Large-scale ocean circulation, air-sea gas exchange, and carbon isotopes in a three-dimensional, computationally efficient ocean model

Müller, Simon A. (2007). Large-scale ocean circulation, air-sea gas exchange, and carbon isotopes in a three-dimensional, computationally efficient ocean model (Unpublished). (Dissertation, Universität Bern, Philosophisch–naturwissenschaftliche Fakultät, Physikalisches Institut, Abteilung für Klima– und Umweltphysik)

Text mueller07phd.pdf - Other Restricted to registered users only Available under License BORIS Standard License. Download (43MB)

This thesis focuses on the intermediate-complexity modelling of the large-scale ocean circulation and the oceanic carbon cycle. The main objective is the development of a model of intermediate complexity by combining, evaluating and extending existing formulations of the single components within a coherent setting. The model should be suited to perform long integrations over the timescales of interest in climate research in short periods of time, also permitting to employ Monte-Carlo and other ensemble techniques.
In Chapter 2 the development of the physical core of the Bern3D ocean model is presented. First, the frictional-geostrophic balance ocean model of Edwards et al. [1998] is reviewed. This model has been designed to be computationally efficient and is used as the oceanic com­ ponent in the coupled atmosphere - sea ice - ocean model "C-GOLDSTEIN" [Edwards and Marsh, 2005] and is also applied as the oceanic component of the grid-based Earth system model GENIE (http://www. genie. ac.uk). The Bern3D ocean model is largely based on the ocean model described by Edwards et al. [1998] with updates as described in Edwards and Marsh [2005]. First, the governing equations describing the large scale ocean circulation and tracer transport are introduced. Then, specific extensions and modifications of the original ocean model constituting the development of the physical core of the Bern3D ocean model are described in detail. The Bern3D ocean model features seasonal boundary conditions and a high vertical resolution. In comparison to the ocean model used in Edwards and Marsh [2005], the mixing scheme for isopycnal diffusion and eddy-induced transport has been enhanced and an altered convection scheme has been implemented. Further, results from the standard simulation and the sensitivity of the steady state circulation to the mixing parameters are examined and presented. Also the simulated barotropic transport for a simulation with opened Indonesian Passage and increased latitudinal resolution is presented and compared to the standard solution.
The large scale circulation and the resulting distribution of water masses as simulated with the Bern3D ocean model has been examined and probed using a set of tracers governed by decadal to multi-century timescales [Müller et al., 2006] (Chapter 3). Modelled tracers are the 39Ar/Ar ratio and natural radiocarbon, both influenced by radioactive decay with different mean life times, as well as bomb-produced radiocarbon, CFC-11, and anthropogenic carbon, all influenced by their transient atmospheric boundary conditions. For all of these tracers exist measurements or data-based reconstructions of their distribution inside the ocean. Simulated distributions and inventories of temperature, salinity and the other tracers are compared with data-based estimates and reasonable agreement is found for a simulation performed with the Bern3D ocean model tuned towards data-based metrics, including the natural radiocarbon signatures of typical water masses and the inventories of CFC-11 for the mid-l 990s in the Indopacific. The model has also been applied to examine the importance of different surface­ to-deep transport mechanisms for the simulated distribution of natural radiocarbon and the uptake of CFC-11. Deep equatorial upwelling has been found to be sensitive to the vertical model resolution, reduced deep equatorial upwelling strength is found for a higher vertical resolution. Furthermore, the carbon budget for the industrial period was closed by simulating an uptake of anthropogenic carbon by the ocean of 18.9 GtC and 19.7 GtC for the 1980s and 1990s, respectively, and inferring land-atmosphere fluxes, in agreement with data-based estimates.
