The rising carbon dioxide concentration of the atmosphere ([CO2]) influences terrestrial ecosystems. Direct physiological responses of photosynthesis and intrinsic water-use efficiency cascade through a hierarchy of ecosystem processes with the potential to increase plant growth and net ecosystem carbon uptake. If these downstream effects of rising [CO2] are substantial at the global scale, the rate of [CO2] increase and thus climate change would be slowed. However, due to the complexity of the processes involved these effects are context-dependent and much of the evidence related to a [CO2]-driven terrestrial carbon sink can appear contradictory. For example, ecosystem-scale Free Air CO2 Enrichment (FACE) experiments show short-term biomass gains, while others show none. Tree-ring studies provide mixed evidence, commonly finding no detectable changes in growth, while forest inventories almost always show increases in biomass. Here we synthesize the current evidence from multiple disciplines—ecosystem manipulation, dendrochronology, forest-inventories, flux-towers, remote sensing, atmospheric measurements, and others—using a relativized response metric, β = log(relative variable change) / log(relative [CO2] change). Evidence is evaluated across a hierarchy of related processes towards establishing a holistic and improved interpretation of historical and future CO2 fertilisation of the global terrestrial carbon sink.
Results/Conclusions
Four integrative processes are key to understanding the CO2 response of the terrestrial carbon sink: gross primary production (GPP), biomass production (BP), vegetation mortality, and litter and soil carbon decomposition. Multiple lines of evidence suggest that GPP increased in proportion to CO2 increase over the 20th century (β = 0.95±0.2 to 1.6±0.9), but the proportion of this trend attributable to CO2 versus other factors of global change is uncertain (a common issue across processes). BP responses to CO2 are mixed, with most historical evidence related to wood BP (-0.45±0.7 to 0.55±0.6) and with elevated CO2 experiments showing a similar range in BP (-0.26±0.6 to 0.49±0.3). Variation in BP responses is linked to nutrient availability and stage of succession. Evidence suggests vegetation mortality is on the rise, but in response to CO2 the sign could be both positive or negative. Soil decomposition rates generally increase in elevated CO2 experiments (0.34±0.2), but litter inputs also increase resulting in little change in soil carbon stocks. Attribution of historical changes to CO2 is a major issue. No experiment can be run at sufficient scale for strong inference of the terrestrial land-surface response, indicating that multiple, independent methods are required to confidently attribute CO2 responses.