[MET]

Distribution of the missing transverse energy (MET) reconstructed from all calorimeter cells with |E|>2sigma(noise) at EM scale. The distributions of cosmic-ray simulation and the 2008 cosmic-ray data are normalized to the 2009 data in the range 100< Missing Et <300 GeV. Only events which have reconstructed jets with Pt>20 GeV are included. The Monte Carlo simulation does not include air showers and also contains an imperfect modeling of electronic noise. eps file

[SumEt]

Distribution of the scalar sum of the transverse energy reconstructed from all calorimeter cells with |E|>2sigma(noise) at EM scale. The same normalization factor as in the figure above is applied. Only events which have reconstructed jets with Pt>20 GeV are included. The Monte Carlo simulation does not include air showers and also contains an imperfect modeling of electronic noise. eps file

[Jet Pt]

Distribution of reconstructed jet Pt from the 2008/2009 cosmic-ray data and cosmic-ray simulation. The same normalization factor as in the figure above is applied.The ATLAS Anti-Kt4 Jet algorithm is used. Only jets with Pt>20 GeV are included. The Monte Carlo simulation does not include air showers and also contains an imperfect modeling of electronic noise. eps file

[Energy ratio of leading cells in jet]

Distribution of energy ratio of leading cells in the jet. The Ratio_LC is defined as the sum of the energy of the 2 most energetic cells in the hadronic calorimeter and 32 energetic cells in the EM calorimeter divided by the total reconstructed jet energy. Only jets with Pt>20 GeV are included. The same normalization factor as in the figure above is applied. The Monte Carlo simulation does not include air showers and also contains an imperfect modeling of electronic noise. eps file

[Jet Rcone]

Distribution of the radius of the reconstructed jet cone. Only jets with Pt>20 GeV are included. The same normalization factor as for the figure above is applied. The Monte Carlo simulation does not include air showers and also contains an imperfect modeling of electronic noise. eps file

[Jet EM fraction]

The jet electromagnetic energy fraction, defined as the ratio of the energy of the jet recorded in the EM calorimeter divided by the total energy of the jet. Only jets with Pt>20 GeV are included. The same normalization factor used for the above figure is applied. eps file

[MET]

Distribution of missing transverse energy in cosmic-ray data, data mixed with simulation events, and dijet simulation. Only events which have jets with Pt>20 GeV and |eta|<2.5 are included. No clear difference is seen in the distribution of MET before and after mixing of cosmic-ray data with simulation events. eps file

[SumEt]

Distribution of SumEt from cosmic-ray data, mixed events, and dijet simulation. Only events which have a reconstructed jet with Pt>20 GeV and |eta|<2.5 are included. A significant difference between cosmic-ray data and mixed events is observed. eps file

[Jet Pt]

Distribution of reconstructed jet Pt from cosmic-ray data, mixed events and dijet simulation. Only jets with Pt>20 GeV and |eta|<2.5 are shown. eps file

[Energy ratio of leading cells in jet, jet Rcone and EM fraction]

Distribution of Ratio_LC, jet Rcone and jet EM fraction from the cosmic-ray data, mixed events and dijets simulation. Only reconstructed jets with Pt>20 GeV and |eta|<2.5 are shown. A slight difference in the distribution of Ratio_LC and jet Rcone between cosmic-ray data and mixed events is observed. The distribution of jet EM fraction in mixed events is wider than cosmic-ray data for the region of low EM fraction. A significant separation between cosmic-ray "fake" jets and real jets from dijet simulation is observed. The three jet variables are used to build Probability Distribution Functions and a Log Likelihood Ratio for rejecting cosmic-ray background. Rcone.eps Ratio.eps EMfraceps

[Log-Likelihood Ratio]

Distribution of the Log-Likelihood Ratio (LLR) obtained with Ratio_LC, jet Rcone and jet EM fraction from the mixed cosmic-ray data and dijets simulation. The ATLAS Anti-kt4 jet algorithm is used. Only reconstructed jets with Pt>20 GeV and |eta|<2.5 are included. eps file

[Cosmic-ray background rejection vs Jet efficiency]

Dijet efficiency versus cosmic-ray background rejection. The ATLAS Anti-kt4 Jet algorithm is used. Only reconstructed jets with Pt>20 GeV and |eta|<2.5 are included. Cosmic-ray background rejection is obtained from the mixed events, while the jet efficiency is obtained from the dijet simulation. The red circles correspond to different cut values of the LLR obtained with all the three jet variables, Ratio_LC, Rcone and EM fraction. The blue squares are from cutting on the jet EM fraction only. The black stars are from a 2-dimensional PDF of Ratio_LC vs Rcone. For the LLR obtained with all the three variables, one can see the improvement on cosmic-ray background rejection over the other two single LLRs. eps file

[Application]

Summary of all cuts on rejecting cosmic-ray background. The ATLAS Anti-kt4 Jet algorithm is used. Only reconstructed jets with Pt>20 GeV and |eta|<2.5 are included. The upper plot shows the distribution of jet Pt in cosmic-ray data without requiring a jet cleaning cut (the blue circles), after applying the cut on the EM fraction (the blue open circles) and after applying the cut on LLR (the blue squares) at a fixed jet efficiency of 98%. The three lines in the same figure show the distributions of the jet Pt from dijet simulation before cleaning cuts, after the EM fraction cut and after the LLR cut, respectively. Here, the LLR is obtained using all three jet variables: jet Rcone, Ratio_LC and EM fraction. Similar plots for mixed cosmic-ray data and simulation events compared to cosmic-ray data are shown in the bottom figure. cosmic.eps mixed.eps