A two-line element set (TLE) is a compact text encoding of a satellite's orbit at a specific moment (its epoch), published by US Space Command via Space-Track and mirrored by CelesTrak. Run through the SGP4 model, it predicts where the object will be at any time.

How often are TLEs updated?

Usually once or twice a day for actively tracked objects, and more often for high-interest or maneuvering satellites. Accuracy decays the further you propagate from the epoch, so using fresh elements matters.

Can a TLE tell me exactly where a satellite is?

It gives a good estimate, not ground truth. SGP4 from a public TLE is typically accurate to roughly a kilometre near epoch and drifts over days. It is not a substitute for an operator's own ephemeris.

Why is a TLE only two lines?

The format dates to the punch-card era, when each line was a fixed-column card. Every value sits at a fixed column position, which is why it looks cryptic but parses unambiguously.

GuideJune 2026

How to read a TLE (two-line element set)

A TLE packs a satellite's entire orbit into two 69-character lines. Here is where it comes from, what every field means, and how to turn it into a position on screen.

If you have spent any time around satellite tracking you have seen a two-line element set, or TLE: two dense rows of digits that somehow describe exactly where a spacecraft is. It looks impenetrable, but it is rigidly structured. Once you know what sits in each column it reads like a passport for an orbit.

What a TLE actually is

A TLE is a snapshot of a satellite's orbit at one instant, called the epoch. It does not store a position. It stores the six orbital parameters (plus a drag term) that a propagation model needs to compute a position at any later time.

The data comes from the US Space Force's 18th Space Defense Squadron, which tracks tens of thousands of objects with radar and optical sensors and publishes their orbits on Space-Track. CelesTrak mirrors and curates the same data. The two-line shape is a relic of the punch-card era: every field lives at a fixed column, which is why it is terse but never ambiguous.

The three lines

Despite the name there are usually three lines. The first is an optional name line; the two numbered lines carry the data. Here is the ISS:

ISS (ZARYA)
1 25544U 98067A   24060.51782528  .00016717  00000-0  10270-3 0  9994
2 25544  51.6416 247.4627 0006703 130.5360 325.0288 15.49814234 22.5

Line 1: identity and drag

25544 is the catalog (NORAD) number; the trailing U marks it unclassified.
98067A is the international designator: launched in 1998, the 67th launch of the year, object A of that launch.
24060.51782528 is the epoch: year 2024, day 60.518… (a fractional day-of-year). This is the moment the orbit is valid for.
The remaining fields are the rate of change of mean motion, a second derivative (almost always zero), and the B* drag term (10270-3 means 0.00010270), which captures how atmospheric drag is decaying the orbit.

Line 2: the orbit geometry

51.6416 is the inclination in degrees (the ISS tilt relative to the equator).
247.4627 is the right ascension of the ascending node (how the orbit plane is rotated in space).
0006703 is the eccentricity, with an implied leading decimal: 0.0006703, i.e. very nearly circular.
130.5360 is the argument of perigee and 325.0288 the mean anomaly, which together fix where the satellite is along the orbit at epoch.
15.49814234 is the mean motion in revolutions per day. The ISS laps the Earth about 15.5 times a day.

The final digit on each line is a checksum, a simple guard against transcription errors.

Turning a TLE into a position

You cannot read a position straight off these numbers, and you cannot just plug them into textbook Kepler equations either. TLEs are mean elements designed for one specific propagator: SGP4 (and its deep-space sibling SDP4). SGP4 folds in Earth's oblateness and atmospheric drag to produce a position and velocity at any requested time. Feed it the same TLE and a time, and you get a coordinate.

That is exactly what powers a live tracker. Vantafort runs SGP4 on the GPU for the entire public catalog at once, which is how it propagates tens of thousands of objects in real time. For the maths behind that, see SGP4/SDP4 on WebGPU.

Accuracy and limits

A public TLE is good, not perfect. Near epoch SGP4 is typically accurate to about a kilometre; over days the error grows, especially for low-altitude objects feeling drag. TLEs are not operator-grade ephemeris, and they cannot anticipate a maneuver that has not happened yet. Always propagate from the freshest element set you can get.

Want to skip the parsing entirely? Open the live ISS tracker and watch a TLE become an orbit and ground track in real time, or read how to track the ISS.

Frequently asked questions

What is a TLE?: A two-line element set (TLE) is a compact text encoding of a satellite's orbit at a specific moment (its epoch), published by US Space Command via Space-Track and mirrored by CelesTrak. Run through the SGP4 model, it predicts where the object will be at any time.
How often are TLEs updated?: Usually once or twice a day for actively tracked objects, and more often for high-interest or maneuvering satellites. Accuracy decays the further you propagate from the epoch, so using fresh elements matters.
Can a TLE tell me exactly where a satellite is?: It gives a good estimate, not ground truth. SGP4 from a public TLE is typically accurate to roughly a kilometre near epoch and drifts over days. It is not a substitute for an operator's own ephemeris.
Why is a TLE only two lines?: The format dates to the punch-card era, when each line was a fixed-column card. Every value sits at a fixed column position, which is why it looks cryptic but parses unambiguously.

All articles