Oxford MIPAS meeting#11
30 Jul 02

• /Lv2PP - corrected code for generating L1C data. Previous versions (up to and including 02JUL02) had 1 grid point shift in spectral axis so all L1C files generated with this or earlier versions should be discarded (Lv2PP version number is in first header record of L1C files - see L1C Format)

• Due to technical problems in ground segment, dissemination of products to Validation PIs delayed until 15th August.
• Commissioning phase review still fixed for 9-13 September at ESTEC (MIPAS on 12th September)

• Perform REC analysis (requires modification to program l2res.pro to allocate OM and day/ngt flag to each extracted residual).
• Generate plots of supplied data - probably just pT as pressure v profile no.

• Contacting BADC about adding starting to send ESA L1B spectra via DDS in the next couple of days (in conjunction with AATSR data)

• Earlier problems due to 1 grid point shift in L1C files (Lv2PP bug)
• pT now retrieved for full orbit - currently 2-3 iterations per MW. [ESA Retrieval] [OPTIMO Retrieval]
• Next: attempt other species

• pT retrieval showed better pressure retrieval than OPTIMO, pointing errors of the opposite sign (odd since both codes use the same microwindows and forward model) [Plot]
• Unfortunately current version of code results in `floating invalid' with BT2 spectra!

• Sorted out problems in O3 1km resolution averaging kernels
• Working on 1D linear diagnostics

I've produced some co-added spectra of orbit 504. I co-added all "Day" spectra and "Night" spectra (32 roughly in each of them). I've looked into the CO region at low tangent heights (following our conversation during the SAG). It seems the CO lines are in absorption at low tangent heights. In the Day spectra we see part of the line (wings) in absorption and the center in emission (coming from higher altitudes). I guess it is going to be difficult to retrieve CO at low tangents heights (below about 15 km). Please, have a look at the spectra in http://www.iaa.csic.es/~puertas/mipas.html

• Region around 960cm-1 suggested for simple radiance threshold test (isolated CO2 lines only)
• No obvious advantage from dual channel approach plotting radiance from one transparent channel against another - simple radiance threshold works best
• Note that the U.Leicester technique relies on the ratio of spectrally averaged radiances for transparent/semi-transparent microwindows. This was developed for CRISTA (resoln ~ 1cm-1) so may not fully exploit MIPAS ability to see `between' lines.

• Either perform cloud detection in Lv2PP and associate Cloud Index or extinction with each tangent altitude in L1C data; or
• Add cloud detection microwindows to L1C data and perform cloud detection in Level 2 code OPTIMO/MORSE

• Either skip cloud-contaminated retrieval levels completely (=variable length profile), or
• Set a priori continuum extinction large for cloud-contaminated levels