Want to measure habitability? You can't.

People. We need to have another chat.

We have all these exoplanets. Some of them are deliciously Earth-sized. And you want to know which are the most habitable. But here’s the thing…

You can’t.

NOT POSSIBLE. CAN’T BE DONE. DENIED. I don’t even care what you read in the Daily Mail last week.

There’s no way of creating a quantitive scale that actually measures how capable a planet is of supporting life. No matter how these habitability metrics are constructed They. Just. Don’t. Measure. Habitability.

Why? I’m so glad you asked. Take a seat. This ain’t going to be difficult.

First let’s crush the ultimate temptation: Why isn’t an Earth-sized planet definitely habitable? Because this:


Our Solar System has two Earth-sized planets: the Earth and Venus. Venus is 95% the size of the Earth, yet its surface would melt lead. And you.

Every time you see “Earth-sized planet” and feel tempted to pack a suitcase, remember this is equivalent to saying “Venus-sized” and see if you still feel like quaffing down an out-of-this-world Starbucks coffee.

Buuuuuuut (I hear you scream) there are so many more planet properties we can measure to calculate the perfect exo-real estate!


While we’ve discovered nearly 4,000 exoplanets, the amount known about each world is pretty sparse. For most of the planets out there, we only know radius and the amount of radiation the planet receives from the star. Just two properties.

There are three important points to realise at this juncture:

  1. Two is a small number. If you had only two jelly beans, that would be sad.

  2. No matter how complex you make your habitability metric, it can only depend on two measurements. Any other quantities will need to be estimated from those two.

    This is like designing a skin-tight superhero suit with exact measurements for torso length, foot size, knee height, shoulder width, arm and leg length, waist, hip and head size but actually just measuring your waist and height and estimating the rest from those two measurements. Sounds doable? Ask any woman about getting jeans to fit.

  3. The two properties you can measure are not directly related to habitability. Whether a planet can support life depends on the surface environment (or potentially sub-surface for life underground). Venus is proof that the radius and level of starlight really tells you jack. Not only does the planet have nearly the same radius as the Earth, but it receives only about twice the amount of starlight. If Venus had an Earth-like atmosphere, then the surface temperature would be somewhere in the 50°C (~ 120°F) range. Certainly a little toasty, but a far cry from the true surface temperature of 460°C (fuck no°F). The difference is Venus’s thick carbon dioxide atmosphere, which traps heat like…

    … seriously, nothing on Earth comes close. That’s the point.

OKAY OKAY OKAY. WHAT IF we could measure more? While our current telescopes have been focussed on finding these planets, the next generation may be able to discover properties such as if they have surface water or identify gases in their atmospheres. If we could measure a few basic properties about the surface environment, could we create an accurate habitability scale?

Here is your problem: this is an accurate diagram of all the planets we know support life:


The only planet we know that supports life is the Earth. That gives us just one data point. And you can’t make any kind of meaningful scale with just one point.

Don’t believe me? Let’s give that a go.

We know a planet can be habitable if it is the same size as the Earth (even if that isn’t always true). But does a planet become more or less able to host life if you change the size? Without more habitable planets of different sizes to compare with, the scale could go up or down as you increase in size.


But but but (you shout quite reasonably) we do know of many uninhabitable planets. Can’t we use this to constrain the radius? You’re absolutely right and we can… a bit.

Based on a handful of planets that have both radius and mass measurements, astronomers spotted that planets larger than about 50% larger than the Earth were typically quite low density. This suggests those planets have very very thick atmospheres, like Neptune.

Neptune is definitely not habitable. Let’s use a “50% larger” as a size limit. The problem is that y’all wanted a habitability scale. So these two points have to be connected by…. what?


The Earth is the largest rocky planet in our Solar System and Neptune is the smallest gaseous world. We don’t have any data points to tell us how the environment would change as the Earth increases in mass. A slight increase in mass might not change habitability at all. Or the extra gravity might flatten our topology, immersing the land under water and cutting off weathering, to result in the planet freezing into a snowball of death while the Sun was still young. Whichever. We’ve absolutely no way of knowing.

The best we can do is say 50% larger = probably bad. Below that…. pray to your gods.

What this means is when you see a fancy-smancy habitability law that looks all like:


You need to be thinking like the embittered, twisted and cynical soul you know you really are and say coldly:

  1. R, M, D and T are likely based on only two actual measurements, neither of which directly relate to the surface property.

  2. The values a, b, c and d supposedly tells you how the habitability (H) changes as you vary that property (R, M, D and T). But there’s no way we have a damned clue what they should be.

  3. The resultant H is as meaningful as my star sign. Now go get this cancer a beer.

When hunting for planets that might be habitable, the best that can be done is to slam down some limits. A planet larger than 50% that of the Earth will probably have choked its surface under a Neptune-like fug. A planet that receives far less starlight than us risks being too cold for talking. Or living. Not good options for scoping out alien neighbours.

What we can’t do is develop a scalable index for habitability and expect a planet with H = 0.7 to be less likely to support life than a planet with H = 0.75.

Can’t be done on just two measurements and one example of life.

Pretty obvious when you think on it, right?