Northern Lights
Bright aurora above a snow covered valley under a clear night sky

Abisko

The strongest reputation in aurora travel, and the numbers that do not confirm it

Abisko has the strongest reputation in aurora travel, and it rests on the blue hole: the Scandinavian mountains stand between this valley and the Atlantic, strip the moisture out of incoming weather, and leave a window of clear sky over Tornträsk when everything around it is shut. The geography is real and documented. The percentages attached to it, 70% clear nights, 80 to 90% success, come from the people selling tours here, and are published without a metric or a period.

So we counted. Ten seasons of Open-Meteo archive (original model unverified), nights with at least two hourly samples, not necessarily consecutive, with cloud cover at or below 50% between 18:00 and 02:00, the same method applied to every destination on this site. Abisko comes out at 18.7% across the season, the lowest of the fourteen places we measure. Kiruna, 90 kilometres east and the town both of these tours actually leave from, comes out at 38.8%.

The original weather-model selection is unverified, so a 28 km grid cannot be attributed to these results. The published values describe the original fixed-location calculation. They do not establish the presence or absence of a local clear-sky window or ice fog, a regional causal explanation, or how a moving tour performs.

The honest position: we cannot tell you the blue hole is a myth, and nobody can tell you it is a fact. Settling it would take station observations, which SMHI publishes in Sweden and which nobody in this business has bothered to cite.

Ratings collected by Viator and Tripadvisor

The numbers, month by month

Darkness and lunar conditions are astronomical calculations. Our implementation uses approximate algorithms and samples positions every twenty minutes; the resulting durations are planning estimates, not exact continuous observations. The standard-time window does not always match local civil time during daylight saving. Previously published Open-Meteo cloud figures, ten September–March seasons (2015–2025). Original model selection unverified; values not recalculated in this correction.

Historical Open-Meteo cloud data, ten seasons; original model unverified, not recalculated. Cloud × darkness is an index, not expected viewing hours. Astronomy values are estimates.
Month Usable-sky nights Darkness per night Moon-free nights Cloud–darkness index
October19.4%6.5 h48%1.26
February18.7%6.7 h54%1.25
January14.2%7.9 h58%1.12
December13.5%8.0 h58%1.08
March25.5%4.0 h45%1.02
November12.3%7.9 h50%0.97
September27.3%2.3 h23%0.63

The thresholds, since they are choices rather than truths. Usable sky means at least two hourly samples, not necessarily consecutive, with cloud cover no greater than 50%, inside the 18:00 to 02:00 window when tours actually operate. Astronomical darkness means the Sun below 18 degrees, which is a standard definition and not ours. Nights without moon interference account for both the illuminated fraction and the Moon's height above the horizon.

The last column is where Abisko's real problem shows. Every month here sits between 0.63 and 1.26 cloud–darkness index. Kiruna, 98 km east, runs between 1.10 and 2.86: its worst month beats Abisko's best. There is no month that rescues this valley on the measurements we can make.

The two months with the best cloud figures, September at 27.3% and March at 25.5%, are fifth and seventh once darkness is counted, because there are 2.3 and 4.0 hours to use them in. October and February are the months, which is not what anyone recommends.

The cloud–darkness product is a comparison index, not measured or expected hours of aurora viewing. It multiplies a monthly share of qualifying nights by average darkness hours without modelling their joint timing. Adding monthly geomagnetic activity does not turn it into a sighting probability.

A person standing on snow looking up at green aurora in the night sky
When the sky here opens, it delivers everything the reputation promises. The question nobody asks is how many nights in ten that happens. · Freepik

What this means in practice

Not that Abisko is a bad destination. Three concrete things.

First, the valley is not the shortcut its reputation suggests. If you are choosing where to spend three nights on the strength of the sky, our numbers put Kiruna ahead in all seven months, by between 12 and 24 points of cloud and roughly double once darkness is counted. It is also the base both of these excursions depart from.

