Meridional eddy diffusion model of the transport of atmospheric carbon dioxide: 1. Seasonal carbon cycle over the tropical Pacific Ocean

Abstract

An extensive set of atmospheric CO2 observations were obtained during the First Global Geophysical Experiment (FGGE) Hawaii to Tahiti Shuttle expedition of 1979–1980. These data have been examined using a one‐dimensional meridional diffusive transport model of the atmospheric circulation. The observed CO2 concentration field was first decomposed into three parts: a seasonal component consisting of two harmonic functions of time, with periods of 1 year and 6 months; a mean annual north‐south concentration profile; and a linear trend with time, independent of latitude. In the first of two papers interpreting these data, we consider which features of the north‐south eddy transport are revealed by the seasonal component of the CO2 concentration field. We assume that the FGGE data represent a zonally and vertically uniform CO2 field, and we restrict our analysis to the latitude belt between 14.5°N and 14.5°S, where seasonal sources and sinks of CO2 have minimal influence on the atmospheric CO2 concentration. The strongest feature of the CO2 data for calibrating the model is the steadily diminishing amplitude of the annual harmonic from north to south. We show that in the absence of sources and sinks the magnitude of the eddy diffusion coefficient as a function of latitude can be prescribed from this amplitude, provided that two additional quantities are specified at a single reference location. These additional quantities are the amplitude of the first harmonic of the seasonal flux and the phase difference between that harmonic and the first harmonic of the local concentration signal. Selecting 14.5°S for this reference location, we find through comparisons of the model prediction with the FGGE CO2 data that the diffusion coefficient depends primarily on the selected amplitude of the seasonal flux and is practically independent of the phase difference. On the basis of the goodness of fit of the model predictions to the CO2 data, we set a lower limit on this flux amplitude, corresponding to an average eddy diffusion coefficient of 4.6 × 109 cm2 s−1. This is equivalent to an upper limit of 1.4 years on the interhemispheric exchange time for CO2. We are unable to set an upper limit on the diffusion coefficient because the phasing of the first harmonic in the FGGE data shifts so slightly with latitude that even very large diffusion coefficients yield satisfactory model predictions of the seasonal CO2 concentration field. The relative variation in the diffusion coefficient with latitude, which is derived solely from the amplitude of the first harmonic of the CO2 data, reveals two zones of resistance to interhemispheric CO2 exchange. One, near 8°N, corresponds to the well‐known intertropical convergence zone of the wind field which exists around the entire earth. The second, near 8°S, correlates with a weaker wind convergence zone in the central Pacific. This regional feature evidently affects the CO2 concentration field in the vicinity of the FGGE data set, but it may not represent a global feature.