Uncertainty and Parameter Estimation

No matter how sophisticated the groundwater model, a model is by definition a simplification of reality—a simplification that tries to accurately characterize mostly unseen water flowing in an unknowably complex earth under future stresses that can only be guessed at. Therefore, it is not surprising that groundwater models can no longer be considered “crystal balls” foretelling the future, but rather are best discussed in the context of their inherent uncertainty. Although general uncertainty concepts are widely understood, best practices for conveying uncertainty estimates are not well established. Parameter estimation techniques provide a quantitative framework that can form a common language and provide metrics for discussing model uncertainty. As a result, parameter estimation approaches are an increasingly indispensable part of 21st century groundwater modeling.