GNGTS 2018 - 37° Convegno Nazionale

GNGTS 2018 S essione 2.1 287 PRELIMINARY NUMERICAL SIMULATIONS OF THE 28 SEPTEMBER 2018 TSUNAMI IN SULAWESI (INDONESIA) A. Armigliato, E. Baglione, G. Pagnoni, M. A. Paparo, S. Tinti, F. Zaniboni Dipartimento di Fisica e Astronomia, Alma Mater Studiorum – Università di Bologna, Italy On 28 September 2018 at 10:02 UTC an earthquake of magnitude 7.5 struck the island of Sulawesi (Indonesia). The earthquake, whose epicentre was located about 78 km north of the city of Palu, was characterized by a prevailing strike-skip faulting mechanism. According to the results available at the USGS web site (https://earthquake.usgs.gov/earthquakes/ eventpage/us1000h3p4/executive), the region of maximum ground shaking is elongated in the N-S direction, from the epicentre to the north down to several tens of kilometres south of the city of Palu. The shaking was responsible for huge and diffuse liquefaction and landslides, but the most devastating phenomenon following the earthquake was a huge tsunami, whose effects were much worse than what could be expected on the basis of the magnitude and of the focal mechanism determination, and which contributed to increase dramatically the level of damage and the number of fatalities. The tsunami was destructive especially inside the bay of Palu, an approximately 30-km long and 6-km wide narrow basin elongated along the N-S direction and with steep bathymetric gradients and maximum depths in the order of 700-800 m. The tsunami impact in different places along the bay of Palu coasts was documented by a large number of videos and reports broadcasted by eye-witnesses over the social networks and successively made available in the web. The destruction documented live over the social networks was in jarring contrast with the performance of the Indonesia Tsunami Early Warning System (InaTEWS): although a warning was issued timely (5 minutes after the earthquake), the combination of the short time interval between the earthquake onset and the arrival of the first tsunami wave, of the underestimation of the expected maximum run-ups, and of the poor efficiency of the downstream communications resulted in a very large number of victims. Moreover, the warning was lifted after about 30 minutes, when a 6-cm amplitude was observed at the tide-gauge of Mamuju (very far from the Palu bay), leading to the wrong evaluation of the tsunami evolution. At the end, more than 2100 fatalities were counted in total in the whole Sulawesi island and more than 600 people are missing, although it is still not clear whether it is possible to separate those caused by the earthquake from those due to the tsunami. From the instrumental point of view, the tsunami was clearly detected only by three tide- gauge stations. Two of them (Bitung and Mamuju) are very far from the epicentral area and from the bay of Palu: the detected sea level anomalies are in the order of few centimetres peak-to-peak. Instead, the tide-gauge of Pantoloan is placed inside the bay of Palu, along its north-eastern coast. The record from this sensor shows a clear first negative movement of the sea occurring 5 minutes after the earthquake, followed by a positive peak of approximately the same amplitude: the peak-to-peak amplitude results in 3.8 m and an estimate of the characteristic period is 6 minutes. Three other wave packets of gradually decreasing, but still relevant (> 1 m) peak-to-peak amplitude are observed during the first hour: the oscillations then continue more or less for other 4 hours. Post-event field surveys have been carried out along the coasts of the Palu bay by an Indonesian team and by a Japanese team in the very first days of October 2018. Other surveys involving international tsunami researchers, including also one of the authors of this paper, are planned to be carried out in November. For the city of Palu, the Japanese team (see Muhari et al. , 2018) reported maximum inundation distances in the order of some hundreds of meters and maximum flow-depths locally larger than 9 m. Along the eastern coastline of the bay, flow- depths decrease from south (4.8 m in Marboro) to north (1.9 m north of Pantoloan). A rather detailed picture of the inundation distance along the coastal areas of the Palu bay can be deduced also through the comparison of satellite images captured shortly before and