimian et al. 2008Garcia et al. 2008( F
BOA ~ 10% agreement )
AERONET estimated aerosol AERONET estimated aerosol forcing in forcing in solar spectrumsolar spectrum
Size distribution
Radiative forcing: ∆FTOA = F0
TOA - FTOA
∆FBOA = F0BOA - F
BOA
Radiative forcing efficiency: ∆FEff
TOA = ∆FTOA/0.55 ∆FEff
BOA = ∆FBOA/0.55
Finding by Derimian et al. 2008: importance of non-sphericity: up to 10% overestimation of ∆FTOA/BOA;
Suggested improvements by Derimian and others: Use net fluxes: ∆FBOA = (F0
BOA- F0BOA) - (F
BOA- FBOA)
Estimate daily forcing Estimates of IR fluxes/forcing
2. Updating scattering model:- including surface roughness for spheroids- expanding ranges of n and k
3. New Inversion developments: - inversion of polarized data (ref: Z. Li talk) - AERONET/MODIS/PARASOL (ref: A. Sinuyk talk)- AERONET/CALIPSO (ref: A. Sinuyk work )- inversion of daily data, combining with PARASOL (ref: O. Dubovik talk )- deriving composition information (ref: G. Shuster work)
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
The Second International Conference of Aerosol Science and Global ChangeAugust, 18-21, 2009, Hangzhou, China