A rise in average global temperatures causes an increase in the time it takes the Earth to rotate according to a new paper. The finding could solve a 20-year-old mystery known as Munk’s Enigma.

It’s already established that the length of a day increases over time because the moon is gradually getting further away. That affects tidal patterns and in turn causes friction on the floor of the seas which slows down the Earth’s rotation. It’s this pattern that means we occasionally have to add a leap second at the end of June or December.

But while the moon is having this effect over time, it’s also believed that the rate of this effect varies in relation to sea levels. In 2002, oceanographer Walter Munk tried to explain the connection between the two but ran into a mystery.

Munk’s figures showed that the Earth’s rotation slows over time and that sea levels changed over time. He could account for a specific relationship between the two going back to the Ice Age, but couldn’t make the figures match up for the 20th century, despite knowledge that increased global temperatures were melting glaciers and in turn causing sea levels to rise.

The mystery was so profound that Munk even developed a theory that melting glaciers might in fact speed up the Earth’s rotation, thus shortening days. He pondered the idea that with the ice melting, the rock beneath could “spring up” and thus add more weight near the surface of the poles. However, Munk openly admitted this idea was uncertain.

Now six researchers led by Harvard geophysics professor Jerry Mitrovica believe they have solved Munk’s Enigma. They’ve produced a paper arguing that Munk made three errors which could have prevented him from establishing the expected “higher sea levels/slower rotation” relationship as having continued throughout the 20th century.

They believe:

1. The figures Munk used for the sea level rise in the 20th century were too big, thus distorting the relationship with day length.
2. The figures Munk used for the Ice Age were incorrect, partly because they didn’t account for the Earth not being a perfect sphere and partly because they didn’t take account of just how much glaciers can deform the rock beneath them,
3. Munk’s calculations didn’t take account of the fact that the Earth has a liquid core which “spins” in the opposite direction to the planet and thus slows down its overall rotation.

According to the researchers, adjusting for these three errors makes it possible to establish a model for a constant sea level-Earth rotation relationship that holds true for both the Ice Age and 20th century and, in turn, allows more confident predictions about future effects.

However, Willieam Peltier, a University of Toronto physics professor quoted by the Washington Post, disputes the findings of the paper. He argues there’s no proof that the liquid core has the effect the researchers assumed when producing their calculations.