New empirical relationships between magnitude and distance for liquefaction

Abstract

Historical research performed in the 1990s has updated previous compilations of liquefaction-induced phenomena that occurred during the last millenium in Italy. Liquefaction indications are reported in Italy for earthquakes with I_o>5–6 (MCS) and M_s≥4.2, 90% of the cases falling within 50 km of the epicenter. The recently re-evaluated seismic parameters of the Italian historical earthquakes, together with the location of 317 indications of liquefaction features, provide a relatively complete database, permitting the author to highlight the distribution of intensity/magnitude values versus epicentral distance. In paleoseismic analyses these relationships may be considered a tool in evaluating the minimum energy of an earthquake that induced liquefaction features.

Introduction

Although the systematic study of liquefaction features is a young discipline, it is widely accepted that the recognition of this phenomenon (e.g. by means of paleoseismological studies) can be assumed as an indicator of strong past earthquakes (i.e. Tinsley et al., 1985). However, since other natural causes can produce or mimic liquefaction features (e.g. structures of syndepositional origin, due to artesian condition, formed by weathering or in periglacial environment), the non-seismic causes should always be investigated (see Obermeier, 1996) in order to avoid hazardous statements about the seismicity of an area.

Liquefaction is the transformation of a granular deposit from a solid state into a liquefied state as a consequence of the increased pore-water pressure determined by cyclic shaking (Youd, 1977). Liquefaction features may vary from place to place in geometry, type, and dimension, due to the anomalous propagation and amplification of the seismic waves at the surface and to the differing site conditions (grain size and density of deposits, position of the ground-water level). It is therefore sometimes difficult to recognize liquefaction in the field, and particularly difficult through the historical description of earthquake-induced effects (Galli and Meloni, 1993).

The most common and conclusive surficial features induced by liquefaction are sand blows that occur both isolated (sand volcanoes; Fig. 1, Fig. 2) or along fissures. Other clear liquefaction-induced phenomena are the lateral spreads of huge masses of soil overlying a liquefied layer or the geometrical settlement of surficial deposits (i.e. craters; Fig. 3, Fig. 4, Fig. 5). A typical effect on anthropic structures is the differential settlement and tilting of buildings, bridges and quays, the swelling of pavement of the ground-floors of buildings or swimming-pools, and the apparent extrusion of pillars or wells above the ground surface, due to the sinking of the surrounding soil (Fig. 6).

All these effects are easily detectable during earthquake field surveys, but some of them are difficult to recognize and/or identify univocally as due to liquefaction in the historical description of earthquakes. This is mainly due to the convergence of forms with other collapse features and damage caused by seismic shaking.

Although the most impressive liquefaction features occur during strong earthquakes (M_s≥6.5), whose epicentral areas are located close to regions highly susceptible to liquefaction (e.g. the alluvial plains of South Carolina during the earthquakes of 1886; the plains of Alaska, 1964; Nijgata, 1964; Kobe, 1995; Izmit, 1999), the database of historical liquefaction in Italy (Galli and Meloni, 1993, Galli et al., 1999) also accounts for a large number of moderate earthquakes (M_s≥4.2) producing liquefaction (as for liquefaction cases induced by moderate earthquakes, see also Papadopoulos, 1993).

In fact, the Italian territory shows a remarkable mismatching between high seismicity areas (mainly located inside the Apennine chain) and broad liquefaction-prone areas (the Adriatic and Tyrrhenian coasts, and the Po River alluvial plain; Galli and Ferreli, 1995). Among those last-mentioned regions, only Calabria, eastern Sicily, northern Apulia and, partly, the western Ligurian coast experienced strong historic earthquakes. Nevertheless, the long tradition of historical written sources in Italy supplies data also concerning areas characterized by moderate seismicity, allowing the recognition of 59 liquefaction features resulting from earthquakes with M_s≤5.9.

The liquefaction database presented here is a completely reviewed and updated version of previous catalogs. It is the result of new systematic historical research performed in archives and libraries utilizing only primary sources (Galli et al., 1999).

The data suggest relationships between the epicentral distance of a liquefaction feature and the magnitude (and intensity) of the earthquake. They also suggest a relationship between epicentral distance and site intensity. The curves obtained replace those reported by Galli and Meloni (1993) and Galli and Ferreli (1995), for which earthquakes were characterized only in terms of MCS (Mercalli–Cancani–Sieberg) intensity, by using different types of magnitude and re-evaluated intensity values from Italian updated seismic catalogs (NT4.1, Camassi and Stucchi, 1997 and CPTI, 1999).

With respect to Ambraseys (1991) and Papadopulos and Lefkopulos (1993), the bounding equations of datapoints distribution presented here account for a larger number of liquefaction features (317) that occurred during 61 earthquakes characterized by focal depths <20 km and mainly related to normal faulting. Landslide descriptions have been a priori excluded from the database in an attempt to reduce ‘contamination’ of the data by liquefaction features of uncertain interpretation.