The CFMIP workshop provided the excuse we were looking for in order to visit Boulder again. Last year this meeting had been held in Tokyo which, on top of being a long way away, also directly clashed with the PMIP workshop that jules was duty-bound to attend. The main focus of the CFMIP workshop is the simulation of clouds in GCMs and how they might be affected by climate change. This is our largest single source of uncertainty in the climate system. We are not really CFMIP people and didn’t plan to present anything, in fact what drew us there more than the workshop content itself was that our two ongoing research collaborations involve a whole bunch of the attendees so it was a good chance for us all to meet face to face for a change rather than via Skype and email. These meetings were fitted in to the odd hour or two here or there in some of the breaks rather than filling a full week as we did in Edinburgh earlier this summer. (Did I not blog that? How remiss of me.) Additionally, jules also tapped up a local for a role as exec editor of GMD, continuing our (their!) policy of rotation and inclusion of new people and ideas. All of that plus the famed Boulder weather and food was sufficient to entice us over for a bit of a holiday that included the workshop.
I had feared that the workshop was going to be strongly focussed on the details of simulating clouds in climate models but the first day, as well as providing an overview of CFMIP, also discussed the notion of feedbacks and climate sensitivity more generally. This is highly relevant to our current research so we paid careful attention. Interestingly, participants seemed resigned to a large proportion of their simulations and analyses not being completed in time for the next IPCC report. In principle the IPCC merely summarises the science and does not commission still less undertake it, but in reality there has usually been a close link between the IPCC and CMIP timetables. With people trying to maximise development time for their climate models, there is limited computing time to do all of the various MIP experiments, especially if some mistake is found and a bunch of simulations have to be repeated. Of course it’s inevitable that the IPCC report is a little out of date because of the lengthy process underpinning it, but it would be unfortunate if lots of new and exciting results were being published around the time that the IPCC produces a report based on much older knowledge. On the other hand most of the climate science the IPCC reports on is mature enough that there probably won’t be many big surprises anyway.
The subsequent days did indeed focus more on details of the behaviour of clouds, and as a result we skipped a fair bit, although this paper (which has just come out) may have wider interest. There was an amusing moment when someone argued that if we wanted to give the best possible predictions for future climate change, we should use the highest possible resolution for an atmospheric model and provide it with prescribed sea surface temperatures in place of a realistic ocean model. It’s a lucky coincidence that he was an atmospheric modeller – just imagine if he’d worked on the ocean for a couple of decades before realising it was completely pointless! 🙂
I also had to arrange a few phone calls from journalists who had got wind of this story: the results of which can be seen here and here (and no, I didn’t tell the 2nd one I had never lost but never mind about that small detail). The timetabling of those also interrupted some of the talks. So while the CFMIP meeting was a bit marginal for us, the side meetings made the week extremely productive.
One of the big attraction of a meeting at NCAR for me is the breakfast that is earned through cycling up the substantial hill.