CALIFORNIA INSTITUTE OF TECHNOLOGY
Despite climate change being most obvious to people as unseasonably warm winter days or melting glaciers, as much as 95 percent of the extra heat trapped on Earth by greenhouse gases is held in the world’s oceans. For that reason, monitoring the temperature of ocean waters has been a priority for climate scientists, and now Caltech researchers have discovered that seismic rumblings on the seafloor can provide them with another tool for doing that.
In a new paper publishing in Science, the researchers show how they are able to make use of existing seismic monitoring equipment, as well as historic seismic data, to determine how much the temperature of the earth’s oceans has changed and continues changing, even at depths that are normally out of the reach of conventional tools.
They do this by listening for the sounds from the many earthquakes that regularly occur under the ocean, says Jörn Callies, assistant professor of environmental science and engineering at Caltech and study co-author. Callies says these earthquake sounds are powerful and travel long distances through the ocean without significantly weakening, which makes them easy to monitor.
Wenbo Wu, postdoctoral scholar in geophysics and lead author of the paper, explains that when an earthquake happens under the ocean, most of its energy travels through the earth, but a portion of that energy is transmitted into the water as sound. These sound waves propagate outward from the quake’s epicenter just like seismic waves that travel through the ground, but the sound waves move at a much slower speed. As a result, ground waves will arrive at a seismic monitoring station first, followed by the sound waves, which will appear as a secondary signal of the same event. The effect is roughly similar to how you can often see the flash from lightning seconds before you hear its thunder.
“These sound waves in the ocean can be clearly recorded by seismometers at a much longer distance than thunder — from thousands of kilometers away,” Wu says. “Interestingly, they are even ‘louder’ than the vibrations traveling deep in the solid Earth, which are more widely used by seismologists.”
The speed of sound in water increases as the water’s temperature rises, so, the team realized, the length of time it takes a sound to travel a given distance in the ocean can be used to deduce the water’s temperature.
“The key is that we use repeating earthquakes–earthquakes that happen again and again in the same place,” he says. “In this example we’re looking at earthquakes that occur off Sumatra in Indonesia, and we measure when they arrive in the central Indian ocean. It takes about a half hour for them to travel that distance, with water temperature causing about one-tenth-of-a second difference. It’s a very small fractional change, but we can measure it.”
Wu adds that because they are using a seismometer that has been in the same location in the central Indian Ocean since 2004, they can look back at the data it collected each time an earthquake occurred in Sumatra, for example, and thus determine the temperature of the ocean at that same time.
“We are using small earthquakes that are too small to cause any damage or even be felt by humans at all,” Wu says. “But the seismometer can detect them from great distances , thus allowing us to monitor large-scale ocean temperature changes on a particular path in one measurement.”
Callies says the data they have analyzed confirm that the Indian Ocean has been warming, as other data collected through other methods have indicated, but that it might be warming even faster than previously estimated.
“The ocean plays a key role in the rate that the climate is changing,” he says. “The ocean is the main reservoir of energy in the climate system, and the deep ocean in particular is important to monitor. One advantage of our method is that the sound waves sample depths below 2,000 meters, where there are very few conventional measurements.”
Depending on which set of previous data they compare their results to, ocean warming appears to be as much as 69 percent greater than had been believed. However, Callies cautions against drawing any immediate conclusions, as more data need to be collected and analyzed.
Because undersea earthquakes happen all over the world, Callies says it should be possible to expand the system he and his fellow researchers developed so that it can monitor water temperatures in all of the oceans. Wu adds that because the technique makes use of existing infrastructure and equipment, it is relatively low-cost.
“We think we can do this in a lot of other regions,” Callies says. “And by doing this, we hope to contribute to the data about how our oceans are warming.”
###
The paper describing the research, titled, “Seismic Ocean Thermometry,” appears in the September 18 issue of Science. Co-authors are Wenbo Wu, postdoctoral scholar in geophysics; Zhongwen Zhan (PhD ’13), assistant professor of geophysics; and Shirui Peng, graduate student in environmental science and engineering, all from Caltech; and Sidao Ni (MS ’98, PhD ’01) of the Institute of Geodesy and Geophysics at the Chinese Academy of Sciences.
Isn’t there a delay between the earthquake starting – and the monitoring station being told the clock has started?
The technique is largely nonsense. The speed of sound in water varies with both temperature and salinity. Without knowing salinity (which varies above 700 meters depth per Argo depending on surface rainfall—in fact Argo salinity probes are calibrated at the 1000 meter drift depth) you cannot infer temperature. Only if the seismometer acoustics were placed below 700 meters depth would the technique have any validity.
