Identification of Paleotsunami Deposits Using XRF and Artificial Intelligence Methods on the Southern Coast of Lebak, Banten

Abstract

The southern coastal area of Lebak, Banten is the southern region of Java which is prone to tsunami disasters. There is evidence of past tsunami events in the southern region of Java. However, not all tsunami deposits have identifiable sedimentological and micropaleontological traces. Geochemical proxies and artificial intelligence with machine learning methods can be used to identify paleotsunami deposits. Machine learning methods that can be used to cluster paleotsunami deposits are Agglomerative Hierarchical Clustering (AHC) and Support Vector Machine (SVM) with validation of model accuracy using the Receiver Operating Characteristic (ROC) and Area Under Curve (AUC) methods. Input data are XRF analysis data and macroscopic core sample description data. The output from data processing is in the form of prediction of tsunami and non-tsunami deposits at each depth of core data sample. The data is correlated and interpreted to identify tsunami events that occurred in the past. The identification results show that the tsunami deposition in the research area, namely the area around the Bagedur coastal coast and Bolang village had tsunami deposit characteristics based on macroscopic description, XRF analysis, and artificial intelligence clusterization. It is also thought to be correlative with the tsunami deposition of previous research in the area around the Binuangeun coastal coast.