Our procedure does not address the issue of how parameterizations can vary for different flow types. However, Edeling et al.  [9] carried out separate calibrations for a set of 13 boundary-layer flows. They summarized this information across calibrations by computing Highest Posterior-Density (HPD) intervals, and subsequently represent the total solution uncertainty with a probability-box (p-box). This p-box represents both parameter variability across flows, and epistemic uncertainty within each calibration. A prediction of a new boundary-layer flow is made with uncertainty bars generated from this uncertainty information, and the resulting error estimate is shown to be consistent with measurement data. This approach is helpful, but it might be extended further by modelling proximity across flows through a distance that would relate to the flow characteristics in order to borrow strength across calibrations instead of splitting the calibrations and then merging the outcomes afterwards. This is a challenging but attractive venue for future research.
