Automatic detection of long-duration transients in Fermi-GBM data

Context. In the era of time-domain, multi-messenger astronomy, the detection of transient events on the high-energy electromagnetic sky has become more important than ever. Previous attempts to systematically search for onboard, untriggered events in the data of Fermi-GBM have been limited to short-duration signals with variability time scales smaller than approximate to 1 min. This is due to the dominance of background variations on longer timescales. Aims. In this study, we aim to achieve a detection of slowly rising or long-duration transient events with high sensitivity and a full coverage of the GBM spectrum. Methods. We made use of our earlier developed physical background model, which allows us to effectively decouple the signal from long-duration transient sources from the complex varying background seen with the Fermi-GBM instrument. We implemented a novel trigger algorithm to detect signals in the variations of the time series that is composed of simultaneous measures in the light curves of the different Fermi-GBM detectors in different energy bands. To allow for a continuous search in the data stream of the satellite, the new detection algorithm was embedded in a fully automatic data analysis pipeline. After the detection of a new transient source, we also performed a joint fit for spectrum and location using the BALROG algorithm. Results. The results from extensive simulations demonstrate that the developed trigger algorithm is sensitive down to sub-Crab intensities (depending on the search timescale) and has a near-optimal detection performance. During a two month test run on real Fermi-GBM data, the pipeline detected more than 300 untriggered transient signals. We verified, for one of these transient detections, that it originated from a known astrophysical source, namely, the Vela X-1 pulsar, showing pulsed emission for more than seven hours. More generally, this method enables a systematic search for weak or long-duration transients.