search

sunpy / sunpy-1.0-paper

The SunPy 1.0 Paper Repo
12 stars 21 forks source link

Various fixes to figures #100

Closed ehsteve closed 5 years ago

ehsteve commented 5 years ago

This depends on PR #96.

ehsteve commented 5 years ago

@ayshih predicted position of Venus is still off by a few arcsec. See the new plot.

ehsteve commented 5 years ago

I realize I made a mistake in the Venus angular extent calculation. Will fix soon.

ayshih commented 5 years ago

@ayshih predicted position of Venus is still off by a few arcsec. See the new plot.

The AIA file you download is a Level 1.0 file. If you run it through SSW's aia_prep – not SunPy's facsimile – to get to Level 1.5, this is the plot you get, which looks much better (although still slightly off): Figure_1 That is, SSW's aia_prep uses pointing information to correct for ~8 arcmin of roll misalignment.

ayshih commented 5 years ago

Actually, if I use get_horizons_coord() to get the SDO spacecraft position (different from the header position by 47 km), plus a minor bug fix for get_horizons_coord(), the agreement is even better: Figure_2

Either way, the discrepancy is <~0.5 arcseconds.

ehsteve commented 5 years ago

Well that’s annoying that you need aiaprep. Since you now have the prepped image could you PR the prepped fits file to the repo please @ayshih?

ayshih commented 5 years ago

Done (#101)

ayshih commented 5 years ago

Well that’s annoying that you need aiaprep.

Did some poking around, and the problem was specific to the file you were working with: it has out-of-date pointing information in its header. (In fact, it might be the case that this incorrect header reflects the state of knowledge prior to analysis of the Venus-transit data!) SSW's aia_prep ignores the pointing information in the header by default and instead accesses up-to-date pointing information. I just now grabbed the same exposure off of VSO, and the header provided has the up-to-date pointing information.