search

duncandc / intrinsic_alignments

project to measure and model the intrinsic alignments of galaxies in both observations and simulations
2 stars 1 forks source link

account for handedness in the implementation of the "hybrid" satellite alignment model #3

Closed duncandc closed 6 years ago

duncandc commented 6 years ago

when aligning with the linear combination of the radial and major axis vector in a halo, a galaxy may be aligned or with a rotated vector that depends on the combination. e.g. when the alignment vector is 0.5radial + 0.5major_axis and the galaxy is located perpendicular to the major axis, the galaxy may be aligned or rotated by 90 degrees.

@c-d-leonard suggested this is related to the treatment of shear in lensing analysis.

jablazek commented 6 years ago

@duncandc : I'm trying to picture this scenario. Is this a case where gamma_x (measured wrt to the main halo center) could be positive or negative, and the physical model has no way to distinguish between the two?

duncandc commented 6 years ago

I have a crude pictorial representation of a this hybrid model in the IA model demo notebook here: https://github.com/duncandc/intrinsic_alignments/blob/master/ia_models/IA%20Modelling%20Demo.ipynb

I think what you said is correct. I think I have a reasonable method to implement it, so I am somewhat more worried about directly constraining it in simulations.

jablazek commented 6 years ago

When there is an ambiguity like that it is a hint that something non-physical is going on. We usually assume that delta correlated with gamma_x (and similar statistics) must be zero due to this sign flip under a parity transformation. Unless the universe has a preferred handedness...

duncandc commented 6 years ago

ahhh, no that is right. sorry, gamma_plus could be positive or negative. gamma_cross should still be zero here.

jablazek commented 6 years ago

As drawn, the perp located satellites have non-zero gamma_x. You mean that, on average, they should have zero gamma_x?

duncandc commented 6 years ago

yes. I can solve this by stochastically aligning satellites such that gamma_cross should be zero.

duncandc commented 6 years ago

I updated the cartoon in the notebook to reflect this.