Three super common measures in space distance are the AU, the parsec and the light year. Although I understand the origins of these units, it seems odd to me that we don't use use standard metric units. For example, I was reading about the Psyche mission to an asteroid about 2-3 AU away. Why not just say 400 to 500 Gm. Or the nearest star is 4.25 light years away. Why not just say 40 Pm?

Surely one advance in science is the use a consistent measurement system across all, so that are aren't always tied up converting chains to furlongs to potrzebie.

  • 21
    $\begingroup$ The important thing about units of measurement is making them useful to the people who use them, and consistent in value. "A consistent measurement system across all" is a "nice to have." but takes a back seat when appropriate. You'll find similar non-SI units in frequent use all over the sciences, not just astronomy. As such, I think this is more of a hsm.stackexchange.com question. $\endgroup$
    – notovny
    Commented Apr 6 at 22:48
  • 9
    $\begingroup$ This is an interesting question, but I think Space SE is not the right site for it - meters (or km) and seconds are the most common units when folks talk about spacecraft. On the other hand, Astronomy SE might be a better place to ask about current usage, since it's astronomers who use all three of those distance units in common parlance. And if you'd like to explore "how did we get to this situation" then as notovny points out, HSM SE is a better site. So if this question is closed as off-topic, don't fret, good chance it would then be migrated $\endgroup$
    – uhoh
    Commented Apr 7 at 0:48
  • 13
    $\begingroup$ The SI system is built upon the magnitude of our lives, so it's only convenient at the magnitude of our lives. $\endgroup$ Commented Apr 7 at 4:54
  • 19
    $\begingroup$ Parsecs and AUs are just what's relatively easy to measure from Earth-based observations. Could you use terameters and petameters instead? Yeah, I guess, but unless you're going to feel a strong need to compare the distance to Mars to a road trip to Tahoe or the size of your bedroom, what actual benefit do you derive from it? $\endgroup$ Commented Apr 7 at 6:16
  • 6
    $\begingroup$ It might horrify you to hear that some astrophysics software packages and such still expect distances in centimeters — the field has had trouble moving away from cgs units. $\endgroup$
    – parasoup
    Commented Apr 7 at 18:18

7 Answers 7


From Kepler to the 1960's, the AU was the best measurement basis available for astronomy. Planetary motions could be expressed precisely in AU, but the relation of the AU to terrestrial units was only known to 4 significant digits at best. It took the development of interplanetary radar to change this.

The AU thus represents hundreds of years of measurements, while the meter has only become useful here within living memory. The parsec is derived from the AU.

  • 3
    $\begingroup$ @Kvothe The CGS (centimetre gram second) system dates to the late 19th century and well predates the adoption of SI (which amongst many other things, shifted the base units to metres kilograms and seconds). ie SI is relatively recent and is not synonymous with the various flavours of the metric system from which it evolved and standardised. Astronomy has been around long enough to still have occasional vestiges of non-SI units (like angstroms instead of the nearby nanometres, and ergs rather than joules). $\endgroup$ Commented Apr 8 at 11:48
  • 13
    $\begingroup$ It might be worth emphasizing that almost all distance measurements outside the solar system are in the end calibrated through parallax measurements. By expressing distances in parsec, the (historical) uncertainty in the relationship between AU and meters factors out. $\endgroup$
    – TimRias
    Commented Apr 8 at 12:20
  • 1
    $\begingroup$ @MichaelMacAskill, so why was CGS chosen in the 19th century. Did it make sense at the time? I was curious whether there was any good reason like there was for AU as explained in this answer. Edit: Okay it seems it was just the arbitrary universal standard of the time and somehow it survived in astronomy while most other sciences replaced it with the next (arbitrary) standard of (kgm and then) SI. $\endgroup$
    – Kvothe
    Commented Apr 8 at 12:28
  • 2
    $\begingroup$ @Kvothe I believe CGS, specifically the "Gaussian" variety, came to astronomy via theoretical atomic physics. Gaussian units are convenient for spherically symmetric electromagnetism, thus for atomic theory. In astronomy, atomic spectra are are large part of how we figure out what's going on. $\endgroup$
    – John Doty
    Commented Apr 8 at 13:01
  • 3
    $\begingroup$ @Kvothe all unit systems are arbitrary, even SI. (Maybe even especially SI, as advances in measurement and the need for continuity have resulted in redefining things that used to be nice tidy ratios, like the mass of 1L of water under standard conditions, to now be untidy rather untidy numbers). $\endgroup$
    – hobbs
    Commented Apr 8 at 13:47

In general, people find really enormous numbers hard to fathom, and this is true even for scientists. When using the metric system, it's not so bad hard to understand kilo-, mega-, and occasionally giga-, but beyond that things get murky. While it's common to use kilometers when referring to distances on or near earth, megameters and gigameters never caught on because they're not needed for everyday experiences.

