An Improved Weighting Algorithm to Achieve Software Compensation in a Fine Grained LAr Calorimeter

An improved weighting algorithm applied to hadron showers has been developed for a fine grained LAr calorimeter. The new method uses tabulated weights which depend on the density of energy deposited in individual cells and in a surrounding cone whose symmetry axis connects the interaction vertex with the highest energy cluster in the shower induced by a hadron. The weighting of the visible energy and the correction for losses due to noise cuts are applied in separate steps. In contrast to standard weighting procedures the new algorithm allows to reconstruct the total energy as well as the spatial energy deposition on the level of individual calorimeter cells. The linearity and the energy resolution of the pion signal in the momentum interval 2 GeV/c <= p <= 20 GeV/c studied in this analysis are considerably improved in comparison to the standard weighting algorithm as practiced presently by the H1 collaboration. Moreover the energy spectra reconstructed with the new method follow in a broad interval a Gaussian distribution and have less pronounced tails.