You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Which for very small angles is ~ SQRT( delta(ra)^2 + delta(dec)^2 )
while in the equirectangular approximation it is ~ SQRT( (delta(ra) * cos(dec1+dec2)/2 )^2 + delta(dec)^2 )
The missing factor of the cosine of the declination yields to an overestimated distance away from the equator.
My impression is that the Gnomonic projection is not suitable in astronomy and is more related to geophysics.
This leads to an actual loss of cross-correlated sources at high latitudes.
The text was updated successfully, but these errors were encountered:
Hi Folks: we're extremely interested in using AXS on a code-to-data platform for Gaia science, but see that some issues have been identified with the spherical trig for computing correctly the matches as one moves away from the equator, or across the RA 0h boundary etc. Are there any plans for pulling in the fixes being worked on by @stargaser and making a new release? Let us know if we can help with testing/validating any branches/forks.
Hi, yes I'm working on the distance computation issue now. I've also brought the code up to spark 3.0, so we will make a new release based on that once the distances are sorted out. Thanks for the offer to help with testing; I'll post back here once we have a release candidate.
We have found an issue with the distance computed by calcGnom for catalogue cross-correlation in AXS :
axs/AxsUtilities/src/main/scala/org/dirac/axs/FrameFunctions.scala
Line 14 in fa3c874
Which for very small angles is ~ SQRT( delta(ra)^2 + delta(dec)^2 )
while in the equirectangular approximation it is ~ SQRT( (delta(ra) * cos(dec1+dec2)/2 )^2 + delta(dec)^2 )
The missing factor of the cosine of the declination yields to an overestimated distance away from the equator.
My impression is that the Gnomonic projection is not suitable in astronomy and is more related to geophysics.
This leads to an actual loss of cross-correlated sources at high latitudes.
The text was updated successfully, but these errors were encountered: