Projected 21st century decrease in marine productivity: a multi-model analysis

M. Steinacher, F. Joos,T. L. Frölicher, L. Bopp, P. Cadule, V. Cocco, S. Coney, M. Gehlen, K. Lindsay, J. K. Moore, B. Schneider, and J. Segschneider Link to full paper

Summary for MEECE Research Highlights by Yuri Artioli

Climate change will lead to a decrease of primary productivity (PP) of the global ocean between 2% and 13% at the end of the century. This is the major output of a multi-model analysis carried out as a joint work of many international projects, including MEECE. Four Earth System Models have been used to evaluate the change in marine PP under the SRES A2 emission scenarios: the IPSL-CM4-LOOP model from the Institut Pierre Simon Laplace (IPSL), the COSMOS Earth System Model from the Max-Planck Institute for Meteorology (MPIM), and two versions of the Community Climate System Model (CSM1.4-carbon and CCSM3-BEC) from the National Center for Atmospheric Research. All the models projected a moderate decrease in PP on the global scale as a combination of two contrasting processes: in the more productive regions (like the North Atlantic and the equatorial area) PP will strongly decrease due to a stronger nutrient limitation, mostly because of the enhanced stratification and slowed circulation, while in the polar region the increase in temperature and light availability will enhance moderately the PP.

Despite of the general agreement on this basic trend, the analysis also show discrepancies between the models highlighting areas where more research is needed. In particular the major differences arise on the evaluation of the impact of the sea-ice retreat (especially in the Arctic Ocean) and on the simulation of the iron biogeochemical cycle (both in terms of atmospheric inputs and biological processes).

The multi-model analysis presented in the paper is very interesting also from a methodological point of view: authors indeed presented a new method to combine results from different model based on regional skill metrics that allow a better representation of the regional features compared to the IPCC-type “one model, one vote” approach.

Vertically integrated annual mean primary production (PP, mgC/m2/day) derived from ocean color (a) (SeaWiFS; Behrenfeld et al., 2006; Behrenfeld and Falkowski, 1997b) and from the multi-model mean under preindustrial conditions (decadal mean 1860–1869) (b) and the projected changes by the end of the 21st century under SRES A2 (c). The changes are shown on an exponential scale and represent the difference between 2090–2099 and 1860–1869 (decadal means). The multi-model means have been computed by using the regional skill scores. The dotted areas indicate that none of the regional skill scores is higher than 0.5. Where no observation based data is available to calculate skill scores (e.g. in the Arctic) the arithmetic mean of the model results is shown. (redrawn from the original figures 2 and 10)