I can see the reasons why they need to be careful. Long range forecasting has been causing them issues for a while now, and the denier-sphere is quick to jump on anything which might be seen as slightly over estimated in projections. For the quiet life estimating with a conservative bias is a lot easier to do. From a risk management perspective its exactly the wrong thing to do, you need to know the worst case scenarios with probabilities) so that you can try to mitigate against them in part or full.
I think that the 3-5 year timeframe is probably right Craigsisland, although by then the ice-free transition (for the summer at least) is likely to be complete. The evidence of the change in the jet stream will be worldwide, so changes over America and Asia will support research nearer to home.
Trying to predict outcomes in the medium term for a chaotic system undergoing change is like trying to hit a moving dartboard from a boat in rough seas. I don't envy their job.