Adding GEO KOMPSAT-2A AMI and Himwari-8/9 AHI

[w-k-jones]

As both AMI and AHI share heritage with ABI, and now have data hosted on AWS, I thought it would be useful to add these datasets to goes2go so they can be accessed in the same manner as ABI data. I have a set of functions duplicating data.py for GK2A working, and am looking into loading AHI binary files using satpy, which should resolve #55. I just wanted to check with @blaylockbk if this is within the scope of goes2go and that you're happy for me to open a PR when these changes are ready.

Todo list:

• Add GK2A AMI download/reader functions
• Add Himawari-8/9 AHI download/reader functions using satpy
• Consolidate common download/read functions shared between GOES, GK2A and Himawari
• Add GK2A/Himawari classes in the same manner as the GOES class
• Add multi-channel L1b reader, conversion of sensor counts to reflectance/BT and creation of "pseudo-MCMIP" datasets from L1b to allow use of RGB accessors

[blaylockbk]

Yes! This is something I'm very happy to add. 😊 This could greatly expand the audience. (side note, are the meteosats available anywhere? I haven't checked. Would like mtg-i1 when it's available, too)

I don't have much time to dedicate to goes2go at the moment, but I will find time to review a PR.

I would request making any new features and dependencies optional, if possible.

Could you include in the Readme/docs something like "goes2go works with more than just GOES-R series data, it helps you get Himawari and GeoKompsat2a too!" with some examples.

[cmo84]

@w-k-jones See section 5.1 here in the GOES-R PUG (product user guide) if you need documention on what goes in the metadata fields while replicating some subset of them in the course of this pseudo-MCMIP production. I'm thinking you may need to add at least the x and y and some of the geolocation related fields to get the accessors working?

@blaylockbk The meteosat data is indeed available through the eumdac library and loadable via satpy. The current active satellites are Meteosat 10 (EO:EUM:DAT:MSG:HRSEVIRI) and Meteosat 9 (EO:EUM:DAT:MSG:HRSEVIRI-IODC). The only caveat is that these require an API key to be used, which is freely available when creating an account at https://eoportal.eumetsat.int/cas/

[vwgeiser]

@w-k-jones Have you made progress on this? I'd also be looking to add compatibility with at least Himawari 8/9. Are we still at Add Himawari-8/9 AHI download/reader functions using satpy ?

[cmo84]

@w-k-jones Have you made progress on this? I'd also be looking to add compatibility with at least Himawari 8/9. Are we still at Add Himawari-8/9 AHI download/reader functions using satpy ?

@vwgeiser There is already some built in capability to do this with the existing satpy package where you can wildcard the object keys and it will pull the files and unpack them into a user friendly scene object. You'd just have to put some datetime formatting around the creation of the wildcarded string to pull different times and bands. Looks like there may be some L2 cloud products like mask/height/phase that are available at different prefixes in the noaa-himawari9 bucket as well.

Relevant satpy readers
https://satpy.readthedocs.io/en/stable/api/satpy.readers.ahi_hsd.html
https://satpy.readthedocs.io/en/stable/api/satpy.readers.ahi_l2_nc.html

Prefixes

AHI-L1b-FLDK/
AHI-L1b-Japan/
AHI-L1b-Target/
AHI-L2-FLDK-Clouds/
AHI-L2-FLDK-ISatSS/
AHI-L2-FLDK-Winds/

Example of reading Band 14 data into scene object and displaying image

from satpy import Scene
storage_options = {'anon': True}
filenames = ['s3://noaa-himawari9/AHI-L1b-FLDK/2024/09/09/1700/*_B14_*.DAT.bz2']
scn = Scene(reader='ahi_hsd', filenames=filenames, reader_kwargs={'storage_options': storage_options})
scn.load(['B14'])
scn.show('B14')

I couldn't get it to load the L2 products the same way with the wildcarded key but it's likely possible.

Disclaimer: I'm not associated with satpy at all, just a user.