ACORN 2.4 influence on Australian temperature trends

(With ACORN 2.5 addendum)

In September 2023 the Bureau of Meteorology again rewrote Australia's temperature history with the release of ACORN 2.4, an update to its homogenised ACORN 2.3 dataset released in December 2021, introducing adjustments that have further distorted the rate of climate warming over the previous 112 years.

Adjusted ACORN 2.4 temperatures currently show 1.29C warming from 1910-1919 to 2013-2022, compared to 0.74C warming with original, unadjusted RAW temperatures.

ACORN is the Australian Climate Observation Reference Network of 112 weather stations across Australia, 104 of which the BoM uses to estimate national/state/territory monthly and annual temperatures since 1910. The rewritten ACORN 2 dataset of 2018 updated the ACORN 1 dataset released in 2011, and immediately increased Australia's per decade rate of mean temperature warming by 23%.

Each ACORN weather station has individual characteristics in terms of site moves, equipment and location or infrastructure changes over the decades since 1910 that warrant temperature adjustment, but the BoM does not adjust for variables such as urban/airport heat islands, 1972 metrication, the introduction of automatic weather stations mostly in the 1990s, an estimated 72.3% artificial warming influence of deteriorating thermometer screens, and site changes either overlooked or misinterpreted (see BomWatch).

ACORN 2.4 extends from 1910 to 2022 and introduces new or reassessed adjustments at 23 of the 112 ACORN stations, and at its long term temperature record page the BoM states that Analysis has shown the changed adjustments in ACORN-SAT 2.4 have had a negligible impact on estimated long-term warming trend in Australia.

In fact, ACORN 2.4 slightly warmed Australia's average temperatures in the first half of the 20th century when compared to ACORN 2.3, although still with significant cooling of unadjusted RAW temperatures.

When ACORN 2.4 absolute daily temperatures are measured, rather than 1961-90 anomalies, it is apparent that ACORN 2.4 has reduced average maxima warming by 0.04C and increased average minima warming by 0.01C from 1910-1919 to 2010-2022 at the 104 non-urban ACORN stations used by the BoM to calculate Australia's national temperature trends. ACORN 2.4 cooled 1910-1919 minima by a further 0.02C compared to ACORN 2.3 at the 104 stations.

Below are charts and tables for minima, maxima and rainfall at the 61 ACORN stations observing temperatures in 1910, unlike all other ACORN stations that opened in following years or decades, and thus provide datasets with the longest timespan (albeit many shifted their location from, for example, a post office to an airport).

australia acorn 2.4 minimum temperatures

104 ACORN v2.3, 2.4 and RAW stations minima Excel spreadsheets

australia acorn 2.4 maximum temperatures

104 ACORN v2.3, 2.4 and RAW stations maxima Excel spreadsheets

In its climate data frequently asked questions (click question 26), the Bureau of Meteorology claims to have made new adjustments in ACORN 2.4 to four different stations (Snowtown, Cape Borda, Wyalong and Wilcannia) and claims to have reassessed new reference data for 11 other stations (Port Hedland, Tennant Creek, Oodnadatta, Palmerville, Georgetown, Charters Towers, Birdsville, Scone, Canberra, Deniliquin, Sale).

This totals 15 stations.

However, ACORN 2.4 has adjusted ACORN 2.3 at 23 different weather stations (Adelaide, Birdsville, Bourke, Broome, Canberra, Cape Borda, Charters Towers, Gabo Island, Georgetown, Mildura, Nowra, Oodnadatta, Palmerville, Port Hedland, Port Macquarie, Rutherglen, Scone, Snowtown, Sydney, Tennant Creek, Walgett, Wilcannia, Wyalong).

Station menu

For specific details of each of the 112 ACORN stations you can either scroll down or click your station of choice from the following list to see charts and tables (with changes from ACORN 2.3 to ACORN 2.4 shown in italics), each followed by a link to return to this list.

Adelaide, Albany, Alice Springs, Amberley, Barcaldine, Bathurst, Birdsville, Boulia, Bourke, Bridgetown, Brisbane, Broome, Bundaberg, Burketown, Butlers Gorge, Cabramurra, Cairns, Camooweal, Canberra, Cape Borda, Cape Bruny, Cape Leeuwin, Cape Moreton, Cape Otway, Carnarvon, Ceduna, Charleville, Charters Towers, Cobar, Coffs Harbour, Cunderdin, Dalwallinu, Darwin, Deniliquin, Eddystone Point, Esperance, Eucla, Forrest, Gabo Island, Gayndah, Georgetown, Geraldton, Giles, Grove, Gunnedah, Halls Creek, Hobart, Horn Island, Inverell, Kalgoorlie, Kalumburu, Karijini North, Katanning, Kerang, Kyancutta, Launceston, Laverton, Learmonth, Longreach, Low Head, Mackay, Marble Bar, Marree, Meekatharra, Melbourne, Merredin, Mildura, Miles, Morawa, Moree, Moruya Heads, Mount Gambier, Nhill, Normanton, Nowra, Nuriootpa, Oodnadatta, Orbost, Palmerville, Perth, Point Perpendicular, Port Hedland, Port Lincoln, Port Macquarie, Rabbit Flat, Richmond (NSW), Richmond (Qld), Robe, Rockhampton, Rutherglen, Sale, Scone, Snowtown, St George, Sydney, Tarcoola, Tennant Creek, Thargomindah, Tibooburra, Townsville, Victoria River Downs, Wagga Wagga, Walgett, Wandering, Weipa, Wilcannia, Williamtown, Wilsons Promontory, Woomera, Wyalong, Yamba

Average of all 112 stations

Wilcannia

The most significant update of ACORN 2.3 was the addition of Wilcannia's temperature data from 1910 to 1956 in ACORN 2.4, with maximums prior to 1950 at this weather station being cooled by -0.43C and minimums prior to 1921 being cooled by -0.87C.

