ssec / polar2grid

Tools for reading, remapping, and writing satellite instrument data.
http://www.ssec.wisc.edu/software/polar2grid/
GNU General Public License v3.0
72 stars 34 forks source link

Default daytime band products should be reflectances #707

Closed kathys closed 1 month ago

kathys commented 3 months ago

Our boss Liam has requested that Polar2Grid create actual reflectances that include division by the cosine of the solar zenith angle for each of the polar orbiting satellites that we support, including AVHRR and MODIS. This should be the default daytime reflectance band output.

djhoese commented 3 months ago

What about geo2grid readers?

kathys commented 3 months ago

I will speak to Graeme about the defaults for G2G.

djhoese commented 3 months ago

My plan for implementing this would be to make "aliases" for all channels requested from the command line that are visible channels. In Satpy terms, they would be an alias for DataQuery(name="<channel name>", modifiers=("sunz_corrected",)). This should mean that any existing composites will not be affected.

In normal software versioning terms, I'm tempted to say this is almost a "major" change deserving of a 4.0 release but I suppose it doesn't have to be. Just something that popped in my head when thinking about this last night.

kathys commented 3 months ago

I understand why you would say this. I think it will also add processing time too. But I think it is the right thing to do, and Liam is right in that our users, especially in the polar regions, will get uniform images across sensors. An important part of this change will be to make sure that I document it well for end users.

djhoese commented 3 months ago

Any idea if there are any readers (AVHRR?) where we don't have access to solar zenith angles and they would have to be computed on-the-fly from the lon/lats? That would take the processing time hit from insignificant to extremely annoying.

djhoese commented 2 months ago

Just to have it written down for future reference, the readers affected by this are:

Any others? Did we ask Graeme about Geo2Grid?

djhoese commented 2 months ago

I should note that for readers/files where SZA is not available at every resolution, there will be a required resampling step that will now take longer because of needing to resample not just band data but also the SZA. One example is mersi2 where SZA is available at 1km but some bands are available at 250.

kathys commented 2 months ago

Understood.

djhoese commented 1 month ago

@kathys I have this functionality, at least the initial draft, implemented in #714. I know you're busy so you probably don't have time to test it but I'm wondering if Liam wants to? If neither of you have time then I might wait to merge the pull request in case other stuff comes up that we want to implement as part of it. Or I could merge it and get you a tarball just to have it when either of you is ready.

kathys commented 1 month ago

Thanks very much Dave. Please provide a tarball. Liam might want to start testing/using it right away.