Hadronic Calibration Page

Information

I am performing checks of the HadrooII hadronic calibration using the H1Lt Analysis Framework and the code can be found in the hadronicChecks cvs package. Now you can check too, with this checkout script! (change the script to specify your $H1ANALYSISDIR)
Latest Results (click on the for the plots)

All plots are made using the full high Q2 inclusive selection (see here) and have the HadrooII calibration applied unless otherwise stated.

The plots on this page use H1oo version 2.4.25. Go here for 2.5.x plots. Or here for 2.2.13 FSCOMB High Q2 NC sample (P.Laycock).
The four main plots: pth/ptda vs ptda, pth/ptda vs gammah^e and the same dependence for yh/yda: Each set of these sets of TProfile plots shows four kinematic balances: pth/pte, yh/yda, pth/ptda, pte/ptda in different regions of pte: all pte, pte < 25 GeV, pte > 25 GeV and in the "golden region" of 12 GeV < pte < 50 GeV and gammah > 15o: Fits are then performed for pte/pth, pth/ptda and yh/yda across bins in gammah, gammah^e and ptda, using 0.5*RMS and 1.1*RMS as the lower and upper bounds respectively: The follwing plots show the mean of each of the fits of each of the kinematic balances, across the bins of gammah, gammah^e and ptda. Note that for the ptda plots the first bin is not used. Bands at 2% and 4% of 1.0 are also shown: Finally, these plots show the ratio of Data/MC for the above plots. Bands at 2% and 4% of 1.0 are also shown: Would now like to do pull distributions.....
Conclusions (so far...)

  • The calibration looks best at high pt and low gammah(^e).

  • The control plots show good agreement between data and MC; at pte > 25 GeV the agreement is especially good.

  • The mean values of the Gaussian fits performed agree betwen data and MC within 2 - 3%, but at very low values of gammah(^e) the mean values are significantly less than 1.0 by more than 2%. This was observed in the previous (FSCOMB) calibration - see here. This trend also occurs at low values of ptda. The distributions also show some sinusoidal behaviour across gammah(^e).

  • The Data/MC plots (derived from the fits) are mostly within 2%, with the gammah^e distributions showing better results than the gammah distributions. Similarly to the mean fit plots, there is a tendancy to lower values (away from 1.0) as gammah(^e) or ptda is decreased. Furthermore, there is a systemtic shift on the absolute scale of about 1% (ie the average of Data/MC is about 0.99) visible moreso in the gammah(^e) plots. This was also seen in the Marseille gammah and ptda distributions. This may be due to iron leakage and will be investigated in the future.

  • However, in the full phase space (ie now not looking at the fits), there is a definite systematic shift, where the data is always below the MC, typically of 1 - 2% (for example in these TProfile distributions).
    • Links

    David South
    Last modified: Wed Feb 18 15:04:35 MET 2004