The Sulawesi earthquake and tsunami: what happened and what we know Now
Davide Mencaroni (ESR6), William Meservy (ESR12)
The September 28th, 7.5 Mw earthquake affecting Sulawesi, Indonesia has been the most catastrophic seismic event in 2018. A report from the ASEAN Coordinating Centre for Humanitarian Assistance on disaster management documents 2,081 fatalities caused by a series of cascading natural disasters concomitant with the quake. These include quick-clay and liquefaction-related landslides and a near-field tsunami that devastated the cities of Palu, Donggala, and Mamuju. Besides generating numerous fatalities, these catastrophes damaged thousands of homes and displaced more than 206,000 people. Tens of thousands of individuals are currently living in temporary shelters and tents as a result.
Images and videos of the shaking, landslides, and tsunami associated with the quake spread virally on social media only minutes after their occurrence, and reporters and bloggers continue to share satellite footage of the landslides and tsunami inundation extent on a near-daily basis. These images not only demonstrate the event’s remarkably destructive nature, they provide incredible value to the scientific community. Experts across the globe use this material—in combination with harder-to-acquire geophysical data and previous studies of the region—to develop a more precise understanding of what transpired in order to more accurately forecast future natural hazards.
Now, more than a month after the event, there is still a lot left unexplained but, also, much we know. This post pieces together preliminary information relating to the earthquake, landslides, and tsunami in order to provide a possible explanation for why this relatively small earthquake has proved so catastrophic and how to prevent a similar catastrophe.
Sulawesi lies within a highly active, complex subduction area at the triple junction between the Australian, Eurasian, and Philippine Sea plates. Several faults accommodate regional deformation, but the main active structure is the Palu – Koro fault. This is a left lateral, NNW-SSE trending, strike-slip fault that traverses the capital and most populated city of Sulawesi’s central province, Palu. Sinistral slip rate of the fault is in the range of 30-40 mm/yr, and a recent study from Watkinson and Hall (2016) recognizes it as the greatest seismic hazard to eastern Indonesia. Indeed, Sulawesi’s September 28, 7.5 Mw earthquake generated along the Palu – Koro fault.
Preliminary modelling of the event show rupture over a length of 150 km, including surface rupture through the city of Palu. The mainshock occurred at 18:02 local time, when blackouts occurred throughout the area and telecommunications went down because of intense shaking.
At 18:07, authorities issued a general tsunami warning for the region via SMS and television, and the people in Palu were told to expect a tsunami of less than 0.5 meters. The message, beyond proving a significant underestimation of the eventual tsunami height, was sent while hundreds of villagers gathered on the beach for the Palu Nomoni Festival, a celebration of the city of Palu’s 30th anniversary. 34 minutes later, the Indonesian authorities responsible for issuing the warning, unable to contact their counterparts in Palu, chose to lift the tsunami warning based on what they expected would be a much smaller and more delayed wave. This cancellation, however, came right after the third tsunami wave and may have had little effect on the overall number of casualties.
In the end, the waves affecting Palu and other nearby coastal villages reached runup heights in excess of 6 meters, something that few predicted. In general, strike-slip earthquakes produce much smaller tsunami runup heights. In fact, a perfectly horizontal strike-slip fault generates essentially no vertical displacement of water.
Moreover, although preliminary studies of the Sulawesi event demonstrate some vertical component of fault movement, this component accounts only for some 50 cm of seafloor displacement—a relatively small amount that should normally result in a maximum tsunami height of less than one meter. Tsunami modelling indicates that this is still nowhere near enough displacement to produce the actual recorded height of the Sulawesi tsunami wave. One suggested explanation for this discrepancy in modelled versus observed tsunami wave heights is that the concave topography of the bay leading to Palu may have significantly amplified the wave. However, most geologists now believe that additional elements contributed to the tsunami wave generation.
Since the origin of the September 28th Sulawesi earthquake was first established, tsunami experts have entertained the idea that concomitant submarine landslides may be the most plausible explanation for the larger-than-expected tsunami waves. There is strong digital evidence in favor of this hypothesis. In a series of posts on his landslide blog, Dave Petley provides an impressive collection of eyewitness videos and satellite imagery demonstrating slope failures near and at Palu during the event, with several slides depositing directly into the sea and generating waves.
Twitter user Gerry Soejatman, shares a remarkable video taken from an aircraft that left Palu just moments before the earthquake happened. In it, it is possible to directly view the generation of a local tsunami from a coastal landslide.
