Photos of the aurora often show vivid green curtains fringed with pink, or entire skies awash in deep red — colours that can look almost unbelievable compared with what a first-time viewer actually sees standing outside. Both the colours and the gap between camera and eye have a specific, well-understood physical explanation, rooted in which atom or molecule is glowing, at what altitude, and under what quantum mechanical rules.
The basic mechanism: collisions, not combustion
Aurora is not combustion, lightning, or any kind of chemical reaction — it’s the result of energetic charged particles, mostly electrons, funnelled down Earth’s magnetic field lines into the upper atmosphere, where they collide with oxygen and nitrogen atoms and molecules. Those collisions transfer energy into the atmospheric particles, temporarily exciting their electrons into a higher energy state. When an excited electron drops back down to its normal state, it releases that energy as a photon of light at a specific wavelength — a specific colour — determined by which element or molecule was excited and by how much energy it absorbed.
This is fundamentally the same physical process (atomic emission) behind neon signs and sodium streetlights, just powered by particles from space rather than an electrical current in a sealed tube. The specific colours aurora produces, though, depend on something more subtle than just “which atom got hit” — they depend on exactly which excited state that atom ends up in, and how quickly (or slowly) that particular state is allowed to release its energy.
Forbidden transitions: the key to understanding altitude
Both of the two dominant aurora colours — green and red — come from the same element, atomic oxygen, but from two different excited states, and both happen to be what physicists call “forbidden” transitions. In quantum mechanics, certain changes in an atom’s electron configuration are strongly disfavoured by the selection rules that govern normal (electric dipole) light emission — not literally impossible, but vastly less probable than an ordinary transition, meaning the atom can remain in that excited state for a comparatively enormous length of time (seconds to minutes, rather than the nanoseconds typical of an ordinary atomic transition) before it eventually happens anyway.
This matters enormously for aurora because the upper atmosphere isn’t a vacuum — it has a real, if thin, density of surrounding particles, and every so often, an excited atom collides with one of them. If a collision happens before the slow, forbidden emission has occurred, the atom loses its excitation energy to that collision instead (a process called collisional quenching) and no photon of that specific colour is ever emitted. Which of these two things happens first — the slow forbidden emission, or a collision — depends entirely on how frequently collisions occur, which depends directly on air density, which depends directly on altitude.
Green: oxygen at low-to-mid altitude
The characteristic aurora green comes from atomic oxygen in an excited state called O(¹S), emitting light at 557.7 nanometres via a forbidden transition with a natural lifetime of a little under a second. At altitudes of roughly 100 to 300 kilometres, the atmosphere is thin enough — and collisions therefore infrequent enough — that this roughly one-second window is usually enough time for the green emission to actually occur before a collision interrupts it. This altitude band also happens to be where the largest number of incoming particles have enough energy to reach and excite oxygen in the first place, which is the second reason green dominates: it’s both the most probable emission at that altitude and the altitude most commonly reached. The human eye’s rod cells, responsible for most low-light vision, are also particularly sensitive to wavelengths in this blue-green range (the eye’s peak scotopic — low-light — sensitivity sits close to 500 nm), which is part of why faint green aurora is disproportionately easy to perceive compared with other colours at the same actual physical brightness.
Red: oxygen at high altitude, and why it needs the extra height
Higher up, above roughly 300 kilometres, a different excited state of atomic oxygen, O(¹D), is responsible for the deep red light seen at 630.0 and 636.4 nanometres — again a forbidden transition, but a far more strongly forbidden one, with a natural lifetime on the order of two minutes rather than under a second. That much longer lifetime is exactly why this red emission is confined to high altitude: down at 100-200 km, the atmosphere is dense enough that a two-minute wait is essentially guaranteed to end in a quenching collision long before the red photon could ever be emitted — the red transition effectively never gets to finish at low altitude. Only above roughly 300 km, where air density has dropped enough that collisions become rare on the timescale of minutes rather than milliseconds, does the O(¹D) state survive long enough to actually emit its red photon before something else interrupts it. This is also why red aurora typically appears as a diffuse glow sitting above and behind an active green display, marking the upper reaches of the same precipitating particle event, rather than as an independent structure of its own — and it generally requires more energetic particle precipitation overall to produce a genuinely bright red display, which is part of why an intensely red aurora is often associated with the stronger end of a substorm (see substorms explained).
Why aurora seen far from the poles often looks red rather than green
This altitude split has a second, less obvious consequence that matters specifically for the mid-latitude, storm-only cities covered on this site. Because Earth is curved, an observer far enough from the aurora — say, someone in London or across much of the northern United States during an exceptional storm — is looking at a display that is actually occurring hundreds of kilometres away, over the horizon to the north. At that distance, the lower, brighter green-emitting layer (100-300 km up) can be entirely hidden below the curve of the Earth from the observer’s vantage point, while the taller red-emitting layer (300+ km) remains just tall enough to peek above the horizon. This is a well-documented reason historic accounts of aurora witnessed from unusually low latitudes — including some descriptions from the 1859 Carrington Event (see what is a CME) — describe an eerie, blood-red glow on the horizon rather than the classic green curtains associated with high-latitude viewing from directly underneath the display.
