South-east Australia analysis shows 0.4C mean temperature increase since the 1800s
The 15 stations are those with adequate daily observations among the 38 analysed in the 2012 study Temperature variations of southeastern Australia, 1860–2011 which was published by the Bureau of Meteorology and based on homogenised (adjusted) monthly temperature averages. The table below calculates the average frequency and original unadjusted temperature of 10% decile temperatures (0-10%, 10-20%, 20-30%, 30-40%, 40-50%, 50-60%, 60-70%, 70-80%, 80-90%, 90-100%) at the 15 stations.
The table below calculates the average of the coldest and hottest single day each year in minima and maxima at the 15 stations, as well as their annual median and averages temperatures.
The BoM doesn't include years before 1910 in its Australian Climate Observation Reference Network (AOCRN) calculation of homogenised national average temperatures, claiming that most weather stations before 1910 used incompatible non-Stevenson thermometer screen shelters which can cause hotter maxima and cooler minima. This is true in many cases, but the charts below show this isn't always the case and the unadjusted hotter maxima and cooler minima nevertheless produce a legitimate mean temperature.
Automatic weather stations The charts below show how the frequency of very hot and very cold days in maxima and minima was stable from the 1800s until 1996 when automatic weather stations became the primary instrument for temperatures within the BoM's official observations. At the 15 south-east Australia weather stations in this analysis these extreme percentiles are exaggerated after 1996, a likely explanation being one second electronic observations which are must faster than the delay time with pre-AWS liquid-in-glass thermometer sensitivity to brief gusts of hot or cool air.
The charts above show significant differences in the frequency of the coldest 10% and hottest 10% of days produced by AWS electronic probes compared to their thermometer predecessors, resulting in 0-10% and 90-100% percentile averages 0.2C warmer in maxima and 0.2C cooler in minima when compared to 1910-1995 temperatures within the same extreme deciles (see top table). Although 1996 was the most common year among the 15 stations when the BoM adopted automatic weather stations as the primary instrument for official observations, several installed an AWS after this year (Adelaide 1995, Bathurst 1996, Bourke 1998, Brisbane 1996, Cobar 1997, Deniliquin 1997, Forbes 1996, Gunnedah 2001, Inverell 2002, Melbourne 1996, Robe 2003, Sydney 1996, Wagga Wagga 1996, Walgett 1997, Wilcannia 2000). The chart below shows the annual average frequency of maximum temperatures within the 90th percentile (hottest 10%) at all 15 south-east Australia weather stations, along with the annual average maximum across all observations, in the 20 years before and after each station had an AWS installed and used as the primary instrument.
The chart below shows the annual average frequency of maximum temperatures within the 90th percentile (hottest 10%) at all 15 south-east Australia weather stations, along with the annual average maximum within only that extreme heat percentile, in the 20 years before and after each station had an AWS installed and used as the primary instrument.
The chart below shows the annual average frequency of minimum temperatures within the 10th percentile (coldest 10%) at all 15 south-east Australia weather stations, along with the annual average minimum across all observations, in the 20 years before and after each station had an AWS installed and used as the primary instrument.
The chart below shows the annual average frequency of minimum temperatures within the 10th percentile (coldest 10%) at all 15 south-east Australia weather stations, along with the annual average minimum within only that extreme cold percentile, in the 20 years before and after each station had an AWS installed and used as the primary instrument.
This analysis suggests average maxima at the 15 south-east Australia weather stations was 0.4C warmer before 1910 than in 1910-1995 and 0.3C cooler than 1997-2021, with minima 0.3C cooler before 1910 than in 1910-1995 and 0.6C cooler than 1997-2021, and mean temperatures 0.1C warmer before 1910 than in 1910-1995 and 0.4C cooler than 1997-2021. Temperature averages and percentile frequencies from the 1800s to 1909 were influenced by the Federation Drought from 1895 to 1902 which reduced the accumulation of rainclouds, and this can be detected in the top table above wherein pre-1909 average maximum temperatures were 0.4C warmer than in 1910-1995 but average minimum temperatures were 0.3C cooler. Regional stations only A significant influence on all averages above is that four of the 15 stations are capital cities where urban heat has increased all temperatures at a greater rate than in regional locations. If these four cities (Adelaide, Brisbane, Melbourne, Sydney) are extracted from the dataset, the influence of urban heat can be reduced.
