Why Earth's Day Is Getting Longer: How the Moon Steals Our Spin

The tidal bulge that never quite catches up

Stand on a beach and the tide seems to be about the Moon lifting water. That's half the story. The Moon's gravity raises a bulge of ocean — and, more subtly, of the solid rock itself — on the side of Earth facing it, plus a matching bulge on the far side where the Moon's pull is weakest. If Earth didn't spin, those two bulges would sit perfectly on the Earth–Moon line and nothing interesting would happen.

But Earth spins fast — once every 24 hours — while the Moon takes 27.3 days to circle us. Our planet rotates out from under the bulge faster than the sluggish Moon can drag it back. Friction between the moving water and the seafloor, especially in shallow seas and continental shelves, carries the bulge ahead of the sub-lunar point. So the near-side bulge is always running slightly early, offset by a few degrees.

This is the whole engine. That leading bulge is a lump of mass sitting off-axis, and it exerts a gravitational tug on the Moon. Two things follow from one geometric fact:

• The bulge's forward pull accelerates the Moon in its orbit, flinging it to a higher, more distant orbit.
• The Moon's backward pull on the offset bulge acts as a brake on Earth's rotation, bleeding away spin.

It is a cosmic tug-of-war fought through a slab of misaligned seawater — and Earth is slowly losing.

Angular momentum: the ledger that must balance

Why does slowing Earth's spin push the Moon away? Because the Earth–Moon system is nearly closed, and its total angular momentum is conserved. Tidal friction converts some of Earth's rotational energy into heat — you can't recover that — but angular momentum can only be moved, not destroyed. Every scrap Earth's spin loses is deposited into the Moon's orbit.

An object in a wider orbit carries more orbital angular momentum, so the Moon responds to its windfall by climbing outward. Laser ranging to the retroreflectors left on the surface by the Apollo missions (1969–1972) — mirrors we still bounce laser pulses off today — pins the recession at 3.8 cm per year, measured by millimeter-precision laser ranging (the rate itself known to about ±0.1 mm/yr). That is comparable to the growth of a human fingernail, yet over the 4.5-billion-year history of the system it has carried the Moon from a searingly close orbit out to today's 384,400 km mean distance.

The energy side of the ledger is brutal by human standards: tidal dissipation drains roughly 3.7 terawatts from Earth's rotation, the bulk of it churned into heat in shallow shelf seas like the North Atlantic and the Bering Sea. That's roughly equal to all the electricity humanity generates on average, quietly warming the oceans and spinning us down.

How fast, exactly — and why the numbers disagree

Here is a subtlety that trips up even careful writers. There are two rates, and they don't match:

• The tidal prediction: if lunar and solar tidal friction were the only forces at work, the day should be lengthening at about +2.3 ms per century. This comes straight from the measured lunar recession.
• The observed rate: the actual long-term average, reconstructed from ancient eclipse records spanning 720 BC to AD 2015, is only about +1.8 ms per century.

The ~0.5 ms/century gap is real and mostly explained. Since the last ice age ended, Earth's poles have been rebounding from the crushing weight of vanished ice sheets — a process called glacial isostatic adjustment. As mass shifts back toward the poles, Earth pulls in like a spinning skater drawing in their arms, which speeds rotation up by roughly −0.5 ms/century, partly canceling the tidal drag.

How do we know the ancient rate at all? Babylonian and Chinese scribes recorded exactly where and when total solar eclipses were seen thousands of years ago. If Earth had kept perfect 24-hour time, those eclipse shadows should have fallen on entirely different parts of the globe. The accumulated timing error — Earth's clock is now roughly 4–5 hours behind where uniform rotation would put it over 2,740 years — is the direct fingerprint of the slowdown. It is one of the most beautiful examples in science of clay tablets constraining planetary physics.

Reading the day in fossils and rock

You don't need a laser or an atomic clock to see Earth spinning down — a fossil coral will do. Corals lay down a fine daily growth ridge as their calcium-carbonate skeletons thicken, and superimposed on these are broader annual bands tied to seasonal warmth. Count the daily ridges between two annual bands and you have literally counted the number of days in a prehistoric year.

