Quantifying the role of logic tree parameters in soil probabilistic seismic hazard assessment: Insights from sensitivity analysis

This paper describes probabilistic seismic hazard assessment (PSHA) incorporating site effects for a deep soil site in Pakistan. Bedrock hazard is computed in FRISK88M using a logic tree approach. Probabilistic site response analysis (PSRA) employs Monte Carlo simulations to account for uncertainty in the soil profile. Site-specific amplifications are derived via regression of equivalent linear site response analysis using STRATA. By incorporating site amplifications, rock ground models are transformed into site-specific models; subsequently, we performed soil PSHA in FRISK88M. To assess the contribution and uncertainty of individual input parameters, sensitivity analysis is required. In this study, sensitivity analysis of logic tree nodes quantifies the uncertainty contributions of input parameters in terms of spectral periods and return periods. The analysis revealed that hypocentral depth distribution and ground motion prediction equations are very high sensitivity (>35%). The b-value distribution is high sensitivity (>25-35%). The magnitude distribution is medium sensitivity (>15-25%). The a-value distribution and source mechanism are low sensitivity (>5-15%). Recurrence methods are insensitive (≤5%) to hazard. •Seismic hazard calculation using FRISK88M Suite by applying rigorous logic tree.•1D equivalent linear SRA for amplification factors estimation utilizing STRATA.•Transformation to site-specific GMPEs by incorporating amplification function.•Compute the Soil PSHA using FRISK88M Suite by applying site-specific GMPEs.•Sensitivity analysis of parameters regarding spectral periods and return periods.