Ivezic, ZeljkoMacLeod, Chelsea Louise2012-09-132012-09-132012-09-132012MacLeod_washington_0250E_10656.pdfhttp://hdl.handle.net/1773/20681Thesis (Ph.D.)--University of Washington, 2012I provide a quantitative analysis of the database of 3.5 million photometric measurements for 80,000 spectroscopically confirmed quasars recently assembled by the Sloan Digital Sky Survey (SDSS). This database is an excellent data set to study quasar variability, and includes 9,000 well-sampled light curves from Stripe 82 and a 25,000-object two-epoch sample from the northern Galactic sky. I show that the damped random walk model provides a robust statistical description of these data. This model characterizes the variability using two parameters, the long-term amplitude and characteristic timescale, and these model parameters show trends with physical parameters such as the luminosity and black hole mass. This mathematical model supports accretion disk instabilities as the dominant variability mechanism. However, there is substantial scatter around the mean trends, as other sources of stochasticity are yet to be discovered. If magnetohydrodynamic models could be improved to reproduce the observed characteristics of variability, they should be able to shed light on the source of stochasticity inherent in quasar light curves. In this dissertation, I discuss the application of a quasar variability model to study the underlying physics of accretion disks. I also discuss the prospect of using variability as a selection method for quasars. Lastly, I use results from SDSS to discuss the prospects of studying quasar variability in upcoming large time-domain sky surveys.application/pdfen-USCopyright is held by the individual authors.Quasars; SDSS; Time Series AnalysisAstronomyAstrophysicsAstronomyThesis