The normal temperature quoted in a forecast is not a long-run average of everything on record. It is a thirty-year average that gets replaced on a fixed schedule, and the replacement changes what counts as unusual.
What a normal actually is
A climate normal is the average of a weather element over a defined thirty-year block, computed station by station for each day or month of the year.
The block is fixed rather than rolling, so every station in the network describes the same span. That keeps comparisons between two cities meaningful.
The convention is decadal. When a decade closes, the oldest ten years drop out of the block and the most recent ten enter it.
Why thirty years and not more
Thirty years is long enough to average out individual hot and cold years, which arrive in irregular clusters driven by ocean and atmospheric patterns rather than at random.
It is also short enough that the average still describes conditions a resident or a utility planner would recognise. A century-long average would blend climates that no longer exist.
The length is a compromise between statistical stability and present relevance, and it was settled as a working convention long before warming trends became the central question.
How the update shifts the baseline
Because each update drops an older decade and admits a newer one, the baseline moves along with the climate it is measuring.
Where temperatures have risen, the new normal sits above the old one. A day that read as above normal under the previous block can read as ordinary under the current one.
The weather has not changed in that comparison. The yardstick has.
Why utilities and planners depend on them
Normals are the reference for the departure figures in daily forecasts, and for the heating and cooling demand estimates that power companies use to plan supply.
They also feed design assumptions in building standards and agricultural planning, where the expected range across many years matters far more than any single season.
A stale baseline would systematically misstate demand and load, so the periodic refresh is a practical necessity rather than a bookkeeping habit.
Why normals are a poor measure of change
Because the baseline moves, comparing a year against the current normal understates how far conditions have shifted over a longer span.
Measuring a trend requires holding one reference period fixed and comparing everything to it, which is why long-term analysis uses a stationary baseline instead of the rolling one.
The two references answer different questions, and confusing them is the usual source of arguments about whether a season was genuinely hot.