I described a collection of methods which use multi-epoch photometry to identify a large number of quasars and estimate their redshifts without spectroscopy. Most galaxy and cosmology science done with quasars requires objects with verified types and redshifts, but there is not enough spectroscopic follow-up for even a small fraction of the sources that current photometric surveys detect. This work was done as a small scale proof of concept for how to study large populations of quasars without spectroscopy, as will have to be done on a much larger scale in future to maximize the science output LSST and other time-domain surveys.

First, using a Bayesian selection algorithm, we determined the optimal combination of color and variability information to identify quasars in current and future multi-epoch optical surveys. We classify the objects in the Sloan Digital Sky Survey (SDSS) Stripe 82 where about 100 repeat observations have been performed, allowing for the study of variability. The color analysis is based on coadded SDSS photometry, and the variability parameters are calculated from power-law fits to the structure functions. We find that simultaneous color and variability classification improves classification over either color or variability selection alone, with particular improvement in the selection of quasars with colors similar to stars. We achieve overall completeness of 96.9% and efficiency of 96.8%. This method identifies 35,820 quasar candidates on Stripe 82, 63% of which had not been previously identified.

Then, to ensure that the sample is useful for doing science, the redshifts of the candidate quasars were estimated using all available bands, weighting each band by smoothing the PDF. To improve upon traditional photometric redshifts methodologies, we incorporated astrometric information into the calculations, using the prismatic effects of the Earth's atmosphere as a low-resolution spectrograph. The astrometric redshift and the photometric redshift probability distribution functions (PDFs) contain complementary information about the quasar redshift. The astrometric redshifts alone are not as accurate, but when combined can break the degeneracies in photometric redshifts, thus improving their accuracy.

The LSST Operations Simulator (OpSim) and Metric Analysis Framework (MAF) make it possible to determine how effective the telescope and processing pipeline will be at reaching the various science objectives of the telescope. Multiple OpSims have been produced by the LSST Collaboration with deviations from the current baseline cadence for the survey. The MAF software package interacts with the OpSims to determine how a given cadence will be for a particular science case with the future LSST data. We wrote a MAF to compare how the proposed OpSims, with different airmass limits and cadences, will affect the astrometric redshift measurements. Initial results show that, after the first year, the number of observations in the main part of the survey should be sufficient that astrometry should be a useful supplement to photometric redshifts, no matter what airmass limit is chosen.