USING A LARGE-SCALE LABORATORY AQUIFER FOR TESTING METHODS OF MEASURING THE VERTICAL HYDRAULIC CONDUCTIVITY OF RIVERBEDS

Different methods for determining the vertical hydraulic conductivity (Kv) of riverbeds have been developed including conventional seepage meters, piezo-seeps (seepage meters that comprise a minipiezometer and use an induced flow across the riverbed) and infiltrometers. These methods are difficult to verify in the field without knowing Kv, but laboratory settings are typically too small for reliable tests. We constructed a large-scale laboratory aquifer tank (2.44-m diameter, 2.44-m high) to test, compare and refine these methods. The tank was filled with a well-sorted medium sand to a depth of 1.75 m and instrumented with 11 nests of three piezometers each. A pipe system allowed withdrawal or injection of water throughout the tank bottom, creating vertical flow and allowing independent measurements of Kv based on hydraulic gradients in the piezometer nests. A steady-state flow with water maintained above the top of the sand allowed simulation of a losing or gaining river. All methods were applied in multiple locations and compared to values derived from constant-head pumping tests.

Kv as determined from the constant-head pumping tests (CHPT), with heads measured in the piezometer nests, varied spatially from 36.3 to 68.8 m/d and averaged 39.5 m/d in the shallowest 1 meter. Conventional seepage metering with hydration bladders yielded Kv values averaging 0.60 times the CHPT values indicating substantial head losses in tubing and bag friction. Piezo-seep measurements, with water flux generated with a peristaltic pump, averaged 31.4 m/d, or about 0.79 times the CHPT average. While more accurate, the piezo-seep is problematic in coarse riverbeds due to the difficulty of pressing the minipiezometer into the bed and obtaining an adequate seal around the screen. We designed a new device, an infil-seep, by attaching a PVC pipe to the top of the seepage-meter bucket. The pipe is used as in a falling-head permeameter test. Like the piezo-seep, the infil-seep is not dependent on natural gradients and long time periods associated with filling seepage bags. However, as with slug tests, there is uncertainty associated with determining the proper length across which the gradient is measured. Comparison to the CHPT data allows proper interpretation of the infil-seep data.