Monitoring trends in biogeochemical indicators such as chlorophyll, dissolved oxygen, and nutrient concentrations is crucial for evaluating the oceans’ response to global climate change and its effects on marine ecosystems. Such field measurements are sparse, however, and the record of remotely sensed ocean color observations—the chief source of global biogeochemical data today—is limited to the sea surface and does not include a number of key variables.
To address these gaps, researchers have begun to integrate satellite-derived data with coupled physical-biogeochemical models, using a method that obtains more realistic estimates by constraining the output to fit the observations. Now Ciavatta et al. extend this “reanalysis” approach from the open oceans to shelf-sea ecosystems. By incorporating chlorophyll data derived from ocean color observations into a northeast Atlantic ecosystem model, the researchers generated the first-ever long-term reanalysis simulation of the biogeochemical conditions in the northwest European shelf.
In contrast to most previous studies, however, the researchers also quantified the uncertainty in these biogeochemical estimates. The team evaluated their model’s skill at matching a reference data set of 10 relevant indicators with the reanalysis output and generating confidence levels, the value of which they highlight in two case studies of policy applications.
In the first, the researchers determined that 325,000 square kilometers of the northwest European shelf’s bottom waters are seasonally prone to oxygen deficiency. However, when using a more conservative criterion of at least 1% confidence (compared to 100%), the size of the oxygen-deficient area increases by an additional 40,000 square kilometers—an area comparable to Switzerland.
In the second case study, the team examined the year-by-year variability in the shelf’s uptake of atmospheric carbon dioxide. They calculated that the net uptake averages 41 teragrams of carbon per year and determined that at the 90% confidence level, this result has an uncertainty of about 25% and an interannual variability of about 20%, reflecting the fact that some shelf areas can serve as either a carbon source or a carbon sink.
Extending this approach to other shelf-sea models would help quantify the uncertainty in more biogeochemical simulations, a step that would ultimately improve the models’ ability to inform better management of sea-shelf ecosystems, the source of more than 90% of the world’s fishing catches. (Journal of Geophysical Research: Oceans, doi:10.1002/2015JC011496, 2016)
—Terri Cook, Freelance Writer
Citation: Cook, T. (2016), Uncertainty evaluations improve biogeochemical simulations, Eos, 97, doi:10.1029/2016EO047765. Published on 15 March 2016.