Satellite imagery strategies for earthquake strong-motion prediction and shakemap generation – SASHA

Although shakemaps are of great interest for large-magnitude earthquakes where the ground motion is mainly controlled by the rupture on the fault, their relevance for moderate-size events is still questionable. So far, the implemented ground-motion prediction does not have the necessary spatial resolution to account for the real impact of the earthquake, because shakemaps use only first-order parameters that are not relevant for the fine city or building scales. This results in their poor efficiency in areas close to moderate earthquakes and/or located on deep sedimentary layers prone to complex site effects. This project aims at improving the reliability and the spatial resolution of the shakemaps and at developing damage maps to predict the impact of an earthquake anywhere, including in complex geological environments and in the lack of seismological data. Our methodology is targeted on rapid and low-cost estimates, but still keeping track of physics-based approaches. We take advantage of satellite imagery in France to develop an innovative method for retrieving sediment thicknesses and predicting sediment resonance periods based on the mapping of subsidence rates in alluvial and urbanized valleys from InSAR data. The collected parameters are compared to observations: earthquake recordings from French seismic stations, from low-cost seismometers hosted by citizens, contemporary and historical macroseismic intensities. The comparison is performed by artificial intelligence to develop a new estimator of the aggravation of the ground motion at fine scale in sedimentary valleys and their periphery. The combination of the high-resolution shakemaps with an innovative damage prediction model will allow to get the damage maps at building scale.