Effect of Input Data on Predicting Plume Location Using MAROS Model

Long-term monitoring plans for groundwater remedial systems use monitored natural attenuation and risk-based goals to reduce the high cost of conventional remediation such as pump-and-treat. Several computer models that evaluate the condition of contaminated plumes and offer information on optimizing the testing plan rely on reasonably organized and accurate data. Uncertainty is interjected from testing methods, use of hydrogeological models used to predict hydraulics of a plume, and the limitations of the computer model. The "user friendly" optimization model MAROS, Version 2.2, is offered by the Air Force Center for Environmental Excellence. This model creates temporal and spatial overviews of the plume with Mann-Kendall, linear regression, and moment analysis using Delaunay triangulation. It further predicts deletion or addition of wells, sufficiency of the wells, and the degree of degradation of the plumes. One specific case study evaluated the condition of existing TCE is shallow aquifer plumes. Initial data was analyzed regarding issues of uncertainty, reliability, frequency of collection, and variations in the overview of the plume conditions. Data that is too random and gapping in time results in inability to analyze plume and varying TCE degrading concentrations. MAROS is considered a generalized management predictability tool to optimize and lower the cost of plume monitoring while meeting appropriate risk levels; it too must rely on certain parameters of data collection and management to provide results that are meaningful and useful.