How GPS Corrects for Relativity

Every GPS satellite carries an atomic clock precise enough that, left uncorrected, Albert Einstein's theories of relativity would make it drift by about 38 microseconds a day compared to an identical clock sitting on the ground. That sounds trivial until you convert it into distance: 38 microseconds of accumulated timing error translates into roughly 10 kilometers (6 miles) of position error per day, according to Ohio State's overview of relativity in GPS. Engineers solved the problem before the first operational satellite ever reached orbit, by deliberately mistuning each clock to run slightly slow on the factory bench, so that once it climbed to orbital altitude and orbital speed, it would tick at exactly the right rate.

This is not a footnote in a physics textbook. It is a load-bearing piece of engineering that roughly four billion smartphones, aircraft, ships, and precision agriculture systems depend on every single day. GPS is one of the clearest cases where a system used by ordinary people, not physicists, only works because its designers took a hundred-year-old theory of relativity seriously as a hard numerical constraint, not an academic curiosity.

Why GPS needs nanosecond-level timing at all

GPS does not work by triangulating angles the way old-fashioned radio direction finding did. It works by timing. Each satellite broadcasts a signal stamped with the exact time it was sent, and a receiver on the ground compares that timestamp against its own clock to calculate how long the signal took to arrive.

Because radio waves travel at the speed of light, roughly 299,792 kilometers per second (186,282 miles per second), that travel time converts directly into distance. A receiver that knows its distance from four or more satellites simultaneously can solve for its own position in three dimensions plus a time correction, a technique called trilateration. The entire system is, at bottom, a very elaborate stopwatch.

That is also why GPS is so unforgiving of timing error. Light moves fast enough that even a one-microsecond mistake in a satellite's clock corresponds to about 300 meters (roughly 980 feet) of position error on the ground, a distance error large enough to put a receiver in the wrong parking lot, let alone the wrong lane of traffic. A car's turn-by-turn navigation, a jetliner's autoland system, and a surveyor's centimeter-grade land measurement all depend on the same underlying clocks staying synchronized to within a few tens of nanoseconds.

A one-microsecond timing error on a GPS satellite corresponds to about 300 meters of position error on the ground, because radio signals travel at the speed of light and even tiny timing mistakes translate into large distances.

Why does GPS need such extreme timing precision in the first place?

GPS calculates your position by timing how long radio signals take to travel from multiple satellites to your receiver, then converting that time into distance using the speed of light. Because light travels roughly 300,000 kilometers per second, even a one-microsecond timing error corresponds to about 300 meters of position error, which is why relativistic effects, negligible in daily life, become critical at this level of precision.

Ordinary clocks, even very good quartz clocks, are nowhere near stable enough for this job. GPS satellites instead carry cesium and rubidium atomic clocks, devices that measure time by counting the extremely regular oscillations of atoms transitioning between energy states, accurate to within a few nanoseconds per day. That level of stability is precisely what makes the relativistic effects, normally far too small to notice in daily life, impossible to ignore.

Two relativistic effects pulling in opposite directions

GPS satellites experience two separate, independently real relativistic effects that partially cancel each other out, but not completely, and the difference between them is what engineers actually have to correct for.

The first is special relativity, the branch of Einstein's theory dealing with objects in relative motion. Special relativity says that a clock moving relative to an observer runs slower from that observer's point of view, an effect called time dilation. GPS satellites orbit at roughly 14,000 kilometers per hour (about 8,700 miles per hour) relative to the ground, fast enough that special relativity alone would make their onboard clocks lose about 7 microseconds per day compared to a stationary clock at sea level.

