A method was developed for assessing the vulnerability of groundwater to contamination from contaminant sources at the soil surface, using a numerical groundwater flow model linked to a digital map database. The method was applied using the pcARC/INFO Geographic Information System (GIS) to input, store, and manipulate base maps, resulting in a database of digital maps for the alluvial aquifer system in the Willamette Valley of western Oregon. Digital elevation maps were created by digitizing topographic maps of land surface (1:250,000 scale), water surface, and the base of the Tertiary-Quaternary sedimentary deposits (1:500,000 scales). Soil association and aquifer unit maps digitized from 1:500,000 scale map sheets were also used. Data were extracted from ARC/INFO to the SURFER software package to create a 3-D surface model for each of the digital elevation maps. An ARC/INFO point coverage was then used to store and overlay these surfaces, allowing the creation of maps of depth to water, saturated thickness, and water table gradient. These data became the input to a numerical finite element groundwater flow model. The model solves a dual formulation problem for the potential function and the stream function to calculate the time-of-travel for water to flow from the surface to the water table and laterally for 100 meters as an index of groundwater vulnerability. A cluster analysis is used to condense the data and form a training data set for a multiple regression model. The regression model is fit to the results of the finite element model with an R-squared of greater than 0.96. The simpler regression model is then used for mapping travel times for the entire study area. When properly calibrated against the finite element model and when combined with the digital map database and Geographic Information System (GIS) procedures described, the regression model can be conveniently used to assess the vulnerability of groundwater to contamination over large areas. / Graduation date: 1989