Determination of Western Anatolia tectonic block model with cluster analysis

Anatolian Plate is an important tectonic structure with different types of faults. In this study, firstly, a block boundary study was carried out with the help of clustering analysis by using the horizontal velocity data of 841 GNSS points on the Anatolian Plate. Secondly, cluster analysis was performed with the horizontal velocity data obtained by evaluating 402 campaign type measurements in the Western Anatolia Region, which has a very complex tectonic structure on the Anatolian Plate. Elbow Method, Calinski-Harabasz Score, Silhouette Score and Davies-Bouldin Index algorithms were used to determine the optimum number of clusters for the Anatolian Plate. With the help of these algorithms, the optimum number of clusters for the study area was determined as 7, 8 and 9, respectively. Then, cluster analysis was applied using the K-Means algorithm. Unlike other studies in the literature, the parameter containing point location information was also included in the analysis. The block boundaries formed as a result of clustering follow the active fault zones and surface ruptures caused by earthquakes. In the block model boundary study on the Anatolian Plate, 3 block boundaries were defined differently from previous studies. When all parameters in the clustering analysis are evaluated, the most appropriate number of clusters is presented as 8 for the Anatolian Plate. In the Western Anatolia Region, 402 campaign type measurements were evaluated with GAMIT/GLOBK software. Then, cluster analysis was performed with K-Means algorithm for the horizontal velocity data. Elbow Method was used to determine the optimum number of clusters in this data set. With the Elbow Method, the most appropriate number of clusters in this region was determined as 14. Then, a synthetic data test was performed to show how well the K-Means algorithm can work with the existing data distribution. Checkerboard test was applied as synthetic data test. Cluster analysis was then performed for the Western Anatolia Region using the K-Means algorithm for the 14 clusters identified. In addition to the horizontal velocity data as in the Anatolian Plate, the point location information parameter was also included in the analysis. The number of micro-blocks formed with the block boundaries defined on the Western Anatolia Region was determined as 15 in this study. In addition to this, 18 micro-block boundaries in the Western Anatolia Region together with other block boundaries proposed in the study are presented in this study.