Characteristic earthquake recurrence and time-dependent seismic hazard modeling for the North and East Anatolian Faults
摘要
Until now, many probabilistic seismic hazard analyses (PSHA) have been conducted for Turkey. However, except in a limited number of cases, characteristic fault source modeling was not used. Since the North and East Anatolian Faults (NAF and EAF) show some evidence of characteristic rupture behavior, the first objective of this research is to develop a sound hybrid characteristic recurrence model for the NAF and EAF. The so-called hybrid model involves a composite characteristic model (i.e., an exponential part for the smaller and a characteristic part for larger magnitudes) developed for each segment combined with a characteristic recurrence proposed for multi-segment ruptures. Two different hybrid earthquake recurrence models are developed with time-independent (or Poissonian) and time-dependent (or renewal) characteristics. The North Anatolian Fault has been known for its capacity to generate multi-segment ruptures, evidenced in the twentieth century by the 1939 Erzincan and the 1999 Kocaeli earthquakes. The EAF, on the other hand, had not produced any multi-segment rupture in the recent past. The occurrence of the February 6, 2023, Kahramanmaraş-Pazarcık earthquake (Mw 7.8, rupture length of 350 km) shed new light on the earthquake hazard-related analyses of this fault zone. In the present study which also allows for multi-segment ruptures on both faults, the effects of characteristic modeling, as well as those of seismic gaps along the NAF and EAF on seismic hazard are assessed, including a pre- and post-2023 modeling comparison for the EAF. The model also allows for a comparison of the results with the fully exponential model of the NAF and EAF, which, generally, produced higher seismic hazard compared to the Poissonian hybrid model. As shown by the occurrence of the 2023 Kahramanmaraş earthquake, slip deficits on some fault segments can increase the seismic hazard dramatically. A renewal seismic hazard model would be more suitable to handle such cases. In any case, the 2023 Kahramanmaraş earthquake has also shown that modeling for the really (even inconceivably) big one is never an overestimation, especially for the major plate-boundary faults.