[1] A mathematical model describing the steady state flows in a forced gradient tracer test between an injection and pumping borehole in a multilayered sandstone aquifer has been developed that includes the effect of vertically variable background heads. A second model describing the recovery of tracer from a layer in which there are discharges due to vertical flow in the injection borehole is also presented. Application of the models to field tracer test data indicates that the observed recoveries, which are not proportional to the abstraction rate in each layer, are consistent with the hydraulic behavior of the aquifer when natural vertical head gradients are taken into account. Investigation with the models illustrates that the vertical distribution of tracer recovery depends strongly upon the background heads and that tracer tests conducted in the same aquifer, but at different times, may interrogate different aquifer layers. It is also shown generally that for a given abstraction rate the vertical distribution of tracer recovery in small-scale tracer tests is controlled largely by the transmissivity distribution but that as the spatial scale of the test increases, the distribution of recovery becomes proportional to the discharges from the injection borehole because of vertical flows within it, which may be natural or induced by pumping in the monitoring borehole. Uncertainties inherent in the design of forced gradient tracer tests in multilayered aquifers and the problems of applying the results of such tests to natural gradient contaminant migration are discussed.

Forced gradient tracer tests between two boreholes can be used to study contaminant transport processes at the small field scale or investigate the transport properties of an aquifer. Full depth tests, in which tracer samples are collected just from the discharge of the abstraction borehole, often give rise to breakthrough curves with multiple peaks that are usually attributed to different flow paths through the aquifer that can rarely be identified from the test results alone. Tests in selected levels of the aquifer, such as those between packer-isolated sections of the boreholes, are time consuming, expensive; and the identification of major transport pathways is not guaranteed. We present a method for simultaneously conducting multiple tracer tests covering the full depth of the boreholes, in which tracer sampling and monitoring is carried out by a novel multilevel sampling system allowing high frequency and cumulative sampling options. The method is applied to a tracer test using fluorescein conducted in the multilayered sandstone aquifer beneath the city of Birmingham, UK, producing six well-defined tracer breakthrough curves.