Resolving theta_{23} Degeneracy by Accelerator and Reactor Neutrino Oscillation Experiments

If the lepton mixing angle theta_{23} is not maximal, there arises a problem of ambiguity in determining theta_{23} due to the existence of two degenerate solutions, one in the first and the other in the second octant. We discuss an experimental strategy for resolving the theta_{23} octant degeneracy by combining reactor measurement of theta_{13} with accelerator nu_{mu} disappearance and nu_{e} appearance experiments. The robustness of the theta_{23} degeneracy and the difficulty in lifting it only by accelerator experiments with conventional nu_{mu} (and nu_{mu}-bar) beam are demonstrated by analytical and numerical treatments. Our method offers a way to overcome the difficulty and can resolve the degeneracy between solutions sin^2 theta_{23} = 0.4 and sin^2 theta_{23} = 0.6 if sin^2 (2 theta_{13}) \gsim 0.05 at 95% CL by assuming the T2K phase II experiment and a reactor measurement with an exposure of 10 GW.kt.yr. The dependence of the resolving power of the octant degeneracy on the systematic errors of reactor experiments is also examined.