Title: Status Operations and Dataprocessing
1Statusof OMI Dataprocessing Jacques Claas
(KNMI)Phil Durbin (Adnet Systems, Inc.)
2Processing Summary
3Data processing status
- Collection 3
- THE dataset for L1B and L2 products. Data are
available on the DISC. Do NOT use older data. - The L1b forward stream as well as the L2 NRT
stream are very stable - NRT data available within three hours after the
overpass - GDPS 1.1.1.1
- The new GDPS version 1.1.1.1 has been running in
forward stream from April 15th 2008 till Oct. 8th
2008 - Improvements gt improved wavelength assignment
gt improved radiance goniometry gt setting noise
levels for individual pixels gt added float
field in irradiance product - GDPS 1.1.2
- Running since Oct. 8th 2008
- Static flagging (XTrackQualityFlags) of CCD rows
affected by row anomaly
NRT image of tropospheric NO2 over Europe
NRT image of tropospheric NO2 over Beijing
4Summary of data processing issues
- FMU anomaly
- Permanent data loss for 2 orbits 18027 18028
(Dec 5, 2007). - No science data for 9 orbits 18235 18243 (Dec
19-20, 2007). - Operations software anomaly
- No science data for 51 orbits 22349 22399 (Sep
27-30, 2008) - Row anomaly
- Two row anomalies exist, affecting part of the
OMI field of view. - Work on correction algorithm in GDPS is on-going.
- L2 developers must carefully read the L1b readme
file on the DISC. - AURA re-phasing
- Due to the AURA re-phasing orbits now have
different geo-location. - Current definition of NOSE parameters is not up
to date. DO NOT USE!
5Impact of FMU anomaly on data processing (1/3)?
- Formatting Multiplexer Unit electronics for
writing to / reading from the Solid State
Recorder. - A FMU anomaly started on Dec 4th 2007 corrupting
the SSR partition 6 data containing S/C
housekeeping data and MLS plus HIRDLS science
data. - The anomaly is most likely caused by a memory
chip. The FMU behaviour was erratic at the
time of the anomaly. - The GBAD data that OSIPS needs for its L0 -gt L1B
processing is derived from the S/C
housekeeping data (attitude/ephemeris). Due to
the anomaly no data processing was possible. - There has been no loss of OMI science data due
to the anomaly. But on Dec 19th no
measurements could be performed for 8 orbits
(18235 18242). - Since Dec 19th 2007 for unexplained reasons, no
anomalies have been observed in the GBAD data
and HIRDLS data. MLS misses only 2-10 science
packets a day. - In case the FMU anomaly shows up again, either
a FMU reset or a FMU swap from the S/C A to the
B-side will be performed. Based on a risk
assessment procedures for this have been
developed.
6Impact of FMU anomaly on data processing (2/3)?
7Impact of FMU anomaly on data processing (3/3)?
- Summary
- The GDPS needs attitude and ephemeris data for
itsL0 -gt L1b processing in order to determine
the geo-location of ground pixels. - Until the FMU anomaly, attitude and ephemeris
data were based on GBAD data - Since the FMU anomaly, attitude and ephemeris
data are based on the OMI specific APID 1836/1837
data. - Differences between old and new attitude and
ephemeris data turn out to be very small.
Differences in geo-location are less than 1 of
the ground pixel size (see KNMI Technical Note
908).
8Row anomaly (1/4)?
- A row anomaly was discovered in September
2007.This anomaly has persisted since then. - On May 19th 2008, a new row anomaly was
discovered. - Both row anomalies impact specific CCD rows (is
specific part of the OMI field of view). - 2007 anomaly 2 rows are affected for complete
illuminated orbit. - 2008 anomaly 4 rows are affected for northern
part of illuminated orbit. - Current tentative explanation is that a small
piece of MLI has moved into the field of view of
the nadir port. - GDPS
- The current GDPS 1.1.2 applies static flagging of
the affected CCD rows in the L1b HDF files - A new GDPS version that will use a correction
algorithm enabling dynamic flagging is under
development. - Data users must carefully read the L1b readme
file on how to treat the affected CCD rows in
their retrieval algorithms.
9Row anomaly (2/4) OMI L0 data
UV image from April 20
X-track
UV image from September 11
wavelength
10Row anomaly (3/4) impact on L2
11Row anomaly (4/4) impact on L2
12Impact of the AURA re-phasing on data processing
(1/2)?
- On NASA instigation, between Dec 5th 2007 and
May 8th 2008, a relocation of the AURA
satellite was performed. - Purpose - Improving the science output by
combining AQUA and AURA science data
(overlapping fields of view for particular
instruments, smaller time difference). -
distance before 5918 km, distance after 3674
km - Consequence - the OMI groundtrack has moved
about 120 km eastwards (about 6 groundpixels).
- the OMI monthly spatial zoom-in measurements
now have a different geo-location. - the NOSE
parameters have to be updated (not done yet!). DO
NOT USE!! - Decision - the timing of the monthly spatial
zoom-in measurements will not be changed, hence
will keep the changed geo-location.
13Impact of AURA re-phasing on dataprocessing (2/2)?
Before re-phasing
OMImonthlyspatial zoom-inmeasurements
After re-phasing
De Bilt
Washington
14PGE to App Transition
- The OMI SIPS is re-designed to the ACPS
(Atmospheric Composition Processing System) - Instead of developing L2 PGEs for the OMI SIPS,
scientists must now develop Apps that run on the
ACPS. - Main differences/advantages
- Can be run directly from TLCF or user
workstations. - Quicker delivery mechanism because it uses
automatic checking of the App structure. - Uses subversion (svn) instead of CVS as a code
repository. - A 3-6 month OSIPS/PGE -gt ACPS/App transition
period began in July.SIPS team does conversion
and training. - Existing and older PGE's will still be available
but source code in CVS will be read only.