MINOS is a long baseline neutrino oscillation experiment. A manmade beam of predominantly muon neutrinos is detected both 1 km and 735 km from the production point by two functionally identical detectors. A comparison of the energy spectra measured by the two detectors shows the energy-dependent disappearance of muon neutrinos characteristic of oscillations and allows a measurement of the parameters governing the oscillations. This thesis presents work leading to measurements of disappearance in the 6% ∇_μ background in that beam. A calibration is developed to correct for time-dependent changes in the responses of both detectors, reducing the corresponding uncertainty on hadronic energy measurements from 1.8% to 0.4% in the near detector and from 0.8% to 0.4% in the far detector. A method of selecting charged current ∇_μ events is developed, with purities (efficiencies) of 96.5% (74.4%) at the near detector, and 98.8% (70.9%) at the far detector in the region below 10~GeV reconstructed neutrino energy. A method of using the measured near detector neutrino energy spectrum to predict that expected at the far detector is discussed, and developed for use in the ∇_μ analysis. Sources of systematic uncertainty contributing to the oscillation measurements are discussed. In the far detector, 32 ∇_μ events are observed below a reconstructed energy of 30 GeV, compared to an expectation of 47.8 for Δm̄²_atm = Δm²_atm, sin²(2‾θ₂₃) = sin²(2θ₂₃). This deficit, in such a low statistics sample, makes the result difficult to interpret in the context of an oscillation parameter measurement. Possible sources for the discrepancy are discussed, concluding that considerably more data are required for a definitive solution. Running MINOS with a dedicated ∇_μ beam would be the ideal continuation of this work.

The IceCube Neutrino Observatory at the South Pole is the world's largest neutrino detector with over 1km^3 of instrumented Antarctic ice. While it has been primarily designed to observe astrophysical neutrinos, this size also allows it to collect vast quantities of atmospheric neutrinos. These high-statistics datasets allow for measurements of the properties of neutrinos, in particular the phenomena of neutrino oscillation. One of the outstanding questions in this field is that of the neutrino mass ordering (NMO). The Precision IceCube Next Generation Upgrade (PINGU) is a proposed low-energy extension to IceCube for which a determination of the NMO is a priority science goal. The current low-energy atmospheric neutrino experiment at the South Pole, DeepCore, has been successfully collecting data since 2011. In this thesis the potential of this existing data to determine the NMO has been explored. While it was not expected to have a large sensitivity, this work has explored a Feldman-Cousins treatment for converting the delta-chi^2 between the two discrete mass ordering hypotheses into the standard Gaussian significance metric. Using 2.7 years of data from the DeepCore detector, the inverted mass ordering was preferred at the level of 0.05sigma. The second aspect of this thesis was to study the impact of the systematic uncertainties on the NMO determination. This particular analysis was actually statistics-limited and so the only impactful systematic uncertainties were the parameters that govern atmospheric neutrino oscillations, theta_23 and Deltam^2_31. Therefore, to improve the NMO results, these parameters were constrained by including the global information on them in the fits, yielding a new NMO sensitivity of 0.29sigma. This new global fit also yields measurements of the oscillation parameters of Deltam^2_32,NO=(2.443+/-0.037)e-3eV^2 and sin^2theta_23,NO=0.442+0.026-0.018 for the hypothesis of the normal mass ordering and Deltam^2_32,IO=(-2.510+/-0.036)e-3eV^2 and sin^2theta_23,IO=0.579+0.019-0.021 for the hypothesis of the inverted mass ordering. In addition to the work on the neutrino mass ordering, this thesis also investigated two issues related to predictions of the flux of atmospheric particles. The first related to the treatment of the predictions of the atmospheric neutrino flux, provided in binned tables. Crucially, these contain values representative of the integral of the flux across that bin and so an integral-preserving interpolation must be used. One such method will be presented along with a discussion of how it performs in the two-dimensional case of the atmospheric neutrino flux. The second issue related to quantifying uncertainties on the background muon distributions observed with the IceCube detector coming from the uncertainties on the initial cosmic ray flux. This involved performing a global fit on the available cosmic ray flux measurements and then propagating these uncertainties in to the muon distributions. To finalise this section, the exact manner in which these uncertainties can be included in the physics analyses of IceCube will be discussed.

