High-Resolution Hydraulic Conductivity Profiling with Direct-Push Methods

Spatial variations in hydraulic conductivity (K) are a primary control on solute movement in groundwater.  Characterization of these variations, however, has proven to be difficult. Over the last two decades, significant progress has been made in utilizing direct-push technology to characterize K in shallow unconsolidated settings. The continuous Direct-Push Injection Logger (DPIL) was designed to provide high-resolution information about relative variations in K. Water is injected continuously through a screened port while the probe is advanced and pressure response is monitored behind the screen. Since the DPIL only provides qualitative K information, methods are needed for transforming DPIL ratios into K estimates. The High-Resolution K (HRK) tool was designed to address this issue by co-locating high-resolution DPIL profiling with the direct-push permeameter (DPP) hydraulic tests; this collocation allows DPIL ratio to be directly transformed into K estimates. The previous HRK tool had an upper DPIL-K limit of 10 m/d, which was primarily imposed by frictional losses caused by the small diameter of DPIL injection tubing. Although increasing the tubing diameter allowed use of higher injection rates, those higher rates resulted in significant nonlinear flow losses that could greatly exceed the formation pressure responses in higher-K zones. In order to correct for these flow losses, we developed an approach based on the step drawdown test procedure commonly used to characterize flow losses in high-capacity pumping wells. The removal of these flow losses allows the DPIL-K limit to be extended upward by a factor of six (60 m/d), thus increasing the utility of the tool in permeable settings. The modified HRK approach was applied in a sand/gravel aquifer at a field site in northeast Kansas, which indicated that this method can be used to characterize K in heterogeneous aquifers at a profiling resolution and speed that has not previously been possible.