TESTING REGRESSION MODELS FOR ESTIMATING NEWMARK CO-SEISMIC DISPLACEMENTS IN MODERATE TO LOW SEISMIC AREAS

The assessment of earthquake-triggered landslide hazard at a regional scale always implies important simplifications. Uncertainties in accurately defining strong ground-motion characteristics at a particular site, as well as in modeling the geotechnical parameters of the slope mass and its dynamic behavior, make earthquake-triggered landslide assessment a very complex matter. The most common procedures followed in regional assessments deal with the well-known Newmark sliding rigid-block method.

The estimation of Newmark displacements in regional hazard assessment is usually done by making use of regression models based on basic earthquake parameters (magnitude and distance) and/or simple strong ground-motion parameters (i.e. PGA, Arias intensity). At present, different authors have proposed regression models from earthquakes occurred around the world, usually of moderate to high magnitude (M_w > 6.0).

In this work, we compare a number of these regression equations to select one that could be used to study areas with moderate to low magnitude earthquakes. This type of seismic scenario is the most common in Spain, where during the last decades several moderate to low magnitude earthquakes (M_w < 5.5) have triggered multiple slope instabilities. In particular, the 2011 Lorca earthquake (M_w = 5.2) triggered far more slope instabilities than any other instrumental earthquake recorded in Spain. We have used the well know data of this seismic event (magnitude, distances, slope instabilities location, geotechnical parameters) to estimate Newmark displacements on the earthquake-triggered landslides location to select the regression equation which offer the best results. This regression equation could be used in similar areas with moderate to low magnitude earthquakes for regional hazard assessments.