“Ka mua, ka muri.” We walk backward into the future, with our eyes on the past. This whakataukī (proverb) represents a New Zealand Māori perspective that has much in common with the way Earth scientists study natural hazards. Understanding and learning from historical events inform our preparedness for and increase resilience against future disasters. Studying past tsunami events, for example, is an important part of better understanding the diverse and complex mechanisms of tsunami generation and for improving natural hazard assessments.
Tsunamis are dangerous natural hazards and are most often caused by earthquakes. Consequently, coseismic tsunamis have drawn most of the focus from researchers and hazard planners. However, several recent tsunamis have been attributed to other sources on which less research has been done, including underwater landslides, as in the case of the Palu, Indonesia, event in 2018, and volcanic eruptions in the case of the tsunami originating from Hunga Tonga–Hunga Ha’apai in early 2022.
The Tasman Sea, located between Australia and New Zealand and known for its notorious storms amid the “roaring forties” latitudes, may have witnessed a series of devastating tsunamis during the past 5 million years (i.e., in the Pliocene and Pleistocene, or Plio-Pleistocene, epochs). These tsunamis likely originated near New Zealand’s western coast and traveled more than 2,000 kilometers to also affect Australia, yet intriguingly, there is little easily observable evidence of these events. This tumultuous history is surprising considering that western New Zealand is not especially exposed to subduction zone processes and their associated seismic activity; such exposure is often the main indicator of how vulnerable a coastline is to a tsunami. However, New Zealand is surrounded by steep and, in some cases, tectonically active submarine slopes, where landslides can occur.
In the past few decades, evidence of six giant underwater landslides dating from the Plio-Pleistocene has been discovered beneath the modern seafloor in the eastern Tasman Sea (Figure 1). The most recent, thought to have occurred about 1 million years ago, is the largest documented landslide in New Zealand, covering more than 22,000 square kilometers—an area larger than Wales. With a volume of about 3,700 cubic kilometers, this landslide was bigger than the famous tsunamigenic Storegga Slide, which involved massive collapses of the continental shelf off the coast of Norway roughly 8,200 years ago.
Can scientists use these landslide deposits to derive credible indications of past tsunamis? If so, how can we assess the modern potential for hazardous tsunamis based only on these ancient, buried remnants? Underwater landslides are not comprehensively included as tsunami sources in New Zealand’s hazard assessments. This data gap exists largely because of a lack of research into underwater landslide return rates (a statistical measure of how often these events are likely to recur) and tsunamigenic mechanisms, as well as of uncertainties introduced by errors in available dating methods and the difficulty and expense of obtaining samples. These key questions and issues are currently being addressed by a trans-Tasman team of researchers, including us, from Australia and New Zealand under the Silent Tsunami project (officially named Assessing Risk of Silent Tsunami in the Tasman Sea/Te Tai-o-Rēhua), which began in 2021.
Search Strategy for Landslide Evidence
Throughout the Plio-Pleistocene, a vast volume of material was eroded from the rapidly uplifting Southern Alps, on New Zealand’s South Island, and delivered to the coast by river networks. Powerful ocean currents then transported the sediment north to the country’s northwestern continental margin. The ocean basin duly accommodated the relentless influx of material, and the margin rapidly prograded (built outward toward the sea) via a series of spectacular, steeply dipping depositional surfaces (up to 1,500 meters tall) called sigmoidal clinoforms, which are the building blocks of deltas and basin margins. The unstable sediment piles, perched precariously at the edge of the tectonically hyperactive interface between the Pacific and Australian plates, inevitably then collapsed in catastrophic fashion several times over.
However, evidence of these Tasman Sea landslides cannot be readily observed, in part because of a lack of detailed seafloor mapping in the area but also because the slides were quickly buried under other sediment. Compounding the difficulty are the erosion and uplift of New Zealand’s dynamic coastline, which have erased potential land-based geological evidence in the form of tsunami deposits. Only past seismic reflection surveying in the area enabled the discovery of evidence for these events (Figure 1), with geologists documenting the landslide deposits while mapping New Zealand’s offshore sedimentary basins.
