Oxford-RAL IASI Group Meetings

Meeting#09 (15Sep15 @ RAL)

[Prev] [Next]

Oxford: Anu Dudhia, Elisa Carboni, Tony Vincent

RAL: Brian Kerridge, Diane Knappett (Skype), Barry Latter, Richard Siddans, Joanne Walker (Skype)

MORSE (AD) [Next]
Started to recode MORSE (not touched in about 3 years!)
With about 250 modules to adapt, it'll be months rather than weeks before the new version is ready.

SO2 Retrievals (EC) [Prev] [Next]
SO2 vertical resolution paper on ACPD
Working with modeller Julius Vira (FMI) for SO2 amount and altitude.
Comparison of IASI and OMI SO2 mass v month over Ecuador Jun07-Dec13. Full iterative IASI retrievals show good agreement, but OMI initially larger then smaller than IASI. Seems to be a sudden change in 2010.

MetOp data (BL) [Prev]
Local RAL reception of MetOp data (GOME2) has improved, although CEDA recently down.

RFM LUTs (TV) [Prev]
Absorption Coefficient LUTs for RFM/IASI. New approach using new spectral grid defined by radiance spectra from RTTOV ensemble. Errors well below IASI noise but often greater than 0.01K nominal error used to generate grid. RFM using new GRD + LUTs now completes full IASI spectral mid-IR range in ~10 seconds.
Noted that RFM LUTs used by David Moore for CO retrievals (talk at RSPSoc last week).

Linear Retrieval (TV) [Prev] [Next]
Joint retrieval of all relevant products, incl CO2/N2O ratio, CO, OCS, O3, Ts, Tatm at 3 levels, H2O at 2 levels.
Try to account for non-linearity by using variety of gain matrices depending on RTTOV ensemble, based on radiance match. Now including additional ensemble truncated at 800mb, but would be better to use surface pressure map, filtered out bad pixels (~0.2%) in this band (2000-2300cm-1) not evident from quality flags, based on Tsfc. Also tried modelling radiance v sec theta to allow for off-nadir scanning without calculating new spectra, but different for different absorption features. CO plots for month similar to MOPITT. OCS plots show oceanic emissions 0-30S, but also Yemen and point in E Africa.

CH4 Retrievals (DK) [Prev] [Next]
Trying to get IASI IR CH4 processing running on CEMS. Takes about 6 days to fully process a year (seems rather fast) - on 800 cores. Still working on getting NRT data working.
SWIR (2560-2750cm-1: look at residuals, separated by day/ngt land/ocean converged/non converged, etc. Pixel-dependent High-freq structure in 2560-2600 region (ghosts? ILS errors? curvature in 2700-2750cm-1 day-time (error in solar spectrum?).

O3 Retrievals (JW) [Prev] [Next]
Looking at O3 part of IASI retrieval scheme, getting it running on CEMS.