When the Australian Bureau of Meteorology issued a statement on October 22, it was meant to close a small chapter on data hygiene. The agency, custodian of the nation’s official climate record, confirmed it had found “a small number of duplicated historical daily temperatures” in four post-1910 months relevant to ACORN-SAT, its adjusted national temperature series. The total, it said, came to “fewer than 200 daily values.” Those values had been corrected. The matter, the Bureau implied, was settled.
The Bureau’s initial acknowledgment was precise in its narrowness. The duplicated values it had identified and corrected were confined to the adjusted ACORN-SAT dataset, a carefully curated series that forms the backbone of Australia’s long-term temperature monitoring. The agency’s statement was followed by an update, which added a confident assertion: “Further investigation has detected no further duplicate months in the entire bureau network other than those quoted in the report and then confirmed by us.”
ACORN-SAT is an adjusted record, meaning it takes in only a portion of the raw historical data and applies corrections for known biases. The raw data, the unvarnished daily readings from stations across the country, is a much larger and messier beast. And it was in that raw archive that Lloyd found the copies.
These were not rounding errors or single-day glitches; they were systematic duplications embedded in the historical record.
Lloyd’s audit identified 44 distinct stretches of duplicated data across 15 weather stations. Together, they totaled 978 days of copied temperatures, spanning from 1861 to 1981. These were not subtle statistical anomalies. They were wholesale repetitions: entire months of maximum and minimum temperatures lifted from one period and pasted into another.
In Cairns, the daily maximum and minimum readings for December 1950 were copied from January of the same year. In Toowoomba, February 1981’s temperatures were a copy of February 1982. At Sydney’s Observatory Hill, 13 identical days between October and November 1890 were recorded. These were not rounding errors or single-day glitches; they were systematic duplications embedded in the historical record.
𝗔𝘂𝘀𝘀𝗶𝗲 𝗕𝗢𝗠 𝗔𝗱𝗺𝗶𝘁𝘀 𝗧𝗵𝗲𝘆 𝗖𝗮𝗻'𝘁 𝗧𝗿𝗮𝗰𝗸 𝗘𝗹 𝗡𝗶𝗻𝗼 𝗘𝘃𝗲𝗻𝘁𝘀 The Australian Bureau of Meteorology (BOM) has admitted they're struggling to accurately track El Niño and La Niña events, blaming climate change for messing up their ocean temperature readings. According to an ABC report, the old system—relying on a 30-year baseline for NINO3.4 sea surface temperatures—is no longer reliable in a warming world. El Niño events are being over-detected, and La Niña might be missed entirely, as rising ocean temps contaminate the data. BOM's fix? A new index that compares temperatures to the "current tropical mean" instead of historical averages. But this sounds like a quick patch rather than a solid solution. The real driver of weather variability is the temperature difference between the western Pacific and the NINO3.4 region, yet they're sticking with models instead of getting boots on the ground. Critics point out this approach seems lazy. Why not invest in direct measurements, like deploying more Argo buoys or chartering boats for real-time data? After all, accurate forecasts are crucial for agriculture across Australia, Asia, and the Americas, where billions depend on predicting rain. If BOM can't get this right now, how can they claim confidence in century-long climate predictions? This overhaul raises questions about the reliability of climate models when basic weather drivers like El Niño are hard to pin down. Time will tell if this "fiddle" improves things. To read the full article go to https://t.co/lE8C8pwgE2
The most uncomfortable finding in Lloyd’s analysis was not the existence of the duplicates, but their persistence. Of the 44 copied stretches he identified, 32—representing 640 days of data—were still present in the raw data files the Bureau publishes today. The Bureau had said it found no further duplicate months. Yet here they were, sitting in the downloadable datasets, untouched.
The distinction matters because the raw data is the foundation on which everything else is built. Adjusted datasets like ACORN-SAT are essential for tracking long-term climate trends, but they are only as reliable as the raw observations that feed them. If 640 days of duplicated data remain in the published raw record, any analysis that draws on that record may be working from a flawed baseline.
None of this suggests deliberate manipulation. Data duplication in historical weather records is a known archival hazard, often the result of digitization errors or legacy data-entry practices. The Bureau’s acknowledgment of the initial 200 duplicated values was a routine act of data stewardship.
For the raw data—the broader, messier, but foundational archive—independent analysis by Patrick Lloyd revealed that 44 duplicated stretches totaling 978 days existed across 15 weather stations, with 640 days remaining in the Bureau’s published raw data.
CCB’s Boss Notified of Suit Challenging Non-release of Obi’s Asset Declaration Forms Alex Enumah in Abuja The attention of the Chairman of the Code of Conduct Bureau (CCB) has been brought to a suit filed against him and the bureau, over their https://t.co/KEtAbFd0Sb
The implications ripple backward and forward. A duplicated stretch in 1890 at Observatory Hill doesn’t just affect our understanding of 19th-century Sydney weather; it subtly warps the baseline against which modern temperatures are compared. If October 1890 looks artificially like November 1890, the historical average for that month shifts, and with it