Investigation of mechanism driving glacial/interglacial variability in stratification/mixing and biological productivity of the Pacific Southern Ocean and potential implications on air-sea CO2-exchange: synthesis of climate and ocean biogeochemical modeling with paleoceanographic reconstructions
Final Report Abstract
The present project set out to advance current understanding and modeling of the Southern Ocean’s role in carbon cycling based on Si-isotopes. It was suggested that glacial/interglacial contrasts in physical and nutrient properties at surface and subsurface water depth are reflected in the Si-isotopic composition of sedimenting opal and measurements could be used to test several hypotheses on (i) the impact of yet unconsidered dust-borne micronutrient deposition on the glacial South Pacific on shifts of primary productivity, Si-uptake rates and carbon export, (ii) the ’silicic-acid leakage’ hypothesis (SALH) and (iii) the formation and extent of surface water stratification. We additionally altered the riverine Si-supply. We started with assessing the viability of coarse resolution models to tackle the task. Accompanying studies on the uncertainty associated with state-of-the-art generic biogeochemical model provided guidance on the level of detail we should resolve in our resource-hungry paleo model runs (Löptien and Dietze, 2019). The simulation of various climate states was then based on the UVIC ECSM, which we extended to explicitly represent silicon-isotopes (H1, BA, LGM and 115+135ka BP). Overall, our reference simulations were inconsistent with the paleo record collection of δ 30 Si provided by Frings et al. (2016). We thus implemented and tested several hypothesis which were suggested in the literature to explain the observed difference in δ 30 Si of BSi between preindustrial climate (PI) and the Last Glacial Maximum (LGM). Our numerical experiments suggest that neither of the following processes effected glacial-interglacial changes in the isotopic composition of BSi consistently with observations: (1) an overall cooling and shutdown of the meridional overturning circulation (our reference experiment for the LGM), (2) a decrease of Si:N quota in diatoms as potentially effected by increased air-sea iron fluxes (our experiment LGMfe), (3) decreasing winds (our experiment LGMslack) and (4) increasing or decreasing Si supply to the ocean within suggested bounds (our experiments LGMtrickle and LGMflush). Out of the seven sensitivity experiments, only the simulation with increased winds (experiment LGM-breezy) and the simulation with a changed isotopic composition of river runoff (experiment LGMlight) reproduce the observed sign of change at the observational sites, specifically in the SO. The experiment with increasing winds (experiment LGMbreezy), however, fails to reproduce the magnitude of observed changes and is, furthermore, inconsistent with the Jaccard et al. (2016) estimate of dissolved near-bottom oxygen concentrations (based on redox-sensitive trace-metal records archived in the sediments in the Atlantic Sector of the SO). The experiment LGMlight is most consistent with observed changes in δ 30 Si of BSi in terms of both, sign and magnitude. This confirms the suggestions by Frings et al. (2016) that changes in the isotopic composition of DSi supplied to the ocean triggered low glacial δ 30 Si of BSi - rather than changes in the oceanic cycling of DSi. Further, the estimated manifestation timescale of changes in δ 30 Si of BSi range between several hundreds to 2500 years at the observational sites. This means that the respective signal should be detectable - despite a global turnover timescale of DSi, which is comparable to the period of glacial-interglacial cycles. Some caveats, however, remain. A major problem of developing a Si module within the framework of an Earth System Model is the high computational cost associated with running test simulations to equilibrium. To this end, a turnover timescale of Si in the ocean of more than 10 000 years is a real handicap. Among the simplifications we chose in order to limit the number of test simulations was discarding the effect of fractionation during BSi dissolution. Further - substantial - uncertainty is added by the generic problem of constraining global biogeochemical ocean models. Further model uncertainties refer to the physical module and here specifically the spatial resolution is subject to an ongoing scientific discussion. Based on our analyses and as far as the Southern Ocean is concerned we rate the biogeochemical problems currently as being more problematic than the uncertainties due to the physical model resolution.
Publications
- (2017) Simulating natural carbon sequestration in the Southern Ocean: on uncertainties associated with eddy parameterizations and iron deposition. Biogeosciences (BG), 14, 1561-1576
Dietze, H., Getzlaff, J. and Löptien, U.
(See online at https://doi.org/10.5194/bg-14-1561-2017) - (2019) Reciprocal bias compensation and ensuing uncertainties in modelbased climate projections: pelagic biogeochemistry versus ocean mixing. Biogeosciences (BG), 16 (9), 1865-1881
Löptien, U. and Dietze, H.
(See online at https://doi.org/10.5194/bg-16-1865-2019) - (2020) Silicon isotopes in an EMIC’s ocean: sensitivity to runoff, iron supply and climate
H. Dietze, U. Löptien, R. Hordoir, M. Heinemann, W. Huiskamp and B. Schneider
(See online at https://doi.org/10.1029/2020PA003960) - (2020). MOMSO 1.0-a near-global, coupled biogeochemical ocean-circulation model configuration with realistic eddy kinetic energy in the Southern Ocean. Geoscientific Model Development 13, 71-97
Dietze, H., Löptien, U., and Getzlaff, J.
(See online at https://doi.org/10.5194/gmd-13-71-2020)
