Modeling the impact of trace metals on biogeochemical cycles

Current models incorporate the effect of trace metal complexation and limitation only in a phenomenological manner. As a result, models must be calibrated for every specific condition/scenario to be investigated. For example, transport models for trace metals may include a binding constant of the trace metal to organic carbon (dissolved organic carbon or a fraction thereof) which has no general validity. Models of the dynamics of phytoplankton in oceans allow for different ammonia/nitrate uptake ratios, but those are not related to the availability of iron or other trace metals. Similarly, the biodegradation of hydrocarbons is represented by a Monod-type formulation for which kinetic rate parameters are environment-specific.

The objective of this component is to integrate the new information about molecular mechanisms into dynamic models of trace metals cycling and of the global carbon cycle.

To obtain a more realistic and, we hope, more robust model, we shall develop mechanistically based formulations that incorporate the enzymatic processes as well as the chelator- and storage-protein—metal interactions. This will be done for metal contaminant transport models, hydrocarbon biodegradation models, and—in collaboration with PEI's Carbon Center—oceanic phytoplankton models that are embedded into models of the global carbon cycle. This will yield a new generation of models that incorporate our new molecular-scale insights and require less site/condition-specific calibrations.

The new model formulations will be tested under a wide range of dynamic conditions. For this purpose, we shall introduce a new set of nonlinear analysis tools to link the experimental data and the models. Such tools are essential to guide the design of complex multivariable biogeochemical experiments, so as to maximize the information obtained from them. The analysis is based on a special Sobol function expansion to relate model output and laboratory observations through a hierarchy of physical variable correlations. These general nonlinear tools will be employed to analyze the laboratory data and refine the model formulations as well as determine whether a model formulation has any essential missing components.

Return Home