Sensitivity Analysis of Seismic Parameters in the Probabilistic Seismic Hazard Assessment (PSHA) for Barcelona Applying the New R-CRISIS

Resumen

One of the most common ways to represent results of the probabilistic seismic hazard assessments (PSHA) are maps of seismic hazard, which usually show values of Peak Ground Acceleration (PGA) in a region for a return period. A common return period is of 475 years. These types of maps are frequently incorporated in seismic codes, which include the minimum requirements to design new buildings. On the other hand, a sensitivity analysis usually gives us additional information about a procedure or result. For instance, a sensitivity analysis about a PSHA can give us information about what variables considered to compute the seismic hazard have a significant influence on the results of seismic hazard. In the present study, we performed a sensitivity analysis related to the PSHA for Barcelona. This analysis was oriented to identify the influence in the results of seismic hazard of the following variables: a) the relationship magnitude-macroseismic intensity chosen to convert values of macroseismic intensity to magnitudes; b) the beta parameter that was used to define part of the seismicity of a seismic source and; c) the ground motion prediction equation (GMPE) which was used to determine intensities values to different distances from the epicenter of earthquakes. For this purpose, we applied the code R-CRISIS, which is the updated version of CRISIS2015. Therefore, the present study had as an additional objective to test the functionality of the new R-CRISIS from the point of view of users. According to the results of the sensitivity analysis of the PSHA of Barcelona both the GMPE and the relationship magnitude-macroseismic intensity are the two variables that have the greater influence on the results of seismic hazard of the present study. For instance, in some results of seismic hazard for the city of Barcelona the values of PGA (for the same return period) differ between them until 82% depending on if we considered the means values of the PGA values of the GMPE or the mean values plus one standard deviationof the PGA values of the GMPE. Finally, according to our experience in the use of the new R-CRISIS in the present study, we can confirm that it is both a powerful and user-friendly software. R-CRISIS has valuable features, for instance, it allows to consider diverse types of criteria to define the type of seismic source. For example, R-CRISIS allows defining different types of geometries of the seismic sources, and different criteria to define the seismicity of each seismic source. An important feature of R-CRISIS is the fact that it includes a database with numerous GMPE ready to be used to compute seismic hazard. Additionally, R-CRISIS has valuables graphical tools which are very helpful during the following two stages: the assigning data and the analysis of results.