The Disp Method for Analysing Large Zenith Angle Gamma-Ray Data

arXiv:1109.6044v1 [astro-ph.IM] 27 Sep 2011 32ND INTERNATIONAL COSMIC RAY CONFERENCE, BEIJING 2011 The Disp Method for Analysing Large Zenith Angle Gamma-Ray Data G ¨UNES¸ D. S¸ ENT ¨URK1 FOR THE VERITAS COLLABORATION2 1Physics Department, Columbia University, New York, NY 10027, USA 2see J. Holder et al. (these proceedings) or http://veritas.sao.arizona.edu/conferences/authors?icrc2011 gunessenturk@gmail.com Abstract: The Disp method is an algorithm that is used for reconstruction of primary gamma ray direction in ground- based atmospheric Cherenkov telescope experiments -measuring very-high-energy (VHE) gamma rays in the energy range between 100GeV and 30 TeV. In general terms, the geometric information obtained from one single shower image is sufficient for the algorithm to find the sky location of the primary. Various versions of the Disp method were implemented and used in the past. In this study, we present a multi-dimensional implementation of the Disp method for the VERITAS instrument and show (using Monte Carlo simulations and the Crab Nebula observations) that it significantly improves the angular resolution for large-zenith-angle (LZA) observations. We also applied the disp method to VERITAS data taken from the galactic center region which is detected by VERITAS. Keywords: Analysis methods, large-zenith-angle observations 1 Motivation The Disp method has been the default direction reconstruc- tion algorithm for single telescope observations (e.g. [1, 2, 3]). The introduction of new generation ground-based gamma-ray instruments operating in array mode with mul- tiple telescopes allowed for new and more accurate tech- niques for direction reconstruction. However, due to a larger (on average) impact parameter of the air showers with respect to the center of the of the IACT array and projection effects, these (geometrical) reconstruction tech- niques do not perform well at LZA. The Disp method com- pensates for this loss in quality of the angular reconstruc- tion. A description of the VERITAS analysis steps prior to direc- tion reconstruction (calibration, image cleaning and image parametrization) can be found in [4]. At the latter step, a second moment parametrization with Hillas parameters is performed on each shower image [5]. See Figure 1b for a detailed representation of relevant Hillas parame- ters. Disp is the angular distance between the image cen- troid and the real source location on the camera plane. An- other important image parameter is size, defined as the total integrated charge in image pixels, representing a measure of brightness for the image. Figure 1a shows the stereo- scopic image of an air shower after parametrization. The default VERITAS algorithm looks for the point with the smallest total distance from each major axis: this is the re- constructed direction. The error associated with this calcu- lation is inversely related to the angles between the major axes. In other words, images with close to parallel major axes (Figure 1c), which more often occur for LZA data, result in large errors in direction reconstrucion. 2 Description of the Algorithm The Disp method makes use of multidimensional lookup tables that contain the image information coming from Monte Carlo simulated air shower data. For each simu- lated shower image, the disp table stores the Hillas pa- rameters size, width, length and disp. These quantities are related to each other by the following geometrical argu- ment: the projected images of air showers on the camera plane should look circular at the center, and become flatter as they move towards the edge. Therefore, disp should in- crease with increasing ellipse flatness, which can be quan- titatively described by width and length parameters. The dependence on size has been verified from parameter distri- bution plots. Additional dimensions present in the lookup tables are zenith and azimuth angles, noise level and tele- scope ID. The way the reconstruction algorithm works is the following: for a given shower image, size, width and length are calculated and the corresponding disp is read from the lookup table. This tells us how far the arrival di- rection is from the image centroid along the major axis, but it does not tell at which side of the ellipse it is located, also known as head-tail ambiguity [6]. To eliminate this problem, the closest cluster of points is picked, one com- ing from each ellipse. In this way, the arrival direction is estimated for each telescope image and a weighted average is calculated. The novelty of this implementation is that S¸ ENT ¨URK et al. THE DISP METHOD Figure 1: a) Representation of an air shower stereo image, seen by three different telescopes. b) Detail from a), Hillas parameters that are used in the Disp algorithm. c) Stereo images for LZA events tend to have mostly parallel major axes of ellipses. In all images, reconstructed and real source locations are represented by green and red dots respectively. width and length paramete

This content is AI-processed based on open access ArXiv data.