While trying to find out why the TOPAZ forecast maps were no longer updated I came into contact with Dr Laurent Bertino, a research director at the Nansen Environmental and Remote Sensing Center who is responsible for the TOPAZ system that is delivered as part of the MyOcean project.
Not only did he direct me to the new dynamic viewer that contains the images of sea ice thickness (and much more), but he also consented in my request for a short interview.
LB: The Nansen Center is a private research foundation (~60 persons) with various activities in climate research, remote sensing, modeling and data assimilation, the latter two are the activities I am personally involved in. See our fresh new web-site for more information: http://www.nersc.no/
What we call "data assimilation" is working out the combination of numerical model forecast and observations that makes a forecast most accurate. This is a field at the crossroads of mathematics, statistics, physics and computer sciences.
N: There are several discrepancies between the sea ice thickness models. I discussed them in a recent blog post and follow-up. PIPS 2.0 for instance is showing a lot more thick ice than TOPAZ. What are your thoughts on this?
LB: The ice thickness in our models is not constrained by observations in any way. So it evolves over years of simulations depending on aspects such as: the atmospheric data used, the ice thermodynamic model, the ocean surface temperature and salinity ...
I do not know enough of the PIPS system to explain the difference. There are anyway very few ground truth observations in the middle of the ice pack and most of the in situ data is only available in delayed mode. So we know that the system was doing OK in reanalysis mode (between 2005 and 2007) but I would not be surprised if the ice thickness were half too thin in the forecast system, not to mention the uncertainties in the trend.
LB: Yes. NERSC has carried out impact studies of CryoSat data almost 10 years ago and developed the assimilation method in order to be ready to include them in the forecast system on the fly, as soon as the data is available. So now just imagine how eager we are...
NERSC follows CryoSat closely, with Stein Sandven being member of the ESA CryoSat science team and with one PhD student working on the calibration of the data.
N: Arctic sea ice extent has been on a downward trend for a few decades now. Do you have any expectations on what will happen in the short-term and medium-term?
LB: Not much. It is a safer bet to say "downwards" but I have nothing precise to say on the short term and medium term (say, less than 10 years) because the sea ice extent is very sensitive to local weather conditions, which are hard to predict. You can check this WCRP news.
N: Final question. Your name sounds French (and a bit Italian too). Do you speak Norwegian, and if so, how on Earth did you manage as a Frenchman? Is it possible for you to buy Camembert and Mâcon over there in Oslo?
LB: Yes, I am a Frenchman, with 100% Italian origins, living in Norway for the last 9 years. This may sound like a mess, but modern Norwegians eat more pizzas than Italians and they serve different things for "Caffè latte" and "Café au lait", so consistency is no longer an issue, I am afraid. Real Camembert and Dijon mustard have made their appearance on the shelves of supermarkets about 6 years ago and the fjords have never been more beautiful since that day.
I still need to bring my own Marsala Mandorla in my luggage, though. Seriously, Bergen is a great place to be if you are interested in data assimilation, oceanography, skiing, hiking, music, and if you don't mind short summers. Oslo has somewhat less charm.
I don't know about you people, but those final remarks have given me an appetite. Before rushing off to the kitchen to finish my fish broth, I'd like to thank Dr Laurent Bertino of NERSC one more time for this interview.