We incorporated accelerator-data-driven tuning for hadronic interaction in our atmospheric neutrino flux calculation which has been used for the analysis of atmospheric neutrino oscillations at Super-Kamiokande. This new approach allows us to evaluate the flux uncertainty more directly compared to conventional tuning using atmospheric muons.
We tuned the hadronic interaction model in our calculation based on recent hadron production data measured by fixed-target accelerator experiments. The neutrino flux calculated with this new tuning is 5--10\% smaller but still consistent with our previously published prediction within its uncertainty.
The uncertainty associated with the new tuning was also evaluated based on the measurement errors of the accelerator data. Flux uncertainty was less than 6\% in 0.2 $< E_{\nu} <$ 10 GeV/$c$ region, which is an improvement over the conventional tuning. We performed the uncertainty evaluation in $< 1$ GeV/$c$ region where the conventional tuning only provided the conservative uncertainty estimation.