Elevated CO₂ effects on vegetation: informing modelling through meta-analysis and targeted experiments

Atmospheric CO₂ concentration (Cₐ) is currently increasing at an unprecedented rate and this increase has important effects on vegetation. During the last three decades, many experiments examined the response of vegetation to the rising Cₐ. However, the results from these experiments have not been fully incorporated into the global models used by the IPCC in predicting the future course of Cₐ and vegetation dynamics. This is partly due to the fact that experimental data are not often analysed in ways that directly relate to model formulations. Therefore, the main aim of this study was to link experimental data more directly with model formulations. The research involved (a) meta-analysis of experimental data using models as a framework and (b) experimental work explicitly designed to address model predictions. In my thesis, I addressed several aspects of modeling Cₐ responses of vegetation. The first question I addressed was whether there is a temperature effect on plant response to elevated CO₂ (eCₐ). Because of the kinetics of the photosynthetic enzyme Rubisco, theory predicts that the Cₐ response should be greater at higher temperatures. Vegetation models incorporating these physiological responses predict that responses of photosynthesis, and consequently net primary productivity (NPP), to eCₐ should increase with rising temperature, and be larger in warm tropical forests than in cold boreal forests. However experimental data do not always show such an interaction. I used meta-analysis techniques to test whether such an interaction is observed experimentally. Firstly, I tested for an interaction effect on plant growth responses in factorial eCₐ x temperature experiments. This analysis showed a mean interaction effect size of 8.2% (95% CI -0.85% to 18.0 %.) for plant above-ground biomass. Although the interaction was not significantly different from zero, it was also not significantly different from the predicted interaction values obtained from leaf-level and canopy-level models. In the second meta-analysis, I examined eCₐ experiments on woody plants across the globe to test for a relationship between the eCₐ effect and mean annual temperature (MAT). This meta-regression analysis gave a positive slope that was again not significantly different from zero or from the slope predicted by global-scale models. With limited factorial studies and insufficient experimental data in tropical regions, there was a lack of statistical power to determine whether or not a positive interaction exists between eCₐ and temperature. The second question I addressed was how stomatal conductance of C4 plants behaves in response to changing Cₐ. Optimal stomatal theory says that stomata should act to maximize carbon gain (photosynthesis, A) while minimizing water loss (transpiration, E). That is, the optimal stomatal behavior would be to maximise the integrated sum of (A - λE), where λ (mol C mol⁻¹ H₂O) represents the marginal carbon cost of water use. The unified stomatal conductance model by Medlyn et al. (2011) captures stomatal responses for the C3 plants. Since C4 plants have different photosynthetic Cₐ responses, we can expect different stomatal responses from them. By using optimal stomatal theory, I predicted how stomatal conductance of C4 plants should change with eCₐ, and tested experimentally whether C4 plants showed this behavior. The theory predicted that stomata of C4 plants should be more sensitive to increasing Cₐ than C3 plants, however my experimental results showed that C4 plants followed the same stomatal behavior predicted for C3 plants. Optimal stomatal theory also predicts that leaf-level water use efficiency (WUE) of plants should be proportional to Cₐ. However, whole-plant WUE is predicted to be somewhat less responsive than leaf-level WUE due to boundary layer effects on canopy transpiration. In the third chapter of my thesis I tested this prediction using meta-analysis techniques, statistically combining all previously published studies on increased Cₐ effects on leaf-level and whole-plant level WUE. I found that at leaf-level, WUE of both C3 and C4 plants responded in proportion to the increase in Cₐ, but that in C3 plants the change in WUE was due to both changes in assimilation and transpiration whereas in C4 plants the change in WUE was primarily due to the reduction in transpiration. At whole plant level, the WUE response was less than proportional to the Cₐ increase, as predicted. The discrepancy was larger in C4 (only 70 – 79% of the Ca increase) than in C3 plants (80 – 99%). The reduction occurred because whole-plant transpiration was less sensitive to Cₐ than leaf transpiration, whereas whole-plant biomass gains were similar in size to photosynthetic responses. This work informs models by analysing the effects of eCₐ on WUE in terms that can be directly compared against model predictions.