Second, October and February are the months to prefer, not the equinoxes and not midwinter. October pairs the second-best cloud figure with six and a half hours of dark; February does much the same. November, which sounds like a sensible aurora month, is the worst on our count at 12.3% and 0.97 cloud–darkness index.

Third, the reputation has a price effect. The two departures here cost 164 and 252 EUR against 153 EUR for the cheapest in Kiruna. You are paying a premium for a promise no public data verifies.

The two departures

The 2 Abisko departures in the catalogue. Lowest published adult rate, read on 10 August 2026.
Departure Price Leaves from What it covers
Guided tour from Kiruna164 EURKiruna, 98 km eastThe round trip and guided time inside the national park
Guided tour with a three-course dinner252 EURKiruna, 98 km eastThe same drive, Sámi guides, dinner, and the country as well as the sky

Both leave from Kiruna because that is where the airport, the beds and the operators are. With one night, buying the transfer makes sense: 98 km of Swedish winter road in the dark is not a job for a hire car and a phone map. With three nights and the aurora as the point of the trip, the Kiruna to Narvik railway stops in the village and the valley is dark the moment you step outside, where the Swedish right of public access does the work for you, free.

Why this page is written this way

A better headline would be "Abisko is the worst", and it would be a worse statement. Our count measures one specific thing at one specific scale, and on a 10 to 20 square kilometre window it declares its own blindness. Publishing the number without that limit would be the same error as the 70%, turned around.

What we can say with a straight face is this: the figure in circulation is not verifiable, ours is but does not answer the same question, and in the meantime the valley next door has better numbers at a lower price. Nobody selling those tours will ever publish any of those three sentences.

The Abisko season, on our numbers

Month 1 2 3 4 5 6 7 8 9 10 11 12
Abisko Every month falls between 0.63 and 1.26 cloud–darkness index, and Kiruna's worst month beats Abisko's best. October and February are the least weak; November at 12.3% is the lowest cloud figure of all fourteen destinations. 1 2 3 4 5 6 7 8 9 10 11 12
Kiruna, 98 km east Six of its seven months are within half an index point of each other, and all six beat anything Abisko offers. Cheaper, with six departures against two. 1 2 3 4 5 6 7 8 9 10 11 12
Peak season Flying Out of season

Open-Meteo, ten seasons; model unverified, no recalculation. ≥2 hourly samples ≤50% cloud, 18–02, not necessarily consecutive. Cloud×darkness: index, not viewing hours or sighting probability.

Abisko, without the sales pitch

Does the blue hole exist?

The original weather-model selection is unverified, so a 28 km grid cannot be attributed to these results. The published values describe the original fixed-location calculation. They do not establish the presence or absence of a local clear-sky window or ice fog, a regional causal explanation, or how a moving tour performs.

When should I come?

October or February, which is not the answer anyone gives. Both pair the least-bad cloud figures with six and a half hours of real darkness. September and March have the best cloud records here and the least dark to use them in, and November, which sounds like a proper aurora month, is the weakest of all at 12.3%.

Then where does the 70% come from?

From people selling tours in Abisko, without stating how it is measured or over what period. That does not make it false, it makes it uncheckable, which in this trade is nearly the same thing.

Abisko or Kiruna?

On our numbers, Kiruna, in all seven months and at a lower price, and by roughly double once darkness is counted. With the caveat above: if the blue hole exists at the scale claimed, our measurement does not contradict it, it simply cannot see it. Kiruna also has six departures against two.

Is it worth sleeping in Abisko rather than Kiruna?

With three nights or more, yes: the Kiruna to Narvik train stops in the village, the valley is dark the moment you leave the building, and the Swedish right of public access lets you stand in a field at two in the morning without asking anyone. With a single night, buy the guided transfer.

Data verified on . Cloud data: Open-Meteo, ten seasons; original model unverified, not recalculated.