Atmospheric muon neutrinos from ANIS and nusim

John Kelley, UW-Madison, March 2008

1. Introduction

I have performed a preliminary comparison of atmospheric muon neutrino flux differences as simulated by ANIS and nusim. The input flux is the Barr et al. ("Bartol") prediction, as implemented in the NeutrinoFlux class. The simulation parameters are as follows:

Both:

• 80 < zenith < 180
• 1 GeV < energy < 100 PeV
• spectral gamma = 1.5
• identical 2005 Amasim settings
• filtered with 2005 filtering + Zeuthen final cuts
• Bartol atmospheric flux weighting

ANIS-specific settings:

• interaction volume: cylinder, h/2=400m, r=400m
• extendrange = 1
• MMC uses mediadef file to simulate both bedrock and ice

nusim-specific settings:

• detector volume: cylinder, h=600m, w=200m
• nusim propagates muons to rock boundary; MMC simulates the ice

2. Results

The number of events and their energy distribution was found to be similar at the final cut level. Specifically:

• 8% fewer events overall in ANIS
• median energy is 3% lower in ANIS (612 GeV in ANIS vs. 630 GeV in nusim)
• zenith distributions are comparable in shape

Here are a few comparison plots:

Figure1: Neutrino energy at final cut level for ANIS and Nusim.

Figure 2: Reconstructed zenith angle for ANIS and nusim.
The left edge is vertical events, right is horizontal.

Figure 3: Number of channels hit for ANIS and nusim.

The E^{-2} spectra are also very similar, with 7% fewer events in ANIS, and similar median energies (55 TeV in ANIS vs. 52 TeV in nusim):

Figure 4: Energy spectra of ANIS vs. nusim for an E^-2 spectrum (final cut level).

4. Behind the scenes

Some of the difference in simulation methodology can be seen via the interaction probability difference. The following plot shows the distribution of event interaction probability weights as a function of energy (Si1 in ANIS, Intwght in nusim):

Figure 4: Interaction probability vs. neutrino energy for ANIS (blue) and nusim (red) at L3.

Understanding this plot requires some explanation. First, there are two distinct bands for neutrinos and antineutrinos. Second, NC interactions along the way through Earth (followed by CC interactions at lower energy) cause the faint low-probability band in nusim (not sure if this is simulated in ANIS). Third, the rolloff at high energies due to Earth absorption is seen in nusim (different roll-offs for different zenith angles), whereas ANIS simply does not output these events.

The difference in the slope in the ANIS interaction plot is because of the energy-dependent interaction volume (scales with muon range).

Any differences do not seem to be coming from the underlying cross sections (CTEQ5 in anis vs. MRSG (1996) in nusim), which are comparable in this energy range:

Figure 5: Cross sections in anis and nusim vs. energy.

Back to home page