These adjustments are indicative of the homogenisation that has cooled Australia's temperature history and its failure to account for all elements that influence the accuracy of historic temperature observations.

Prior to metrication of Australia's temperature observations in September 1972, most observations were recorded as rounded x.0F with a majority of observers considering decimal readings on their Fahrenheit thermometers unimportant and instead recording a full number such as 64F or 86F and considering this more accurate than rounding up if they could not clearly see the decimal position of the alcohol or mercury.

Three dataset comparisons by the BoM confirmed an anomalous and probably artificial 0.1C increase in Australian temperatures at the time of 1972 metrication but, with odd reasoning, these findings were dismissed and not included in ACORN homogenisation on the basis that there may have been natural warming due to record national cloud cover and rainfall in the early to mid 1970s.

Metrication influence on Australian temperature trends, analyses all ACORN stations including Wilcannia's historic temperatures when benchmarked against Extreme Temperature Events in Australia, a 2001 PhD thesis by ACORN-SAT architect Blair Trewin from the Bureau of Meteorology.

His study estimated that among 94 of the 112 ACORN-SAT stations, the average proportion of Fahrenheit temperatures rounded by observers every day from 1957 to 1971 was 51.5%.

Independent analysis by this website estimates the same 94 stations had a rounding average of 51.3% from 1957 to 1971, the similarity to 51.5% suggesting concurrence between analyses.

The analysis by Trewin estimated that 99.7% of all temperatures at Wilcannia were rounded .0F from 1957 to 1971, and this website's independent analysis also calculated the proportion to be 99.7%.

The likely artificial cooling influence of rounded Fahrenheit temperatures at Wilcannia is ignored in the station's update within ACORN 2.4, which instead further cools the original observations by -0.43C (max <1950) and -0.87C (min <1921).

Tabulated below is a comparison between the estimates of Trewin and independent analysis of the 94 ACORN stations, and on the right the annual 0.F rounding percentage at Wilcannia since 1910.

wilcannia temperature rounding

Variable adjustments

ACORN 2.4 anomalies, which are the difference compared to 1961-90 averages, show the maximum anomaly averaging -0.396C in 1910-1919 and +1.053C in 2010-2022, suggesting 1.449C warming of maxima between those periods. ACORN 2.4 anomalies show minimum averages at -0.518C in 1910-1919 and +0.678C in 2010-2022, suggesting 1.196C warming of minima between those periods.

ACORN 2.3 maxima averaged 24.75C in 1910-1939 and ACORN 2.4 averaged 24.80C, a warming of 0.05C. ACORN 2.3 minima averaged 12.44C in 1910-1939 and ACORN 2.4 averaged 12.42C, a cooling of 0.02C.

The graphics below help to illustrate the timing and amount of difference between ACORN 2.3 and ACORN 2.4 absolute average annual temperatures at the 104 non-urban ACORN stations within the timeline of 1910 to 2021 :

ACORN 2.3 australia acorn 2.4 temperatures

Although ACORN 2.4 in total makes little difference to ACORN 2.3 absolute temperatures or anomalies when compared to 1961-90 averages, it is worth noting the ACORN bias toward cooling historic trends by comparing absolute temperatures in RAW, ACORN 2.4 and the original ACORN 1 homogenised dataset released in 2011 which adjusted temperatures from 1910 to 2017 :

acorn and raw temperature differences

The ACORN updates from version 1 to version 2.4 have added 0.44C mean temperature warming if comparing 1910-1919 with 2010-2019 (final year of ACORN 1).

It is worth noting that these results accord with separate analysis on this website showing that Australia's mean temperature has been cooling for well over a decade prior to 2024.

Accuracy of ACORN adjustments

Apart from variations between the two ACORN datasets, different stations have had their original historic observation adjusted either up down, with averages showing a bias to cooling in the early 1900s.

Apart from 1972 metrication, the introduction of automatic weather stations and site changes either overlooked or misinterpreted, another variable that may influence warming since 1910 is rainfall trends.

The chart below shows that the 61 ACORN stations that were observing temperatures in 1910 had significantly more rainfall than the 21 other ACORN stations that started observing temperatures in years or decades later but which were nevertheless recording rainfall measurements in 1910 (the remaining ACORN station were not recording rainfall levels in 1910).

acorn weather station rainfall

As 61 stations v 21 stations is not an even comparison, the chart below compares the 61 ACORN stations that were observing temperatures in 1910 with 61 other stations, both ACORN and non-ACORN, that were widely distributed across Australia and recording rainfall measurements in 1910, showing a lesser difference but still suggesting that ACORN stations in the first half of the 20th century had above average rainfall which has since declined (possibly due to a disproportionate number of such stations located in the south-east quadrant of Australia).

This above average but thereafter declining rainfall is likely to have caused some natural cooling in the early 1900s which has since dissipated, resulting in cooler averages in the first half of the ACORN timeline and potentially distorting temperature trends attributed to artificial warming caused by carbon dioxide.

acorn and non-acorn weather station rainfall

Alternatively, the 61 ACORN stations observing temperatures in 1910 can be compared with the BoM's national average annual rainfall estimates from all weather stations across Australia, charted below.

acorn and national weather station rainfall

Individual stations

Minima and maxima at each of the 112 ACORN stations are charted and tabulated below as decadal averages for ACORN 2.3, ACORN 2.4 and RAW absolute temperatures.

Average of all 112 stations

Adelaide

acorn temperatures adelaide

ACORN 2.4 influence on Australian temperature trends

ACORN 2.5 influence on Australian temperature trends