Trying to fix past mistakes through homogenization
Lots of things can muck up a perfect thermometer spot, like shade, new roads, new screens, or old paint. In order to remove these annoying non-climatic effects, the BOM compares each station to those around it to look for odd changes. In theory this sounds like a good idea. In practice it’s more like hepatitis – bad news that spreads. It’s a rogue code, sweeping through records, trying to find undocumented changes, and enabling any amount of revisionism.
The BOM “detects” these mysterious shifts at each site through thermometers that may be hundreds of kilometers away, even across a mountain range or the Bass Strait.
Among other sites, Cape Bruny in far south Tasmania has been corrected with the help of Ballarat 812 km away on the mainland, over mountains and across the Bass Strait. In 1991 Cape Bruny was found to be “statistically” wrong, and adjusted down by over half a degree.
…
All these sites marked in red were used to correct the record at least once at Cape Bruny, a distant island in the far south of Tasmania.
It’s a tough life for old screens: Their wooden houses get cracked, their lawn doesn’t get mowed, they get moved around the corner, and then the old thermometer gets swapped for a new one. Sometimes careless people build buildings and carparks in inconvenient places, bringing shade and windbreaks. Sometimes these are recorded, sometimes not.
No one is saying that the raw data doesn’t need any adjustments, but the BOM should be doing the historical research first, as the unpaid BOM audit team does — and for these kinds of changes, especially Bill Johnston. Only after the BOM has all the documentary photos and site moves should it even think about letting an algorithm sweep through the raw data trying to make up for all the inadequacies of the recording equipment or the lack of good information about the site.