Pink, purple and blue: nitrogen joins in
Along the rapidly moving lower edge of a bright, energetic display — usually below about 100 kilometres, where only the most energetic incoming particles penetrate — molecular nitrogen (N₂) and ionised molecular nitrogen (N₂⁺) contribute additional emission that oxygen alone doesn’t produce. Excited neutral N₂ contributes red and pink-toned light through its “first positive band” system, often blending with any oxygen red present, while ionised N₂⁺ contributes a distinctive blue-violet emission around 391.4 nanometres through its “first negative band” system. Combined, this nitrogen emission tends to appear as pink-to-purple fringes or rippling edges at the very base of an otherwise green curtain, and it’s strongly associated with the more energetic, fast-moving substorm phase of a display rather than a calm, static glow, simply because reaching sub-100 km altitude at all requires unusually energetic precipitating particles.
Why your eyes and your camera disagree
Human vision in low light relies mostly on rod cells, which are far more sensitive to dim light than the eye’s colour-sensing cone cells, but rods are essentially colour-blind — they register brightness, not hue. That means a faint aurora display can appear as a pale grey-green or even just a washed-out white-grey band to the naked eye, even though it’s genuinely emitting the same green and red light a camera would register in full colour; you are, in that moment, functionally colour-blind in the same way anyone is under sufficiently dim illumination. A camera sensor doesn’t have this limitation: given a long enough exposure (commonly several seconds, discussed in aurora photography: camera settings that actually work), it accumulates light over time and reveals colour saturation the eye simply isn’t built to perceive in real time, especially for fainter displays.
This is a genuine, well-documented gap in human versus camera vision — not a sign that photographers are exaggerating or that a “real” aurora looks nothing like the display you saw. A bright, fast-moving substorm, though, genuinely can show vivid colour to the naked eye, since a strong enough display provides enough total light for human cone cells (and therefore colour vision) to engage at least partially — which is part of why the difference between a faint glow and a genuinely active display, whether you’re watching from Fairbanks, Abisko or Kiruna, matters so much to how memorable a night turns out to be.
Frequently asked questions
Why is aurora usually green?
Green comes from oxygen atoms at roughly 100-300 km altitude, the most common altitude band that incoming particles reach with enough energy to excite oxygen into emitting light at 557.7 nanometres — a wavelength the human eye is also particularly sensitive to, which is part of why green looks so bright.
Why do some displays look red instead of green?
Red aurora comes from oxygen at a higher altitude, above roughly 300 km, where the atmosphere is thin enough that oxygen atoms take longer to release their absorbed energy — long enough that the deep-red 630 nanometre emission has time to occur before a collision interrupts it. This typically requires more energetic particle input.
Is the aurora ever really pink or purple to the naked eye?
Occasionally, yes, particularly along the lower edge of a bright, fast-moving display, where nitrogen molecules contribute blue-violet and pink-red light. More often, though, cameras — especially phone cameras with long exposures — render colour, including pink fringes, far more vividly than the naked eye perceives it in the moment.
Why do photos of the aurora look more colourful than what I saw with my own eyes?
Human night vision relies heavily on rod cells, which are colour-blind and most sensitive in low light, while a camera sensor integrates light over a longer exposure and can register colour the eye simply cannot process fast enough or sensitively enough to see. This gap is real and well understood, not exaggeration or fakery.
What is a "forbidden" atomic transition, and why does it matter for aurora colour?
It's a transition between energy states that quantum mechanical selection rules make very unlikely (hence 'forbidden'), so the excited atom sits in that state for an unusually long time -- seconds to minutes -- before finally emitting a photon. Both the green and red oxygen lines are forbidden transitions; it's exactly that long-lived, easily interrupted excited state that makes their appearance so altitude-dependent, since a collision with another particle can knock the atom out of the excited state before the slow forbidden emission has a chance to happen.
Why do aurora seen from far mid-latitude, storm-only cities often look red rather than green?
Earth's curvature is the reason. From somewhere like southern England or the northern US, the lower, brighter green-emitting layer of a distant aurora sits below your local horizon, physically blocked from view by the curve of the planet -- only the higher-altitude red-emitting layer is tall enough to still be visible above the horizon at that distance, which is why historic accounts of aurora seen far from the poles often describe an eerie red glow rather than green curtains.
Does the aurora ever appear blue?
Yes, though it is less common and generally fainter than green or red -- blue and violet light comes from ionised molecular nitrogen (N2+) at the lowest altitudes reached by the most energetic incoming particles, typically below about 100 km, and tends to appear as a fringe at the very base of an intense, fast-moving display rather than as a dominant colour on its own.
Why does the human eye struggle to see aurora colour even when a camera captures it clearly?
Human colour vision (via cone cells) requires more light than most aurora displays provide at night, so low-light vision instead relies on rod cells, which are far more light-sensitive but register no colour information at all -- only brightness. A faint aurora that a camera renders in vivid green and red may appear to the naked eye as simply a pale, colourless grey-white glow, because you are, in that moment, functionally colour-blind.