FIELD METHODS FOR PREDICTING SOIL MOBILITY

Landslides with looser than critical-state soil porosities can transform into rapidly moving liquefied flows when contraction rates exceed diffusive pore-pressure equilibration rates allowing excess pore pressures to develop. Others have expressed the ratio of these rates by a dimensionless parameter (R) which characterizes the inclination for pore-pressure development:

R=kE/vμδ

where k is intrinsic hydraulic permeability, E is Young’s modulus, v is velocity of intergranular sliding, μ is the viscosity of the pore fluid, and δ is the average grain diameter. Pore-pressure fluctuations are likely for R<1 indicating that pore contraction or dilation rates exceed pore equilibration rates.

We present a field method for estimating R and identifying soils susceptible to flowing. Field measurements of k and δ are required while E and v are estimated. These estimates are a function of the porosity and grain size distribution of the field soil and are based on laboratory measurements for a wide variety of coarse-grained soils.

Laboratory tests (20 flows, 8 partial flows, 17 slides) were performed on 45 samples to evaluate the mobility of natural soils as a function of porosity. Failure mode was classified as a slide if failure occurred along discrete failure planes or as a flow if fluid-like behavior was observed. R<1 was consistent with 53% of the flows (R_med(median R)=0.9), 13% of the partial flows (R_med =2.3) and 0% of the slides (R_med =4.8). Among the flows with R<1, excess pore pressures were measured in 78% of the experiments (R_med =0.6). About half of the flows had R<1 and we attribute this low rate to artificially high k rates due to piping in the loose soil.

Laboratory measurements were combined with field measurements to estimate R for 30 natural landslides (16 flows, 3 partial flows, 11 slides). R<1 was measured for 75% of the flows (R_med =1.2x10^-2), 33% of the partial flows (R_med=3.0) and 18% of the slides (R_med=12.8).

With caution, this method can be used in the field to identify soils susceptible to flowing. The method is applicable to shallow landslides (<5m) that initiate in soils that are well-graded (C_u>8) with few to moderate fines (<18% silt + clay) and with low liquid limits (LL<40). When all of these criteria are met, R<1 was measured for 89% of flows (R_med =1.1x10^-4), 0% of partial flows (R_med =3.8) and 0% of slides (R_med =54.8).