Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 1 DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING Occultation and limb scattering Erkki Kyrölä Finnish Meteorological Institute 1. Instruments 2. Occultation: GOMOS inversion 3. Limb scattering: OSIRIS inversion 4. Summary 5. References
36
Embed
DRAGON ADVANCED TRAINING COURSE IN …earth.esa.int/dragon/D3_L3_Kyrola.pdf · Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 1 DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE
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
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 1
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
Occultation and limb scatteringErkki Kyrölä
Finnish Meteorological Institute
1. Instruments2. Occultation: GOMOS inversion
3. Limb scattering: OSIRIS inversion4. Summary
5. References
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 2
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
Occultation: SAGE, MSX UVISI, GOMOS, SCIAMACHY
Limb: OSIRIS, SOLSE/LORE, SCIAMACHY, GOMOS
INSTRUMENTS
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 3
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
Iref
occI
T(λ)= Iref(λ)Iocc(λ)
OCCULTATION
calibration free principle
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 4
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
GOMOS: Measured Sirius reference spectrum
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 5
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
GOMOS: Measured Sirius transmitted spectrum
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 6
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
GOMOS: Calculated Sirius transmissions
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 7
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
T(λ,z) = exp(− σ j (λ,T(z(s))ρ j∫∑ (z(s))ds) Beer-Lambert law
Occultation inversion is simple because...
But ...
Occultation inversion
sσ = cross section
z(s)
ρ = number density
= temperatureT
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 8
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
Weak scintillations: intensity maxima and minima
Density fluctuation
Strong scintillations: multiple stars
Stellar occultations: dilution & scintillations
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 9
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
T(λ,z,t2)
T(λ,z,t1)
different times
Stellar occultations: chromatic time delay
Different colors different refraction angles
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 10
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
T(z,λ) = Tref Text
We can, however, write
Transmission from refractive effects can be estimated from ray tracing calculations (dilution, chromatic effects). In addition, we need photometer measurements to estimate the random part (scintillations).
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 11
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
GOMOS: Horizontal transmissions 5-100 km
O3 inmesosphere
O3 instratosphere
NO2 instratosphere
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 12
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
T(λ,z) = exp(− σ j (λ)N∑ j(z))
N j (z) = ρ j∫ (z(s))ds
Occultation inversion using Beer-Lambert: Two step
Spectral inversion
Vertical inversion
This separation is not true if cross sections depend on temperature. In these cases we can use iteration over spectral and vertical inversion or one-step inversion.
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 13
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
C = covariance matrixT = transmission vector (all wavelengths)N = column density vector (different constituents)
We aim to minimize
Solution by Levenberg-Marquardt algorithm
Spectral inversion
S(N) = (Tobs − Tmod (N))T C−1((Tobs − Tmod (N))
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 14
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
Aspects of spectral inversion in UV-VIS
• Linearization• Non-linear approach
• Spectrally global• Spectral windows
• Absolute cross sections• DOAS
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 15
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
10-23
10-22
10-21
10-20
10-19
10-18
10-17
Cro
ss se
ctio
n (c
m 2
)
6000500040003000Wavelength (Å)
O3
NO2
O3
NO3
OClO
BrO
Cross sections
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 16
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
Transmission components at 27 km
ozone
NO2
NO3
Rayleigh
aerosol
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 17
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
An example of LM iteration
red=model to be fittedblue =data
black=residual
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 18
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 19
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 20
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 21
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
Chi2 vs iteration number
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 22
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
GOMOS vertical inversion
K =
d11
2d21 d22
2d31 2d32 d33
⎡
⎣
⎢ ⎢ ⎢ ⎢ ⎢ ⎢
⎤
⎦
⎥ ⎥ ⎥ ⎥ ⎥ ⎥
Discretize N(z) = ρ∫ (z(s))ds
N = Kρ
where the kernel matrix is
Onion peel solution
d11d22 d21
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 23
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
Tikhonov regularization
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 24
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
• Data extraction• Datation•Geolocation (ECMWF+MSIS90)• Wavelength assignment• Spectrometer samples correction• Photometer data processing• Central band background estimation• Star spectra computation• Transmission computation• Products generation
GOMOS Level 1b
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 25
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
GOMOS level 1Raw data
Geolocation& ray tracing
Instrumentalcorrections
Photometerdata
Transmissiondata
Limbdata
ECMWFprediction/analysisMSIS90
Calibrationdatabase
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 26
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
GOMOS level 2
Crosssections
Spectralinversion
Local densitiesO3, NO2, NO3aerosols, AirT, H2O, O2
Verticalinversion
Line densities
Level 1transmissions
Dilution& scintillation
corrections
Level 1photometer
data
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 27
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
LIMB SCATTERING
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 28
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
OSIRIS radiances
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 29
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
Scattered limb radiances
Total radiance= single scattering + multiple scattering
I = Isun Tsun∫ (s)(ρa (s)σ a (λ)Pa + ρR (s)σ R (λ)PR )Tdet (s)ds + Ims
s
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 30
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
Single and multiple scattering
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 31
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
Difficulties in limb radiative transfer
• MS time consuming• Albedo• Clouds• Aerosols• Polarization• Raman scattering
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 32
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
Day 3 Lecture 3 Retrieval techniques - Erkki Kyrölä 36
DRAGON ADVANCED TRAINING COURSE IN ATMOSPHERE REMOTE SENSING
ReferencesThis presentation has followed:
Kyrölä, E., E. Sihvola, M. Tikka, Y. Kotivuori T. Tuomi, and H. Haario, Inverse Theory for Occultation Measurements 1. Spectral Inversion, J. Geophys. Res., 98, 7367-7381, 1993.
Oikarinen, L., E. Sihvola, and E. Kyrölä, Multiple scattering radiance in limb-viewing geometry, J. Geophys. Res., 104, 31261-31274, 2000.
Auvinen, H., L. Oikarinen and E. Kyrölä, Inversion algorithms for limb measurements, J. Geophys. Res., 107, D13, 2001JD000407, ACH 7-1: 7-7, 2002