Search

Global reanalyzes are widely used for investigations of Antarctic climate variability and change. The European Centre for Medium‐Range Weather Forecasts 5th generation reanalysis (ERA5) is well regarded and spans 1940 to today. We investigate whether ERA5 reliably represents the 2‐m air temperature trends across the 1940–2022 (83 years) period at seasonal and annual time scales. We compare ERA5 temperatures with an observation‐based temperature reconstruction for Antarctica (RECON) that has monthly resolution for 1958–2022, the period of reliable observational availability. Results for individual stations are also examined. ERA5 anomalously warms Antarctica in relation RECON especially for the period prior to 1979 when satellite observations over the Southern Ocean were sparse. Trend hotspots that are shown to be artifacts are found at three locations and are present until today. The results demonstrate that ERA5 temperature trends can be questionable even today, but variability is well captured after 1979. Plain Language Summary: Reanalyzes are especially useful for describing and understanding the climate of regions where observations are few and far between. Reanalyzes are not actual conditions but approximations that have strengths as well as weaknesses. A continuing reanalysis challenge for higher latitudes of the Southern Hemisphere that has caused many previous artifacts is the transition from few satellite observations over the Southern Ocean prior to 1979 to the relative abundance since. Here we examine the 2‐m air temperature from the latest global reanalysis from ECMWF for Antarctica and show that it is still challenged by the 1979 transition. ERA5 warms Antarctica much too rapidly especially prior to 1979 and contains artificial trend hotspots at three locations right up to the present. ERA5 cannot be used for Antarctic temperature climate change applications prior to 1979 but can be applied for this purpose after 1979 with appropriate caution. Key Points: ERA5 air temperatures have spurious trends for most regions of Antarctica prior to 1979Artificial trend hotspots are present at three low elevation sites until the presentERA5 cannot be used for temperature climate change studies over Antarctica earlier than 1979 but can be selectively applied post 1979 [ABSTRACT FROM AUTHOR]