mation about natural variability and rapid climate change on time scales as short as decadal.
The paper in this chapter by Cole et al. (1995) describes the shorter, but often high-quality, records that can be found in the chemical composition of corals. Coral records spanning hundreds of years can be pieced together from a single region to provide information regarding sea surface temperature, upwelling, rainfall, and winds. Because corals often grow at rates three orders of magnitude greater than the rate of sediment accumulation in the deep sea, their temporal resolution is exceptionally good, providing high-fidelity records of natural variability on time scales as short as seasonal.
Finally, in addition to the oceanic proxy indicators preserved through time that are discussed above, some researchers have proposed that the spatial and temporal distributions of plankton and fish populations are intimately linked to ocean and climate conditions. The paper in this chapter by McGowan (1995) reviews the current state of our knowledge and discusses the problems in using such data as proxies for ocean/climate variabilities. McGowan's view of the relationships between climate and various proxies contrasts with the findings of Dickson (1995) that appear in Chapter 3. Their differences highlight the uncertainties surrounding this relatively new endeavor and the challenges associated with using highly complex biologic indicators.
Scientists have been examining a number of noninstrumental atmospheric proxy data sources in search of clues to past climate conditions. Each type of record contains information on one or more aspects of climate. Among these are historical documents (almost all aspects of climate), tree rings (temperature, precipitation, pressure patterns, drought, and runoff), ice cores (temperature, precipitation, atmospheric aerosols, atmospheric composition, and more), lake levels (runoff and drought patterns), and varved sediments (temperature, precipitation, and solar radiation). With the exception of the historical documents, perhaps, the climate information in these proxy indicators must be isolated from the nonclimate portion of the signal and any accompanying noise.
Of the various sources of atmospheric proxy indicators, the greatest attention has so far been given to the use of historical documents, tree rings, and ice cores. (Of these, tree rings have been subjected to the most rigorous testing as sources of information on past climate.) A number of papers in this volume make reference to non-instrumental records, primarily tree rings (Jones and Briffa, 1995; Cook et al., 1995b, and Reifsnyder, 1995, all in this chapter; and Diaz and Bradley, 1995, in Chapter 2.) The uses of ice cores (Grootes, 1995, in this chapter), which reflect past temperatures, and of lake levels (Street-Perrott, 1995, in this chapter), which reflect overall net moisture supply to the landscape, are presented as well.
The paper in this chapter by Jones and Briffa (1995) sets out to show the potential value of proxy records by looking at dendroclimate reconstructions of summer temperatures of four regions: northern Fennoscandia, the northern Urals, Tasmania, and northern Patagonia. In none of the thousand-year reconstructions did the twentieth century stand out as the warmest century, although it was among the warmest. Jones and Briffa also provide a useful discussion of the potential limitations of single-site dendroclimate reconstruction, and of the difficulties and uncertainties associated with comparison of multi-site records. Similar limitations can be expected to apply to other site-specific proxy indicators, so they represent a general caution about the interpretation of climate proxy records.
Despite the limitations, these long tree-ring records clearly document large and pronounced climate signals, such as the century-long cold period beginning in 1550. Tree-ring evidence is of critical importance in establishing the magnitude and duration of natural climate anomalies, since the instrumental records are too short to do so and also may reflect anthropogenic contamination. In addition, the tree-ring data can provide key information for the interpretation of gradual trends in climate. For example, they show that the period from 1880 to 1910 was anomalously warm throughout much of the Northern Hemisphere. Thus, the apparent magnitude of the present warming trend will depend on when the selected period begins.
The contribution to this chapter of Cook et al. (1995b) examines the power spectrum of a 2290-year reconstruction of warm-season Tasmanian temperatures to detect signs of periodic decadal-scale fluctuations. The paper suggests that the decadal-scale temperature anomalies over Tasmania during the twentieth century, both warm and cold, have been driven in part by long-term climate oscillations. It neither supports nor eliminates greenhouse warming as a possible contributor to the recent temperature increase in Tasmania. However, their study also investigates mechanisms of climate change that are testable (e.g., internal forcing related to deep-water formation, or external forcing by long-term solar variability) and provides specific explanations of synoptic-scale variability (e.g., the expansion and contraction of the circumpolar vortex).
Grootes (1995, in this chapter) discusses the potential role of ice-core records in reconstructing decade-to-century-scale climate variations. He seems confident that the new ice-core records being obtained from the summit of the Greenland ice sheet provide an accurate and remarkably detailed history of changes in climate over the North Atlantic basin as far back as 200,000 years ago. Indeed, there is little doubt as to the potential utility of the high-resolution, fast-response climate records from ice cores. At present, however, they can be obtained only from high-latitude or