On how the ergodic assumption influences seismic risk

The Monte-Carlo (MC) simulation approach has been widely applied to the problem of probabilistic seismic hazard assessment (PSHA) for geographically distributed systems. Although the conventional integral approach is well-established in PSHA, the MC simulation approach is an efficient and flexible alternative when complicated factors, such as spatial correlation of ground shaking are involved. The objective of this study is to assess how the ergodic assumption influences the economic loss results for a region. To meet this objective, MC simulation and multi-scale random fields techniques are used to assess the probabilistic seismic hazard of a region, while incorporating estimated site-terms (δS2Ss) based on available strong-motion records at specific sites and spatially correlated δS2Ss at others in addition to single-station sigma to the PSHA process. The proposed method is applied to the Fatih district of Istanbul in the case of a moment magnitude 7.3 earthquake on the Marmara Fault. Results indicate that the probability for total economic loss being underestimated due to the ergodic assumption by at least 29% is 0.50 in the case of a moment magnitude 7.3 earthquake on the Marmara Fault.