Systematics

Foreground Physical Effects and Ongoing Efforts

A number of physical effects are under active study by members of the LSST Project and  Dark Energy Science Collaboration.  From one kiloparsec down, these include the four sources listed on the left sidebar and links.

Of course, there are likely unknown systematics.  A worthy goal of this workshop is to imagine what these may be.

To date, efforts to characterize these effects and their residuals after correction have largely focused on individual sources of systematic uncertainty, but optimal mitigation will only occur by considering the effects jointly.  For example, each of the above effects exhibits some degree of chromaticity, meaning that the effects are subtly different for objects with different spectral energy distributions, adding a layer of complexity that needs to be handled coherently in the analysis pipeline. Better understanding of residual foregrounds could be used as Bayesian input to an improved shear reconstruction algorithm; the best we now have (SFIT) marginalizes over these unknowns.

In addition to the effects themselves, we note at least three areas of ongoing efforts that can lead to mitigation of foreground systematics:

  1. Operations model.  The optimal cadence, which includes sky coverage strategy, filter changing, dithering pattern (including rotation angle), and exposure time distribution can be tuned to measure and/or mitigate effects such as CCD cross-talk, survey depth variation, chromatic effects, etc.
  2. Combined probes.  Different cosmological probes will have different sensitivities to systematic effects, which can be used to break degeneracies between systematics and science.  To be fruitful, systematics must be consistently incorporated into analyses and simulated data.
  3. Joint survey analyses. Combining imaging and spectroscopy of the same galaxies seen by LSST and other surveys (e.g., WFIRST-AFTA, Euclid) will help break degeneracies between intrinsic galaxy properties, galactic dust, and instrumental systematics.