Abstract
The Dark Energy Survey Data Management System: The Coaddition Pipeline and PSF Homogenization
J. Anthony Darnell (Astronomy Dept. Univ. of Illinois)
Emmanuel Bertin (Institut d' Astrophysique), Eric H. Neilsen (Fermi National Accelerator Laboratory), Chow-Choong Ngeow (Univ. of Illinois), Joseph J. Mohr (Univ. of Illinois), Darren Adams (Univ of Illinois), Cristina Beldica (Univ of Illinois), Y. Dora Cai (Univ of Illinois), Gregory E. Daues (Univ of Illinois), Michelle Gower (Univ of Illinois)
The Dark Energy Survey (DES) is an optical survey of 5000 deg2 of the South Galactic Cap to ~24th magnitude in multiple filter bands using a new wide field CCD camera, DECam, mounted on the Blanco 4-m telescope at Cerro Tololo Inter-American Observatory (CTIO).
The DECam is a large focal plane array with a short readout time which will collect approximately 300 GB of science images per night. These images are collected from an array of 62 CCDs that will be combined, after processing, to build sets of square, 1 degree tiles to cover the entire survey region. This process of combining multiple images of the same region of sky (and of the same band) into deeper images is called coaddition. The resulting 5000 coadd images, one for each square degree of sky, are arranged on a predefined grid and periodically reprocessed throughout the life of the project.
Because of changing observing conditions from exposure to exposure, combining frames of different quality leads to a PSF that varies discontinuously over the coadd image. This affects star galaxy separation and contributes to variation in the completeness at a given photometric depth as a function of position in the image. In addition to PSF variation, there is a inhomogeneous noise field that results from the coaddition of many independent, partially overlapping exposures of a given portion of the sky.
Here, we introduce our method of reducing PSF variation effects by processing the images to bring them to a common PSF within an image and from image to image within a coadd tile. This is accomplished by applying position dependent smoothing kernels determined using power spectrum weighting functions that adjust the relative contributions of large scale and small scale power within an image in such a way as to bring the PSFs within and among the image samples into agreement. This approach has been applied successfully in image differencing algorithms, and we plan to apply it within our coaddition framework. We also present a summary of our coaddition pipeline.
Mode of presentation: oral