Over the past decades, the overall risk profile of the Loviisa NPP has been steadily reduced through continuous improvements, including both plant modifications and reductions of excessive conservatism in PRA models. As the internal contributions have declined, the relative importance of external hazards such as earthquakes has increased, despite Finland’s low-seismicity environment. Recent sensitivity analyses have highlighted several components of probabilistic seismic hazard analysis (PSHA) that carry substantial uncertainty, largely due to the limited seismic data available for the region.

This paper applies a tailored approach for defining a maximum‑magnitude distribution for Finland and its surrounding areas as part of the national PSHA development. The method builds on the commonly used Bayesian framework by constructing a prior distribution from analogue domains that are geologically similar to Fennoscandia, using an updated stable continental region (SCR) earthquake catalogue.

The resulting distribution places significantly more probability mass on lower maximum‑magnitude values than more globally oriented models, reflecting the view that Fennoscandia likely exhibits lower seismic potential than the SCR average. Nevertheless, the overall seismic hazard at Loviisa decreases only modestly, indicating that the hazard is dominated by low‑to‑moderate magnitude events rather than the upper tail of the magnitude distribution.