Our main recommendation is to focus research efforts on understanding and better simulating the response of the carbon cycle, both on land and on marine energy, on CO2. It is the main driver of natural carbon sinks, the greatest source of uncertainty in the future fraction of air, and contributes much more than γ to uncertainty in the TCRE. If we want to reduce uncertainty in future carbon budgets, that must be our main concern. While the country contributes the most to the diffusion of models, ocean wells also contribute to the carbon sink and present significant regional uncertainties (Hewitt et al 2016) and are becoming increasingly important on time scales beyond 2100 (Randerson et al 2015, Jones et al 2016b). Here, we conduct a new analysis of three generations of modeling results for the Earth system over a decade to examine whether existing simulations and analyses are well positioned to meet the growing demands of policy makers on the carbon cycle research community. In Section 2, we present a new analytical framework that allows us to quantify the sources of uncertainty in carbon budgets in order to increase the response to CO2 or the terrestrial or oceanic climate. Analysis of the results in Section 3 shows that it is the response of the carbon cycle to CO2, not its response to climate, that dominates uncertainty in TCRE households and hence carbon budgets. We conclude with recommendations for the carbon cycle research community. These similarities between the carbon cycle and climate feedback formalisms have led to early analyses to focus on the positive reactions of the system.
These were analogies with climate feedback and were often considered the greatest source of uncertainty (Matthews et al 2005, Raddatz et al 2007). Even before the first generation of C4MIP modelling results (Friedlingstein et al.2003) presented an uncertainty analysis to show that the co2 cycle response to climate (γ) contributed to most of the differences between the early Hadley Centre and the IPSL carbon cycle feedback experiments (Cox et al 2000, Friedlingstein et al 2001). While this remains true, γ contributes to most of the uncertainty in the carbon climate cycle, but overlooks the underlying uncertainty in the unchanged wells themselves. Lenton, T.M. et al. Millennial Timescale carbon cycle and climate change in an efficient terrestrial system model. Clim. Dynam 26, 687-711 (2006). Addressing the key uncertainties in the carbon cycle is essential to provide greater clarity on the mitigation measures needed to achieve the Paris Agreement`s goal of limiting global warming to "a level well below 2 degrees Celsius" and continuing efforts at 1.5 degrees Celsius to reduce the effects of dangerous climate change. There is therefore an urgent need to better understand and model the processes that fuel observed atmospheric CO2 variability in seasonal to secular scales, in order to improve climate forecasts and inform on climate prevention and adaptation. They also showed that the return gain of the carbon cycle, g, could be expressed in the form of these quantities: although carbon-carbon cycle co2 and climate-cycle feedbacks have been identified for more than a decade, we still have limited capacity to quantify these returns and confidently attribute past changes to the carbon cycle and thus anticipate its future evolution.
There is therefore an urgent need to better understand and model the processes that fuel observed atmospheric CO2 variability at seasonal to secular scales, in order to improve climate forecasts and inform about climate and adaptation. What does this analysis mean for projections for the 21st century? To date, almost all climate-carbon cycle analyses have focused on high monotonous CO2 scenarios: from IS92a (Cox et al 2000) to SRES-A2 (Friedlingstein et al 2006) to idealized RCP8.5 and 1% CMIP5 experiments. In these scenarios, CO2 emissions will increase rapidly and continuously from 700 to 1100 ppm by 2100 (Figure 6).