In the 1960s, paleontologist John W. Wells did exactly this with Devonian corals about 380 million years old and counted close to 400 growth lines per year. The arithmetic is inescapable: the year is set by Earth's orbit and has changed little, so 400 days means each day was shorter — about 22 hours long. Later work on rhythmites (tidally banded sediments) and on Ediacaran-era corals pushes the record back to roughly 620 million years ago, when a day was near 21.9 hours and a year held about 400 days. Note that these deep Precambrian/Ediacaran proxy day-lengths are scattered and model-dependent, so the near-equality of the 21.9-hour Ediacaran and 22-hour Devonian values sits within proxy uncertainty rather than marking a precisely flat rate across those 240 million years.

Extrapolate far enough and the picture becomes dramatic. Not long after the giant impact that formed the Moon, Earth may have spun once every 5 to 6 hours, with the Moon looming a huge, tide-raising disc perhaps ten times closer than today. Interestingly, the slowdown was not perfectly smooth: recent modeling suggests a solar-driven atmospheric tide may have held the day near a constant ~19.5 hours for over a billion years in the Proterozoic, a tug-of-war between lunar braking and solar spin-up before the lunar tide finally won out.

The wobble on top of the trend: why some recent days got shorter

Here is the headline that confuses everyone: even though the day is getting longer over geological time, Earth has recently been spinning faster. Since about 2020 it has repeatedly broken short-day records. The shortest day yet reliably measured fell on 5 July 2024, which ran about 1.66 ms shorter than a standard 86,400-second day. This is not a contradiction — it is a matter of scale.

The steady 1.8 ms/century tidal trend is a whisper buried under much louder, faster fluctuations of milliseconds that come and go over years and decades:

• Core–mantle coupling: Earth's liquid outer core sloshes and exchanges angular momentum with the solid mantle, nudging the surface faster or slower over decades.
• Mass redistribution: melting glaciers, groundwater depletion, and shifting ocean currents move mass toward or away from the axis — measurable by the GRACE gravity satellites (2002–2017) and their successor GRACE-FO.
• Seasonal atmosphere: the jet stream and seasonal winds trade angular momentum with the solid Earth, making the day wobble by ~1 ms within a single year.

These wobbles are so significant that timekeepers may soon need something unprecedented: a negative leap second, subtracting a second from Coordinated Universal Time (UTC) rather than adding one — potentially around 2029. In 27 leap seconds since 1972 we have only ever added time. The long-term Moon still wins; it's just being briefly out-shouted.

The far future, and the limits of the slowdown

Run the tape forward and the day keeps growing — but not forever, and not into anything as tidy as the myths suggest. A few honest bounds:

• Not headed for a 24-hour = 1 month lock. A common claim is that Earth will eventually become tidally locked to the Moon, always showing it the same face, with a day equal to a month. In principle the endpoint is a mutual lock at a day/month of about 47 of today's days. In practice the Sun will swell into a red giant in roughly 5–6 billion years — long before the lock completes — so it will almost certainly never happen.
• The Moon's escape has a deadline too. As the day lengthens, the tidal bulge lags less, the braking weakens, and recession slows. The whole process is self-limiting, not a runaway.

A useful near-term marker: in roughly 200 million years, the day should reach about 25 hours. That's the same tick-rate that pushed a Devonian day down to 22 hours, simply run in reverse. And a caution on precision — the 3.8 cm/yr recession we measure today is anomalously fast because our current continental geography, with its resonant shallow seas like the North Atlantic, dissipates tides unusually efficiently. Averaged over the past billion years the rate was slower; naively multiplying today's rate backward would collide the Moon with Earth only ~1.5 billion years ago, which is wrong. The system's history is written in the drifting continents as much as in the sky.

Earth's day length across deep time versus the recession of the Moon

Era	Day length	Days per year	Moon's distance
~4.5 billion years ago (post-impact)	~5–6 hours	~1,400+	~10× closer than today
620 million years ago (Ediacaran)	~21.9 hours	~400	slightly closer
380 million years ago (Devonian)	~22 hours	~400	~10,000 km closer (~374,000 km)
Today	24 hours (86,400 s)	365.25	384,400 km (mean)
~200 million years future	~25 hours	~350	farther, still receding