Earth's Spin Is Slowing at a Rate of 1.8 Milliseconds per Century
The latest findings in Earth science are brought to you by ancient astronomers who observed the heavens as much as 2,700 years ago.
Thanks to hundreds of records of lunar and solar eclipses carved in clay tablets and written into dynastic histories, modern scientists have determined that the amount of time it takes for Earth to complete a single rotation on its axis has slowed by 1.8 milliseconds per century, according to a report published Wednesday in Proceedings of the Royal Society A.
It may not sound significant, but over the course of 2 1/2 millenniums, that time discrepancy adds up to about 7 hours.
In other words, if humanity had been measuring time with an atomic clock that started running back in 700 BC, today that clock would read 7 p.m. when the sun is directly overhead rather than noon.
"There is time and then there is how fast the Earth spins," said Duncan Agnew, a geophysicist at Scripps Institution of Oceanography in San Diego, who was not involved with the work. "Traditionally those things are closely linked, but they are not the same."
Our earliest ancestors measured time based on the position of celestial bodies in the sky, such as the rising and setting of the sun or the changing shape of the moon. Scientists refer to this as Universal Time, and it is governed by the dynamic gravitational motions of the Earth, moon and sun.
Terrestrial Time, on the other hand, is measured by clocks and is independent of the laws of physics. Since the 1960s, it has been tracked by exquisitely precise atomic clocks. According to our modern take on Terrestrial Time, there are exactly 86,400 seconds in a day and each second is defined as exactly 9,192,631,770 oscillations of a cesium-133 atom.
But our planet does not keep perfect time, so Universal Time and Terrestrial Time do not always line up.
In our modern world, governed by atomic clocks, the International Earth Rotation and Reference Systems Service calls for a leap second to be added whenever Universal Time is on track to be out of sync with Terrestrial Time by more than 0.9 of a second. Leap seconds traditionally are added on June 30 or Dec. 31. (In fact, one will be added this New Year's Eve.)
The Earth's rotational rate, which determines Universal Time, is affected by many factors.
Large weather systems and atmospheric winds can exert enough force on Earth's surface to cause it to slow down or speed up by thousandths of a second over a single season. Large volcanoes and earthquakes may also cause Earth to speed up or slow down, but these changes are hard to detect. In 2011, a researcher at the Jet Propulsion Laboratory in La Cañada Flintridge calculated that the magnitude 9 quake that struck off Japan may have shaved about 1.8-millionths of a second off the calendar year.
The dominant force affecting Earth's spin rate over longer time periods is the interplay of gravity between the oceans and the moon. Scientist have known for decades that this phenomenon is causing Earth to spin more and more slowly. However, there are other, more subtle factors at play on this timescale as well.
Changes in the amount of polar ice affect the shape of Earth and can impact how fast or slow the planet turns on its axis. In addition, movements of molten rock in the planet's core can affect the speed of its rotation, Agnew explained.
To determine how much all these forces have changed the planet's spin rate over the centuries, a small group of British scientists took on the Herculean task of gathering hundreds of astronomical records made by ancient scribes from around the world.
The team's goal was to pinpoint when and where lunar and solar eclipses occurred in antiquity, then compare them to computer models of when and where they should have occurred based on the current rate of Earth's spin. By measuring the difference between these two sets of data, researchers can determine how much Earth's rotation rate has changed over time.
It wasn't easy to assemble. Richard Stephenson, an astronomer who recently retired from Durham University in England, started working on this project 40 years ago when he was a young researcher at the University of Essex.
Because nearby Durham University had a good Chinese studies department, he began with ancient texts from China. To aid him in his search, he memorized 1,500 Chinese characters -- too few to read a newspaper, but enough to help him decipher astronomical records written into dynastic histories. Over the years he was able to find about 50 reliable dates and times of solar and lunar eclipses recorded between AD 434 and AD 1280.
"During that period they were using water clocks to tell time," he said. These devices work by dripping water into a vessel at a constant rate. "While those are not terribly accurate, if you get 50 observations over a period of time you can get a good average."
A treasure trove of data came from translations of clay tablets written in cuneiform by Babylonian astrologers. Although many of the tablets were damaged in the excavation process, Stephenson found more than 150 useful entries dating from 720 BC to 10 BC.
Other observations came from Islamic astronomers working in the Arab world between AD 800 and 1000, as well as from mentions of solar and lunar eclipses in medieval European chronicles, mostly written in Latin.
"The Arab timing of solar eclipses in particular is superb," Stephenson said. (He could not say the same for the European data.)
Stephenson said he is grateful to all these ancient sky watchers.
"People recording these things never had the slightest notion that what they were doing would lead to people in our generation actually studying changes in the Earth spin," he said. "We are very much at the mercy of these ancient chroniclers and astronomers."
He added that despite the crude instruments many of these observers had to work with, they were often remarkably accurate both in their observations and their ability to keep track of time.
"In 90% of the cases in the Chinese and Babylonian records, their dates reproduce exactly to the date of the calculated eclipse," he said. "This gives us a lot of confidence in our data set."
Stephenson's interest in ancient astronomical observations stops around 1600. That's about when the first telescope was invented and astronomical measurements became significantly more precise.
Luckily, he teamed up long ago with Leslie Morrison, a recently retired astronomer from the Royal Greenwich Observatory in London who has collected thousands of records of lunar occultations from the last 400 years.
Lunar occulations occur when the moon appears to block certain stars from the perspective of an observer.
