The time-averaged shear wave velocity in the upper 30 m (VS30) is commonly used as explanatory variable for site characterization. Although measuring shear wave velocity (VS) to 30 m or greater depth is the most direct and robust way to compute VS30, oftentimes VS profiles are shallower than 30 m. For such cases, various methods have been proposed to extrapolate the time-averaged VS to a specified depth (VSZ) and/or VS at a specified depth to VS30. These VSZ-to-VS30 extrapolation methods typically provide a greater predictive power than methods based on geotechnical investigation (e.g., correlations with penetration resistance) and on proxies (e.g., surface geology, terrain categories, or topographic slope). In this study, we investigate the functional forms used in five VSZ-to-VS30 extrapolation models: (1) Boore (2004), (2) Boore et al. (2011), (3) Midorikawa and Nogi (2015), (4) Dai et al. (2013), and (5) Wang and Wang (2015). We then validate model performances using five regional VS profile datasets: (1) Japan - 289 profiles from the KiK-Net and PARI arrays; (2) California - 71 profiles from Caltrans reports; (3) Oregon and Washington - 450 profiles from reports of state and federal geologic surveys (Oregon DOGAMI, Washington DNR-GER, USGS, and CGS) and various studies performed by university research groups; (4) central and eastern North America - 200 profiles from the reports of Nuclear Regulatory Commission and university research group studies; (5) Beijing plain, China - 463 profiles from unpublished technical and research reports. For each selected model, we perform regression analyses developing model coefficients for those five datasets and then investigate potential regional differences. Midorikawa and Nogi (2015) and Dai et al. (2013) models provide the lowest model bias and dispersion relative to measured VS30 values. We also developed non-region-specific models which provide comparable model bias and dispersion to the region-specific models. The developed models are applicable to any VS profile for depths < 30 m.
Performance evaluation of VSZ-to-VS30 correlation methods using global VS profile database
Zimmaro P.;
2017-01-01
Abstract
The time-averaged shear wave velocity in the upper 30 m (VS30) is commonly used as explanatory variable for site characterization. Although measuring shear wave velocity (VS) to 30 m or greater depth is the most direct and robust way to compute VS30, oftentimes VS profiles are shallower than 30 m. For such cases, various methods have been proposed to extrapolate the time-averaged VS to a specified depth (VSZ) and/or VS at a specified depth to VS30. These VSZ-to-VS30 extrapolation methods typically provide a greater predictive power than methods based on geotechnical investigation (e.g., correlations with penetration resistance) and on proxies (e.g., surface geology, terrain categories, or topographic slope). In this study, we investigate the functional forms used in five VSZ-to-VS30 extrapolation models: (1) Boore (2004), (2) Boore et al. (2011), (3) Midorikawa and Nogi (2015), (4) Dai et al. (2013), and (5) Wang and Wang (2015). We then validate model performances using five regional VS profile datasets: (1) Japan - 289 profiles from the KiK-Net and PARI arrays; (2) California - 71 profiles from Caltrans reports; (3) Oregon and Washington - 450 profiles from reports of state and federal geologic surveys (Oregon DOGAMI, Washington DNR-GER, USGS, and CGS) and various studies performed by university research groups; (4) central and eastern North America - 200 profiles from the reports of Nuclear Regulatory Commission and university research group studies; (5) Beijing plain, China - 463 profiles from unpublished technical and research reports. For each selected model, we perform regression analyses developing model coefficients for those five datasets and then investigate potential regional differences. Midorikawa and Nogi (2015) and Dai et al. (2013) models provide the lowest model bias and dispersion relative to measured VS30 values. We also developed non-region-specific models which provide comparable model bias and dispersion to the region-specific models. The developed models are applicable to any VS profile for depths < 30 m.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
