Water compositions in a pit lake developing at the Jamestown Mine, southern Mother Lode Gold District, California, can be modeled as the result of a series of geochemical reactions constrained by wall rock mineralogy, and chemistry of influent waters (rainfall, runoff, and ground waters). A series of inverse modeling calculations was performed using the USGS computer code PHREEQC. Mineral compositions from electron microprobe analyses were used as input constraints for wall rock minerals. Potential precipitating phases were chosen based on mineral saturation indices from speciation calculations of unfiltered water samples, and reasonable kinetic assumptions. Error ranges for water compositions were incorporated based on charge imbalance computations from analytical data, and laboratory measurement uncertainties.

Scenarios modeled are: early filling of the pit, development of summer stratification, a rainfall/runoff event during which arsenic concentration in the pit lake increased due to flushing of arsenic-rich sulfide weathering products from the pit walls, and development of winter mixis. For most scenarios, multiple minimal models were possible because of similarities in compositions of potentially reacting phases. The best model for each scenario can be chosen by verifying reasonable mixing ratios of incoming waters, taking into account seasonality in rainfall and evaporation, and evaluating relative reactivity of potential mineral reactants.