Double Chooz est une expérience dédiée à la mesure de l'angle de mélange θ₁₃ caractérisant le phénomène d’oscillation des neutrinos. Elle consiste en l’installation de deux détecteurs identiques respectivement installés à 400 m et 1050 m des deux réacteurs à eau pressurisée de la centrale nucléaire de Chooz dans les Ardennes. Les réacteurs nucléaires sont en effet à l’origine d’un flux intense d’antineutrinos électroniques (de l’ordre de 10²¹ ⊽ₑ/s pour un réacteur de 1GWe)qui peut être détecté par réaction bêta inverse dans le liquide scintillant des détecteurs : ⊽ₑ + p −> e⁺ + n. Le paramètre θ₁₃ peut ensuite être déterminé en cherchant une réduction du nombre d’antineutrinos et une distorsion du flux mesuré dans le détecteur lointain par rapport au détecteur proche. La première phase de l’expérience pour laquelle uniquement le détecteur lointain prend des données a débuté en avril 2011. En l’absence du détecteur proche dont l’installation sera terminée en 2014, une prédiction du flux d’antineutrinos non oscillé attendu dans le détecteur lointain est nécessaire à la prédiction de θ₁₃ . Dans ce manuscrit, nous présentons le travail de simulation réalisé en vue de prédire les taux de fission des deux cœurs de Chooz à l’origine des antineutrinos émis par les réacteurs. Pour cela des simulations de cœur complet des réacteurs ont été développées à l’aide du code de simulation MCNP Utility for Reactor Evolution (MURE). Les résultats de ces simulations ont permis de déterminer les taux de fission et les erreurs systématiques associées durant les périodes de prise de données et d’aboutir à la première indication d’un angle θ₁₃ non-nul en novembre 2011. / The Double Chooz experiment is designed to search for a non-vanishing mixing angle θ₁₃ characterizing the ability of neutrinos to oscillate. It consists in two identical detectors located respectively at 400 m and 1050 m of the two pressurized water reactors of the Chooz nuclear plant in the French Ardennes. Indeed, nuclear reactor are huge electron antineutrino emitters (about 10²¹ ⊽ₑ/s for a 1GWe reactor). In Double Chooz, antineutrino sare detected by the inverse beta decay process in the liquid scintillator of the detectors : ⊽ₑ + p −> e⁺ + n. The θ₁₃ parameter can be investigated searching for ⊽ₑ disappearance and ⊽ₑ energy distortion in the far detector with respect to the near detector. The first phase of the experiment during which only the far detector is taking data has started in April 2011. In absence of far detector whose installation will be completed in 2014, a prediction of the non-oscillated antineutrino flux and spectrum shape expected in the far detector is mandatory to measure θ₁₃ . In this manuscript, we present the simulation work performed to predict the fission rates of both Chooz cores responsible for the reactor antineutrino flux. In this view, a complete core model has been developed with the MCNP Utility for Reactor Evolution (MURE) simulation code. The results of these simulations were used to determine the fission rates and associated systematic errors since the beginning of data taking and led to the first indication for a non-zero θ₁₃ mixing angle in November 2011.

The T2K experiment is a long-baseline neutrino oscillation experiment, in which a muon neutrino beam is directed over a 295 km baseline from the J-PARC facility to the Super-Kamiokande detector. This allows neutrino oscillation to be studied in two channels: disappearance of &nu;_&mu; and appearance of &nu;_e. The T2K beam can be run either in neutrino or antineutrino configuration, for a neutrino beam which is predominantly composed of &nu;_&mu; or anti -&nu;_&mu; respectively. This thesis presents the first oscillation analysis to combine neutrino-mode and antineutrino-mode data sets, giving the first ever sensitivity to the CP-violating phase &delta;_CP from T2K data alone, as well as the most precise T2K measurement of the other neutrino oscillation parameters. The analysis uses a Markov Chain Monte Carlo method to construct a sample from the Bayesian posterior distribution, from which the oscillation parameters of interest and their uncertainties are estimated. Data samples from the T2K near detector, ND280, are fit simultaneously with data from the far detector in order to reduce the uncertainty in the far-detector prediction. When fitting the T2K data alone, the best-fit oscillation parameter values are sin² &theta;₂₃= 0.519^+0.031<sub style='position: relative; left: -2.8em;'>-0.059</sub>, sin² &theta;₁₃= 0.0257^+0.0106<sub style='position: relative; left: -3.2em;'>-0.0440</sub>, &Delta;m²<sub style='position: relative; left: -0.4em;'>32</sub>= 2.54^+0.12<sub style='position: relative; left: -2em;'>-0.10</sub> x 10^-3eV²,and &delta;_CP=-1.91 rad. The 90&percnt; credible interval for &delta;_CP excludes values around &pi;/2: &delta;_CP &notin; [0.38, 2.60] rad. When fitting the T2K data with a constraint on sin² &theta;₁₃ from measurements by reactor experiments, the best-fit oscillation parameter values are sin² &theta;₂₃= 0.519^+0.061<sub style='position: relative; left: -2.8em;'>-0.029</sub>, sin² &theta;₁₃= ^+0.0014<sub style='position: relative; left: -3.2em;'>-0.0014</sub>, Dm²<sub style='position: relative; left: -0.4em;'>32</sub>= 2.54^+0.12<sub style='position: relative; left: -2em;'>-0.12</sub> x 10^-3eV², and &delta;_CP=-1.72 rad. The 90&percnt; credible interval for &delta;_CP contains values &delta;_CP &isin; [-3.10,-0.17] rad, excluding the CP-conserving values 0 and ± &pi; at 90&percnt; probability.

