vrame: vertically resolved aerosol model for europe from a synergy of earlinet and aeronet data...
TRANSCRIPT
VRAME:Vertically Resolved Aerosol Model for Europe from a Synergy of EARLINET and AERONET data
Elina Giannakaki, Ina Mattis, Detlef Müller, Olaf Krüger
EARLINET ASOS Symposium, Geneva, 20 September 2010
Outline
The idea behind VRAME
Atmospheric correction
Development of a new aerosol model
Activities of VRAME
Combination of EARLINET and AERONET data
EARLINET ASOS Symposium, Geneva, 20 September 2010
The idea behind
Satellites monitoring the oceans in the visible range of electromagnetic spectrum
The primary goal is to extract concentrations of marine phytoplanktonPhyto (φυτό) +plankton (πλανκτόν)
EARLINET ASOS Symposium, Geneva, 20 September 2010
Satellite Ocean Color A satellite observes both oceans
and the atmosphere
The atmosphere is approximately 90% of the measured signal in the visible and must be accurately modeled and removed
A 1% error atmospheric correction will result in a 10% error in water-leaving radiances
EARLINET ASOS Symposium, Geneva, 20 September 2010
5
Signal
Derived bio-optical products, e.g. chl-a
Top-of-the-atmosphere calibrated radiances
Water-leaving radiances
geolocation and calibration
atmospheric correction
bio-optical algorithms
Ocean Color Retrieval
Level-0 ( L0 )
Level-1B ( L1B )
Level-2 ( L2 )
Level-2 ( L2 )
EARLINET ASOS Symposium, Geneva, 20 September 2010
Shettle and Fenn (1979) introduced a set of basic aerosol models
Continental (rural and urban)
Maritime (oceanic and continental)
d’ Almeida et al. (1991) provide a more comprehensive classification
Maritime (clean, mineral, polluted)
Ocean color implementation, Gordon and Wang (1994) tropospheric, coastal, maritime and oceanic
Antoine and Morel (1999) include operational models for desert dust
Up to now…
EARLINET ASOS Symposium, Geneva, 20 September 2010
MERIS
SeaWiFS
New generation of aerosol models, Ahmad et al. (2010)
bimodal lognormal distribution
Eight relative humidity values (30 – 95 %) Varying fine mode fraction (from 0 to 1)Same spectral dependence of SSA as observed in AERONET
Up to now…
EARLINET ASOS Symposium, Geneva, 20 September 2010
Develop a new aerosol model
EARLINET ASOS Symposium, Geneva, 20 September 2010
Comprehensive, quantitative and statistically significant data base
Develop a new aerosol model
Synergy of EARLINET and AERONET datasets for VRAME
EARLINET ASOS Symposium, Geneva, 20 September 2010
Lidar data already finilized are only used which are included in the
ESA-CALIPSO database
Develop a new aerosol model
Synergy of AERONET and EARLINET datasets for VRAME
The first and currently primary application of the new aerosol model will be the atmospheric correction of MERIS data over the ocean
WAVELENGTHs (MERIS):
443, 510, 560, 709, 778, 865 nm
VERTICAL RESOLVED:
Extinction Coefficient
Single Scattering Albedo
Phase Function
Assymetry Parameter
AEROSOL TYPES:
Marine polluted
European anthropogenic pollution
Continental background aerosols
Saharan dust
Volcanic aerosols in the stratosphere
Aged and young forest fire smoke
EARLINET ASOS Symposium, Geneva, 20 September 2010
VRAME
1.Development of the non maritime aerosol database
Identification of aerosol layers
Aerosol – type analysis
Profiles of optical and microphysical properties
Extinction Coefficient
Single Scattering Albedo
Phase Function
Assymetry Parameter
As input to RT model
EARLINET ASOS Symposium, Geneva, 20 September 2010
VRAME
2.LUT generation
RT calculations for individual and for the generalized
EARLINET/AERONET datasets
The results of these calculations will be the radiances on
TOA at different wavelengths and viewing geometries
1.Development of the non maritime aerosol database
EARLINET ASOS Symposium, Geneva, 20 September 2010
VRAME
1.Development of the non maritime aerosol database
2.LUT generation
3.Sensitivity Study
Is it possible to retrieve the aerosol type from TOA
radiances?
measurable radiances of a satellite sensor show
significant differences in the presence of different
aerosol types
How profitable is the knowledge of the aerosol
type?
climatological mean aerosol data/specific aerosol
type
How profitable is the knowledge of the profile
information?
