Empirical regressions for distributed faulting of dip-slip earthquakes

Abstract

Probabilistic Fault Displacement Hazard Analysis (PFDHA) is pivotal in assessing the probability of surface fault displacement during seismic events, with critical implications for infrastructure systems like pipelines and lifelines. Recent earthquakes have underscored the importance of fault displacement as a cause of damage. This study aims to improve predictive models and regression analyses to assess the probability and expected amount of throw on distributed ruptures (DR) from normal and reverse earthquakes. The research leverages the SURE 2.0 database, which classifies ruptures into different ranks, including primary and various typologies of distributed ruptures. Logistic regression models are employed to assess the probability of DR occurrence, considering factors such as earthquake magnitude, proximity to the principal fault, faulting style, and DR location (hanging wall/footwall) as predictor variable. In addition, a fault displacement model is formulated to estimate the median throw on DRs based on these parameters and on a mean throw on the principal fault. The outcomes offer valuable insights into the characteristics and likelihood of DRs, carrying implications for PFDHA applications. The study introduces a decision-tree algorithm designed to assist PFDHA practitioners in selecting the most suitable approach based on the available geological knowledge. This study contributes to an improved understanding of fault displacement hazard analysis and provides a versatile framework for its application across diverse geological contexts.