Dew Point: What It Is, and How the Calculator Gets It

The formula, and which coefficients these are

Magnus–Tetens is not one equation but a family. The shape is always the same — saturation vapour pressure as an exponential of a·T/(b + T) — and what changes between versions is the pair of constants, which are fitted, not derived. This tool uses a = 17.625 and b = 243.04 °C, the set Oleg Alduchov and Robert Eskridge published in the Journal of Applied Meteorology in 1996 after refitting the classical Magnus form against the accepted reference formulations. Over a working meteorological range it stays within a few tenths of a per cent.

The other set you will meet is 17.27 and 237.7, from Tetens by way of a 1980 paper, still embedded in an enormous amount of code. It is not wrong; it is older, and it drifts further from the reference at cold and hot extremes. If two dew point calculators disagree by a couple of tenths at the same inputs, this is usually why, and neither of them is broken. What would be a real disagreement is a page that computes over ice rather than over water below freezing — a distinction this tool does not make, and that matters for frost forecasting rather than for comfort.

Why it beats relative humidity

Relative humidity is a ratio, and a ratio whose denominator moves is a poor thermometer of anything. Warm air can hold much more water than cold air, so the same percentage means completely different amounts at different temperatures. Run it through the calculator and the point is unarguable: 50 % humidity at 50 °F gives a dew point of 32 °F and reads as dry; 50 % at 86 °F gives 65 °F and reads as uncomfortable; 50 % at 104 °F gives 82 °F and is barely survivable. One number on the display, three different days.

The reverse test is just as clean. Three quite different days — 68 °F at 83 %, 86 °F at 45 %, 95 °F at 35 % — all land on a dew point near 63 °F, and all three feel like the same slightly clammy air, because they contain the same water. That is what an absolute measure buys you. The catch is that dew point tells you about the moisture and nothing about the heat: 95 °F with a 63 °F dew point is dangerous for a different reason, and that is what a heat index is for.

The tool sorts its answer into five bands, and it is worth knowing that they are cut on the Celsius figure whatever units you are working in: dry below 50 °F, comfortable to 60.8 °F, a bit sticky to 64.4 °F, uncomfortable to 69.8 °F, oppressive above that. Those odd-looking Fahrenheit boundaries are conversions of round Celsius ones — 10, 16, 18 and 21 °C — and they are close to, but not the same as, the 55 °F and 65 °F breakpoints the National Weather Service uses in its own summer guidance. Read the band as a label, not as a threshold with legal force.

Condensation: the practical use

Here is the reading that earns the calculator its place on a phone. Any surface colder than the dew point of the air touching it will collect water. Not might — will, as long as air keeps arriving. That single sentence explains the window, the cold pipe, the cistern that drips in August, the camera lens that fogs the moment you carry it out of an air-conditioned car, and the beer glass that leaves a ring on the table.

Put numbers on it and it becomes a tool. A room at 70 °F and 40 % humidity has a dew point of 44.6 °F: nothing in the room is that cold, so nothing sweats. Raise the humidity to 60 % and the dew point climbs to 55.5 °F — now the single-glazed window on a winter night is well under it, and it runs. That is why the answer to condensation on the inside of a window is almost never the window: it is the moisture load, and the fix is a bathroom fan, a lid on the pan, and airing the room, not a squeegee.

The same arithmetic settles the lens. Carry a cold camera into a room at 78 °F and 65 % and the dew point is 65.3 °F, so anything at 40 °F is more than 25 degrees under it and will fog instantly — inside the barrel as well as outside, which is why the fix is to leave it sealed in its bag until it has warmed up, rather than wiping the front element. And the direction never reverses: the dew point of a body of air is a property of that air, and cooling a surface can only bring it below the line, never lift the line.

Two things the calculator checks, and one it does not

Humidity is guarded at both ends. Zero or less is refused, because ln(0) has no value and air with literally no water in it does not have a dew point; anything above 100 % is refused too, and that second guard is what makes the whole thing consistent. Feed it 100 % and the algebra returns the air temperature itself — the tool prints 25.0 °C for 25 °C air — so with supersaturation excluded, the dew point can never come out above the temperature you typed. If you ever see a dew point above the air temperature on any page, that page has a bug.

Temperature is not guarded at all, and that is the one thing to know before trusting an extreme value. The Alduchov–Eskridge fit is stated for roughly −40 °C to +50 °C, but the tool applies it to whatever number you type. Enter 500 °C and it reports a dew point of 420.2 °C without comment. Enter −273.15 °C — absolute zero — and it returns −273.3 °C, a temperature that does not exist. Nobody will do this by accident on a weather question, but it is worth knowing that the range check exists on one input and not the other, and that the answer degrades quietly rather than refusing.

One inconsistency inside the tool itself is worth flagging, because it will confuse anyone who reads both parts of the screen. The tip under the inputs says that above a dew point of 18 °C the air feels oppressive. The tool's own classifier disagrees: it calls 18 to 21 °C uncomfortable and reserves oppressive for 21 °C and above. Trust the band on the result card, which is what the arithmetic actually produced, over the sentence in the hint, which is a rounder statement written separately.

Frequently asked questions

Which Magnus coefficients does this calculator use?

a = 17.625 and b = 243.04 °C — the Alduchov–Eskridge set, published in the Journal of Applied Meteorology in 1996 as a refit of the classical Magnus form. It computes α = ln(RH/100) + a·T/(b + T) and then Td = b·α/(a − α). The other common pair is 17.27 and 237.7, and a page running it will disagree with this one by a couple of tenths of a degree at normal temperatures and by more at the extremes. Neither is a bug; they are different fits of the same functional form.

Can the dew point ever be higher than the air temperature?

No. At 100 % relative humidity the two are equal — the calculator returns 25.0 °C for air at 25 °C — and below 100 % the dew point is strictly lower. Anything above 100 % would be supersaturated air, which this tool refuses as an input, so the case never arises. If a page shows you a dew point above the air temperature, it has either taken a humidity above 100 or made an arithmetic error, and everything else it says is suspect too.

What dew point makes a window run with water?

Whichever one is above the glass's own temperature. Compute the dew point of the room air and compare: a room at 70 °F and 40 % has a dew point of 44.6 °F, so a window at 50 °F stays dry; push the humidity to 60 % and the dew point is 55.5 °F, and the same window now runs. This is why the practical fix is almost never the glass. Lower the moisture — extractor fan, lid on the pan, dry the laundry elsewhere, open a window for five minutes — and the dew point falls below the coldest surface in the room.

Why do forecasters quote dew point instead of humidity?

Because dew point is absolute and relative humidity is not. A dew point of 63 °F is the same amount of water in the air whether the thermometer reads 68 °F or 95 °F, so it can be compared across the day and across the country; 50 % humidity means one thing at breakfast and another at three in the afternoon in the same place, because the air warmed up and the denominator moved. Humidity also swings wildly overnight with no water having gone anywhere, which is why an evening reading of 90 % rarely means the muggy day you might expect.

Is there any input the calculator does not check?

Yes: temperature. Humidity is bounded on both sides and anything at or below zero, or above 100, is refused outright. Temperature is not bounded at all, so the tool will happily apply a fit stated for roughly −40 °C to +50 °C to any number you give it — 500 °C returns a dew point of 420.2 °C, and absolute zero returns something below absolute zero. In ordinary weather this never matters. It matters if you are pasting values from an industrial sensor or a furnace log, where the answer will be quietly wrong rather than refused.