Innovative use of groundwater modelling tools for holistic assessment of remedial options for mine tailings in a marine intertidal environment at Toquaht Bay

Magnetite iron ore was mined from the abandoned Brynnor Mine in the 1960s. It was crushed and magnetically separated to produce sand-sized tailings that were deposited in Toquaht Bay to produce sandy upland, intertidal beach and subtidal areas. In the Maa-Nulth Final Agreement (Treaty), the upland area was transferred to Toquaht Nation and the Province of British Columbia committed to remediating the area. Environmental investigations found groundwater contaminated with arsenic and cobalt discharges to the intertidal area. Porewater investigations found unacceptable toxicity to intertidal aquatic life. Subtidal tailings do not present unacceptable ecological risks. Three preferred remedial alternatives were evaluated: • Partial Solidification, involving ex-situ solidification of some tailings with cement to form a cap. • Landfill Disposal, involving removal of tailings to below sea level and backfilling. • Commercial Option, involving placement of asphalt/geomembrane cover. Three numerical groundwater models were used to evaluate remedial options considering different aspects of the complex tidal groundwater-seawater flow system. A 3D MODFLOW model estimated the average annual reduction of contaminated groundwater discharge through tailings to the lower intertidal area and used to assess effectiveness of remedial options. A 2D SEAWAT model evaluated the influence of remedial alternatives on density-driven groundwater flow. A 2D HYDRUS model simulated transient cycling/exchange of seawater and groundwater in response to hourly tidal conditions in the intertidal zone. It also calculated contaminant concentrations in the remedial cap based on conservative contaminant transport processes and hydrodynamic mixing. Each model addressed different modelling objectives and were considered in conjunction with the results of the other two models and with an understanding of the models’ capabilities and limitations.Non UBCUnreviewedOthe