In the HIRS FCDR, I'm noting the time per scanline in "seconds since ...", and use an integer for it. This is problematic, because a HIRS scanline lasts 6.4 seconds, and the add_offset essentially indicating the start of the orbit time is in milliseconds, as is orbit_start_time (which should be called time_coverage_start according to #294). I should probably have a scale_factor of 10 to the time, so I can store the time per scanline in units of 0.1 second. Needs discussion with @TomBlock and the other FCDR sensor teams as I believe we had a project decision on this, but looking at the FCDR file format specification, the microwave and AVHRR use a double with seconds since 1970, not a scaled integer.
gerritholl
commented
5 years ago
In the HIRS FCDR, I'm noting the time per scanline in "seconds since ...", and use an integer for it. This is problematic, because a HIRS scanline lasts 6.4 seconds, and the add_offset essentially indicating the start of the orbit time is in milliseconds, as is orbit_start_time (which should be called
time_coverage_startaccording to #294). I should probably have ascale_factorof 10 to the time, so I can store the time per scanline in units of 0.1 second. Needs discussion with @TomBlock and the other FCDR sensor teams as I believe we had a project decision on this, but looking at the FCDR file format specification, the microwave and AVHRR use a double with seconds since 1970, not a scaled integer.