UC Berkeley

In the study of astronomical transients, deriving knowledge from discovery is a multifaceted process that includes real-time classification to identify new events of interest, deep, multi-wavelength follow-up of individual events, and the global analysis of multi-event catalogs. Here we present a body of work encompassing each of these steps as applied to the study of gamma-ray bursts (GRBs). First, we present our work on utilizing machine-learning algorithms on early-time metrics from the Swift satellite to inform the resource allocation of follow-up telescopes in order to optimize time spent on high-redshift GRB candidates. Next, we show broadband observations and analysis of the early-time afterglow of GRB 120119A, utilizing data obtained with a dozen telescope facilities both in space and on the ground. This event exhibits extreme red-to-blue color change in the first few minutes after the trigger at levels unseen in prior afterglows, and our model fits of this phenomenon reveal the best support yet for the direct detection of dust destruction in the local environment of a GRB. Finally, we present results from the PAIRITEL early-time near-infrared (NIR) afterglow catalog. The 1.3 meter PAIRITEL has autonomously observed 14 GRBs in under 3 minutes after the burst, yielding a homogenous sample of early-time JHK light curves. Our analysis of these events provides constraints on the early-time NIR GRB afterglow luminosity function and gives insight into the importance of dust extinction in the suppression of some optical afterglows.