Incorrect refraction elevation cutoff in ephemeris solar position algorithm

[AdamRJensen]

The refraction correction used in the ephemeris solar position algorithm has a cutoff at -1 degree, which is not mentioned in the original publication or the provided Fortran code:

pvlib-python/pvlib/solarposition.py

Line 839 in 6d886dc

Refract[(Elevation > -1) & (Elevation <= -0.575)] = -20.774 / TanEl

To be consistent with the publication I suggest that we remove the lower limit and apply refraction correction to all negative solar elevation angles.

[AdamRJensen]

ICYMI @IoannisSifnaios @kandersolar

[cwhanse]

The reference publication is here: https://pvpmc.sandia.gov/app/uploads/sites/243/2022/10/Engineering-Astronomy.pdf

It doesn't mention correction below -0.575 degrees, that I can find.

My guess is that the -1 degree limit was added to the Matlab code because we didn't think refraction correction was meaningful when the sun's disc was below the horizon.

I have no objection to removing the limit, or adding a comment pointing back to the Matlab code. I also have no objection to removing the ephemeris function as long as there is an equally fast solar position algorithm in the library.

[kandersolar]

+1 to removing ephemeris eventually, and leaving it as-is in the meantime. But good to document the issue regardless.

A bit of context: @AdamRJensen, @IoannisSifnaios, and I have a project underway comparing solar position algorithms. We plan to contribute implementations of one or more algorithms to pvlib-python, depending on the results of the comparison. We expect some of them to be better than ephemeris in all respects (speed, accuracy, period of validity, clear reference).