THOMAS R. KARL
During the 1970s, the 1980s, and into the 1990s atmospheric scientists have accelerated research aimed toward identifying and explaining the presence of decade-to-century scale climate fluctuations. The motivation for special consideration of atmospheric climate fluctuations on these longer scales has its roots in the numerous major climate events whose occurrence is so well documented in the instrumental record. Probably the best-known of these fluctuations occurred about 1970, when rainfall suddenly decreased over the Sahel; the causes of this jump are discussed in this section by Nicholson (1995) and Shukla (1995). Other examples include the Dust Bowl years in North America during the 1930s, the diminished intensity of tropical storms affecting the East Coast of the United States during the 1960s, 1970s, and much of the 1980s, and the wet weather of the 1970s and 1980s over much of the United States.
As access to climate records has broadened during the past few decades, researchers such as Hurst (1957), Mandelbrot and Wallis (1969), Mitchell (1976), Douglas (1982), and Lorenz (1986) have presented evidence suggesting that the notion of a static climate is no longer tenable, even on less-than-geological time scales. We have come to realize that 30 years of data, the length of time that has been used to compute temperature and precipitation "normals" (Court, 1968), is inadequate to define climate. It does not provide us with sufficient information either to minimize adverse climate impacts within such sectors as energy, water supply, transportation, environmental quality, construction, agriculture, etc., or to maximize the availability of the climate-governed resources, such as water and energy, on which both natural and man-made systems depend.
Several climate fluctuations that have affected the United States serve to illustrate this point. Beginning about 1975, the interannual variability of mean winter temperatures and total precipitation averaged across the contiguous United States substantially increased (Figures 1a and 1b); this increase persisted at least through 1985. In contrast, the interannual variability had been very low during the previous 20 years. The mean temperature increased dramatically over the United States during the 1930s, coincident with a large decrease in summer precipitation (Figures 1c and 1d), before returning to more typical conditions. The wetness of the 1970s and the first half of the 1980s, which is clearly evident in Figure 1e, resulted in record-setting high lake levels and caused considerable economic damage and human suffering (Changnon, 1987; Kay and Diaz, 1985). Over the past decade another climate fluctuation has been evident over the United States; as Figure 1f shows, temperatures increased in a jump-like fashion during the early 1980s. Interestingly, the temperature discontinuity that is apparent between January and June is not reflected by the temperatures during the second half of the year.
Due to the limited span of the instrumental climate record, evidence for climate fluctuations on decade-to-century time scales is biased toward higher frequencies. (The situation is much worse in the oceans; as Wunsch (1992) points out, the absence of comprehensive oceanographic data has led many oceanographers to focus on identi-