The new project takes a three-pronged approach to carry earlier findings forward. First, we’re combining tools and techniques from the playbook used to analyze the formation and evolution of sedimentary basins, especially how the basin filling process interacts with tectonic processes. These methods include the conversion of time series seismic reflection data into depth measurements using seismic wave velocities measured from drill holes (meaning the depth for each data point is known) and virtually stripping away overlying sediments (back stripping) using computational models. This approach allows us to unearth accurate original volumes (areal extent and thickness) of the landslides before their burial and compaction.
Second, we’re applying these new physical descriptions of landslides, along with knowledge of where they occurred, to inform computational models. The models, run using the cutting-edge fluid dynamics modeling tool Basilisk, simulate landslide motion, tsunami generation, and hazard metrics like inundation extents, wave amplitudes, wave arrival times, and current velocities.
Third, during two research voyages, we have collected new geophysical data—multibeam bathymetry, subbottom profiles, and high-resolution multichannel seismic reflection profiles—and sediment samples from rock dredges and sediment cores from the site of the landslides. Data from the voyages are perhaps most critical to the outcomes of the project. The modeling builds a picture of the likely impacts of the Tasman Sea landslides, but probing the sites of their origin in the real world draws tangible ties between these ancient events and the present day.
So what about the present day? During sea level highstands, when sea levels rise above the edge of a continental shelf, as is the case today, delivery of sediment to the deep ocean is thought to decrease. However, a paucity of information from the Tasman Sea region means that no one knows how much, how fast, and exactly where sediment is accumulating at present. It is not clear whether the conveyor belt of northward sediment delivery is still operating or what could trigger a future landslide event.
In October 2021, on the first of the two research voyages, a small science party of five boarded the R/V Tangaroa for an 11-day voyage to map some 5,000 square kilometers of the Tasman Sea for the first time and to identify targets for a sampling campaign to be conducted during the second voyage (Figure 1). The preexisting seismic reflection data set for the region (Figure 2), comprising data gathered during numerous explorative surveys over several decades, appeared to show evidence of “megablocks” peeking up through the modern seabed from within the most recent Tasman Sea landslide deposit. These megablocks are large clasts or “rafts” of material that were transported within a landslide and that have remained mostly intact. Such blocks often form highly irregular seafloor topography in the immediate aftermath of an underwater landslide and can create localized sediment traps when normal sedimentation resumes. Heading into the voyage, it was uncertain whether these features would be visible or prominent on the seafloor or whether we could identify viable targets for sampling.
True to form for the Tasman Sea, after leaving the shelter of Wellington Harbour, a howling southerly wind and 8-meter swells pummeled our ship during the 20-hour transit. Once on site, however, about 100 kilometers off New Zealand’s North Island, above the continental shelf break and rise, conditions calmed, allowing the ship’s multibeam echo sounders to run—and map the seafloor at high resolution—uninterrupted.
As the data came in, we spent long hours poring over them in the bathymetry lab aboard the Tangaroa. The time was highly rewarding. First came images of canyons, numerous pockmarks, and evidence of recent small-scale slope failures as the ship passed over the shelf break and traversed the continental slope (Figure 3). Then we saw an astonishing area of deep seafloor littered with numerous angular, often elongated ridges and peaks up to 100 meters in relief, some with surrounding “moats” winnowed by the action of recent ocean currents. These ridges are the exposed tops of megablocks from the most recent Tasman Sea landslide, still making their mark on the seafloor roughly 1 million years later.
We decided the megablocks, now that we’d observed them firsthand, were viable targets for rock dredging, offering the tantalizing possibility of sampling landslide material itself. If we could achieve it, this sampling could allow us to characterize the sedimentology and physical properties of the landslide and thus to refine our fluid dynamics models. In addition, areas between blocks would be good targets to sample covering sediments to help constrain the minimum age of the most recent and largest landslide and to determine the rate and patterns of modern sediment accumulation.
