EGU General Assembly 2020
© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

The influence of velocity gradients choice in deep alluvial basin seismic site response

Valeria Cascone and Jacopo Boaga
Valeria Cascone and Jacopo Boaga
  • Department of Geosciences, University of Padova, Italy (

The characterization of seismic site response represents one of the most important issues of seismic hazard assessment and risk mitigation planning. Characterizing the site conditions involves the measurement of several soil properties such as the shear-wave velocity (Vs), density and damping properties as a function of depth. Therefore, most of the site-effect studies in earthquake ground motions are based on the properties of the upper 30 meters and the anti-seismic building codes propose in most cases a simplified analysis based on shear wave velocity of the shallow subsoil. From a seismological perspective, the upper 30 meters would almost never represent more than 1% of the distance from the source. This should be taken into account especially for large and deep alluvial basins, representing the most inhabited geological environment of the world, where could be difficult to estimate the thickness and the velocity profile of the soft sediment overlying the rigid seismic bedrock.

The common approach adopted to characterize greater depths is then an extrapolation of shear wave velocity in depth, considering a selected linear or non-linear velocity gradients till the depth of the considered seismic bedrock (usually set to Vs ≥ 800 m/s). These gradients are generally derived from geological information or from literature, but how much the gradients choice affects the final site response analyses is often a neglected aspect.

In this work we try to investigate the generic case of deep alluvial basins. We consider the shallow subsoil as characterized by several in-situ tests in northern Italy. We extrapolate the deeper soil structure considering different literature velocity gradients obtained for deep basins in different geological contests: tectonic basins (Lower Rhine Basin and Po Plain) and Alpine basins (Grenoble and Lucerna Basins). We perform one-dimensional analysis of shear waves with the Linear Equivalent Method. The study demonstrates how relevant can be the role of velocity gradient choice for the ground response scenario. Starting from the same shallower Vs structures, the computed seismic motion at surface can present variation in the order of 50% varying the velocity gradients in depth. The results are of relevant interest for the analysis of seismic hazard in the deep alluvial basins environments, which host the main urban areas around the world.

How to cite: Cascone, V. and Boaga, J.: The influence of velocity gradients choice in deep alluvial basin seismic site response, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-5123,, 2020

Comments on the presentation

AC: Author Comment | CC: Community Comment | Report abuse

Presentation version 1 – uploaded on 30 Apr 2020
  • CC1: Comment on EGU2020-5123, Hans-Balder Havenith, 03 May 2020

    Dear Valeria,


    your results seem to show that very deep investigations are necessary to assess the velocity gradients, ...even inside the bedrock (if the bedrock does not reach a certain Vs level).

    Or should we mainly try to catch the upper (in your suimulations highly curved) part of the gradient.

    It would be interesting to know down to which depth we need to specify the gradient .. as reaching 100 m is possible (with arrays ..) while 350 m not. Maybe, if you had specifid the gradient down to 100 m, the differences in amplification levels wouldn't have been so strong.

    ... if you participate tomorrow in the chat - leave the answer for tomorrow.



    • AC1: Reply to CC1, Valeria Cascone, 04 May 2020

      Dear Hans-Balder 

      Thank you for your interest. 

      In the soil profile model shown in our poster we consider an unknown bedrock depth, so the model reaches different bedrock depth (Vs ≥ 800 m/s) depending on the type of gradient.  

      Of course, if I consider only the 100m in depth the influence of the type of gradient couldn't be so evident. However in a future work we will consider separate datasets which have common bedrock depth.


  • CC2: Comment on EGU2020-5123, Daniel Bowden, 04 May 2020

    Interesting figures!
    Emphasizing the forward problem is nice -> deeper velocities can have a big impact! Some conventional wisdom for GMM's is that we mainly just need Vs30 / shallow profiles, but you nicely show that this is not sufficient, especially for periods ~1 second.
    Does your work imply anything about solving for deeper structure? Could we use Spectral Acceleration to help uniquely constrain the right power-law coefficient? Or do we need some other tomographic method?