Modellgestützte Untersuchungen zum Einfluß mariner Vermischungsprozesse auf glazial-interglaziale Änderungen des atmosphärischen CO2 Gehalts.
Final Report Abstract
The project resulted in the derivation of a robust method to diagnose effective diffusivities from tracer release experiments in z-level models of the ocean circulation. In numerical z-level models diapycnal diffusion is not only determined by the parameterization of the vertical diffusion, but also artificially by numerically induced diffusion. This is an effect evolving from discretization errors of the advection terms and other along-isopycnal processes. Our approach follows real-ocean tracer release experiments, i.e. a dye tracer will be injected into a given isopycnal layer and its spreading over time wil be used to infer diapycnal mixing rates. Although the method is limited to the analysis of spatial mean values over a dye tracer patch, the results can be considered regional estimates as long as the patch occupies a volume much smaller than an ocean basin. From the results of this novel method we could establish that FCT-advection induces only a small amount of numerical diffusion, different to the advection schemes used in first generation GCMs. The development of this novel method turned out to be more challenging than expected during the writing of the proposal. Nevertheless, it addresses one of the unsolved problems in numerical modeling it is therefore an important contribution to the discussion of the outstanding problem of how to diagnose the spurious mixing problem. To reach the initial main goal of the project, we used a coupled ecosystem-circulation Earth system model of intermediate complexity (UVic, 2.8) to study the sensitivity of simulated glacialinterglaciel CO2 changes to idealized changes in upper ocean nutrient uptake. In the study of Archer et al. (2000), the box models show a reduction of the atmospheric CO2 by ⇠120 ppm, whereas in the first generation GCMs a reduction of only ⇠45 ppm is found. It turns out that all simulations with nutrient depletion yield a reduction of the atmospheric CO2 levels by 65–95 ppm. This reduction is significantly higher than that simulated by the first-generation GCM results reported by Archer et al. (2000). Additionally it is also somewhat higher than the results of simulation nutrient depletion south of 30 S described by Marinov et al. (2006), who reported a reduction of atmospheric CO2 by 70 ppm. An increase of the vertical diffusivity also increases the sensitivity of the atmospheric CO2 changes in the biological pump’s efficiency of the UVic model. The results show that in contrast to the conclusions inferred from the first generation GCMs, high latitude changes in the biological pump cannot be ruled out as significant contributor to glacial-interglacial changes in atmospheric CO2 concentrations. These results are, we think, important for a better understanding of glacialinterglacial changes in atmospheric CO2 and will be published in due time.
Publications
- 2010. Diagnostics of diapycnal diffusivity in z-level ocean models part I: 1-Dimensional case studies. Ocean Modelling, 35(3)
Getzlaff, J., Nurser, G. and A. Oschlies
(See online at https://doi.org/10.1016/j.ocemod.2010.07.004) - 2012. Diagnostics of diapycnal diffusion in z-level ocean models. Part II: 3-Dimensional OGCM. Ocean Modelling, 45–46
Getzlaff, J., Nurser, G. and A. Oschlies
(See online at https://doi.org/10.1016/j.ocemod.2011.11.006)
