Evaluating the Utility of Watershed-Scale Numerical Models in the U.S. Geological Survey's Mineral Environmental Assessment Project

The Development of Mineral Environmental Assessment Methodologies Project (MEAP) is a new project within the U.S. Geological Survey’s Mineral Resources Program. One objective of the MEAP is to predict potential environmental impacts at a watershed scale, with associated uncertainty, from future mining activities. One initial task is to determine the role of numerical modeling in this effort. More specifically, the objective of this task is to identify numerical modeling strategies that can predict metal concentrations at a selected downstream compliance point. Questions to be addressed include: 1) What scale is appropriate, 2) Which numerical models, if any, are appropriate and useful under a variety of different conditions, 3) Can we simulate metal transport through the system, and 4) How uncertain are model predictions likely to be given realistic and limited data coverage?

Two sites in Colorado with hydrothermally altered bedrock (Prospect Gulch and Handcart Gulch) are being used to develop the MEAP’s numerical modeling effort. Data on ground-water flow, surface-water flow, geology, and geochemistry are available at both sites for use in integrated modeling efforts. Efforts will focus on simulating current conditions before simulating possible future mining influences. At these sites, the “watershed scale” is on the order of 5 km². Ground-water flow models have been developed to assist in understanding flow in fractured bedrock at this scale (MODFLOW and FEFLOW). The aqueous geochemistry modeling program PHREEQC will be applied to assist in understanding geochemical reactions. Inverse modeling will be used to evaluate the sensitivity of model parameters and test prediction uncertainty based on available data. Research at Prospect Gulch and Handcart Gulch will provide the initial basis for the MEAP in scaling up to larger watersheds (up to 200 km²) where model selection, more limited data, and greater prediction uncertainty will be more important.