K-Means Cluster for Seismicity Partitioning and Geological Structure Interpretation, with Application to the Yongshaba Mine (China)

Abstract

Seismicity partitioning is an important step in geological structure interpretation and seismic hazard assessment. In this paper, seismic event location ( , , ) and Euclidean distance were selected as the -Means cluster, the Gaussian mixture model (GMM), and the self-organizing maps (SOM) input features and cluster determination measurement, respectively, and 1516 seismic events ( > −1.5) obtained from the Yongshaba mine (China) were chosen for the cluster analysis. In addition, a Silhouette and Krzanowski-Lai-(KL-) combined S-KL index was proposed to obtain the possible optimum cluster number and to compare the cluster methods. Results show that the -Means cluster obtains the best cluster "quality" with higher S-KL indexes on the whole and meaningful clusters. Furthermore, the optimal number for detailed geological structure interpretation is confirmed as eleven clusters, and we found that two areas probably have faults or caves, and two faults may be falsely inferred by mine geologists. Seismic hazard assessment shows that C5 and C7 ( = 11) have a high mean moment magnitude ( ) and C1, C2, C3, and C4 ( = 11) have a relatively high , where special attention is needed when mining. In addition, C7 ( = 11) is the most shear-related area with a mean S-wave to P-wave energy ratio ( Es/Ep ) of 41.21. In conclusion, the -Means cluster provides an effective way for mine seismicity partitioning, geological structure interpretation, and seismic hazard assessment.

Data Field-Based K-Means Clustering for Spatio-Temporal Seismicity Analysis and Hazard Assessment

“…Morales-Esteban et al [10] proposed an efficient adaptive Mahalanobis-based K-means algorithm and this has been applied to study the seismic catalogues of Croatia and the Iberian Peninsula. Shang et al [11] proposed a Krzanowski-Lai and Silhouette combined index to select the optimum number of clusters for K-means and they interpreted the geological structure in the Chinese Yongshaba mine. The K-means cluster is useful for a large dataset cluster, and some studies have been done to reduce the effect of initial cluster centers.…”

“…Besheli et al [21] performed a K-means clustering along with a SOM on an Iranian foreshock database and they found a foreshock zone which has a very close relation with large earthquakes. Nonetheless, research results [11,22] have shown that the SOM clustering may only be valid for high seismic activity areas, while for low seismic activity zones it may have bad cluster results. Furthermore, the SOM clustering may have discontinuous zones, which makes it hard to interpret cluster results [11,22].…”

“…Nonetheless, research results [11,22] have shown that the SOM clustering may only be valid for high seismic activity areas, while for low seismic activity zones it may have bad cluster results. Furthermore, the SOM clustering may have discontinuous zones, which makes it hard to interpret cluster results [11,22].…”

“…Martínez-Álvarez et al [22] used a TriGen-based clustering algorithm [28,29] for seismic zoning in the Iberian Peninsula. A comparison of using different seismic clusters can be found in [11,30,31].…”