- JoNova - https://joannenova.com.au -

Historic documents show 30-40%* of Australia’s warming trend is due to “adjustments”

UPDATE: *Chris has been over the entire dataset again, and makes a correction that adjustments account for 30-40% of the rise. A bit less than half. Headline updated. See his site for the newer stats. March 9, 2015

Adjustments that cool historic temperatures have almost doubled Australia’s rate of warming.

CSIR published “Meteorological Data” 1855 – 1931

 There was a time back in 1933 when the CSIRO was called CSIR and meteorologists figured that with 74 years of weather data on Australia, they really ought to publish a serious document collating all the monthly averages at hundreds of weather stations around Australia. Little did they know that years later, despite their best efforts, much of the same data would be forgotten and unused or would be adjusted, decades after the fact, and sometimes by as much as one or two degrees. Twenty years later The Commonwealth Bureau of Census and Statistics would publish an Official Year Book of Australia which included the mean temperature readings from 1911 to 1940 at 44 locations.

Chris Gillham has spent months poring over both these historic datasets, as well as the BoM’s Climate Data Online (CDO) which has the recent temperatures at these old stations. He also compares these old records to the new versions in the BOM’s all new, all marvelous, best quality ACORN dataset. He has published all the results and tables comparing CDO, CSIR and Year Book versions.

He analyzes them in many ways – sometimes by looking at small subsets or large groups of the 226 CSIR stations. But it doesn’t much matter which way the data is grouped, the results always show that the historic records had warmer average temperatures before they were adjusted and put into the modern ACORN dataset. The adjustments cool historic averages by around 0.4 degrees, which sounds small, but the entire extent of a century of warming is only 0.9 degrees C. So the adjustments themselves are the source of almost half of the warming trend.

The big question then is whether the adjustments are necessary. If the old measurements were accurate as is, Australia has only warmed by half a degree. In the 44 stations listed in the Year Book from 1911-1940, the maxima at the same sites is now about half a degree warmer in the new millenia. The minima are about the same.

Remember that these sites from 1911-1940 were all recorded with modern Stevenson Screen equipment.  Furthermore, since that era the biggest change in those sites has been from the Urban Heat Island (UHI) effect as the towns and cities grew up around the sites. In some places this effect may already have been warming those thermometers in the first half of the last century, but in others UHI can make 5 to 7 degrees difference.

If Australian thermometers are recording half a degree higher than they were 70 – 100 years ago, we have to ask how much of that warming is the UHI effect? Common sense would suggest that if these older stations need any correction, it should be upward rather than downward to compensate for the modern increase in concrete, buildings and roads. Alternatively, to compare old readings in unpopulated areas with modern ones, we would think the modern temperatures should be adjusted down, rather than the older ones.

The Official Year Book 1953

Chris Gillham discusses the potential size of the UHI changes:

“In 2012 and 2013 it was anticipated that UHI warming in south-eastern Australia will continue to intensify by approximately 1C per decade over and above that caused by global warming (Voogt 2002), with tests in 1992 showing a UHI influence up to 7.2C between the Melbourne CBD and rural areas. [PDF]

Smaller but significant UHI influences were found in regional towns, with a 1994 test observing a UHI intensity up to 5.4C between the centre of a Victorian town and its rural outskirts.”  [PDF]

 

The historic CSIR data:  226 stations from 1855 – 1931

The situation with adjustments stays roughly the same if we go back even further. Gillham compared 226 stations during the period from 1855 -1931 and the average is about half a degree less than what it is now — from 2000-2014.

The first station in the CSIR record, Melbourne, starts in 1855. Each year, new stations came online. By 1865 there are ten stations and by 1880 there are nearly 30.

Ideally we could compare 50 stations which didn’t move or start and stop over the same period, but even the ACORN dataset in the 1900s doesn’t do that, introducing new stations up to the 1970s.

It is hard to draw conclusions from the CSIR record as is. But neither can it be ignored. Roughly two thirds of the temperatures were recorded on Stevenson screens, but much of the data in the 1800s was recorded on screens, sheds and shades until Stevenson screens were introduced across Australia over the 20 year period from 1887 – 1907. And scientists in the 1930s were very much aware of the effect of slight changes in screens as one long running comparison of different screens side by side had already been going for over 30 years in Adelaide. (I’ll write more on that soon).

It’s rough but, as rough guides go, it’s the only data we have. Other peer reviewed papers have estimated Australia’s average temperature change to 0.09C  in 1000AD based on two groves of trees in Tasmania and New Zealand. Wouldn’t thermometers be kinda useful?

One small piece of good news is that at least the early CDO records maintained by the BoM online appear to match the averages within the Year Book and CSIR tables. At least the copies of the original data put online are accurate as far as these rough tests go.

The Bottom line

There is a treasure trove of information in these historic documents for people interested in long-term climate.

The difference between the original records and the adjusted ACORN dataset suggests that the adjustments cooled original temperatures by 0.4C between 1910 and 1940, which means that around 45% of the modern “warming” trend is due to these homogenisations and adjustments which have not been independently justified and oddly appear to go in the opposite direction to what common sense would suggest might be necessary. In the older and larger CSIR tables, there is an overall cooling adjustments of 0.5C.

Thanks to Chris Gillham for the massive amount of data crunching and tracking it takes to provide meaningful numbers.

Chris Gillham’s Conclusions:

Downward ACORN adjustment of historic temperature records from weather stations before 1940 adds 0.3C or 0.4C to Australia’s rate of climate warming since 1910 but the reason for the downward adjustments is unclear.

Various timescale and station comparisons show insignificant changes or warming up to 0.5C from <1931 to 2000-14. These temperatures from 1855 to 1940 are compared to what the BoM describes as the hottest decade ever recorded in Australia (2014 claimed as the third hottest).

Other historic documents add weight to the evidence that pre-1910 temperatures were not significantly cooler than current readings.

For example, On the Climate of the Yass-Canberra District published in 1910 by Commonwealth Meteorologist Henry Hunt shows temperatures at 10 locations were on average 0.1C warmer in all years before 1909 than in 2004-2013. Hunt also presents <1909 summer and winter mean temperatures at six northern Australia locations which average 0.2C warmer than those locations in 2004-2013 (download PDF).

The CSIR and Year Book temperature datasets are unadjusted records compiled by Australia’s leading scientists and weather experts in the mid 20th century and are accurate but differ from BoM records that are adjusted in both RAW and ACORN.

Their dataset timescales include the first 85 years of temperature recording at most weather stations across Australia in a network more than twice as large as ACORN, and their averages are a legitimate historic record indicating climate warming has been significantly less than calculated with adjusted data since 1910.

REFERENCES

Historic temperature sources:

Current temperature sources:

 

9.6 out of 10 based on 134 ratings