column average aerosol properties / vertical
information
EARLINET ASOS Symposium, Geneva, 20 September 2010
VRAME
1.Development of the non maritime aerosol database
2.LUT generation
3.Sensitivity Study
4.Aerosol distinction algorithms development
Distinguish between different aerosol types by the ratio of
radiances measured in different MERIS channel
EARLINET ASOS Symposium, Geneva, 20 September 2010
VRAME
1.Development of the non maritime aerosol database
2.LUT generation
3.Sensitivity Study
4.Aerosol distinction algorithms development
5.Validation
Include in-situ measurements of water leaving reflectance
spectra and demonstrate if an improvement of the
atmospheric correction with the new aerosol model can be
achieved EARLINET ASOS Symposium, Geneva, 20 September 2010
VRAME
1.Development of the non maritime aerosol database
2.LUT generation
3.Sensitivity Study
4.Aerosol distinction algorithms development
5.Validation
EARLINET ASOS Symposium, Geneva, 20 September 2010
EARLINET and AERONET for VRAME
Super-site stations Cabauw, The Netherlands (2008-2009) Leipzig, Germany (2001-2009) Potenza, Italy (2006-2009) Athens, Greece (2008-2009)
EARLINET ASOS Symposium, Geneva, 20 September 2010
EARLINET and AERONET for VRAME
Super-sites Cabauw, The Netherlands (2008-2009) Leipzig, Germany (2001-2009) Potenza, Italy (2006-2009) Athens, Greece (2008-2009)
High performance Thessaloniki, Greece (2003-2009) Potenza, Italy (2004-2006) Barcelona, Spain (2006-2009) Granada, Spain (2006-2009) Hamburg, Germany (2006-2009) Minsk, Belarus (2006-2009)
EARLINET ASOS Symposium, Geneva, 20 September 2010
EARLINET and AERONET for VRAME
Super-sites Cabauw, The Netherlands (2008-2009) Leipzig, Germany (2001-2009) Potenza, Italy (2006-2009) Athens, Greece (2008-2009)
High performance Thessaloniki, Greece (2003-2009) Potenza, Italy (2004-2006) Barcelona, Spain (2006-2009) Granada, Spain (2006-2009) Hamburg, Germany (2006-2009) Minsk, Belarus (2006-2009)
Basic stations Belsk, Poland (2006-2009) Lecce, Italy (2006-2009)
EARLINET ASOS Symposium, Geneva, 20 September 2010
EARLINET and AERONET for VRAME
EARLINET ASOS Symposium, Geneva, 20 September 2010
22 July 2004: Smoke / Anthropogenic
Super-sites Cabauw, The Netherlands Leipzig, Germany Potenza, Italy Athens, Greece
10 June 2010: AnthropogenicHigh performance
Thessaloniki, Greece Potenza, Italy Barcelona, Spain Granada, Spain Hamburg, Germany Minsk, Belarus
Leipzig as only bp at 532 nm Basic stations Belsk, Poland Lecce, Italy
Leipzig, 22 July 2004
Time (UTC)
EARLINET ASOS Symposium, Geneva, 20 September 2010
Consistency of two datasets
EARLINET ASOS Symposium, Geneva, 20 September 2010
Aerosol Type, 22 July 2004
EARLINET ASOS Symposium, Geneva, 20 September 2010
0
1
2
3
4
5
6
7
8
9
0 1 2 3 4 5 0 50 100 20 40 60 80 -1 0 1 2 3 0.01 0.02 0.03 0.04
BACKSCATTERCOEFF., sr-1 Mm-1
HE
IGH
T,k
m
355 nm 532 nm 1064 nm
EXTINCTIONCOEFF., Mm-1
LIDAR RATIO, sr
åa,VIS-UV
åb,VIS-UV
åb,IR-VIS
ÅNGSTRÖMEXPONENT
DEPOLARIZATIONRATIO
The optical profiles, 22 July 2004
EARLINET ASOS Symposium, Geneva, 20 September 2010
0
1
2
3
4
5
6
7
8
9
0 20 40 60 80 100 120 140 160 180
EXTINCTION COEFFICIENT [Mm-1]
HE
IGH
T, a
sl [k
m]
355 nm 532 nm 443 nm 510 nm 560 nm 709 nm 778 nm 865 nm 1064 nm
EARLINET ASOS Symposium, Geneva, 20 September 2010
Extinction CoefficientWAVELENGTHs (MERIS channels):
443, 510, 560, 709, 778, 865 nm
443,510 and 560 nmAerosol Extinction Coefficient at 532 nm [LIDAR]Ångström exponent between 355/532 nm [LIDAR]
709,778 and 865 nmAerosol Extinction Coefficient at 532 nm [LIDAR]Ångström exponent between 500/870 nm [CIMEL]
0
1
2
3
4
5
6
7
8
9
0 1 2 3 4 5 0 50 100 20 40 60 80 -1 0 1 2 3 0.01 0.02 0.03 0.04
BACKSCATTERCOEFF., sr-1 Mm-1
HE
IGH
T,k
m
