A Quantitative Framework for Comparing Observed and Simulated Pit-Lake Geochemistry

Abstract

Pit lakes are commonly studied using advanced geochemical models. However, quantitative post-audits of these studies to assess the degree of agreement with observations are rare. This study used a quantitative evaluator (error variance parameter; Δ) to compare predicted and observed geochemistry of nine pit lakes across three continents and five ore-deposit types. The compiled modeling studies applied a variety of methods to incorporate geochemical (sulfide oxidation and mineral precipitation) and hydrologic processes (groundwater flow and hydrodynamics). Several common ions (Ca, Mg, SO4) and pH were shown to compare well with observations. In contrast, comparisons of model simulations and observations for trace metals and metalloids such as Fe and As were less congruent. The discrepancies for these and other species is potentially linked to inappropriate application of equilibrium thermodynamics and temporal model discretization. Model complexity was also shown to be related to the reproducibility of observed geochemistry. Quantitative analysis using Δ in a wider range of studies may increase understanding of prevalent geochemical processes invoked in pit-lake modeling.