Oxford MIPAS meeting#24
04 Feb 03

Present

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

L2 Data (CP) [Prev] [Next]

In /home/crun/eodg/mipas/L2/ [Plots of profile locations]

Data received up until 2nd February

Starting to produce L2 antarctic profiles again

Software Updates (VP) [Prev] [Next]

Software in /home/crun/eodg/mipas

/optimo Source code added

Auxiliary Data Updates (AD/CP) [Prev] [Next]

New data in /home/crun/eodg/mipas

/MW/, /CS/, /grd/
- New microwindow databases, LUTs and irregular grid data for NH3, OCS and SF6 microwindows

Occupation Matrix Statistics (CP) [Prev] [Next]

Still detecting unidentifiable OMs - suggest contact Sven Bartha with a specific example.
24 and 27 km sweeps appear to be corrupt more often than other altitudes (approx 3% of scans)

Missing Antarctic Data [Next]

Very few L2 profiles of pT generated from early December until late-January, although now starting to reappear.

Detailed discussion with Michael Hoepfner, summary:

No nominal OM pT retrievals (ie using OM 001) below 65S from early December 2002 until late January 2003, some failed retrievals using different OMs, implies ESA L2 processing is detecting a problem with L1B spectra
IMK detect corrupt spectra , often at 24/27, which appear to be symptomatic of a phase error. These occur frequently at southern latitudes throughout January (but not December) and also over the equator.
Current speculation is that high radiances associated with high antarctic stratospheric temperatures and/or high equatorial clouds lead to error in L1 processing

OPTIMO (VP/AB) [Prev] [Next]

(VP) Currently implementing cloud detection

(AB) Attempting to run OPTIMO on Linux machines, but obtaining large differences in results compared to Alpha machines.

Data Assimilation (MJ) [Prev] [Next]

Assimilating MIPAS O3 and H2O in an isentropic model.

Recently obtained SAGE III O3 data and using assimilation model to compare SAGE III and MIPAS ozone, currently at 850K (~10mb).
Generally agree within random error bars, but suggest MIPAS biased 3-4% high wrt SAGE III

SO2 Retrievals (AB) [Prev] [Next]

(AB) Retrieving SO2 from ESA L2 residual spectra using REC analysis

Currently analysing single day, using only CH4_0001 MW
Generally shows less SO2 than predicted, but excess SO2 coincides with high cloud locations, so probably a false signal

Next: incorporate more microwindows and select individual altitudes

Retrieval Non-Linearity (CR/CP) [Prev] [Next]

Examining the effect of non-linearity on retrieval convergence and use of singular vectors.

Take mid-latitude, day-time H2O profile, x0
Perturb by retrieval error 1 sigma, xp
Compute Forward Model difference F(x0)-F(xp)
Compute Jacobian difference K(x0)-K(xp)

Still need to scale Jacobian difference by (xp-xo) for comparison.

Oxford MIPAS meeting#24 04 Feb 03

Oxford MIPAS meeting#24
04 Feb 03