Comparison of national and local building inventories for damage and loss modeling of seismic and tsunami hazards: From parcel-to city-scale

In this paper, we compare the National Structure Inventory, a recently released building inventory for the United States, with a local tax assessor building inventory for use in damage and loss modeling of seismic-tsunami hazards. The city of Seaside, Oregon located in the North American Pacific Northwest and subject to seismic-tsunami hazards from the Cascadia Subduction Zone is used as a testbed. Input attributes – such as spatial footprint, year built, structure type, and number of stories – are compared at the parcel, block, block group, and tract levels. The input attributes show large differences at the parcel level and compare favorably when aggregated at increased spatial scales. The National Structure Inventory consistently underestimates structure value and number of stories compared to the tax assessor data. Expected damages across 7 mean recurrence intervals, from 100-yr to 10,000-yr, are computed using IN-CORE, an open-source community resilience model. Expected damage to the National Structure Inventory tends to be slightly larger than that of the tax assessor data and errors are reduced when aggregated at increased spatial scales. The National Structure Inventory underpredicts total economic losses and risks compared to the tax assessor data, particularly for recurrence intervals associated with high economic risks. A variance-based sensitivity analysis is performed to identify how uncertainties in the input attributes propagate to uncertainties in expected damage. Structure type, design-level, and building location all influence expected damages for seismic-tsunami hazards, highlighting the importance of accurate building inventories in multi-hazard damage and loss modeling.