The Mystery of Hidden Temperature express

Original: "Fundamental temperature in the superstatistical description of non-equilibrium steady states"
· Sergio Davis
arXiv:2604.17035 · 2026-04-18 · CC BY 4.0 · 2 min
Scientists have found a way to measure invisible temperature in complex systems.

Hot tea cools down, and its temperature is the same throughout the cup. But in many systems (for example, in the rarefied gas between stars or in plasma) temperature is not uniform: in one place particles are faster, in another slower. Physicists call such systems "superstatistical": it's as if the tea consisted of many small regions, each with its own temperature.

The problem is that this "local" temperature cannot be measured directly: it is not tied to a single quantity that you can probe with an instrument. So scientists have to guess its distribution, which is unreliable. The authors of the new work found a way out: they showed that the hidden temperature can be linked to the system's energy. Energy can be measured, and thus temperature can be computed.

A bag of candies of different sizes: you can't see the candies, but you can weigh the bag. Knowing the total weight and the number of candies, you can guess what the candies are like — small or large. Similarly, here: the system's energy is the weight of the bag, and the temperature is the size of the candies.

The authors tested their method on a concrete example — the so-called q-canonical ensemble, which describes systems with unusual statistics. They computed how temperature is distributed at a given energy and obtained the full distribution without complicated mathematical transformations. It's as if, instead of emptying the bag, you learned to determine the size of the candies by the sound they make when poured.

Now physicists have a tool to look inside complex non-equilibrium systems — from plasma in the lab to distant star clusters. And perhaps better understand how order emerges from chaos.

In superstatistics, temperature can be not just a number but an entire probability distribution, like a weather forecast that doesn't say "it will be 20 degrees" but "with 60% probability it will be between 18 and 22."

Scientists
John von NeumannLudwig BoltzmannJosiah Willard GibbsJacob BekensteinWilhelm WienCharles Bennett
Concepts
thermal state effective temperature thermal equilibrium temperature anomaly Boltzmann equation Boltzmann distribution equation of state function partition function quantum thermodynamics probability distribution thermometer statistical mechanics hypothesis testing linear inverse model heat heat map thermalization statistics permutation test quantum mpemba effect