Investigating high‐resolution spatial variability of site response in the Los Angeles Basin using a dense sensor network

Abstract

We characterize seismic site response across a densely instrumented region of the Los Angeles Basin by combining data from the Community Seismic Network (CSN) and regional networks in the Advanced National Seismic System (ANSS). Mixed-effects regression is used to partition ground-motion model residuals and compute pseudo-spectral acceleration site residuals (η s ) from the NGA-West2 ground motion database and a CSN dataset. For short periods ( T < 0.3 s), CSN and ANSS site residuals agree closely, while observed differences at longer periods are likely related to signal-to-noise limitations of the CSN sensors rather than sensor accuracy. Correlations between CSN and ANSS site residuals are high, and the dense CSN network reveals localized amplification not resolved by the ANSS network. Along the northeastern basin margin, PSA at 1 second reaches amplitudes up to approximately twice those predicted for reference rock conditions, consistent with basin-edge amplification effects. Systematic variations in site response are observed across major fault zones, particularly between [BSAS1.1]areas north of the Lower Elysian Park fault and between [BSAS2.1]the Lower Elysian Park and Puente Hills faults. These findings demonstrate the value of dense seismic networks, such as the CSN, for resolving localized amplification and improving seismic hazard assessment in urban sedimentary basins.