The uncertainty of models is still very high for the time horizon relevant for policymakers. I guess there's work underway to cross-sheck the C3S data with other (Copernicus, ground or other source) data to reduce it? If yes, pls. give a few links!

I would say that the comparison between climate model outputs and present-day observation does help us understand what is working well (at process level) and what is working less well in the current generation of climate models. This -unfortunately though- is only one of the many sources of uncertainty and thus such a comparison does not necessarily reduce the overall uncertainty in the predictions.