Physical Modelling of Laterally Spreading Soil with and without Plant Root Analogues

A key problem experienced when soils liquefy during earthquakes is that gently sloping ground can translate downslope. Most recently, this lateral spreading was seen to affect large portions of coastline around Palu Bay in the 28 September 2018 Indonesia earthquake. Soil observed to experience lateral spreading frequently contains plant roots that in principle should have a reinforcing effect on the soil. However, field data is difficult to interpret due to the many uncontrollable variables, so there has not previously been an attempt to quantify roots effects on lateral spreading. Physical modelling offers an opportunity to test soil in real site conditions with controlled variables (soils, slopes, stratification etc.). In this study, a small scale model with the potential to undergo lateral spreading is created and tested on the geotechnical centrifuge. Considerations of how the model could exhibit lateral spreading without boundary interaction, and how both fibrous roots (like grass) and more woody roots (like a shrub) could be modelled in asmall scale model, are discussed. The model is shown to recreate a lateral spread during centrifuge testing when the soil contains no roots. When tested in moderate-sized earthquake events, the lateral displacement was reduced by ~60% by fibrous root analogues and 70% by woody root analogues. However, in stronger events where liquefaction occurred to greater depths, then the near-surface roots offered less resistance to displacement. These results indicate that without roots, spreading could be worse, but the roots alone may not be enough to prevent damage completely, particularly in larger earthquakes