Sensibility analysis in Probabilistic Seismic Hazard Analysis (PSHA), Chile as case of study

We performed a sensitivity analysis to assess the impact of some of the input parameters and methodological decisions on the calculation of probabilistic seismic hazard (PSHA). We work on continental Chile, between 14° and 46° south, a country characterized by high seismic activity, which can be classified into three main regimes: interface, inslab, and crustal. 
The interface regime corresponds to the boundary between the South American and Nazca plates up to a depth of 60 km. Intraslab seismicity occurs within the Nazca Plate between 60 and 200 km depth, while crustal seismicity develops on the South American Plate. One of the main challenges in classifying seismic events in these regimes is differentiating between crustal and interface seismicity in the first 30 km depth, especially when focal mechanisms are not available. To analyze the effect of this classification, we tested 20 different scenarios defined by the horizontal distance from the trench in degrees (1.7–2.0° in 0.05° increments) or by the perpendicular distance to the trench in kilometers (5–20 km in 1 km increments).
To determine the recurrence parameters, two previously published zonal models for Chile were used (Martin, 1990 and Molina et al., 2021). The calculation of the recurrence parameters for each seismogenic zone followed the usual steps: (1) declustering the seismic catalog using various space-time windows (e.g., Reasenberg, 1985; Gardner & Knopoff, 1964); (2) estimating the magnitude of completeness by means of maximum curvature and completeness analysis (Stepp, 1972); and (3) calculating a- and b-values using means of least squares (MMCC), maximum likelihood and Weichert methods. While MMCC is less favored in current practice, it remains in use in some regions (e.g., Benito et al., 2010; Nuñez, 2014; Gamboa-Canté et al., 2024). Therefore, its impact was also evaluated.
Preliminary findings reveal that the boundary separating crustal and interplate regimes has minimal influence on completeness magnitude, completeness analysis, or b-value estimation. However, the choice of space-time windows for declustering significantly affects the a-value, producing variations from 5 to 6.7 in certain seismogenic zones. These differences have a pronounced effect on PSHA results, highlighting the importance of careful parameter selection in seismic hazard studies.