A method for untriggered time-dependent searches for multiple flares from neutrino point sources

TL;DR

This paper introduces a novel time-dependent search method for detecting multiple weak neutrino flares from astrophysical sources, combining time-clustering with an unbinned likelihood to improve sensitivity over various time-scales.

Contribution

It presents a new approach that enhances detection of multiple small neutrino flares by integrating clustering with likelihood analysis, especially for short-duration flares.

Findings

Sensitivity is about 15% worse than standard methods for single flares.

More sensitive than time-integrated analysis for short flares.

Effective for detecting weak, distributed neutrino signals.

Abstract

A method for a time-dependent search for flaring astrophysical sources which can be potentially detected by large neutrino experiments is presented. The method uses a time-clustering algorithm combined with an unbinned likelihood procedure. By including in the likelihood function a signal term which describes the contribution of many small clusters of signal-like events, this method provides an effective way for looking for weak neutrino flares over different time-scales. The method is sensitive to an overall excess of events distributed over several flares which are not individually detectable. For standard cases (one flare) the discovery potential of the method is about 15% worse than a standard time-dependent point source analysis with unknown duration of the flare. However, for flares sufficiently shorter than the total observation period, the method is more sensitive than a time-integrated analysis.