Transcript Geolocation status (M Bates

Level 1.5 Processing Issues
Martin Bates
RAL
Level 1.5 Processing Issues
• Geolocation
–
–
–
–
Current Status
TSOL Jitter (CSOL-TSOL) correction
Longterm TSOL Jitter Solution
Comparisons with RGP Processing
• Upcoming Updates to Processing
• Processing Validation
Current Status
• Level 1.5: Adhoc E-W
geolocation correction
– sinusoid (amplitude, phase, offset
tunable)
– ESU dependent
– accurate to ~1 pixel, a bit worse NS.
• Level 2: fit to SEVIRI data
(formerly MS7A)
– decoupled from Level 1.5
– accurate to ~ 0.2 pixels (< 0.5
pixels)
09 Nov 2003
TSOL Jitter
• Predicted TSOL – actual TSOL
– as used by IMPF
– predicted TSOL (CSOL) from rotational model
• Temporary scheme
– data available on ftp server at EUMETSAT since 16
February
• coverage > 99.6 % (dropouts at sun-satellite colinearity)
– copy across and use as correction to E-W
geolocation
– ggsps correction for early April.
TSOL Jitter
data spikey –
needs filtering
14 Nov 2003
Smoothed TSOL Jitter
Result of median
filter ( 1 minute)
One spike in
close-up
Corrected Data (15 Nov 2003)
no filtering
am/pm Systematic
15 Nov 2003
TSOL Jitter Correction
23 March 2004
TSOL Jitter Correction
23 March 2004
am/pm systematic
still present
23 March 2004
offset slightly
different
Long term solution
• Robustness of ftp server solution in question.
• Longer term solution:
– append (TSOL Jitter) value to GERB packet
– correction algorithmically no different
– improved reliability and timeliness of information
• Timescale: 6-12 months.
– requires Engineering Change Proposal (for IMPF
and UMARF)
– still to be approved by EUMETSAT
• SEVIRI reprocessing ?
– can then get values for before 16 Feb.
RGP Comparisons
• Process data from 14-16 Nov 2003 using
TSOL jitter correction.
• Compare to RGP geolocation (high CPU
fit) comparing to SEVIRI.
North-South Compression
15 Nov 2004 11:56:58 (SW)
Still there !
N/S offset not perfect for these data
Quasi-Rotation
15 Nov 2004 11:56:58 (SW)
Also still there ...
Retune Optical Model ?
1. Could be due to local
deformations in one of
the mirrors – but no
estimate of possible
sizes of effects
2. Image ratio
plots  more
structure in
GGSPS plot
GGSPS
RGP
Geolocation Plans
• Interim TSOL Jitter Correction (April)
• Long term TSOL Jitter Correction
– for improved robustness
– still to be approved, end of year ?
• Optical Model – tune using RMIB info
• Axis misalignments
– MSG spin to MSG structure
• large (~0.3 ?)
• changes at each manoeuvre
– action on EUMETSAT to investigate
• Mirror Side Effects (non-parallel N/S ?)
Upcoming Processing Updates
• New BAT, BBRAD tables
– but may be coordinated with data release
• Offline gains (stray light & moon avoidance)
– depends on gain stability:
• if update < once per month
– use offline gains (data table)
• if update more frequently
– avoid problem regions or
– use gains calculated at good times of day
» startup problems & time dependent processing
• Stray Light Correction TBD
• Numerous minor updates
– NCR fixes
– additions (diagnostics) to NANRG product
– confidence flag definitions
Processing Validation
• Geolocation Validation
– Comparisons with SEVIRI / RGP geolocation (RMIB)
• Overall Validation
– series of unit and integration tests done (test data,
limited set of circumstances)
– step by step validation using real data and checkpoint
files suggested (validation plan) – TBD
– monitoring of calibration parameters using engineering
report (Imperial).