Solving the puzzle of the 1996 Biak, Indonesia tsunami

Abstract

Abstract On February 17, 1996, an earthquake ( $${M}_{\text{w}}=$$ M w = 8.2) occurred northeast of Biak Island, Indonesia, and generated a tsunami. Interestingly, the tsunami runup on the southwest side of Biak Island, which did not face the epicenter, was higher than that on the side of the island facing the epicenter. It was earlier suggested that the earthquake triggered submarine landslides. However, this hypothesis remains unexplored and unconfirmed. In this study, tsunami arrival times obtained from eyewitness accounts were used to perform backward tsunami travel time modeling. Based on the backward tsunami travel time results and multibeam bathymetry, five submarine landslide candidates were identified: three located to the southwest of Biak Island and two to the south of the island. Our analysis indicates that tsunamis from a large submarine landslide off the southwest coast of Biak Island (SL 2) and a small submarine landslide off the south side of Biak Island (SL 4) contributed to the 1996 Biak tsunami event and came earlier than the main tsunami from the earthquake fault. Tsunami sources were earlier modeled only by fault parameters without considering submarine landslide sources. As a result, the models could not explain the observed runup and eyewitness tsunami arrival times for the southwest and south of Biak Island. To address this problem, we propose an approach that combines a submarine landslide model with a modified version of a previously proposed fault model. Our combined model explains the observed runup and eyewitness tsunami arrival times well (Aida’s index values of the model are K = 1.00 and $$\kappa$$ κ = 1.44). Graphical abstract

Keywords

• Tsunami simulation
• Submarine landslide
• Backward tsunami travel time
• Biak Island
• Indonesia