Isotopes
Foraminiferal carbon isotopes are widely measured in their own right, and as a by-product of δ18O measurements. Many hundreds of Pleistocene records have been collected in marine cores spanning the globe. Known artefacts have been identified, including problems with infaunal benthic species, the “Mackensen effect”, and “vital effects”. Through an exhaustive search of online databases and printed literature we will collate this data, together with related information such as species, age, laboratory, and collection technique. Efforts will be made to assess data quality based on analytical uncertainty, measurement protocols, species, and, where data density allows it, by comparison with nearby data of the same age. This compilation will form the major dataset to be modelled in WP2.
Carbonate δ13C will be augmented by Cd/Ca and 231Pa/230Th proxy data. Like δ13C, these two proxies are controlled by ocean productivity and circulation. Existing compilations of core-top and water-column 231Pa/230Th and Cd/Ca will be quality controlled and extended into the past. We will also collate and assess the quality of water-column measurements of δ13C, the quantity of which has increased during the WOCE programme.
We will review existing and proposed parameterizations for the behaviour of water-column and sediment 13C and 14C, Cd/Ca and Pa/Th, in order to define a consistent, combined parameterisation scheme for this suite of tracers with the minimum complexity required to represent the key dynamical processes at G-IG timescales. Where possible, remineralisation profiles will be replaced by a directly calculated velocity-dependent sinking scheme.
Incorporation of tracer dynamics into the GENIE ocean model is already in progress. The closely-related Bern3D model has already successfully simulated recent 14C and 231Pa/230Th based on imposed particle fields and ocean/atmosphere fluxes. The BIOGEM ocean biogeochemistry module of GENIE already incorporates the basic carbonate system dynamics and sediment dynamics necessary to provide inorganic particle field sources and sinks. Both the Bern version (which is ocean-only) and the new version of BIOGEM incorporate substantial changes compared to the current version of GENIE some of which are currently untested or make mutually conflicting assumptions. We will complete and unify these developments, and proceed to the incorporation and testing of the complete tracer suite.
Forward integrations of ~5000 yrs will result in equilibrium fields of the modelled tracer suites in the water column and in recent sediment. Process-based sensitivity studies varying key uncertain parameters such as gas exchange, remineralization rates, and settling rates will be used to evaluate model performance. Accurately mimicking the distribution of all three tracers will provide a stringent test for the model’s ocean circulation, ocean productivity, and particle dynamics.
Concepts which have been proposed to explain glacial-interglacial cycles in atmospheric CO2 typically make specific predictions about the nutrient state of surface and (through deep-water formation) deep waters, and therefore about the distribution of δ13C, Pa/Th, and Cd/Ca. The hypotheses will be tested by a forward modelling approach, and through data assimilation. In the former, specific model features will be altered (e.g. Southern Ocean productivity; NADW flow) and model proxy response tested against observations. Automated filtering and assimilation techniques will also be used to tune the model to optimally reproduce past data and, if possible, complete time series. This process will refine estimates of uncertainty in model parameters, and of the plausibility of conceptual hypotheses.