A spatial panel approach to the statistical assessment of seismic impacts and building damages: Case study of Taichung, Taiwan

Abstract A set of seismic-related statistical models is developed, using pseudo-data generated by an earthquake-engineering simulation model. The Peak Ground Acceleration (PGA) is the principal measure of seismic hazard, the Peak Ground Displacement (PGD) represents secondary impacts, and land-use patterns characterize urban vulnerability. A PGA model based on earthquake magnitude, epicenter-to-site distances, and source depth is formulated as a spatial lag panel (SLP) model to account for PGA spatial interactions (neighborhood effects). A PGD spatial lag panel model is also specified to account for soil liquefaction effects. Finally, a model of seismic damages to buildings is formulated, relating monetary damages (loss in property values) to seismic hazards (PGA and PGD) and urban vulnerabilities (residential, commercial, industrial, and facility building stocks). Pseudo-data are generated under 22 simulated historical seismic events for the city of Taichung, Taiwan. These data are then used for model estimation. By combining the three models, monetary damages can be estimated as a function of land-use patterns, PGA, PGD, their neighborhood effects, and other seismic characteristics. These models can therefore provide seismic information for policy making and help in allocating land to new activities while minimizing potential seismic damages.