Stats
How often can you actually see it?
Plenty of sites will tell you the aurora is “possible” from your city. Far fewer tell you how often.
We estimate the number of clear aurora nights per year for a place by combining three long-run averages: how many nights get properly dark, how often the Kp index climbs high enough to reach that latitude, and how often the sky is actually clear.
It's a climatological estimate — not measured history — but it makes the gap between a true aurora base and a hopeful mid-latitude city impossible to miss.
~100+
clear aurora nights a year at the best bases
The strongest aurora bases sit under the oval, keep long dark seasons and enjoy dry, clear skies — so a good display is a most-weeks event. From mid-latitudes, where you need a rare geomagnetic storm to push the oval overhead, it's only a handful of nights a year.
The best aurora bases, ranked
Sorted by estimated clear aurora nights per year. Follow any place for its live forecast and the Kp it needs.
Nights the oval reached each place since 2026-08-11 — geomagnetic activity only, before cloud and moon.
- YellowknifeCanada
148 nights
- FairbanksUSA (Alaska)
125 nights
- IlulissatGreenland
102 nights
- ReykjavíkIceland
88 nights
- RovaniemiFinland
68 nights
- TromsøNorway
58 nights
- AbiskoSweden
41 nights
- TasmaniaAustralia
5 nights
- Stewart IslandNew Zealand
3 nights
- DunedinNew Zealand
2 nights
- InvercargillNew Zealand
1 nights
| Place | ~Clear aurora nights/yr | Kp needed | Best season | Not overcast % | Oval reached |
|---|---|---|---|---|---|
| YellowknifeCanada | 148 nights | Kp 1+ | Nov – Apr | 49% | 42/42 nights |
| FairbanksUSA (Alaska) | 125 nights | Kp 1+ | Aug – Apr | 45% | 42/42 nights |
| IlulissatGreenland | 102 nights | Kp 1+ | Sep – Apr | 41% | 42/42 nights |
| ReykjavíkIceland | 88 nights | Kp 1+ | Sep – Apr | 31% | 42/42 nights |
| RovaniemiFinland | 68 nights | Kp 2+ | Sep – Mar | 31% | 38/42 nights |
| TromsøNorway | 58 nights | Kp 1+ | Sep – Mar | 23% | 42/42 nights |
| AbiskoSweden | 41 nights | Kp 1+ | Dec – Mar | 16% | 42/42 nights |
| TasmaniaAustralia | 5 nights | Kp 6+ | Mar – Sep | 60% | 0/42 nights |
| Stewart IslandNew Zealand | 3 nights | Kp 6+ | Mar – Sep | 38% | 0/20 nights |
| DunedinNew Zealand | 2 nights | Kp 7+ | Mar – Sep | 54% | 0/20 nights |
| InvercargillNew Zealand | 1 nights | Kp 7+ | Mar – Sep | 42% | 0/20 nights |
…and how fast it collapses further south
Drop off the auroral oval and the same maths turns brutal. These mid-latitude cities only see the lights when a strong storm drags the oval unusually far south — so the estimate falls from three figures to almost nothing.
Edinburgh
Scotland
4 nights/yr
needs Kp 6+
London
England
1 nights/yr
needs Kp 8+
Berlin
Germany
<1 nights/yr
needs Kp 9+
Mid-latitude figures assume an average clear-sky fraction. The point isn't the exact digit — it's the order-of-magnitude gap. A storm big enough to light up Berlin is a once-or-twice-a-year headline event, not a travel plan.
How we estimate this
Every number on this page is the product of three independent factors. We keep the method simple and out in the open:
- Dark nights — nights with at least an hour of sky dark enough for the aurora (sun 6° or more below the horizon), computed month by month from the sun's geometry at each place. Far-north sites lose weeks to the midnight sun, so this count actually falls toward the poles.
- Kp probability — the chance a given night reaches the Kp the oval needs to sit overhead at this location's geomagnetic latitude (not its map latitude), by calendar month, from GFZ Potsdam's definitive Kp index for 2015–2024. High Kp is exponentially rarer, which is why mid-latitude aurora is so scarce.
- Clear-sky odds — the share of days that were not overcast (mean cloud cover 70% or less) at these coordinates, month by month, from ERA5 reanalysis for 2015–2024. Dry continental interiors like Yellowknife and Fairbanks comfortably beat cloudy maritime coasts like Reykjavík and Tromsø.
This is an estimate, not a logbook. We treat the three factors as independent, and a day's mean cloud cover stands in for the night's. Real weather and space weather are correlated in ways this ignores.
Use these figures to compare places, not to promise a number for any single trip or year.
We're building per-night logging against our live verdict engine, so future versions can report measured hit-rates alongside this estimate. Until then, treat every figure here as an order-of-magnitude guide.
See tonight's odds for your spot
Stats tell you where. The live forecast tells you whether it's worth stepping outside right now — for your exact location, tonight.