MERCURY CONTAMINATION OF THE SOUTH RIVER, VIRGINIA, RECONSTRUCTED USING RESERVOIR THEORY

Our study area is the high energy, gravel-bedded, bedrock-controlled South River of Virginia, where mercury was released from an industrial source from 1929-1950. To guide remediation planning, we document changes in the concentration of mercury on suspended sediment carried by the river since 1929, and we also determine characteristic reworking timescales for contaminated fluvial deposits. Frequent monitoring was only initiated in the late 1970s, so we rely on data from fluvial deposits that store suspended sediment. We address a frequently encountered but largely unsolved problem: how to chronologically order geological samples without relative or absolute age dating. Our data consist of a relatively small number of radiometric ages (15) and a large data base of measured sediment mercury concentrations (187). We argue that the mercury concentrations we measure on sediment in deposits record what has been transported by the river in the past. Thus, the distribution of past mercury concentrations on suspended sediment is known. However, the limited number of radiometric dates cannot define the ages of most of our samples, and dates cannot be inferred by extrapolation or interpolation from known ages using superposition other commonly used stratigraphic methods. So, although we know the values of past mercury concentrations on suspended sediment, our data do not define when those values occurred: the time series of mercury concentrations is unknown. We solve this problem by combining reservoir theory and Monte Carlo modeling. The known distribution of ages defines timescales of storage and reworking of sediments and associated mercury, and Monte Carlo simulations allow us to define mercury concentration histories that are consistent with the presently observed mercury concentration distribution. Although our results do not define a unique history of mercury contamination, they allow us to document 1) a reduction in mercury concentration on suspended sediment by nearly two orders of magnitude in recent decades, and 2) a similar reduction in the total mass of mercury stored in deposits of the river channel. We also constrain future mercury sediment contamination in response to different remediation strategies.