Temperature Trends at the South Pole and McMurdo Sound

A detailed analysis of atmospheric temperatures at the South Pole and McMurdo Sound is presented. Missing data are common, especially in the stratosphere, and the usual practice of computing monthly means as an average of all available observations produces unreliable results because the annual cycle is aliased onto the interannual variations and longer term trends. A methodology to rectify this involves computation of the smoothed mean annual cycle for each day of the year and then subsequently analyzing the anomalies. The persistence of the anomalies within each month reveals that regular observations about every three days are required to produce a reliable climate record throughout the troposphere and lower stratosphere. A comparison of station data with lower stratospheric analyzed values from the National Meteorological Center (NMC) reveals big discrepancies at times arising mainly from methods used to produce the NMC analyses. The mean annual cycle of temperature features the coreless winter at low levels at both stations. The largest amplitude annual cycle occurs at 10 mb with maximum temperatures in December. The maximum occurs progressively later at lower levels down to 200 mb, where it occurs in February. The seasonal transition occurs more rapidly in spring than in autumn. Interannual fluctuations are dominated by a quasi-biennial variation. Noticeable downward trends in temperature were found mainly near the time of greatest variability, in late spring. Temperatures decreased between 50 and 100 mb in October and November and at 100 mb in December and January; but only from 1985 to 1987 have monthly means gone outside the range of previous variations. The most extreme anomalies, of as much as −21°C, were found in November 1987. The downward trends arise from a delay in the spring warming apparently brought about by the diminished solar heating due to low ozone amounts associated with the “ozone hole.” Trends from 1979 to 1986 or 1987, which is the period of satellite data on temperatures and ozone, are not representative of the overall record which begins in 1956.