One of the many tasks required of astronauts during the six successful moon landings was the positioning of lunar seismometers to probe the interior of the Moon.
These seismometers measured tiny shifts in the lunar surface, moon-quakes, and meteorite impacts until around 1977. Using this data, researchers learned a lot about the composition of the moon, but not as much as they would have liked. Seismologists analyse seismic waves travelling through the Earth (or in this case, the Moon) to probe its inner structures. These waves can either be natural, as from an earthquake, or artificial (seismologists detonate small charges under the surface or fire slugs into the ground for this purpose). As the waves propagate through different layers of rock or magma, they speed up or slow down, or even reflect, according to the properties of that layer. By monitoring the waves with seismometers placed in strategic locations, they can deduce what the wave must have passed through.
Unfortunately for lunar seismologists, the data recovered was very “noisy” – there were so many spurious and complex signals that it was extremely difficult to tease out the original signal. Part of the problem was the rugged broken surface of the Moon, causing the seismic waves to reflect repeatedly in unpredictable and chaotic ways.
But now a team of researches have gone back to the old data and applied the very newest techniques used in terrestrial seismology to clean up the signal. In a process called Array Stacking, they were able to combine the same data from multiple sources to smooth out the discrepancies. In this way, the extremely faint signals could be amplified while leaving the noise behind. The result of their work? The moon has a hard iron-rich core about 660km in diameter (roughly a fifth of the Moons size), surrounded by a 90km thick semi-molten shell (similar to the Earth’s inner and outer core). Many researchers had previously inferred a similar structure before, but since they lacked hard data they could never be certain. Finally, 40 years after the data was first collected, we have the answer.