From Carl's email

When Kevin and Zilong tuned Perugia 2012 to match the STAR 200 GeV pi^+/- cross
section, they initially looked at a couple of parameters that influence the MB
pion spectra before settling on PARP(90).  Another "interesting" parameter is
PARP(91) that specifies the intrinsic kT.  Perugia 2012 sets it at 1.0 GeV/c.
But by rights, one would expect the (non-perturbative) intrinsic kT to be down
closer to Lambda_QCD, and recent SIDIS measurements tend to support that.  So it
wouldn't hurt to see what effect reducing PARP(91) to 0.5 GeV/c (or perhaps even
somewhat less) would have on the pi^+/- cross section, UE generation, and jet
properties.

The things that make this parameter particularly "interesting" are that: (a)
  There is a good physics reason why it should be smaller, and now data to
  support that reason, and (b) This is one of the few parameters that is easy to
  change via event reweighting after a simulation sample has been produced.  The
  only harm is that the reweighting increases the statistical fluctuations
  relative to a sample of the same size that is produced up-front with the
  modified parameter value.

If (big IF!!!) a reduced intrinsic kT preserves the good description of jet
properties while providing a better match to the observed UE rate, that would
let you reduce another systematic uncertainty in the energy scale.