Kinematics Related Systematic Uncertainties via MCEEP

1
Kinematics RelatedSystematic Uncertainties via
MCEEP

• P.E. Ulmer
• Old Dominion University
• 12/11/02
• Hall A Analysis Workshop

2
Breaking NewsCross sections depend on
kinematics! Uncertainties run wild!

• Coincidence cross section can vary strongly with
  kinematics.
• Results in systematic uncertainties
• need to evaluate changes in cross section for
  variations of kinematical quantities.
• Account for kinematical constraints
• For example, fixed missing mass.
• May include constraints from various calibration
  measurements, such as H(e,ep)
• Must acceptance average derivatives.
• Its a snap with new MCEEP tools
• Cross sections handled at present.
• Other observables could be added easily.
• Satisfaction guaranteed or your money back.

3
Procedure

• MCEEP Hbook file

4
Process NtupleFortran Program systerr

• Start with MCEEP Ntuple, containing Transport
  coordinates at target
• Vary nine quantities, in turn
• (beam, scatt. electron, ejectile)
• x
• (delta, phi, theta)
• Produce new Ntuple, consisting of original
  variables plus 10 cross sections (nominal and
  nine shifted).
• Program links to MCEEP subroutines and has access
  to its physics models.

5
Acceptance AveragePAW systerr.kumac

• Error sum positive definite quantity
• Must first acceptance average.
• Sum the weights
• Produce vectors of summed cross sections (10 in
  total).
• Bin vectors in terms of any kinematical quantity
  within Ntuple.
• Diagnostic histograms
• Fractional derivatives of cross section with
  respect to each of the nine varied kinematical
  quantities.

6
Combine Errors Fortran Program toterr

• Produce cross section uncertainty, given
  kinematical uncertainties and correlations. For
  each bin, form

7
More Information

• www.physics.odu.edu/ulmer/mceep/mceep.html
• Includes
• Sources
• Installation Instructions
• User manual
• In particular, see /mceep/systerr/README
• JLAB-TN-02-015
• Systematic Uncertainties in E89-003
• (K. Fissum P.E. Ulmer)
• See http//hallaweb.jlab.org/publications/Techno
  tes/files/2002/02-015.ps

8
Figures

• Both figures are based on Experiment E94-004
  D(e,ep)n cross section vs. Pm
• Cross section derivatives for each of the nine
  quantities
• Kinematics centered on Pm100 MeV/c
• Units fractional derivatives per 1 mr or per
  10-3 in momentum.
• Total error vs. Pm
• All kinematics for E94-004 included, from Pm0 to
  Pm500 MeV/c (central values)
• Uncorrelated analysis
• Correlated analysis assumes constraints from
  1H(e,ep)