Simulating LSST images to test analysis algorithms

Cosmologists will use LSST to explore the structure of the Universe.  The patterns of galaxies across the sky can tell us about the evolution and contents of the cosmos, the laws of gravity, and the Big Bang.  Our mathematical and computational models of physics can make predictions for what patterns we will see, and we can confront those with LSST observations.

To do this we need exquisite measurements of both the brightness of galaxies, at different wavelengths, and their shapes.  We use the brightness to model the distance to the galaxy, and the shapes to understand how they have been "lensed" - distorted by the gravitational fields between us and them.  With these two ingredients we can start to model the 3D structure of the "cosmic web", the network of galaxies we can observe.

Figure 1: a false-colour rendering of a small region of one of the large image simulations being generated by the LSST Dark Energy Science Collaboration to test the analysis algroithms that they will run on LSST data. UK researchers are participating in this simulation programme, using computing resources funded by STFC through the IRIS and GridPP projects. 

Doing this accurately enough will be an huge challenge, and to prepare for LSST we make galaxy simulations, like the one pictured above, so that we can test our algorithms and tools at the enormous data scale that the project will generate.