Rud
They seem to be putting their chips on a depth of greater than 2,000 meters.
“Depending on which set of previous data they compare their results to …” says it all.
In the 70s I was employed with a marine seismic survey company that used Shoran (and similar systems) as a navigation tool for dynamically locating survey ships as they undertook grid surveys offshore. Atmospheric propagation properties could easily change often depending on time of day weather and local flora – tropical forests sucked the hell out of signals! So I’m sure, while this method in principle is sound, there are too many unmeasurable variables. One being how can they be sure that an earthquake occurring in the same location several times over a period is actually has the exact same focus as it’s previous iterations. I would doubt that after the strata have settled after one earthquake the next pressure/fracture point is unlikely to be in the same spot. Then you have the intervening ocean environment: changes in salinity, depth, pressure, temperature, flora and fauna. Not a happy recipe for success.
From the above article: “. . . as much as 95 percent of the extra heat trapped on Earth by greenhouse gases is held in the world’s oceans.”
Ummm . . . I guess you are referring to that net 0.9 W/m^2 that suddenly appeared in the Kiel & Trenberth 2008-2009 revision of their diagram of Earth’s “energy budget”, you know, the one that shows the balances of W/m^2 power fluxes (not energy flows). This is also the 0.9 W/m^2 that so alarms Bill McKibben (he assess it to be “about three-fourths of a watt”; see https://wattsupwiththat.com/2020/09/08/bill-mckibben-talks-up-the-alleged-climate-change-extreme-weather-link/ ).
So, the above two paragraphs, taken together, leave only about .05*0.9 W/m^2 = 45 milliwatts/m^2 “extra heat” to warm Earth’s total land surfaces and total atmosphere. And—need it be mentioned?—this would assume one can track the W/m^2 fluxes to a precision of .045 W/m^2 out of 341 W/m^2 incoming, or about 1 part in 10,000.
Is it any wonder that I’ve started to feel cold lately?
“So, the above two paragraphs, taken together, leave only about .05*0.9 W/m^2 = 45 milliwatts/m^2 “extra heat” to warm Earth’s total land surfaces and total atmosphere. ”
And don’t imagine Bill McKibben including Geothermal heat.
Feeling even cooler.
But let’s face it, 15 C is chilly.
And we in an Ice Age- which has more temperature extremes.
Can only hope that it gets warmer.
This is very interesting indeed.
A new fully independent way to measure ocean temperatures in 3D? This could cause some serious fear and insecurity among the activist-custodians of current ocean temperature data. It could put Josh Willis into prison for example.
Somehow I suspect it will be like the CO2 satellite measurements. The results will not be kosher so it will fade from view. Sad but true.
The traditional thinking is that the visible and shorter wavelengths penetrate much much further and cause more heating. Near and far IR from the sun may heat the very top inches but that is all. Back radiation from CO2 is miniscule compared to the sun.
What this really means is that for detectable temperature increases, CO2 is not the issue, sunlight or some other effect is.
The heat capacity of the entire Earth’s atmosphere is about 1/1000 the heat capacity of the global ocean.
A column atmosphere of 1 square meter area will contain about 10000 kilograms.
A column of seawater with the same mass will extend about 10 meters down from the surface.
The specific heat capacity of seawater is about 4 times the air above.
So the heat capacity of the entire atmosphere above the ocean is matched by seawater to a depth of 2.5 meters.
Taking the actual global masses of atmosphere vs ocean, the oceans will have 1000 times the heat capacity.
If you added enough energy to heat the atmosphere by 1K, the same amount of energy would increase the ocean temp by 0.001 K.
For this to work the velocity of propagation of the earthquake energy through the earth would have to be a constant. If this is not the case, then the exact time that the earthquake event took place is unknown thus affecting the difference in time between the arrival of the seismic energy in the earth as against in the water. I do not know if seismic energy velocity through the earth is constant but it would seem unlikely given the different densities of strata through which it must travel.
Peter
Seismic velocity through rocks is NOT constant, varying with the type of rock, and in the case of sedimentary rocks, with the compaction or induration. In general, velocity increases with density, meaning the velocity increases with depth in the mantle.
Rough trend of 0.035K per decade which would be like 0.35K from year 2000 to 2100. But that’s just a rough visual estimate not accurate because I don’t have the full print or data.
The time period and clock accuracy etc. would also affect the accuracy.
