This repository containts the results of the paper "UNIONS: The impact of systematic errors on weak-lensing peak counts".
Authors: Emma Ayçoberry, Virginia Ajani, Axel Guinot, Martin Kilbinger, Valeria Pettorino, Samuel Farrens, Jean-Luc Starck, Raphaël Gavazzi, Michael J. Hudson
The Ultraviolet Near-Infrared Optical Northern Survey (UNIONS) is an ongoing deep photometric multi-band survey of the Northern sky. As part of UNIONS, the Canada-France Imaging Survey (CFIS) provides r-band data with a median seeing of 0.65 arcsec, which we use to study weak-lensing peak counts for cosmological inference. This work aims to assess systematics effects for weak-lensing peak counts and their impact on cosmological parameters for the UNIONS survey. In particular, we present results on local calibration, metacalibration shear bias, baryonic feedback, the source galaxy redshift estimate, intrinsic alignment, and the cluster member dilution. We expect constraints to become more reliable with future (larger) data catalogues, for which the current pipeline will provide a starting point. This paper investigates for the first time with UNIONS weak-lensing data and peak counts the impact of the local calibration, residual multiplicative shear bias, redshift uncertainty, baryonic feedback, intrinsic alignment and cluster member dilution. The value of matter density parameter is the most impacted and can shift up to ~ 0.03 which corresponds to 0.5 sigma depending on the choices for each systematics. We expect constraints to become more reliable with future (larger) data catalogues, for which the current code provides a starting point.
To be able to run the example, the following python packages should be installed with their specific dependencies in a python==3.6 environment.
You can follow the instructions to install Anaconda here or miniconda here and create an environment as
conda create --name <your_env> python==3.6.13
then install the following packages:
- numpy
- astropy
- lenspack
- scipy
- scikit-learn==0.20
- matplotlib
- emcee==3.1.1
- joblib
- tqdm
- chainconsumer
- getdist
- zenodo_get
- jupyter
Note about redshifts: as we found in the paper that the change in the peaks ditribution (and on the corresponding constraints) between redshift z=0.65 and z=0.68 is not significant, we provide here only the arrays at z=0.65. If interested in the distributions at z=0.68 please contact emma.aycoberry@iap.fr or vajani@phys.ethz.ch.
-
List of the peaks from the simulations at redshift z=0.65: list_cosmo_peaks_z065.txt
-
Peaks distribution coming from simulations: these need to be downloaded from here, DOI: 10.5281/zenodo.6344515. This step is already present in the example notebook where you create a folder called peaks_z065/ in the repo where the peaks distribution coming from simulations have to be stored and download the peaks as:
zenodo_get 10.5281/zenodo.6344515in the folder
input/peaks_z065 -
Baryonic correction: Fid_correction.npy, HighAGN_correction.npy and LowAGN_correction.npy
-
Peaks from CFIS-P3 data with different parameters: peaks_mean_global.npy = peaks obtain with the global calibration, peaks_mean_Xdeg.npy = peaks obtain with the calibration on X square degree, peaks_mean_dm_1deg.npy = peaks obtain with the calibration on 1 square degree, and the multiplicative shear bias delta m = 0.007
-
Peaks for the covariance: convergence_gal_mnv0.00000_om0.30000_As2.1000_peaks_2arcmin_0.65_b030_snr_min_max_ngal_7.npy
-
param_z: the redshift of the desired simulations Default is ’065’ (if interested in the distributions at other redshifts, please contact emma.aycoberry@iap.fr or vajani@phys.ethz.ch.)
-
param_z_cov: the redshift of the desired simulations, different of the previous file to respect the name given by the simulations Default is ’0.65’
-
param_cal: the way the calibration is done Can be ‘global’ (global calibration), ’05’ (calibration on 0.5 square degree), 1’, ‘2’, ‘4’, ‘dm_1’ (calibration on 1 square degree, with the multiplicative shear bias delta m = 0.007)
-
param_baryonic_correction: which baryonic correction we want to apply Can be ‘no’ if we don’t want to correct the simulations, or ‘Fid’, ‘HighAGN’, ‘LowAGN’ to correct the simulations
-
param_cut: up to which index of SNR we are going Can be 19 if we want to cut the last bins or 30 if we want to keep all the bins.
-
Fig. 7: test the calibration size param_z = '065' param_z_cov = '0.65' param_cal = ‘global’ or ’05’ or ‘1’ or ‘2’ or ‘4’ for the different cases param_baryonic_correction = 0 param_cut = 30
-
Fig. 8: test the residual shear bias param_z = '065' param_z_cov = '0.65' param_cal = ‘global’ or ‘1’ or ‘dm_1’ for the different cases param_baryonic_correction = 0 param_cut = 30
-
Fig. 11: test the baryonic correction param_z = '065' param_z_cov = '0.65' param_cal = ‘global’ param_baryonic_correction = 0 or ‘LowAGN’ or ‘Fid’ or ‘HighAGN’ for the different cases param_cut = 30
-
Fig. 12: test the intrinsic alignment and boost factor param_z = '065' param_z_cov = '0.65' param_cal = ‘global’ param_baryonic_correction = 0 param_cut = 30 or 19 for the different cases
-
Fig. 13: ideal model vs conservative model Ideal model: param_z = '065' param_z_cov = '0.65' param_cal = ‘global’ param_baryonic_correction = 0 param_cut = 30
Conservative model: param_z = '065',
param_z_cov = '0.65',
param_cal = ‘dm_1’,
param_baryonic_correction = 'Fid',
param_cut = 19,
To run the example you need to check that you have propery installed the Dependencies listed above in your environment and then git clone this repository as
git clone https://github.com/CosmoStat/shear-pipe-peaks.git
ad then go to the example folder and launch the jupyter notebook constraints_CFIS-P3.ipynb by doing:
cd example
jupyter notebook
and run it following the instructions in the description of the cells.
Follow the Configuration needed to obtain the different figures of the paper to get the results of the paper or play with the different input parameters to get different configurations.
Contact vajani@phys.ethz.ch or emma.aycoberry@iap.fr for questions concerning the repo.