modis_GEO.py for real-time TERRA

By jasmine Date 2012-07-23 19:33

Could you tell me the correct convention to call modis_GEO.py when processing real-time TERRA data? I'd like to use the embedded onboard real-time attitude ephemeris.

I have tried the following:

$OCSSWROOT/run/scripts/modis_GEO.py --disable-download --log -d -v $l1a_file
(where the L1A file is a real-time file with no available attitude and ephemeris files on the oceancolor site)

This results in a PCF with
600280|Source for spacecraft kinematic state|MODIS Packet (as expected)
and the lines 10501 and 10502 are missing (the att/eph lines)

The error I'm getting is:

/home/swdev/ocssw/build/src/geogen_modis/GEO_ephem_attit.c, GEO_interp_ephemeris_attitude():MODIS_W_SAMP_TIME_OOR:288779331
Mon Jul 23 18:50:20 2012
Sample time 617222145.636005 for samp = 1 is more than 2.048 seconds outside the range from 617221532.260000 to 617222143.588000

(repeated many times)
then

PGS_PC_GetNumberOfFiles():PGSPC_W_NO_FILES_FOR_ID:76811
The Logical ID (10501) does not contain a physical File Name.

PGS_EPH_GetEphMet():PGSEPH_E_NO_SC_EPHEM_FILE:44547
error in accessing spacecraft ephemeris file(s)

/home/swdev/ocssw/build/src/geogen_modis/GEO_write_granule_metadata.c, GEO_write_granule_metadata():MODIS_W_GEO_EPHEMERIS:288779324
Mon Jul 23 18:50:20 2012
No s/c ephemeris/attitude files could be found for input time 2012-07-23T18:25:24.306930Z

PGS_PC_GetNumberOfFiles():PGSPC_W_NO_FILES_FOR_ID:76811
The Logical ID (10501) does not contain a physical File Name.

/home/swdev/ocssw/build/src/geogen_modis/GEO_write_input_metadata.c, GEO_write_input_metadata():MODIS_E_GEO:288779785
Mon Jul 23 18:50:26 2012
Error returned by function PGS_PC_GetNumberOfFiles(10501)

PGS_PC_GetNumberOfFiles():PGSPC_W_NO_FILES_FOR_ID:76811
The Logical ID (10502) does not contain a physical File Name.

Thanks!
Jasmine.

By gwyn Date 2012-07-24 17:06

Hi, Jasmine -

Only predicted ephemeris is included in the onboard data; you'll still need to have at least the predicted attitude file (AM1ATTNF).
See this thread for more explanation.

It is possible to process geolocation data even without that; geogen_modis writes a valid file but exits with an error. Geolocation values will be off by a small amount, as you'd expect - except in the last scan, where differences are large.

Also, certain metadata fields are set incorrectly or not at all:
AUTOMATICQUALITYFLAG="Failed"
EQUATORCROSSINGDATE, EQUATORCROSSINGTIME, ORBITNUMBER, EQUATORCROSSINGLONGITUDE all set to default values

By gwyn Date 2012-07-25 19:35

I was mistaken; the Terra data stream includes attitude data as well.

The MODIS geolocation code is developed by another group, then tweaked for our use. They always use external attitude & ephemeris files, so the code paths using entrained data aren't well tested. Returning a fail status for "missing" att/eph files, and then not writing metadata correctly, are areas where the code needs improvement.

Further, the modis_GEO.py script does not allow use of the Terra entrained data when the att/eph files are available in our database. That's something else that needs improvement.

Thank you (and Liam Gumley) for identifying these issues. Thanks also for actually looking at the pcf and log files before posting - few people do!