LE 8B PRODUIT DANS LE COEUR DU SOLEIL A TRAVERS LA REACTION 7BE(P,) 8B EST LA SOURCE MAJEURE SINON UNIQUE DES NEUTRINOS DE HAUTES ENERGIES DETECTES DANS LA PLUPART DES EXPERIENCES DE DETECTION DES NEUTRINOS SOLAIRES, EXCEPTE GALLEX ET SAGE. CES EXPERIENCES ONT TOUTES MESURE UN FLUX DE NEUTRINOS INFERIEUR A CELUI PREDIT PAR LES MODELES SOLAIRES. PLUSIEURS EXPLICATIONS ONT ALORS ETE INVOQUEES POUR TENTER DE COMPRENDRE CE DEFICIT MAIS TOUTES NECESSITENT UNE CONNAISSANCE PRECISE DE LA VALEUR DE LA SECTION EFFICACE DE LA REACTION 7BE(P,) 8B, PUISQUE LE FLUX DES NEUTRINOS DU 8B EST DIRECTEMENT PROPORTIONNEL A CETTE DERNIERE. LA MESURE DIRECTE DE LA SECTION EFFICACE DE CETTE REACTION A L'ENERGIE SOLAIRE EST IMPOSSIBLE A CAUSE DE SA TRES FAIBLE VALEUR (DE L'ORDRE DU FEMTOBARN). POUR CONTOURNER CE PROBLEME, LES SECTIONS EFFICACES SONT MESUREES A PLUS HAUTE ENERGIE PUIS EXTRAPOLEES A L'ENERGIE SOLAIRE EN UTILISANT UNE DEPENDANCE EN ENERGIE THEORIQUE. LES SIX DETERMINATIONS EXPERIMENTALES PRECEDENTES DE LA SECTION EFFICACE SE DIVISAIENT EN DEUX GROUPES BIEN DISTINCTS PRESENTANT DES ECARTS DE L'ORDRE DE 30%, CE QUI IMPLIQUAIT UNE INCERTITUDE DU MEME ORDRE SUR LE FLUX DES NEUTRINOS DE HAUTES ENERGIES. REMESURER AVEC UNE MEILLEURE PRECISION LA SECTION EFFICACE DE CETTE REACTION EST DONC APPARU TRES IMPORTANT. DANS UN PREMIER TEMPS, NOUS AVONS EFFECTUE DES MESURES DIRECTES DE LA SECTION EFFICACE DE CETTE REACTION DANS LA GAMME D'ENERGIE COMPRISE ENTRE 0.35 ET 1.4 MEV (CM). CES EXPERIENCES ONT FAIT L'OBJET D'UNE MESURE PRECISE DE CHACUN DES PARAMETRES INTERVENANT DANS LA DETERMINATION DE LA SECTION EFFICACE. DANS UN SECOND TEMPS, NOUS AVONS ENTREPRIS DES MESURES DE LA SECTION EFFICACE AUPRES DE L'ACCELERATEUR PAPAP, A 185.8 KEV, 134.7 KEV ET 111.7 KEV, L'ENERGIE DANS LE CENTRE DE MASSE LA PLUS BASSE JAMAIS ATTEINTE A CE JOUR. LES RESULTATS SONT EN EXCELLENT ACCORD AVEC CEUX OBTENUS A PLUS HAUTES ENERGIES. LA VALEUR TROUVEE POUR LE FACTEUR ASTROPHYSIQUE S 1 7(0), PAR L'EXTRAPOLATION DE NOS DONNEES EST EGALE A 19.21.3 EV-B, CE QUI ENTRAINE UNE REDUCTION SENSIBLE DE L'INCERTITUDE SUR LE FLUX DES NEUTRINOS DE HAUTE ENERGIE DU 8B.