STEP 2: Shapelet Simulations
STEP home Simulations Data Mailing List

The second STEP project assessed the impact of complex galaxy morphologies on weak lensing analyses using a large set of Shapelet simulations from Richard Massey. To reduce shape noise, these simulations consisted of two sets; an original set and a rotated set. In the rotated set, galaxies in the original set of images were rotated by 90 degrees before the shearing and PSF smearing was applied. Please feel free to use these simulations to optimise your cosmic shear pipelines, including acknowledgements to Richard Massey in any publications.

STEP2 simulations be downloaded from, along with the input shears and catalogues. Information about the shapelet simulations can be found here.

The results of the STEP2 analysis are online and detailed in the STEP2 publication . The minutes of the April STEP2 telecon, where these results were discussed, can be found here .

Rules for the analysis

  • All simulations have constant shear and constant PSFs and also have the same set of galaxies.
  • These measures have been taken to speed up the production of the simulations and also to simplify the final comparison analysis. Knowing that the simulations have these properties is however very different to what one would know about a real data set!
  • Therefore, please do not use extra information about the properties of each image simulation in your analysis and run your pipelines as you would normally on a set of real data.
  • Anisotropic PSF modelling should be performed on each starfield simulation image . Spatial stability of the PSF across the image should not be assumed.
  • In the cases where the PSFs are known to be pure e1 or pure e2, both components should still be measured from the simulations as you would with real data.
  • Galaxy positions should be determined from each simulation image with your normal source detection software, not from a composite catalogue.
  • Please do not use your prior knowledge that the new set of images are rotated versions! i.e. do not look at the images in pairs, compare galaxies with their rotated counterparts, or obtain a population ensemble from a catalogue combining original and rotated galaxies.
  • If you've used other short-cuts that use prior knowledge about the simulations, and feel they are valid, let me know with a short explanation.

Last modified 18th Jan.
STEP pages maintained by Catherine Heymans: heymans[at]