Noosa's climate record

The Bureau of Meteorology has been recording weather at Tewantin since 1895. Split those raw observations into daytime maximums and overnight minimums and they tell two different stories — which is the part usually left out.

Rainfall, 1896–2025

Noosa averages 1680 mm a year across 125 complete years of record. The striking feature is not a trend — it is the range. The wettest year on record, 1898, delivered 3062 mm. The driest, 1902, managed 511 mm — less than a third as much.

Over the full 125 years the trend is -4.181 mm per decade. Across the whole record that amounts to roughly -54 mm against a 1680 mm average — far smaller than the year-to-year swings, and not a meaningful signal.

Temperature, and why there are three lines

Noosa's temperature record comes from two sites about 100 m apart: Tewantin Post Office, which closed in 1996, and Tewantin RSL Park, which took over. It is tempting to join them into one 131-year line. That would be wrong, and the data says so.

What the three lines say

Where the warming actually is

Split the record into daytime maximums and overnight minimums and the answer stops being ambiguous. Across the long Tewantin Post Office record they behave completely differently:

Over 38 complete years from 1907 to 1995, Tewantin's daytime maximum trends at -0.03 °C per decade — flat, with a confidence interval straddling zero. Over the very same years the overnight minimum rises at +0.34 °C per decade, and the confidence interval is nowhere near zero.

The warming is in the nights. The days did not warm.

The decade means say it plainest

From the 1900s to the 1990s the daytime maximum moves from 25.77 to 25.89 °C — +0.12 °C in about ninety years. The overnight minimum moves from 13.77 to 16.70 — +2.93 °C. The gap between day and night closes from 12.01 to 9.19 °C.

A diurnal range narrowing by nearly three degrees is a very large change, and larger than greenhouse warming alone is expected to produce. Compressed range is the classic signature of things happening around the thermometer as much as above it — vegetation growing in, buildings and paving arriving, screen and instrument changes, a site slowly becoming more sheltered. It can also be genuine climate. What it is not is something you can wave away in either direction, and it is why the raw record cannot settle this on its own.

Note the gap in that table: no decade between the 1900s block and the 1960s qualifies, because those years are missing too many days. The comparison is really early-century against late-century, not a continuous march.

The modern pair, split the same way

Do the same max/min split on the two modern stations and the 2014 level shift shows up in both elements at both sites — but not equally:

The step is larger in the daytime maximum (+0.61 and +0.71) than in the overnight minimum (+0.44 and +0.47). At Tewantin, not one of those within-segment trends is distinguishable from zero — the whole modern movement at that site is the step.

Which element is the problem

Now the test that separates a real signal from an instrument. Compare the two stations on days both recorded, and do it separately for maximum and minimum. If a genuine regional difference were shifting, both elements should shift. Only one does:

The overnight difference holds near +1.1 °C for eighteen straight years. It wanders by a few tenths and finishes essentially where it started. The two sites agree about nights, consistently, throughout.

The daytime difference does not. It sits near −0.3 °C through 2008–2012, stretches to −0.93 °C by 2015, then closes and re-levels near zero from 2019 — and stays there. That is a permanent change of roughly 0.35 °C in how these two instruments relate to each other, in daytime readings only, on 355–365 paired days a year.

Which splits the modern step in two. The overnight part — +0.44 °C at Tewantin, +0.47 °C at the airport — appears at both sites with a stable relationship between them, which is what a real regional signal looks like. The daytime part sits on top of a documented drift between the instruments, and should not be read as clean evidence of anything until someone accounts for that drift.

So: where was the warming?

In the nights, both times — and that is the only part of it that survives scrutiny.

Across the historic Tewantin record the overnight minimum rises +0.34 °C per decade while the daytime maximum does nothing at all (-0.03). Ninety years on, Noosa's afternoons are about as warm as they were in the 1900s; its nights are nearly three degrees warmer.

In the modern record both elements step up around 2014. The overnight step shows at both stations while the offset between them stays put — that behaves like a real regional signal. The daytime step sits on top of a 0.35 °C permanent drift between the two instruments that appears in daylight readings only, and until someone accounts for that drift it cannot carry much weight.

So the honest single sentence is: Noosa's nights have warmed substantially and consistently; its days have not, and the one stretch of record that suggests otherwise is the stretch where the instruments stopped agreeing with each other.

Anyone quoting a single Noosa temperature trend at you — in either direction — is compressing all of that into one number.

How complete is the record, really?

This is the part usually left out. The daily record starts in 1895, but "starts in 1895" is not the same as "is usable from 1895". A year only earns an annual figure here if it has at least 350 daily readings. Many early years do not come close.

Tewantin Post Office temperature readings. The 1930s average 142 readings a year — well under half a year of observations. Any claim about how Noosa's 1930s compared to its 2020s is standing on that.

Method

• Source: Bureau of Meteorology daily observations, mirrored locally. Stations 40264, 40908, 40861.
• These are raw observations as published by the Bureau, not homogenised or adjusted by us.
• Annual figures use only years with 350+ daily readings. Trends are ordinary least squares on those annual values.
• Temperature is never spliced across the 1996 site change. Rainfall is combined across the two Tewantin sites, because a rain gauge is far less sensitive to a 100 m move than a thermometer.
• Check any of it yourself at bom.gov.au/climate/data.