DOI: https://doi.org/10.24028/gzh.0203-3100.v35i4.2013.111423

Estimation of magnitude and the depth of the subcrustal source of Vrancea zone in quasi-real time scale by the connectionist simulation

M. Lazarenko, O. Gerasimenko

Abstract


The heading portions of Vrancea three component earthquakes, recorded by Odessa seismic station, are used for development of neural network model of quake’s magnitude and hypocentral depth prediction. The validity of prediction is demonstrated for different components and time­window length used for neural network training. The magnitude assessment by the leading time interval of seismic signal can be used for seismic hazard mitigation in the spot of recording.

References


Лазаренко М. А., Герасименко О. А. Використання нейронних мереж для побудови годографів сейсмічних хвиль // Вісн. Київ. ун­ту. Геологія. — 2004. — Вип. 30. — С. 93—97.

Рихтер И. Ф. Элементраная сейсмология. — Москва: Изд. иностр. лит., 1963. — 670 с.

Leach R. R. J, Dovla F. U., Vergiuo E. S. Yeld estimation using bangpass­filtered seismograms: Prelimenary results Using Neural Networks with mb(Pn), Shorttime, Long­time, and Coda energy Measurments // Bull. Sеism. Soc. Amer. — 1993. — 87, № 2. — P. 488—508.

Magotra N., Ahmed N., Chael E. Seismic Event detection and source Location Using Single­Station (Three­Component) Data //Bull. Seism. Soc. Amer. — 1987. — 77, № 3. — P. 958—971.