Our work on trapped gases in ice cores has focused on the isotopic composition of O2 and N2, and the O2/N2/Ar ratio, of the trapped gases. Four environmental characteristics are encoded in these gas properties. The δ15N of N2 records gravitational fractionation. Gases in glacial ice are trapped 50-120 m below the surface of an ice sheet, as burial leads to densification and the sintering of ice grains. The uncompacted ice above the trapping depth (or closeoff depth) is a porous medium allowing molecular diffusion with little or no advection through most of its length. Under these conditions, the partial pressure of each gas (or isotope) will increase with depth according to the barometric equation, and the partial pressure of heavy gases (or isotopes) will increase faster than the light. Reconstructing atmospheric properties requires correcting for gravitational fractionation, and this correction is made by measuring the δ15N of N2, a gas whose atmospheric composition has not changed for many millions of years. At times of rapid temperature change, δ15N (and all gases) records another environmental characteristic: the temperature gradient from the surface to the closeoff depth. In a diffusive medium, isotopes of gases will fractionate according to temperature gradients, with heavier isotopes generally enriched at the cold end. Snow is an effective insulator, so that, after temperature changes rapidly, there is a temperature gradient between the surface to the closeoff depth for about 100 years, the length of time required for the new temperature to penetrate to the closeoff depth. Gases in the firn reach their equilibrium profiles in about a decade. Hence at times of rapid temperature change, there is a change in the isotopic composition of gas trapped at the closeoff depth that records the surface variation. This isotopic change adds to the gravitational fractionation when the surface warms, and subtracts from it when the surface cools. The influence of temperature on δ15N (or other gases of constant atmospheric composition) was first recognized by Jeffrey Severinghaus and Todd Sowers.
The third environmental characteristic recorded by the gas properties is written in the isotopic composition of O2. In a nutshell, the isotopic composition of O2 records a number of interesting properties about the status of the global biosphere. The triple isotope composition of O2 gives a fairly robust estimate of the fertility of the global biosphere through time. The δ18O of O2 reflects a number of important properties of the biosphere, including relative rates of production by the terrestrial and marine realms, and the global response of the land biosphere to water availability. The challenge of δ18O studies is to tease apart the various influences, and this endeavor is a work in progress.
The fourth characteristic is local insolation, imprinted in the O2/N2 ratio of ice cores on the East Antarctic Plateau (Vostok, Dome Fuji, and Dome C). When gases are trapped in glacial ice, O2 is preferentially excluded (the O2/N2 ratio of ice core trapped gases is typically 0.5-1.5% lower than the atmospheric ratio). O2 is excluded preferentially to N2 because O2 is the smaller molecule, and more easily escapes when bubbles close. The degree of exclusion is somehow related to surface insolation; we know this because the O2/N2 record of Vostok and Dome Fuji are highly coherent with local summer solstice insolation. The mediating mechanism is not known, but we can say that insolation influences some property of ice at the surface that, in turn, determines the extent to which O2 is excluded during closeoff. The faithful link between O2/N2 and insolation allows us to date polar ice cores by orbital tuning, arguably the most accurate means of dating the deep ice cores of the East Antarctic Plateau when annual layers are absent.
δ18O of O2 also turns out to be an important dating tool, in 2 ways. Given the long O2 turnover time in the atmosphere, about 1,200 years between production and respiratory consumption, O2 is well mixed and its isotopic composition is constant throughout the atmosphere. Ice core can be correlated by matching up their δ18O of O2 records. Also, δ18O of O2 covaries coherently with June insolation at 65° latitude. This covariation provides another tool for orbital tuning of ice core chronologies, and the results agree very well with chronologies inferred from O2/N2 ratios. Applications of δ18O of O2 to ice core dating include correlating the Vostok, Byrd, GISP2, Taylor Dome, and Siple Dome cores over the past 100,000 years, and establishing an absolute chronology for the deep Vostok ice core.
As indicated above, the robust contribution of the isotopic composition of O2 to studies of the biosphere involves estimates of past productivity from the triple isotope composition of O2. O2 is useful in this respect because it is mass-independently fractionated in the stratosphere. Normally, 17O is fractionated half as much as 18O, and the δ17O of any compound is 0.5 times δ18O. Atmospheric O2 is an exception. In the stratosphere, there is an isotope exchange (or isotope scrambling) reaction with CO2, mediated by ozone. As shown by Thiemens, Mauersberger, Boering and colleagues, 17O is anomalously fractionated during this reaction: the δ17O of O2 decreases by 1.7 times as much as δ18O, rather than 0.5 times as much. This stratospheric anomaly mixes throughout the global atmosphere. How big is the anomaly in air O2? It depends on the relative rates of reactions in the stratosphere, which produce the anomaly, and the biosphere: respiration consumes anomalous O2 in air, and photosynthesis replaces the lost O2 with new molecules that are normally fractionated. Given estimates of the stratospheric rates in the past (which depend primarily on the CO2 concentration) and data on the O2 isotope composition, we can reconstruct global productivity through time.
Dating the Vostok ice core from O2 variations and δ18O of O2:
As described above, these two properties of the ice core gas record both vary with insolation curves. δ18O of O2 varies with insolation because the hydrosphere and the biosphere respond, along with ice volume and other climate properties, to changes in northern hemisphere insolation. O2/N2 varies with local insolation, which affects ice properties that govern O2 exclusion when gases are trapped in ice. The two properties give independent orbital tuning chronologies that have been shown, for the period prior to 200 ka, to be nearly identical. An important part of the Ph. D. dissertation of grad student Makoto Suwa is extending the O2/N2 record of the Vostok core to younger ages, and improving the resolution of the entire record. The result will be an orbitally tuned chronology for the Vostok record prior to 100 ka with an accuracy of about ± 2 kyr.
We have also compiled a record of O2/N2 for the Greenland GISP2 ice core, showing that the link with insolation is present at this site as well.
Dating tropical and mid-latitude ice cores by gas stratigraphy:
Through heroic efforts, Lonnie Thompson (Ohio State University) has collected ice cores from tropical and temperate glaciers that record climate history of these areas in high resolution back to the last ice age. The tops of these cores can be accurately dated by counting annual layers. We plan to contribute to this work by applying gas stratigraphy to date the older sections of the cores, where annual layers can no longer be distinguished. The relevant gas properties are δ18O of O2, 17Δ of O2, and CH4 concentration.
Studying the geochemistry of ice cores from the Dry Valleys, Antarctica:
David Marchant, Boston University, has accumulated evidence over many years that ice as old as 9 Ma is preserved in the rock glaciers of Mullins Valley and other tributaries of Beacon Valley, Antarctica. Extraordinary preservation is possible because, as ice flows down the valleys, it passes into a zone of net ablation where dirt and rocks carried within the glacier are exposed and blanket the ice, suppressing further ablation. Our interest in this work is twofold. First, we have recently determined, by analyzing Ar isotopes in samples from Vostok and the EPICA Dome C core, the rate at which the 40Ar/36Ar ratio is rising due to outgassing of radiogenic 40Ar from the crust and mantle. We can now date trapped gases by analyzing their Ar isotope composition. This work is complicated by the large gravitational fractionation of Ar isotopes in ice core trapped gases, but we still achieve an uncertainty of about ± 150 kyr. Second, we will investigate the possibility of reconstructing ancient greenhouse gas concentrations by analyzing the trapped gases. This attempt may be complicated by metabolism, which can produce CO2 and CH4. An alternative is to reconstruct ancient atmospheric CO2 concentrations by measuring the triple isotope composition of O2, which depends on the atmospheric CO2 burden.