Atmospheric CO2 Concentrations from the Commonwealth Scientific and Industrial Research Organization (CSIRO) GASLAB Flask Sampling Network (March 1991 - December 2006)

• Main Authors: Steele, L., Krummel, P., Langenfelds, R.
• Format: Dataset
• Language: English
• Published: Environmental System Science Data Infrastructure for a Virtual Ecosystem; Carbon Dioxide Information Analysis Center (CDIAC), Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (USA) 2008
• Subjects: 54 ENVIRONMENTAL SCIENCES; Alert, NWT, Canada; Cape Ferguson, Australia; Cape Grim, Australia; Casey, Antarctica; Estevan Point, BC. Canada; Macquarie Island, Australia; Mauna Loa, Hawaii; Mawson, Antarctica; Shetland Islands, Scotland; South Pole, Antarctica; Antarctic; Canada; Ferguson; Grim; Scripps; South Pole; The Antarctic; Antarc*; Antarctica; Macquarie Island; South pole
• Online Access: https://dx.doi.org/10.3334/cdiac/atg.018; https://www.osti.gov/servlets/purl/1389352/

Individual measurements have been obtained from flask air samples returned to the CSIRO GASLAB. Typical sample storage times range from days to weeks for some sites (e.g. Cape Grim, Aircraft over Tasmania and Bass Strait) to as much as one year for Macquarie Island and the Antarctic sites. Experiments carried out to test for changes in sample CO2 mixing ratio during storage have shown significant drifts in some flask types over test periods of several months to years (Cooper et al., 1999). Corrections derived from the test results are applied to network data according to flask type. These measurements indicate a rise in annual average atmospheric CO2 concentration from 357.72 parts per million by volume (ppmv) in 1992 to 383.05 ppmv in 2006, or an increase in annual average of about 1.81 ppmv/year. These flask data may be compared with other flask measurements from the Scripps Institution of Oceanography, available through 2004 in TRENDS; both indicate an annual average increase of 1.72 ppmv/year throuth 2004. Differences may be attributed to different sampling times or days, different numbers of samples, and different curve-fitting techniques used to obtain monthly and annual average numbers from flask data. Measurement error in flask data is believed to be small (Masarie et al., 2001).