Data from: High Accuracy 14C Measurements for Atmospheric CO2 Samples from the South Pole and Point Barrow, Alaska by Accelerator Mass Spectrometry

Accelerator mass spectrometry (AMS) was used to obtain 14CO2 measurements from flasks collected at the South Pole and Point Barrow, Alaska, USA, at the Scripps Institute of Oceanography network station (Keeling et al., 1989) (see also http://cdiac.ess-dive.lbl.gov/trends/co2/sio-bar.html). Air samples for years 1985-1991, originally collected in 5-liter flasks, had been stored in Pyrex sealed-off tubes after stable isotope analysis (Roeloffzen et al., 1991). These samples were more or less evenly distributed over time. Typically, the extracted CO2 from 2-5 flasks from consecutive weekly measurements were put together in one flame-off tube. The tubes were broken in the standard inlet arrangement of a dual inlet stable isotope ratio mass spectrometer, and a stable isotope measurement was performed using the same instrument used in the original analyses (Roeloffzen et al., 1991). Then, the CO2 was cryogenically trapped and converted to graphite (Aerts-Bijma et al., 1997). For the majority of the samples the amount of CO2 was sufficient for our regular, 2 mm diameter AMS targets, containing about 1.5 mg of C. For most of these, two targets could be produced from one CO2 flask. In some cases, however, the amount of CO2 was only sufficient for a smaller target. Individual standard samples were graphitized along with the corresponding atmospheric CO2 samples, so that any day to day variability in the graphitization circumstances would be visible in the standard samples. The 14C is reported as ∆14C as per Stuiver and Polach 1977. More information about 14C standards can be found at: http://www.c14dating.com/agecalc.html. More information about sample preparation and analysis at the Groningen AMS can be found at: http://www.rug.nl/ees/onderzoek/programmas/radiocarbonams/sampleTreatmentAms The primary literature reference for the material presented here is Meijer et al. (2006). The statistical analysis included curve fitting described by Cleveland (1979). A seasonal pattern was then fit to the residuals, and the data are given in terms of these trends and departures from them