Monitoring wells in clay: the apparently static water level and its influence during variable-head permeability tests

A permeability test in a monitoring well (MW) is used to assess the hydraulic conductivity, k, of clay. To interpret the test data, the piezometric level (PL) during the test must be known. A previous paper has concluded that due to a long time lag, the water level in the pipe is never a PL and any extrapolation of water level versus time gives an unreliable prediction of the PL. As a result, an incorrect PL yields an incorrect k value. This paper shows that a variable-head permeability test can provide a fair estimate of both the PL and the k value when it is interpreted using the central portion of the velocity graph. The first portion of the graph, when the head difference and the gradients are high, is curved by the expansion of the injection zone (undrained and drained wall displacement with eventually some creep), and thus cannot be used. The last portion of the graph primarily reflects, with some time lag, the natural variations of hydraulic heads in adjacent aquifers, and thus must not be used.