About the service

Rengasennuste turns the weather and road surface forecast into an answer for each municipality: does the road stay above freezing, are winter conditions coming during the forecast period, or have they already started. The map covers all 309 municipalities in Finland, and each one gets its own colour.

This page describes the method behind the colour, what the estimate rests on and where its limits are. Your own area is on the map. The law's requirements (when tyres are mandatory, the studded tyre window, tread depth and trailers) are on winter tyre rules.

The colour tells you when the tyres have to change. Behind it is always the same thing: what the road surface does. The calculation is the same in autumn and in spring, and the season changes only the answer (in autumn the same window is a deadline, in spring a wait).

Colour and road surface	In autumn	In spring
Stays above freezing for all 27 days	Over 27 days to swapSummer tyres are enough	You can swap nowSummer tyres are enough
May freeze within 27 days (long outlook)	Under 27 days to swapThe turn is visible, but there is time	Wait a while yetWinter can still return
May freeze within 10 days	Under 10 days to swapWinter conditions approaching	Not yetWinter conditions may return
Freezes within 3 days, or snow expected	Under 3 days to swapWinter has arrived	Winter tyres still neededWinter is not over
No forecast available	The municipality stays grey on the map. A missing forecast is never assumed to be green.

The near term decides red: it takes two freezing days, or one measured freeze (that one has already happened, so one is enough). The whole period (roughly the notice a tyre shop appointment needs) decides amber: one freezing day is enough, when the model puts it at 60 % or more. A single cold night far from the rest of the period does not colour a municipality on its own.

The long outlook decides yellow, and it is a different kind of question from the other two. Three weeks out there is no road surface to compute and no single night to trust, so the rule reads two slow quantities instead: the three-day rolling mean of the daily mean temperature drops to +7 °C or below, and the period holds at least two expected frost nights. Neither can be moved far by one day. The windows are nested: a municipality always takes the shortest colour that can be claimed about it.

Red and amber are based on the temperature of the road surface, not of the air. On a clear, still night the road cools below the air and can freeze even while the air still reads above zero. The model estimates the road surface temperature from the weather for each day. The mechanism is set out under what the model is based on.

The current day combines the first two tiers: the part of the day that has passed is measured, the rest is forecast, and the card shows the colder of the two.

Each day is also given a probability of freezing: the more the weather models agree about the night's temperature, the narrower the uncertainty. The accuracy has been measured against a set of 554 observations (the figures are under what the model is based on).

The so-called "+7 degree rule" (summer tyre rubber begins to harden somewhere around +7 °C) is read as an average over several days, not from one cold morning. On top of that, the forecast must hold at least one night on which the road surface can drop below freezing.

The legality of studded tyres is its own signal and does not affect the colour described above. The window comes from the calendar, but outside it the conditions can still permit studs: when studded tyres are allowed.

Every card in the sidebar is explained below under the sidebar cards explained.

Once a municipality is selected, several cards appear in the sidebar. Here is what each of them shows.

All of Finland

Shown before a municipality is selected: the number of municipalities, the age of the forecast, a proportional green/amber/red/grey bar and the count for each colour.

The municipality's colour

The selected municipality's colour, a short "why", and a three-part scale showing where it sits. How it is computed is above under how the forecast works.

Change by / Summer tyres no earlier than

The date to act by, the days remaining and a short explanation. If the first freezing night is uncertain, the date is shown as an estimate rather than as a certainty.

Today

The nearest station's readings and the lowest road surface and air temperature for the whole current day (the elapsed part measured, the rest forecast, the colder shown). Weather icons mark ice, freezing rain or other slipperiness. See the three tiers.

10-day forecast

An hour matrix: a column per day, 24 cells down it, with snow and frost marks. Every colour is a probability of freezing — above freezing under 40 %, borderline 40–60 %, freezing at least 60 %, and a deepest step over 90 % — not a temperature. This is the card that drives red (near term) and amber (whole period). See two time windows.

Long outlook (three weeks)

A curve of the daily minimum air temperature, not the road surface: the near days as a precise forecast on a solid line, the further ones as MET Norway's long-range outlook on a dashed line, out to about 26 days. The line above it states the direction fitted to that curve — when it reaches freezing — rather than a forecast for any one night. This outlook is what turns a municipality yellow: when the three-week mean drops below +7 °C with night frosts expected, the municipality's deadline becomes 27 days. The date on the card is still a different thing from the colour — the colour is a window, the date an estimate from inside it, which is why it is marked “about”. The outlook speaks about air, because the road surface cannot be modelled that far ahead.

Studded tyres

Whether studded tyres are legal, exceptional or prohibited in the municipality: its own signal, separate from the colour. The window comes from the calendar, the exception from the road. See when studded tyres are allowed.

The phenomenon is ordinary road weather science: on a clear night the road radiates heat to the sky and cools below the air. The same phenomenon is computed as a proper energy balance by the Finnish Meteorological Institute's RoadSurf model (Kangas, Heikinheimo & Hippi 2015, open-sourced in 2024) and by the international METRo (Crevier & Delage 2001). Rengasennuste is not an energy balance model but a rough approximation of these: it knows nothing of the road's structural layers, solar radiation, traffic or salting, none of which a ten-day forecast covering a whole country can obtain.

