Physics-Based Modelling of Wave Propagation in Basins

Spectral Aggravation Factor

To investigate the basin-induced aggravation in addition to the 1D site amplification, rather than the absolute ground motion in a sedimentary basin, we (Zhu et al., 2018a, b) constructed a total number of 82 heterogeneous trapezoidal basin models. Most of those basin models are based on real 1D soil profiles compiled from the KiK-net database. Nine strong ground-motion recordings (rock site) are selected from the NGA-West2 database. Both the 1D and 2D site responses of each basin subjected to these nine seismic records are modeled using the finite difference approach. 2D simulations model the vertical propagation of plane SH waves in 2D basins while 1D computations simulate the vertical propagation of SH waves in local soil columns. We then use the spectral aggravation factor (SAG) to quantify the aggravation in basins. SAG at a surface receiver is defined as the ratio of its spectral acceleration computed using a 2D model to that computed using the local 1D model at this location. 

In all these basins, significant aggravations are only found in the period ranging from 0.1s to Th, where Th is the 1D fundamental site period at the basin center. Beyond this period range, 1D modeling gives quite comparable results to a 2D analysis. However, the spatial distributions of aggravation are found to be different in relatively deep and shallow basins. In shallow basins (Zhu et al., 2018a), it is the region close to the basin edge that experiences significant ground motion aggravation, increasing 1D amplification by a factor of which the 16th and 84th percentiles are 1.2 and 1.5, respectively. In contrast, in deep basins (Zhu et al., 2018b), the whole basin width is significantly aggravated by a factor of which the 16th and 84th percentiles are 1.3 and 1.8, respectively. The above results are obtained using basin models with regular geometries. To verify these results, we (Zhu et al., 2018c) then utilize real sedimentary basins alluvial valleys with highly irregular sediment-bedrock interfaces, i.e., Mygdonia valley in Greece, Upper Rhine Graben (Karlsruhe) in Germany, Bovec Basin (B-B) in Slovenia, Bovec Basin (C-C), Valais (Monthey) Valley in Switzerland, and Tagliamento Valley in Italy. Results show that previous conclusions are applicable to these real basins (valleys). 

These findings have significant engineering implications. Current 1D-based approaches to assess site effects can significantly underestimate the seismic hazard in the close-to-edge region of a shallow basin, and in the whole surface region of a deep basin. Applying a constant aggravation factor to the basin as a whole to account for basin effects is inappropriate for a shallow basin but might be feasible for a deep basin. 

 Zhu, C., F. J. Chávez-García, D. Thambiratnam, C. Gallage (2018a). Quantifying the edge-induced seismic aggravation in shallow basins relative to the 1D SH modeling. Soil Dynamics and Earthquake Engineering 115, 402-412. https://doi.org/10.1016/j.soildyn.2018.08.025 

Zhu, C., D. Thambiratnam, C. Gallage (2018b). Statistical analysis of the additional amplification in deep basins relative to the 1D approach. Soil Dynamics and Earthquake Engineering 104, 296-306. https://doi.org/10.1016/j.soildyn.2017.09.003  

Zhu, C., D. Thambiratnam, J. Ritter, J. Zhang (2018c). Can site effects be accounted for using a single aggravation factor in irregular-shaped Alpine valleys? 16th European Conference on Earthquake Engineering, Thessaloniki, Greece, 18-21 June 2018.