Knowledge of aquifer characteristics, including transmissivity distributions, heterogeneities, and the vertical distribution of hydraulic head, is essential for the construction and calibration of regional groundwater flow models. Dynamic flow logging, in which vertical flowmeter data are collected while a well is pumped, offers a cost-effective alternative to traditional pumping and packer tests for obtaining such information.

We have successfully conducted dynamic flow tests on several deep (1500-2000 ft), large-diameter (10-24 in.) production wells in support of a regional modeling project in southeast Wisconsin. The Cambrian-Ordovician sandstone aquifer, which consists of a series of sandstone and carbonate units more than 2000 ft thick, is the major source of groundwater for municipal and industrial use in this area. Heavy pumpage concentrated in the western suburbs of Milwaukee has created a large cone of depression with a decline in head of more than 500 ft from predevelopment conditions. Total dissolved solids (TDS) levels are increasing in several municipal wells near the center of this cone. Calibration of a three-dimensional groundwater flow model of this region requires information on the vertical distribution of transmissivity and hydraulic head across multiple aquifer units.

The dynamic flow tests utilized a full suite of downhole logs (gamma, normal resistivity, temperature, fluid conductivity, caliper, video) as well as impeller and heat-pulse flowmeters. In all cases the logging tools were lowered through small-diameter drop pipes to avoid tangling with the pumping equipment inside the well. The heat-pulse flowmeter was most effective for measuring ambient borehole flows under static conditions. The impeller flowmeter was more effective during pumped conditions and clearly indicated the zones contributing water to the wells. Combining the static and pumped logs through mass-balance calculations allows estimates of the transmissivity and relative hydraulic heads of the various aquifer units penetrated. The use of flow inversion algorithms greatly enhances these interpretations.