ElasticTruth

Special Report: The Rutherglen Proof: When the Algorithm Overrides the Thermometer

A single chart proves homogenisation can invent cooling

The Bureau of Meteorology has a standard defense when its climate data is questioned: "The adjustments are necessary to correct for station moves and instrument changes. We use neighboring stations to verify the corrections."

For many adjustments, that defense holds up. 

But at Rutherglen in January 2007, it collapses completely. 

The chart below is the perfect visual proof of how the Bureau's algorithm can manufacture a warming trend where the raw data shows none.

The "Statistical" Adjustment

In January 2007, the Bureau applied a -0.61°C cooling adjustment to all data before that date at Rutherglen. Because there was no documented site move or instrument change, this was labeled a "Statistical" adjustment. The algorithm simply decided a break existed.

But the neighbors say otherwise.

Reading the Chart

 The Top Chart (Raw data): This shows the temperature difference between Rutherglen and its 10 nearest neighboring stations. Look at the blue vertical line (January 2007). The black line is smooth. The neighbors show a step of just-0.01°C. Nothing happened. The weather at Rutherglen was perfectly normal relative to its region.

 The Bottom Chart (ACORN-SAT): Look at the orange line at the same vertical line. There is a massive, artificial step. The Bureau's algorithm applied a -0.61°C cooling adjustment to the past. 

The Manufactured Trend

By cooling the past by 0.61°C and leaving the present alone, the Bureau artificially created a warming trend at Rutherglen that does not exist in the raw data or the neighboring stations. 

This is not a correction. This is the algorithm manufacturing a warming trend by cooling the past, despite the raw data and the neighboring stations proving no such change occurred.

The Bottom Line

The Rutherglen adjustment is one of 1,575 adjustments the Bureau has made. If the algorithm can invent a -0.61°C cooling step where the neighbors show nothing, how many other "statistical" adjustments are equally unsupported? 

This single chart proves the Bureau's algorithm is not just correcting data; it is capable of overriding the raw evidence and manufacturing trends. 

Thanks to Patrick Lloyd and his assistant, Claude AI for the analysis.