The second is general relativity, the branch dealing with gravity as a curvature of spacetime. General relativity says the opposite in this situation: a clock farther from a large mass, where the local gravitational field is weaker, runs faster than a clock deeper in that mass's gravitational well. GPS satellites orbit about 20,200 kilometers (12,550 miles) above Earth's surface, in a mid-altitude orbit region often called medium Earth orbit, far enough out that the weaker gravity there makes their clocks gain about 45.9 microseconds per day relative to an identical clock on the ground, a figure documented in Neil Ashby's widely cited review paper, "Relativity in the Global Positioning System", published in Living Reviews in Relativity.

Both effects are physically real. Neither one is an approximation invoked for convenience; both have been independently verified in laboratory tests unrelated to satellite navigation, from atomic clocks flown on aircraft to clocks compared at different floors of the same building. GPS did not invent either phenomenon. It simply became the first large-scale engineered system forced to account for both simultaneously with extreme precision, because its entire function depends on nanosecond-level timing.

Effect	Cause	Direction	Magnitude (per day)
Special relativity	Satellite's orbital velocity (~14,000 km/h relative to the ground)	Slows the satellite clock	~7 microseconds
General relativity	Weaker gravity at ~20,200 km altitude compared to Earth's surface	Speeds up the satellite clock	~45.9 microseconds
Net effect	Combination of both	Satellite clock runs fast	~38.7 microseconds

General relativity's speeding-up effect dominates by a wide margin, leaving GPS satellite clocks running roughly 38 microseconds per day faster than identical clocks on the ground, a figure engineers had to design around before launch, not patch afterward.

The two relativistic effects do not cancel out. General relativity's speeding-up effect wins by a wide margin, leaving GPS satellite clocks running about 38 microseconds per day faster than identical clocks on the ground, a net drift that engineers pre-correct in hardware before every satellite ever launches.

How engineers pre-correct for it before launch

Rather than trying to correct the time signal after the fact, in software, on the ground, GPS engineers built the correction directly into the satellite hardware itself, before the satellite ever left the factory. Each satellite's atomic clock is manufactured to tick at a frequency of 10.22999999543 megahertz instead of the nominal 10.23 megahertz used by ground-based reference clocks, a deliberate, calculated offset chosen specifically to counteract the roughly 38-microsecond daily gain the clock will experience once it reaches orbital speed and altitude.

Think of it like setting a wristwatch a few minutes slow before a long hike into thin mountain air, because you know from prior trips that the watch's mechanism runs measurably fast at altitude and low pressure. GPS engineers performed the equivalent calculation, but instead of trial-and-error from past hikes, they used Einstein's field equations and the specific orbital parameters of the GPS constellation to compute the exact frequency offset needed, then built that correction directly into the clock's factory settings rather than relying on a software patch uploaded after the satellite was already flying.

This pre-correction approach has a practical engineering logic behind it. Ground stations can still apply small, secondary corrections to account for orbital eccentricity, atmospheric effects, and other minor variations during operation, but the overwhelming majority of the relativistic offset, the part caused by the constant combination of orbital speed and orbital altitude, is fixed and permanent, baked into the clock the day it was manufactured. That means the correction cannot be forgotten, misconfigured, or accidentally disabled by a software update; it is a physical property of the hardware itself.

How was the exact correction frequency determined?

Engineers calculated the combined special and general relativistic time dilation for the GPS constellation's specific orbital velocity and altitude, arriving at a net gain of about 38.7 microseconds per day. They then lowered the clock's manufactured frequency from 10.23 MHz to 10.22999999543 MHz, an offset of about 4.57 parts in 10 billion, precisely calibrated so the relativistic speed-up brings the clock back to the correct rate once in orbit.

Did engineers get this right on the first try?

Not entirely by blind faith. The first Block I GPS satellite, launched with a cesium clock in the late 1970s, reportedly flew for an initial period with its relativistic correction switchable, allowing ground controllers to compare the clock's behavior with and without the offset engaged. Sources describing the early GPS program, including summaries referenced in GPS World's coverage of GPS and relativity, note that the observed drift matched Einstein's predictions closely enough that engineers activated the permanent correction with confidence, effectively turning the entire GPS satellite fleet into an ongoing, real-world confirmation of general relativity.

