Control Function Measures for Hydrodynamic Problems

Abstract

This short paper shows that different choices made for control functions of information in hydrodynamic flow problems can have significant implications for interpretations of the system. Using as a simple illustration the case of a steady-state, one-dimensional flow model with no internal sources or sinks and with the hydraulic conductivity depending on a single parameter and the distance from the origin, it is shown that, even when a continuous, error-free head data set is provided, statements about the uniqueness or not of the inverse solution are conditioned on the choice of the control function. Care has to be exercised in obtaining physically meaningful results and, depending on the model assumptions and the data available, there may not be acceptable models. It is also shown that there may be more than one model behavior that is acceptable. The results have implications for the hydrodynamic upscaling problem for flow in permeable media, for ensemble averaging methods, and for parameter determination for deterministic models of permeable flow.