Scale and Scheme Variations in Unitarized NLO Merging
Precision background predictions with well-defined uncertainty estimates are important for interpreting collider-physics measurements and for planning future high-energy collider experiments. It is especially important to estimate the perturbative uncertainties in predictions of inclusive measurements of jet observables, that are designed to be largely insensitive to non-perturbative effects such as the structure of beam-remnants, multi-parton scattering or hadronization. In this study, we discuss possible pit-falls in defining the perturbative uncertainty of unitarized next-to-leading order multi-jet merged predictions, using the PYTHIA event generator as our vehicle. For this purpose, we consider different choices of unitarized NLO merging schemes as well as consistent variations of renormalization scales in different parts of the calculation. Such a combined discussion allows to rank the contribution of scale variations to the error budget in comparison to other contributions due to algorithmic choices that are often assumed fixed. The scale uncertainty bands of different merging schemes largely overlap, but differences between the "central" predictions in different schemes can remain comparable to scale uncertainties even for very well-separated jets, or be larger than scale uncertainties in transition regions between calculations of different jet multiplicity. The availability of these variations within PYTHIA will enable more systematic studies of perturbative uncertainties in precision background calculations in the future.