Wednesday :: Aug 20, 2008


by Christina Hulbe

I'm waiting for a flight at the Albuquerque airport, on my way home from a community ice sheet model development workshop at the Los Alamos National Lab. The small ice sheet group here at LANL has been making rapid progress implementing a model developed in the UK. The workshop objectives this week were to prioritize additions to that model, with a keen eye to what must be done in time for the next round of climate system model runs for the Intergovernmental Panel on Climate Change (IPCC). The next IPCC report is scheduled for delivery by 2013, which means that any changes to the models contributing to the effort must be finalized by late 2009 (September is the limit for the Community Climate System Model they work on here at LANL). The results of the model runs, which consume months of computer time, must be delivered by the end of 2010 in order for the report to be written, reviewed, edited, and approved in time.

Ice sheets are one of many components in a climate system model. Until recently, the usual assumption was that ice sheet dynamics (the flow of the ice) was a low priority, something to be addressed when time allowed. The models dealt mainly with effects of and processes at the surface of the ice (how reflective it is, how much it melts, and so on). The first IPCC assessment report, issued in 1990, gave ice sheet dynamics a pass, evaluating the time scales for change to be too long to be of concern over the 21st century. The 1995 report noted that the West Antarctic Ice Sheet might be capable of rapid change, but concluded that this was a high risk/low probability scenario but that was about it. Not a lot changed in the 2001 report, despite concerns raised by the glaciology community. Ice sheet dynamics and the possibility of rapid change in polar ice sheets finally went prime time in the 2007 Fourth Assessment Report but the big models were not up to the task of producing such changes so the IPCC (emphatically) didn't include them in their sea level rise estimate (and thus perhaps significantly underestimated the sea level rise projection).

My small discipline has been pushed into something of a spotlight. Everybody wants us to "do something." The thing is, the ice sheet components of the big coupled climate models are not the same as the kinds of models research scientists develop in order to answer specific research questions. There are a number of reasons for this, ranging from simple implementation issues (the spatial scales we care about for ice sheets are small compared to the scales at which coupled climate models operate) to unresolved but fundamentally important science issues. So here we are.

Over the course of the workshop, we have talked about aspects of our individual research efforts that bear on the problems at hand and a range of technical issues. We've discussed model validation and evaluation, standardization of data sets and model initialization states (the starting values for all the variables in the model equations), and our individual opinions about what's easy and what's not, and what we have confidence about and what we don't. These discussions have been open to all voices and yet remained focussed (I find this amazing and wonderful, and suppose it results from having great organizers here at LANL and a strong community sense that there is important work to be done).

With all of that in our heads, we sorted ourselves into more focussed groups, 2 on the first day and 2 others on the second, to identify and prioritize the work to be done (improvements to the models, development of standard data sets, software issues), develop work plans. On the third day (today), we did our best to formalize the work plans and associate specific names with specific work items. Overarching it all is the intent to build the most useful, most credible community ice sheet model possible on the CCSM timeline so that the 100 year prediction of sea level rise will be improved. Built within the community climate system model, the product will be available to everybody.

This has been an interesting and enjoyable process. We are working in a framework adopted long ago by the climate modeling community but for us, it's new. Glaciology is a relatively small discipline and ice sheet modeling is a small subset of that. We tend to work as individuals or in collaboration with "data people," funded to answer specific research questions. It's just the nature of the funding structure within which we work. It is also certainly the case that these individual efforts have moved the collective knowledge forward. It's that fundamental work on which we are building now. Volunteering for additional work is not easy, we are already about as busy as we can be but we've finally been invited to the ball and we're getting ready as fast as we can.

One other comment, regarding the guest list for the ball. Two numbers were hard (for me) to ignore as I sat in the seminar room this week: 3 and 0. At a similar workshop held in Princeton about a year and a half ago, the numbers were 1 and 0. The total number of participants here was something like 35, in Princeton the total was a bit higher, that is, less than 10% of the participants were women and 0% were from other underrepresented groups. The geosciences are not the only disciplines with pronounced gender and ethnic imbalance but we're certainly near the back of the pack. It should be an embarrassment.

Christina Hulbe :: 4:40 PM :: Comments (10) :: Digg It!