Closing the global budget for CO2

The available scientific knowledge tells us that the two large carbon reservoirs – the terrestrial biosphere and the oceans – have taken up the excess CO2 in approximately equal proportions. However these CO2 “sinks” are not fixed, in fact they are highly variable and respond to elevated atmospheric CO2 levels and changes in the climate.  To account fully for all the CO2 emitted, we need to know the size of the land and ocean CO2 sinks and their evolution in time. And we need to know this with an accuracy that is higher than the uncertainty in CO2 emissions themselves. Full accounting of CO2 emissions is a necessity if we are to monitor the transition from a CO2-intensive to a low CO2 economy. Unaccounted CO2 emissions open the door to all kinds of abuses, from inaccurate accounting and declaration of countries’ emissions to misleading demands for carbon credits associated with Clean Development Mechanisms and geoengineering options.

Can we quantify the global CO2 sinks to such accuracy? To answer this question, we (Le Quéré et al. 2009) have put together a global CO2 budget of all the major sources and sinks of CO2 for every year from 1959 to 2008 (Figure 1).  The aim is to provide information on the year-to-year changes in all aspects of the global CO2 budget, and to identify the drivers of variability and trends. We used economic data to estimate CO2 emissions from fossil-fuel combustion and land-use change. Atmospheric CO2 was measured directly from a network of around 100 stations.  We quantified the evolution of the land and ocean CO2 sinks with the help of models. All models included key processes like plant productivity and respiration on land and circulation, chemical reactions and biological productivity in the ocean. All models were forced by increasing atmospheric CO2 and the meteorological conditions corresponding to the time period of study. To minimise the errors, we used the mean of all models and estimated error using model spread.

Known problems in the CO2 budget explain most of the mismatch.  For instance, in the 1970s, it seems there was more additional CO2 in the system than emissions indicate. We know that La Niña-like conditions led to unusual cool, wet conditions prevailing in the tropics throughout the 1970s causing more carbon to move to the land sink. The models overestimated the land-CO2 uptake in response to these conditions.
Similarly, the massive Mount Pinatubo volcanic eruption in 1991, which affected climate globally by injecting large volumes of small particles into the upper atmosphere, helps explain the missing CO2 in the early 1990s. The land models did not account for the increase in available light in the vegetation canopy from enhanced diffusion during and after the eruption.
Finally, the excess CO2 of the late 1990s appears to be partly a signature of political incentives to clear land in Indonesia that took advantage of the ongoing drought conditions.

Our analysis of the trend in airborne fraction from this global CO2 budget shows a likely positive trend of 0.3 percent per year, with a 90 percent probability the trend is above background variability and additional uncertainty due to poorly quantified land-use CO2 emissions. The models reproduce such a trend and suggest it is a response of the land and ocean sinks to climate variability and climate change for the past 50 years. If the model results are correct in how they represent the processes that reproduced past trends, this supports the existence of a positive feedback between climate and the carbon cycle that was predicted by many carbon-cycle models.

The range of model results is representative of the uncertainty in the known processes. The range of results is smaller than the uncertainty in emissions, which supports the possibility that full accounting of emitted CO2 is possible, even with existing models. However, to reach such a state requires major improvements in models’ year-to-year estimation of CO2 sinks.

The land models could be improved right away by including the known missing processes. The land and ocean models could also be improved further if the mismatch with observations can be constrained not only globally, but also spatially. Regional information is available from direct measurements in the ocean, and from inverse methods that provide information on the regional variability in the CO2 fluxes.
We could achieve further improvements if the sinks could be quantified directly from observations. For the ocean, this may be possible with increased data coverage and improved analysis tools. There is little prospect of estimating the land CO2 sink directly, and thus it will always have to rely on models. Model validation in this context becomes crucial as it ensures the quality of the model estimates.  

Economic slump
In the past decade, CO2 emissions increased at a rate of three percent per year. The emissions are projected to decrease in 2009 in response to the global economic downturn. However this decrease should only bring the global emissions down to their 2007 levels. The key to decreasing global emissions in the long term is to decouple energy use from wealth. With the large reorganisation of the world’s energy system that is required to stabilise CO2 in the atmosphere, the global change in CO2 emissions will be closely scrutinised in the future. There would be enormous benefits to society if the world’s scientific community pulls its expertise together to provide information to account for all the CO2 emitted to the atmosphere. This could assist a peaceful transition to a different economy, but also provide an early warning of how the natural carbon cycle responds to CO2, climate and other environmental changes