Thinking too hard about time is a little like thinking too hard about blinking; it seems natural and intuitive until suddenly you’re sweating and it makes no sense at all. At least, that’s how I felt when I came across an incredible new study published in Nature this afternoon by Duncan Agnew, a geophysicist at the Scripps Institution of Oceanography, suggesting that climate change might be affecting global timekeeping.
Our internationally agreed-upon clock, Coordinated Universal Time (UTC), consists of two components: the one you’re familiar with, which is the complete rotation of the Earth around its axis, as well as the average taken from 400 atomic clocks around the world. Since the 1970s, UTC has added 27 leap seconds at irregular intervals to keep pace with atomic clocks as the Earth’s rotation has gradually slowed. Then that rotation started to speed up in 2016; June 29, 2022, set a record for the planet’s shortest day, with the Earth completing a full rotation 1.59 milliseconds short of 24 hours. Timekeepers anticipated at that point that we’d need our first-ever negative leap second around 2026 to account for the acceleration.
But such a model doesn’t properly account for the transformative changes the planet is undergoing due to climate change — specifically, the billions of tons of ice melting from Greenland and Antarctica every year.
Using mathematical modeling, Agnew found that the melt-off, as measured by gravity-tracking satellites, has again decreased the Earth’s angular velocity to the extent that a negative leap second will actually be required three years later than estimates, in 2029.
While a second here or there might not seem like much on a cosmic scale, as Agnew explained to me, these kinds of discrepancies throw into question the entire idea of basing our time system on the physical position of the Earth. Even more mind-bogglingly, Agnew’s modeling makes the astonishing case that so long as it is, climate change will be “inextricably linked” to global timekeeping.
Confused? So was I, until Agnew talked me through his research. Our conversation has been edited and condensed for clarity.
How did you get involved in researching this? I’d never have expected there to be a relationship between climate change and timekeeping.
Pure accident. I’m a geophysicist and I have an avocational interest in timekeeping, so I know all about leap seconds and the history of atomic clocks. I thought about writing a paper figuring out statistically what the next century would bring in terms of leap seconds.
When I started working on the paper, I realized there was a signal that I needed to allow for, which was the change induced by melting ice — which has been studied, there are plenty of papers on this satellite gravity signal. But nobody has, as far as I can tell, related it to rotation. Mostly because, from a geophysical standpoint, that’s not very interesting.
Interesting. Or, well, I guess not interesting.
I mean, there is geophysical literature on this, but it’s largely, Okay, we see this signal, and gravity doesn’t mesh with what we think we know about ice melt. Does it measure what we think we know about sea level change? How does the geophysics all fit together? And the fact that it changes Earth’s rotation is kind of a side issue.
I did not know about this when I got started on this project; it appeared as I was working on it. I thought, “Wait, I need to allow for this.” And when I did, it produced the — I don’t want to use the words “more important” because of the climate change part, but it produced a secondary result, which was that this potential for a negative leap second became clear.
Walk me through how the ice melting at the poles changes the Earth’s rotation.
This is the part that’s easy to explain. Ice melts. A lot of water that used to be at the poles is now distributed all over the ocean. Some of it is close to the equator. The standard picture for what’s called change of angular velocity because of moment of inertia — ignore all the verbiage — but the standard picture is of an ice skater who is spinning. She has her arms over her head. When she puts her arms out, she will slow down — like the water going from the poles to the equator. And that’s it. This is the simple part of the problem.
So what’s the hard part?
The hard part is explaining the part about the Earth’s core. If you have two things that are connected to each other and rotating and one of them slows down, the other one has to speed up. I have not been able to think of an ice skater-like-metaphor to go with that, but the simple one is if you were to put a bowl of water on a lazy Susan and you spin the bowl, then the water will start to spin. It won’t spin initially, but then it will start.
If you started stirring the water in the other direction, that would slow the Lazy Susan down. And that’s the interaction between the core and the solid part of the Earth.
And is that causing the negative leap second to move back three years?
That’s why the leap second might happen at all. On a very long timescale, what’s happening is that the tides are slowing the Earth down. The Earth being slower than an atomic clock means that you need a positive leap second every so often. That was the case in 1972, when they started using leap seconds. The assumption was that the Earth would just keep slowing down and so there would be more positive leap seconds over time.
Instead, the Earth has sped up, entirely because of the core, and that’s not something that people necessarily anticipated. When you take the effect of melting ice out, it becomes clear there’s this steady deceleration of the core; the core is rotating more and more slowly. If you extrapolate that — which is a somewhat risky thing to do, you can’t really predict what the core is going to do — then you discover that there is a leap second, in 2029. The ice melting is going in the other direction; if the ice melting hadn’t occurred, then the leap second would come even earlier. Is this all making sense?
