OCA_REL

The next stage determines the relative acrylic tile extinctions as a function of wavelength and the relative wavelength dependence of the PMT efficiencies. For a given laserball position within 200cm of the center of the AV (so photons are not too far from normal incidence on AV and PSUP surfaces), sets of PMTS are chosen for which recorded photons traverse a particular acrylic tile. This amounts to $\sim100$ PMTs per tile. After a first-order correction for the relative PMT efficiency above 400nm, it is assumed that the median PMT efficiency at lower wavelengths based on any such set of $\sim100$ tubes should be the same. Thus, after further corrections to the observed prompt-peak intensity at each wavelength for solid angle, D2O/H2O extinction and the Poisson probability for multiple hits within a single laser pulse, observed variations in the median values corresponding to each acrylic tile are taken to be the result of differences in the relative optical extinction of that tile. These values and their estimated uncertainties are assigned to the variables OCA_Rel_Acr(IW,NTile,1) and OCA_Rel_Acr(IW,NTile,2), respectively, where $IW$ is the wavelength index of the laserball and $NTile$ is the acrylic tile number. Furthermore, differences between each tube of a given set and the corresponding median value is taken to represent small variations in the wavelength dependence of the PMT/reflector efficiency. These values and their estimated uncertainties are assigned to the variables OCA_Rel_QE(IW,NPMT,1) and OCA_Rel_QE(IW,NPMT,2), respectively, where $NPMT$ is the PMT number.