Planet colors: all nine measured from published albedo spectra

Every diagram of the solar system tells you the same thing. Mars is red, Jupiter is banded orange, Saturn is butterscotch, Uranus is pale cyan and Neptune is a deep royal blue. It is such a settled picture that nobody thinks to check it.

It is also wrong, and not by a little. The planets are among the least colorful objects anyone has ever photographed. I ran all nine through the same pipeline used for the moon and the sun: published geometric albedos in twelve photometric bands, the extraterrestrial solar spectrum as the illuminant, the CIE 1931 observer, and a perfect white Lambertian disc in the same sunlight as the reference white. Here is what came out.

What color are the planets

These are disc averages at zero phase angle, which is the fully lit face you would see from directly Sunward. Each swatch is the true brightness as well as the true hue, so they are directly comparable to one another and to the grey of the page.

Nothing there is vivid. The chroma column runs from 3.05 to 22.04 with a mean of 11.59. For scale, the blue primary on the screen you are reading this on has a chroma of 135.3, the green primary 120.6 and the red primary 101.8. Mars, the most saturated object in the solar system, does not reach a quarter of the weakest of those.

Put it the other way round. Sample the sRGB cube evenly and 91.2 percent of the colors in it are more colorful than Mars. 98.8 percent beat Mercury. 99.8 percent beat Venus. The planets live in the sliver of color space that color pickers put in the corner and nobody selects on purpose.

They are not even using the room they have

Low chroma can be an artefact of low or high lightness, since a very dark or very pale color cannot be saturated in the first place. So I asked a fairer question: at each planet’s own lightness and its own hue angle, what is the most saturated color sRGB can produce, and what fraction of that does the planet actually use?

• Mean across the nine: 20.5 percent of the available chroma.
• Best case, Mars: 37.3 percent.
• Worst case, Venus: 3.6 percent, which is a white cloud deck doing exactly what a white cloud deck does.

Even after correcting for brightness, the planets sit at a fifth of the saturation their own lightness permits. This is not a trick of the exposure. It is what reflected sunlight off rock, cloud and gas looks like.

Where the numbers come from

The input is Mallama, Krobusek and Pavlov’s 2017 reference photometry, which reports geometric albedos for all eight planets in seven Johnson-Cousins bands and five Sloan bands. A geometric albedo is exactly the quantity needed here: the ratio of the planet’s flux to that of a perfectly reflecting Lambertian disc of the same size at the same distance. It is a reflectance, band by band.

Six of those bands have effective wavelengths inside the visible range, at 438, 469, 545, 617, 641 and 700 nm, with two more just outside on each side. I placed the albedos at their effective wavelengths, interpolated with a shape preserving monotone spline so the curve cannot ring between points, multiplied by the ASTM G173 extraterrestrial solar spectrum, and integrated against the CIE 1931 two degree color matching functions. Pluto is not in the Mallama tables, so it comes from Buratti and colleagues’ B, V and R albedos of 0.44, 0.52 and 0.58.

The white point is a perfect white disc in the same sunlight. That is the right choice rather than a convenience, because it is what your visual system would do if you were out there: sunlit white is white, wherever you are standing.

The check that matters

Two of those photometric systems are independent. Johnson-Cousins and Sloan were measured with different filters, different detectors and partly different observing campaigns. If the method is sound, running each system on its own should land in the same place.

• Largest disagreement between the two systems, Saturn: 5.33 CIEDE2000 units.
• Smallest, Mars: 1.43.
• Median across the eight planets: about 2.7.

The median gap between two planets, by contrast, is 19.9 units. So the measurement noise is roughly seven times smaller than the signal it is being used to read. For what those units mean in practice, the CIEDE2000 explainer covers the formula and the piece on just noticeable difference puts the threshold for two patches side by side at about 2.3.

Uranus and Neptune are the same color

This is the result that should end an argument. Measured from the albedo tables and compared in CIEDE2000, the two ice giants are 3.78 units apart. That is under two just noticeable differences. Strip the brightness out and compare hue and chroma alone and the gap is 3.98.

Neptune is very slightly darker and very slightly bluer. Nothing else separates them. Set the two swatches on a wall and most people would call them the same paint.

