Simplified empirical method for predicting earthquake-induced settlements and its application to a large area in Spain

Abstract

Most of the commonly used existing empirical methods to predict earthquake-induced settlement (S) in sandy soils require numerous iterations or the use of charts, tables and diagrams. In general, these methods estimate the one-dimensional settlement of dry sandy soil on level ground by using a well-known step by step procedure based on Standard Penetration Test (SPT) values, which is particularly effective for practical applications. A review of the state-of-the-art methods shows that seismic settlement in all cases increases as the layer thickness (h) of sandy soils increases and corrected SPT blow count (N1)60 decreases. With that in mind, we propose a novel simple way to estimate S based on the h/(N1)60 ratio for a reference earthquake magnitude. This approach provides a tool that can rapidly obtain S in numerous sites and can be applied to large areas. In the last fifty years, the population of the Metropolitan Area of Granada (MAG) has doubled. The amount of developed land has increased by approximately 4650ha and the areas with the greatest population and construction growth are located on sedimentary deposits. The land beneath the urbanized areas of the MAG is located on alluvial, colluvial, silt and clay deposits with different thicknesses of granular soils and varying water table depths. The MAG is acknowledged to be the most seismically active zone in Spain and seismically-induced phenomena such as liquefaction and ground settlement were reported in specific zones during moderate (1806) and strong (1431) historical local earthquakes. The present study focuses on differential vertical displacement assessment in this large area of Southern Spain for two earthquakes of magnitude Mw 6.6 and 7.0. The maximum expected settlement due to earthquake shaking of alluvial soils, sandy soils and fine soils (clay and/or silt) was obtained by correlating the mentioned h/(N1)60 ratio with the S predicted by two well-known methods. Vs values have been estimated from (N1)60 data using methods proposed in the literature and tested with Vs local data from SPAC and refraction profiles. The results from the new formula proposed here show predictable settlement ranging from 0.1 to 21.4cm and up to 24.7cm for the 6.5 and 7.0 earthquakes, respectively. These were greater than 2.0cm and 3.3cm in the north-central and north-western sectors of the study area, especially in the town of Atarfe and along the road between Pinos Puente and Atarfe, the same zones where settlements were observed in the 1806 earthquake. Zones where earthquake building damage may appear have been detected by comparing the maximum S and the maximum permitted settlements derived from different angular distortion values considering 5 and 6m as typical distances between supporting structures.