Thus, the increase of the spread could mean that, by reducing the uncertainty in each individual model, we now quantify sources of error (related to resolution) that were previously unmeasured by past ensembes. In other words, higher resolution does not increase uncertainty per se, but improves the estimation of the “global” uncertainty (at the collective scale of the ensemble).
Your question is complex. Not sure we have the answer yet as the ensemble of very high resolution models is still very new. We keep it for later.
My feeling is that model spread is often said to quantify uncertainty, and that an increase (or reduction) of the model spread mirrors an increase (or reduction) of uncertainty. But an increase of model spread can also just mean that uncertainty is now better estimated. I’d love to know what you think about the book! Haha. Would be very nice to talk again soon.
