Hygrometer
Organised by what is physically sensed
Instruments get marketed by product category, which is not very informative. Sorted by what actually does the sensing, there are five approaches and everything else is a variation.
1. Absorption — something takes up water and changes
The oldest idea. A material in equilibrium with the air absorbs water in proportion to the relative humidity, and something measurable about it changes.
- Mechanical: hair lengthens by roughly 2.5% between dry and saturated. Levers amplify that into a dial pointer. This is the antique brass instrument on the wall.
- Capacitive: a thin polymer film between two electrodes changes its dielectric constant as it absorbs water, changing the capacitance. This is essentially every modern sensor.
- Resistive: a salt or conductive polymer changes electrical resistance. Cheaper, more drift-prone, common in low-cost devices.
2. Evaporation — the psychrometer
Two thermometers, one with a wetted wick over the bulb. Evaporation cools the wet one, and how much it cools depends on how readily the air accepts water. The difference between the two readings gives the humidity.
Purely thermodynamic, needs no calibration of a sensing material, and stays the reference method for checking other instruments.
3. Condensation — the chilled mirror
Cool a polished mirror until an optical sensor detects dew forming, and read the mirror's temperature at that instant. That temperature is the dew point, by definition.
The most accurate practical instrument, around ±0.2°C, and the standard others are calibrated against. Expensive, and the mirror needs cleaning.
4. Spectroscopy — infrared absorption
Water vapour absorbs specific infrared wavelengths. Shine light through a sample, measure what arrives, infer the vapour concentration. Used in industrial process control and atmospheric research, rarely anywhere else.
5. Gravimetric — weighing the water
Pass a known volume of air through a desiccant and weigh what it gained. The primary standard, in the metrological sense: it is how the other methods are ultimately traceable. A laboratory procedure, not an instrument.
Which one gave you your number
| Where you saw it | Almost certainly | Typical accuracy |
|---|---|---|
| Home weather station | Capacitive | ±3–5% RH |
| Cheap digital display | Resistive or low-grade capacitive | ±5% or worse |
| Brass dial on the wall | Hair tension | ±5–10%, and drifting |
| Official weather observation | Capacitive in a radiation shield | ±2–3% |
| Weather app on a phone | None — a service's interpolation | Depends on station distance |
| Laboratory reference | Chilled mirror | ±0.2°C dew point |
The radiation shield matters as much as the sensor
An outdoor sensor in direct sun reads its own heated enclosure rather than the air. Official observations sit inside a radiation shield — the louvred white screen you see at weather stations — which blocks sun and rain while letting air circulate.
This is not a detail. A sensor in the sun can be several degrees warm, and since relative humidity depends on temperature, a few degrees of error in the temperature produces a large error in the humidity. A great deal of the difference between a hobby station and an official one is the shield and the siting, not the sensor.
And when nothing measured it at all
The reading in a weather app is not a measurement of your location. It is a weather service's estimate for your coordinates, built from stations that may be miles away plus model output.
That is a legitimate and useful thing — it is what this site's live pages and the app both do — but it is a different kind of number from an instrument in your hand, and worth understanding as such.