Even though liquefaction in gravelly soil is well documented in many earthquakes since 1891, most of the “simplified procedures” and national buildings codes still consider only sandy soil liquefaction in seismic hazard assessment. In this study, 109 sites of gravel liquefaction related to 27 historical earthquakes from 1891 to 2020 are reported, with a wide range of moment magnitudes, Mw (5.3 to 9.2) and focal depths (5.4 to 33 km), highlighting the potential for liquefaction of gravelly soils even during moderate earthquakes. Although gravels are often thought to have hydraulic conductivities high enough to preclude liquefaction, gravels that have liquefied are generally well-graded sandy gravels. The sand content is typically 30% or more so that the hydraulic conductivity is governed by the sand size making them clearly liquefiable. Even for gravels with lower sand contents, a low permeability surface layer has often been observed that could restrict drainage and produce excess pore pressure during strong shaking.
The epicentral distance calculated for gravel liquefaction sites plotted vs the magnitude of the earthquake event shows a pattern which closely follows similar curves provided in the literature for sandy soils. However, field observations of liquefaction in gravelly soils are less frequent in the historical record.
In addition to examining gravel liquefaction sites in general, this paper provides a case history illustrating the difficulties of liquefaction assessment in gravels at a site in Santa Rufina (Sulmona basin, Central Italy). This site, characterised by high vertical and lateral stratigraphic variability, was selected for gravel liquefaction assessment using a combination of in-situ tests, laboratory geotechnical analysis, and geological studies. We found that, even if SPT- and DPT- based in-situ methods provide conflicting results, the availability of a borehole log, along with standard laboratory test results, proved to be fundamental to achieving a reliable assessment of the liquefaction hazard.