Here are some questions from my friend Andrew, posted on Twitter.
1. How do you present multivariate uncertainty in climate to policy makers/journalists/the public?
2. How do you represent uncertainty in a climate prediction when the inputs and physics are uncertain?
3.How do you estimate the 'true' uncertainty in climate predictions when many of the models share the same components?
4. How do you estimate uncertainty in a climate model when the focus is on increasing resolution at the expense of speed?
5. How do you deal with the fact that the present climate is uncertain in every location at all time points?
6. How do you account for the real concern being in climate extremes which are much harder to estimate with uncertainty than means?
Good, open questions all
Doug McNeall's blog 