The global atmosphere

Although CO2 is well mixed in the global atmosphere, it is not perfectly mixed, and there are small horizontal gradients in CO2 concentration that are driven by the spatial pattern of sources and sinks for carbon dioxide. For example, CO2 concentrations are higher in the northern hemisphere (the source of most fossil fuel combustion) than in the southern hemisphere. In principle, if the transport of CO2 in the atmosphere can be accurately simulated by atmospheric transport models, it should be possible to use observations of atmospheric CO2 concentrations to derive the spatial pattern of CO2 sources and sinks. As the spatial pattern of CO2 sources from fossil fuel combustion can be accurately predicted from economic data, the remainder will be the spatial pattern of biosphere CO2 sources and sinks (including the effects of land use change). Further information can be derived by also including observations of the stable isotopes of CO2 (biosphere exchange has a different isotopic signature than fossil fuel combustion), and observations of O2 concentrations.

A number of research groups are using this "inverse modelling" approach. The primary requirements for these studies are atmospheric transport models, which developed in the 1980s with the expansion of computing power, and a global network of observation stations. However, most of these observation stations are in North America and Europe, with very little coverage over the oceans and no coverage over tropical land masses. This dearth of observations is the main problem that currently constrains this approach.

A number of research groups around the world are currently applying this atmospheric inversion technique. There is significant disagreement between the results, but some points of agreement. The TRANSCOM3

experiment compared the results from applying 16 different transport models to the same dataset [Gurney et al. (2002)]. The scientists found that results were relatively robust for the northern and southern extratropics, but very poorly constrained in the tropics. There appeared to be a significant carbon sink uniformly distributed across northern land regions, with a total northern land sink of 2.3 ± 0.7PgCyear~1. The study suggests that the tropical lands are probably a net source of carbon, of magnitude +1.0 ± 1.3PgCyear~1, implying that the source from tropical deforestation more than compensates any carbon sink in intact forests. However, some more recent analyses [Rodenbeck et al. (2003)], with more sophisticated use of monthly weather data, estimates tropical carbon balance of —0.8 ± 1.3PgCyear-1, with a high probability of being outside the range of the TRANSCOM values, implying that the tropics are a net carbon sink, and that the intact tropics must therefore be a very strong carbon sink of about 2 Pg C year-1. The uncertainties in this method are clearly still large, and primarily caused by the sparseness of data and problems in modelling the turbulent land-atmosphere interface.

0 0

Post a comment