Skill of the two 20th century reanalyses in representing Antarctic near-surface air temperature

The performance of the 20th century reanalysis (20CR) and of the pilot reanalysis of the 20th century (ERA-20C) for Antarctic near-surface air temperature is investigated based on quality-controlled monthly temperature measurements with a time coverage of at least 15years from 42 automatic and manual weather stations, and temperature reconstructions. On the inter-annual timescale, ERA-20C and 20CR agree well with the temperature observations since 1957/1958, but there are systematic biases relative to absolute values with cold bias at East Antarctic coastal stations, and warm bias over the inland stations. An abrupt upwards shift in Antarctic annual and seasonal (except summer) mean temperature averaged over the East Antarctic ice sheet and the Antarctic ice sheet occurs around 1950 in 20CR and around 1979 in ERA-20C, which is probably related to the increasing counts of assimilated observations. Moreover, the significantly negative trends of seasonal and annual mean temperature in 20CR since the late 1970s are found to be largely spurious, which most likely results from the poor representation of variability in sea surface temperature (SST) around Antarctica in the 20CR's boundary forcing. These results imply and urge that caution must be taken when applying the two global reanalyses for evaluating long-term climate changes at high southern latitudes.