This is not a novel claim on my part, which is the encouraging bit. Irwin and colleagues reached it in 2024 by an entirely different route, reprocessing Voyager 2 and Hubble imagery against modern spectral models, and concluded the two planets are a similar shade of greenish blue. The familiar deep azure Neptune comes from Voyager 2 frames whose contrast was stretched to bring out cloud features. The stretch was never hidden. It just outlived the caption.

Their physics is shared too, and it shows up in the reflectance curve. Methane in both atmospheres absorbs hard in the red. Between 500 and 700 nm the albedo falls by a factor of 2.68 on Uranus and 2.76 on Neptune, against 1.06 on Jupiter over the same span, where the methane sits below a deck of cloud that hides it. Chop the red off a white spectrum and you get cyan. That is all the ice giants are doing.

Which is why neither of them is blue

Hue angle is the cleanest way to say this. Uranus lands at 226.1 degrees, which matches the CSS color skyblue to within 1.3 degrees. Neptune lands at 238.2. Earth lands at 260.9.

Earth’s hue is 22.7 degrees further round toward blue than Neptune’s and 34.8 further than Uranus’s. The planets sold as the blue ones are cyan, sitting near the boundary that the teal piece found separates green from blue by lightness rather than hue, and the only planet whose hue is genuinely blue is the one with the oceans on it.

Earth pays for that hue in chroma. At 8.79 it is fifteen times less saturated than the sRGB blue primary, and the measured disc, #ACAEB9, is a cool grey that happens to lean blue. Carl Sagan’s phrase was better calibrated than the posters that followed it. He said pale blue dot, and pale is doing most of the work. The mechanism behind what little blue there is gets its own treatment in the ocean piece and the sky piece, since both contribute.

The pair nobody expects

Uranus and Neptune are the famous near-twins. They are not the closest pair in the solar system. Of the 36 possible pairings, the smallest CIEDE2000 gap is between a gas giant 143,000 km across and a dwarf planet you could drive around in a fortnight.

2.80 units. Around one JND. Jupiter and Pluto are, to the eye, the same pale warm grey, and they arrive at it from completely different chemistry: ammonia cloud tops on one, nitrogen and methane ice with a tholin haze on the other. It is a nice small instance of the thing metamerism is really about, which is that color is a three number summary and very different spectra can collapse onto the same three numbers.

Eight of the 36 pairs sit under 10 units. Two sit under 5. The solar system is not a palette. It is a handful of near neighbours with Mars and Mercury off on one side and nothing much anywhere else.

The whole system is two colors

Sort the nine by hue angle and they fall into two groups with a hundred and twenty five degree hole between them.

• Warm: Mars 73.7, Saturn 77.4, Mercury 80.9, Pluto 85.8, Venus 96.0, Jupiter 100.8. Six planets inside a 27 degree window.
• Cool: Uranus 226.1, Neptune 238.2, Earth 260.9.

Six of nine occupy 7.5 percent of the hue circle. There is no green planet, no purple planet, no magenta planet, nothing in the yellow green, nothing at all between 101 and 226 degrees. Compared with the even spacing of a twelve spoke color wheel, the solar system is a badly drawn wheel with two spokes.

The reason is that there are only two ways to be a planet color. Either your reflectance rises toward the red, which is what rock, oxidised iron and organic haze all do, or methane takes the red away and you go cyan. The slope from 400 to 700 nm tells the whole story: 4.22 on Mars, 2.32 on Saturn, 1.88 on Mercury, 1.43 on Pluto, 1.39 on Venus, 1.29 on Jupiter, then 0.71 on Earth, 0.38 on Uranus and 0.32 on Neptune. A rising line reads warm, a falling line reads cool, and there is no third option out there because there is no planetary process that makes a hump in the middle of the visible band.

Mars is not the red planet

Mars has the steepest slope of the nine and still does not get to red. At hue angle 73.7 degrees it is in the orange-yellow, and the measured disc, #927252, sits 3.5 CIEDE2000 units from raw umber. Coffee with milk. A muddy brown a decorator would call greige.

The color does come from iron oxide, so the folk explanation is right about the chemistry and wrong about the outcome. Fine dust scatters enough across the whole visible band to wash the red out, and the reflectance at 400 nm is 0.068 against 0.288 at 700 nm, which sounds dramatic until you notice both numbers are small. Mars is dark first and orange second. On a clear day the sky above that dust is the inverse of ours, which is its own story and the same physics the sunset piece works through.

