UC Berkeley

Current levee analysis and design standards require the assumption of fully developed steady-state flow conditions for estimation of susceptibility of a levee to underseepage and/or piping via calculation of average uplift gradients. The use of steady state analysis conditions is generally considered the most critical scenario in terms of the development of pore pressure through and under the embankment; however, steady state seepage is not necessarily representative of real flooding conditions. Therefore, the objective of this research was to explore the transient development of pore pressure within and underneath a levee embankment, and to evaluate the influence of roots and animal burrows on the potential for development of preferential seepage and/or piping.

The first part of this dissertation focused on an assessment of the importance of transient seepage analysis in the estimation of heave or uplift gradients in levees for cases where a fine-grained blanket overlays a coarse aquifer layer. A series of analyses was performed to compare steady state and transient uplift gradients as a function of site geometry, aquifer and blanket hydraulic conductivity and flood duration. The results show that for high conductivity aquifers the transient seepage gradients exceed 90% of the steady state values within a few days during a flood event. However, for other geometries and flood durations, steady state seepage is overly conservative. Therefore, transient seepage analyses should be considered for newly designed levees or levees that are being re-constructed, assuming the sites are well characterized with respect to the geometry and properties of the foundation deposits.

The second part of the research focused on the assessment of the effects of woody vegetation and animal burrowing on levee performance, particularly seepage through the embankments. The results of two field seepage tests and the mapping of burrows at two test sites were used to provide prototype conditions for the transient seepage analyses. In general, the seepage tests showed that tree roots/root balls” tended to inhibit seepage front propagation and live roots did not serve as conduits for water as has been postulated in previous studies. The animal burrows, on the other hand, clearly presented readily available seepage paths when open to the waterside of the levee. The density of the burrows and their linear extent was found to be a function of the available food source, vegetation cover, and the frequency with which the burrows were eliminated. However, animal burrows that are not open to the water side do not fill with water until full saturation is reached at the burrow/tunnel wall. This outcome can be shown analytically or through numerical simulations as shown herein. Consequently, the results show that the continuity, the length of the burrows, location of the burrows, and degree of saturation all play a role in the development of through seepage with the through going burrows/pipes clearly providing a direct seepage path and an opportunity for piping to occur. On the other hand, burrows from the land side of the levees that do not penetrate completely do not start to flow with a significant velocity until after the adjacent soil mass is completely saturated.