سال انتشار: ۱۳۸۶

محل انتشار: پنجمین کنفرانس بین المللی زلزله شناسی و مهندسی زلزله

تعداد صفحات: ۷

نویسنده(ها):

Hashemi – Assistant Professor, Faculty of Earth Sciences, Damghan University of Basic Sciences, Damghan, Iran
Mehdi Zadeh – M.Sc. student, Faculty of Earth Sciences, Damghan University of Basic Sciences, Damghan, Iran

چکیده:

The Zagros region of Iran is known for its typical tectonic setting as a collision zone. This region shows high tectonic activity, especially from the seismicity viewpoint. Although, the tectonic and seismotectonic zoning of this region has been the subject of many researches during the past decades, but this study introduces a new objective method for the zoning. In this research, the Zagros region has been selected for presenting and demonstrating the ability of the cluster analysis method as a computer-aided pattern recognition technique to perform a seismotectonic pattern recognition study. For this purpose, the study area firstly divided into a number of 137 non uniform sub-areas each one separated from the neighboring, by natural geological and geophysical features. Then, for each sub-area a number of 18 geophysical and geological properties such as seismicity, gravity anomalies, lithology, topography, folding, and faulting were calculated and recorded as quantitative variables. After preparation of data matrix, it was analyzed by suitable statistical software for finding similarities between sub-areas. Results obtained have been used for constructing the seismotectonic zoning maps of the region in different levels of similarities by joining the similar sub-areas. The results indicate that among different solution zoning maps, the 6-zone solution map shows the best agreement with the overall geological and structural characteristics of the area. Finally, these 6 seismotectonic zones were compared to each other quantitatively and the general characteristics of them were discussed. Since our results show good agreement with the previous qualitative studies, it can be said that cluster analysis as a statistical pattern recognition technique can be very useful in identification of hidden patterns based on both values and changes of the measured data. This new approach can be accounted only as a starting point and it is expected to be improved and refined by using more accurate data in future.