Antarctic temperature variations for 1957°82 have been objectively analyzed by gridding monthly data, from 16 stations, onto a 5° latitude by 10° longitude grid, from 65 to 90°S. These gridded data were used to calculate monthly values of the spatial mean temperature south of 65°S. The uncertainty in the area average is estimated to be 0.22°C for the annual values prior to 1970. After 1970 there is an additional uncertainty of about 0.10-0.16°C due to the cessation of Byrd station. The annual mean and summer areas averages show significant linear warming trends amounting to 0.74 and 0.77°C respectively. Spatial characteristics of the annual and seasonal temperature variations are described using principal components analysis of the station anomaly data. The first two principal components of the annual and winter data are similar PC1-winter is also similar to the winter pattern for linen trend found by van Loon and Williams for 1956–73. The warming trend associated with this pattern ceased in the mid-1970s. Relationships between Antarctic temperatures and various parameters are investigated using linear trend and correlation analyses. Antarctic temperatures cannot be inferred from the long Orcadas record and the relationships between Antarctic temperatures and sea ice extent are complex. The most significant correlations between mean Antarctic temperature and sea ice extent averaged around Antarctica am found in spring; warm springs tend to be associated with anomalously large maximum sea ice extent. Lower Antarctic temperatures occur during summers and winters with strong westerlies (significant at the 0.1% level in summer and at the 0.1% level in winter).
|Number of pages||13|
|Journal||Monthly Weather Review|
|Publication status||Published - 1984|