Nus 2019 -- FHC covariance¶
The absolute calibration shift is treated as a single systematic, with 5% energy scale up treated as the +1 sigma shift and 5% energy scale down treated as the -1 sigma shift. To be conservative, the two detectors' covariance matrices are considered to be uncorrelated.
The calibration shape shift is a one-directional systematic, which applies a shift for positive sigma and has no impact on the spectrum for negative sigmas. This systematic is treated as uncorrelated between the two detectors.
The light level systematic is treated as a single systematic, with the upward shift applied for positive sigma and the downward shift applied for negative sigma. This systematic is treated as uncorrelated between the two detectors.
The Cherenkov systematic is a one-directional shift applied for positive sigma shifts; no shift is applied for negative shifts. Since this systematic is a model parameter, it is treated as fully correlated between the two detectors.
The kaon systematic is a two-directional systematic, treated as fully correlated between the two detectors.
The cross-section tune systematic is a one-directional systematic. As such, it is applied for a positive sigma, with no spectrum reweighting for negative shifts.
The neutron systematic is a two-directional systematic, and is treated as fully correlated between the two detectors.
Each individual component of the beam transport systematic is considered as a separate two-directional shift. A more detailed accounting of each individual component will be forthcoming.
All GENIE knobs will be considered as separate systematics, each fully correlated between the two detectors. For the time being, each is implemented as a separate two-directional systematic, although a more detailed categorisation will be performed shortly to classify non-Gaussian and one-directional systematics.
A matrix is populated according to the average shift across all 100 shifted universes, for each beam component. There is currently no ISyst associated with this systematic, so a set of PCA ISysts should be produced for cross-referencing. This systematic is correlated between the two detectors.