Summary
Programs for the measurement of atmospheric carbon dioxide using infra-red gas analysis techniques, have now been conducted by CSIRO in the Australian/Antarctic region of the Southern Hemisphere for over fifty years. In this report, changes in equipment and operating methods are described for the period from 1981 to 1995. Time series of monthly mean baseline CO2 values are tabulated for data from flask sampling in the south-eastern Australian mid- and upper-troposphere, the upper-troposphere over the Great Australian Bight and the surface at Kennaook/Cape Grim (40° 41'S; 144° 41'E; site code CGO), Macquarie Island (54° 29'S; 158° 58'E; site code MQA) and Mawson (67° 37'S; 62° 52'E; site code MAA), and from in situ monitoring at CGO and MQA. Data based on flask samples of air from the surface in three tropical areas and from ships in the Southern Ocean are also presented as listings of individual samples. Correction procedures used to bring the entire data sets as closely as possible into concordance with the 1985 WMO CO2 Calibration Scale, are reported. In particular, a study of the apparent drift rate of the Australian Interim CO2 Calibration Scale (defined in this report) during 1987-1995, involving a number of intercalibrations with other monitoring groups, enabled a linear correction to be deduced for this period. In addition to these intercomparisons, a series of other experiments were performed after 1981, to help determine the degree of precision and intercomparability of our data. From the results of these experiments and assessment of the data sets, a number of conclusions regarding the effectiveness of the operating and data selection procedures of the various CSIRO programs, were drawn. Also, the level of intercomparability of the data with those of other groups was deduced. Overall, the conclusion drawn was that, for the period being considered, inter-program precision was in the order of 0.5 ppm (parts per million mole fraction in dry air) unless specific measures were taken to ensure that comparable equipment and calibration techniques were used.