Using the new classmethod "assume_spherical_screen" in Helioprojective in conjunction with solar differntial rotation correction

[Chris Bethge]

Hi,

I was wondering if someone could help me to use the new classmethod "assume_spherical_screen" in Helioprojective in conjunction with the solar differential rotation correction described in this example here:

https://docs.sunpy.org/en/latest/_downloads/fa31464c65aca307090b96037f642dd1/reprojected_map.py

The background: we are using the differential rotation correction for SUVI, but since the off-disk parts do not have a HeliographicStonyhurst coordinate, they get set to NaN in the output image. We do need the off-disk parts for context though, so right now we are clumsily puzzling the on-disk and off-disk parts back together. From what I understand, the "assume_spherical_screen" method could mitigate this problem, but I don't understand how to combine it with the reproject example above. When I just try a simple

with Helioprojective.assume_spherical_screen(suvimap.observer_coordinate):
    with transform_with_sun_center():
        arr, _ = reproject_interp(suvimap, out_wcs, out_shape)

I only get an odd hourglass shape of the data, with the rest set to NaN.

Any help greatly appreciated!

Cheers,

Chris

[Albert Y. Shih]

Hi, Chris,

     Heh, I've never actually tried using those features simultaneously!

     I see what's going awry in the code, but I'm not sure whether I can easily fix it.  Essentially, `assume_spherical_screen()` assumes that the screen is spherical in both the source and destination coordinate frames, which is the case for a simple change in observer location.  However, once you mix in differential rotation, the true screen does not stay spherical across the transformation, and consequently the reprojection code rejects nearly all of the pixels (aside from pixels where the discrepancy is minor).

     Feel free to create a GitHub issue to track this, as I mull over possible fixes.

Thanks,

Albert

--
You received this message because you are subscribed to the Google Groups "SunPy" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sunpy+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sunpy/0354aa9a-915e-4434-976d-5029dbc5daean%40googlegroups.com.

[Chris Bethge]

Hi Albert!

Thanks for looking into this already! No worries, I was just wondering if I did something wrong when using it. It would be great if we could get this working at some point - when I read the release notes for the new SunPy version, I thought "Yay, that's exactly what we need!". Anyway, we can live with the "crutch solution" for SUVI for now. Once your solution is working, I definitely want to implement this properly in the SUVI pipeline. I might write a GitHub issue against it next week, just to not lose track on it, as you said. Other people might try and do the same thing with AIA, who knows.

Cheers and have a good weekend,

Chris