Abstract
The behaviour of a numerical model for the global carbon cycle is eluidated by a simple analytical model for the biosphere. In the period 1980—1990 the ocean is estimated to have absorbed 33% of the total CO2 emission to the atmosphere in that same period. Net deforestation was responsible for 12—17% of this total emission rate, whereas the CO2-fertilization effect caused a re-absorption of 20—25%. Aggregation of the above-ground biosphere into a single pool in the model caused an oversitmation of the CO2-fertilization effect. Also, the estimate of this rate increased when the fraction of carbon assumed to remain after the transformation of litter into humus was increased, but the rate was little influenced by the model structure for soil organic carbon. A larger estimate for carbon uptake in the biosphere (Tans, Fung, and Takahashi, 1990) must be compensated by a reduced uptake in the ocean to arrive at a carbon balance. To do this, either the exchange rate between the upper mixed ocean layer and deep sea, or between ocean surface and atmosphere, should be reduced. In addition, a good match to the observed time-course of 14C carbon in the atmosphere must be preserved by the model. The 14C time-course did not remain well-matched if the atmosphere—ocean surface exchange was reduced, but it was hardly distrubed at all if the exchange rate with the deep sea was reduced.