Africa and the global carbon cycle

Abstract

The African continent has a large and growing role in the global carbon cycle, with potentially important climate change implications. However, the sparse observation network in and around the African continent means that Africa is one of the weakest links in our understanding of the global carbon cycle. Here, we combine data from regional and global inventories as well as forward and inverse model analyses to appraise what is known about Africa's continental-scale carbon dynamics. With low fossil emissions and productivity that largely compensates respiration, land conversion is Africa's primary net carbon release, much of it through burning of forests. Savanna fire emissions, though large, represent a short-term source that is offset by ensuing regrowth. While current data suggest a near zero decadal-scale carbon balance, interannual climate fluctuations (especially drought) induce sizeable variability in net ecosystem productivity and savanna fire emissions such that Africa is a major source of interannual variability in global atmospheric CO2. Considering the continent's sizeable carbon stocks, their seemingly high vulnerability to anticipated climate and land use change, as well as growing populations and industrialization, Africa's carbon emissions and their interannual variability are likely to undergo substantial increases through the 21st century.

Figure 1

Latitudinal distribution of mean annual precipitation [72], soil [78] and plant [84] carbon density, annual net primary production (NPP) per unit ground area from CASA [23, 26] and the Potsdam 17-model intercomparison [89], total soil and live carbon, total annual NPP, shown with the spatial distribution of land cover [106] (colors). Means and totals were calculated from published data using all terrestrial locations in 5° latitude zones.

Figure 2

The African carbon cycle. Annual fluxes and pools (shown in parentheses) all in units of 10¹⁵ g C, where NPP is net primary production, and Rh is heterotrophic respiration. Estimates as reported in Table 1

Figure 3

Standard deviation of net ecosystem carbon exchange (NEE) estimated with three ecosystem models, High Resolution Biosphere Model (HRBM), Terrestrial Ecosystem Model (TEM), and Lund-Potsdam-Jena model (LPJ) as reported by McGuire et al. [24].

Figure 4

Terrestrial carbon source and sink estimates for Africa, tropical and global lands. (a) Net carbon dioxide flux totals, and (b) net carbon dioxide flux per unit area. Positive values indicate a surface source. Boxes show the range of +/- 1 standard deviation from the IPCC report [6] for global and tropical land during the 1980s (dark) and 1990s (light), whereas symbols report results from inverse analyses cited in Appendix I. Triangles and error bars indicate mean flux estimates from individual inversion studies and associated posterior uncertainties. Squares indicate the average, and pluses indicate the standard deviation, of mean flux estimates from a group of inverse solutions. Circles indicate the average uncertainty estimates among the group of inverse solutions. Atmospheric inversion results for Africa are taken from Bousquet [18] (B99), Ciais [107] (C00), Rödenbeck [51] (R03), Gurney [17] (G02), and Gurney [19] (G04), with years spanned in each analysis shown below literature source abbreviations of (a).

Figure 5

Root mean square of annual net carbon flux obtained from time-dependent inverse solutions [20] for the period 1990 – 2001 and for regions based on TransCom [20, 108]. NH (Northern Hemisphere land) includes temperate and boreal Asia, temperate and boreal North America, and Europe, SH (Southern Hemisphere land) includes temperate South America, Australia and New Zealand, Tropical includes tropical America, Africa and tropical Asia, and Amer. abbreviates America.