Research
I have actively participated in many of the time-domain surveys of the last decade. My work has focused on recognizing, extracting, classifying, and studying variable astronomical phenomena in near real-time from survey-class data streams. An important aspect of many of these surveys has been the dissemination of information to the astronomical community to enable global follow-up and parallel science. I have been a strong advocate of this position, and have helped to realize this goal when possible.
The important scientific problems that have motivated my past research include : the study of Galactic dark matter through gravitational microlensing; a census of the orbits and composition of solar system objects; the history and eventual fate of the Universe as inferred from supernovae studies; the real-time study of rare events from the distant Universe; and the discovery of truly unknown and unexpected phenomena. Descriptions of my research with, and contributions to, various time-domain astronomical surveys are detailed below.
Large Synoptic Survey Telescope (LSST)
LSST is a next generation astronomical survey that will fulfill the promise of ``cosmic cinematography'', observing the entire available sky every few nights (Tyson 2002). My work with LSST so far has been in the areas of algorithmic research and development with a team of scientists from the University of Washington, Princeton University, Brookhaven National Lab (BNL), and the National Center for Supercomputing Applications (NCSA). Our team is assessing the quality of current state-of-the-art photometry and difference imaging packages, and ascertaining if they are able to meet LSST's stringent science requirements. We expect that this effort will establish the University of Washington as an important member of the LSST Image Processing Pipeline, with involvement continuing into the construction and operational stages of the project. Our collaboration with NCSA will yield an operational infrastructure to quickly reduce large-scale astronomical data sets using the TeraGrid.
Sloan Digital Sky Survey (SDSS) II Supernova Search
This effort is part of the SDSS II extension and intends to fill in the ``redshift desert'', between redshifts of 0.1 < z < 0.3, that lacks well-measured Type Ia SNe. We initially expected to have ~200 well-measured SNe lightcurves after three years of observations. However, we verified more than 120 events as Type Ia SNe in a very successful first year of observations. My responsibilities include adapting and maintaining a version of Photpipe that reduces SDSS images, making sure that the analysis software is delivering quality candidates to the classification stages, adding deep co-added templates to the difference imaging pipeline, reviewing candidate events, as well as undertaking follow-up of detected supernovae on the APO 3.5 telescope. I expect to start developing a method of ``final photometry'' on the difference images that will be generalized and portable, as well as to work towards a high-quality set of low redshift (z < 0.1) lightcurves that will supersede the commonly used but 10-year-old Calan-Tololo set (Hamuy et al. 1996). These lightcurves provide the anchor for all dark energy studies using SNe, and are therefore of fundamental importance.
SuperMACHO Collaboration
The SuperMACHO collaboration is a next-generation microlensing survey undertaken on the CTIO 4m telescope. By using a larger telescope, larger focal plane, and better site than MACHO, we expect to detect a factor of several more events, which should allow us to determine the nature of the microlensing signal seen towards the LMC (Alcock et al. 2000) and M31 (Uglesich et al. 2004). I was involved in the initial strategy and field-selection process, and have multiple integrated months of on-site observing experience from this project alone. For this project I helped to develop (along with Dr. Armin Rest) the Mscpipe and Photpipe automated data reduction pipelines. These are fault-tolerant, generalized image reduction (Mscpipe) and difference imaging (Photpipe) pipelines that completely automate the flow of information from the telescope to the internet (Smith et al. 2002). These pipelines are currently in use by a variety of collaborations, and are also being used as testbeds by various Large Synoptic Survey Telescope (LSST) working groups. For this project I designed a web-based method of candidate event review and classification that allows collaborators at distant institutions to review the data in near real-time, alleviating the burden on those astronomers at the telescope. As a prime example of taking advantage of unexpected, parallel science in time-series data, we recently uncovered light echoes from several-hundred-year-old supernovae that were initially classified as background noise in our search (Rest et al. 2005).
ESSENCE Collaboration
This supernova search is the sister project to SuperMACHO, sharing collaborators and image reduction software. The goal of ESSENCE is to constrain the dark energy equation of state parameter to 10% accuracy by detecting Type Ia supernovae (SNe) over a broad range in redshift (Miknaitis et al, in preparation). My primary responsibilities have been to adapt Photpipe for the different field conditions compared to SuperMACHO as well as to adapt the underlying infrastructure for the display of candidate information for use by this particular search.
