Variability in the carbon cycle over the past millennium (7810)
Measurements of gases in polar ice sheets provide the best available evidence of changes in atmospheric composition over the industrial period and the previous millennia. Records from Law Dome, Antarctica in particular benefit from high age resolution and overlap with recent direct atmospheric measurements. They are widely used as forcings in climate model simulations (eg. IPCC AR5) and in biogeochemical model interpretations of trace gas budgets and carbon-climate feedbacks. As such, it is essential that the records represent the original atmospheric composition as closely as possible.
We have revised our records of CO2, CH4, N2O and carbon-13 in CO2 (13C-CO2) from measurements of Law Dome ice cores and firn air. We include measurements of a new ice core from Law Dome, an ice core from Dronning Maud Land (Antarctica) and firn air from South Pole. Previous measurements (spanning 2 decades of measurement campaigns) were reselected using a consistent rule-based system and calibrations to most recent scales. Dating has also been improved with new ice chronology and firn air modelling. The main features and decadal variability of the Law Dome records are confirmed by these revisions and by new data. The Dronning Maud Land ice core measurements, and new records emerging from the West Antarctic Ice Sheet Divide core, show small differences that can be explained mainly by their wider air age spread.
We demonstrate improved consistency in 13C-CO2 measurements between Law Dome ice, South Pole firn and the Cape Grim (Tasmania) atmospheric data, providing evidence that our new data reliably extend the atmospheric record back in time. We then present improved CO2 and 13C-CO2 measurements of the last 1000 years which show trends since 1800 AD consistent with the release of anthropogenic carbon, and significant variability on decadal timescales. We use a Kalman Filter-Double Deconvolution to infer the net CO2 fluxes between the atmosphere, ocean and land. Measurements of additional atmospheric tracers, such as carbonyl sulphide, and simple carbon-climate modelling, help identify the causes of the variations.