Can a supernova be located by its neutrinos?

A future core-collapse supernova in our Galaxy will be detected by several neutrino detectors around the world. The neutrinos escape from the supernova core over several seconds from the time of collapse, unlike the electromagnetic radiation, emitted from the envelope, which is delayed by a time of the order of hours. In addition, the electromagnetic radiation can be obscured by dust in the intervening interstellar space. The question therefore arises whether a supernova can be located by its neutrinos alone. The early warning of a supernova and its location might allow greatly improved astronomical observations. The theme of the present work is a careful and realistic assessment of this question, taking into account the statistical significance of the various neutrino signals. Not surprisingly, neutrino-electron forward scattering leads to a good determination of the supernova direction, even in the presence of the large and nearly isotropic background from other reactions. Even with the most pessimistic background assumptions, SuperKamiokande (SK) and the Sudbury Neutrino Observatory (SNO) can restrict the supernova direction to be within circles of radius $5°$ and $20°,$ respectively. Other reactions with more events but weaker angular dependence are much less useful for locating the supernova. Finally, there is the oft-discussed possibility of triangulation, i.e., determination of the supernova direction based on an arrival time delay between different detectors. Given the expected statistics we show that, contrary to previous estimates, this technique does not allow a good determination of the supernova direction.