Metop-B coastlines offset

[lobsiger...@gmail.com]

Hi all,

Metop-B coastlines from France and UK seem shifted. Metop-C and NOAA-20 coastlines look good. TLEs updated,  SatPy Version 0.36. Do Metop coastlines depend on TLEs (and my Metop-B TLEs are kind of bad) while NOAA sats have lon/lat per pixel (or at least per tie-point) ? Any other ideas?

TBH I havn't noticed that before as I have only taken Aqua/Terra and NOAA-20/Suomi-NPP passes over tha UK up to now.

Best regards,

Ernst

[Raspaud Martin]

Hi Ernst!

That is the data format of your Metop-B data?
Processing the data through AAPP for example depends on the TLEs for
sure to generate the pixel positions.

I can recommend using the TLEs provided by eumetsat for the metops, eg
https://service.eumetsat.int/tle/data_out/latest_m01_tle.txt they seem
to work well for us.

Anyway, if the pixel positions are bad, than the resampling will be
off, and then the coastlines (which do not rely on the tles) will be
off too.

Best regards,
Martin

[lobsiger...@gmail.com]

Hi Martin,

the data is from EUMETCast channel E1B-EPS-10. It has to be read with the Satpy 'avhrr_l1b_eps' reader.

When I first saw the coastline offset I updated my TLEs to no avail. I now changed the Metop-B TLEs

with the two lines from EUMETSAT you pointed me to. I still see no difference.

Does this mean there is a problem with the Metop-B EUMETCast data?
Is that something that should maybe reported to EUMETSAT OPS?

Regards,

Ernst

[Raspaud Martin]

Hi Ernst,

Yes, in this case the data is already geolocated so the tles you have
won't help.

It points indeed to a problem in the data, so yes, you can definitely
report it to OPS.

Best regards,
Martin