is perhaps the most well behaved, with remarkable stability over the first half of the record. However, an increase in amplitude is indicated in the A.D. 900-1200 interval (during the Medieval Warm Period), followed by greatly reduced amplitude in the A.D. 1300-1800 period (during the Little Ice Age). Since about 1800, the 204-year waveform has increased in amplitude to a level not seen since A.D. 1300. The other waveforms are much more complicated, with each showing varying degrees of amplitude and phase modulation.
The statistical evidence for persistent decade-to-century-scale oscillations in warm-season Tasmanian temperatures is provocative and begs for a physical explanation. We do not have one yet, but some relationships between sea level pressures, sea surface temperatures, and land air temperatures over Tasmania for the past 100 years may provide some clues for where to look.
As noted earlier, Cook et al. (1992) suggested that the 31- and 55-year oscillations could be related to 20-to-30 and 40-to-60 year fluctuations in zonally averaged July sea level pressure (SLP) data in the 40° to 50°S latitude zone (Enomoto, 1991). Specifically, the 20-to-30-year SLP fluctuation was considered by Enomoto (1991) to be a standing oscillation of wave-number-zero structure related to the expansion and contraction of the circumpolar vortex. Inter
estingly, one of the centers of maximum variance was located in the Tasman Sea region ( 130°E to 170°W). The 40-to-60-year fluctuation was considered by Enomoto (1991) to be an oscillation with wave-number-one structure, meaning an eccentricity of the circumpolar vortex. Each of these SLP fluctuations showed large amplitude changes in the 1880 to 1920 interval in the Tasmania-New Zealand sector, with anomalously low pressure being indicated for the 1890 to 1905 period. This implies an expansion of the circumpolar vortex at that time, with a greater tendency for cool, southwesterly winds and below-average temperatures over Tasmania.
A comparison of Hobart SLP data and Tasmanian actual and reconstructed temperatures for the November-to-April warm season supports this interpretation. Figure 7 shows the