Deep Lens Survey (DLS)
The Deep Lens Survey is a weak lensing survey that was undertaken on the CTIO 4m and KPNO 4m telescopes. To build up the depth required to do weak lensing cosmology, a given DLS field was imaged 20 times in each of the B, V, R and z passbands. These observations were obtained in sets of 5-image dithers, the 5 images being taken back-to-back. This strategy provided unprecedented sensitivity to short timescale (600-900s) phenomena at faint (~23rd) magnitudes. The cadence of field reacquisition was staggered to provide sensitivity to variability across 5 decades of event timescale (103 - 108 seconds). I designed a difference imaging pipeline that reduced imaging data in near real-time, with latencies of 2-3 hours. I also wrote a graphical user interface to view events extracted from the pipeline - an astronomer at the telescope was allowed to view the sequence of difference images containing the candidate events and reject them as noise or classify their nature, such as a ``Supernova'', ``Asteroid'', or an otherwise unknown ``Transient''. Information on all events was automatically copied to a publicly available web site. This was a noteworthy transformation in the way astronomical variability data were treated - traditionally, supernovae studies have ignored variable stars in their data, microlensing studies have ignored asteroids, etc. We reported all events without bias, which built a practical foundation for the realization of synoptic variability science. This research allowed us to resolve a potentially new class of astronomical events that flare on short (1000s) timescales and at faint magnitudes (Becker et al. 2004). It also yielded a flexible difference imaging package still in use by many of the collaborations above (including SuperMACHO, ESSENCE, and the SDSS II Supernova Search).
Global Microlensing Alert Network (GMAN)
This effort was undertaken to follow-up MACHO-detected microlensing events on a nightly basis, and formed the bulk of my Ph.D. thesis work. My responsibilities included the coordination of, and allocation of targets to, a global array of 1-m class telescopes, including the UTSO 0.6m telescope, the CTIO 0.9m telescope, the CTIO 4m telescope, the MSO 0.8m telescope, the Wise Observatory 1.0m telescope, and the MJUO observatory 0.6m telescope. I designed a data reduction pipeline in Perl that automated the complete reduction of these data, from raw image calibration to photometry using the DAOPHOT package, with latencies of several hours. I was responsible for final photometry of the ensembles of data using the ALLFRAME package, joint lightcurve fitting, and extraction of the science contained in the lightcurves. These data were integral to resolving exotic microlensing effects in dozens of events, in many cases for the first time. The lightcurve measurements obtained by MACHO, GMAN, and MPS proved sensitive to more than 6 decades in lens mass, from Earth-mass planetary companions (Rhie et al. 2000) to ~10 M_sun isolated black hole lenses (Bennett et al. 2002) .
Microlensing Planet Search (MPS)
I also joined the MPS Collaboration, who were (independent of GMAN) allocated microlensing follow-up time on the MSO 1.9m telescopes and CTIO 0.9m telescopes. The goal of MPS was to detect short timescale deviations in on-gong events; thus the strategy was more dense follow-up of fewer events than was undertaken by GMAN (e.g. Rhie et al. 1999). My responsibilities included on-site observing (including at CTIO in Chile and MSO in Australia), design and maintenance of image reduction software, and real-time lightcurve analysis.
MACHO Collaboration
The MACHO project was one of the first truly large-scale photometric time-domain surveys, and developed and utilized one of the first multi-chip focal planes. We observed nightly for several years the Galactic Bulge and Large and Small Magellanic Clouds, searching for the 1-in-10-million signature of gravitational microlensing. My responsibilities in MACHO included the early design, eventual implementation, and daily monitoring of the MACHO Alert System. This system was designed to automatically sift through the nightly addition of data points (to millions of lightcurves!) for events that appeared similar to gravitational microlensing. I monitored candidate events daily, and if they were well fit in functional form to microlensing, released alerts to the astronomical community with a latency of 12-24 hours to enable the global followup of events. I maintained a web-page that contained the most recent fit parameters, finding charts, and other information that assisted follow-up teams. On-going events were also monitored daily for signatures of ``exotic'' effects, which manifested themselves as sometimes slight deviations from the smooth, symmetric, idealized microlensing lightcurve. These exotic effects included lens binarity (Alcock et al. 2000; Alcock et al. 1999), source binarity (Alcock et al. 2001), parallax due to the Earth's motion around the Sun (Bennett et al. 2002), and effects due to the finite size of the lensed source star (Alcock et al. 1997). Resolution of these effects generally allowed some breaking of the degeneracies on the lens parameters of mass, distance, and velocity that are encoded in the standard microlensing fit. Importantly, resolving these second-order effects helped to verify the microlensing interpretation for these incidents of stellar brightening, as well as to mature the field of microlensing studies overall.