Facebook
Twitter
Pocket
Reddit
Email
Climate change is transforming our measurements of many things—summer temperatures, the level of the seas, the strength of hurricanes, for instance. But now, as polar ice caps melt with more regularity, it’s even messing with how we keep time.
According to a recent paper in Nature, that polar thaw has progressed to such an extent that it’s slowing the rotation of the planet. As the melt runs into the ocean, the extra water gathers around the Earth’s belly, fattening it a bit around the equator and reducing the planet’s angular velocity—or rate of spin. Think of an ice skater slowing her axel by extending her arms and you get the picture.
A slower rate of spin, in turn, alters the time it takes for the planet to rotate once—what we think of as a day—which means the rising and setting of the sun are not happening quite when expect them to, putting a kink in the vast international network of infrastructure that tells us what time it is.
A negative leap second is a serious business.
The problem is, even tiny deviations in time-keeping can wreak havoc with the systems we most depend on to orient us throughout our day. Things like financial markets, airline reservation systems, social media servers, and search engines all rely on hyper-accurate synchronizations of time across the globe, and when the timing is off, things can get pretty weird.
To smooth over recent discrepancies, by 2029, we’ll need to add something called a negative leap second—a day where the clocks would go from 11:59 and 58 seconds directly to midnight, skipping 11:59:59—to allow our timekeeping devices to catch up with Earth, writes geophysicist Duncan Agnew of the Scripps Institute for Oceanography in the Nature paper.
It’s something that has been anticipated—with no small amount of unease—for some time, Agnew says. One big problem is that we’ve never dealt with a negative leap second before. We’ve only ever added leap seconds—and the prospect of subtracting one has those responsible for keeping us on schedule nervous.
“A negative leap second is a serious business,” says physicist Judah Levine, one of the world’s primary timekeepers. For the past half century, Levine has worked at the National Institute for Standards and Technology in Boulder, Colorado where he helps oversee NIST-F1—the cesium atomic clock that serves as the beating heart of time in the United States. It’s one of about 450 cesium clocks from all over the globe that report data to the United Nations’ International Telecommunications Union—which, in turn, disseminates the correct time to the rest of the planet via radio frequencies and the internet.
“There is a serious concern that the software to implement the negative leap second has bugs because it has never been tested,” says Levin. “There are still problems with adding positive leap seconds, even though it has been done for 50 plus years.”
The worst of these problems came on June 30, 2012, when the addition of a leap second broke some fundamental pillars of the Web. As soon as the atomic clocks inserted a 61st second at midnight of that day, a timekeeping kernel in the open-source Linux operating system—which forms the backbone for much of the code on the internet as well as on Wall Street—popped. (A Linux subsystem called “hrtimer” couldn’t reconcile the time differences, which sparked hyperactivity on servers running Linux, which then locked up the hardware of those machines.)
The cascading effects of that one hiccup caused Qantas Airlines reservation system to crash, stranding passengers worldwide. Reddit, the enormous internet discussion platform, went dark. Java quit working. Mozilla internet browsers went a little haywire. Gawker, the once snarky pop-culture blog spot, dropped from view, as did Yelp, LinkedIn, and other monuments of the Web’s adolescent years. And rather than risk billions of dollars going astray, the New York Stock exchange and Nasdaq simply halted after-hours trading sessions to let the leap second pass.
But let’s step back for a … leap second, to explain why leap seconds came into use at all. Since 1972, the international community has precariously balanced two methods of keeping time. One method, based on the Earth’s daily rotation, is as old as timekeeping itself, and relies on the position of the sun and other stars in the sky. The other more precise method relies on the steady, unerring frequency of light emitted by the cesium atoms within atomic clocks, which supply the regularity so essential to the digital underpinnings of our lives.
The trouble is that, over time, the two different kinds of clocks—atomic and astronomic—tend to diverge. Astronomic time, called Universal Time, or UT1, has, until recently, tended to fall a few ticks behind atomic clock time, known as Universal Atomic Time, or TAI. While TAI runs at a steady 86,400 seconds a day, the planet itself is influenced by tides, the moon’s gravity, and the currents of its molten liquid core, which aren’t so exact or predictable.
Having to deal with leap seconds drives me crazy.
The effect of these combined forces means that every few years we’ve had to bring TAI and UT1 back together by inserting a leap second—pausing the atomic clocks for a few beats to allow the astronomical clock to catch up. When the two clocks stray by nine tenths of a second, the International Earth Rotation and Reference Systems Service in France schedules a leap second, which is added on December 31 or June 30 in whatever year it is needed. Once the two clocks are again in accord, they create what’s called Coordinated Universal Time, or UTC—the time as we know it until the clocks, again, part ways.
This was all working just fine, said Agnew, until a few decades ago when activity in the Earth’s molten liquid core caused the outer crust of the planet—where we live—to spin more rapidly. As the core spins faster, it gradually transfers some of its rotational momentum to the outer layers over long timescales. As a result, we recently began adding leap seconds less frequently than we used to. At first, they were needed almost annually—we added one 23 times between 1972 and 1999—but since the turn of the century, we have added only four, the last in 2016. Little by little, Agnew told me, it became clear that we could have to subtract one.
Originally, the dreaded negative leap second was predicted to fall in 2026. But data Agnew analyzed from satellites mapping Earth’s gravity show that, since the 1990s, our planet has become less spherical and more oblong as ice in Antarctica and Greenland melts and moves toward the equator. This has subtly applied the brakes to the planet’s rotation and delayed the negative leap second—and the screw-ups it could cause—by three years.
In the meantime, the Linux code blamed for the rippling mishaps in 2012 has been rewritten. Other computer giants that form the digital landscape like Google, Microsoft, and Meta now use their own proprietary systems to account for leap seconds. Meta, for instance, “smears” the 61st second across many hours so that Facebook, Instagram, and WhatsApp don’t lose their bearings.
In Levine’s view, big tech running separate code to grapple with leap seconds only creates more confusion. Their methods, he said, “are not consistent with the official definition [of time] and the different systems are not consistent with each other.” Further, Levine adds, how any of those corporations plan to handle subtracting, rather than adding, a second remains unknown. (Google, for instance, did not respond to my questions on the topic).
“Having to deal with leap seconds drives me crazy,” says Levine. The rest of the world feels the same way. In November of 2022, timekeeping representatives from the world’s governments gathered at the General Conference on Weights and Measures in Paris, and voted to abandon leap seconds from 2035 on.
What will take its place thereafter—and keep the atomic and astronomical clocks aligned—is, Levine said, a big unknown. Levine himself has proposed something not unlike Meta’s practice of smearing. In his method, an entire leap minute would be amortized over the course of a decade in an algorithmically guided sequence.
Getting everyone to agree on that is yet another headache as vexing as the leap second itself. “The political considerations are very often more important than the technical issue,” Levine tells me.
For now, the International Earth Rotation and Reference Systems Service has granted us a bit of a reprieve. According to its latest bulletin, “NO leap second will be introduced at the end of June 2024.”
The announcement comes just in the nick of time. 
Lead image: Mopic / Shutterstock
Get the Nautilus newsletter
Cutting-edge science, unraveled by the very brightest living thinkers.
