Wind Chill: What the “Feels Like” Number Actually Measures

Which formula, and how to recognise it

There are two wind chill indices in circulation and they do not agree. The 1945 one came out of an Antarctic experiment in which Paul Siple and Charles Passel timed how long it took water in a plastic cylinder to freeze on a roof, and reported a rate of heat loss that was later rewritten as an equivalent temperature. The 2001 one, worked out jointly by the American and Canadian weather services with a human-tissue model and volunteers in a wind tunnel, models heat loss from a bare face at head height rather than water in a bottle at roof height. It is the one every official chart now uses, and it is markedly less alarming than its predecessor at the same conditions.

You can tell which one a page is running without reading its code. The 2001 index has the wind raised to the power 0.16 and four coefficients: 13.12, 0.6215, 11.37 and 0.3965 in Celsius and km/h, or 35.74, 0.6215, 35.75 and 0.4275 in Fahrenheit and mph. This tool has them. It stores the metric form and nothing else: when you switch it to US units it converts your °F and mph into °C and km/h, runs the same four coefficients, and converts the answer back — which is why its output tracks the published Fahrenheit equation to about five hundredths of a degree rather than matching it digit for digit.

That exponent of 0.16 is the whole shape of the thing. Wind enters the formula under a fractional power, so its effect saturates fast: at 20 °F, going from 5 mph to 15 mph costs you 6.7 degrees, while going from 25 mph to 35 mph costs 2.5. The first breath of wind does most of the damage, because what it removes is the thin warm layer clinging to your skin, and once that is gone there is nothing left to take.

The two bounds, and what happens outside them

The National Weather Service is explicit: the index is defined only for temperatures at or below 50 °F and wind speeds above 3 mph. That is a narrow box, and it is where almost every wind chill widget on the internet goes wrong — it applies the regression to whatever it is handed and prints the result. This one does not. Above 50 °F it says so and stops; below 3 mph it says so, and then shows you the number it would have printed, which is the most persuasive argument available. At 23 °F with no wind at all, the formula returns 50 °F.

One correction is owed to the tool's own note, which credits Environment Canada for the range. The formula is Canadian and American jointly, but those two bounds are the American statement. Environment and Climate Change Canada's climate glossary is narrower still: it applies this equation only when the air is at or below 0 °C and the wind is 5 km/h or more, and for winds above zero but under 5 km/h it uses a second, entirely different equation — W = T + [(−1.59 + 0.1345·T)/5]·V, a straight line rather than a power law. This calculator has neither the 0 °C ceiling nor the light-wind equation. Between 0 °C and 10 °C it is answering by the American rule, which is a defensible choice, but not the one its footnote describes.

It models bare skin, not you

The 2001 index is computed for a specific, stated subject: an exposed adult face, in shade, on a person walking at about 3 mph, with the wind adjusted from the ten-metre anemometer height down to head height. Every one of those conditions is a decision that was made for you. In sunlight the same air feels several degrees warmer. In a coat and a balaclava almost nothing of what the index describes is exposed at all. On wet skin — sweat, sleet, a splash — heat leaves far faster than the dry-skin model allows, which is exactly the situation the number will flatter.

That is why the useful output is not the equivalent temperature but the hazard band underneath it, which answers a question the equivalent cannot: how long can exposed skin stay out. The tool ships Environment Canada's published table — no risk down to 0, low risk to −9, uncomfortable to −27, then 10 to 30 minutes to freeze to −39, 5 to 10 minutes to −47, 2 to 5 minutes to −54, and under 2 minutes below that. Those breakpoints are whole degrees Celsius and the tool rounds the index before classifying it, so a −27.4 lands in the same row as −27. Switch to US units and the numbers on screen turn into Fahrenheit but the classification is still made on the Celsius index, so the band boundaries fall on 15.8 °F, −16.6 °F, −38.2 °F, −52.6 °F and −65.2 °F — conversions of Canadian breakpoints, not the round numbers on an American chart.

It says nothing about an object

This is the most common misreading and it is worth being blunt about. A parked car does not cool below the air temperature because the wind is blowing. Neither does a water pipe, a bottle on a balcony, or a bag of shopping in the boot. Wind chill measures how fast a body that is generating heat loses it, and an object generates none. What wind does to an object is get it to the air temperature sooner — an engine that would have taken two hours to cool takes one — and then stop. There is no mechanism by which moving air at 23 °F can take anything below 23 °F.

