Predictability of the terrestrial carbon cycle

Background/Question/Methods: Modern civilization is largely driven by fossil fuel energy. Fossil fuel burning releases carbon dioxide (CO₂): the main driver of anthropogenic climate change. Given the potential consequences of climate change for human civilization, there has been major research investment to understand and predict how carbon cycles throughout the biosphere, atmosphere and oceans. It is an imperative responsibility of scientists to deliver credible estimates of carbon sequestration in land and oceans.  Terrestrial ecosystems play a crucial role in the global carbon cycle and the regulation of climate change. In the past 50 years, terrestrial ecosystems absorbed nearly 30% of those emissions¹¹. This magnitude of the terrestrial carbon sink has been deduced indirectly: combining analyses of atmospheric carbon dioxide concentrations with ocean observations and working out what the net terrestrial carbon flux must be. When knowledge about the terrestrial carbon cycle itself is integrated into land models, they generate uncertainty that is usually too large to effectively constrain global land carbon sink. Although many research programs are underway to improve understanding of the terrestrial carbon cycle through further observations, experiments, and modeling.  They have shown limited success in reducing uncertainty in model predictions.  We need to explore alternative approaches to improve our predictive understanding of in the terrestrial carbon cycle.

Results/Conclusions: In this talk, we explore the issue of predictability of the terrestrial carbon cycle. We asked a question: To what extent is the terrestrial carbon cycle intrinsically predictable? If intrinsic predictability is low, further research should make limited improvements to the accuracy of future projections despite increasing our understanding. However if intrinsic predictability is high, then why do the projections from state of the art models continue to differ so widely? We first examined fundamental properties of the terrestrial carbon cycle, which form a basis upon which its intrinsic predictability can be analyzed. We show that there is abundant empirical and modeling evidence that the terrestrial carbon cycle is actually highly predictable. We discussed the predictability under five types of external forcing, which encompass almost all possible scenarios the terrestrial ecosystems may experience in the Earth system. With the perspective of predictability, we identified major knowledge gaps in carbon cycle research and proposed strategies to improve the predictive ability of the carbon cycle models.