Retrieving daytime mesospheric O3 from OSIRIS
measurements of O2 (a1∆)
Anqi Li et al.
Chalmers University of Technology
Gothenburg, Sweden
A brief story of ...
Odin components
OS S
IRI
SMR
Optical Spectralgraph
InfraRed Imaging System
Sub-Millimetre Radiometer
since 2001
IRI in a nutshell
128 pixels vertically
O2(1Δ) at 1.27 μm
O_2 (^1 \Delta)
Odin-IRI-O2 channel
Observation
N pole
S pole
EQ
EQ
Optimal estimation method
a.k.a.
Bayesian statistical approach
Inversion step 1
Limb radiance -> Volume emission rate
\mathbf{y = F (x)}
Limb
Volume
\mathrm{O}_2(^1\Delta)
\mathrm{O}(^1D)
\mathrm{O}_3
\mathrm{O}_2(^1\Delta)
\mathrm{O}_2(^1\Sigma)
\mathrm{O}_2
Optimal estimation method
a.k.a.
Bayesian statistical approach
Inversion step 2
\mathbf{y = F (x)}
\mathrm{O}_2(^1\Delta)
\mathrm{O}_3
Non-linear
Levenberg-Marquardt method
Ozone result
Ozone from Odin compare
N pole
S pole
EQ
EQ
Question:
Why so many negative O3 values ?
\mathrm{O}_2(^1\Delta)
\mathrm{O}_3
Morning sunrise
Odin 6-18h orbit
(Mr. Sloth in Zootopia)
Equilibrium assumption
>> underestimation of O3
Invalid O3
After a very special treatment on the retrieval scheme ...
