NUMERICAL GROUND-WATER FLOW MODELING OVER THE LAST 30 YEARS—INNOVATIONS IN DATA INPUT, OUTPUT, AND DISPLAY (AND UNDERSTANDING?)

Modern numerical ground-water flow modeling began in the late 1960s with the advent of digital computers. Finite-element and finite-difference methods were developed almost simultaneously. Finite-difference methods were relatively easy to use, but at the time they did not allow for a moveable seepage face (water table), which finite-element methods did. Through the 1970s, the most popular modeling codes were the Prickett and Lonnquist (1971) and Trescott, Pinder, and Larson (1976) finite-difference codes. These codes were capable simulating two-dimensional transient flow in heterogeneous, anisotropic aquifer systems. These codes required line (or card) entry of data input; output usually consisted of “green-bar” paper, which would then be transferred to a hand-drawn map of the modeled area, with a grid overlay. If the user desired any additional capabilities, such as pre- or post-processing, additional Fortran code had to be either integrated into the existing modeling code or written as separate stand-alone code. For several years, this method of modeling was little changed, even with the development of MODFLOW in 1983. Computers that were accessible to most people in the early 1980s had primitive graphics capabilities and generally were not used for graphical analysis of model output.

Geographic information systems (GIS), although they had been in general use since the early 1980s, were not commonly used by ground-water modelers until the late 1980s and early 1990s. Current graphical user interfaces for numerical ground-water flow models are integrated with GIS capabilities and input/output editing and manipulation, are PC-based, and are relatively easy to use for computer-literate modelers. New modeling routines, such as parameter estimation and optimization, are also easily incorporated using the current technology. However, as technology facilitates professional and public understanding of the intricacies of ground-water flow, it also facilitates creation and propagation of mistakes in data entry, conceptualization, and understanding of ground-water flow systems.