Site Response of Southern California Sedimentary Basins and Other Geomorphic Provinces

Chukwuebuka Nweke, Jonathan P. Stewart, & Scott J. Brandenberg

Published October 25, 2020, SCEC Contribution #10965

Seismic site response can be influenced by a variety of physical mechanisms that include amplification due to resonance, nonlinearity, topographic effects, impedance contrasts, and contributions from two- or three-dimensional wave propagation in sedimentary basins. Current ground motion models use ergodic procedures that average these effects over many sites globally by conditioning on the time-averaged shear wave velocity in the upper 30 m (VS30), and in some cases, on the depth to a shear wave velocity isosurface (zx) that is also known as the basin depth parameter.

Current site amplification models conditioned on VS30 reflect, in an average sense, most of the aforementioned physical mechanisms, including basin effects. The site response contributions from basin effects are associated with a differential depth parameter taken as the difference between depth for a particular site and average basin depth conditioned on VS30. The basin amplification models are “centered”, in the sense that they predict changes in ground motion amplification for non-zero differential depths. The changes in ground motion amplification are positive and negative for sites with positive and negative differential depths, respectively. The models predicting this behavior are derived using data from both northern and southern California, and for sites situated within sedimentary basin structures but also other geomorphic provinces (e.g., sedimentary structures of different scales and sites with shallow soil overlying rock).

We investigate the benefits of regionalizing basin response in ergodic ground motion models. Using southern California data we consider the following questions: (1) how should basin and non-basin locations be classified?; (2) how does mean site response and the associated ground motion variability differ for basins compared to non-basin geomorphic provinces?; and (3) what are the variations in basin response between different basin structures and how can this be modelled for predicting ground motion intensity measures?

We recommend a site classification scheme that distinguishes basins, basin edges, valley (sedimentary structures smaller in scale than basins), and mountain/hill areas. Moreover, we distinguish basins in southern California based on their geologic origin: coastal basins with varied depositional histories and large depths (e.g., Los Angeles Basin); inland, fault-bounded basins with relatively shallow sediments derived from neighboring mountains (e.g., Chino Basin); Imperial Valley, a basin on the transform fault plate boundary at the location of a graben located in the step-over between the San Andreas and Imperial Faults.

To address the second and third questions, we compile a large ground motion database for southern California that significantly expands upon the data available in the NGA-West2 project, and which has the benefit of significantly increasing the number of recorded events per site. We verify that an NGA-West2 ground motion model has unbiased source and path terms relative to the dataset, and we make minor modifications to the global VS30-scaling function to fit the mean of southern California data. Using the slightly adjusted ground motion model, we compute site terms for 670 sites, which approximately represent the mean difference between the actual site response and the ground motion model prediction.

We find that the combined data (i.e., site terms) from all sites exhibit trends with differential depth that are qualitatively similar to those in NGA-West2 models (de-amplification for negative differential depth, amplification for positive). We find basin and basin edge categories to be similar to each other, but different from the combined data set in the sense that de-amplification at negative differential depths is generally absent. Moreover, the depth-invariant mean amplification for this condition is positive, indicating under-prediction from the VS30-scaling model. We find the valley and mountain/hill categories to exhibit similar trends in which amplification scales with differential depths and has both positive and negative values. The depth-invariant mean amplification for these conditions is negative.

Among basin sites, we find differences for coastal and inland basins. The response of coastal basins essentially matches that for the overall basin category (amplification scales up with increasing differential depth). A similar pattern is followed by the Imperial Valley Basin. However, the response of other inland basins is different, with no appreciable dependence on differential depth, but apparent uniform shifts that are specific to individual basins (but which are poorly constrained by the data).

Models are proposed to capture the mean behavior of the recommended groups -- coastal basins, inland basins, and Imperial Valley Basin. Site-to-site standard deviation terms (S2S) are found to vary strongly across geomorphic provinces, with basins and valleys having notably lower dispersions than mountain/hill sites and the reference ergodic model.

Key Words
Site response, basins

Nweke, C., Stewart, J. P., & Brandenberg, S. J. (2020). Site Response of Southern California Sedimentary Basins and Other Geomorphic Provinces. Los Angeles, US: Natural Hazards Risk and Resiliency Research Center, B. John Garrick Institute for the Risk Sciences, UCLA . doi: 10.34948/N3159F.

Related Projects & Working Groups