Regional probabilistic modeling of required slope resistance to co-seismic failure including site amplification effects

Identifying slopes most susceptible to co-seismic failures at a regional scale is essential for prioritizing resources aimed at mitigating earthquake-induced landslide hazards. To improve reconnaissance-level assessments of such hazards, this study revisits a probabilistic framework of Del Gaudio et al. (2003) that relates a slope's intrinsic critical acceleration ac to the resistance demand imposed by regional seismicity, with particular emphasis on refining the latter. The resistance demand is expressed as the critical acceleration (Ac)x required to keep the probability of Newmark's displacement DN exceeding a critical threshold x below a prescribed level. We introduce two refinements to the original approach. First, a new empirical DN predictive equation is developed using 224 real and synthetic accelerograms representative of seismic sources affecting the landslide-prone Daunia Mountains (southeastern Italy). Although the region-specific model improves the accuracy of DN predictions relative to equations calibrated elsewhere, its impact on the estimated slope resistance demand is modest. Second, site amplification effects are incorporated in the estimation of (Ac)x by deriving site-specific amplification factors from seismic microzonation data through simplified 1D site response analyses. This refinement yields considerably higher (by up to 0.13 g) resistance demand estimates than those obtained using empirical amplification relationships. Application of the updated probabilistic framework to the Daunia Mountains produced regional-scale maps that represent the spatial variability of seismic demand in a more conservative manner. The proposed approach provides a sound basis for prioritizing investigations and mitigation measures in other landslide-prone regions where seismic microzonation data are available.