Turn the exposure up and the hue appears

There is a fair objection to all of this, and it is worth answering properly. The swatches above are dark, and every photograph you have seen of these planets was brightened. So maybe the dullness is just the exposure. It is not, and the reason is worth seeing rather than being told: here is each planet twice, the left half at its true brightness and the right half scaled until its brightest channel is at full, which is all an exposure adjustment can do.

Brightening makes them lighter, not richer. Every right hand half is a pastel. Exposure moves a color along the lightness axis and cannot manufacture chroma, so the only way to get the Neptune of the posters is to multiply the saturation, and the multiplier needed to reach something as ordinary as CSS royalblue is 4.3 times. That is not a brighter photograph. That is a different color.

Your eyes would still work out there

The other thing worth ruling out is whether any of this is visible in person, because the far planets are a long way from the lamp. From the same solar spectrum, full sunlight above the atmosphere is 133,100 lux, and it falls with the square of distance.

• Jupiter: 4,917 lux, disc luminance about 851 cd/m².
• Saturn: 1,463 lux, about 246 cd/m².
• Uranus: 361 lux, about 53 cd/m².
• Neptune: 147 lux, about 19 cd/m².
• Pluto: 85 lux, about 14 cd/m².

Cone vision needs something above roughly 3 cd/m² to work properly. Even at Pluto, where sunlight is 1,559 times weaker than at Earth, the lit disc clears that by more than a factor of four. Sunlight at Pluto is about as bright as a well lit corridor, and daylight on Neptune is around the level of a supermarket aisle. Color vision holds up all the way out. The washed out palette is genuinely the palette, not a side effect of the dark.

Why the wrong picture stuck

Three things, and none of them is anybody lying.

The first is that spacecraft imagery is built for science. Filters get chosen to isolate a methane band or an ammonia feature, then composited into false color, then contrast stretched so that a two percent variation in cloud brightness becomes a visible band. The result is published with a caption explaining exactly that, the caption is dropped, and the image becomes the planet.

The second is that a poster of nine grey-beige circles is a bad poster. Illustrators need the planets to be distinguishable at thumbnail size, and the honest version is eight thumbnails you cannot tell apart plus Mars. Saturation is the only lever available.

The third is the interesting one, and it is about memory rather than publishing. Color memory drifts toward the category name. Once you have filed Neptune under blue, what you recall on demand is a prototypical blue, not the pale cyan you saw. The pull is measurable and it is the same effect behind memory colors, where people reliably remember bananas as more yellow than any banana has ever been. The posters exaggerate, and then memory exaggerates the posters.

You can feel the size of that drift directly. Look at the Neptune swatch above for five seconds, close the tab, and try to pick it again an hour later. Almost everyone reaches for something far too saturated. That gap between the color in front of you and the color you can retrieve is the whole premise of the color memory game, and the gradient mode is the version that drills exactly this, near neighbours of one hue held in your head and matched back. If you want the deliberate practice version, training your eye for color lays out how.

The table, for reference

All nine, hex codes measured rather than picked. If you want to know how to read the codes themselves, the hex explainer covers it, and the hex mode turns it into a game.

• Mercury #77695C, a dark warm grey. Airless rock, weathered.
• Venus #E3D6CD, near white with the faintest cream. Sulfuric acid cloud.
• Earth #ACAEB9, cool grey leaning blue. Half cloud, half ocean.
• Mars #927252, milky brown. Iron oxide dust.
• Jupiter #CEC1B1, pale warm grey. Ammonia cloud tops.
• Saturn #D9BB9F, light tan. The most butterscotch of the nine, and still barely.
• Uranus #9EB9C2, dusty cyan. Methane absorption.
• Neptune #93B0C1, dusty cyan, marginally deeper. Same methane.
• Pluto #CEBDAD, pale warm grey. Nitrogen ice under tholin haze.

Nine hex codes, eight of which you would struggle to name without a chart, which is roughly the position the name that color mode puts everyone in. The solar system is beige and cyan. It has always been beige and cyan. The paint was never the thing that made it worth looking at.