5:15 PM - 7:15 PM
[SSS14-P05] Study on the fault model parameters for strong ground motion prediction: An example of the 2024 Noto Peninsula Earthquake
Keywords:Noto Peninsula Earthquake, Fault model
1. Introduction
On January 1, 2024, the Noto Earthquake (Mw7.6) occurred due to offshore active faults near the northern coast of the Noto Peninsula. In the Sea of Japan, several organizations and research projects have constructed databases of offshore active faults and developed source fault models for tsunami prediction. However, the fault models assumed in those projects were consequently inadequate for predicting the strong ground motion observed during the Noto Earthquake. This study revisits and reevaluate fault models for the Noto Peninsula, as part of an effort to improve the construction of predictive models large, shallow crustal earthquakes including near-coast active faults.
2.Fault model for the 2024 Noto earthquake
To understand the characteristics of the Noto earthquake, we set a reproducible model from previously mapped surface fault traces and ~150 km long aftershock distribution. The Noto earthquake primarily occurred along the preexisting fault traces. However, the fault at the northeastern end of the aftershock zone dips to the north, in contrast to the main fault segments, which dip southeast. Additionally, although the distribution at the western end of the aftershock zone follows a N-S trend, no fault trace has been mapped in this area with the same orientation. Despite incorporating this post-earthquake information and optimizing the fault to account for the maximum values of relevant parameters, we found it impossible to fully reproduce the prolonged shaking observed during the Noto earthquake. To address this issue, a previous study by Asano and Iwata (2024: SSJ) proposed a model with two hypocenters. By applying this approach to our study, we have come closer to capturing the characteristics of prolonged shaking process. The rupture of the Noto earthquake did not propagate concentrically from a single hypocenter. Instead, it propagated separately in both eastward and westward directions from two hypocenters that occurred at different times.
3. Predictive model and probability of fault linkage
In our study, we developed two models sequentially to assess whether an earthquake similar to the Noto earthquake could have been predicted using only pre-event information. Stage 1 utilized only information available before the Noto earthquake, while Stage 2 incorporated insights gained from the event. To evaluate the likelihood of multi-segment fault rupture based on fault geometry and alignment, we estimated rupture propagation probabilities using the method of Biasi and Wesnousky (2017). Although the method by Biasi and Wesnousky (2017) is originally designed to evaluate strike-slip fault interaction, we extended its application to reverse faults following by the approach of Walsh et al. (2023).
4.Results and Speculation
Considering the uncertainties in observational data, it appears possible to make reasonable assumptions about the dip angle and depth of the fault bottom. The magnitude of the Noto earthquake could also have estimated by considering the various possibilities taking into account of multi-segment rupture probabilities of the offshore active faults in this region. On the other hand, we confirmed it is challenging to assume the N-S strike of the western end of the 2024 fault zone and the rupture propagation from two hypocenters with time delay. The probability of rupture propagation to the faults on the west side of the Noto Peninsula is higher for N-S striking faults that ruptured during the Noto earthquake than for E-W trending faults that were previously mapped in this area, even though the magnitudes and the intensities of strong ground motion in the northern peninsula are roughly identical. In contrast, the ground motion along the west coast of the peninsula is significantly underestimated when the pre-event E-W striking fault model is simply applied. Further comparisons are required to learn from the Noto earthquake and incorporate these insights into the ground motion prediction framework.
On January 1, 2024, the Noto Earthquake (Mw7.6) occurred due to offshore active faults near the northern coast of the Noto Peninsula. In the Sea of Japan, several organizations and research projects have constructed databases of offshore active faults and developed source fault models for tsunami prediction. However, the fault models assumed in those projects were consequently inadequate for predicting the strong ground motion observed during the Noto Earthquake. This study revisits and reevaluate fault models for the Noto Peninsula, as part of an effort to improve the construction of predictive models large, shallow crustal earthquakes including near-coast active faults.
2.Fault model for the 2024 Noto earthquake
To understand the characteristics of the Noto earthquake, we set a reproducible model from previously mapped surface fault traces and ~150 km long aftershock distribution. The Noto earthquake primarily occurred along the preexisting fault traces. However, the fault at the northeastern end of the aftershock zone dips to the north, in contrast to the main fault segments, which dip southeast. Additionally, although the distribution at the western end of the aftershock zone follows a N-S trend, no fault trace has been mapped in this area with the same orientation. Despite incorporating this post-earthquake information and optimizing the fault to account for the maximum values of relevant parameters, we found it impossible to fully reproduce the prolonged shaking observed during the Noto earthquake. To address this issue, a previous study by Asano and Iwata (2024: SSJ) proposed a model with two hypocenters. By applying this approach to our study, we have come closer to capturing the characteristics of prolonged shaking process. The rupture of the Noto earthquake did not propagate concentrically from a single hypocenter. Instead, it propagated separately in both eastward and westward directions from two hypocenters that occurred at different times.
3. Predictive model and probability of fault linkage
In our study, we developed two models sequentially to assess whether an earthquake similar to the Noto earthquake could have been predicted using only pre-event information. Stage 1 utilized only information available before the Noto earthquake, while Stage 2 incorporated insights gained from the event. To evaluate the likelihood of multi-segment fault rupture based on fault geometry and alignment, we estimated rupture propagation probabilities using the method of Biasi and Wesnousky (2017). Although the method by Biasi and Wesnousky (2017) is originally designed to evaluate strike-slip fault interaction, we extended its application to reverse faults following by the approach of Walsh et al. (2023).
4.Results and Speculation
Considering the uncertainties in observational data, it appears possible to make reasonable assumptions about the dip angle and depth of the fault bottom. The magnitude of the Noto earthquake could also have estimated by considering the various possibilities taking into account of multi-segment rupture probabilities of the offshore active faults in this region. On the other hand, we confirmed it is challenging to assume the N-S strike of the western end of the 2024 fault zone and the rupture propagation from two hypocenters with time delay. The probability of rupture propagation to the faults on the west side of the Noto Peninsula is higher for N-S striking faults that ruptured during the Noto earthquake than for E-W trending faults that were previously mapped in this area, even though the magnitudes and the intensities of strong ground motion in the northern peninsula are roughly identical. In contrast, the ground motion along the west coast of the peninsula is significantly underestimated when the pre-event E-W striking fault model is simply applied. Further comparisons are required to learn from the Noto earthquake and incorporate these insights into the ground motion prediction framework.