Geospatial Site Amplification Model with Geotechnical Adjustments for the Basin and Range Province of the Western United States

ABSTRACT Ground-motion model (GMM) site terms developed from continuously available geospatial data and enhanced with site-specific measured geotechnical data have been demonstrated to be an effective approach for site term development in California (Roberts et al., 2025; Roberts et al., 2026). This study extends this methodology to the Basin and Range physiographic province in the western United States. Ground-motion data from 16,852 recordings at 433 stations in the Next Generation Attenuation-West3 database (Buckreis and Stewart, 2025) are used for model development. A base site term model is developed using mapped geospatial variables (e.g., sediment thickness, elevation, and surficial geologic units) to capture trends in soil stiffness and basin effects. The target site amplifications of peak ground acceleration, peak ground velocity, and pseudospectral accelerations from 0.01 to 10 s are decomposed from the residuals of the Boore et al. (2014) (BSSA14) GMM. A linear mixed-effects regression model is then developed to predict each target site amplification using geospatial variables. The resulting model is a linear geospatial site term that provides a consistent site term for all locations in the region. The geospatial site term shows a substantial reduction in site-to-site variability; on average, an 8.5% reduction is achieved compared with BSSA14. This base geospatial model is then enhanced with local geotechnical information where available. The proposed geotechnical site term adjustment models are developed using microtremor horizontal-to-vertical spectral ratio data (Anbazhagan et al., 2025) and measured VS30 data (Buckreis and Stewart, 2025). Additional reductions in the site-to-site variability are achieved when the geotechnical adjustments are applied. This study illustrates that the approach of incorporating broadly available geospatial data before site-specific geotechnical data is effective in regions outside of California and demonstrates how geospatial site amplification models with explicit uncertainty characterization can be developed where ground-motion data are sparse.