2.2.2. Flask Sample Carbon Dioxide Measurements

Air samples collected at the CMDL global cooperative air sampling network sites (Figure 2.2) enable measurement of the atmospheric variations and distribution of trace gases (CO₂, CH₄, CO, H₂, N₂O, and SF₆) and stable isotopic ratios in CO₂ (d¹³C and d¹⁸O). These measurements are used to document trends and help constrain estimates of sources and sinks (budgets) for these climatically and chemically important species. Carbon dioxide mole fractions were measured in 8100 samples in 1996 and 7700 samples in 1997. Methods have been described by Conway et al. [1994]. In January 1997 sampling was begun at Bobabeb, Namibia (23°34'S, 15°2'E) in cooperation with the Desert Research Foundation of Namibia. In June 1997 a trial air sampling program was begun at Summit, Greenland (72°35'N, 38°29'W) in cooperation with the University of Arizona Department of Hydrology and Water Resources and the National Science Foundation. This program will continue for 1 year and possibly longer if the site continues to be occupied year-round. Samples have also been collected from two sites in Kazakstan since October 1997. One of these sites, Plateau Assy (44°35'N, 77°45' E), is in the desert and the other, Sary Tkum (43°15'N, 77°45' E), is on a mountain. The cooperating agency is the Kazakh Scientific Institute of Environmental Monitoring and Climate. Sampling ended at Mould Bay, Canada (76º15'N, 199º21'W) in May 1996 when the Canadian government discontinued operations at this site. An excellent record of measurements has been obtained at Mould Bay starting in 1980. The 1995-1997 annual mean CO₂ mole fractions for the 49 network sites active in 1997 are given in Table 2.3.

Fig. 2.2. Network of continuing measurements by the Carbon Cycle Group

TABLE 2.3. 1995-1997 Annual Mean CO₂ Mole Fractions From Network Sites

The square brackets indicate insufficient data to calculate annual mean.

The 1997 annual means are provisional.

TABLE 2.4. 1995-1997 Annual Mean CO₂ Mixing Ratios From Pacific Ocean Cruises

In 1996 and 1997 we began converting the shipboard sampling programs from evacuated 3-L flasks to flushed and pressurized 2.5-L flasks. An automated sampling system was installed on the Brisbane Star of the Blue Star Ship Management Ltd. in cooperation with the University of Rhode Island. An Air Kitzis Sampler (Airkit) was installed on the M/V Frontier Express, a Caltex tanker, (South China Sea) in November 1997. It is hoped that the change in sampling method will lead to a higher percentage of good pairs and provide sufficient air for all analyses. The 1996-1997 annual mean CO₂ mole fractions for the Pacific Ocean cruises and for the South China Sea are reported in Tables 2.4 and 2.5.

TABLE 2.5. 1995-1997 Annual Mean CO₂ Mole Fractions From South China Sea

To increase the efficiency of air sample analysis a new analytical apparatus was constructed and implemented in July 1997. This apparatus combines two gas chromatographs and an NDIR analyzer with a common sample inlet manifold to consolidate three separate analytical operations into one. This new system called MAGICC (Measurements of Atmospheric Gases Influencing Climate Change) is capable of measuring CO₂, CH₄, CO, H₂, N₂O, and SF₆ in 34 samples per 9-hour day.

The MAGICC system incorporates changes from the old CO₂ flask measurement apparatus design which appear to have eliminated small systematic losses of CO₂ that occurred in the old system. On the basis of a thorough analysis of intensive experiments and daily measurements of test flasks, we have applied corrections to all the CO₂ measurements made on the old system. These corrections are +0.10 ppm for samples measured on the old system from July 1987 through November 18, 1993, and +0.24 ppm for samples measured from November 19, 1993, until the old system was retired on April 10, 1998. We have detected no significant CO₂ offsets on MAGICC from July 1997 through April 1998.

Global fossil fuel emissions and the globally-averaged CO₂ uptake is shown in Figure 2.3. The solid line is based on the assumption that the measurements at the Mauna Loa Observatory, Hawaii (MLO) represent the global atmosphere, the gray band (including a Monte Carlo error estimate) is based on the cooperative flask sampling network. The difference between the fossil fuel emissions and the global net uptake is the atmospheric increase.

Fig. 2.3. Global rate of fossil fuel emissions and global net uptake of CO₂ by the oceans and terrestrial ecosystems combined, as determined from atmospheric measurements. Solid curve: assuming that the MLO record represents the entire atmosphere perfectly; shaded band: as determined from the global air sampling network and a two-dimensional (latitude, height) atmospheric transport model, including Monte Carlo error estimates in which the sampling network was varied.

The average atmospheric increase rate over the period of the flask network (1980-1997) is 1.47 ppm yr^-1. It is evident that there is significant interannual variability in uptake that is much larger than the variability in fossil fuel emissions. Understanding the processes that produce these variations is a primary goal of the air sampling program. For instance, the high rate of uptake observed in 1992-1993, particularly in the northern hemisphere, appears to have resulted from increased carbon storage in the terrestrial biosphere during the cool period following the Mount Pinatubo eruption. This conclusion is based on the spatial pattern of CO₂ during those years and on measured ¹³C/¹²C isotopic ratios. Some of this temporarily stored carbon was released back to the atmosphere in 1994 and 1995, followed by another period of as yet unexplained low CO₂ increase in 1996.

[BACK] [CONTENTS] [NEXT]