These 11 regional locations in south-east Australia had an 0.2C unadjusted mean temperature increase in 1997-2021 compared to the locations before 1910, and a 0.4C increase since 1910-1995. The regional results also show a 0.1C maximum temperature increase since before 1910 (0.6C since 1910-1995) and a 0.3C minimum temperature increase since before 1910 (0.1C since 1910-1995), which better reflects the Federation Drought's reduction in raincloud cover which normally results in cooler nights and warmer days. The charts below show the frequency of 10th percentiles at the 11 regional stations when their official observations were at manual thermometer stations before 1996 and when AWS observations were used after 1996.
The chart below shows the annual average frequency of maximum temperatures within the 90th percentile (hottest 10%) at the 11 south-east Australia regional weather stations excluding capital cities, along with the annual average maximum across all observations, in the 20 years before and after each station had an AWS installed and used as the primary instrument.
The chart below shows the annual average frequency of maximum temperatures within the 90th percentile (hottest 10%) at the 11 south-east Australia regional weather stations excluding capital cities, along with the annual average maximum within only that extreme heat percentile, in the 20 years before and after each station had an AWS installed and used as the primary instrument.
The chart below shows the annual average frequency of minimum temperatures within the 10th percentile (coldest 10%) at the 11 south-east Australia regional weather stations excluding capital cities, along with the annual average minimum across all observations, in the 20 years before and after each station had an AWS installed and used as the primary instrument.
The chart below shows the annual average frequency of minimum temperatures within the 10th percentile (coldest 10%) at the 11 south-east Australia regional weather stations excluding capital cities, along with the annual average minimum within only that extreme cold percentile, in the 20 years before and after each station had an AWS installed and used as the primary instrument.
Capital city stations only The data above from 11 regional weather stations can be compared with only the four capital city stations (Adelaide, Brisbane, Melbourne, Sydney).
These four south-east Australia capital cities had a 1.1C unadjusted mean temperature increase in 1997-2021 compared to the cities before 1910 (11 regionals 0.2C increase), and a 0.7C increase since 1910-1995 (11 regionals 0.4C increase). The charts below show the frequency of 10th percentiles at the four capital city stations when their official observations were at manual thermometer stations before 1996 and when AWS observations were used after 1996.
The chart below shows the annual average frequency of maximum temperatures within the 90th percentile (hottest 10%) at the four south-east Australia capital city weather stations excluding capital cities, along with the annual average maximum across all observations, in the 20 years before and after each station had an AWS installed and used as the primary instrument.
The chart below shows the annual average frequency of maximum temperatures within the 90th percentile (hottest 10%) at the four south-east Australia capital city weather stations excluding capital cities, along with the annual average maximum within only that extreme heat percentile, in the 20 years before and after each station had an AWS installed and used as the primary instrument.
The chart below shows the annual average frequency of minimum temperatures within the 10th percentile (coldest 10%) at the four south-east Australia capital city weather stations excluding capital cities, along with the annual average minimum across all observations, in the 20 years before and after each station had an AWS installed and used as the primary instrument.
The chart below shows the annual average frequency of minimum temperatures within the 10th percentile (coldest 10%) at the four south-east Australia capital city weather stations excluding capital cities, along with the annual average minimum within only that extreme cold percentile, in the 20 years before and after each station had an AWS installed and used as the primary instrument.
Note : Maximum temperature calculations, tables and charts for all 15 weather stations in the analysis above can be downloaded as an Excel (xls 5.7mb), a CSV (555kb) or an Apple Numbers (10.2mb). Alternatively and if you don't like using spreadsheets, all of the results can be viewed in a PDF (26.6mb). Note : Minimum temperature calculations, tables and charts for all 15 weather stations in the analysis above can be downloaded as an Excel (xls 5.7mb), a CSV (555kb) or an Apple Numbers (10.2mb). Alternatively and if you don't like using spreadsheets, all of the results can be viewed in a PDF (26.6mb).
This website produced by Scribeworks 2008-2023 |
Back To Top
