Penn became a collaborating institution in DES in 2007, including a substantial institutional financial commitment and a commitment to develop several key elements of the science software pipeline. Professor Jain and Dr. Jarvis have made significant contributions to the weak lensing program and its connections to image quality and data management—they have worked with DES since Spring 2005, well before Penn formally became a member. Professor Bernstein, Professor Sako, postdoc T. Qian and other Penn personnel are ramping up contributions to the project.

DES proposes to build a wide-field camera for the existing 4-meter Blanco Telescope in Chile and execute an imaging survey in four filters of _ 5000 square degrees. The survey will provide four complementary probes of dark energy: galaxy clusters, baryon oscillations, supernovae and weak lensing. It will be carried out between 2011-2016. The current activities of most relevance to the Penn effort are: the development of the software pipeline; the lensing science requirements on image quality and therefore camera and telescope design; interface of lensing modules with the data management system; and the development and testing of the lensing and Supernovae science codes.

Jarvis, Bernstein and Jain obtained dark-energy constraints (Jarvis et al. 2006) from an end-to-end analysis of a 75 square degree imaging survey carried out with the Blanco telescope on CTIO. Since this is the telescope to be used for DES, the DES collaboration sought our input to understand and reduce PSF (point spread function) systematics from this telescope. The analysis and forecasts we have made (described below) helped make the case for the lensing program of DES, as reflected in the DES white papers submitted to the Dark Energy Task Force and the DES proposal to DOE/NSF in the last year.

For lensing with DES (and even more for LSST), potentially the largest systematic error arises from interpolation of the anisotropic PSF. Jarvis and Jain’s principal component analysis (PCA) method solves this problem, not just for existing data, but scaling up to the survey size planned for DES. It does so by optimally utilizing information about the PSF measured from stars in all survey exposures to better estimate the PSF at the galaxy locations on each individual exposure. With increasing survey size and number of exposures, this technique is able to use more PSF measurements for the correction, leading to lower residuals. It is the only existing method and software that tackles the critical problem of PSF interpolation for DES sized datasets at the precision that is required.