Can we develop a more targeted approach to mitigating seismic risk?

The devastating impact of recent major earthquakes underscores the urgent need for effective seismic risk mitigation. The high death tolls from events like the 2015 Nepal earthquake, the 2011 Japan earthquake, the 2010 Haiti earthquake, and the 2008 Wenchuan earthquake, along with the 2023 Turkey-Syria earthquake, highlight the continued threat posed by earthquakes. Rapid urbanization in seismically active areas has exacerbated the problem, with increased population density and wealth concentration leading to higher physical vulnerability. The construction of new buildings and infrastructure often prioritizes speed over quality and safety, further amplifying this risk. Achieving the sustainable development goals and disaster risk reduction targets in the Sendai Framework necessitates a more proactive and targeted approach to seismic risk mitigation. Probabilistic seismic hazard maps, widely used since the 1960s, help regulate seismic codes for new buildings but often neglect the systematic improvement of existing building resilience. Cost-benefit analyses demonstrate that pre-disaster mitigation, including optimized disaster risk reduction strategies, building retrofitting, and public education, is far more cost-effective than post-earthquake emergency response. However, limited resources necessitate prioritization. This paper focuses on identifying "priority zones" of highest potential seismic damage to maximize the efficiency of limited mitigation budgets.

Existing literature highlights the challenges in systematically improving seismic resilience of existing buildings. While probabilistic seismic hazard maps are widely used, there's a lack of in-depth discussion on systematically enhancing the seismic resilience of existing structures. Studies emphasize the cost-effectiveness of pre-disaster mitigation compared to post-earthquake responses. However, limited budgetary resources constrain the implementation of these mitigation measures. Therefore, the pre-identification of high-risk zones is crucial for effective resource allocation. The literature also demonstrates the significant impact of uncertainties in seismic hazard estimations on overall risk assessment. Previous work, focusing on areas with relatively low seismic hazard, has shown the dominant contribution of hazard uncertainty to total risk uncertainty. In contrast, for high-hazard areas, the complexity of large earthquake rupture patterns amplifies the impact of ground motion variation on risk estimates. Studies suggest that the uncertainty in building response is more sensitive to ground shaking variation than to changes in structural modeling. These findings underscore the importance of accurately characterizing seismic hazards for reliable risk assessment. Existing ground motion prediction methods, while widely used, have limitations such as the ergodic assumption, insufficient consideration of spatial correlations, and limitations in near-source regions due to data scarcity. Physics-based simulation (PBS) methods offer an alternative approach that can address some of these limitations, though their reliability is dependent on robust model inputs. The integration of PBS methods with empirical approaches is an area of ongoing research.

This study proposes a strategy to delineate regional high seismic risk zones by comparing seismic risk calculated from two distinct ground shaking prediction methods: empirical Ground Motion Prediction Equations (GMPEs) and physics-based simulation (PBS). The strategy assumes that exposure and vulnerability uncertainties are reducible with more data, thus the focus is on the impact of seismic hazard variations. The approach is demonstrated using seismic hazard and loss estimation results for Mw 7.5 earthquake scenarios on the Jiaocheng fault in the Shanxi Rift System, China. This fault is selected due to its high seismic activity and proximity to the densely populated city of Taiyuan. The methodology involves several steps: 1. **Ground shaking simulation:** Peak Ground Acceleration (PGA) maps are generated for four Mw 7.5 scenarios using both PBS (from Zhang et al. 2017) and four empirical GMPEs (BA08, CB08, BSSA14, CB14). The consistency of these maps was previously explored. 2. **Site effect rectification:** The PGA maps are rectified to account for local site amplification effects, using site classes from Li et al. (2020). Due to limited data, site classes II is used where no information is available. 3. **Exposure modeling:** A high-resolution residential building stock model is used, modifying the unit construction prices from Xin et al. (2021) based on data from the Department of Housing and Urban-Rural Development of Shanxi Province. The model considers building type, story class, and construction year to estimate replacement values. 4. **Vulnerability assessment:** Vulnerability curves are developed for representative building types and seismic design code levels from the second, third and fourth probabilistic seismic hazard maps (1977, 1990, 2001) to reflect the varying seismic code levels of the buildings. 5. **Seismic risk modeling:** Seismic risk models are developed by combining the rectified PGA maps, the exposure model, and the vulnerability curves, creating seismic loss distribution maps. 6. **Model validation:** The reliability of the risk models is evaluated by comparing modeled seismic losses with those estimated from an empirical loss model specifically regressed from historical earthquake damage information. 7. **High-risk zone delineation:** Finally, the validated seismic risk models are used to delineate the regional high seismic risk zone.

The analysis reveals that the delineated high-risk zone accounts for approximately 7% of the regional land area but is responsible for approximately 85% of the total financial loss in the earthquake scenarios considered. The comparison between PBS-based and GMPE-based seismic loss models reveals a high degree of consistency, validating the robustness of the proposed methodology. Seismic loss distributions generated from both methods show a high spatial overlap in high-loss areas, further strengthening the reliability of the high-risk zone delineation. Intensity-based analysis shows that the loss ratio (loss/exposure) increases significantly with intensity. The study also examines the impact of building type on seismic losses, highlighting the disproportionate impact on less resilient structures. A comparison of losses estimated in the current study with previous studies is made using a set of intensity values and shows a reasonable agreement between these losses. By combining the PBS-based loss maps and the GMPE-based loss maps for the high loss grids, the top 10% of loss grids represent 88.4% -93.7% of the total loss, demonstrating a high concentration of risk in a small area. The top 10% of the loss grids concentrate in 30 counties/districts, with Taiyuan’s Qingxu County having the most high risk grids (240). The study finds that if all buildings were retrofitted to high code levels, the total financial loss could be significantly reduced.

The findings highlight the effectiveness of a targeted approach to seismic risk mitigation. The significant concentration of risk within a relatively small area (7% of land area accounting for 85% of financial loss) underscores the potential for maximizing the impact of limited resources by prioritizing mitigation efforts in high-risk zones. The consistency between the PBS and GMPE-based loss estimates enhances confidence in the delineated high-risk zone. The proposed strategy can be applied to other densely populated cities in seismically active regions, providing valuable information for resource allocation and mitigation planning. The identification of specific high-risk areas aids decision-makers in deploying emergency response resources and prioritizing pre-disaster mitigation actions. This targeted strategy can contribute to enhanced preparedness and resilience against future earthquakes, reducing potential fatalities and financial losses.

This study proposes a robust strategy for delineating regional high seismic risk zones, demonstrated using a case study of the Jiaocheng fault in China. The results highlight the significant concentration of seismic risk within a small geographical area, emphasizing the need for targeted mitigation strategies. The consistency between physics-based and empirical model results adds confidence to the approach. Future research should focus on expanding the frequency range of simulated ground motions, enriching earthquake scenarios (considering wider magnitude ranges and nucleation positions), incorporating non-residential buildings and infrastructure, and exploring the sensitivity of the results to various input data changes. Access to detailed building-level data would further refine the exposure model and enable more precise risk zone delineation.

The study's scope is limited to the Jiaocheng fault and considers only Mw 7.5 earthquake scenarios. The exposure model is limited to residential buildings, neglecting other vulnerable assets. The site effect rectification is simplified due to the lack of comprehensive Vs30 data for the entire study area. While the comparison with existing loss estimations provides validation, this validation is based on the available historical data. The study does not directly model casualties. Further research is needed to address these limitations and expand the scope of the analysis.