The location of the center of the earthquake must be known with high precision for this method to work. Since this is done by triangulation using several seismic signals, there is some uncertainty in calculating the center. Then the sound wave travels through the rock for some distance, and then through the water. The depth below the sea floor must be known as well as the path of the sound wave so the differing velocities can be accounted for. Water depth and salinity also affect the accuracy. Did they do some error analysis?
“as much as 95 percent of the extra heat trapped on Earth by greenhouse gases is held in the world’s oceans.”
This is utter garbage. IR can not penetrate water, the oceans can not absorb global warming. The oceans can not hinder CO2s ability to t warm the troposphere.
Come on, we all know this, this has been discussed many times, even Real Climate did a piece on it.
Basic problem of the method – first arrival times from earthquake to sensor will be through rocks not through the water. This is because the water velocity is about 1500 m/s whereas typical crustal rocks lie in the range 2500 (shallow) to about 3500 m/s (deep sediments). Limestones and evaporites are much faster, as are crystalline rocks, some approaching 6,000 m/s. The first arrival is always through the fastest medium (obviously).
In order to reliably measure an arrival through water from an earthquake you first need to establish the distance to the quake – this only comes approximately from triangulation. Then you need to identify and observe the water-borne signal. That will never be a water only arrival as earthquakes are in rocks, so some of the signal will always be through the rock medium. Identifying the water borne signal within the seismogram will also be difficult. Finally without a repeatable source in a fixed location over many years you need the point of origin to get the distance in order to compute the velocity from v = d/t. You also need to be able to apply tomography to invert the travel time data to velocity – again you need the distances and a very large number of (accurate) travel time arrivals.
Complete dud as a method. You could only (possibly) make this work to detect velocity changes within the ocean with controlled sources in the ocean, permanently deployed. Even then, the influence of temp on velocity is pretty small – other factors such as salinity will be far more important (see Batzle and Wang, 1992 for the physics).
Its a dud. Chasing grant money is what it smells like.
See also velocity calculations above.
Source to recording instrument one would clearly be to imprecise.
Yet us that what they are doing?
Would it not be more logical to measure time between recording instrument one, in the water, and recording instrument two? ( And are those locations fixed )
To get long enough travel distances would need really big charges in precisely known positions. Nuclear depth charges would be perfect.
“as much as 95 percent of the extra heat trapped on earth by greenhouse gases is held in the world’s oceans”
According to an Open University (UK) book I have on Seawater :Its composition, properties and behaviour
“The long- term stability of the distribution of temperature within the ocean means that sections and profiles of average temperature do not change significantly from year to year. This stable thermal structure is maintained by the continuous three-dimensional motion of the global system of surface and deep currents”
Another factor nobody seems to have mentioned: doppler effect. The travel time is c. 2200 seconds, the difference about 0.1 second, equal to 1/22000 of 3300 km = c. 150 meters distance, so an uncetainty in the location of the eathquake of 150 meters is enough to wipe out any temperature signal.
But this distance uncertainty could also be due to ocean currents. For the travel distance to change 150 metes over a 37 minute period only requires that the average velocity vector of the water along the propagation path changes by 0.13 knots. And note that this was in the Indian Ocean where ocean currents vary strongly during the year and between years due to the monsoon.
To summarize, the travel time is affected by:
1. The exact position of the quake (which will not be the same for different quakes)
2. The depth of the quake (dito)
3. The structure of the rocks a) along the path from the quake up to the ocean bottom and b) along the underground path (dito)
4. The salinity of the ocean along the propagation path (dito)
5. The movement of the ocean water along the propagation path (dito)
6. The temperature of the seawater along the propagation path.
To measure (6), (1) through (5) must be known with extremely high precision.
And 6 Turbidity, which has a seasonal element.
It seems like a nice idea, earthquakes a decently known distance away that can be triangulated with other seimometers. Measure the difference in time between the rock wave and the water wave which is easy with just a quartz crystal controlled timer, assuming the data storage has a resolution of a hundredth of a second or better.
The real problem is going to be the signal path. The path through the rock should be constant barring any major seismological event between the source and seismometer. But the path through the water is questionable. The first thing I would look for is the El Nino, La Nina Pacific oscillation in the data. I would also want to see data on the depth changes of the thermocline from the Argus floats.
If the water wave path is deep water then this might show a temperature rise in the deep ocean. It should still be rising from the little ice age.
” It should still be rising from the little ice age.”
Or it could be sinking as water from the MWP upwells and is replaced by LIA water. The turnover time for the deep sea is 800-1,600 years, it varies geographically.
The proposition assumes a point source where P waves convert to sound/water waves.
The rupture zone/source for the 2004 Sumatra earthquake was perhaps 1400km long.