Northern Lights forecast: how to read one

How far ahead the aurora can be forecast

The Finnish Meteorological Institute (FMI) is blunt about it: a single display is a random event, and only the amount of activity and how visible it is likely to be can be estimated. What can be estimated depends on how far ahead you look.

• Years: the Sun’s activity rises and falls over a cycle of about 11 years.
• Weeks: the Sun turns about once every 27 days as seen from the Earth, and a large hole in its corona can come round again in much the same shape.
• One to three days: images of the Sun show material heading for the Earth. FMI puts the error in its arrival time at about half a day either way.
• About an hour: satellites at a point called L1, upstream of the Earth, measure the solar wind as it passes. NOAA’s short-range aurora forecast is built on those readings and looks 30 to 90 minutes ahead.
• Now: FMI’s magnetometers across Finland show the disturbance as it happens. FMI calls this monitoring, and does not call it a forecast.

None of those horizons reaches the week you are booking. So choose dates by what can be known months ahead: the dark season, which runs from late August to early April, and the number of evenings you keep free. Our Northern Lights guide explains why four nights is the least we plan, and when to go sets out the months.

What the Kp index measures, and what it leaves out

The Kp index comes from NOAA’s Space Weather Prediction Center, which defines it as an index from 0 to 9 of how much the Earth’s magnetic field is disturbed over a three-hour interval, averaged from 13 observatories between 44 and 60 degrees geomagnetic latitude. A reading of 5 or more is a geomagnetic storm, which NOAA also reports on its G scale, from G1 at Kp 5 to G5 at Kp 9.

NOAA’s viewing tips explain what the number does to the aurora. As Kp rises, the auroral oval moves further from the pole and the aurora gets brighter and more active. At Kp 0 the equatorward edge of the oval sits at roughly 66 degrees of magnetic latitude, and it moves towards the equator about 2 degrees for each step up.

Every line of that describes how far towards the equator the aurora reaches, which is why a low number matters less in Lapland. FMI’s own aurora page says auroras are common in Lapland even when space weather is calm, and its probability words apply to southern and central Finland. FMI’s guide to reading the forecast puts the line at roughly Rovaniemi: its estimate covers the area south of the usual aurora zone. FMI’s English page goes further: in Lapland, it says, auroras are quite common always, even at solar minimum. A Kp of 2 on a clear night in Inari or Kilpisjärvi is a night to be outside.

FMI’s forecast for Finland, section by section

FMI publishes its aurora service in Finnish, Revontulet ja avaruussää, and an English page with the live readings, auroras and space weather. The Finnish service has two parts.

The first is a short written outlook, in Finnish, with a probability for each of the next three nights on a four-step scale: pieni, pienehkö, kohtalainen, suuri (small, fairly small, moderate, large). Each step is read against how common auroras are on average at the place. The text is updated once each weekday, and more often when space weather is especially interesting. FMI says two things about it that change how you read it. It does not take cloud into account. And, as above, it is written for the country south of Rovaniemi. On the morning of 25 September 2026 it gave pienehkö for all three nights ahead, and added in the same breath that auroras in Lapland are common even in calm space weather.

The second part is the live readings from FMI’s magnetometer stations: a bar chart of the last 24 hours and a map with the latest value at each station. The value is the R-index, how fast the magnetic field is changing. Each station has two thresholds.

FMI stresses that the link between the readings and the aurora is statistical. A single bar over the line can pass with nothing to see, and several in a row are a better sign. When the station nearest you turns red, FMI’s advice is that it is time to go out and look up. Displays are most likely around midnight and for a couple of hours either side, though in a big storm they can come at any time of the dark.

FMI also runs a separate space weather site, ISES RWC Finland, with an activity forecast for the next hour, a probability map and an all-sky camera in Muonio, in northern Lapland. The camera image is not updated in daylight.

Cloud is the part of the forecast to check first

The aurora glows far above the weather, so a sky full of cloud hides the strongest display, and FMI leaves cloud out of its aurora probabilities. You have to check it separately. FMI points to two places. Its satellite cloud map animates EUMETSAT’s Meteosat images over the Nordic countries, with high, cold cloud shown white and low cloud in grey. Its local forecasts, such as the one for Rovaniemi, show with their weather symbols whether a clear or a cloudy sky is expected.

FMI’s free weather app joins the two. For places in Finland, FMI says, the app’s forecast for the next few hours also says whether auroras can be seen. We describe it the way FMI does and have not tested it.

NOAA’s 30-minute aurora forecast draws the oval over each pole, green where it is forecast and red where it will be more intense, and does not show cloud. Its aurora dashboard adds a view line for tonight and tomorrow night, and NOAA labels it experimental.

One evening in Lapland, checked in order

The same checks in words. After dark, look at the cloud first, on the satellite map and the local forecast. If a gap is coming, open FMI’s station map and look at the station nearest you: yellow means faint auroras are possible, and red means go out now. NOAA’s 30-minute forecast shows whether the oval is over you. Then go to open ground away from lights, with a clear view to the north, which is the spot FMI recommends.

On a guided night the guide does this for you. One Rovaniemi operator, for example, says its guides watch the forecasts, the cloud cover and the aurora activity and drive to a dark spot. Our page on Northern Lights tours covers what they include, and how to photograph the Northern Lights covers the camera.

Four ways people misread an aurora forecast

Waiting for a storm. On NOAA’s scale a storm starts at Kp 5, and storms are what push the oval towards the equator. In Lapland you are already at the usual aurora zone, and FMI’s point about calm nights stands.

Reading the three-day Kp as a timetable. NOAA’s text forecast gives the expected Kp for each three-hour block of the next three days, in UTC, so convert to Finnish time before you read it against your evening. The one issued at 00:30 UTC on 25 September 2026 expected a highest Kp of 3.67 over 25 to 27 September, below storm level. A number like that is an estimate of activity. It says nothing about cloud over your cabin.

Planning around the 27-day return. A coronal hole can come back after one turn of the Sun, and FMI uses that for its outlook weeks ahead. FMI says a large hole can keep much the same shape for several turns, which is a long way short of a promise, so treat it as a hint for people who live here and not as a reason to move a flight.

Treating an app’s percentage as the chance of seeing it. When an app shows a probability or a Kp, find out which service the number comes from, where it applies and whether cloud is in it. FMI’s yellow and red readings are measured in Finland; a Kp number is a planetary average.

For Tromsø, which has its own forecast from a Norwegian research centre, see Tromsø and the Northern Lights.