Abstract
Abstract
The Deep Underground Neutrino Experiment (DUNE) and Hyper-Kamiokande (Hyper-K) will measure neutrino oscillation parameters with an unprecedented precision that requires neutrino energy estimation to be controlled at the few-MeV level. A central challenge in achieving this is the modelling of the reinteractions of hadrons produced in neutrino-nucleus scatters with the residual nuclear medium, or final-state interactions (FSI). In this work we use state-of-the-art neutrino interaction event generators to review the impact of FSI modelling on the kinematic and calorimetric neutrino energy estimators used by Hyper-K and DUNE respectively, considering both the semi-classical intranuclear cascades (INCs) that dominate current simulations and a microscopic treatment based on a relativistic mean field calculation. We find that plausible variations of the FSI model introduce uncertainties on the neutrino energy estimation proxies that are at or above the precision on the energy scale control required for Hyper-K and DUNE projected neutrino oscillation sensitivities, highlighting the importance of careful FSI modelling to allow robust near detector constraints. We further demonstrate that the different detector technologies, and the consequent different approaches to neutrino energy reconstruction, at the two experiments render them sensitive to different aspects of the FSI modelling. Neutrino energy estimation at Hyper-K is most impacted by pion absorption and nuclear effects beyond the semi-classical paradigm, whilst the DUNE energy estimation is more affected by the modelling of how hadronic energy is shared between sources of visible and invisible energy in the detector. We discuss the implications of these findings for neutrino oscillation analyses and outline some of the key experimental and theoretical developments needed to bring FSI modelling uncertainties under control.