maho3/ltu-gobig

Test impact of lightcone effects using BORG-PM

Closed this issue · 3 comments

The goal of this task is to produce tests of gravity solvers with and without on-the-fly lightcones, using the CMASS NGC redshift slice of 0.4-0.7.

The implementation can be done by using the same initial field for BORG-PM and BORG-PM lightcone, and then measure summary statistics on the resulting DM density field.

If the differences are minor, then we can keep our current, single-snapshot implementation. However, if they are large (likely on the larger scale modes) then we need to implement lightcone versions of all the gravity solvers we have implemented.

Definition of done:

  • Plots of comoving DM Pk and other summaries with and without on-the-fly lightcones

@maho3 The lightcone in BORG-PM/tCOLA main was highly experimental. I can fix that fairly quickly though. Please bear in mind to ping me on this issue.

@leanderthiele has a lightcone code (https://github.com/leanderthiele/nuvoid_production) which extrapolates halo/galaxy positions across different redshift snapshots.

All it requires is halo pos/vel and HOD galaxy pos/vel at different snapshots. And they need not be linked across snapshots.

He found that it requires ~10 snapshots to converge the summary statistics across the CMASS NGC redshift range. We will try a test with the ltu-cmass, and compare to the simplified cuboid_remapping script

Closing issue on account of the fact that Leander's results implied lightcones are necessary, up to 10 snapshots. Adding an issue to implement this in ltu-cmass#21