PML/National Centre for Earth Observation (NCEO) scientists have generated the first-ever long-term reanalysis simulation of the biogeochemical conditions in the northwest European shelf, highlighting the importance of measuring uncertainty in these indicators, in order to better inform future marine policy.
This unique study by Ciavatta et al (2016), selected Research Highlight in Journal of Geophysical Research, sought to address current gaps in knowledge, helping us build a better picture of how our oceans will respond to future global climate change. To do this, the scientists integrated satellite-derived data with coupled physical-biogeochemical models.
What makes this study different from others of a similar nature however, is that the researchers also quantified the uncertainty in these biogeochemical estimates, evaluating their model’s skill at matching a reference data set of 10 relevant indicators and creating ‘confidence levels’, the value of which they highlight in two case studies of policy applications. Focusing on shelf-sea ecosystems, the team incorporated chlorophyll data derived from Ocean Colour of the European Space Agency’s Climate Change Initiative, into a northeast Atlantic ecosystem model which is based on the European Regional Seas Ecosystem Model (ERSEM). In this way, the first-ever decadal reanalysis simulation of the biogeochemical conditions in the northwest European shelf was generated.
Importantly, the reanalysis comes with confidence levels that quantify the uncertainty of the biogeochemical estimates. The crucial implications of this supplementary information were evident in two case studies, where it was assessed that:
1. An area as large as 325,000 km2 was vulnerable to oxygen deficiency at the bottom of the North West European shelf, but additional 40,000 km2 are included when using a strict 1% confidence criteria.
2. The North West European shelf is a net sink of atmospheric CO2, but the simulated uptake can range between 36 and 46 Tg C yr21, when applying a 90% confidence level for the estimates.
The confidence levels are an added value of the reanalysis with respect to the model output alone, because estimates of reliability are much needed for model applications in marine policy [Hyder et al., 2015].
In future studies this approach could be extended to other shelf-sea models, helping to quantify the uncertainty in the health conditions of more marine ecosystems. This process would aid scientists in their ability to inform management of sea-shelf ecosystems through use of model simulations in “operational ecology”, e.g., in the framework of the Copernicus Marine Environment Monitoring Service.