Uncertainty in PSHA related to the parametrization of historical intensity data

M Mucciarelli

doi:10.5194/nhess-14-2761-2014

Abstract

Abstract. In recent times a great deal of research aimed to reduce of uncertainties in probabilistic seismic hazard analysis (PSHA). Most attention was paid to the role of ground motion prediction equations (GMPEs; see, e.g., Strasser et al., 2009), while no studies were devoted to a possible larger source of uncertainties: the historical catalogues of earthquakes. In areas where historical catalogues provide a record many centuries long and surface geology does not permit us to obtain complete catalogues of seismogenic faults at the moment, their use is unavoidable for estimating seismicity rates required for PSHA. Their use is also gaining popularity as an independent tool for the estimation of PSHA (D'Amico and Albarello, 2008) and for validation purposes (Stirling and Petersen, 2006; Mucciarelli et al., 2008). After proposing an alternative parametrization of historical macroseismic intensity, this paper is devoted to discussing what the real impact of starting uncertainties in intensities on the final uncertainties in PSHA is.

Highlights

In recent times a great deal of research aimed to reduce of uncertainties in probabilistic seismic hazard analysis (PSHA)
For each known historical earthquake, the first step of research is to assign a macroseismic intensity to any point of the area affected by what is recognized to be a single event
It is common practice to assign values such as VII–VIII, which are subsequently treated as 7.5 in further calculations. This is not the correct way to proceed, since intensity scales do not provide for the use of half-degrees

Summary

State of the art and a new proposal

For each known historical earthquake, the first step of research is to assign a macroseismic intensity to any point of the area affected by what is recognized to be a single event. It is common practice to assign values such as VII–VIII, which are subsequently treated as 7.5 in further calculations This is not the correct way to proceed, since intensity scales do not provide for the use of half-degrees. The vectorial intensity was subsequently proposed for site hazard estimates by Albarello and Mucciarelli (2002) This is commonly done when using probabilistic attenuation relationships but could be a great opportunity for the revision of catalogues, allowing the historical seismologist to express his feeling that an event is “almost certainly” of an intensity VIII (e.g. 0.9 probability) while it cannot be completely ruled out that the effects

Absolutely false

The effect of intensity uncertainty on PSHA

Conclusions