A fundamental goal of the science of the carbon-climate-human system is to understand and eventually reduce the Earth's vulnerability to dangerous changes in climate. This agenda requires that we understand the mechanisms that drive climate change, develop strategies for minimizing the magnitude of the climate change that does occur, and create approaches for coping with the climate change that cannot be avoided. Successful pursuit of this agenda is simpler when the carbon-climate-human system generates negative feedbacks (that tend to suppress further climate change), and it is more complicated when the system generates positive feedbacks (Figure 1.1). Positive feedbacks are especially challenging if they occur suddenly, as threshold phenomena, or if they involve coupled responses of the atmosphere, land, oceans, and human activities.
We are entering an era when we need not—and in fact must not—view the ques tion of vulnerability from any single perspective. The carbon-climate-human system generates climate change as an integrated system. Attempts to understand the integrated system must take an integrated perspective. Mechanistic process models, the principal tools for exploring the behavior of climate and the carbon cycle on land and in the oceans, are increasingly competent to address questions about interactions among major components of the system (Gruber et al., Chapter 3). Still, many of the key interactions are only beginning to appear in models or are not yet represented. For these interactions, we need a combination of dedicated research and other tools for taking advantage of the available knowledge. In assessing the vulnerability of the carbon cycle to the possibility of large releases in the future, we combine results from mechanistic simulations with a broad range of other kinds of information.
Several new lines of information suggest that past assessments have underestimated the vulnerability of key aspects of the carbon-climate-human system. Several of these concern climate-carbon feedbacks. Simulations with coupled climate-carbon models demonstrate a previously undocumented positive feedback between warming and the terrestrial carbon cycle, in which CO2 releases that are stimulated by warming accelerate warming and further CO2 releases (Friedlingstein, Chapter 10, this volume). The experiments to date are too limited to support an accurate quantification of this positive feedback, but the range of results highlights the importance of further research. The behavior of two models of comparable sophistication is so different that, with similar forcing, they differ in atmospheric CO2 in 2100 by more than 200 parts per million (ppm).
The models that simulate the future carbon balance of land are still incomplete. At least three mechanisms either not yet represented or represented in the models in a rudimentary way have the potential to amplify positive feedbacks to climate warming (Gruber et al., Chapter 3). The first of these is the respiration of carbon currently locked in permanently frozen soils. General Circulation Model (GCM) simulations indicate that much of the permafrost in the Northern Hemisphere may disappear over the next century. Because these soils contain large quantities of carbon (Michaelson et al. 1996), and because much of this carbon is relatively labile once thawed, potential releases over a century could be in the range of 100 PgC (Gruber et al., Chapter 3). Wetland soils are similar, containing vast quantities of carbon, which is subject to rapid decomposition when dry and aerated. Drying can allow wildfires, such as those that released an estimated additional 0.8 to 3.7 PgC from tropical fires during the 1997-1998 El Niño (Langenfelds et al. 2002). Drying wetland soils might result in a decrease in methane emissions, along with an increase in CO2 emissions, requiring a careful analysis of overall greenhouse forcing (Manne and Richels, Chapter 25). A third aspect of the terrestrial biosphere with the potential for massive carbon releases in the future is large-scale wildfire, especially in tropical and boreal forest ecosystems (Gruber et al., Chapter 3). Climate changes in both kinds of ecosystems could push large areas past a threshold where they are dry enough to support large wildfires (Nepstad et al. 2001), and a fundamental change in the fire regime could effectively eliminate large areas of forest. None of these three mechanisms is thoroughly addressed in current ecosystem or carbon-cycle models. As a consequence, it is not yet feasible to estimate either the probability of the changes or the likely carbon emissions. Still, ignoring the potential for these large releases is not responsible, and the vulnerability of the climate system to them should be explored.
Vulnerability of ecosystems used for carbon management highlights other aspects of the need for an integrated perspective on the carbon-climate-human system. Ocean fertilization and deep disposal both create altered conditions for ocean ecosystems (Bakker, Chapter 26; Brewer, Chapter 27). To date, the consequences of these alterations are poorly known. Ecosystem alteration is also an issue for terrestrial sequestration through afforestation. Especially where afforestation involves plantations of a single tree species or non-native species, it is important to assess how any extra vulnerability to loss of ecosystem services alters the overall balance of costs and benefits (Raupach et al., Chapter 6).
Was this article helpful?