So when referring to enormous distances, astronomers prefer to use more traditional units so they're dealing with smaller numbers. 1 AU = about 1.5e8 km, so it's a convenient unit for measurements within the solar system. For interstellar distances, light-years and parsecs are easier to work with. It's much nicer to say that Alpha Centauri is 4.3 light-years than 40.6 trillion kilometers (or 40.6 petameters).

One of the benefits of the metric system is how easy it is to switch among scales -- going from milimeters to meters is dividing by 1000, which just moves the decimal point. But this doesn't come up much when dealing with astronomical distances (although they do use metric prefixes to scale up to galactic and intergalactic distances, by using kiloparsecs).

Note that this differs between scientific disciplines. Particle physicists are happy referring to giga-electron volts. The differences are likely historical -- the metric system was well established by the time it became necessary to deal with these quantities. But astronomy is a much older science.

  • 1
    $\begingroup$ Even the electron volt isn't an SI unit, the SI unit is the joule, but even when talking about GeV that's still 10^-10 J which is an inconveniently small number $\endgroup$ Commented Apr 8 at 5:48
  • $\begingroup$ @AlanBirtles and by this point, using nanojoules or picojoules would feel weird as well. I generally feel that the SI-prefixes beyoud pico/Tera often are not helpful at all compared to just using an exponential notation (femto/Exa are already much less useful, and beyond that, there is mental conversion required all the time, with mixups as well). Naming things after these is fine, but actually using them is cumbersome. $\endgroup$
    – Chieron
    Commented Apr 8 at 11:24
  • 3
    $\begingroup$ @Chieron but all this is just a matter of what you're used to. Femtoseconds are completely normal to people working with pulsed lasers. If astronomers would stick to peta/exa/zettametres for a couple of years, they would soon find them just as intuitive as parsecs. With AU it's a different matter, more comparable to the Ångström in chemistry or the year in meteorology. $\endgroup$ Commented Apr 8 at 11:38
  • $\begingroup$ Regarding your first paragraph, it's worth noting that distances that could be expressed in megameters are more commonly expressed in thousands of km. So the Moon is said to orbit at an average distance of 384 400 km, not 384.4 Gm. While the latter is shorter and should be more practical, in real life, it requires more mental agility than the former (because we are going to convert it to something we know better anyway). $\endgroup$
    – jcaron
    Commented Apr 8 at 14:21
  • 1
    $\begingroup$ The distance of Alpha Centauri is known with a precision of only three digits, writing five digits does not make sense. So 4.34 light-years or 40.6 Pm, conversion to metric units does not increase precision. $\endgroup$
    – Uwe
    Commented Apr 8 at 17:28

Let's take the discussion to the distances of objects which astronomers are currently studying. Alpha Centauri at 40 Pm doesn't get a lot of attention anymore. The Galatic center is at 220 Em and the closest galaxy, Andromeda, is at 21 Zm. The Coma Supercluster of galaxies is around 2.8 Ym. Anything further out and you have to start using redshift to measure distance.

Don't know about you, but I had to look those last 3 prefixes up. You can find them in Haliday & Resnick, but not in Cutnell & Johnson (both popular first year physics textbooks) which tops out at tera. I don't carry the conversion for peta around in my head like I can for the smaller ones.

It's all a question of scale and the numbers that we can keep in our head. This is why we use kiloparsecs within the galaxy and megaparsecs for extra-galatic objects.

  • 2
    $\begingroup$ A couple of decades ago, I attended a talk on the electrodynamics of hypernovae by Roger Blandford. He was modeling accretion disks as electrical machinery with YA currents yielding ZV voltages. This seemed to vex the theorists in the audience ツ $\endgroup$
    – John Doty
    Commented Apr 8 at 17:36
  • 1
    $\begingroup$ As long as you treat astronomical units as "mind-bogglingly large distances" and not compare them with everyday's metres and kilometres, there's little difference between using (kilo- and megaparsecs) and (zetta- and yottametres). They are just names for units used only in astronomy. $\endgroup$
    – Trang Oul
    Commented Apr 9 at 6:41
  • 2
    $\begingroup$ Why would be any harder to learn these prefixes than learn AU, Parsec, and light year? $\endgroup$ Commented Apr 9 at 8:16
  • 3
    $\begingroup$ @JackAidley Without a mnemonic, I find it tricky to remember the seldom-used SI prefixes. On the other hand, the astronomical unit, parsec, and light year are all descriptive names that allude in some way to what's actually being measured. I might mix up a yottameter and a zettameter, but I know that a light year is much bigger than the astronomical unit we use to measure things on the scale of the solar system. There really is no outside information that can help you order the SI prefixes, while knowing what the non-SI units refer to can help you understand the scale. $\endgroup$ Commented Apr 9 at 13:41
  • 2
    $\begingroup$ What most people are missing here is that the distances are not measured in meters to begin with. Stellar distances are measured in parsecs (not light years) which are defined on the AU and that becomes your ruler by which you measure objects further away. If the value of AU changes due to better measurement, all of your old values in decades of published papers have changed. Distances are not converted to parsecs to make them easier to use. They start as parsecs and we avoid the conversion. (For the public, multiply by 3 and call it light years) $\endgroup$
    – Boyd
    Commented Apr 9 at 20:55

