How are distances measured in space? Kilometres, au and light-years

One text gives a distance as 149.6 million kilometres; another calls it about 1 astronomical unit. The numbers look very different, yet the length can be the same. How do astronomers choose these units — and how can they know a distance when they cannot stretch a tape measure to a star?

The kilometre, astronomical unit and light-year are units of length. They set the size of the “division” used to express a distance. The distance itself is determined from observations: for example, the return time of a light pulse, or the geometry of a star’s apparent shift. Choosing a unit and choosing a measuring method are separate decisions.

A larger division gives a smaller number

A route 3 kilometres long can also be written as 3000 metres. The route has not grown: we have changed the unit we compare it with. One kilometre is exactly 1000 metres. The same rule applies far from Earth. BIPM: SI decimal prefixes.

Kilometres are convenient for distances near Earth. Between planets, the numbers already contain many zeros. We can keep using kilometres, but a larger unit makes proportions easier to see: twice the scale of Earth’s orbit, thirty times that scale, and so on.

This resembles choosing between metres and kilometres on a city map. An astronomical unit, however, does not require a physical ruler laid out in space. It is an agreed length used to express a result.

The astronomical unit is a fixed length

The astronomical unit, with the international symbol au, is exactly 149 597 870 700 metres, or 149 597 870.7 kilometres. The International Astronomical Union adopted this definition in 2012. For a rough estimate, people often say “about 150 million kilometres”. IAU: Resolution 2012 B2.

An au is close to the characteristic separation of Earth and the Sun. That gives us the useful introductory description “the average Earth–Sun distance”. It is not the unit’s modern exact definition. Earth does not follow a perfect circle, so its distance from the Sun changes during the year while the au stays fixed. Also, an “average” requires us to specify how the averaging was done. NASA: elliptical orbits, USNO: mean distance.

Thus, “a probe is 2 au from the Sun” means a separation equal to two of these units — about 299 million kilometres. It does not yet tell us how far the probe has travelled, how far it is from Earth, or how its distance was determined. Those questions require separate information about position, path and measurements.

A light-year also measures length

For interstellar distances, even an au becomes a small division. A light-year, symbol ly, is the distance light travels in a vacuum in one Julian year. Here a year means exactly 365.25 days of 86 400 seconds each, rather than an arbitrary calendar year with 365 or 366 days. One light-year is about 9.46 × 10¹² km, or 63 241 au. USNO: light-year and Julian year, JPL: time intervals and constants.

The name contains “year” because a time interval was used to define a length. Nevertheless, “two light-years” answers “how far?”. “Two years” answers “how long?”. A spacecraft will not necessarily cover two light-years in two years: its speed and the target’s motion matter.

Look at these three expressions for one segment. The small decimal fraction in light-years does not make it shorter — the unit itself is much larger.

One span — three ways to write it

Length from A to B: 1 au

Kilometres: 149 597 870.7 kmExact conversion
Astronomical units: 1 auExact definition
Light-years: ≈ 0.0000158 lyRounded conversion

All three expressions describe the same length. A larger unit gives a smaller number. This is a hypothetical segment, not a measured distance or a diagram of Earth’s orbit. Definitions: IAU, USNO.

Branch · With equationsOptional · you can skip this

Where does 9.46 × 10¹² km come from?

Let $c$ be the speed of light in a vacuum, $T$ the duration of a Julian year, and $L$ the length travelled. Then $L=cT$ . In SI, $c$ has the exact value 299 792 458 m/s. BIPM: definition of the metre.

First express the time interval in seconds:

\begin{aligned} T &= 365.25\times86\,400\ \text{s}\\ &= 31\,557\,600\ \text{s}. \end{aligned}

Now multiply speed by that interval:

\begin{aligned} L &= cT,\\ &\approx 9.46\times10^{15}\ \text{m},\\ &\approx 9.46\times10^{12}\ \text{km}. \end{aligned}

Metres per second multiplied by seconds give metres: units provide a check on the reasoning. Under this definition, the exact result is 9 460 730 472 580.8 km. The short expression 9.46 × 10¹² km is rounded. This is the precision of a unit conversion; it says nothing about the accuracy of a measured distance to a particular star.

A unit cannot replace observations

Send a short laser pulse to a special reflector on the Moon, and some light returns to the telescope. An instrument measures the interval between sending and receiving it. Knowing the speed of light gives the total path travelled; because the light went there and back, divide by two to obtain the one-way distance. This is the basic idea of laser ranging. The resulting length can be expressed in either kilometres or au. ILRS: laser ranging.

Nearby stars offer another possibility. Observe how a star’s apparent position against distant objects changes as the observer moves along an orbit. The measured angular shift and a known observing baseline allow its distance to be calculated geometrically. This is parallax. More distant stars have smaller shifts, which are harder to measure. A result expressed in light-years does not mean somebody timed light’s journey from that star with a stopwatch. ESA: how parallax works.

Branch · A numerical exampleOptional · you can skip this

Why divide by two?

Take a hypothetical return time $t=2.56$ s, rather than a measurement made here. Suppose we can neglect station and reflector motion and atmospheric effects at this level of approximation. The distance $d$ from the station to the reflector is:

\begin{aligned} d &\approx \frac{ct}{2},\\ &\approx 300\,000\times1.28\ \text{km},\\ &\approx 384\,000\ \text{km}. \end{aligned}

For this estimate, the speed of light is rounded to 300 000 km/s. Without dividing by two, we would obtain the entire light path, about 768 thousand kilometres. Real observations must account for additional effects and for the locations of the station and reflector: this is not yet a centre-to-centre Earth–Moon distance.

Precise methods have their own limits, corrections and uncertainty. A single universal technique does not cover stars and distant galaxies; we will examine geometry and other methods separately. Here the essential distinction is between what was observed and the units used to express the inference.

More digits do not improve a measurement

Imagine a hypothetical distance estimated as 2.0 ± 0.1 au. Here ± gives an uncertainty specified for the example: about five per cent. Converted to kilometres, this is about 300 ± 15 million kilometres. The unit has changed; the relative uncertainty remains about five per cent.

A calculator may display many digits after multiplying by the exact length of an au. Those digits cannot add information missing from the original measurement. A unit definition, rounding a written value, and observational uncertainty are different things. NIST: measurement results and uncertainty.

Try the idea: a distance is 0.5 light-years. Is that half a year of time or a length? Will expressing it in kilometres make the measurement more accurate?

It is a length — about 4.73 × 10¹² km. In a simple, stationary geometry, light in a vacuum would take half a Julian year to cross it. Converting the units does not itself improve the observation.

Astronomers therefore choose a unit that makes the scale easy to grasp: kilometres for nearby distances, au for planetary distances, light-years for interstellar ones. This is a convenient choice, not a ban on other units. When you encounter a number, ask separately: between which points, by which method, and with what uncertainty? Return to the structure of the Solar System to picture the region where the au is especially useful.