Climate dependence of the CO2 fertilization effect on terrestrial net primary production

The quantitative formulation of the fertilization effect of CO2 enrichment on net primary production (NPP) introduced by Keeling and Bacastow in 1970s (known as Keeling's formula) has been recognized as a summary of experimental data and has been used in various assessments of the industrial impact on atmospheric chemistry. Nevertheless, the magnitude of the formula's key coefficient, the so-called growth factor, has remained open to question. Some of the global carbon cycle modelers avoid this question by tuning growth factor and choosing the value that fits the observed course of atmospheric CO2 changes. However, for mapping terrestrial sinks induced by the CO2 fertilization effect one needs a geographical pattern of the growth factor rather than its globally averaged value. The earlier approach to this problem involved formulating the climate dependence of the growth factor and the derivation of its global pattern from climatic variables (whose geographical distribution is known). We use a process-based model (TsuBiMo) for this purpose and derive the values of growth factor for major biomes for comparison our approach with the earlier studies. Contrary to the earlier prevailing opinion, TsuBiMo predicts that these values decrease with mean annual temperature (excluding biomes of limited water supply). We attribute this result to the effect of light limitation caused by mutual shading inside a canopy, which was considered earlier as unimportant, and conclude that current hypotheses about CO2 fertilization effect (and thus projections of the related carbon sink) are very sensitive to the choice of driving forces taken into account.