In this specific case, AU and parsec are directly relevant to how astronomers measure the distance to objects using parallax. You can research this yourself, but astronomers measure the position in the sky of an object, and then again six months later, when the Earth is on the opposite side of the Sun, and use the equation:

$$ d=~\frac{1~\rm{AU}}{\theta}$$

where $\theta$ is the angle between the two observed positions, and $d$ is the distance from Earth. If $d$ is measured in parsecs, the equation is simply

$$ d=~\frac{1}{\theta}$$

When astronomers write papers for other astronomers to read and analyze, it is more sensible to publish distances in pc, kpc, Mpc, etc – rather than pointlessly converting to some number like 120 trillion meters, only to have the reader convert back to a unit that he actually uses.

More generally, I think in speaking and writing about astronomy, we should actually lean more heavily into relative distances. Saying "the Moon is 384,000 km from Earth" conveys essentially no information to the reader, other than "really far away." It is comparing the positions of celestial bodies to the distance between my hometown and the neighboring town. But if I say, "the Moon is about 29 Earth diameters away" that gives you a pretty clear picture of the relative positions, and even the ability to build your own scale model of the Earth-Moon system.

Likewise, "the Voyager 1 probe is 24.3 billion km from the Sun" vs. "Voyager 1 is 164 AU from the Sun"..."Voyager 1 is 17 times further from the Sun than Saturn."

Astronomical distances are so beyond human imagination that the only way we can form any meaningful picture at each scale is in relative terms to the next-lowest scale. And even this can barely suffice in cases. E.g. The Milky Way diameter is about 100 million times the solar system diameter. That means if the solar system (to Neptune's orbit) were the size of a baseball, the Milky Way would be about the size of the Earth. No one can really picture the whole planet Earth or its size as a matter of experience – so even this direct ratio "object in question to object at next lowest scale" does not give us much of an intuition. But it is certainly better than comparing the size of the Milky Way to the size of a meter stick.


It depends on the context and scientific branch. Astronomy and related fields deal with distances so vast, that measuring them in run-of-the-mill Earth units, such as kilometers or miles, would result in an extremely large number! Likewise, the masses of celestial bodies are too big for the various types of ton used in contexts other than astronomy. Thus astronomy (and the other astro fields) has its own system for measuring the insanely immense sizes and distances involved. In other sciences, like quantum physics, sizes, distances, times, etc are so small, that even the tiniest of regular units of measurement are far too big to be useful without having long chains of decimals or fractions.

  • 3
    $\begingroup$ Extremely large numbers could be trimmed by using larger SI prefixes. For example distance to Proxima Centauri, about 4×10^13 km, could be neatly expressed as 40 Pm. During our lifetimes we've switched and very likely will switch again to larger prefixes for bytes, for the same reason. $\endgroup$
    – Trang Oul
    Commented Apr 8 at 12:04
  • $\begingroup$ @Trang Oul: Except that it would make all the historic astronomy literature that much less accessible to people newly entering the field. They would have to get used to measures on a scale they aren't used to (as they already do), as well as still having to pick up the historic units otherwise they'd have no hope of understanding material that in historic terms is not that old. $\endgroup$ Commented Apr 9 at 1:23
  • $\begingroup$ @SoronelHaetir Of course, such a sudden change will cause a lot of disruption. I just wanted to point out that it's not big numbers that would be an issue. $\endgroup$
    – Trang Oul
    Commented Apr 9 at 6:33
  • 1
    $\begingroup$ @TrangOul Actually, I believe that computing is the best public Si-Prefix training program we've ever had. Nowadays, a lot of people know that 1000 GB is the same as 1 TB which was definitely not the case before harddrives surpassed the 1 TB barrier. And we all know that a 1 Gbit connection is ten times better than a 100 Mbit connection. $\endgroup$ Commented Apr 9 at 9:25
  • $\begingroup$ @cmaster-reinstatemonica, it's good that computing spread the knowledge of larger SI prefixes, but at the same time spread the misconception that 1 kB = 1000 B instead of 1024. And while new 1024-based prefixes were invented (kibi-, mebi-, ...) to fix that, I've never encountered them in non-technical writings, let alone everyday speech. $\endgroup$
    – Trang Oul
    Commented Apr 9 at 10:23

I think it is some kind of mental inertia of the field.

