Confidence in the neutrino mass hierarchy

The number of sigma of confidence in a determination of the neutrino mass hierarchy may be obtained from the statistic Delta chi squared. However, as the hierarchy is a discrete variable, this number is not given by the usual square root formula. We review a simple Bayesian formula for the sensitivity to the hierarchy that can be obtained from the median experiment as a function of Delta chi squared. We compare this analytical formula to 6 years of simulated data from JUNO together with a 4% (1%) determination of the effective atmospheric mass splitting from the disappearance channel at MINOS (NOvA). We find a Delta chi squared of 11 (20) yielding 2.6 sigma (3.9 sigma). However when the unknown nonlinear energy response of the detector is included in our analysis this significance degrades considerably. This degradation can be eliminated by dividing the single detector into a near and far detector of the same total target mass. A further advantage of a second detector is that, even while the reactor neutrino experiment runs, the decay at rest of a single, high intensity, continuously running pion source close to one of the detectors, such as that described by the DAEdALUS project, may determine the leptonic CP-violating phase delta.