Skyglow

Why can every outdoor lamp be individually reasonable and yet nobody can see the stars?

Pick any single outdoor lamp and ask whether it should exist. A porch light, a car park pole, a floodlit forecourt, a bollard on a footpath. Each one is defensible on its own terms, and each one, switched off tonight, would restore almost exactly nothing to the sky.

Yet the sky over most of the inhabited world is now bright enough that the Milky Way is invisible from where the great majority of people live. Every step in the chain was reasonable, and the outcome is one nobody chose. That gap between defensible parts and an unwanted whole is the thing worth understanding.

Start with the physics, because the sky is not glowing by itself. Light that leaves a lamp and travels upward meets air. Air molecules scatter it — Rayleigh scattering, strongly favouring short wavelengths, which is why blue-rich sources contribute more skyglow per lumen than warm ones. Aerosols, dust and water droplets scatter it too, less selectively but often more strongly near a city. Some fraction of that scattered light comes back down towards your eye from every direction at once. A luminous dome, not a beam.

Ask where the upward light comes from, since lamps are meant to point down. Two sources. Some is emitted directly above the horizontal by an unshielded fixture. The rest is light that did its job, landed on ground, walls or vehicles, and bounced. Asphalt reflects something like a tenth of what hits it, pale concrete rather more. So even a perfectly shielded installation returns light to the sky, and shielding reduces the problem without abolishing it.

Now the crucial step. What does one lamp add to that dome? Essentially nothing measurable — its contribution is buried far below the natural airglow and starlight background. But what does the dome equal? To a good approximation, the sum of every lamp's upward flux, weighted by distance. Skyglow adds. It does not saturate, and it does not average.

Hold those two facts together, because that is the whole trap. The quantity each decision-maker can perceive is their own marginal contribution, which is genuinely negligible. The quantity that determines whether anyone sees stars is the total, which no single decision-maker controls or is billed for. So "is my lamp a problem?" and "are lamps a problem?" have opposite honest answers, and the first is the only question anyone is ever asked.

Two features make it worse rather than better. The dome travels: a city's glow is visible from tens of kilometres away, falling off with distance but not respecting any boundary, so the cost lands mostly on people who installed nothing. And there is a feedback. A brighter background makes a given lamp less effective, and the reasonable response — brighter, whiter, more of them — feeds the background it was responding to. The switch to LEDs, which made a lumen far cheaper, was widely expected to cut emissions and instead often coincided with more lumens installed; satellite measurements over the 2010s found lit area and radiance growing, though the instruments then in orbit were blind to the blue part of white LED output, so the true trend is likely understated rather than exaggerated.

So the structure is not a physics failure or a moral failure. It is a collective-action structure: individually rational choices, a summed and shared outcome, a benefit that is private and a cost that is not.

Three corrections worth making.

The first is that skyglow is not the same as glare or light trespass, and the fixes differ. Glare is direct light in your eye; trespass is light landing where it was not wanted. Both are local and both have an obvious complainant. Skyglow is the diffuse remainder, and its distinguishing feature is precisely that it has no local complainant — which is why it is the one that persists.

The second is that "just switch things off" misreads the structure. The load is not evenly spread: a modest number of over-bright, unshielded, blue-rich, all-night installations do disproportionate work, because contribution scales with upward lumens rather than with fixture count. That is what makes shielding, warmer colour temperature, lower levels and curfews effective without darkening anything anyone needs lit.

The third is that this is not an argument that any individual was wrong. In a summed commons, the individual reasoning is sound and the aggregate is still bad — which is why the remedies that work are the ones that operate on the aggregate directly, in the form of ordinances, standards and procurement rules rather than exhortation. Note the honest gap: how much artificial light is needed for safety, and whether more of it reduces crime or collisions at all, remains genuinely contested in the evidence.

• How the Bortle scale and sky-quality meters put numbers on what has been lost.
• Why blue-rich white LEDs scatter more per lumen, and what colour temperature buys back.
• The ecological side: migration, insect populations, and nocturnal predation under an artificially bright sky.
• Whether skyglow is genuinely reversible — unusually among pollutants, it stops the night the lamps do.

Every term the collection defines is gathered in the glossary.