There is a real effect that does put an object below air temperature, and it is worth separating from this one because they pull in opposite directions. On a clear night a surface radiates heat to a cold sky and can end up a few degrees below the air — which is why frost forms on a windscreen when the forecast said two degrees above freezing. Wind works against that, not with it: it mixes warmer air down onto the surface and keeps it near the air temperature. Frost is a still-night phenomenon. If you want a rule of thumb that survives contact with a driveway: wind chill for skin, clear calm sky for glass.

Reading the number in practice

Read the drop, not the equivalent. The calculator prints how many degrees below the thermometer you are as a separate figure, and that is the one that tells you what to change: eight degrees of drop is a hat, twenty is a hood and a scarf, and the band underneath tells you whether it is minutes or half an hour before bare skin is in trouble. The equivalent temperature is a number to compare against other equivalent temperatures, and comparing it against your indoor thermostat means nothing at all.

Then be honest about the wind you typed. The formula wants the wind at ten metres, which is what an airport anemometer reports and what a forecast quotes — and it is not the wind between two buildings, in the lee of a hedge, or on an exposed ridge. It is also not your own airspeed: cycling at 20 km/h into a 20 km/h headwind is a 40 km/h problem for your face, and the tool has no idea you are moving. If you want the honest figure for a ride, add your speed to the headwind before you type it.

Frequently asked questions

Which wind chill formula does this calculator use?

The 2001 one, the joint American–Canadian revision, in its metric form: 13.12 + 0.6215·T − 11.37·V^0.16 + 0.3965·T·V^0.16. It does not use the 1945 Siple–Passel index, which came from timing water freezing in a plastic cylinder rather than modelling a face, and which is harsher at the same conditions. If you switch the tool to °F and mph it still runs the metric equation, converting in and out, so its answers sit within about five hundredths of a degree of the Fahrenheit form printed on official charts, 35.74 + 0.6215·T − 35.75·V^0.16 + 0.4275·T·V^0.16.

Why does it refuse to answer at 60 °F, or with no wind?

Because the index is not defined there, and printing a number anyway is how these pages mislead people. The National Weather Service states the range plainly: at or below 50 °F, and above 3 mph. Outside it the regression is being extrapolated past the data it was fitted to. The tool demonstrates why rather than asserting it: at 23 °F with no wind, the V^0.16 term becomes zero and the formula returns 50 °F. Anything that tells you a calm winter day feels warmer than the thermometer has stopped modelling the world.

Does wind chill make my car, my pipes or my plants colder?

No. Wind gets them to the air temperature faster and then stops. A pipe in a 23 °F draught will freeze sooner than a pipe in still 23 °F air, but neither ends up below 23 °F, and the wind chill number tells you nothing about how soon. Living plants are a partial exception in that they hold moisture and lose it to dry moving air, but that is desiccation, not wind chill. The one thing that genuinely puts an object below air temperature is radiative loss to a clear night sky, and wind reduces it by mixing warmer air down.

Is the “feels like” what I will feel wearing a coat?

No, and this is the single biggest thing to know about the index. It is computed for exposed skin — a bare face, in shade, on someone walking at about 3 mph — and what it predicts is how fast that skin freezes, not how a clothed body experiences the day. Under insulation the wind mostly cannot reach you, so a well-fitted jacket removes most of the effect; a jacket that lets air through removes very little. Sun adds several degrees back. Wet skin loses heat far faster than the index says, which is why the number is least trustworthy in sleet.

Why does doubling the wind barely change the answer?

Because the wind enters as V raised to the power 0.16, not as V. A fractional exponent flattens fast: at 20 °F, 5 mph gives 12.9 °F, 15 mph gives 6.2 °F, 25 mph gives 2.6 °F and 35 mph gives 0.1 °F — the first ten miles an hour are worth almost three times the last ten. Physically this is the boundary layer: the wind's job is to strip the thin sheet of warm air held against your skin, and once that sheet is gone, more wind has nothing left to remove.