The Bern3D ocean model has been complemented with standard formulations of the abiotic, organic matter, and calcite carbon cycle (Chapter 4), as formulated for the phase two of the Ocean Carbon-Cycle Model Intercomparison Project (OCMIP-2) [Orr, 2000]. Further, the formulations of the OCMIP-2 have been extended to include the radiocarbon fluxes in the organic matter and calcite cycles and the 13C fluxes for the whole oceanic carbon cycle formulation. Results have been compared with data-based estimates and are found to reasonably reproduce the main large-scale features as observed in the data. The oceanic carbon-cycle model component has been complemented with a 4-box formulation for the land biosphere following Siegenthaler and Oeschger [1987], in order to calculate the radiocarbon production rate by inversion of atmospheric Ll14C reconstructions from tree rings [Reimer et al., 2004, McCormac et al., 2004] with a global carbon cycle model.
The magnitude of the global mean air-sea gas exchange rate for CO2 has been discussed and debated over the last few decades in the scientific literature. In a contribution to this discussion (manuscript in preparation, Chapter 5) the natural radiocarbon distribution, basin-wide inventories of excess radiocarbon, the Earth system budget of radiocarbon, and basin-wide inventories of CFC-11 have been simulated using the gas-exchange formulation of the OCMIP-2 in an extensive sensitivity study varying the magnitude of the gas-exchange rate and oceanic transport strength. The results are compared to recent data-based estimates. It is demonstrated that the rate-limiting process for the uptake of excess radiocarbon is the gas-exchange rate and that the uptake of CFC-11 is dominated by oceanic transport and mixing. A target consisting of the data-based basin-wide inventories of excess radiocarbon in the Pacific, Indian Ocean, and Southern Atlantic for different points in time, the natural radiocarbon distribution in the surface ocean, and the Earth system budget of radiocarbon is defined and is best matched if the gas-transfer field from the OCMIP-2 is scaled down by (26 ± 16)%. Simulated column inventories are found to be similar in all basins, in contrast to data-based estimates, where the North Atlantic inventories are significantly higher than in the other basins. Estimates for inventories of excess radiocarbon, using the downscaled gas transfer rate of (15.7 ± 3.3) cm hr-1, are presented.
In a publication by Muscheler et al. [2007] (Chapter 6), the radiocarbon production rate records calculated by different carbon-cycle models and reconstructions of 10Be from ice­ cores are used together with records of the geomagnetic field intensity to infer the solar magnetic modulation, which is linked to the solar activity. The solar modulation parameter record is compared to irradiance records. In this publication, the radiocarbon production rate calculated with the Bern3D model is presented and compared with other radiocarbon production rate reconstructions using carbon-cycle models featuring a box model representation of the ocean. In comparison to the box models of the ocean, the Bern3D model provides a more realistic boundary condition at the surface ocean for the air-sea gas exchange of radiocarbon. ∆14C reconstructions for the northern and southern hemispheres, which differ owing to a gradient in atmospheric ∆14C at the equator, can thus be included as boundary conditions of the respective oceanic domains. In the box models of the ocean, which do not resolve the two hemispheres, the atmospheric ∆14C boundary condition has to be an averaged value of the two hemispheric records. The results using the different modelling approaches to calculate radiocarbon production are found to agree well.
An outlook regarding ongoing and possible future model development extending the Bern3D ocean model into various directions and applications thereof is given in Chapter 7. The Appendix details the construction of smooth atmospheric pC02 records as applied in the studies presented in this thesis and also their projection into the future in the form of idealised stabilisation scenarios as used for projections of climate change commitment using Earth system models of intermediate complexity for the upcoming Fourth Assessment Report (AR4) of the Intergovernmental Panel on Climate Change (IPCC) and e. g. in Plattner et al. [submitted] and Knutti et al. [2005]. Furthermore, a publication [Muscheler et al., 2005] contributing to a discussion that was initiated following the publication of a paper by Solanki et al. [2004] is attached in the Appendix. There, using radiocarbon production rate based reconstructions of the solar modulation parameter, Muscheler et al. [2005] find that the solar modulation parameter was comparable to or even exceeds today's values during several periods in the last 500 years.

Interest & Impact

Downloads

2 since deposited on 18 Apr 2024
2 in the past 12 months

Citations

5 Citations in Web of Science ®
6 Citations in Scopus

Search

in Google Scholar™

Services

Actions (login required)

Edit item

Actions (login required)

Edit item