Articles | Volume 18, issue 2
https://doi.org/10.5194/nhess-18-499-2018
https://doi.org/10.5194/nhess-18-499-2018
Research article
 | 
15 Feb 2018
Research article |  | 15 Feb 2018

The effect of alternative seismotectonic models on PSHA results – a sensitivity study for two sites in Israel

Matan Avital, Ronnie Kamai, Michael Davis, and Ory Dor

Abstract. We present a full probabilistic seismic hazard analysis (PSHA) sensitivity analysis for two sites in southern Israel – one in the near field of a major fault system and one farther away. The PSHA analysis is conducted for alternative source representations, using alternative model parameters for the main seismic sources, such as slip rate and Mmax, among others. The analysis also considers the effect of the ground motion prediction equation (GMPE) on the hazard results. In this way, the two types of epistemic uncertainty – modelling uncertainty and parametric uncertainty – are treated and addressed. We quantify the uncertainty propagation by testing its influence on the final calculated hazard, such that the controlling knowledge gaps are identified and can be treated in future studies. We find that current practice in Israel, as represented by the current version of the building code, grossly underestimates the hazard, by approximately 40 % in short return periods (e.g. 10 % in 50 years) and by as much as 150 % in long return periods (e.g. 10E−5). The analysis shows that this underestimation is most probably due to a combination of factors, including source definitions as well as the GMPE used for analysis.

Download
Short summary
We perform a hazard sensitivity study for two sites in Israel, exploring effects of uncertainty on the calculated seismic hazard. We account for uncertainty in the earthquake source properties, such as their geometric representation and seismic activity. We also account for uncertainty in the wave propagation path, by using alternative models to describe the ground motion calculations. We conclude that the current practice in Israel should be updated as it is probably underestimating the hazard.
Altmetrics
Final-revised paper
Preprint