Recently, new reactor antineutrino spectra have been provided for 235 U, 239 Pu, 241 Pu, and 238 U, increasing the mean flux by about 3 percent. To a good approximation, this reevaluation applies to all reactor neutrino experiments. The synthesis of published experiments at reactor-detector distances < 100 m leads to a ratio of observed event rate to predicted rate of 0.976±0.024. With our new flux evaluation, this ratio shifts to 0.943±0.023, leading to a deviation from unity at 98.6% C.L. which we call the reactor antineutrino anomaly. The compatibility of our results with the existence of a fourth non-standard neutrino state driving neutrino oscillations at short distances is discussed. The combined analysis of reactor data, gallium solar neutrino calibration experiments, and MiniBooNEν data disfavors the no-oscillation hypothesis at 99.8% C.L. The oscillation parameters are such that |∆m 2 new | > 1.5 eV 2 (95%) and sin 2 (2θnew) = 0.14 ± 0.08 (95%). Constraints on the θ13 neutrino mixing angle are revised.

Precise predictions of the antineutrino spectra emitted by nuclear reactors is a key ingredient in measurements of reactor neutrino oscillations as well as of the recent applications to the surveillance of power plants in the context of non proliferation of nuclear weapons. We report new calculations including the latest information from nuclear databases and a detailed error budget. The first part of this work is the so-called ab initio approach where the total antineutrino spectrum is built from the sum of all β-branches of all fission products predicted by an evolution code. Systematic effects and missing information in nuclear databases lead to final relative uncertainties in the 10 to 20% range. A prediction of the antineutrino spectrum associated with the fission of 238 U is given based on this ab initio method. For the dominant isotopes 235 U and 239 Pu, we developed a more accurate approach combining information from nuclear databases and reference electron spectra associated with the fission of 235 U, 239 Pu and 241 Pu, measured at ILL in the 80's. We show how the anchor point of the measured total β-spectra can be used to suppress the uncertainty in nuclear databases while taking advantage of all the information they contain. We provide new reference antineutrino spectra for 235 U, 239 Pu and 241 Pu isotopes in the 2-8 MeV range. While the shapes of the spectra and their uncertainties are comparable to that of the previous analysis of the ILL data, the normalization is shifted by about +3% on average. In the perspective of the re-analysis of past experiments and direct use of these results by upcoming oscillation experiments, we discuss the various sources of errors and their correlations as well as the corrections induced by off equilibrium effects.