Home / Articles / Measurement Systems

Measurement Systems · 6 min read

Celsius, Fahrenheit and Kelvin: Which Scale to Use and Why

Three temperature scales, three different zero points. Understanding what each zero represents explains why temperature conversion is not simple multiplication.

Temperature is the one conversion that trips up otherwise careful calculations, because it is the one where the scales do not share a zero. Every other everyday conversion is a multiplication. Temperature is a multiplication and an offset, and forgetting the offset produces answers that are wrong in ways that look plausible.

Fahrenheit: the oldest of the three

Daniel Gabriel Fahrenheit proposed his scale in 1724. His reference points were a freezing brine solution at zero, and roughly human body temperature near 96 — a number chosen because it is divisible by two six times, convenient for marking a thermometer by repeated bisection.

Later refinement fixed water's freezing point at 32 °F and boiling at 212 °F, exactly 180 degrees apart. Body temperature drifted to 98.6 °F as a consequence, which is why that oddly specific number is really just 37 °C converted.

Fahrenheit's practical virtue is resolution in the range humans inhabit. Between a cold winter day and a hot summer one you have roughly 100 gradations without decimals. Its supporters argue that 0 to 100 °F maps more usefully onto human comfort than 0 to 100 °C does.

Celsius: built around water

Anders Celsius proposed a 100-degree scale in 1742 — originally inverted, with 0 at boiling and 100 at freezing. It was flipped shortly after his death.

Modern Celsius is no longer defined by water at all. Since the 2019 SI revision it is defined from the kelvin, with the relationship °C = K − 273.15. Water's freezing and boiling points are now measured consequences rather than definitions, and at standard pressure water freezes at 0.000 °C and boils at 99.974 °C — very slightly off the round numbers everyone learns.

Kelvin: absolute zero

Kelvin is the SI base unit and the only one of the three that is absolute: its zero is the actual physical floor, the point at which classical thermal motion ceases. Nothing can be colder, so kelvin temperatures are never negative.

A kelvin and a Celsius degree are the same size, so the conversion is pure offset: K = °C + 273.15.

Note the convention: it is "300 kelvin", not "300 degrees kelvin", and the symbol is K with no degree sign. Kelvin is a unit of thermodynamic temperature, not a position on a graduated scale.

Absolute scales matter because physics needs ratios. Doubling absolute temperature doubles average kinetic energy. This is meaningless on Celsius: 20 °C is not "twice as hot" as 10 °C, because zero is arbitrary. Gas laws, thermodynamics and blackbody radiation all require kelvin.

Why the conversion is not a simple factor

Converting Celsius to Fahrenheit requires both a scale change and an offset:

°F = (°C × 9/5) + 32

The 9/5 accounts for Fahrenheit degrees being smaller — 180 of them span what 100 Celsius degrees span. The +32 accounts for the different zero.

Miss the offset and you get nonsense: 20 °C becomes 36 °F instead of 68 °F. That is cold instead of comfortable, and it is a plausible-looking number, which is what makes it dangerous.

The interval trap

This is the subtle one, and it catches engineers.

Converting a temperature reading and converting a temperature difference are different operations.

If a process runs 10 °C hotter than ambient, that difference in Fahrenheit is 10 × 9/5 = 18 °F, not (10 × 9/5) + 32 = 50 °F. The offset cancels when you subtract two readings, so intervals use the ratio alone.

This is why our site keeps a separate temperature interval converter alongside the standard temperature converter. Thermal expansion coefficients, specific heat capacities and temperature tolerances are all intervals, and running them through an offset conversion is a classic and expensive error.

The one point they agree

Celsius and Fahrenheit cross at −40. It is the single temperature where the two scales give the same number, and it makes a useful sanity check: if your conversion does not return −40 °F for −40 °C, your formula is wrong.

Which to use

  • Kelvin for any physics, chemistry or thermodynamic calculation.
  • Celsius for weather, cooking and general use nearly everywhere.
  • Fahrenheit for weather and cooking in the US, and a handful of Caribbean territories.
  • Rankine (absolute, Fahrenheit-sized degrees) survives in some US thermodynamics and aerospace work.