Does a full moon ruin the northern lights? What our camera data shows

Does a full moon ruin the northern lights?

No. A bright moon washes out faint aurora, but a moderate or strong display still shows clearly. The moon matters most when the aurora is weak, and least when it is strong. A full moon low on the horizon costs less than one high in the sky.

What the cameras saw under a bright moon

Our forecast is trained on all-sky cameras in northern Norway and Sweden, whose images are classified automatically. Under a bright moon, the classifier files a lot of aurora as "bright sky" rather than aurora. Counted naively, that makes moonlit nights look far worse than they are.

So we looked at moonlit camera frames ourselves and graded each one from 1 to 10 for how the aurora would look to a person standing there.

• About half of the moonlit aurora the cameras recorded was still worth seeing for a person (graded 5 or higher).
• The grades fell as the moon got brighter: under a dim moon 8 of 10 frames were worth seeing, under a bright moon 4 of 10.
• The faintest moonlit detections, below the cut we now use, were never worth seeing in our check.

How much does the moon change your odds?

The forecast now counts a moonlit aurora detection at about half the weight of one on a dark night, and it takes into account how bright the moon makes the sky, not only whether the moon is up. In practice:

• On an otherwise identical night, a full moon low in the sky (around 8°) gives a clearly better chance than a full moon high up (around 40°).
• A new moon week gives the darkest skies, so faint aurora counts too.

How should you plan around the moon?

1. Check the moon phase and the northern lights calendar. Each night shows how many dark hours have no moonlight.
2. If your trip falls around full moon, don't cancel. Look for the hours after moonset or before moonrise, and for nights with stronger activity.
3. Face away from the moon and keep it behind a building or a hill. Your eyes adapt better without it in view.
4. A phone camera on a long exposure picks up aurora the eye barely sees in moonlight.

Where the numbers come from

The camera classification is the UEC Tromsø AI dataset (Nanjo et al. 2022). The eye check covered 42 moonlit frames from two independent rounds in October 2026; with that sample, the "about half" figure has a wide margin (roughly 40–70%). Details are on How we know.