CO2 Fertilization: What Models Can Talk to Observations

Of the nearly 10 PgC carbon human beings currently emit into the atmosphere every year, only ~50% stays in the air while the rest is absorbed by the world's oceans and lands. This so-called airborne ratio has stayed surprisingly stable in the past five decades, strongly suggesting that natural carbon sinks strengthening with the increasing atmospheric CO2 concentration. For terrestrial carbon sinks, it implies the fertilization of CO2 on vegetation primary production. However, measurements on-the-ground seem to deliver inconsistent messages regarding the magnitude and the spatio-temporal extent of the fertilization effect, whose identification is likely obscured by concurrent changes in climate and human activities as well as other biological processes (e.g., respiration). In contrast, carefully designed ecosystem-model experiments provide a controlled environment to disentangle the different effects exerted by all the interacting processes. In this study, we use ensemble model results from the Multi-scale Synthesis and Terrestrial Model Intercomparison Project (MsTMIP) to evaluate the effect of CO2 fertilization on global terrestrial gross primary production (GPP). By comparing the model simulations sequentially with controlled climate, atmospheric CO2 concentration, nitrogen deposition, and other factors, we find that a spatially rather homogenous fertilization of CO2 on in the model simulated GPP. However, the net effects of the CO2 fertilization are complicated, in particular, by different ecosystem responses to changes in temperature and precipitation at different geo-locations. As such, the model results coming out from MsTMIP may help in guiding a more reliable evaluation of the CO2 fertilization effect in the observations.