The existence of decadal-scale temperature and precipitation fluctuations in continental, hemispheric, and global instrumental records is well documented (e.g., Folland et al., 1984; Karl and Riebsame, 1984; Bradley et al., 1987a; Ghil and Vautard, 1991), if still poorly understood. Mechanisms believed to be responsible for some of the observed multi-year climatic fluctuations include explosive volcanic eruptions (Bradley, 1988), solar variability (Newell et al., 1989), lunar tidal effects (Currie, 1981), and the ocean thermohaline circulation (Stocker and Mysak, 1992). Indeed, the difficulty in determining the causes of decadal-scale climatic fluctuations has relegated the 10-to-100-year bandwidth of climatic variability to the ''gray area of climatic change" (Karl, 1988), where physical theory is at present inadequate to explain the observations.
Understanding the causes of interdecadal climatic variability is a challenging task, made all the more difficult by the brief span of the instrumental climate records. Fluctuations on the time scale of 10 to 100 years are difficult to study if the instrumental data extend back only 80 to 100 years. Therefore, extending climate records back in time, particularly in poorly covered regions of the world like the SH, is critical to the study of interdecadal variability. One of the few regions of the SH that has the potential for millennia-long, high-resolution proxy climate time series is Tasmania, with its temperate rainforests of long-lived conifers. Recently, Cook et al. (1991) reported the development of a climatically sensitive Huon pine (Lagarostrobos franklinii C.J. Quinn) tree-ring chronology from western Tasmania that extended back to A.D. 900. The yearly variations in ring width were shown to correlate well with changes in warm-season (i.e., November to April) temperature. Cook et al. (1991) then went on to draw inferences about the significance of the present-day warming trend apparent in the series since 1965, and speculated about the occurrence of the Little Ice Age and Medieval Warm Period in Tasmania.
In a later paper, Cook et al. (1992) described more fully the development and characteristics of this tree-ring chronology and extended their analyses to the quantitative reconstruction of November-to-April average temperatures for Tasmania back to A.D. 900. In so doing, they produced the first well-verified estimates of past temperature change in the Australia-New Zealand sector of the SH that extend back through the Medieval Warm Period and the Little Ice Age as broadly defined by Lamb (1965) and Grove (1988), respectively. The Huon pine temperature estimates explained approximately 37 percent of the temperature variance and correlated well with actual data withheld from the regression-based reconstruction model. Cross-spectral analysis also indicated that the temperature reconstruction was especially good at estimating decadal-scale temperature fluctuations, with magnitude-squared coherencies between actual and estimated temperatures exceeding 60 percent for periods longer than 12 years.
Since the Cook et al. (1992) paper was written, the Tasmanian temperature reconstruction has been extended back to 300 B.C. through the inclusion of tree-ring series from several old Huon pine stumps found at the same site where the original series was developed. This extended reconstruction is shown in Figure 1. (Although no year zero actually exists for the B.C. to A.D. time scale, our series