Another video filmed onboard a boat in the gulf just North of Palu depicts the same event. Austin Elliott has somehow tied down the location of this boat on his Twitter account as -0.830349, 119.813329.
This second video shows the generation of a tsunami by what it looks to be a very large landslide. As the camera turns, it is possible to see the generation of a second tsunami (possibly spawned by a second landslide event), and then a third one.
These videos are probably the most direct and detailed observation of tsunamis generated by coastal landslides during an earthquake ever recorded. The relationship between the waves and landslide events shown in the video seems to be straightforward; however, future offshore and onshore surveys of the area will help reveal whether one, large, underwater slump is primarily responsible for the main tsunami generation and associated fatalities, or whether a contribution of several localized landslides and tsunamis are to blame.
In addition to direct landslide and tsunami footage, an attempt by Prasetya et al. (2001) to model historical earthquakes and tsunamis generated along the Palu-Koro fault demonstrates that faulting alone cannot possibly explain the similarly-high tsunami runup heights observed in past events in the region. Nearly two decades earlier, they too proposed, “submarine slumping as the secondary mechanism for tsunami generation” in the region, explaining that a seismic profile of the southern Makassar Strait indicates “slumping material from the east (Sulawesi side)”.
Although images of the tsunami have attracted considerable global attention, three very large landslides in Palu are most responsible for the earthquake’s high number of fatalities. There, landslides were widely documented by locals, and satellite imagery has enabled an initial quantitative estimation of the slide volumes and dynamics involved. This YouTube video puts together some of the most fascinating and tragic images relating to the slides.
Besides three major landslides that affected the city of Palu, a number of smaller events are visible everywhere in the proximity of the fault trace. However, an extreme mobility of sediments often associated with liquefaction processes characterizes the events in the Balaroa district of Palu. This does not come as a big surprise: the sandy-silty composition of the soil and the elevated condition of sediment saturation, during Sulawesi’s rainy season, makes the slopes vulnerable to failure during strong shaking. Field data, lab testing, and geotechnical modelling over the next months will provide further details on why the slopes in the city of Palu initiated flow under the effect of the earthquake shaking. A visually impressive article from Reuters provides a decent explanation of the extent of the landslide damage.
A terrifying combination of earthquake, tsunami, and landslide disasters is ultimately responsible for the unusually high number of victims. CNN released a series of before-and-after earthquake images of the Palu area that demonstrate the impact of the disaster. Although each of these events transpired rapidly, their effects will last a lifetime for the people who were in harm’s way. To paraphrase Dave Petley, we must consider that landslide disasters like these may be far worse than normal building collapses due to earthquakes. In the latter case, people may lose houses and possessions; while in the case of such a catastrophic combination of events, they may even lose their lands.
One of the primary concerns for rebuilding in this area—a current subject of debate among landslide experts—is that at least some of the landslides are thought to have failed along layers of quick-clay, or layers of fine marine or fluvial-deposited mud comprised of clay minerals. If proven the case, then it may be impractical and unsafe for locals to rebuild in the same area, as they will risk similar catastrophic failures due to shaking in the future.
Tsunami early-warning system failures and a better approach
Now, a month later, the most important question scientists, policy-makers, and the press are asking is “how can a similar catastrophe be avoided?” Although much will be said, there are few easy and cost-effective answers, especially for Indonesia.
However, there is at least one thing Indonesians should do to avoid tsunami fatalities: they should decrease their dependency on “top down” tsunami warnings (i.e. government alerts based on faulty tsunami early warning systems). According to Hall et al. (2017) , locals should learn to recognize the natural warning sign of a tsunami: the earthquake.
Since its creation, and, indeed, during this recent event in Sulawesi, Indonesia’s tsunami early-warning system has repeatedly failed to properly alert locals to a possible tsunami threat. A variety of technical problems (i.e damaged or stolen tsunami buoys, broken tsunami sirens, downed phone lines, etc…)—not, necessarily, the actions of the BMKG—are to blame. However, even when Indonesia’s tsunami early-warning system does work correctly, it is completely tectonic-based. Thus, depending on the earthquake’s proximity to shore, or whether a landslide generated the tsunami (as may be the case in Sulawesi), it may not always provide sufficient warning time to flee to high ground.
The best solution to this problem is to empower individuals. Everyone should be taught that the earthquake is the warning: if the ground shakes for ~20 seconds or more, regardless of the intensity, then people should immediately seek high ground without waiting for a warning from the government. However, they should maintain a safe elevation until they receive notice to return, as tsunami wave trains may last for several hours.