Multi-Scale Modeling Framework for Mercury Biogeochemistry

Mercury (Hg), and most of its compounds, are extremely toxic to wildlife and humans, causing both chronic and acute poisoning. Natural sources of Hg, such as volcanoes, are responsible for approximately half of atmospheric mercury emissions, with the other half coming from human activities, such as coal-fired power plants. The Oak Ridge Reservation is a site of particular Hg contamination, due to past weapons production activities. Efficient remediation of contaminated sites and prevention of additional contamination requires a predictive understanding of Hg biogeochemical transport and transformation in surface and subsurface environments. The planned research aims to construct a multiscale modeling framework that connects the molecular scale to local (mesoscale), reach and watershed scale models.

The ORNL Mercury Scientific Focus Area (SFA) is presently establishing reach- and watershed-scale modeling capabilities for the East Fork Poplar Creek (EFPC). While it is generally recognized that linking atomistic knowledge to macroscopic scales has considerable potential for improving overall model accuracy, deficiencies in our understanding of basic processes and the lack of development of molecular scale modeling methods has hindered the generation of reliable data and thus the integration of these data across scales to obtain a model of Hg cycling.

The University of Tennessee has recently established a molecular-scale, quantum chemical computational methodology that has been demonstrated to calculate the structures, rates and binding constants of Hg complexes with organic and inorganic species with chemical accuracy. There is a need to link atomistic and mesoscale modeling to impact field-scale studies and improve the accuracy with which processes that control Hg fate can be modeled.

Research within this project will involve computing the mechanisms, kinetics and thermodynamics of a comprehensive list of Hg chemical processes critical to EFPC. The results of the molecular-scale modeling will be combined with information from existing thermodynamic databases and used as input for local-scale thermodynamic speciation and reactive flow modeling. The computer simulations will make use of DOE supercomputers.

Expected outcomes include mesoscale models that incorporate new atomistic process understanding. The meso-scale modeling will focus on integrated local-scale model systems containing ions, natural organic matter and microbes, and will include the modeling of critical geochemical gradients. The results will be tested by comparison with key speciation and kinetic experiments existing in the literature and being performed within the ORNL SFA program. When integrated with the work being performed in the SFA, the present work will establish a generally-available modeling framework for deploying knowledge gained from molecular-scale Hg models up to the field scale. The established computational framework will be in the future generalizable to other metals, and will be able to be continuously iterated with experiment over all scales to improve modeling accuracy.