"After the advent of the telescope and the introduction of pendulum clocks, it was possible to watch the moon approach a star in its path, suddenly block it out, and then see the star reappear again on the other side," Morrison said. "This allowed us to measure difference in time no longer in hours or minutes, but in seconds."
Astronomers in the 1600s and 1700s made these observations mainly to try to pinpoint the orbit of the moon. Today, scientists know where the moon and stars were hundreds of years ago. If they know what time the observations were made, they can use all this information to determine how much faster Earth was spinning at the time.
Stephenson and Morrison have continued to work on this project in retirement, and they say they are not necessarily done yet. Stephenson would like to find observations from the Incas and the Maya that could be added to their data set. So far, he said, they have struck out.
He would also like to find more observations from AD 200 to 600. That would help fill their biggest hole.
In the meantime, geophysicists like Agnew said they will use the astronomers' work to better understand the forces speeding up and slowing down the spinning of our planet.
"Nothing else on Earth cares about when an eclipse happens, so we have to rely on these observations," Agnew said. "The most astounding thing about this paper is the fact that we have this information at all."
© 2017 Los Angeles Times under contract with NewsEdge/Acquire Media. All rights reserved.
Posted: 2016-12-10 @ 8:29am PT
In my calculation in post below I forgot to multiply by days in each year, so:
1) (1.8ms X (27 centuries / 2)) = Average milliseconds lost each year for 2700 years = 24.3 milliseconds per year.
2) (24.3ms X 2700years) = 65,610 milliseconds
3) (65,610ms X 365days) = 23,947 seconds = 399 minutes = 6.6 hours
So it does work using basic math, same result as "Another Chris" that squares and integrates the area of a triangle.
These astronomers have been at this a long time. Google:
"Leslie Morrison" "Richard Stephenson"
See this detailed article from 1998 that better describes their technique:
Edwin Berry, PhD, Physics:
Posted: 2016-12-10 @ 6:07am PT
(1.8 ms/100 yr) 2500 yr = 45 ms slowing in 2500 years.
(1.8 ms/100 yr)/(3600 sec/hour) =(1.8x10^(-5)s/yr)/(.36x10^4 s/hr) =
(1.8/.36)x10^(-9) hr/yr = 5x10^(-9) hr/yr = 5 hr in 10^9 years = 5 hour rotational slowing in one billion years =
1 hour rotational slowing in 200 million years
Another Chris Again:
Posted: 2016-12-10 @ 1:21am PT
Oops, I guess the math wasn't as simple as I thought. Upon further reflection I realize that the simple math I and others had done only gave us the additional length of a day if each day increased by 1.8 ms each century. To find the total time difference over 2700 years you have to add up (or integrate) all the accumulated extra time from each day (really each second) during that period. Thanks to Neil's tip I realized this is similar to solving a physics problem for distance traveled from rest in time t with constant acceleration. In other words, it can be solved by finding the area of a triangle since the rate of change of time is constant. So after I converted the rate of change of time from 1.8 ms per day per century (Thanks again, Neil) to 6.6 x 10^-18 s/s/s I just had to multiply that by the number of seconds in 2700 years squared and divide by 2 and then convert from seconds back to years and indeed I got 6.6 hrs or about 7 hrs. I stand corrected. My apologies.
Catalino Garcia Jr.:
Posted: 2016-12-09 @ 3:52pm PT
Next thing you know Al Gore and his scamming disciples will blame all of these slowing down to "climate change".
Posted: 2016-12-09 @ 3:37pm PT
I see others have noticed the obvious problem with the extremely simple (7th grade in US, probably earlier in the rest of the world) math. I wish someone involved with the article would respond and explain.
Posted: 2016-12-09 @ 1:05pm PT
I could be wrong - but this is my math:
(1.8ms X 13.5) = Average milliseconds lost each year for 2700 years = 24.3 milliseconds per year.
(24.3 X 2700) = total time lost due to slower earth = 65610 milliseconds?
Really six seconds rather than 7 hours?
Posted: 2016-12-09 @ 8:17am PT
The "Principles of Atomic Gravity" explain how gravity works in nature.
Posted: 2016-12-09 @ 7:41am PT
The headline is confusing, the earth's rotation is actually slowing by 1.8ms per day per century. According to my calculations, that equals just under 4 hours after 2500 years (912500 days): Integrate C.x^2 from 0 to 912500 (C=45m/912500^2)
Posted: 2016-12-09 @ 5:10am PT
Anybody considered the longer term consequences of wind or wave power. Renewable - or will the moon crash into the earth.
Posted: 2016-12-09 @ 4:42am PT
Posted: 2016-12-08 @ 3:57am PT
Cannot figure out how to work out the 7hr difference over 2700 years.
Posted: 2016-12-08 @ 2:09am PT
1.8 Milliseconds per Century, over 2700 years, is 1.8 x 27. That is 48.6 milliseconds. Not quite 7 hours!
Posted: 2016-12-08 @ 2:02am PT
700 BC means 27 centuries ago.
27 x 1.8 milliseconds is nowhere near 7 hours !
Its not even one tenth of one second.
Posted: 2016-12-08 @ 1:42am PT
There's something wrong with your math.
2 1/2 millenniums = 2500 years = 25 centuries
1.8 ms/century * 25 centuries = 45 ms
45 milliseconds is 419955 milliseconds shy of 7 hours.
Since this happens to be in the first 2 paragraphs I assume the reset of the article is just as false and not worth the time to read. You might want to double check your other statements.
Posted: 2016-12-08 @ 12:41am PT
If 1.8 milli seconds slowdown per century is correct, then the conclusion that will result in a 7 hrs change in 2 and a half millennia is incorrect.