Seismic microzonation by means of cluster analysis

Abstract

The identification of areas with uniform behaviour in seismic perspective is a very complex task that involves the simultaneous assessment of many parameters. In this paper we propose an automatic procedure for the seismic microzonation based on cluster analysis and performed by means of an agglomerative hierarchical algorithm. It exploits a proximity matrix specifically defined on the parameters space and the average linkage criteria. The procedure has been used to carry out the expedite seismic microzonation of some high seismic hazard areas, only using the data inferred by the extensive application of the HVSR technique. The results of the clustering procedures have been assessed from a quantitative point of view, by means of the variance decomposition. This procedure could be also used to identify homogeneous areas in a different parameters space or in a multi-dimensional space, to obtain a robust and exhaustive seismic microzonation, based on a quantitative method.