Skip to main content
European Commission logo print header

Capitalizing on Gravitational Shear

Final Report Summary - COGS (Capitalizing on Gravitational Shear)

My project sought to develop shear measurement tools in collaboration with computer scientists to allow my team to play a central role in the key results of the Dark Energy Survey. The specific objectives of the funded part of the grant are listed below, with a bulleted summary of the outcome for each.

(i) set future GREAT challenges and interact further with the computer science community
* CoGS PDRA Rowe co-ran the third GREAT Challenge. This involved writing the GREAT3 Challenge Handbook (Mandelbaum et al 2014) and writing a piece of software now widely used internationally, called GalSim (Rowe et al 2015).

(ii) successfully enter future GREAT challenges by developing new shear estimation methods in collaboration with computer scientists;
* We wrote a new shear measurement code from scratch, im3shape, which is now publically available. We tested this shear measurement code on simulated images of the Dark Energy Survey and participated in the GREAT10 shear measurement challenge which ran from December 2010 until August 2011 (Kitching et al 2011; Kitching et al 2012).
* For the first time we described a new shear measurement bias on galaxy model-fitting methods which arises due to the non-linear relation between noise on an image and the measured shear (Refregier et al 2012). We quantified the size of this bias on realistic images and found it to be very significant if uncorrected, and we demonstrated a successful calibration scheme inspired by the approach taken to this problem in the statistics literature (Kacprzak et al 2012).
* We also entered the third GREAT Challenge (Mandelbaum et al 2015) which involved significant code development for multi-epoch exposures and improve shear calibration techniques.
* We published on our new shear measurement code developed in the COGS project: im3shape (Zuntz et al 2013). We showed that noise and model bias (see previous reporting period) are additive to a very good approximation (Kacprzak et al 2013).

(iii) produce a reliable shear catalogue from the Dark Energy Survey that can be used to constrain the nature of dark energy or modifications to gravity.
* In reporting period 3 we spent a significant amount of time working on the first Dark Energy Survey Science Verification period data. This resulted in the first publication of weak lensing results from the Dark Energy Survey (Melchior et al).
* In the final reporting period we delivered a high quality shear catalogue to the DES Collaboration on the Science Verification area, which formed a bedrock of the publication Jarvis, Sheldon, Kacprzak, Zuntz, Bridle, et al 2015 describing the DES shear catalogues. The shear catalogues are now publically available. We carried out high quality tests of the shear catalogue at the two-point level which are described in detail in Becker, Troxel, MacCrann, et al 2015.

(iv) measure intrinsic alignments from WiggleZ galaxies
* We set upper limits on the intrinsic alignments of blue galaxies at relevant redshifts (WiggleZ) for the first time in Mandelbaum, Blake, Bridle et al 2010.

(v) measure intrinsic alignments of red galaxies using photozs for the first time using MegaZ-LRG
* For the first time, we detected the galaxy intrinsic alignment signal in galaxies at distances similar to those probed by upcoming gravitational shear surveys, using a new technique we developed using approximate galaxy distances (Joachimi et al 2010).

(vi) develop a halo model for intrinsic alignments
* We updated the most popular model for intrinsic alignments and graphically showed the impact of ignoring them or using the wrong model (Kirk et al. 2012). We showed that the dark energy equation of state could easily be measured incorrectly by many times the statistical precision of the upcoming surveys.

(vii) use realistic photometric redshift distributions in cosmology from cosmic shear in the presence of intrinsic alignments
* In MacCrann et al we re-analysed the CFHTLenS cosmic shear constraints using a more sophisticated model for intrinsic alignments and using their real photometric redshift distributions.
* In the final reporting period we verified the photometric redshift catalogues, in Bonnett, Troxel et al 2015. We used these real photometric redshift distributions in the first cosmological constraints from DES, from cosmic shear, marginalizing over intrinsic alignments (see objective (x) below).

(viii) develop lensing cluster selection into tomographic peak statistics on DES
* In the final reporting period we collaborated with former COGS member Kacprzak to constrain cosmology using shear peak statistics in DES SV data.

(ix) combine galaxy-shear and galaxy position correlations with cosmic shear
* We demonstrated that deviations from General Relativity can be measured even when simultaneously mitigating the effect of intrinsic alignments (Laszlo et al 2012) and that surveys designed to measure dark energy are also well optimized for measuring deviations from General Relativity (Kirk et al 2013).

(x) obtain cosmological constraints from cosmic shear in DES
* We developed the cosmological analysis framework CosmoSIS (Zuntz et al 2015) which is now used widely in DES, LSST and beyond.
* The work in the final reporting period went smoothly and culminated in the publication of the first cosmology constraints from the Dark Energy Survey (DES). We led the first cosmology paper from the Dark Energy Survey (The DES Collaboration et al 2015) which has an alphabetical author list, but the two corresponding authors are COGS team members Zuntz and MacCrann.