Seismic source modeling

Plate tectonics cause stresses to accumulate in the Earth's crust. When these exceed the crust's yield strength, the most fragile areas (faults) rupture and an earthquake occurs. The fault suddenly slides and a rupture spreads, generating waves inside the Earth. The more extensive the ruptured surfaces, the higher the earthquake's magnitude (amount of energy released during rupture), and the greater the movement on the ground at the Earth's surface. Thanks to the waves recorded by several seismic sensors, we can explain the rupture mechanism and assess the parameters of the fault along which the rupture propagated (direction, length, depth, etc.), as well as the parts of the fault that ruptured the surface, their displacement and their rupture velocity.


We can observe waves that followed different paths depending on the sensor's position in relation to the earthquake.

When close by (<1500 km), we mainly record waves that are propagated through the crust. At these distances, waveforms are generally complex and reflect the multitude of interactions with the heterogeneities of the propagation environment.

At teleseismic distances (when the station is located more than 1500 km away), the seismic wave path mainly traverses the mantle, a more homogeneous environment than the crust. We can model and reproduce the behavior of the waves that cross it with relative accuracy. Accordingly, the shape and amplitude of the teleseismic seismograms for earthquakes of moderate to high magnitudes (ML>6) can be used to explain the rupture mechanism. The parts of the fault that ruptured the surface are modeled by elliptic surfaces known as "slip patches" [Vallée and Bouchon, 2004] and each surface slips at its own rupture velocity.
 
Example of application to the Boumerdès earthquake (Algeria, May 21, 2003, Magnitude = 6.8)
 

We first choose a set of teleseismic stations distributed as uniformly as possible at azimuths around the earthquake (Figure 1). The red star at the center of the figure shows the earthquake location, and the green squares the chosen stations.



Figure 1
 
We then construct numerically computed synthetic seismograms for each selected station and for different parameter sets characterizing the seismic source. The model ultimately adopted will be the one that can generate the synthetic signals that are closest to the real signals. In Figure 3, the signals represented in the outer ring are the P waves (pressure waves) and in the inner ring the S waves (shear waves).

click to enlarge the map
Figure 2: Real signals appear in black and synthetic signals in red.

Lastly, we visualize the model that enabled the best synthetic seismograms to be computed (Figure 3). We see here that 2 slip patches were required to explain the signal shapes. The color codes the final displacement (in meters) of the point of the fault under investigation. The green star represents the hypocenter, i.e. the initial rupture point.

Figure 3