Early Groundwater Model
By Darrel Dunn, Ph.D., PG, Hydrogeologist
The purpose of this web page is to place an early finite-difference groundwater modeling experiment on the World Wide Web. The value of this experiment is only historical. The modeling was published in my University of Illinois Ph.D. Thesis (138) . As far as I know the model was the first attempt to calibrate a groundwater model to actual water level data. This modeling occurred because my Petroleum Reservoir Engineering professor suggested that the method cold be applied to groundwater reservoirs. Previously, it had been applied only to petroleum reservoirs. Consequently, I studied the application to petroleum reservoirs and researched ways of obtaining solutions to large sets of finite-difference equations. I concluded that the simplest approach would be Gauss-Seidel iteration. When I used this method in my thesis project, I applied extrapolation to increase the rate of convergence. The Gauss-Seidel method with extrapolation is now called successive over relaxation (SOR).
My Ph.D. thesis involved the hydrogeology of an area around Stettler, Alberta, Canada. Stettler produced its water from wells scattered around the town. The water came from an inhomogeneous aquifer. Consequently, I had an opportunity to use a finite-difference groundwater model to study the wellfield. I think this was the first use of a finite-difference model to simulate a water well field. The following excerpts from my thesis describe the model and the results. Click the links.
Recent groundwater models are much more complex. Computers can now handle more computationally intensive code without utilizing excessive CPU time.
After being awarded my Ph.D., I completed many projects involving ground water modeling and surface water modeling. I have used MODFLOW, SWIFTIII, MIKE SHE/MIKE 11, AEM , MODPATH, MT3D, UTCHEM, and other programs including ones I have developed with FORTRAN, PYTHON and other languages.