Improving Estimates of Ground-Motion Site Response in the New Madrid and Wabash Valley Seismic Zones

Ground-motion site response, i.e., the modification of ground-motion in terms of its duration, frequency content, and amplitude by the near-surface low velocity rocks and soils, is one of major research topics in engineering seismology. Site response not only affects estimates of ground motion, but also causes potential damage to susceptible infrastructure and buildings. Classic examples of significant damage caused by near-surface soft-soil-induced amplified ground motions include the lake sediments in Mexico City during the 1985 Michoacán earthquake (M 8.1) and the bay muds in the Marina District of San Francisco during the 1989 Loma Prieta earthquake (M 6.9). These phenomena have also been observed in the central and eastern United States (CEUS) such as the Ohio River valley (i.e., Maysville, Ky.) during the 1980 Sharpsburg, Kentucky M 5.2 earthquake, and the Wabash River valley during the 2008 Mt. Carmel, Illinois earthquake. Although there are many factors that contribute to site response, including incident motion, subsurface structure and material properties, ground water level, and topography, the level of input motion, shear-wave velocity profile, and soil nonlinearity are the three key parameters that are considered in earthquake engineering. For example, in current engineering practice, the site coefficient for design ground motion is determined by the level of incoming ground motion and time-weighted average of shear-wave velocity for the top 30m of soils (VS30). Specifically, the coefficient increases with decreasing average shear-wave velocity, but decreases with increasing input ground-motion level due to the soil nonlinear response. Therefore, assessment of site response is complicated and cannot be quantified adequately by a single parameter. VS30, in particular, is inadequate because it does not account for resonance effects, which can exert strong control on site response, particularly in the central and eastern U.S. due to large sediment-bedrock impedance contrasts. Further, accounting for site response is important to constrain ground motion models for seismic hazard assessment. Recent studies indicated that the site terms in ground motion models are more accurate when they make use of a frequency-dependent parameter, fpeak, which is sensitive to the response of the entire sediment thicknesses, and not just the upper 30 m. This proposed project will evaluate in detail the skill of site fundamental frequency and bedrock-sediment velocity contrast with predicting site responses at 12 USGS ANSS and other regional seismic stations in the New Madrid (NMSZ) and Wabash Valley (WVSZ) seismic zones. The bedrock-sediment velocity contrast and f0 parameters will be calculated from shear-wave velocity (Vs) profiles developed at each site using theoretical relationships and site response software (e.g., DEEPSOIL). Specifically, Vs profiles will be developed at each site from surface SH-wave reflection/refraction soundings. When needed to constrain layer thicknesses for deeper (> 200 m) layers and bedrock, P-wave soundings will be acquired and Vs will be constrained by earthquake HVSR curves. The methods proposed will be validated by observations at three sites with surface and borehole seismometers where velocity profiles to bedrock and empirical site transfer functions are available: two sites in the Mississippi Embayment (NMSZ), one with 100m and the other with 585m thick sediment; and one with 14m of Ohio River valley sediment (WVSZ). In addition, the relationships between the theoretical and empirical (from HVSR) f0 and between the first peak of the HVSR curve and the amplification at the fundamental frequency will be evaluated. The results are anticipated to reveal under what circumstances S-wave HVSR curves in the New Madrid and Wabash Valley seismic zones estimate site response, and can be useful for improving ground-motion models. This proposed research contains elements from the proposal title “Dynamic Site Characteristics at Selected EarthScope TA and FA, KSSMN, and ANSS Stations in the New Madrid and Wabash Valley Seismic Zones” submitted for FY 2019 Funding Opportunity G18AS00021, which was recommended for funding, but significant modifications have been made, including focusing on testing new site-effect prediction parameters. This proposal will improve our understanding of site-effect in the New Madrid and Wabash Valley Seismic Zones by using instrumental recordings of key regional seismic stations. This proposed project is in the research priorities for the CEUS Research Area in the Fiscal Year 2020 NEHRP Program Announcement, and will directly contribute to regional earthquake hazards assessments (Element I).