Combining data and models - statistical analysis and data assimilation in biogeochemical oceanography at IFM-GEOMAR
Kiel , Germany , 10/09/2007 – 14/09/2007
The goal of this summer school was to teach data-analysis and data-assimilation methods to PhD students and young PostDocs working in biogeochemical oceanography. Bringing together observational and modelling aspects of oceanography to train the next generation of scientists to synergistically exploit both. This one-week summer school was composed equally of lectures and computer labs, the latter of which were taught as a special course offered by Fastopt. Topics that were covered included Bayesian statistics, model-data comparison, tangent linear model, ecosystem models assessment, handling uncertainties, adjoint method, review of data assimilation techniques, and Eur-Oceans model database. The computer labs were very demanding and they have been taught early on in the week (Tuesday and Wednesday) in order to give us the chance to revisit the computer programs after the lectures on the consecutive days.
In particular I enjoyed the discussion presented by Samar Khatiwala on a new strategy for the efficient simulation of geochemical tracers in ocean models, using the Transport Matrix Method the proposed method is considerably more efficient than the offline tracer models and automatically includes all parameterizations of unresolved processes present in the most sophisticated GCMs.
Overall the summer school was extremely well organized and structured, with time for further interest discussions easy and welcome. I gained a better understanding of the possibilities and limitations of data assimilation methods applied to marine biogeochemistry and got to meet some interesting people at the forefront of this field of investigation.
This link shall provide some links to sites on the web which provide additional material in the context of this summer school: http://euroceans.ifm-geomar.de/school/ (now obsolete).