Flavor and energy inference for the high-energy IceCube neutrinos

TL;DR

This paper develops a Bayesian inference method to estimate the energy and flavor of high-energy neutrinos detected by IceCube, providing the first detailed posterior probability distributions for these properties based on six years of data.

Contribution

It introduces a novel Bayesian framework combined with MCMC techniques to infer neutrino properties from IceCube observables, improving understanding of neutrino interactions.

Findings

First posterior distributions for neutrino energy and flavor in IceCube data

Enhanced accuracy in neutrino property reconstruction

Method applicable to future neutrino detection analyses

Abstract

We present a flavor and energy inference analysis for each high-energy neutrino event observed by the IceCube observatory during six years of data taking. Our goal is to obtain, for the first time, an estimate of the posterior probability distribution for the most relevant properties, such as the neutrino energy and flavor, of the neutrino-nucleon interactions producing shower and track events in the IceCube detector. For each event the main observables in the IceCube detector are the deposited energy and the event topology (showers or tracks) produced by the Cherenkov light by the transit through a medium of charged particles created in neutrino interactions. It is crucial to reconstruct from these observables the properties of the neutrino which generated such event. Here we describe how to achieve this goal using Bayesian inference and Markov chain Monte Carlo methods.