What happens if the correction is ignored

If engineers had ignored relativity entirely and launched GPS satellites with clocks running at the ordinary, unmodified 10.23 megahertz frequency, position errors would compound at roughly 10 kilometers per day, and the error would keep growing every subsequent day because the underlying clock drift never stops accumulating.

That is not a small inconvenience for a navigation system whose entire value proposition is meter-level or better accuracy. Ground receivers would need progressively larger correction factors just to remain marginally usable, and within a very short window of continuous operation, some estimates cited in the Ohio State summary put it at around two minutes, an uncorrected GPS fix would already be too degraded for practical navigation. A car's dashboard map would show the vehicle driving through buildings; an aircraft's instrument approach would place the runway hundreds of meters from where it actually is.

This is one of the few places in ordinary engineering where general relativity, a theory usually considered irrelevant to anything built on Earth's surface, has to be treated as a first-order design requirement rather than a rounding error. Bridges, buildings, and cars are engineered entirely within Newtonian mechanics; GPS satellites cannot be, because their function depends on timing precision fine enough to expose effects that are otherwise buried far below the threshold of everyday experience.

Without the relativistic correction, an uncorrected GPS fix would become too degraded for practical navigation within roughly two minutes of continuous operation, according to the Ohio State summary of Ashby's analysis, because the roughly 10-kilometer-per-day drift rate never stops accumulating.

Why this counts as real-world evidence for relativity

Relativity is often taught as an abstract, almost philosophical theory, something involving trains, light beams, and thought experiments about twins aging at different rates. GPS turns that abstraction into an operational engineering constraint with a dollar cost attached to getting it wrong.

Every one of the roughly 30 to 32 active satellites in the GPS constellation, operated by the United States Space Force, is a functioning experiment that continuously validates both special and general relativity, around the clock, for anyone who cares to check the math against the observed clock behavior. This is part of why physicists frequently cite GPS, alongside phenomena like gravitational lensing and the precession of Mercury's orbit, as one of the most commonly referenced real-world confirmations of Einstein's theories, precisely because it is not a laboratory experiment run once under controlled conditions. It is a continuously operating, commercially critical system that would visibly and immediately break if relativity were wrong, or even off by a modest percentage.

Similar relativistic timing considerations apply to other satellite navigation systems as well, including Europe's Galileo, Russia's GLONASS, and China's BeiDou, each of which independently has to solve the same underlying physics problem for its own constellation's specific orbital parameters. The consistency of these independent engineering solutions across multiple countries and space agencies, all converging on corrections consistent with Einstein's equations, adds further weight to relativity's practical validity outside the context of any single national program.

Common misconceptions about GPS and relativity

A common misconception is that GPS relies only on general relativity, or only on special relativity, when in reality both effects are physically real, both are calculated with precision, and the net correction applied to the satellite clocks is the difference between the two, not either effect considered in isolation. Framing GPS as validating only one half of Einstein's work understates the actual engineering challenge and the actual physics involved.

Another misconception is that the correction happens continuously, in real time, through software running on the ground segment during normal operation. The primary correction is a fixed, permanent adjustment baked into the satellite clock's manufactured frequency before launch; it does not change day to day and does not require ongoing recalculation. Additional, much smaller corrections do occur in the ground segment, primarily to account for the small variations introduced by each satellite's slightly elliptical orbit rather than perfectly circular orbit, but the bulk of the relativistic offset, the roughly 38 microseconds per day, is handled entirely at the hardware level and never touched again after launch.

A third, subtler misconception is that this correction was some kind of lucky guess or rough approximation that happened to work well enough. In reality, the frequency offset of 10.22999999543 megahertz was calculated directly from Einstein's equations applied to the GPS constellation's known orbital velocity and altitude, not estimated empirically after the fact through trial and error. The initial Block I satellite's switchable correction period existed specifically to confirm that calculation against real orbital data, not to search for the right answer through experimentation.

