Bias-adjusted analysis of global natural disaster records and an assessment of seismic hazard in Sweden

Abstract: Natural disasters pose significant challenges today and demand efficient allocation of society’s limited resources for disaster risk reduction. This relies on analysis of natural disaster records, which are prone to (reporting) biases that can affect the inferences drawn from their analysis. Data incompleteness is common in earthquake seismology, and, with a starting point in the Gutenberg-Richter law, this thesis studies the power-law behaviour between fatalities and the frequency of global natural disasters recorded by EM-DAT from 1900 to 2020. Analysing, first, the power-law behaviour of global earthquake frequency and magnitude reveals overlapping patterns, and apparent improved reporting completeness over the studied periods, implying any increase in recorded earthquakes is a reporting artefact. Similarly, “frequency-magnitude” analysis of earthquake disasters shows power-law behaviour and an almost-unchanged distribution of fatalities over time, which also implies improved reporting. Similar analysis of hydro-meteorological disasters in EM-DAT shows their substantially increased number with time to be a reporting artefact driven by improved reporting of low-fatality disasters. The increasing gradient of the power-law part of the frequency-magnitude graphs implies that the worst events have become less fatal on average.  Climate-related disaster risk reduction actions have thus been very successful, resulting in a continuous decline in aggregate associated fatalities, to the extent that earthquakes have been the most fatal natural disasters over the last few decades.  Notably, while the earthquake magnitude distribution is dominated by earthquakes in high-seismicity zones, fatalities predominantly occur in low-seismicity intraplate regions. This implies that seismic hazard in Sweden, an intraplate area, may be significant, especially as earthquakes as large as M8 have occurred in its recent deglaciation phase. Hazard analysis based on data from the recently expanded Swedish seismic network clearly unveils significant hazard posed by Post-Glacial faults in the North. This research advances our understanding of natural disaster dynamics, emphasizing the necessity for non-conventional methodologies to analyze historical data. It offers important insights to help form effective disaster risk reduction strategies, particularly in the context of seismic hazard assessment for specific regions like Sweden.