Rapid Estimation of Direct Economic Losses Caused by Significant Earthquakes: An Evidence-Based Model and its Applications

抄録

This paper presents a rapid or real-time estimation method of the economic value of direct stock damages caused by significant earthquakes in Japan. The result will contribute to both the government and private sectors’ early decision-making, particularly for provisional budget allocation. First, we developed a simple but evidence-based model for estimating stock losses explained by a representative earthquake hazard factor and an exposure factor, i.e., seismic intensity and existing stock of physical assets. The key characteristic of our estimation model is that the dependent variable is prefectural damage amount. Still, the explanatory variables come from municipal sources: we overcome this data availability problem through our estimation process. Second, we carefully checked the model’s specification, estimation, and performance to be soundly applied to a real-time assessment of future earthquake events. We also explain the automated measuring of the prefectural direct loss value and its distribution to every 250 m mesh. Finally, we show two examples of the application of our model; one is the case of the 2018 Northern Osaka Earthquake, and the other is the anticipated Tokyo inland earthquake.