Stable Isotopes of N and Ar as Tracers to Retrieve Past Air Temperature from Air Trapped in Ice Cores

Abstract

Ice cores are paleoclimatic archives that permit the reconstruction of past local precipitation temperature (from the measurements of water isotopes) and past atmospheric gas concentration (from the analysis of the air trapped in the ice) over the past 800,000 years. However, water isotopes are not a quantitative tracer for past temperature in Greenland ice cores. Moreover, because of the entrapment process, air is always younger than the surrounding ice so that the temporal phasing between temperature and atmospheric concentration changes is difficult to estimate. Here, we present a recently developed method that permits us to infer the amplitude of past abrupt temperature changes from the air trapped in an ice core. This method is based on very accurate measurements of δ15N of N2 and δ40Ar of Ar in the air trapped in ice cores. Abrupt temperature changes at the surface of the ice-sheets create a transient temperature gradient in the firn (unconsolidated snow in the top 100 m of the ice sheet), which leads to isotopic fractionation of nitrogen and argon. The variations of δ15N and δ40Ar arising from surface temperature change are then trapped at the bottom of the firn as air bubbles close off. Then, combining the isotopic measurements with a firnification model including heat diffusion permits us to determine the amplitude of the past temperature change with an accuracy of ± 2.5°C. This method has demonstrated in particular that the Dansgaard Oeschger events that punctuated the last glacial period had an associated amplitude of up to 16°C in less than 100 years and that methane and Greenland temperature increases were synchronous (within 50 years). Finally, there is hope that such a method may be used in Antarctic ice cores to constrain the phasing between changes in local temperature and in atmospheric concentration. This is however still an ongoing task, because in Antarctica, the temperature changes are too slow to induce a strong thermal gradient in the firn and the measured evolution of δ15N and δ40Ar is only partly understood. Still, promising first results using δ15N and δ40Ar measurements in the air trapped in an Antarctic ice core have shown that over a glacial to interglacial transition, CO2 was lagging East-Antarctic temperature by 800 ± 200 years.