Tuesday, March 17, 2009

Beach Ridges, Swamps and Tsunami

Last year, geologist Katrin Monecke and co-authors published a paper in Nature that showed the last tsunamis in Aceh occurred in 1907, ~600 years ago, and ~1000 years ago (Monecke et al., A 1,000-year sediment record of tsunami recurrence in northern Sumatra, Nature 455, 30 October 2008).

One of the areas described in the paper is shown above. This is a landscape of beach ridges separated by low-lying, swampy swales. It is in these swales that the paleotsunami deposits are found. In March 2009, we went back into the field to collect data to help determine how these ridges were formed.

One of the interesting aspects of these deposits in this area is that the oldest deposits (from 1000 years ago) occur only in the swales furthest from the modern beach. So, where was the beach when these sediment were deposited? How much has the coastline grown outwards since their deposition?

To answer these questions, we will attempt to date the ridges using thermolumniescence. Some minerals, when exposed to light, store heat in their crystal structures. When these samples are heated again, they emit light (luminesce) in a way that is related to how long ago they were removed from the light source. Really smart people can then determine the age of a sample based on the light the mineral emit when heated. Pretty cool!

We collected several samples from a village north of Meulaboh, NAD (Aceh Province), to see if this method is feasible here. If so, a group of researchers from the Netherlands will come back and do a full dating campaign, and by determining the timing of ridge formation, we can possible determine how the ridge was formed.

When you see a repeating pattern in nature, such as the ridges pictured above, there is often a recurring event that is responsible for their formation. Terraces step down the mountains around Santa Cruz, California, that we formed by changes in sea level. Rings around corals in the islands offshore Sumatra are formed when earthquakes lift the coral partially out of the water, killing the top animals of the colony. Following the 2004 earthquake, the coastline in this area subsided several tens of centimeters, so it begs the question- Could subduction zone earthquakes somehow be responsible for beach ridge formation?

Doubt it.

James Goff, director of the Australian Tsunami Research Centre, showed that in New Zealand, beach ridges are formed when pulses of sediment are delivered to the coast following massive landsliding triggered by earthquakes. There was not widespread landsliding following the 2004 earthquake, so it is unlikely that there would be the sediment available to create another ridge. Furthermore, as the coast subsided, a lot of land was actually lost, presumably to the offshore. Please, prove me wrong- it would be really helpful if these ridges were actually related to the subduction zone earthquakes!

It might be that the beach ridges are formed following earthquakes on the Great Sumatra Fault- the big strike-slip fault that runs through the Sumatra highlands. If the dates from the ridges correlate with this earthquake history, then it is likely that landslides in the highlands provided the sediment for their formation, and that they are just a convenient "catcher's mit" for tsunami sediment.

Or, perhaps they were formed by volcanic activity, as suggested by Stanford geology student Mindi Summers in her senior thesis on beach ridges in New Zealand. Another day....

Sunday, March 1, 2009

Offshore Landslides and Tsunami

Could this landslide from the continental slope offshore Oregon generate a tsunami?

The answer is simple. Maybe...

It depends on several factors. First off, is it big enough? Tsunamis are generated by rapide movements in lots of water, be it from earthquakes, asteroids, volcanoes or landslides. A pebble thrown into the ocean is certainly no asteroid.

And while even a small event can generate a very small tsunami, it depends on the location of whatever it was that moved the water as to whether the tsunami will be noticed or not. Landslides, volcanoes and even asteroids are "point-sources", while and earthquake moves a very large area of seafloor (hence a lot of water). Waves generated by point source attenuate or decay much faster than those generated by a lifting up a vast area of seafloor. In other words, it is very difficult for a tsunami generated by a landslide, volcano, or asteroid to make it all the way across an ocean.

Yet another complication, as with many things, is timing. Most very large tsunamis are very old- in nature, big events tend to happen infrequently. Should very large offshore landslides occur as frequently as large-magnitude earthquakes, the continental slope would quickly eat its way onto land! So, while technology has allowed us to image all sorts of amazing landslides on the world's continental slopes, many of which appear very young, they tend to be old, and pose little danger to communities on shore.

(One might ask then, when do they occur? Most offshore landslides occur either during sea level lowstands or during times of changing sea level. During glacial periods, sea level is lower, and rivers dump heaps of sediment directly on the continental slope, providing fuel for landslides. During highstands, like now, sediment is sequestered in estuaries and in foreland basins on the continental shelf. If earthquakes trigger landslides, then the hypothetical first earthquake after a sea level lowstand would shed all the sediment off, leaving nothing left to fail! But, this is just a digression by this marine geologist....)

Perhaps the last unconstrained variable (i.e. something we have no idea of!) is once a landslide begins, how fast does it actually fail? Most models that predict tsunamis assume that the landslides are frictionless- that is, once they get going, there is nothing slowing them down. This isn't likely the case, but nonetheless, landslides occur offshore, and they generate tsunamis.

There are two clear examples of offshore landslides that have generated tsunami. The first (that I know of) was 7,200 years ago offshore Norway called Storegga. Following the last sea level lowstand, glaciers provided sediment to the continental slope, and as the land rebounded from having the weight of the glaciers released, earthquakes fired off, shaking loose the sediment. The failure generated a tsunami that laid a deposit in many places surrounding Norway, including underneath the greens at the famous golf course at St. Andrews.

The second example was much more recent- in 1929 a magnitude 7.2 earthquake triggered a landslide offshore Newfoundland, Canada in the Grand Banks. We know there was a landslide because it severed submarine communication cables going between North American and England. Not long after the earthquake, a 7.5 m tsunami hit Newfoundland's Burin Peninsula, killing 28 people in Canada's worst earthquake-related disaster.

So, the most important thing we can do to plan for landslide generated tsunami is to constrain the properties that trigger landslides. What is the strength of the sediment in question? Has there been landslides in this region in the past? What is the frequency and magnitude of earthquakes in the region? Is there a weak layer that might provide a sliding surface for a failure?

To address these questions, geotechnical engineers, geologists, geophysicists and seismologists must come together to address this complicated, technical question. Not one of these scientists can address this working in a vacuum, nor does any one of these possess the necessary knowledge and skills to really be effective. Yet a community working together, making the necessary data and information freely available, can make advances towards protecting life and property in leaps and bounds. A kind of scientific open-sourcing.