Constraining neutrino interaction uncertainties for oscillation experiments

Neutrinos rarely interact, but sometimes they do. How much? And does it matter? Charged-Current Quasi-Elastic (CCQE) neutrino-nucleus interactions are the main signal channel for T2K and other currently running neutrino oscillation experiments, but in recent years, with a number of experimental results, it has become clear that these interactions are not as well understood at few-GeV neutrino energies as was previously thought. This thesis focuses on the impact that this uncertainty will have on neutrino oscillation measurements. Published CCQE and the closely related Neutral-Current Elastic (NCEL) cross section results from the MiniBooNE experiment are used to fit a 3+1 sterile neutrino model, and the uncertainty over fundamental parameters in the cross section model used is shown to badly bias the results. Conservative sterile neutrino limits which treat the cross section uncertainty correctly are used, and a general note of caution is sounded over sterile neutrino results which make tacit and unfounded assumptions about the neutrino cross section model. New theoretical models of CCQE scattering which try to explain the experimental situation have become available, some of which have been implemented into NEUT, T2K's primary interaction generator. Two candidate models are used in a fit to all published CCQE data on nuclear targets to select a default model and constrain the parameters of that model in order to increase the sensitivity of T2K oscillation analyses. This CCQE parametrisation will be used for T2K oscillation analyses from the summer of 2015 onwards. The fit framework developed for this work has been integrated into the T2K oscillation analysis framework for future iterations. Additionally, the implementation of one such model, the Effective Spectral Function is described and compared with the available neutrino-nucleus scattering data