The High cadence Transient Survey (HiTS): real-time detection of transient events

The High cadence Transient Survey (HiTS) is a survey looking for fast optical transients in the southern sky using the Dark Energy Camera (DECam). HiTS pioneered the real-time analysis of DECam data, managing to analyze a data stream of more than 33 Mbps coming from CTIO via optical fiber link (REUNA), using the supercomputer leftraru at the National Laboratory for High Performance Computing (NLHPC), hosted at CMM, with a custom built pipeline which uses advanced techniques from image analysis, high perfomance computing, machine learning and visualization.

HiTS is composed by an interdisciplinary team led by CMM scientist Francisco Förster, including mathematicians, astronomers, computer scientists and electrical engineers from different institutions (CMM, MAS, DAS, DIE, UNAB, ESO, CTIO).

HiTS has been running since 2013 in three different observing runs during contiguous nights close to new Moon. In 2013A we observed 120 square degrees with a cadence of two hours in the u’ band during 4 contiguous nights. In 2014A we also observed 120 square degrees with a cadence of two hours, but in the g’ band and during 5 contiguous nights. In 2015A we observed 150 square degrees with a cadence of 1.6 hours in the g’ band and during 6 contiguous nights, followed by three half nights separated from the main survey from days to weeks.

In 2014A and 2015A we were able to process the data in real time, i.e. each observation was processed in less than one exposure time after the observation was performed. In 2014A we discovered 12 young supernovae. In 2015A we used an improved pipeline which included a more precise image treatment and a better machine learning classifier and visualization webpage, which allowed us to found 18 more supernovae in the 2014A dataset, but also to improve the visual inspection significantly, allowing us to do this process before the arrival of the next image candidates. In total, in 2015A we discovered 90 new supernovae, many of them very young, which will become an important dataset for future high cadence surveys.

Our work was presented in an invited talk in the DECam Community Science Workshop 2015.