Finally, some assume the relativistic correction is the only source of GPS timing error engineers have to manage. It is one of the largest and most famous, but GPS receivers and the ground segment also correct for atmospheric delay as signals pass through the ionosphere and troposphere, satellite clock drift unrelated to relativity, multipath interference from signals bouncing off buildings and terrain, and orbital position uncertainty. Relativity is a foundational correction applied once at the hardware level; these other corrections are applied continuously, in real time, by the receiver and the ground control segment working together.

What this means for the systems that depend on GPS

Precision timing from GPS is not only used for figuring out where you are. GPS-derived timing signals synchronize cellular network towers, financial trading systems that require microsecond-accurate transaction timestamps, and portions of the electrical power grid that use timing to detect faults and coordinate load balancing across long-distance transmission lines. All of these downstream systems inherit the same relativistic correction baked into the GPS satellite clocks, whether or not the engineers using GPS timing signals for their own unrelated purposes are even aware that relativity is involved.

This dependency chain is worth noting because it illustrates how a correction calculated for one purpose, accurate position fixes for navigation, ends up quietly underpinning infrastructure that has nothing to do with maps or driving directions. A stock exchange's trade-timestamping system and a smartphone's turn-by-turn navigation app are, at a fundamental level, both relying on the same 1970s-era engineering decision to pre-slow a set of atomic clocks by a few parts in ten billion.

The takeaway

GPS is a rare case where a system used by billions of people every day only works because engineers took Einstein's relativity seriously as a practical, load-bearing engineering constraint rather than a theoretical curiosity confined to physics classrooms. Special relativity slows the satellite clocks down by about 7 microseconds a day because of their orbital speed, general relativity speeds them up by about 45.9 microseconds a day because of weaker gravity at altitude, and the net difference, roughly 38 microseconds a day, is corrected permanently by manufacturing each clock to tick at a deliberately slower frequency before it ever leaves the ground. The corrected clock frequency built into every GPS satellite is a small, permanent, and continuously operating monument to general relativity's real-world accuracy, quietly running underneath everything from turn-by-turn directions to stock market timestamps.

---

Read next:

• What is a photon, explained
• How NFC payments work securely
• What if Earth had two moons

FAQ

Q: How much does relativity affect GPS satellite clocks?

Special relativity slows GPS clocks by about 7 microseconds per day due to their orbital speed, while general relativity speeds them up by about 45.9 microseconds per day due to weaker gravity at altitude. The net effect is a gain of about 38 microseconds per day, which engineers correct for before launch.

Q: How does GPS correct for relativistic time drift?

Engineers manufacture each satellite's atomic clock to run at a slightly lower frequency, 10.22999999543 megahertz instead of 10.23 megahertz, so that once the satellite reaches orbital speed and altitude, the relativistic effects bring it back into sync with ground-based time.

Q: What would happen if GPS didn't account for relativity?

Position errors would accumulate at roughly 10 kilometers per day, and a GPS fix could become unreliable for practical navigation within about two minutes of continuous operation, because the underlying clock drift never stops compounding.

Q: Why does GPS need such precise timing?

GPS calculates distance by measuring how long a radio signal takes to travel at the speed of light, so even a one-microsecond timing error translates into roughly 300 meters of position error on the ground.

Q: Is the relativistic correction applied continuously or just once?

The main correction is a one-time, permanent adjustment to the satellite clock's manufactured frequency, made before launch. Smaller secondary corrections are applied in the ground segment during operation to account for orbital variations, but they are separate from, and much smaller than, the core relativistic offset.

Q: Do other satellite navigation systems besides GPS also need relativistic corrections?

Yes. Galileo, GLONASS, and BeiDou each face the same underlying physics problem for their own constellations and apply their own relativistic corrections calculated for their specific orbital altitudes and velocities.