Antarctic sea ice hits lowest winter maximum on record: US data

> "This is the lowest sea ice maximum in the 1979 to 2023 sea ice record by a wide margin," said the NSIDC, a government-supported program at the University of Colorado at Boulder.

> Sea ice extent data from multi-channel [1979–2012] passive microwave sensors provided one of the longest-running climate records from satellites in the form of the sea ice extent index [1]. However, while the more than 30-year record is now long enough for typical climate-scale change, it is still relatively short. Even adding in the Nimbus V single-channel Electrically Scanning Microwave Radiometer (ESMR) data extends the record by only another 6 years and that data is not as complete or consistent [2]. Earlier data are only available from operational ice charts that are of variable quality and do not cover all ice-covered regions. This was particularly true in the Antarctic, where there were few ice charts before 1972.

Sections IV. ERROR ANALYSIS and V. ANTARCTIC SEA ICE EXTENT ESTIMATES, 1964 VS. 1966 then go on to explain how calculating ice extents before the passive microwave sensor era is unreliable, with gems like:

> The 1964 satellite data provide fairly complete coverage over the Antarctic, allowing the ice edge to be clearly seen in most areas in at least one image over the three weeks. This provides a reasonably robust estimate of sea ice extent for the period, though some uncertainty results from the variation of the ice edge location during the time period. The Nimbus I data had at most eight images of any area in the three weeks of data collection, so it was difficult to distinguish homogeneous clouds and homogeneous ice.

> Coverage of Nimbus II AVCS images for the Antarctic for 1966 was fairly complete but the film record ends on August 31 so a direct comparison with the 1964 estimate from Nimbus I is not possible. [...] The results for May 1966 had higher uncertainty because images often lacked clean ice edges. For both May and June there were longitudes with no ice edge detection, leading to further uncertainty

> The fact that our August 1966 is such an outlier indicates potential issues in the Nimbus August data. In addition, the drop in extent from July to August is not seen in the passive microwave record and is not physically likely.

> Unfortunately, we have been able to uncover any corroborating information on Antarctic sea ice in 1964 or 1966, including any data or imagery from US Navy aircraft over-flights.

Basically, any measurements from before 1979, and especially before 1972, are not particularly reliable. Some researchers have produced the best numbers they could from the data they did have, but there needs to be a big asterisk by them.