Articles | Volume 12, issue 3
Research article
04 Feb 2015
Research article |  | 04 Feb 2015

Evaluating Southern Ocean biological production in two ocean biogeochemical models on daily to seasonal timescales using satellite chlorophyll and O2 / Ar observations

B. F. Jonsson, S. Doney, J. Dunne, and M. L. Bender

Abstract. We assess the ability of ocean biogeochemical models to represent seasonal structures in biomass and net community production (NCP) in the Southern Ocean. Two models are compared to observations on daily to seasonal timescales in four different sections of the region. We use daily satellite fields of chlorophyll (Chl) as a proxy for biomass and in situ observations of O2 and Ar supersaturation (ΔO2 / Ar) to estimate NCP. ΔO2 / Ar is converted to the flux of biologically generated O2 from sea to air (O2 bioflux). All data are aggregated to a climatological year with a daily resolution. To account for potential regional differences within the Southern Ocean, we conduct separate analyses of sections south of South Africa, around the Drake Passage, south of Australia, and south of New Zealand.

We find that the models simulate the upper range of Chl concentrations well, underestimate spring levels significantly, and show differences in skill between early and late parts of the growing season. While there is a great deal of scatter in the bioflux observations in general, the four sectors each have distinct patterns that the models pick up. Neither model exhibits a significant distinction between the Australian and New Zealand sectors and between the Drake Passage and African sectors. South of 60° S, the models fail to predict the observed extent of biological O2 undersaturation. We suggest that this shortcoming may be due either to problems with the ecosystem dynamics or problems with the vertical transport of oxygen.

Short summary
We compare how two global circulation models simulate biological production over the year with observations. Note that models simulate the range of biological production and biomass well but fail with regard to timing and regional structures. This is probably because the physics in the models are wrong, especially vertical processes such as mixed layer dynamics.
Final-revised paper