Does the growth response of woody plants to elevated CO 2 increase with temperature? A model‐oriented meta‐analysis

The temperature dependence of the reaction kinetics of the Rubisco enzyme implies that, at the level of a chloroplast, the response of photosynthesis to rising atmospheric CO 2 concentration (C a ) will increase with increasing air temperature. Vegetation models incorporating this interaction predict that the response of net primary productivity ( NPP ) to elevated CO 2 (eC a ) will increase with rising temperature and will be substantially larger in warm tropical forests than in cold boreal forests. We tested these model predictions against evidence from eC a experiments by carrying out two meta‐analyses. Firstly, we tested for an interaction effect on growth responses in factorial eC a  × temperature experiments. This analysis showed a positive, but nonsignificant interaction effect (95% CI for above‐ground biomass response = −0.8, 18.0%) between eC a and temperature. Secondly, we tested field‐based eC a experiments on woody plants across the globe for a relationship between the eC a effect on plant biomass and mean annual temperature ( MAT ). This second analysis showed a positive but nonsignificant correlation between the eC a response and MAT . The magnitude of the interactions between CO 2 and temperature found in both meta‐analyses were consistent with model predictions, even though both analyses gave nonsignificant results. Thus, we conclude that it is not possible to distinguish between the competing hypotheses of no interaction vs. an interaction based on Rubisco kinetics from the available experimental database. Experiments in a wider range of temperature zones are required. Until such experimental data are available, model predictions should aim to incorporate uncertainty about this interaction.