Seismic shadow zone

Why does a large earthquake leave a wide band of the planet where its shear waves never arrive?

A great earthquake shakes the whole planet, and instruments on the far side record it. But not everything arrives everywhere: past a certain angular distance from the epicentre, one of the two kinds of body wave stops turning up, over a vast band of the Earth's surface.

Why a band? Absence over a patch might mean a bad instrument or an unlucky path. Absence over a band the same width every time, for every earthquake, wherever it happens, is a fact about the medium the waves crossed. This is the companion question to why earthquakes happen at all: here the waves are no longer the thing being explained but the instrument doing the explaining.

Begin with the difference between the two waves. A compressional wave, the fast one, works by squeezing and stretching the material — it needs the medium to resist a change of volume, and everything does, including air and water. A shear wave works by sliding neighbouring layers past one another. Ask what it needs, and the answer is rigidity: a material that pushes back when you try to distort its shape without changing its volume.

Which materials do that? A solid does. A fluid does not — displace a layer of water sideways and nothing restores it. So a shear wave cannot propagate through a fluid at all: not attenuated, but with nothing to travel on.

Put those two facts together with the observation. If the deep Earth contained a large fluid region, shear waves from any epicentre would be blocked from reaching whatever part of the surface lies behind it, and since the region is roughly spherical and centred, that blocked area would be a cap of fixed angular size around the point opposite the source. That is exactly what is seen: direct shear waves are absent beyond roughly a hundred and three degrees from the epicentre. The size of the shadow is a measurement of the size of the fluid region, and it puts its boundary at a depth of about two thousand nine hundred kilometres — a figure I am recalling rather than deriving, though it follows from the geometry.

Then a second observation, which looks like a complication and turns out to be the confirmation. The compressional wave has a shadow too, but a different one: a ring between roughly a hundred and three and a hundred and forty-three degrees where the direct wave is weak or absent, beyond which it returns. Why should a wave that can cross a fluid go missing?

Because it does not only cross — it refracts. The compressional speed drops sharply on entering the core, from something like thirteen kilometres a second at the base of the mantle to around eight at the top of the outer core. A wave entering a slower medium bends toward the boundary's normal, so rays that would have emerged in that ring are swung deeper and emerge further round instead. The shadow is deflection, not blockage, and the energy piling up just beyond a hundred and forty-three degrees is where those rays land.

Ask now what would have to be true for this to be wrong. If the shadow were a property of the source — a peculiarity of how faults radiate — it would move with the fault's orientation, not with the epicentre. It does not. If the fluid region were not centred, the shadow would sit at different angular distances for earthquakes in different places. It does not. And a single clean direct shear arrival at a hundred and thirty degrees, from any earthquake anywhere, would end the account. None has been found.

One more thing was found instead: faint compressional arrivals inside the supposedly dark ring, from which Inge Lehmann concluded in 1936 that a further boundary existed — an inner core. That is the shape of the whole method, the structure of the planet read from where signals fail to arrive.

The word "shadow" invites a misconception worth heading off. A shadow zone is not silent. Instruments there record plenty: surface waves that never went near the core, waves reflected off the underside of the surface, diffracted energy creeping around the core boundary. What is absent is the direct arrival at the time straight-line travel would predict. Reading the shadow means reading a specific expected arrival against the clock, not listening for quiet.

That distinction is what makes the evidence robust, and it is worth being explicit about why no single record settles anything. An earthquake does not radiate equally in all directions — there are nodal directions along which shear radiation is nearly nil, so one station can miss a shear wave for reasons that have nothing to do with the core. Instruments fail, and noise buries small arrivals. The claim is not built on any one seismogram; it is built on the systematic pattern across thousands of source-station paths, in which the failures line up by angular distance and by nothing else. A handful of anomalies is variability. A boundary that appears at the same angle for every epicentre on the planet is structure.

Two honest limits. First, the inference from the shear shadow is that the outer core cannot support shear — that is what liquid means here — but the shadow alone does not say what it is made of; the composition comes from other lines of argument, chiefly the density and speed required to match the seismic profile against what iron alloys do at those pressures. Second, the inner core is generally held to be solid, which implies shear waves should travel through it; the phase that would demonstrate this directly has been claimed but is faint and contested, and the stronger evidence comes instead from how the whole Earth rings after a great earthquake, since those free oscillation periods depend on rigidity at depth.

Every term the collection defines is gathered in the glossary.