Comparative analysis of SN1987A antineutrino fluence

We discuss the electron antineutrino fluence derived from the events detected by Kamiokande-II, IMB and Baksan on February 23, 1987. The data are analysed adopting a new simple and accurate formula for the signal, improving on the previous modeling of the detectors response, considering the possibility of background events. We perform several alternative analyses to quantify the relevance of various descriptions, approximations and biases. In particular, we study the effect of: omitting Baksan data or neglecting the background, using simplified formulae for the signal, modifying the fluence to account for oscillations and pinching, including the measured times and angles of the events, using other descriptions of detector response, etc. We show that most of these effects are small or negligible and argue, by comparing the allowed regions for astrophysical parameters, that the results are stable. We comment on the accordance with theoretical results and on open questions.