In this issue, we present two questions for debate. The first is a technical question related to the use of modeling to predict and evaluate MNA, and the second addresses the social issues of embedded impacts of active remediation versus MNA. For the past two years, the journal has published this forum. To date, ten questions have been considered and a total of 34 responses published. The questions are based on topic ideas we received from the MNA Panel, the journal's advisory board, and our readers. We do our best to present questions related to both the technical and social issues surrounding current natural attenuation topics and obtain responses from diverse backgrounds. We encourage you to send your questions or comments on these topics to raise further discussions for the journal. Please send your thoughts to Bob Norris at nrkathleen@qwest.net or Mark Ferrey at mark.ferrey@pca.state.mn.us.Question 1: Natural attenuation remedies are based largely on predictions drawn from sitespecific models, whether they are general conceptual models or more detailed fate and transport models. How reliable are these tools in evaluating natural attenuation as a remedy that will protect human health and environmental quality? Which of these two types of models are appropriate for most monitored natural attenuation projects?
PANEL MEMBER-JOHN T. WILSON, PHD, US EPA/R. S. KERR LABORATORY, ADA, OKLAHOMAIn general, models are used to forecast natural attenuation and create an expectation for natural attenuation, but long-term monitoring data are used to evaluate whether natural attenuation is protective of human health and environmental quality. A model is neither reliable nor unreliable. The model is nothing more than a tool that helps organize the information available about a site. A model will enlighten or mislead depending on who uses it, how, and for what purpose. I have been watching people use (and abuse) models for regulatory purposes for more than 25 years. Enforcement actions work better when there is a consensus about the behavior of contaminants between the regulators and the people that are responsible for a site. Models are a good way to achieve that consensus. In almost every case, the person that calibrates the model will compare the output of the model to their personal internal mental model of the site. If the output does not "look right," they vary the values used to calibrate the model until it matches that internal mental model. Ideally, the formal model