How accurate is the redshift conversion of AstroPy.coordinates.Distance function?
It appears to be useful only to the thousandths digit (much less precise than floating point number precision issues):
from astropy import units as u
from astropy.coordinates import SkyCoord, Distance
from astropy.cosmology import Planck15
z1 = 0.05598
z2 = 0.31427
dist1 = Distance(unit=u.pc, z = z1, cosmology = Planck15)
dist2 = Distance(unit=u.pc, z = z2, cosmology = Planck15)
dist1.z #prints 0.05718
dist2.z #prints 0.31916
I am using this to compute 3D distances between extragalactic sources, and these discrepancies are on the order of Mpc, which is very large for what I am studying. Is this an unavoidable limitation of AstroPy?
This works in astropy 3.2.1 for python 3.7.
from astropy import units as u
from astropy.coordinates import SkyCoord, Distance
from astropy.cosmology import Planck15
z1 = 0.05598
z2 = 0.31427
dist1 = Distance(unit=u.pc, z = z1, cosmology = Planck15)
dist2 = Distance(unit=u.pc, z = z2, cosmology = Planck15)
dist1.z
Out[9]: 0.055979999974738834
dist2.z
Out[10]: 0.31427000077974493
It looks like the calculation has precision to the around 7 significant digits.
z3 = 1.31427987654321
dist3 = Distance(unit=u.pc, z = z3, cosmology = Planck15)
dist3.z
Out[23]: 1.3142798808605372
z4 = 900.31427987654321
dist4 = Distance(unit=u.pc, z = z4, cosmology = Planck15)
dist4.z
Out[29]: 900.3142861453044
Somewhere close to z=1000 this will return an error saying the value is maxed out, since at that point you're getting close to CMB territory.