The Daily Temperature in the Record Books Is a Midrange
Gepubliceerd op 13-11-2026 · 3 min leestijd · Dagelijkse calculators
Antonio Delgado — Redacteur tradities en tijdverdrijf bij OneKitly
Symboliek · Kalenders
Getoetst aan 2 bronnen
The midrange is the halfway point between the smallest and largest values, and it uses nothing else. For 4, 8, 15, 16, 23, 42 it is 23, well above both the mean of 18 and the median of 15.5, because the single value of 42 sets half of it. That fragility is why statistics teaching mentions the midrange once and moves on — and it is also why the number matters, because the daily mean temperature in most long climate series is exactly this: the day's maximum plus its minimum, divided by two.
Add the hottest and coldest readings and halve them: that is the daily mean in most climate records. Two numbers out of 1440 decide the century's trend.
Why weather stations settled on it
A maximum-minimum thermometer records two numbers and needs visiting once a day, which is exactly what an observer with a bicycle and a logbook could sustain for a century. Continuous recording arrived far later, and switching to a true 24-hour average would break comparability with everything measured before it. Climate series are therefore kept on the old definition deliberately, and the modern hourly data is converted backwards to match. Consistency across a hundred and fifty years is worth more than accuracy on any single day.
What the two-point average misses
A day that is hot for one hour and mild for the rest gets the same midrange as one that is hot all afternoon, since only the peak counts. Comparisons of a coastal station with an inland one are affected in a systematic direction, because the inland site swings further in both directions and its midrange sits closer to the true average than its shape deserves. Researchers correct for this with adjustment models, and the size of the correction — usually a few tenths of a degree — is roughly the size of the trend being studied, which is why the methodology papers are longer than the result papers.
| Measure | Value | Points used |
|---|---|---|
| Median | 15.5 | 2 middle |
| Mean | 18 | All 6 |
| Midrange | 23 | 2 extremes |
Veelgestelde vragen
- Is the midrange ever the right choice?
- Yes, when the data is genuinely uniform between two limits, because then the extremes really do carry the most information about where the centre is. Tolerances on a machined part, arrival times spread evenly across a window, and readings from an instrument that clips at both ends all fit. Outside that case it is the least stable centre available, moved entirely by the two values most likely to be wrong.
- How different is it from a true daily average?
- Usually by a few tenths of a degree, and not in a constant direction. Temperature spends more of the day near its minimum than near its maximum, since afternoons peak briefly and nights cool slowly, so the midrange typically runs slightly warm. The bias also depends on cloud, season and latitude, which is precisely why it cannot simply be subtracted once and forgotten. Any series that switches from one definition to the other mid-record needs an explicit adjustment, and unadjusted switches are a well-known source of artificial jumps.
Artikelen die je misschien interesseren
Alle gidsen →Vergelijkbare tools
Bronnen
Een fout in dit artikel gezien?