I think I’m grasping it.
Just so you know, one of the two reviewers of this paper was someone in geophysics who said, “I know all this stuff. I wasn’t familiar with the rotation part. This paper has an awful lot of moving parts.”
So, it’s just a difference of a second. Why does this even matter?
We are all familiar with the problem of not being synchronized — we just went through it. If you forget that we did Daylight Savings Time, then you’re an hour off from everybody else and it’s bewildering and a nuisance.
Same problem with leap seconds, except for us, a second is not a big deal. For a computer network, though, a second is a big deal. And why is that? Well, for example, in the United States, the rules for stock markets say that everything that is done has to be accurately timed to a 20th of a second. In Europe, it’s actually to the nearest 1,000th of a second. If we were all just farmers or something, it wouldn’t be a problem, but there’s this whole infrastructure that’s invisible to us that tells our phones what time it is, and allows GPS to work, and everything else.
The easiest thing to do would be to not have a negative leap second. Indeed, there are plans not to have leap seconds anymore because for computer networks, they’re an enormous problem. They arrive at irregular intervals; some human being has to put the information into the computer; the computer has to have a program that tells it when the leap seconds are; and most computer programs can’t tell whether it’s a plus or minus second because there’s never been a minus before. From the computer network standpoint, it would be simplest to just not do this.
So, you ask, why are we doing this? In 1972, when leap seconds were instituted, there were two communities that cared about precise time. One was the people who cared about the frequency of your radio station and other kinds of telecommunications. They wanted to use atomic clocks with frequencies that didn’t change, but that didn’t mesh with what the Earth was doing.
Who cares about time telling you how the Earth is rotating? Well, the answer then was that there were people who used the stars for celestial navigation. Back then, celestial navigation was used not just for ships, but for airplanes — if you flew across the ocean, there was a guy in the cockpit, an actual navigator, who would use a periscope to look at the stars and locate the plane, if only as sort of a backup system. That community is now gone. Almost nobody uses celestial navigation as a primary, or even a secondary, way of finding out where they are anymore because of GPS.
My own personal view — and I can warn you, there’s a huge amount of dispute about this — is that we’d be fine if we just stopped having leap seconds at all.
Is there a … governing body of time? That forces us to do leap seconds?
There’s a giant tangle of international organizations that deal with this, but the rules were set by the people in charge of keeping radio stations aligned because radio broadcasts were how time signals were distributed back in 1972. So the rule was created. Who makes that decision is something called the International Earth Rotation and Reference Systems Service, which uses astronomy to monitor what the Earth is doing. They can predict a little bit in advance where things are going to be, and if within six months things are going to be more than half a second out, they will announce there will be a leap second.
Back to climate change: It seems pretty amazing that something like melting ice can throw things off so much.
All the stuff about negative seconds is important, but it’s only important because of this infrastructure, because we have all these rules. Strip all of that away and the most important result becomes the fact that climate change has caused an amount of ice melt that is enough to change the rotation rate of the entire Earth in a way that’s visible.
How do you talk to people about the gigatons of ice that Greenland loses every year? Do you talk about “water that could cover the entire United States to the depth of X” to get it into people’s minds? Yes, these are small changes in the rotation rate, but just the fact that we can say, “Look, this is slowing down the entire Earth” seems like another way of saying that climate change is unprecedented and important.
How do we proceed, then, if climate change is messing with our system?
There was a lot of resistance to even introducing atomic time. Time was thought of as being about Earth’s rotation and the astronomers didn’t want to give it up. In fact, in the 19th century, observatories would make money by selling time signals to the rest of the community. Then, in the 1950s, the physicists showed up, ran atomic clocks, never looked at the stars, and said, “We can do time better.” The physicists were right. But it took the astronomical community a while to come around to accepting that was how time was going to be defined.
If we get rid of leap seconds then we’d really have cut the connection between the way in which human beings have always thought of time as being, say, from noon to noon, or from sunrise to sunset, and we’d be replacing it with some bunch of guys in a laboratory somewhere running a machine. For some people, it’s very troubling to think of severing the keeping of time from the Earth’s rotation.
You lose a bit of the romance, I think. But clearly, tying our way of describing the linear passage of sequential events to the Earth’s rotation is going to be messy.
Exactly right. There’s a quote from, of all people, St. Augustine, saying, “I know what time is, but if somebody asked me, I can’t tell them.”
Walmart
