Studies of Thresholds and Level 2 Filters

Threshold Study

The question of how much of how accurately do we need to know our thresholds was posed. I conducted a study of signal monte carlo where I first decreased the threshold on every OM by 30 percent (m_30). Next, I increased the threshold by 30 percent (p_30). I also tried decreasing every other module's threshold by 30 percent (m_30.ev) and increasing every other module's threshold by 30 percent (p_30.ev). The files were generated using the geometry files and pulse shapes as given on Stephen Hundermark's web page. The table below sumarizes the results and compares them to "normal" thresholds (normal). The results have been weighted for 1 day of atmospheric signal monte carlo.

Atmospheric Neutrino Trigger Rates
File Trigger Level(# events) Level 1(# events) Level 2(# events)
numu.m_30
30.8 ± 0.86
29.2 ± 0.84
26.4 ± 0.77
numubar.m_30
10.0 ± 0.11
9.4 ± 0.11
8.5 ± 0.10
numu.m_30.ev
30.9 ± 0.89
28.1 ± 0.79
25.7 ± 0.74
numubar.m_30.ev
9.1 ± 0.09
8.8 ± 0.09
8.0 ± 0.08
numu.p_30
24.8 ± 0.60
22.8 ± 0.56
20.7 ± 0.51
numubar.p_30
8.5 ± 0.09
8.0 ± 0.08
7.3 ± 0.08
numu.p_30.ev
26.4 ± 0.66
24.6 ± 0.63
22.0 ± 0.57
numubar.p_30.ev
8.2 ± 0.08
7.7 ± 0.07
7.1 ± 0.07
numu.normal
29.2 ± 0.76
27.3 ± 0.71
24.5 ± 0.64
numubar.normal
9.2 ± 0.09
8.7 ± 0.09
7.9 ± 0.08

I also produced some graphs of various characteristics of the signal monte carlo. In these graphs the results of monte carlo files generated by nus and nubars are combined. The black lines represent "normal" thresholds, red lines represent thresholds reduced by 30% for ALL OMs, green lines are thresholds reduced by 30% for EVERY OTHER OM, blues lines are where thresholds were increased by 30% for ALL OMs and the yellow lines are where thresholds were increased by 30% for EVERYOTHER OM.

log jkchi.ps

log jkprob.ps

log jkrchi.ps

log nch.ps

log zenith.ps

All distrubutions are shown on log plots. Jkchi, jkrchi and zenith distrubutions are of the best fit. Jkprob is of the line fit.

Level 2 Reconstruction Study

We were curious to know what effect different choices for the number of iteritations of minimizations had on the number of triggers for Level 2 reconstruction. So, I took a different number of iterations for both the line fit and Peter Steffan's direct walk method and compared them for 1 day of atmospheric weighted signal monte carlo and data. Cuts were made on the zenith angle of the best fit in each case > 80.

Level 2 Reconstruction Study
Reconstuction Atmospheric Signal MC(# events) Data(# events)
2 Minimizations, Line Fit
32.92 ± 0.74
7151
4 Minimizations, Line Fit
33.07 ± 0.75
5467
8 Minimizations, Line Fit
33.17 ± 0.75
4524
16 Minimizations, Line Fit
33.08 ± 0.75
4246
2 Minimizations, Direct Walk
32.06 ± 0.71
4817
4 Minimizations, Direct Walk
32.53 ± 0.73
4148
8 Minimizations, Direct Walk
32.53 ± 0.73
3738
16 Minimizations, Direct Walk
32.87 ± 0.74
3666