In the second of two papers interpreting atmospheric CO2 observations obtained during the First Global Geophysical Experiment (FGGE) Hawaii to Tahiti Shuttle expedition of 1979-1980, we consider features of the atmospheric CO2 cycle revealed by the mean annual component of the CO2 concentration field. For this purpose the FGGE data, after decomposition into seasonal, secular, and north-south varying components, were extended to 71¿N and to the south pole by included smoothed mean annual data based on CO2 observations at seven land stations. The resulting mean annual north-south profile was referred to a datum of January 1, 1980. An additional profile for January 1, 1962, was derived from observations at five land stations, at an Arctic ice floe station, and from ships, during the period 1960-1963. Both profiles have been examined using a one-dimensional meridional diffusive transport model of the atmospheric circulation in which the latitudinal dependence of the eddy diffusion coefficient between 14.5¿N and 14.5¿S has been determined from seasonal variations in atmospheric CO2, and its mean value estimated from halocarbon and 85Kr data. The difference between the two CO2 concentration profiles is explained as being due almost entirely to the combustion of fossil fuels, which caused 2.7 ¿ 1015 g more carbon to be injected into the air in 1980 than in 1962, predominantly north of 14.5¿N. A residual profile was obtained by subtracting the predicted effect of the injection of fossil fuel CO2 from the 1980 profile. This residual profile has a peak concentration near the equator which, according to the model, is a result of the release of 5.0 ¿ 1015 g yr-1 of carbon to the atmosphere between 14.5¿N and 14.5¿S, balanced by an equal removal from the atmosphere poleward of these latitudes.

The source-sink couple inferred to produce this CO2 exchange is consistent with the distribution of CO2 partial pressure in the equatorial ocean surface water, as observed on the FGGE Shuttle Expedition, provided that the air-sea exchange rate of CO2 there is 30 mol m-2 yr-1. This exaggerated rate probably reflects the lack of vertical resolution in the model, such that the CO2 concentration near the sea surface is assumed to apply to the entire air column. The residual profile also shows a higher average concentration in the southern hemisphere than in the northern. The cause of this difference, of the order of 1 ppm, could not be resolved by the model owing to lack of information on the character of sources and sinks and transport behavior poleward of 14.5¿N and 14.5¿S; it may be owing to either oceanic or land biospheric CO2 exchange with the air or even to time dependent atmospheric transport.