Parts 1 and 2 of this series discuss the rocks that rise from the mantle into our hands. They are valuable for science, of course, but of limited use for picturing the whole mantle in any detail. Here I'll talk about how we're creating images of the mantle the same way we use medical ultrasound to create images of unborn children.
The xenoliths collected from kimberlites and lamproites (see Part 1) and other ultrahigh-pressure rocks tell us about conditions as deep as 300 kilometers beneath the oldest continents. That sounds pretty good, but the mantle is 3,000 km deep, and the most ancient continents cover only about 10 percent of the Earth. Hotspots and mantle plumes (see Part 2) are thought to boil up all the way through the mantle, but the rocks that result are so mixed and altered during this slow journey that their stories are almost unreadable. Moreover, we have to assume that they are unusual and not typical. So to really get a global picture of the mantle, we have to give up on sampling it—and listen to it instead.
Listening to the Mantle
What we listen to is earthquakes—seismic waves. Think of seismometers as stethoscopes laid upon the Earth, like those a doctor lays on a patient's body. If you put a seismometer on the exact opposite side of the Earth from an earthquake, it detects seismic waves from that quake that have traveled straight to it through the mantle and the central core. If the seismometer is close to the quake instead, less than maybe 200 km away, it feels waves that traveled through just the crust. At in-between distances, seismic waves travel through various parts of the upper and lower mantle and upper core.
By close study of many earthquake records from around the world, we can map out the parts of the Earth's interior where seismic waves arrive sooner or later than average. Those that come sooner travel faster. Those that come later are slowed down by something along the way. In scientific shorthand, the variations in travel time reflect variations in seismic velocity. That's the basis of seismic "ultrasound imaging," because many specific parts of the mantle (and core) have different seismic velocities. Once you know how to pinpoint those places, you can turn earthquake data into pictures of the Earth's interior.
There are two simple complications. First, seismic waves don't move in a straight line through the Earth but instead bend, or refract, in response to changes in rock density. Second, seismic waves can bounce, or reflect, off of sharp density boundaries. One of these is the Earth's surface, the boundary between the crust and the atmosphere. Another one, even sharper than that, is between the mantle and the core. And there are other lesser ones.
In exactly the same way, doctors send high-frequency sound through a mother's belly to map out the density boundaries between bones and skin and organs to take a baby picture. The image is actually a tomogram ("slice picture"), a cross-section of the mother and child. The same procedure on the Earth is called seismic tomography.
Mantle Maps
We have a century worth of seismograms by now, and lately our maps of the mantle are getting very interesting. Adam Dziewonski, of Harvard University, invented seismic tomography. A set of typical mantle maps is shown in this article from GSA Today.
These are weird maps compared to the ordinary maps we know. Not only are they crude and blobby-looking, but they are maps of seismic velocity and nothing else. Moreover, maps based on P-waves differ from maps based on S-waves. Figuring out what differences in seismic velocity mean in terms of real rocks, with real chemical compositions, is a thorny problem.
In the case of medical ultrasound, we know what's in the human body, and we can easily match up variations in ultrasound velocity with different tissues. In the case of Earth's body, we can barely peek beneath the outermost skin. But we learn more about its organs and bones, as I report in Part 4, by treating rock samples to the tremendous pressures and white heat of the mantle's deepest places.
PS: Explosions work as well as earthquakes for tomographic purposes. Although they have less energy than quakes, explosions can be precisely controlled in time and place. Of course, the bigger the explosion the better. In 1986 some scientists proposed, a little too hopefully, that the U.S. government set up some seismometers on the exact opposite side of the Earth from the Nevada Test Site so as to get some use out of the nuclear bombs being tested there. But as it happens, seismology is part of the nuclear test ban mechanism ensuring that this proposal will never bear fruit.
Next page > The Big Squeeze > Page 4, 5, 6
