search

igmhub / picca

set of tools for continuum fitting, correlation function calculation, cosmological fits...
GNU General Public License v3.0
31 stars 22 forks source link

sub-sampling covariance is overestimated for picca_co #672

Open londumas opened 5 years ago

londumas commented 5 years ago

When computing the correlations involving two objects, e.g. auto correlation of DLA or of quasars, or their cross-correlation, the sub-sampling covariance is overestimated at small rt, when comparing to the mock-to-mock covariance.

The following plots are done using london mocks v9.9, with ~36 DLA auto-correlation. The autoc-orrelation of quasars has the same behavior

coDLA_4

coDLA_3

co_DLA_1

coDLA_1

londumas commented 5 years ago

On a side note, the agreement is very good for the auto, cross, autoLyb and crossLyb. The following two plots present the auto and the cross.

auto_compare_var cross_var