Corrigendum to “An alternative method of determining the neutrino mass ordering in reactor neutrino experiments” [Phys. Lett. B 772 (2017) 179–183] (Physics Letters (2017) 772 (179–183), (S0370269317305208), (10.1016/j.physletb.2017.06.044))

There was an error in incorporating the expression of the probability, given in eq. (16) of our article [1], in the code used for the statistical analysis of the simulated JUNO data performed in [1]. As a consequence, Fig. 1 and the sensitivity on [Formula presented] reported in the article are incorrect. They are corrected here using updated best fit values and 3σ allowed ranges of the relevant neutrino oscillation parameters, obtained in the recent global analysis of the neutrino oscillation data in ref. [2]. The other ingredients of the statistical analysis are the same as in [1]. In Fig. 1, left panel, we show the 1σ, 2σ, 3σ and 4σ allowed regions (for 1 degree of freedom) in the plane [Formula presented] assuming normal ordering. The red point represents the best fit point, which corresponds by construction to the best fit values of [Formula presented] and [Formula presented] from [2]. In Fig. 1, right panel, the presence of a second local [Formula presented] minimum at [Formula presented] C.L. at [Formula presented] (and [Formula presented] eV2), associated with spectrum with inverted ordering, is also seen. For the difference between the [Formula presented] minima in the NO and IO cases we find [Formula presented]. This implies that the type of neutrino mass ordering can be established at 3.8σ C.L. Performing a similar analysis but assuming the inverted ordering to be the “true” one we obtain [Formula presented], i.e., that the type of neutrino mass ordering can be established at [Formula presented] C.L. These confidence levels are somewhat higher that the 3σ C.L. which the analyses using the standard method [4] (and performed under the same conditions) typically give (see, e.g., [3]). The reason the alternative method employed in our analysis leads to a somewhat stronger rejection of the “wrong” ordering can be related to the fact that i) we have neglected matter effects in the oscillations of reactor [Formula presented], which, although small, have to be taken into account given the exceptionally high precision of JUNO experiment, and ii) we have not taken into account the small difference between the baselines of the reactors which provide the flux of [Formula presented] for JUNO. Our results show that the sensitivity to the neutrino mass ordering that can be achieved employing the proposed alternative method of determination of the ordering in any case is not worse than the sensitivity that can be achieved using the standard approach [4]. The precision with which [Formula presented] can be determined using the proposed alternative approach matches the precision that can be reached utilising the standard method [5]: we find that [Formula presented] can be determined with 1σ relative uncertainty of 0.49% in both cases of “true” spectrum considered. The precision on [Formula presented], which is determined simultaneously with [Formula presented] (or [Formula presented]) is also exceptionally high: assuming NO (IO) spectrum to be the true one we get for the 1σ relative uncertainty 0.15% (0.14%). These results confirm the conclusion reached in [1] that the considered novel method of determination of the neutrino mass ordering (spectrum) can be used as a complementary method of the ordering (spectrum) determination independently of, and on a equal footing with, the standard method. We would to thank Yuanyuan Zhang for useful correspondence regarding the problematic Fig. 1 in ref. [1].