Rapid Oscillations (Decades)

Ice cores provide the primary record with high enough resolution to allow exploration of decadal-scale system variability. Ice core records were used to discern decadal-scale variability as early as the 1970s [ Thompson, 1977], although cores date back to 1950 [ Bradley, 1985]. In the 1980s, studies of cores from Greenland revealed abrupt and massive climate change at these timescales [ Hammer et al., 1986; Dansgaard et al., 1989]. Heavy isotope and dust concentration profiles from the Greenland cores indicate that at the end of the Younger Dryas, southern Greenland warmed at a rate of 1.4C per decade (7C over 50 years) [ Dansgaard et al., 1989]. The authors noted that while rapid, it was not unprecedented, as warming occurred as fast as 4C per decade over North Greenland during this century (the 1920s) [ Dansgaard et al. 1989]. In 1993, several groups published the results of detailed measurements obtained during the European Greenland Icecore Project (GRIP) and the U.S.-Greenland Ice Sheet Project 2 (GISP2) [ Alley et al., 1993; Dansgaard et al., 1993; GRIP Members, 1993; Mayewski et al., 1993; Taylor et al., 1993]. Measurements included annual snow and dust accumulation, chemical and isotopic composition, and electrical conductivity, with high enough resolution to identify even 3- to 5-year oscillations [e.g., Alley et al., 1993].

Dansgaard et al. [1993] used stable isotope measures from the full length of the 250 kyr (thousand-year) Greenland record to examine climate stability during glacial periods. They found that apart from the most recent 10,000 years, instability dominated the North Atlantic climate over the last 230,000 years, with O changes as great as 10% over 1 kyr (1.4C per century). Mayewski et al. [1993] focused on chemical composition of the ice at the time of the Younger Dryas (12.8-11.5 kyr BP [before present]). The GISP2 record indicated an abrupt drop of calcium levels by 300 ppb over 20 years at the end of the Younger Dryas [Figure 1]. Calcium serves as a measure of crustal dust loading; lower levels indicate less intense circulation over continental regions, changes in source area, or decreased aridity [ Mayewski et al. 1993]. The extreme rapidity of the changes of chemical composition and dust in core ice (1-3 years) implies that events at the end of the last glaciation potentially were responses to a threshold in the North Atlantic climate system [ Alley et al., 1993]. Similarly, Taylor et al. [1993] concluded that the observed rapid climate oscillations, as reflected in changes in electrical conductivity on the order of 60 Å (micro Angstroms) over 5 years, might reflect ``flickering'' between two preferred states of atmospheric circulation.

Rates of change in the earth climate system over decadal timescales have also been observed in records of sea sediments, sea surface temperature, and turnover of tropical forests [ Lehman and Keigwin, 1992; Parrilla et al., 1994; Phillips and Gentry, 1994]. Studies of sea sediment cores are often not of high enough resolution to identify patterns of variability on timescales more rapid than centuries [e.g., Bond et al., 1992]. In the Norwegian trench, however, deposition rates are exceptionally high (5 m per thousand year) and allow resolution of a few decades [ Lehman and Keigwin, 1992]. Based on faunal and O variations, Lehman and Keigwin [1992] conclude that the southeast Norwegian Sea was alternately influenced by cold and temperate water masses, reflecting sudden changes in the poleward advection of warm subtropical surface waters. The corresponding rates of sea surface temperature change in this region were 2C per decade. Direct measurements of sea temperature were examined by Parrilla et al. [1994], who compared Atlantic Ocean records from 1957, 1981, and 1992. The authors noted that over the 35-year period, ocean temperatures between 800-2,500 m depth warmed consistently across the North Atlantic at a maximum rate of 1C per century, with cooling of 0.09C per century below a depth of 3,000 m. A deepening of isotherms (regions of constant temperature), isohalines (constant salinity), and isopycnals (constant density) by 50-70 m in the depth range 1,000-1,500 m was also noted.

Significant decadal rates of change in terrestrial ecosystems have also been measured. Phillips and Gentry [1994] measured annual rates of basal area increment and mortality of individual trees over the period from 1934 to 1993 as an estimate of forest turnover at 40 tropical sites (Figure 2). Their analysis revealed significant increases in turnover since 1950, with higher rates after 1980 (nearly a 125% increase over the last forty years). The authors proposed several alternative hypotheses for the increased rates of turnover, all of which involve coupling to the atmosphere, including fertilization by increasing carbon dioxide, the effects of progressively more extreme weather, and micrometeorological effects of nearby deforestation. It is unclear what the long-term implications of changing forest turnover are for carbon storage or biophysical exchanges [e.g., Dickinson, 1992], but they may be significant because of effects on standing biomass, albedo, surface roughness, and leaf area index.