Groundwater Flow: Models, Data, and Uncertainties

Abstract

This paper addresses three major problems in groundwater flow determinations. First, a quantitative method is provided to show how data can be used to help disentangle more appropriate models of groundwater flow from less appropriate, and a quantitative definition is provided for such resolution. An example is given from steady-state one-dimensional flow to illustrate the method using a continuous, perfectly measured head data field. Second, the method is generalized to show how it operates with finitely sampled, but perfect, data. Third, the effects of data uncertainty are examined and a procedure provided which allows determination of parameters, and their uncertainties, in models from the uncertain data field. An example is given to illustrate when it is appropriate to collect more data of higher quality (less uncertainty) and when it is appropriate to improve model resolution before one embarks on further data collection.

The importance of the quantitative procedure is that ist allows one to address when one has sufficient data resolution that models can be distinguished one from another, and the procedure also provides information on when the models are not adequate, so that better representation of the hydrogeological flwo regime is necessary.