Enlarge /. A seismometer on the atoll of Diego Garcia (left) can calculate the sea temperature during earthquakes near Sumatra (right).
Geophysics has shown that precise measurements and a little modeling can work wonders, such as the detailed structure of the Earth's interior, despite being inaccessible buried under hundreds of kilometers of rock. This is possible because seismic waves generated by earthquakes subtly change speed or direction as they pass through different materials. A new paper shows that something similar can actually measure small temperature changes in the deep sea.
The idea of using sound waves from artificial sources actually started circulating a few decades ago, but died out after a few attempts. A team led by Wenbo Wu at the University of Toronto realized that earthquakes could be exploited in the same way by removing the expensive logistics of constantly triggering booms for measurements, as well as concerns about the impact on marine life.
There are actually different types of seismic waves caused by earthquakes, and each one behaves slightly differently. The P wave (P for "primary" because it is the first to arrive) is analogous to a sound wave in that it compresses the rock in the same direction as it is moving. Where this wave reaches the sea floor or the ground surface, the rock can act like a massive loudspeaker and generate a very low frequency sound wave in the air or in the water.
The same process can be reversed. Suppose there is an earthquake under the ocean floor just offshore. The movement of the sea floor can create a sound wave that moves across a sea basin and re-penetrates the sea floor rocks of an opposite coast. This can actually be recognized by seismometers as a vibration that arrives extremely late – as the wave moves much slower through water than through rock. As becomes apparent after the primary and secondary seismic waves, this strange acoustic straggler is known as the tertiary or T-wave.
T-wave cruising speed is sensitive to water temperature, with warmer water slowing it down. (To make matters worse, it is also slightly sensitive to changes in salt concentration or current movement. However, the researchers say that the temperature effect dominates.) Seismometers are sensitive enough to detect very small differences in time, so changes in much are measured can be less than 1 ° C.
Of course, to calculate a change you need at least two measurements. That means you need an earthquake almost identical to an earlier one – what the researchers call a "repeater". However, it doesn't require these earthquakes to be large, so it's not as difficult as you might think.
It's not an atoll problem
To show that this works, the researchers used a seismometer station on Diego Garcia, a small atoll in the Indian Ocean about 3,000 kilometers from Sumatra. The tectonic plate boundary there is incredibly active, so there is no shortage of earthquakes to work with. Between 2004 and 2016, over 4,000 earthquakes of magnitude 3.0 or greater occurred near the island of Nias in Sumatra. The researchers carefully processed all of these events to find repeaters that are similar enough for the temperature calculation. They found over 2,000 such pairs based on 900 earthquakes.
Enlarge /. Here's an example of two nearly identical earthquakes, but it took the T waves a little longer to arrive in the second because the ocean was warmer.
If this part of the Indian Ocean warmed by 1 ° C, T waves from these earthquakes would take 5.4 seconds longer to reach this seismometer. The changes observed are smaller but coherent – there is both an annual cycle and a gradual warming trend that is similar to other, more traditional, data sets.
The trend calculated by this “seismic ocean thermometry” is, however, somewhat larger. The researchers compare with estimates from the automated Argo float array and with NASA's ECCO dataset, which combines data from various sources. Over the same period, Argo in this area shows a warming trend of 0.026 ° C per decade, while ECCO shows 0.039 ° C per decade. The seismic estimate is 0.044 ° C per decade.
Enlarge /. So, the seismic temperature record (blue) is aligned with two sets of data based on measurements of things like swimmers and satellites.
The point here is not to declare an estimate to be the best. The point is to demonstrate that the technology is plausibly useful. Measuring the change in temperature in a completely independent manner is valuable in itself. And instead of relying on a limited number of point measurements by a swimmer, for example, it integrates the average temperature change over the entire volume of water. This is unique.
It also measures a deeper part of the ocean that is harder to access, and that depth can be somewhat adjustable depending on the specific frequency of the T-wave you are working with.
The researchers mention that underwater microphones (hydrophones) could be even more sensitive if smaller earthquakes were exploited. (There are already a number of hydrophones in existence in support of contracts banning nuclear testing.) Most interesting, however, is that existing seismometer data could be captured using this technique. So with a pegged method, it is possible to look for historical measurements, not just future ones. Retrieving data without providing a set of new tools is usually much cheaper than providing a set of new tools.
As it turns out, if your instruments are precise enough and you know how to use them, physics can tell you a lot.
Science, 2020. DOI: 10.1126 / science.abb9519 (About DOIs).