Himawari 8/9

[craigim]

Japan has a nearly identical satellite to GOES covering the Pacific. The data is in AWS, managed by NOAA. How difficult would it be to add in the ability to pull data from that bucket as well?

https://aws.amazon.com/marketplace/pp/prodview-eu33kalocbhiw?sr=0-1&ref_=beagle&applicationId=AWSMPContessa

[blaylockbk]

Hi @craigim,
This is a nice idea. I am aware of the Himawari data. I have never looked at that data, probably because it looks like the data is in binary format and not NetCDF.

Adding this capability to goes2go could be a useful feature, but it would take some effort. I make several assumptions in the goes2go.data.py file about the GOES data that don't apply to the Himwari dataset. It might be best to start from scratch and make a goes2go.data_himawari.py file and a HIMAWARI class that has the same methods as the GOES class.

Other things that would need to be addressed:

• The directory structure is different; AHI has different products than ABI and different file names.
• Himawari files are in a binary format. I would like to see a reader to read the files into an xarray DataFrame.
• The ABI channels are different and the RGB recipes would need customization. For example, AHI has a "green" channel while ABI does not.
• Does AHI have a "multichannel cloud moisture imagery" product like ABI has?

[craigim]

Exploring the AWS link, they do have a directory in which NOAA has put the data into netCDF format, although it's tiled which makes it annoying (noaa-himawari8/AHI-L2-FLDK/ISatSS). You would think that since NOAA is already reformatting the data from the original binary data, they'd make it as close as possible to the GOES format, but here we are.

The bands between the AHI and ABI instruments are effectively the same except that Himawari swaps a 0.51 micron green band for the GOES 1.37 micron cirrus band, so you can use the same recipes that directly use the bands. The ones that use derived or higher level imagery would have to be excluded or reformulated, though.