of years. These changes are well correlated with variations in the Earth's orbit around the Sun.
It was recognized in the nineteenth century that variations in the Earth's orbit would cause changes in incoming solar radiation that could be important in controlling ice ages. Theoreticians first calculated how these variations would interact, and the study of deep-sea sediments has yielded persuasive evidence that the recurring ice ages of the past million years are indeed closely associated with orbital cycles. These cycles cause subtle changes, particularly in high latitudes, in the seasonal variation of the incoming solar radiation, called insolation, and may be reliably calculated from celestial mechanics. The ice ages themselves are recorded in the ratio of oxygen isotopes in deep-sea sediments. This is because the elevated fraction of light isotopes in fresh water evaporated from the ocean surface and stored in ice sheets is reflected by an increased fraction of heavy isotopes in the precipitated carbonate skeletons of microorganisms living in the remaining ocean water. Variations in the oxygen isotope (18 O/16O) ratio with depth in a sediment core are widely interpreted as indicating total land-ice volume as a function of time.
The glacial record, as revealed in ice cores, sedimentary sequences, landforms, and other related phenomena, is especially useful for understanding past changes and anticipating the characteristics of future changes. In some instances, fossil pollen and other specific environmental indicators are also present in stratigraphic records.
As better cores are examined and dating procedures are refined, it has become apparent that the changes in insolation correlate with subsequent changes in ice volume. However, a correspondence between the two records requires allowance for the slow buildup of ice sheets over several tens of thousands of years in contrast with their relatively rapid decay, which introduces a degree of nonlinearity into the system response. Presumably the periodic changes in insolation are the ultimate cause, but the precise mechanism remains obscure. The presence of ice, however, does not appear essential for a cyclical response. Recently compiled geological records of 200-million-year-old lake sediments in the eastern United States show a sequence of cycles of approximately the same intervals as the present orbital cycles, spanning a period of 40-million-years. These lakes were then in the tropics. No evidence of continental glaciers or sea ice exists for this period, but local climate and lake levels were apparently influenced by a strong stable external control.
The relationship of the orbital cycles to climatic variation is a fertile research field. The deep-sea sediment record indicates that about 900,000 years ago the governing periodicity of cycles switched from 40,000 years to 100,000 years. This sudden change has not yet been explained. High-resolution records of the glacial cycles come from fast-sedimentation-rate deep-sea cores, from cores in the Greenland and Antarctic ice sheets, and from cores of mountain glaciers at low and mid-latitudes. Cores from the Vostok station, high on the Antarctic ice cap, have extended the record back to about 160,000 years ago, so a remarkably complete record is now available of how temperature varied through the whole of the last glacial cycle. Analyses of air bubbles in ice cores show that temperature and atmospheric CO2 content generally varied sympathetically. Geologists are currently investigating whether orbital variations drove the system and whether changes in oceanic circulation and biology affected the atmosphere's trace gas content, amplifying the climatic oscillations.
Abrupt changes in environmental conditions are recorded in the Greenland record, where the Dye-3 core indicates a switch from glacial to interglacial conditions within one century. Some researchers suggest that such sudden changes during the last glacial period could have been triggered by major diversions of meltwater draining from the Laurentide ice sheet. In a more recent time frame, dust and oxygen isotopic records associated with an ice record from the Peruvian Andes indicate that local transition from the Little Ice Age to current conditions could have occurred about 100 years ago and in as short a time span as a few years.
These examples show that large climatic changes can occur on many time scales, including those of critical relevance to modern society. Climate change may be initiated by variations in atmospheric and oceanic circulation patterns driven by feedback connections to other terrestrial environmental factors, such as changes in vegetation cover or in physical composition of the atmosphere influenced by volcanic or human activity. Therefore, the geomorphic history, geographic distribution, and rates of glacial advances and retreats need to be documented to permit understanding of the interconnected global associations of environmental change and to seek causal connections. These data can provide very important independent tests of the atmospheric general circulation models (GCMs)