The study of fast radio transients is important for a number of fundamental problems in astrophysics, including the search for new classes of object, the detection of the hidden baryonic matter in the intergalactic medium, and as tests of General Relativity via the formation and evolution of compact objects. These systems are rare, so large surveys are conducted for their discovery. However, because of the events’ rarity, with the right type of measurement even a single discovery can produce a powerful science payoff, offering a unique view of the physics at play.
Recent technological advances, exemplified by the development of the Jansky Very Large Array (VLA), have opened access to fast transients through millisecond timescale radio interferometric imaging. While large, single-dish telescopes have pioneered the field, interferometers will transform it through their ability to precisely localize, efficiently survey, and fully characterize sources. For example, the 100-meter Green Bank Telescope and VLA have comparable sensitivities, but the VLA surveys 16 times more sky while localizing sources 10 to 300 times more precisely. However, this powerful capability can only be exploited if data rates larger than 1 TB hour\({}^{-1}\) can be managed. At the VLA, collecting 1 hour of such data requires 3 hours simply to move it to the archive and, once written, the data volume overwhelms astronomers. A new paradigm is needed to solve the “needle in the haystack”-style problem with such massive data streams.
We propose building a GPU-accelerated cluster at the VLA to search for fast transients in real-time, an effort we call realfast. Searching the data in real time allows triggered data recording only for those rare moments when a transient is detected. At the VLA, this will open access to higher data rates and smaller integration times than has been achieved before, giving higher sensitivity to millisecond-scale radio transients. Moreover, by integrating this system with a commensal data stream, we will turn every VLA observation into a portion of a massive transient survey. A new, dedicated compute cluster is the key to accessing thousands of hours of observing time per year and fulfilling the latent science potential of the VLA.
Our team has developed the concept of fast imaging and real-time processing over the past five years through a series of projects, including the largest interferometric search for millisecond transients (200 hours, 200 Terabytes). Most recently, we have built a real-time, fast transient search pipeline for dedicated observations at the VLA. We are now poised to apply those lessons to the challenge of real-time, commensal transient surveys.
Our off-line and real-time fast transient searches have prototyped major components of the proposed system. We have shown that triggered data recording will reduce the recorded data volume by a factor of 1000. The proposed system will observe at any frequency and antenna configuration to capture all transient events stronger than \(8\sigma\), or roughly 100 mJy at timescales of 1–20 ms. Over the course of this program, the proposed system will:
Detect \(\gtrsim 12\) cosmological “fast radio bursts” with arcsecond localization and potentially find their host galaxies for pioneering tests of cosmology and the intergalactic medium.
Search for transients on time scales from milliseconds to minutes commensally with other observations, including the 6000-hour VLA Sky Survey.
Detect transients within minutes of observation, permitting rapid, multiwavelength follow up.
This proposal is timely, since the VLA can lead development of this novel concept as next-generation radio radio interferometers are under construction throughout the world. Major investments are being made in telescopes such as MeerKAT and through projects such as the ngVLA. However, their potential will remain untapped if their design is limited by costs of data transportation or they ignore the potential for commensal science. Real-time processing and “data triage” will be critical to accessing high data-rate science in these projects and others, both in astronomy and beyond.