When ancient and medieval astronomers dutifully jotted details about lunar and solar eclipses onto clay tablets and parchment, little did they know that information would one day be used to calculate the Earth’s spin – and how it’s slowed over the millennia.
A trio of physicists in the UK gathered records from 720 BCE to 1600 CE which describe partial or total eclipses. From these, they calculated the planet’s rate of rotation at the time.
Writing in the Proceedings of the Royal Society A, they found that an Earth day grew by an average of less than two-thousandths of a second each century.
That the Earth’s spin is slowing isn’t new. It should be, after all – a little of the planet’s rotational momentum is constantly transferred to the moon’s orbital momentum.
This means the moon speeds up slightly and slowly pulls away from Earth – at around four centimetres each year – while the planet takes a little longer to complete one full rotation (or a day).
Richard Stephenson from the University of Durham and Leslie Morrison have examined historical eclipses for clues about the Earth’s rotation for nearly four decades.
Their sources came from civilisations the across the world over nearly 3,000 years.
For instance, from around the eighth century BCE to 100 CE, Babylonian astronomers etched eclipse details in cuneiform script on clay tablets. Chinese dynasty histories contained a special section for astronomical records. Ptolemy’s Almagest recorded the timings for eclipses in ancient Greece. Three medieval Arab astronomers – Ibn Yunus, al-Battani and al-Biruni – kept detailed records from around 830 CE to 1020 CE.
Some of these observations were timed. Those without timing details still served a useful purpose – if the team knew the date of the eclipse and the site of the observations, they could calculate upper and lower time limits.
After 1600 CE, telescope observations provided accurate eclipse information.
Along with Catherine Hohenkerk, Stephenson and Morrison built upon previous work, adding more records and adjusting for those which were incorrect.
Using their own software, they found fluctuations in rotation rate on a scale of around 1,500 years. Averaged, the length of Earth’s day – that is, one rotation – increased at a rate of 1.8 milliseconds per century – somewhat lower than the predicted value of 2.3 milliseconds.