NERC Urgency Sulawesi tsunami 2018

At 6pm local time on the September 28th 2018, a strike-slip earthquake, Mw 7.5, struck the west of Sulawesi Island in Indonesia. Earthquake shaking immediately destroyed large areas of Palu City, the nearest large town of 135,000 people. Further damage was caused by liquefaction of underlying, fine-grained fluvial sediments, which mobilised mudslides that swept through the city. Soon after the earthquake, tsunamis, with elevations of up to 11 m inundated local coastlines causing further destruction, especially to Palu City, and Donggal farther north. At present, 2,100 people are known to have died in the event with over 700 missing. A significant number of these died in the tsunamis.

Recent tsunamis provide critical information on mechanisms and impacts which provide essential data on informing on hazard and risk. The mechanism of the Sulawesi earthquake is significant because it is different to recent devastating events, such as Papua New Guinea (1998), Indian Ocean (2004) and Japan (2011), which were triggered by thrust faulting. Sulawesi is a strike-slip rupture with dominantly horizontal movement. Present understanding of earthquake tsunami generation suggests that, because of the absence of significant vertical seabed elevation, it should not have generated the recorded (up to 11m) tsunamis. Preliminary numerical tsunami modelling confirms this understanding. An additional tsunami mechanism is required, and submarine landslides are most likely.

Numerical modelling of tsunamis require validation from field observations to confirm their accuracy. These observations have to be made as soon as possible after impact because much of the evidence on the tsunami and its' scale is extremely fragile and short-lived. For example, the directions of flow may be identified by flattened grass; tsunami flow depth and elevation may be identified from water marks on buildings or building damage, vegetation stripping and debris caught in trees or bushes. Sediment deposited, especially in monsoonal conditions, is rapidly eroded or removed. It is vital therefore that after a major tsunami, field observations are made and the research carried out as soon as practicable.

Here, for Sulawesi, we propose an urgent response field survey, based on pre-survey interpretations of before and after high-resolution satellite imagery. The results will identify in detail geographical variations on tsunami inundation which will inform on the impact, potential tsunami mechanisms and offer validation for numerical tsunami models. This methodology (desk study followed by field work) was successfully applied to the Japan 2011 tsunami, and is a first critical step towards developing an integrated system for interoperable digital field data collection.

From the present field surveys there is information over much, but not all, of the area impacted by the tsunami (see 2). There is a close correlation between the offshore strike-slip rupture (mapped from remote - inSAR data) and field evidence for a tsunami from inundation measurements. There is also correlation between the offshore rupture and coastal subsidence, which supports coastal/submarine landslides as the tsunami mechanism. For the proposed survey, PI Tappin leads an international and multidisciplinary initiative (UK, Indonesia, US and Poland) to study the Sulawesi tsunami. This approach is most likely to fully understand the event, and is required because, at present tsunami warning in Indonesia is based on far-field earthquakes. Locally triggered events, such as Sulawesi, where there was minimal warning, are not at present addressed. More broadly, the project will contribute to an improved understanding of locally, strike-slip triggered, submarine landslide tsunamis and their hazard. The project will inform on improved mitigation strategies in this context.

The research will form the basis for future, more in depth research.

Initially, the results will be of use to tsunami scientists studying tsunami impact in relation to tsunami source and levels of impact inundation. The research improves our understanding of the tsunami in relationship to the likely source mechanisms (in this case the earthquake triggered landslides) and in improving and developing future mitigation strategies.

We hope the results will be used by government hazard managers in Indonesia for developing new mitigation strategies because it provides a different context to tsunami hazard in the country, which suffered such a major catastrophe in 2004.

Tsunami mitigation following the 2004 tsunami was to develop a warning system to address megathrust events, mainly in west Indonesia - Sumatra and Java, where there are major population centres. The warning system was mainly focussed on events where there was time to allow evacuation of coastal communities. One major realisation from the Sulawesi tsunami was that since 2012, much of the infrastructure (such as DART buoys) had become inoperable (none of the 22 buoys was operational) and the infrastructure (seismology, tide gauges) of the system was out of date. A major aspect of warning is that local populations need to know how to respond when a warning is issued, and in Sulawesi, this was not the case. In addition, the tsunami struck so soon after the earthquake, there would have been little time at many locations to respond.

With hindsight, there had been other analogous tsunami events previous to Sulawesi, Java in 2006, but the most important was Flores Island in 1992. Here an earthquake tsunami killed over 1800 people. There were similar anomalous (high) tsunami runups, which were considered to be from coastal or submarine landslides, but this was never proved. One of the most serious aspects of the Sulawesi disaster was that a publication in 2016 (Cipta et al., 2016) recognised the major hazard along the fault that slipped, identifying the local hazard in a nearfield context, from liquefaction, landsliding and tsunamis, with previous events in 1927 and 1997. Understanding the mechanism of the Sulawesi tsunami - most likely from landslides - will highlight the landslide hazard and that this is very different to megathrust events, sourced quite a long way offshore and with more warning time. As has been highlighted, the lack of simple educational programmes is an essential part of mitigation, but a validated explanation is required to support developing these. Proving the non-seismic mechanism for 2018 is an essential contribution to this.

Hopefully, after the 28th September event major changes in mitigation will take place in Indonesia, and in the east of the country where the tsunami hazard could be different to that in the west, with little warning time, and more of an emphasis on reaction through education. These changes need to be based on a solid foundation of observational evidence on the impact of the tsunami. The data acquired during the forthcoming surveys will thus be essential for use by government hazard managers in developing these new strategies in preparation for the saving of lives when the next tsunami strikes, as it will.

Cipta, A., et al., A probabilistic seismic hazard assessment for Sulawesi, Indonesia. Geological Society, London, Special Publications, 2017. 441(1): p. 133-152.