Performance, Analysis, and Simulation of a Two‐Well Tracer Test at the Mobile Site

Abstract

During the late summer of 1984 a two‐well tracer test was performed which involved travel distances of 38.3 m to 90 m. A bromide tracer slug was injected at one well while groundwater was simultaneously pumped from the other well. The two‐well test, which required 32 days to complete, was then simulated by GeoTrans, Inc., using a three‐dimensional advection‐dispersion model with a longitudinal dispersivity of 0.15 m, a transverse dispersivity of 0.05 m, and the permeability distribution being measured independently in a single‐well test. (Single‐well test results, which involved a travel distance of 5.5 m, indicated that the hydraulic conductivity varied by about a factor of 4 throughout the aquifer thickness.) The two‐well test simulation, made without knowledge of the test results, and without any calibration, was remarkably good. Local hydrodynamic dispersion was negligible, so that the concentration versus time breakthrough at the withdrawal well could be simulated just as well with a quasi‐three‐dimensional advection model having zero hydrodynamic dispersion. This implies that over the travel distances applicable to the two‐well test the spreading of the slug in the aquifer depended mainly on macroscopic velocity variations that were quantifiable in terms of the inferred permeability distribution. This result, which pertains to an aquifer in a fluvial terrace deposit of Quaternary age, is similar to that observed by Pickens and Grisak in a sandy glaciofluvial aquifer near the Chalk River Nuclear Laboratories in Canada. Such flow systems can result in fully three‐dimensional plumes with high concentrations of contaminant moving over large horizontal distances in the higher‐permeability zones.