Surprisingly, the World War I era temperatures are still changing. Mornings that seemed nippy at the time are now susceptible to frosts.
Someone should warn the farmers — except they’re all dead.
Thanks to Chris Gillham for independently and laboriously going through the new unannounced changes in another cycle of BOM’s hidden revamp of Australia’s history. ACORN 2.2 is the latest version of the Australian Climate Observation Reference Network of “the best” 112 weather stations across Australia.
Graham Lloyd, The Australian
“The bureau has now remodelled the national temperature dataset three times in just nine years,” Dr Jennifer Marohasy said.
In the last five years the ACORN re-revisions by the BOM have discovered another quarter of a degree of warming that we didn’t know about from the last hundred years. It’s not clear why the BOM doesn’t want to tell the world how good they are at correcting thermometer records from 1913. It seems like a remarkable skill.
The minima just keep getting cooler
Chris Gillham plots the longest running stations from the ACORN 2.2 set against the old raw readings:
Independent analysis of the latest changes show they added 0.06C to maximum warming and 0.11C to minimum warming from 1910-19 to 2010-19.
A series of updates to the ACORN data have added 0.228C mean temperature warming if comparing 1910-19 with 2010-17 (2017 being the final year of ACORN 1)
Clearly, the cheapest way to stop global warming would be to sack everyone at the BOM.
The Bureau admitted to Lloyd that they changed 25 sites in the latest round, most of them because of statistical revelations:
The bureau said changes were made to 20 sites on the basis of statistical analysis. According to the bureau, statistical analysis is used to identify an abrupt warming or cooling at a particular site, relative to other sites in the region.
The BOM refuses to tell us exactly how they do this — admitting that they can’t train anyone outside their sacred guild. Which reminds us of the old axiom, “if it can’t be replicated, it isn’t science”.
Apparently the future of life on Earth depends on their data, but it would take too long to train someone outside BOM to check it. Tell us again, how much does climate change matter?
Here’s the BOM explanation of the need for adjustments:
“A significant change relative to other sites indicates a non-climatic driver, which sometimes has an easily identifiable cause (e.g. a new building near a site) and sometimes does not (often these will relate to local vegetation or land surface changes)”, the bureau said. “In carrying out this statistical analysis, the bureau uses 10 years’ worth of data from multiple sites to quantify the size of the change. Adjustments are only applied where a significant change has been identified.
So instead of researching the site history as deeply as they can, the BOM use statistical
trickery analysis to figure out local site changes using some other thermometers “within a thousand miles”. What they don’t do is full historical research and documentation of the sites first. (Like the ones here, here and here, thanks Bill Johnston). So it’s possible that if, say, 10 sites got similar changes at the same time (like changing to a new small screen) the other better sites in non-changing screens could be slowly “corrected” one-by-one by the junkier sites with the right statistical algorithm. Now, perhaps the BOM have ways to make sure that doesn’t happen — but they won’t tell us what they are, they don’t publish the detailed documentation of either the full site histories or the methods they use to correct them.
As Chris Gillham notes, the BOM say the new adjustments didn’t change the long term trend. But a lot of little changes in the same direction would. The new adjustments might have also made a few local hottest ever headlines much easier to get. Suddenly the modern contenders for the longest heatwave, warmest night, and hottest summer all got a head-start.
The mysterious process of homogenization
The Bureau “homogenizes” a site according to a mystical method that uses an inexplicable group of sites that are in the same region, or at least in Australia somewhere. For example, lighthouse readings off the Bass Strait have been corrected with help from Hay, 600 kilometers away, on the other side of the Great Dividing Range. A thermometer in Ballarat has been used to correct one at Bruny Island off the South coast of Tasmania. The BOM has also corrected Alice Springs using measurements in Collarenabri nearly 1,600 kilometers away. (Thanks to Ken Stewart for that tip). When I said “a thousand miles” I wasn’t joking.
Congratulations to Chris Gillham, Jennifer Marohasy and thanks to Graham Lloyd.