Considerations on the diffuse seismicity assumption and validity of the G-R law in stable continental regions (SCR)

• Autor(a) principal: Riera, Jorge Daniel
• Data de Publicação: 2015
• Outros Autores: Iturrioz, Ignacio
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Institucional da UFRGS
• Texto Completo: http://hdl.handle.net/10183/189212
• topic: Terremoto; Análise de risco; Método dos elementos discretos; Diffuse seismicity; Stable continental regions; Seismic risk; Gutenberg richter law; B-value; Weibull distribution

It is well known that seismic activity is much higher along inter-plate boundaries, decreasing markedly in intra-plate regions. Although a few locations in so-called Stable Continental Regions (SCRs) around the globe, like the New Madrid region in the USA, have been subjected to earthquakes with magnitudes above Mw = 8, the largest events in most SCRs do not exceed Mw ≈ 7, and their prediction for engineering purposes presents great difficulties on account of the scarce available evidence on seismic activity in intra-plate regions. The situation led in the last two decades to extensive studies promoted by EPRI. In view of the difficulty to identify seismogenic sources in most SCR areas, in such regions the assumption of diffuse seismicity is often accepted. The South American Plate is used in this paper as an illustrative example. The acceptance of the assumption of diffused seismicity justifies the adoption of a Poisson process to describe the distribution in time of the occurrence of seismic events in such locations. Note that, if in addition to the assumption of a Poisson process in time to describe the occurrence of seismic events, the distribution of the magnitudes of these events is assumed to be exponential, the result is the well-known Gutenberg-Richter relation. In the paper the authors examine available seismic data for a 1200km square region in the South American SCR, showing conclusively that the distribution of amplitudes is not exponential, but may be satisfactorily approximated by a Weibull (minimum) distribution, giving rise to a frequency-magnitude relation that differs from the G-R relation and that presents improved fit to available data, since the G-R law is just an specific case of the former. Moreover, the approach permits the identification of large seismic events that should not be included in the same data base as the background seismicity, since they are characterized by a different rate of occurrence.