The climate science community went a bit bonkers last month. People got thin and pale and stressed. The reason? There was a deadline! (as Gavin has already noted) The idea is that, in order to make the jobs of the IPCC AR5 authors tractable, papers to be quoted in second draft of the report should have been submitted to a journal by the end of July. I would have thought that at this 5th iteration of the IPCC report, climate scientists would be used to this, and be able to set sensible targets - but apparently not. It seems that many people overestimate what they will do, and then work themselves into the ground. How properly finished or well thought out are the papers that result?
As coordinator of JUMP (Japan Uncertainty Modelling Project) which has a strong "towards IPCC" component, my main strategy towards this deadline has been to encourage everyone to publish their work over the last 5 years. This might sound odd but generally there is not a much pressure to publish in Japan - rather promotion is gained by pleasing your boss (shining his shoes, tidying his desk etc). The strategy seemed to work OK, with about 15 papers cited in the first order draft of the report. As James has previously mentioned, this compares rather well with other groups in Japan, but then, remember that they are under no pressure to publish...
After Gavin & Axel's meeting in Hawai'i in March we made a plan for which papers we needed to write in time for the deadline. Part of me wanted to just go on holidays, but we had requests from two IPCC authors to write up particular pieces, one by jules and one by James. I imagine it must be awkward for the authors to see relevent work presented at meetings and be unable to cite it, so we wrote these two papers into the plan. Our colleague (Tokuta Yokohata) in NIES was also working on an evaluation of the CMIP5 ensemble using our methods, so we put aside some time to help with that. Then there was the extra piece of work that I offered to do at the workshop to go towards the paper that Gavin was going to write. By late May the plan became a week-by-week schedule. Key to success was stopping James starting any new work after he had submitted his paper so he could contribute to the other papers. Extra things always pop up. This time it was revising 2 papers which came back from review. Despite that, the plan was completed with a few days to spare, which meant that, on the day of the deadline, jules and James were already off walking along the sunny ridges of the Japanese alps.
The only not yet submitted paper is Gavin's one, but really that is his responsibility - we gave him our contribution in mid-June. :-) I do, however, think that Gavin is correct - that papers should not be submitted in a shoddy state, as these merely waste the valuable time of reviewers.
I return from holidays to find my email inbox stuffed full of papers to be edited!
6 comments:
Congratulations; it'd be a shame if you got diverted to teaching people how to manage their time instead of doing science, but you've mastered something very few people imagine doing.
> on the day of the deadline, jules
> and James were already off walking
That's the key -- knowing what you'd rather be doing instead of desperately revising and retyping while short on sleep and almost late.
I got through my grad school years by moving every deadline to the preceding Friday, promising myself I'd go camping that weekend after handing in the work. I did, too.
Hank: luckily there is little chance of us being promoted to the level of our incompetance as long as we remain here. We don't even get given the chance to polish the boss's desk! :-)
Hmm, desk... you don't even mention the sword. :)
Still without deadlines nothing would get done. Ask Eli
I followed the link to Tamsin's blog and was reading this
http://www.maths.bris.ac.uk/~mazjcr/climPolUnc.pdf
Suppose that you were one of a group of climate scientists, interested in playing an active role in climate policy, and able to meet the three strictures outlined in section 4. You have all embraced subjective uncertainty, and have been summoned, willingly, to a carefully facilitated expert elicitation session. After two intense but interesting days your 95% equi-tailed credible interval for the maximum height of water in the Thames Estuary in 2100 is 0.5m to 2.75m higher than today. This is wider than your initial interval, as you came to realise, during the elicitation process, that there were uncertainties hich you had not taken into account.Suppose that this has recently happened, and you are reflecting on the process, and wondering what information might have made a large difference
to your uncertainty assessment, and that of your fellow experts. In particular, you imagine being summoned back in the year 2020, to re-assess your uncertainties in the light of eight years of climate science progress. Would you be saying to yourself, “Yes, what I really need is an ad hoc ensemble of about 30 high-resolution simulator runs, slightly higher than today’s resolution.” Let’s hope so, because right now, that’s what you are going to get. But we think you’d be saying, “What I need is a designed ensemble, constructed to explore the range of possible climate outcomes, through systematically varying those features of the climate simulator that are currently ill-constrained, such as the simulator parameters, and by trying out alternative modules with qualitatively different characteristics. ”Obviously, you’d prefer higher resolution to the current resolution, but you don’t see squeezing another 0.25 out of the solver as worth sacrificing all the potential for exploring uncertainty inherent in our limited knowledge of the earth system’s dynamics, and its critical ecosystems. We’d like to see at least one of the large climate modelling centres commit to providing this information by 2020, on their current simulator, operating at a resolution that permits hundreds of simulator runs per scenario (a resolution of about 2 , we hazard). Research funders have the power to make this happen, but for some reason they have not yet perceived the need.
what do you think ?
I'm not as optimistic as they sound concerning the value of a large ensemble of a single model with different parameter values. In our experience, that tends to give you a lot of bad models, which rather resemble the original tuned model but are generally (not all) less realistic.
I'm also not as pessimistic as they seem to be concerning the value of a lot of competing models all produced by different research centres, trying to improve on each others' ideas.
(Various recent papers of ours back up these thoughts with real analysis.)
They also seem to be ignoring the realpolitik that each nation's climate research centre is actually fulfilling a somewhat political and strategic role in building and maintaining a national capability. They won't readily give that up to some supra-national authority.
We could potentially see a European climate research centre, I suppose. And that might be a worthwhile enterprise - the anaogous weather prediction centre has been a good thing IMO. But it acts as an adjunct to the major national centres, not a replacement, even on this limited stage.
Post a Comment