We set off on the 3-week-long second voyage, again aboard the Tangaroa, on 15 March 2022. Taking advantage of a spell of calm weather, we deployed the ship’s brand-new 96-channel solid seismic streamer to collect reflection data, then waited anxiously for the first data. Our worry was unnecessary, as the data looked beautiful, with much-improved resolution compared with the preexisting data set.
The biggest highlights from this voyage came as we turned our attention to sediment sampling and targeted several megablocks with the rock dredge. We recovered a lot of sticky mud thought to be the “mud drape” formed by the continuous rain of fine-grained sediment that accumulates normally over many years. We also recovered fist- to paving slab–sized clasts of more consolidated mud and fine sand, which we cautiously assumed to be landslide material.
After deciding to target a flat-topped megablock at roughly 1,500 meters depth for coring, we again waited nervously to see what, if anything, we’d recover. To the team’s excitement, we indeed recovered a 4-meter core from the megablock. Does it contain landslide material, or is it all mud drape? Time will tell. Now back on dry land, we are awaiting the results of nondestructive preliminary scanning before we split and subsample the core to determine in detail what we recovered. In all, 79 meters of core were successfully recovered on the voyage, including from the areas between megablocks, which we are confident will enable us to characterize modern sediment deposition and properties.
From Data to Knowledge to Application
We expect our project to generate new knowledge that builds a picture of modern-day conditions at the site of the Tasman Sea landslides; to refine our understanding of the return rate of large, potentially tsunami-generating landslides; and to develop credible scenarios of the specific hazards related to them. Pathways to assessing the usefulness of the information gained and to guide its uptake in national hazard assessments involve working with a hazard scientists’ advisory group, territorial authorities, and civil defense agencies.
The most likely conduit to implementation in New Zealand is the Review of Tsunami Hazard in New Zealand, a probabilistic risk assessment that quantitively estimates maximum tsunami heights along the country’s coastlines (Figure 4). The model underpins more detailed site-specific hazard assessments and emergency management planning and is continually refined and updated with new information. In Australia, new information from this project could be incorporated into state-based hazard assessments and education programs led by the country’s Emergency Management authorities.
An exciting prospect is the potential to apply the same approaches used in our project to other areas of New Zealand, Australia, and elsewhere. Many of the world’s continental margins have been imaged using seismic reflection surveying—often during exploration for offshore energy resources—creating a vast repository of information about the subsurface. Most of what is known about tsunamis generated by underwater landslides comes from computational models, with few observed examples of such slides to validate them. But existing data sets may hold a wealth of data related to numerous examples of ancient underwater landslides now buried beneath the seafloor.
Translating knowledge from examples of subsurface landslides into information to support hazard assessment is rarely done because of a lack of information on the ages of the landslides and the complexities of assessing their size introduced by their burial, compaction, and incomplete preservation. We hope that results and learning from our early-stage research will help scientists better understand regional tsunami hazards. We also hope that these results will pave the way for future endeavors to develop constructive tools to support refined tsunami hazard assessment and emergency management planning, helping safeguard people around and beyond the Tasman Sea.
The project described above is funded by the New Zealand Ministry of Business, Innovation and Employment Endeavour Fund, with additional support from the New Zealand Strategic Science Investment Fund, the Tangaroa Reference Group, and the University of Newcastle, Australia.
Suzanne Bull (firstname.lastname@example.org), Institute of Geological and Nuclear Sciences Limited (GNS Science), Wellington, New Zealand; Sally J. Watson, National Institute of Water and Atmospheric Research (NIWA), Wellington, New Zealand; Jess Hillman, GNS Science, Wellington, New Zealand; Hannah E. Power, University of Newcastle, Australia; and Lorna J. Strachan, University of Auckland, Auckland, New Zealand