Rapid changes in geotechnical and geological ground conditions lead to significant ground motion variability. This condition mainly occurs at the so-called basin edges, where there is an abrupt transition between soft highly compressible soils and stiffer materials. This problem becomes more relevant in areas where ground subsidence drastically changes the dynamic response of high plasticity clay deposits, such as those found in Mexico City, due to fundamental site period evolution with time. This paper presents site response analyses at an abrupt transition area in the southeast Mexico City region, along the edges of the Xochimilco-Chalco lakes. Considerable damage associated with three-dimensional wave propagation effects was observed in this zone during the September 2017 Puebla-Mexico earthquake. A series of three-dimensional finite difference numerical models of the basin edge were developed to evaluate ground motion variability, considering topographic effects and soil non-linearities. Good agreement between the computed response and the observed damage during the 2017 Puebla-Mexico earthquake reconnaissance was found. In addition, several normal and subduction events with a return period of 250 years were considered to evaluate the effect that frequency content, and strong ground motion duration have on the soil response variability. From the results gathered here, it was established the relevance of accounting for three-dimensional wave propagation fields to assess site effects at basin-edge zones properly and to be able to implement proper risk mitigation measurements at these zones.

Due to the unique soil, morphological, and subsurface topographical conditions, amplified and prolonged seismic demand traces were observed in historical strong ground motion records from Bayrakli-Izmir-Turkiye. A vivid example of this response was recorded during the Mw 7.0 Samos event on October 30, 2020. After the event, structural damage and loss of life were unexpectedly concentrated in Bayrakli-Izmir, even though the fault rupture was located 70 km away. The presence of strong ground motion stations (SGMS) located on rock (#3514) and soil (#3513) sites enabled a quantitative assessment of the amplified and prolonged seismic demand traces. The seismic response of SGMS #3513 site was assessed by using 1-D equivalent linear and analytical methods. The idealized 1-D soil profile and input parameters were calibrated and fine-tuned by using the 2020 Samos earthquake accelerograms. Then, the calibrated equivalent linear site response model was further validated by the recordings from historical events. Alternatively, an analytical wave propagation-based model was proposed, the input parameters of which were probabilistically estimated based on, again, historical recordings. Finally, the seismic responses of the site during future earthquakes were predicted based on the calibrated and validated site response models. The predicted intensity-dependent amplification spectral responses were compared with the provisions of the TEC (2018). Even though limited in number in all five future seismic scenario events, amplification ratios suggested by TEC were exceeded by a factor of 2-4 at periods falling in the range of 0.5 to 1.2 s. This clearly suggested the need to further quantify the Bayrakli seismic basin responses with basin-specific models, rather than code-based, intensity-dependent generalized amplification factors.