Earthquakes represent one of the most significant destructing natural hazards worldwide. Recent devastating earthquakes near cities have shown that urban areas should become the centerpiece of seismic hazard and risk studies.
The global urban population has increased in recent years and it will continue in the oncoming decades. While earthquake prediction is still in its early stages, seismic ground motion prediction has advanced quite considerably mainly due to the deployment of dense arrays in Japan and in the USA. These data are used to construct empirical ground motion prediction models that are worldwide employed in seismic hazard studies. In spite of these networks, located in seismically active zones, there is a lack of ground motion recordings near the seismic source for all kind of soil conditions. In addition, local geology may strongly affect the incident wavefield producing amplification, longer duration and spatial variability on the ground motion.
Moreover, if the surficial soil is water saturated, its strength is not high and the seismic ground motion is strong, the material may present nonlinear behavior. This means that the instantaneous elastic properties of the medium change (i.e. reduction of the shear modulus and increase of the damping), strongly affecting the medium response to earthquake activity. Velocity changes have been studied at the scale of the crust using ambient seismic noise before, during and after an event as it was shown for the 2011 Tohoku earthquake. It is observed that co-seismic crust velocity changes during the 2011 Tohoku earthquake affected the volcanic regions in Japan.
At a smaller scale, in the order of 100-200 m depth, shear wave velocity variations using ten years of KiK-net data show that near-surface velocities are modified by earthquake and rain activity. These studies show that earth materials behave as a nonlinear elastic body. Indeed, after several weeks or months, the material recovers its behavior prior to strong earthquake shaking.
However, laboratory soil tests show clear nonlinear hysteretic behavior during cyclic loading. Does this mean that soil nonlinear hysteretic behavior is a transient phenomenon within a larger nonlinear, slow dynamics behavior? Does the fast soil dynamics have a healing? If so, what is the relation between fast and slow dynamics behavior?
Understanding nonlinear soil behavior using in-situ data is a major task. The majority of soil nonlinear rheology studies come from laboratory experiments. There are few field observations of clear nonlinear soil behavior, and their comprehension is almost non-existent. Furthermore, the incident wave-field is strongly affected by nonlinear material response and therefore the buildings constructed on these types of soils.