Abstract
Transient seepage analyses, which are becoming more common in practice, carry inherently more complexity compared with traditional saturated steady-state seepage analyses. The results of a four-year remote monitoring investigation were used to investigate common practices used in transient seepage analyses. Initial pore water pressure distributions were found to correspond to predicted infiltration distributions, which were less than typically assumed. The laboratory-measured drying soil water retention curve was found to provide an upper bound to field measurements. Field-measured soil water retention data were found to better correspond to a mean between the laboratory wetting and drying curves. Transient seepage and stability analyses showed that using a drying soil water retention curve resulted in lower factors of safety compared with using a wetting curve. However, a mean curve between the wetting and drying curves proved to be more accurate when compared with representative field measurements. Using unsaturated shear strengths along with conventional saturated shear strengths for levee embankments was found to minimally contribute to the stability factor of safety. Incorporating the findings from this investigation into a transient seepage analysis will help to improve the reliability of the results.