Comparing seismic hazard maps and real seismicity: objective assessment of different models

Comparing different seismic hazard maps over the Italian territory A reliable and comprehensive characterization of expected seismic ground shaking, eventually including the related time information, is essential in order to develop effective mitigation strategies and increase earthquake preparedness. However, it is important to bear in mind that we know little about how earthquake hazard maps describe the shaking that will actually occur in the future and have no agreed way of assessing how well a map performed in the past, and, thus, whether one map performs better than another. Moreover, we should not forget that different maps can be useful for different end users, who may have different cost-and-benefit strategies; therefore, the performance of a map may depend on the particular application. Quantitative assessment of maps performances is a key issue towards an increasingly reliable seismic hazard assessment (SHA), and it must become an essential step in the scientific process of their revision and possible improvement (Kossobokov et al., 2015). Objective testing against available observations, however, is not an easy task and it implies a careful application of statistics to data sets of limited size and different accuracy. Cross-checking of probabilistic models with observations from past earthquakes and independent physics-based models is also recognized as a major verification procedure (e.g. Zuccolo et al., 2011; Nekrasova et al., 2014, 2015). Maps can be made using a wide variety of assumptions, which lead to different predicted shaking; comparing their performance according to selected performance measures, provides valuable insight into key map parameters and assumptions. The existing maps from the classical probabilistic seismic hazard analysis (PSHA), as well as those from the neo-deterministic analysis (NDSHA), which are available for the Italian territory and for several countries worldwide (Panza et al. 2022), are considered to exemplify the possibilities of the cross-comparative analysis in spotting out limits and advantages of different methods. Neo-deterministic refers to a scenario-based approach, proposed by Panza et al. (2001), which allows for the consideration of a wide range of possible earthquake sources to construct ground motion bedrock scenarios via realistic full waveforms modeling at national scale. Panza et al. (2012) and references therein published a new variant of these maps obtained as a result of a comprehensive set of parametric tests.[...]