Burketown - But Which One Is It?

Burketown (029004 - [data] [graph]) is the representative station for Queensland District 2 (Gulf Country) in the Australian RCS Network. In the RCS, this station goes by the name "Burketown Post Office", but it is difficult to find in the underlying BoM data.

A look at the map shows that there are (apparently) eight stations in the Gulf Country District:

029012 - Croydon Township - [data] [graph]
029038 - Kowanyama Airport - [data] [graph]
029039 - Mornington Island - [data] [graph]
029063 - Normanton Airport - [data] [graph]
029077 - Burketown Airport - [data] [graph]
029139 - Sweers Island - [data] [graph]
029167 - OZ Century Mine - [data] [graph]

Plus the eighth station that goes by the name "Burketown".

But - where is this "Burketown" station? It doesn't show up on the list of stations on the Queensland Weather Observations page. And if we click on the "Burketown" legend on the map, it takes us here:

A non-station, apparently. But given that the data starts in 1890 for 029004 - Burketown PO, the assumption is that "Burketown PO" and "Burketown" are one and the same station. But would you bet the house on it?

After looking at the data for the rest of the stations in this District, the answer is "probably not". Here is a brief summary of the discrepancies between Station start date and Data start date for the remaining seven stations (oldest to newest):

029012 - Croydon Township. Station date: 1889; Data date: 1912 (23 years missing)
029139 - Sweers Island. Station date: 1893; Data date: 2001 (108 years missing)
029038 - Kowanyama Airport. Station date: 1912; Data date: 1965 (53 years missing)
029039 - Mornington Island. Station date: 1914; Data date: 1972 (58 years missing)
029063 - Normanton Airport. Station date: 2001; Data date: 2001
029077 - Burketown Airport. Station date: 2001; Data date: 2001
029167 - OZ Century Mine. Station date: 2003; Data date: 2003

Clearly, apart from the first station (029012 - Croydon Township), the remaining stations are useless for assessing long-term climate trends. They give a whole new meaning to "the dog ate my homework".

Consequently, the choice for the RCS representative station comes down to two candidates, geographically separated by 290km:

A) 029012 - Croydon Township. A station with a 98-year history, albeit one with data missing from the middle years, which shows a pronounced cooling trend over the period:

B) 029004 - "Burketown Post Office". A poorly-sited station of questionable origin (which doesn't appear in the day-to-day list of weather stations), but does profess to have a slightly longer data history (again with missing data in the middle). And does indicate a period of warming between 1950-1990.

Answer: Let's go with the "Dodgy Post Office" - the one that can have a warming "signal" teased out of it.

Summary:
The data from the stations in the No. 2 Queensland District are a shambles, to say the least. Across all the stations, there are no continuously unbroken data history prior to 1990, which means that any contribution to "climate records" from this District is - at best - guesswork. There also appears to be very little quality control of the basic station records - "Burketown" and "Burketown PO" seem to be used interchangeably. And finally, when in doubt about which station to include in the RCS, the default position appears to be to select the one that indicates a warming trend - however tenuous.