| Summary: | The European Seismic Hazard Model 2013 has been updated to the 2020 version ESHM20. There are systematic differences for Fennoscandia between the two versions, where western Norway, southwestern Sweden and Denmark see a decrease in hazard while the hazard has increased in most of Sweden and Finland. In this report, the origin and significance of these differences are investigated, with a special emphasis on the sites of nuclear power plants in Sweden and Finland. The European seismic hazard model 2020 (ESHM20) supersedes the 2013 version (ESHM13) while following the same principle of state-of-the art procedures homogeneously applied for the entire pan-European region, without country-borders issues. ESHM20 includes updated datasets (earthquake catalogues, active faults, ground shaking recordings), information (tectonic and geological) and models (seismogenic sources, ground motion). For Fennoscandia, the earthquake dataset increased by eight events, but the seismic source zones were updated in the north, there was a complete change of ground motions models and the logic tree was significantly expanded. We uncovered a mistake in the update of recurrence parameters in one source zone in ESHM20 which likely has led to an overestimate of the hazard at Ringhals. We find that the hazard, in terms of mean peak-ground-acceleration (PGA), has increased from ESHM13 to ESHM20 at the nuclear power plants (NPPs) in Olkiluoto, Forsmark, Oskarshamn and for long return periods in Loviisa. It has decreased at Ringhals and for short return periods at Loviisa. In addition, the standard deviations of the PGA distributions have increased considerably at all locations except at Ringhals, where the increase is more modest. The differences between ESHM13 and ESHM20 are likely due mostly to the complete update of ground motion models, the significantly expanded logic tree and improved methodologies and algorithms. Assessing the significance of the differences between two hazard models is a long-standing problem in seismic hazard. We ...
|
|---|