Update- more detailed explanation below.
I have been monitoring minimum temperatures and rainfall since 2003 for 10 locations in South and Central Queensland.
These are: Rockhampton, Clermont, Longreach, Charleville, Cunnamulla, Surat, Oakey, Taroom, Maryborough, and Amberley.
If you thought August was cold, you’re right.
I have been calculating temperature anomalies using my own method, which is:
The difference between daily raw Tmin and the running 30 day mean of long term monthly means, both converted to Kelvin (to reduce the relative size of winter variation), expressed as a percentage of the monthly mean. This smooths the transition from month to month of the daily anomalies. I then calculate a 30 day running mean of these anomalies for the 10 locations. (See below.)
On 21 August this value reached
-1.273 -1.282. The next lowest was on 25 August 2008 when it reached -1.252 -1.261. (Made a small correction!)
The 4th order polynomial trend line shows the past, not the future, but there is no doubt that this is the coldest month and year of this decade.
The trends (for this decade) are similar.
Update: My method is rather idiosyncratic and needs further explanation.
Download from Climate Data Online raw daily Tmin.
Convert to Kelvin (Celsius + 273.2) (TminK)
Enter BOM’s long term monthly mean for each month. Convert to Kelvin.
For each day of each month calculate the centred 31 day running mean of the long term monthly mean. (MeanK).
Calculate daily % anomalies ((TminK – MeanK)/ MeanK %).
Calculate 30 day, 90 day, and 365 day means.
Vertical gridlines are at 30 day intervals.