Expressing distances in km (like 9.461e+12 km) is equally wrong, would be better written 9.461 Pm (peta meter). Maybe people are not used to handle mental comparisons in large SI prefixes, for example 1000 light years would be 9.4Em and a million light years would be 9.4Zm. Likely people need the same unit (light year) to perform comparisons, would be "hard" to operate with Z/E/P multiplier prefixes.

When I've learned physics I was told that a major source of mistakes is unit conversion, sometimes we confuse 1Km for a meter, second for hour, inch for a cm, and so on, especially when working with constants, like a density, the gravitational acceleration or submultiples of units, such as yards, feet and inches. SI provides a better handling of this class of mistakes.

And it didn't happen in school only, happens to the professionals too: https://www.latimes.com/archives/la-xpm-1999-oct-01-mn-17288-story.html

Computer science took the SI prefixes and transformed them in the power of two. A KB is 1024B, a MB is 1024*1024B and so on. That was caused by practical reasons, when building a memory chip it is convenient to align its size to a power of two. (1024=2^10). Later, the hard drive manufacturers (these are "linear" magnetic devices, they don't have the need to align to power of two), repurposed the old SI prefixes to advertise their capacities. In hard drives terms, 1MB is one million bytes. However, computing people don't have a problem to convert between EB, PB, TB, GB, MB and KB, I've never heard somebody saying "my database is five million MB", but rather they will say "5 TB". That's because we are equally used to all prefixes, especially in the big data field (some maybe don't use EB and PB so often).


Answer: Because those scientists are humans.

One of the reasons for the plethora of units is that human brains handle small and large numbers differently. There is an extensive discussion of this in “Wild Minds” by Marc Hauser. I’ll summarize: all animals, including humans, can intuitively count up to about 5. Beyond that, they resort to systematic naming or other formal strategies.

Animals need to count. If 3 lions walk behind a bush but only 2 walk out the other side, the gazelle they are stalking needs to know there is still one lion behind the bush.

If someone drops a few coins (up to 5) you don’t need to count them. You can just “see” there are, for instance, 4. But larger groups (say 12) need a different process. You could name them in a learned sequence (1,2,3,…) or visually divide them into 3 groups of 4. But it would be very difficult to just “see” it as a group of 12.

5 or 6 is the upper boundary of objects most people can “see” without reverting to counting. Dice have six sides, but the dots are in patterns to aid recognition. 5 is radially symmetric while 6 is bilaterally symmetric. If 5 and 6 were both dots in a circle (like the EU flag) identification errors would increase.

Same with fractions. Humans are good at estimating fractions down to about 1/5. That’s why measurements with traditional analogue instruments are accepted down to 1/5 of the scale marks (unless a Vernier scale is added).

Traditional units were usually developed so that quantities could be converted numerically between 1/5 and 5. That’s where all those weird units like chains, stones and gills came from. A standard British bar shot is 1/5 of a gill (1 oz) and a gallon in 4 quarts. It is easy to picture 5 shot glasses or 4 quart bottles.

Metric units usually use multipliers of 1000 between units. This makes conversions as easy as shifting a decimal point. But it makes it difficult to picture the resulting quantities. I can’t picture 183cm. But I can picture 6 feet. The person described may weigh 168 lb, which I can’t picture, but 12 stone is a bit easier.

I have nothing against the metric system. It is very rational and excels for many of the purposes I use it for. I cook in metric, build my house in imperial fractions and use machine tools in decimal inches. And I describe solar system distances in AU.

No matter how far we go in space, we are still human and still limited by our evolved mental abilities. People will continue to use unit systems which suit themselves as well as the application.

  • $\begingroup$ The phenomenon you are describing is called subitizing. And speaking of imperial units (lb, ft), I can't picture them, because I grew outside the US and I'm used to the metric system. On the other hand, I can easily picture someone 183 cm high (by comparing to my height). It's just a matter of familiarity. $\endgroup$
    – Trang Oul
    Commented Apr 10 at 6:56

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.