The bars show how often the forecast held in a set of 554 measured road station observations: above the 60 % line a forecast of freezing was nearly always right, and below it only in a third of cases. That is why the threshold sits where it does. Over the whole set the mean error for the current day is 1.65 °C and the hit rate for freezing about 91 %. The observations come from the Nordic and Baltic road weather station network, mostly outside Finland's borders, and describe accuracy on a maintained main road. On a smaller or unmaintained road there can be ice even when the model is right.

More background: the model's corrections, coefficients and calibration

The cooling figure is held back by two corrections, most visible in summer. Night length: the base figure matches October to January nights, which hold well over ten hours of darkness. At Utsjoki on 12 August there is no darkness at all, and the road does not then cool as it would in November, so the cooling figure is reduced in proportion to how long the sun stays below the horizon. From October to March the night is long enough everywhere in Finland that the correction changes nothing. Ground heat: under the asphalt is a heat store that keeps the surface warm. At Näätämö +22.3 °C was measured under the road on 7 August 2026. Warm ground damps the cooling figure, based on the stations' own ground temperature at 4–7 cm sensor. Neither correction can cool the estimate or raise it above the air minimum: they can only remove cooling the model has no grounds for.

The model's coefficients are our own estimates, not taken from a publication. That a clear night drops the surface 3.0 °C below the air minimum and a rainy one 0.3 °C was chosen to be of the right order for the physics, not fitted to measurement data. The same goes for the wind and dew point limits, and for how warm the ground has to be before it starts damping the cooling. They are reasoned rules of thumb, meant to be refined as observations accumulate. Dry nights between these extremes are computed from the forecast's cloud cover, so that low cloud slows the cooling more than thin high cloud does. Night length is the exception: it is computed from latitude and date, and has no coefficient to guess.

Night length is taken into account in both directions, so in spring Lapland's short night damps the cooling figure just as it does in August: at Utsjoki on 15 April, by about a third. The spring estimate therefore moves slightly earlier, towards summer tyres. That is a deliberate choice, because an error in that direction shifts the timing of the change back rather than the change to winter tyres. Both ends will be reviewed once measurement data has accumulated over a winter.

The "+7 degree rule" is likewise an established rule of thumb in the tyre trade, not a threshold determined by research. Rubber stiffening in the cold is a real phenomenon, but the figure +7 °C in particular is an industry convention. That is why it is read here as a soft, multi-day average rather than as a precise threshold.

A ten-day forecast is not accurate enough to name the exact day of freezing, but it is accurate enough to time an appointment. A municipality's colour is computed from a single representative point, so within one municipality the coast and the interior can differ from each other.

The forecast updates twice a day from September to December and in March and April, once a day in January and February, and on Mondays from May to August. The top of the map states when the forecast on screen was computed, and the page says so itself if an update has failed to arrive.

The sources come in two layers. The air forecast comes from a meteorological service. The road surface figure is measured, modelled or estimated (how the forecast works), and in Finland road surface measurements come from Fintraffic alone.

The air forecast's figures and its worded descriptions come from the Finnish Meteorological Institute. MET Norway supplies the same forecast's uncertainty and the long-range outlook, covers the areas the Finnish Meteorological Institute's query does not reach, and acts as a fallback if that query fails.

Two services are not two different models. Both rest on ECMWF's forecast model, and the Nordic high-resolution model is shared between MET, SMHI and the Finnish Meteorological Institute. When two sources arrive at the same figure, that tells you only that two services agree. It does not mean two independent models have confirmed the same thing.

• Finnish Meteorological Institute (Ilmatieteen laitos): weather forecast, open data, licensed under CC BY 4.0
• MET Norway (Meteorologisk institutt, api.met.no): forecast uncertainty, the three-week outlook and the fallback source, open data, licensed under CC BY 4.0
• Fintraffic / digitraffic.fi: road weather station observations and the road condition forecast for road sections (road surface temperature), open data, licensed under CC BY 4.0
• Statistics Finland (Tilastokeskus): municipal boundaries (309 municipalities)
• Leaflet: map library

The service has no user accounts. The "Show my area" function resolves your area in your own browser. Location data is not stored and is not sent anywhere.

The site uses Google Analytics 4 through Cloudflare Zaraz. This lets us see which municipalities are viewed and whether people come back. Analytics starts only once you accept it in the cookie banner, never by default. You can change your choice at any time through the "Cookie settings" link in the footer.

The sidebar may show a tyre shop booking link, always marked with a "Commercial partnership" label. The link is kept separate from the weather forecast: the forecast does not know who has bought visibility in an area, and commercial interests cannot influence the colour.

Feedback, error reports and enquiries about advertising can be sent to contact@rengasennuste.fi. The same address travels with the data queries the service makes, so the maintainers of our data sources can reach us through it too.