Oxford MIPAS meeting#20
03 Dec 02

Present

Oxford: Alastair Burgess, Anu Dudhia, Martin Juckes, Francesco Lama, Vivienne Payne, Chiara Piccolo, Clive Rodgers.
RAL: Victoria Jay, Jolyon Reburn

Instrument Status [Prev] [Next]

Operating normally

Auxiliary Data Updates [Prev] [Next]

New data in /home/crun/eodg/mipas/

MW/oldmw (AD) MW database files for `old' (ie hand-selected, no masks) microwindows, renumbered #50-59
CS/oldmw (AD) CS files (LUTs) for old MWs.

L1/L2 Data Delivery [Prev] [Next]

No CDs received since 01 Oct - ESA looking into why but say that CDs will only be delivered up until Nov 5th
(CP) retrieving MIPAS L2 data via ftp from Kiruna up until 2nd Dec, although missing 5 days in November (trying to get a full month).
ESRIN processing now started so all MIPAS orbits should be available either on the Kiruna or ESRIN ftp sites.

RAL Retrievals (JR) [Prev] [Next]

Joint retrieval of O3 and H2O using ESA pT profiles

Redone for `correct' scans for AMIL2DA
Residuals will be sent to Oxford for analysis

Data Assimilation (MJ) [Prev] [Next]

Running simple single-level assimilation scheme for MIPAS O3, driven by ECMWF winds.

[O3, 23 Sep] [O3, 26 Sep] [H2O, 23 Sep] [H2O, 26 Sep] (crosses show location of MIPAS profiles included in assimilation, colours of crosses show actual MIPAS value)

Notes

So far using 850K (~10mb) surface, 8 days at end of Sep 2002
Note unusual O3 depletion in anticyclone as well as vortex.
Will try looking at GOMOS O3 profiles for validation

Gain Oscillation (VP) [Prev] [Next]

OPTIMO-ORM temperature differences showed chequerboard pattern, due to (visible) vertical oscillations in the OPTIMO retrieval rather than the ORM.

Traced to +/- 1% (max) oscillation in L1B data, possibly associated with scan mirror direction and gain calibration
However direct temperature sensitivity (3-4%/K) too small to explain +/- 2K retrieval oscillation, so retrieval probably unstable.
Might be possible to retrieve gain?

Minor Species (AB) [Prev] [Next]

Investigating potential to retrieve minor species from MIPAS spectra for orbit 2081 (24 Jul 02) using OPTIMO

[ClONO2] [CFC-11] [CFC-12] [SF6]

Getting some diurnal variability in N2O5 retrieval, as expected, although retrieval still difficult
No explanation for peak in SF6 at high altitudes - try retrieval using HITRAN line data rather than cross-section data?
Retrieval plots often show unrealistic structure into mid-stratosphere, probably due to influence of a priori. Suggest removing points with large a priori contributions.

REC analysis (AD) [Prev] [Next]

Spectral Shift: [Prev] [Next] REC analysis also shows a spectral shift varying from -0.001cm^-1 to +0.001cm^-1 from 68 to 12 km tangent altitude.

Look for correlation between spectral shift and 1st or 2nd derivative of the vertical gradient of radiance in case shift is related to uniformity of illumination of the FOV. Results show no evident correlation.

Jitter ghost: [Next] suggestion from Manfred Birk at the previous week's CalVal meeting that periodic oscillation in pointing elevation could produce "ghost" spectra at +/-13.5cm-1 with amplitude of the order of 1% of real spectra.

Check microwindow residuals for orbit 2081 at 30km altitude as follows:

Calculate the expected signature for a 1m tangent height displacement, displaced by +/- 13.5cm-1.
Find the best least squares fit to this signature in the residuals of the A+AB band microwindows and compare this with the best fit for the B+C band microwindows (9 and 8 MWs respectively) for each 30km sweep. (using Vivienne's OPTIMO retrievals for this since she has separate residuals for each scan, but it shouldn't make much difference.)
If there is any real signature from a jitter spectrum there should be a good correlation between the fitted amplitudes calculated from these two sets of microwindows.
However, the results show a fairly random distribution, indicating no real jitter signal present in the residuals.
As a control, perform the same analysis for a jitter spectrum displaced by +/-13.0cm-1 (where no signal is expected).
This shows a similar distribution, further confirming the conclusion that the fits at 13.5cm-1 are random.

Instrument Line Shape: [Prev] [Next] REC analysis of signature of 2nd derivative of spectrum in residuals.

Results for Orbit 2081 show AILS width underestimated at low altitude (which may also be due to an offset in the pressure retrieval, ie looking at higher tangent pressures than modelled) and overestimated at high altitude.
Similar analysis applied to the Perturbed ILS error spectra suggests significantly more uncertainty in the AILS width than allowed for in the microwindow generation/error budget, perhaps up to a factor 10.

H2O Retrievals (AD) [Next]

AD has to summarise the state of MIPAS H2O retrievals at the end of the Validation Workshop.

Error budget suggests HITRAN uncertainties are the dominant systematic error source at high altitudes
Results from CalVal meeting suggest H2O may be ~20% too high at high altitude (H2O+CH4 increasing with height, comparison with climatology)
Comparison with climatology also suggests MIPAS too dry over equator
ECMWF comparison shows MIPAS values greater than ECMWF, but ECMWF known to have a low bias, so inconclusive.
Probably useful to try retrieval using different microwindows

NO Retrievals (CDR) [Prev] [Next]

Running MWMAKE microwindow selection, obtained up to 11 bits of information, contradicting conclusions of earlier study by VP. To be investigated.

Oxford MIPAS meeting#20 03 Dec 02

Oxford MIPAS meeting#20
03 Dec 02