Statistical Methods and Modeling of Seismogenesis. Eleftheria Papadimitriou. Читать онлайн. Newlib. NEWLIB.NET

Автор: Eleftheria Papadimitriou
Издательство: John Wiley & Sons Limited
Серия:
Жанр произведения: Социология
Год издания: 0
isbn: 9781119825043
Скачать книгу
version of the code, the seismogenic source is approximated as a rectangle, composed of a number of cells with assigned dimensions in along-strike and down-dip directions. The rectangular fault is then divided into an arbitrary number of segments, without constituting any barrier to rupture growth. Their only role is to make it possible to assign different slip rates associated with tectonic loading onto each of them. Each cell is randomly assigned an initial stress budget within a given interval around an arbitrary average value. This is to accommodate our lack of knowledge about the initial status of stress and strength on each point of the fault. The stress budget on a cell is progressively changed in three ways:

       – it is increased at every time step (e.g. one day) through equation [2.18], according to the long-term slip rate (tectonic loading) mainly constrained by geodetic measurements;

       – it is decreased at the occurrence time of every rupture, by a given amount (e.g. 3.3 MPa); the same cell can rupture more than once in the same earthquake;

       – it is increased by a Coulomb stress change associated with a point source at the center of any other cell that ruptures during an earthquake; as all cells are assumed to rupture approximately with the same mechanism and the same fault plane, this stress change is always positive (see the appendix in section 2.6 for details).

      Particular attention has been given to the part of the simulator code that tunes the conditions of stopping an already initiated rupture. We obtained reasonable results by introducing a pair of “heuristic” rules to modulate the stress threshold to be exceeded for expanding an ongoing event into new cells or repeating the slip on an already ruptured cell. These rules, which have a relevant impact on the magnitude distribution of the synthetic catalogs, are:

       1) The stress threshold adopted for the nucleation of an event is decreased, after the initial rupture of the nucleation cell, by a quantity which is proportional to the square root of the number of the already ruptured cells, multiplied by a free parameter called the “strength reduction coefficient” (S-R). This feature mimics the sharp decrease of strength at the edges of an expanding rupture, through a sort of weakening mechanism. Increasing this parameter encourages the growth of ruptures, thus decreasing the b-value in the frequency-magnitude distribution. This parameter has a similar role to the η free parameter in the Virtual Quake simulator developed for California (Schultz et al. 2017).

       2) The square root of the number of already ruptured cells used in the previous rule is limited to a number equal to the width of the fault system, divided by the size of a cell, and multiplied by a free parameter called “fault aspect ratio” (A-R).

      Although the first of these two empirical rules enhances the capability of an already nucleated event to expand into a larger rupture, the second one limits this enhancement to a size that does not exceed by many times the width of the fault system. As has been proved by numerous tests, the strength reduction coefficient (of the order of a few percent) influences the proportion of the seismic moment released by small and by large earthquakes: the smaller this parameter, the larger the number of small events. On the contrary, for the fault aspect ratio, it has no influence on the magnitude distribution of the background activity, but affects the shape of the magnitude distribution in the large magnitude range. This simple algorithm ensures a stable process, during which the stress budget is maintained below the nucleation threshold and never vanishes, if a suitable initial value is chosen. The earthquake rate is modulated by the slip rate assigned to each fault segment.

Schematic illustration of rupture nucleation and propagation adopted in this earthquake simulator.

      The simulation algorithm, in its simplicity, provides preference for new ruptures to nucleate at the points of the fault where the stress budget is higher, i.e. where the time elapsed since the latest event is longer. Once it is nucleated, the rupture expands in the directions where the stress budget is still higher, thus simulating a preference for filling pre-existing gaps and epicenter migration. Moreover, because of the stress transfer included in the model, earthquakes are more likely to occur close to the borders of the rupture of a preceding large earthquake, simulating a feature similar to aftershock production.

Graph depicts frequency-magnitude distribution of the earthquakes in the synthetic catalogs obtained from the simulation algorithm 


                  <div class= Скачать книгу