2005 AMANDA Monte Carlo using Photonics

John Kelley, UW-Madison, October 2006

1. Introduction

Production of 2005 AMANDA Monte Carlo is nearing completion, using the Photonics software package [uu.se] for photon propagation in South Pole ice. This page will archive technical details of this simulation production and will be updated as the simulation progresses.

• AMASIM version: aluminum-fix [physto.se]. Compilation instructions that worked for me (in Madison) are available here.
  
• AMASIM 2005 calibration files: available on M. Walter's web page [desy.de] (use the aluminum versions!).
  
• scatserver [physto.se] (the interface between AMASIM and Photonics) version 2.
  
• Photonics tables: fully detailed Millennium ice (dz=20m, dtheta=10deg) wl06v200 [uu.se] , AMANDA coordinates, AMANDA depth span.
  
• For dCORSIKA:
  • version 6.5
  • SIBYLL high-energy hadronic interaction model
  • Hörandel parametrization
  • includes string trigger (with downscaling = 1)
  • DSLOPE = -1 (harder spectrum: see my notes on reweighting (PDF, 70K))
  • oversampling of 2x near the horizon (detector zenith angle > 85 degrees)
    
• For nusim:
  • includes string trigger (with downscaling = 2: HOWEVER, according to Johan et al., string trigger downscaling is not functional in the Aluminum-fix release. See his wiki [physto.se]. I have not checked this.)
    
• For coincident muons:
  • Photonics 1.5_r1e
  • Same as dCORSIKA above unless specifically mentioned.
  • Preliminary: time window of +/- 10 us.
  • Notes on reweighting to follow soon.
    

2. Trigger-level Comparison with PTD

A preliminary comparison has been performed with integral PTD (mamint) ice, using O(100) nusim upgoing neutrino / antineutrino MC files with an E^{-1} spectrum (no weighting: for weighted comparisons, see this page). The number of files has been normalized in order to compare trigger rates. The plots below are at trigger level, with no reconstructions performed. Significant differences are visible in the COGZ structure. More subtle differences show up in the true energy of triggered events, and in the low Nchannel and Nhits spectrum: AMANDA+Photonics is apparently more sensitive to low-energy events than PTD. I do not see as dramatic difference in the number of hits with high Z as reported previously.

Here are a few of the same plots (Nch and Nhits) in log scale:

For ongoing comparisons of dCORSIKA and nusim with data after filtering, please see Jim Braun's page.

3. Production with simuperl

Paolo Desiati's simuperl package has been updated to work with Photonics. Much of the work was completed by Rafael Lang (see his web pages here [desy.de]). I have cleaned up the final few details to allow use with released versions of AMASIM and scatserver, and to optimize Photonics table loading. Details on how to install the new simuperl can be found here.

For those interested, simulation involves splitting an input F2K file into zenith angle ranges, spawning a scatserver which loads only the tables needed for that zenith angle range, iterating, and then gluing all the results back together at the end. All of this is transparent to the user.

4. Filtering

Filtering of nusim is complete, using the 2005 filtering scheme for data (see David Boersma's web page here). Generation and filtering of dCORSIKA continues using the GLOW and NPX clusters.

Please feel free to contact me if you have any questions or comments.

Back to home page