Extended Data Fig. 7: Scrutiny of the θ₁₃ Measurement.

The nominal θ₁₃ measurement (top) can be decomposed into a) the rate-only and shape-only contributions, b) FD-I (no ND) and FD-II (isoflux) contributions. A measurement without marginalising over \(| \Delta {m}_{ee}^{2}|\) as (2.484 ± 0.036) × 10⁻³eV² is also shown. These numbers demonstrate that the nominal θ₁₃ measurement is dominated by the rate-only information (systematics limited) of the best FD-II isoflux data sample. Furthermore, releasing the constraint on \(| \Delta {m}_{ee}^{2}|\) does not impact the measured central value of θ₁₃. Two alternative θ₁₃ measurements are also shown for comparison: a) the Data-to-Data and b) the Reactor Rate Modulation (RRM). Both are expected to be immune to the reactor model spectrum distortion while excellent agreement is found (details in text). Last, the FD-I+FD-II single detector θ₁₃ measurements are also shown using two uncertainty prescriptions. The new data-driven prescription uses an increased 4σ reactor model shape uncertainty. The standard reactor model prescription is also shown, indicating a bias on the result. The agreement of the single and multi detector, with the more conservative uncertainty and the much better χ²/d.f., suggests the new prescription provides a better treatment of the data. The previous FD-I only SD θ₁₃ measurement¹² (blue) is shown for reference. Bugey4 must be used in all single detector measurements to protect the rate normalisation. The 1σ uncertainty stands for 68% frequentist probability: both statistics (red error bar) and total (black error bar, including systematics).