Fitting with FD covariance matrices Seb Jones Department of Physics - PowerPoint PPT Presentation

Fitting with FD covariance matrices Seb Jones Department of Physics & Astronomy University College London April 29, 2019 S. Jones (UCL) DUNE LBL April 29, 2019 1 10

Previous fitting attempts 2 χ 5 4 3 2 FD CovMx & no FD EScale 1 No FD CovMx & no FD EScale FD Escale & no FD CovMx 0 − − 1 0.5 0 0.5 1 δ π / Last week showed a variation of this plot showing CPV sensitivity change when fitting with FD covariance matrix Using FD covariance matrix lowers sensitivity relative to no CovMx case when also not profiling over FD EScale systematics (green) However, sensitivity unexpectedly greater than case where we profile over FD EScale systematics (red) S. Jones (UCL) DUNE LBL April 29, 2019 2 10

Covariance matrix with unoscillated spectrum FD covariance matrix with unoscillated spectra 0.3 60 0.25 50 0.2 40 0.15 0.1 30 0.05 20 0 10 -0.05 -0.1 10 20 30 40 50 60 This is covariance matrix that was previously being used, produced with unoscillated ν µ and ν e distributions From left to right, samples are ν µ FHC, ν µ RHC, ν e FHC and ν e RHC S. Jones (UCL) DUNE LBL April 29, 2019 3 10

With oscillated spectra (2 neutrino approx.) FD covariance matrix with oscillated spectra, 2 neutrino approx. 0.6 60 0.5 50 0.4 0.3 40 0.2 30 0.1 20 0 10 -0.1 -0.2 10 20 30 40 50 60 Covariance matrix is fractional – we thought it should not vary with oscillation parameters However, when the matrix is remade with the approximate oscillated spectra it shows significant differences S. Jones (UCL) DUNE LBL April 29, 2019 4 10

Comparison with same Z axis scale FD covariance matrix with unoscillated spectra FD covariance matrix with oscillated spectra, 2 neutrino approx. 0.6 0.6 60 60 0.5 0.5 50 50 0.4 0.4 0.3 0.3 40 40 0.2 0.2 30 30 0.1 0.1 20 20 0 0 10 10 -0.1 -0.1 -0.2 -0.2 10 20 30 40 50 60 10 20 30 40 50 60 To calculate the oscillated spectrum weighted ν µ events by 1 − sin((1 . 27 × 0 . 00232 × 1300) / E ) 2 and ν e events by sin((1 . 27 × 0 . 00232 × 1300) / E ) 2 S. Jones (UCL) DUNE LBL April 29, 2019 5 10

Fitting with oscillated covariance matrix 5 4 3 2 CovMx with osc No FD CovMx 1 CovMx with no osc − − 1 0.5 0 0.5 1 The oscillated covariance matrix case (black) gives lower CPV sensitivity than unoscillated matrix (blue) but still greater than expected (red) S. Jones (UCL) DUNE LBL April 29, 2019 6 10

Further tests with CAFAna The covariance matrix is produced using the CAFs but externally to CAFAna. I have attempted to remake the matrix at different oscillation parameters to check for noticeable variation S. Jones (UCL) DUNE LBL April 29, 2019 7 10

Oscillation parameter comparison δ = 0, θ 23 lower octant, IH δ = π , θ 23 upper octant, NH 60 60 0.1 0.05 40 40 0.05 0 0 20 20 -0.05 -0.05 0 0 0 20 40 60 0 20 40 60 Can see just by eye there are differences between the two matrices Still unable to produce a matrix similar to the ones produced by Chris’s method Going to try and run my previous fits again with these matrices to see effect S. Jones (UCL) DUNE LBL April 29, 2019 8 10

Unoscillated spectra FHC ν µ RHC ν µ 45000 40000 40000 35000 35000 30000 30000 25000 25000 20000 20000 15000 15000 10000 10000 5000 5000 0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10 ReconstructedE ν ReconstructedE ν FHC ν e RHC ν e 35000 35000 30000 30000 25000 25000 20000 20000 15000 15000 10000 10000 5000 5000 0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10 ReconstructedE ν ReconstructedE ν S. Jones (UCL) DUNE LBL April 29, 2019 9 10

Oscillated spectra FHC ν µ RHC ν µ histCV FD FHC numu histCV FD RHC numu 25000 Entries 398102 Entries 384537 35000 Mean 4.028 Mean 4.722 Std Dev 2.618 Std Dev 2.705 30000 20000 25000 15000 20000 15000 10000 10000 5000 5000 0 0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10 ReconstructedE ν ReconstructedE ν FHC ν e RHC ν e histCV FD FHC nue histCV FD RHC nue 2200 Entries 355036 Entries 335730 1800 Mean 2.967 Mean 3.074 2000 Std Dev 1.135 Std Dev 1.247 1600 1800 1400 1600 1400 1200 1200 1000 1000 800 800 600 600 400 400 200 200 0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10 ReconstructedE ν ReconstructedE ν S. Jones (UCL) DUNE LBL April 29, 2019 10 10