Oxford-RAL IASI Group Meetings

Meeting#01 (25Nov14)

Oxford: Elisa Carboni, Anu Dudhia, Catherine Hayer, Tony Vincent
RAL: Richard Siddans, Brian Kerridge

SO2 Retrievals (EC) [Next]

SO2 retrievals using 1100-1410cm-1, also an ash retrieval using longer wavelengths. Initial 'detection' using linear retrieval. Full retrieval:

• forward model: modified RTTOV with SO2
• state vector: assume Gaussian SO2 profile - retrieve pressure of peak amplitude, column amount, spread (currently fixed), and surface temperature.
• Temperature and H2O from ECMWF
• Measurement error covariance allows for cloud-contamination, single global covariance matrix.

Plan to incorporate detection in NRT processing on CEDA and, eventually, retrieval.

RAL Retrievals (RS) [Next]

Currently 2 IASI retrieval schemes at RAL

• Within NCEO, mainly used for CH4 but also works for O3, CO and HNO3.
  • Forward model: RTTOV
  • State vector: CH4, N2O (used to fit effective cloud parameters CEF, CTH), Ts, H2O, HDO.
  • Temperature, a priori Ts and H2O from ECMWF
  CH4 retrieved from entire IASI dataset, 1 day in 2, 1 pixel in 4. Also used for EUMETSAT studies to define IASI OE CH4 retrieval scheme, and look at using principal component version of L1C data. Also the basis of an IASI O3 NRT scheme, add-on to GOME O3 scheme within CEDA.
• Temperature and H2O retrieval, part of EUMETSAT study to look at AMSU/MHS for T, H2O, O3 if added to IASI. Uses RTTOV version of Wisconsin emissivity data. Aim to use this retrieved T to improve CH4 retrieval.

Diane Knappet starting on IASI will use 3.7um CH4 band to improve CH4.

Retrieval Levels (TV) [Next]

Currently investigating selection of pressure levels for retrievals of O3, H2O, CH4, CO and T. Comparing 3 methods

1. Uniform pressure increments
2. Selection of levels at increments along diagonal of averaging kernel matrix
3. New method: starting with fine grid and selectively merging levels while maximising DOF.

Previously examined effect of apodisation on IASI noise - now has IDL procedure to produced apodised error covariance matrix.

IASI NRT web-site (CH) [Next]

Creating IASI NRT web-site, hosted at CEDA, initially with SO2 linear and full retrieval results, then extend to more species (flagging and retrievals). Plan is to use IASI NRT and ECMWF forecast data first (ECMWF data required for T,H2O for SO2 full retrieval), then later replace this with IASI off-line and ECMWF analysis. Current problem at CEDA is that there is no IASI off-line data since March.

Reading IASI Data [Next]

Tools used to read IASI data

• EC uses BEAT
• TV uses CODA, which can also extract AVHRR data (unlike BEAT).
• RAL have EUMETSAT IDL reader
• CH has difficulty reading IASI NRT data (BUFR format) - RS said he had some IDL software to do this