Natural and anthropogenic changes in atmospheric CO 2 over the last 1000 years from air in Antarctic ice and firn (original) (raw)

1996, Journal of Geophysical Research

A record of atmospheric CO 2 mixing ratios from 1006 A.D. to 1978 A.D. has been produced by analysing the air enclosed in three ice cores from Law Dome, Antarctica. The enclosed air has m•paralleled age resolution and extends into recent decades, because of the high rate of snow accumulation at the ice core sites. The CO 2 data overlap with the record from direct atmospheric measurements for up to 20 years. The effects of diffusion in the fun on the CO 2 mixing ratio and age of the ice core air were determined by analyzing air sampled from the surface down to the bubble close-off depth. The uncertainty of the ice core CO 2 mixing ratios is 1.2 ppm (1 o). Preindustrial CO 2 mixh•g ratios were in the range 275-284 ppm, with the lower levels during 1550-1800 A.D., probably as a result of colder global climate. Natural CO 2 variations of this magnitude make it inappropriate to refer to a single preindustrial CO 2 level. Major CO 2 growth occurred over the industrial period except during 1935-1945 A.D. when CO 2 mixing ratios stabilized or decreased slightly, probably as a result of natural variations of the carbon cycle on a decadal timescale. CO 2 levels is required to improve the understanding of the natural and human-perturbed carbon cycle and the climatic impact of CO 2 as a greenhouse gas. Measuring air extracted from polar ice is the most direct way of reconstructing past atmospheric CO 2 mixing ratios. Less direct techniques are based on measurement of the isotopic ratios of carbon preserved in organic matter. They suffer from larger measurement error, uncertainties caused by physiological influences on photosynthetic fractionation, and by the assumptions required to derive atmospheric mixing ratios from isotopic values in a carbon exchange model. Ice cores, in particular those from cold Antarctic sites, offer the following advantages: (1) whole air is enclosed in bubbles in the ice; (2) the ice is a relatively inert storage medium for CO 2 and many other atmospheric trace gases; (3) the enclosed air generally represents the "background" atmosphere, remote from biological or anthropogenic CO 2 sources or sinks; (4) timescales ranging from tens of years to hundreds of thousands of years can be investigated by selecting appropriate sites; and (5) other relevant physical and chemical i•fformation resides in the same stratigraphy, such as trace acids for dating [Hammer, 1980], or water isotopic ratios as climatic indicators (in the case of Law Dome, Morgan [1985]). There are several possible difficulties in the ice technique: (1) reactions involving CO 2 may occur if the ice approaches melting [Neftel et al., 1982], or if it contains high concentrations of impurities, conditions that may have affected the CO 2 in some sections of Greenland ice cores [Delmas, 1993; Staffelbach et al., 1991]; (2) cracks in cores may release air or allow contamination to enter; (3) ice core samples must be carefully refrigerated to avoid post coring melting (PCM) which may change the composition of the air bubbles [Pearman et al., 1986]; (4) the pressure in the ice at about 500 m and deeper may be sufficient (depending on temperature) for the bubbles to disappear and form clathrates 4115 4116 ETHEPdDGE ET AL.: CHANGES IN, CO 2 OVER THE LAST 1000 YEARS [Miller, 1969] which may complicate air extraction; (5) dating the ice by annual stratigraphy becomes difficult at sites with low-accumulation rate or where ice flow disturbs the layered sequence at depth; (6) the age resolution of the enclosed air can be limited by the progressive closure of the air bubbles and, to a lesser degree, diffusion of air from the ice sheet surface through the firn layer to the closure depth; and (7) associated with the diffusion are possible fractionation effects [Craig et al., 1988; Schwander, 1989] which are small for CO 2 mixing ratios but significant at the precision level of carbon isotope ratio measurements. Most of these difficulties can be avoided by appropriate selection of the drilling site and careful handling of the ice sample. Desirable characteristics of ice core sites for CO 2 studies are negligible melting of the ice sheet surface, low concentrations of impurities, regular stratigraphic layering which is undisturbed at the surface by wind or at depth by ice flow, and high snow accumulation rate.