UVA on cloudy days: how much gets through

"It's overcast, I'm fine" is one of the most reliable ways to underestimate your ultraviolet exposure — and it fails specifically for UVA. Cloud is very good at removing the things you can sense, brightness and warmth, and comparatively poor at removing the long-wave ultraviolet you cannot. On a grey day you have lost the warning signals while keeping much of the dose.

Transmission by cloud type

"Cloudy" covers everything from a faint high veil to a black storm base, and the difference between them is more than an order of magnitude. What matters physically is not how much sky is covered but the optical depth of what is covering it — how much liquid water sits in the column between you and the sun.

Note the third row. Under scattered cloud with the sun in a gap, the figure can sit above 100% — that is the enhancement effect, covered below and not a typo.

Why UVA gets through cloud better than UVB

Cloud attenuates UVA less than it attenuates the sunburning UVB, and less again than it attenuates visible light. Three mechanisms are at work.

• Cloud droplets scatter rather than absorb. Water droplets are large compared with UV wavelengths, so they mostly redirect ultraviolet instead of destroying it. A photon that enters a cloud has a good chance of eventually emerging downward — just from a different direction. This is why an overcast sky is uniformly bright rather than dark: the energy is still arriving, it has simply lost its direction.
• Scattering lengthens the path, and a longer path hurts UVB more. Every bounce inside a cloud increases the total distance a photon travels through the atmosphere, and therefore its chance of meeting an ozone molecule. Ozone absorbs strongly in the UVB and only weakly in the UVA, so the detour is far more costly for UVB. The band that penetrates a straight path better also survives a crooked one better.
• UVA is already mostly diffuse. Even under a perfectly clear sky, a large share of surface UVA arrives as scattered skylight from the whole dome rather than straight from the solar disc. Cloud primarily removes the direct beam — so it takes away a smaller proportion of UVA than of visible light, most of which does come straight from the sun. The same fact explains why shade offers less UVA protection than it appears to: standing under an umbrella removes the sun, not the sky.

The cloud enhancement effect

This is the part that surprises people: a partly cloudy sky can deliver more UVA than a clear one.

The geometry is simple. When the sun sits in a clear gap between bright cumulus clouds, you receive the entire direct beam, exactly as you would under a clear sky — plus additional radiation scattered off the brilliant sunlit flanks of the clouds around the gap. Nothing is subtracted and something is added. Measured enhancements are commonly in the 5–25% range, with short-lived spikes recorded higher, particularly when a bright surface such as snow is also bouncing light back up from below.

The effect is episodic and short-lived: it applies while the sun is in a gap, and gives way to genuine attenuation when a cloud passes in front. Averaged over an afternoon, broken cloud usually nets out a little below clear sky. But the peak instantaneous values are real, and they are not what anyone expects from a cloudy forecast.

What this changes in practice

• Overcast is not a day off. At 30–60% transmission, an outdoor afternoon under grey skies in midsummer can still deliver more UVA than a clear day in midwinter. See why UVA changes so little with the seasons for the arithmetic.
• You will not get the usual warning. Cloud strips UVB faster than UVA, so the sunburn that normally signals overexposure is suppressed more than the UVA that causes photoaging and pigmentation. Cloud shifts the exposure mix toward UVA in the same way glass and low sun angles do.
• Cloud cover percentage is a weak predictor. "90% cloud cover" in a forecast describes how much sky is obscured, not how thick the obscuring cloud is. Ninety percent thin cirrus and ninety percent nimbostratus differ by a factor of five in what reaches you.
• Photosensitive people should treat cloud as almost irrelevant. If your threshold is low, a 40% day still crosses it. That is covered in UVA photosensitivity.

How the UVA Index handles cloud

Cloud is the single most variable input to this site's UVA Index, and the model treats it in two ways depending on what data is available for your location.

Preferred: derived from live UV data. Open-Meteo publishes both the actual UV Index and a clear-sky UV Index for the same hour. Their ratio is a measured erythemal cloud transmission for the real sky overhead — far better than any parameterisation of a cover percentage. Because UVA penetrates cloud better than erythemal UV, that transmission is then lifted for the UVA band:

The exponent is below 1, so any attenuation is softened, and the gap widens as cloud thickens — which matches the observed spectral behaviour. A heavy overcast that cuts erythemal UV to 30% leaves roughly 38% of the UVA; light cloud cutting erythemal UV to 70% leaves about 75%.

Fallback: parametric, from cloud cover. When the live UV pair is unavailable, the model falls back to the reported total cloud cover:

The curve is deliberately gentle, because a reported "100% cover" is a statement about sky coverage, not optical depth — so the right average behaviour at full cover is a moderate reduction to about 60%, not a blackout. The full derivation, along with every other coefficient, is in how the UVA Index is calculated.

One honest limitation: the model does not reproduce cloud enhancement. Both routes are attenuation-only, so a factor above 1 never occurs. Under broken cumulus with the sun in a gap, the real UVA can exceed what the UVA Index reports. It is listed among the factors deliberately left out, and it is the one case where the number is more likely to under-read than over-read.