A perturbation simulation of CO₂ uptake in an ocean general circulation model

Abstract

The uptake of anthropogenic CO2 by the ocean is simulated using a perturbation approach in a three‐dimensional global general circulation model. Atmospheric pCO2 is prescribed for the period 1750–1990 using the combined Siple ice core and Mauna Loa records. For the period 1980 to 1989, the average flux of CO2 into the ocean is 1.9 GtC/yr. However the bomb radiocarbon simulation of Toggweiler et al. (1989b) shows that the surface to deep ocean exchange in this model is too sluggish. Hence the CO2 uptake calculated by the model is probably below the actual value. The observed atmospheric increase in 1980 to 1989 is 3.2 GtC/yr, for a combined atmosphere‐ocean total of 5.1 GtC/yr. This is comparable to the estimated fossil CO2 production of 5.4 GtC/yr, implying that other sources and sinks (such as from deforestation, enhanced growth of land biota, and changes in the ocean carbon cycle) must be approximately in balance. The sensitivity of the uptake to the gas exchange rate is small: a 100% increase in gas exchange rate gives only a 9.2% increase in cumulative oceanic uptake. Details of the penetration into different oceanic regions are discussed.

Evaluation and intercomparison of three-dimensional global marine carbon cycle models

“…In both cases, the Southern Ocean exerts considerable influence on atmospheric CO, because its surface waters have a great area1 extent and are in intimate contact with the large volume of the deep ocean through upwelling and convection Also, the abundance of surface nutrients in the Southern Ocean means that biological processes have the potential to dramatically modify the surface carbon balance in this region Toggweiler, 1984, Knox andMcElroy, 1984;Siegenthaler et al, 1984, Sarmiento andOrr, 1992). Future scenarios of the air-sea partitioning of CO, further emphasize the likely importance of the Southern Ocean.…”

“…JGOFS, particularly the SMP phase, has a number of interim goals as well, including the determination of fluxes and inventories of carbon in the modern ocean that are germane to the airsea partitioning of CO, Models have a role to play here too, because many of these fluxes and inventories-such as the distributions of anthropogenic dissolved inorganic carbon (DIC), new primary production and aphotic zone remineralization-while not amenable to direct observation on the large scale, can be determined using a variety of modeling approaches (Siegenthaler and Oeschger, 1987;Hasselman, 1987, Bacastow andMaier-Reimer, 1990;Sarmiento et aL, 1992.…”

Model constraints on the anthropogenic carbon budget of the Arctic Ocean

“…Instead, to correct results from our biogeochemical simulations, we essentially add in the difference between two additional simulations made in the same circulation model but coupled to the more efficient single-tracer perturbation approach (Sarmiento et al, 1992) rather than the full PISCES biogeochemical model (24 tracers). By definition, the perturbation tracer is initialized to zero everywhere; it accounts for only the perturbation (C ant ), assuming it is independent of the natural carbon cycle.…”

“…Because of computational limitations, it was necessary to start the anthropogenic CO 2 perturbation of our reference ORCA05-PISCES simulation in 1870 as opposed to the traditional earlier reference of 1765 (Sarmiento et al, 1992), a more realistic approximation of the start of the industrial-era CO 2 increase. A similar compromise was adopted for CMIP5 (Taylor et al, 2012).…”

“…The perturbation approach of Sarmiento et al (1992) avoids the computationally intensive standard CO 2 system calculations by focusing only on anthropogenic carbon and exploiting a linear relationship between the anthropogenic change in oceanic 5 pCO 2 and its ratio with the ocean's corresponding change in C T (C ant ):…”