355 nm 532 nm 1064 nm
EXTINCTIONCOEFF., Mm-1
LIDAR RATIO, sr
åa,VIS-UV
åb,VIS-UV
åb,IR-VIS
ÅNGSTRÖMEXPONENT
DEPOLARIZATIONRATIO
The optical profiles, 22 July 2004
EARLINET ASOS Symposium, Geneva, 20 September 2010
Time (UTC)
EARLINET ASOS Symposium, Geneva, 20 September 2010
Single Scattering Albedo
0
1
2
3
4
5
6
7
8
9
0.7 0.8 0.9 1.0
SINGLE SCATTERING ALBEDO
HE
IGH
T [
km]
355 nm 532 nm 1064 nm
With inversion of optical properties [3 backscatters and 2 extinctions] we retrieve the profile of SSA for 355, 532 and 1064 nm
Linear approximation to estimate the desired wavelengths
WAVELENGTHs (MERIS channels):
443, 510, 560, 709, 778, 865 nm
0 20 40 60 80 100 120 140 160 1801E-3
0.01
0.1
1
10
100
1000
SCATTERING ANGLE (degree)
PH
AS
E F
UN
CT
ION
- 4
43 n
m
Layer 1 Layer 2 Layer 3 Layer 4 Layer 5 Layer 6 Layer 7 Layer 8 Layer 9
EARLINET ASOS Symposium, Geneva, 20 September 2010
Phase Function and Asymmetry Parameter
g
coscoscos2
1 1
1
dPg
Refractive index and size distribution from inversion algorithm are used in a Mie code to find the phase function in several wavelengths for each layer
Then the asymmetry parameter is being calculated:
The results: Input for RT model
EARLINET ASOS Symposium, Geneva, 20 September 2010
400 500 600 700 800 9000.00
0.05
0.10
0.15
0.20
EX
TIN
CT
ION
CE
OE
FF
ICIE
NT
[km
-1]
WAVELENGTH [nm]
layer1 layer2 layer3 layer4 layer5 layer6 layer7 layer8 layer9
400 500 600 700 800 900 1000 11000.75
0.80
0.85
0.90
0.95
SIN
GL
E S
CA
TT
ER
ING
AL
BE
DO
WAVELENGTH, nm
Layer1 Layer2 Layer3 Layer4 Layer5 Layer6 Layer7 Layer8 Layer9
0 20 40 60 80 100 120 140 160 1801E-3
0.01
0.1
1
10
100
1000
SCATTERING ANGLE (degree)
PH
AS
E F
UN
CT
ION
- 4
43
nm
Layer 1 Layer 2 Layer 3 Layer 4 Layer 5 Layer 6 Layer 7 Layer 8 Layer 9
Thessaloniki, 10 June 2010
EARLINET ASOS Symposium, Geneva, 20 September 2010
The optical profiles
EARLINET ASOS Symposium, Geneva, 20 September 2010
0
1
2
3
4
5
6
7
8
0 100 200 300 0 2 4 6 0 20 40 60 80 100 120 0 1 2 3 40
1
2
3
4
5
6
7
8
[a]
EXT. COEF. [Mm-1]
HEIG
HT, a
sl [km
]
355 nm
[b]
BSC. COEF. [Mm-1sr-1]
355 nm 532 nm
[c]
LIDAR RATIO [sr]
355 nm
THESSALONIKI, 10 JUNE 2010
[d]
ÅNGSTRÖM EXP.
b355
/b532
EARLINET ASOS Symposium, Geneva, 20 September 2010
532532355 aLRLR
10201020/532
532/355
1020/532
532/355 aA
A
A
Atotal
total
layer
layer
Following Balis et al. (2009) the missing information in lidar profiles are approximated with the synergy of sunphotometer data:
In this way the inversion algorithim is being applied and the sequence of the previous steps could be applied
Additional assumptions to reach MWL stations
CIMEL
Consistency check
EARLINET ASOS Symposium, Geneva, 20 September 2010
161 162 1630.00
0.25
0.50
0.75
1.00
AERONET Input of the RT
443nm 510nm 560nm 709nm 778nm 865nm
443nm 870nm
Thessaloniki, 10.06.2010, anthropogenic aerosols
1020nm 870nm 675nm 440nm
A
ero
sol O
ptic
al D
ep
th
Julian day
RT results
Leipzig, 22 July 2004 as basic station
EARLINET ASOS Symposium, Geneva, 20 September 2010
Apply the information of backscatter contribution to the total backscatter into the total aerosol optical depth to retrieve extinction coefficients at several layers and several wavelengths
Assume same microphysical properties though column
Summary
A 1% error in atmospheric correction will result in a 10% error in water-leaving radiances
• The main objective of VRAME is to develop a dynamic, vertically resolved aerosol model to be delivered to the sattelite community for accurate atmospheric correction.
With the synergy of AERONET and EARLINET data a dynamic aerosol model will be developed
Different assumptions need to be introduced for each group of dataset
EARLINET ASOS Symposium, Geneva, 20 September 2010
Thank you for your attention