Completed final presentations on EUMETSAT studies, draft final reports
produced. Might be an extension to link the two studies, i.e., pass
the retrieval of T,H2O, Surface emissivity data into the CH4 retrieval.
3 days of test data generated (can be processed overnight
using around 50 nodes). Noted that EUMETSAT have a more definitive
estimate of the IASI noise characteristics than the spectrally flat models
commonly used. TV: Hilton et al, BAMS, 2012
[http://dx.doi.org/10.1175/BAMS-D-11-00027.1]
note that IASI noise is very stable with time.
Working on temperature retrieval from IASI CO2 bands (jointly with
H2O and O3), using ECMWF pressure levels.
Using RFM forward model with LUTs - processing a few hundred pixels.
Investigating how to handle surface pressure, which is intermediate between
ECMWF pressure levels.
Trying to get ash retrievals ready for SCOPE intercomparisons. Now using
latitude-dependent covariance matrices, and running retrievals for specific
volcanic eruptions. Next task is to convert data to required format -
RAL can help with this (GT probably). Tried different two
different types of ash, but only a limited selection of infrared ash spectra
available.
Fixed a couple of problems so improved CEMS/LOTUS processing of
SO2 linear retrievals - 4 years now processed. Main problem is juggling
of disk space, but also some unexplained retrieval failures.
Iterative retrievals running at Oxford and about to transfer these to CEMS.
Writing up SO2 paper.
(JP) Looking for low-level SO2 emission signatures in time-averaged
IASI SO2 linear retrievals, 2008-9. Some clear volcanic signatures, also
Beijing pollution. Looked at iterative retrievals over Ecuador volcanic arc -
find persitent negative SO2 signal - connected to surface
elevation/emissivity?
Now has a working internal version (within Oxford) but no data
displayed yet.
IDL bug creating netCDF4 files using 'STRING' variables - particularly a
problem for DJANGO which expects date/time information as string format.
Also a problem that SO2 flag seems not to be working. Aiming for public
release by Easter.
Part of ongoing work with UKMO, creating
database of RFM high resolution optical depth spectra
for ECMWF set of 83 profiles, 450-3260 cm-1 @ 0.001 cm-1 resolution.
Converted to netCDF format.