Impacts of Cascadia Subduction Zone M9 Earthquakes on Bridges in Washington State: SDOF Idealized Bridges

Abstract

The Cascadia Subduction Zone has the potential to generate Magnitude 9 earthquakes that couldhave large impacts on bridges throughout western Washington State. The likelihoods of damage to bridges during an M9 event at various locations in Western Washington were estimated by the following process: 1. Ten locations in Washington State were selected to reflect a wide variety of fault distances and sedimentary basin depths. At each location, the ground motions for an M9 event were simulated for 30 earthquake scenarios by a United States Geological Survey (USGS) and University of Washington (UW) team, with the support of the National Science Foundation (Frankel et al. 2018). The baseline motions were modified to account for four sets of 30 site profiles, corresponding to four subcategories of the NEHRP site classes. 2. WSDOT, UW, and Washington State University engineers compiled a detailed database of key properties of 609 WSDOT bridges along key lifelines in the Puget Sound region. About three quarters of these bridges were constructed before 1976. 4Nearly 70% of the bridges documented in this database were supported by at least one reinforced concrete column at intermediate supports. 3. The bridges with reinforced concrete columns were idealized as single-degree-of-freedom systems in which the effective stiffness and lateral strength were derived from the column properties, and the resistance provided by the abutments wasneglected. The cyclic force-deformation responses of the columns were modeled with an Ibarra-Medina-Krawinkler deterioration model that was calibrated using force-deformation histories from the UW-PEER Column Performance Database (UW-PEER 2020).4. The deformation demands for pre-1976 and 1976-present bridges were estimated for 30 M9 scenarios, the 10 locations, N-S and E-W directions, 120 site profiles (sorted into four site categories) and 18 periods. Both the response of systems with mean strengths and reduced strengths were simulated. The researchers found that the lateral strengths of most of the lifeline bridges exceeded the strengths expected from design procedures, even after accounting for material overstrength and strain hardening. As a result, many of the bridges, even older bridges designed before the mid-1970s, performed better than expected. In particular:  Bridges located along the west coast of the Olympic Peninsula (e.g., Ocean Shores,Forks) had median displacement ductility demands in the range of 2 to 4. Collapse probabilities above 15-20% were only predicted for older pre-1976 bridges with periods below 0.3s. Averaged over many events, the likelihood of column concrete spalling range from 40-60% for a wide range of periods, and the likelihood of buckling for most periods was in the range of 5% to 25%. 5 The likely performance of bridges located on sedimentary basins in the Puget Sound region (e.g., Seattle, Tacoma, Everett) depended strongly on the bridge’s effective period. o Bridges with periods below 0.5s had median displacement ductilities below or near 2. o Between periods of 0.5s and 3s, the displacement ductilities were larger than at lower or higher periods, and they reached values of 5 for older bridges located on soil profiles corresponding to the softer end of Site Class D (Subclass D3). o New bridges had negligible likelihoods of collapse, and the likelihood of bridge collapse for older bridges only exceeded 20% for soil subclass D3.  Bridges located at similar distances from the coast as Seattle, but located outside of sedimentary basins (e.g., the Seattle basin), had median displacement demands below 2, and negligible collapse probabilities and likelihoods of bar buckling. Some spalling might be experienced in bridges located on softer soils. From the ductility demand data, fragility curves were constructed using both Sa and Sa,eff for various levels of ductility demand. The fragility curves produced using Sa showed significant regional variation, whereas those produced using Sa,eff were highly consistent across the 10 locations. In addition, the fragility curves produced using Sa,eff were much steeper and showed less scatter than those produced using spectral acceleration alone. The results of these analyses come with several caveats. Namely, the analyses conducted as part of this study:  Neglected the resistance provided by abutments 6 Did not consider the behavior of very soft soils (e.g., site classes E and F). The interaction between the site profiles with long periods might exacerbate the amplification of long-period components of motions observed in sedimentary basins.  Neglected the likelihood of span unseating, under the assumption that previous WSDOT retrofit efforts have precluded this failure mode  Neglected the likely correlation between ductility demand and performance levels for a given M9 event. In other words, some events might lead to much larger (or smaller) demands and damage that others, leading to much larger (or smaller) damage levels for a particular event.  Most importantly, shear or foundation failure were not considered. Both of these failure modes might be critical in older bridges.