Modelling carbon uptake of Australian evergreen ecosystems under rising CO2 concentration and water limitation

Abstract

The current rise in the atmospheric CO2 concentration (Ca) provides both challenges and opportunities to terrestrial plant communities. Higher Ca provides a benefit to plants by allowing them to achieve higher photosynthetic rates at lower stomatal conductance (gs). On the other hand, the negative impact of rising Ca is global warming. Rising temperatures directly affect plants but also increase the dryness (vapour pressure deficit, D) of the air. Higher D could reduce gs and thus photosynthesis, leading to a loss of plant fitness. Terrestrial vegetation models can be used to quantify the combined impact of these environmental changes but need to be evaluated for their performance against observations. This thesis focuses on evaluating Ca responses of Australian ecosystems, which feature evergreen trees adapted to frequent water deficits. In the following chapters, I focus on three major components of terrestrial vegetation models: leaf area index (LAI); the response of gs to D; and the response of gs and photosynthesis to elevated Ca. These three components are particularly important for the modelling of rising Ca because the leaf scale response is captured by the responses of gs and photosynthesis to water deficit and Ca, while LAI is particularly important for the up-scaling of leaf level carbon and water fluxes to the whole ecosystem. Improvements in these components are thus likely to reduce the uncertainties in current terrestrial vegetation models. In Chapter 2, I test the concept of ecohydrological equilibrium for its ability to predict key traits of Australian evergreen ecosystems. This theory posits that long-term equilibrium LAI (Lequ) is determined by water availability. The predicted LAI values and the response to Ca both compared well to those of satellite-derived data. These results indicate that Lequ could be a useful alternative to satellite-derived data to terrestrial vegetation models to guide foliage carbon allocation. In the second research chapter (Chapter 3), I compared existing gs models and commonly used assumptions (i.e., hydraulic and non-stomatal limitations) for their ability to predict leaf and canopy-scale carbon and water fluxes under high D. I found that incorporating an empirical non-stomatal limitation of apparent photosynthetic capacity with increasing D improved model performance against data and outperformed models incorporating hydraulic limitation . The results suggest that future models should consider non-stomatal limitations to photosynthesis, especially in high-D environments. The Chapter 4 of this thesis aimed to determine the gross primary productivity (GPP) under ambient and elevated Ca (+38%; 150 µmol mol-1) at the Eucalyptus Free Air CO2 Enrichment (EucFACE) experiment. I parameterised the process-based model, MAESTRA, with a suite of in situ measurements of canopy structure and plant physiology shared with me by the EucFACE scientific community. I also conducted an attribution analysis to explore the determinants of the response of GPP to elevated Ca. My findings indicate a relatively small elevated Ca response of GPP (+8%) in the evergreen woodland. My results are key to understanding the response of this ecosystem to elevated Ca. In summary, the findings from this thesis provide some key insights into current gaps in the modelling of terrestrial vegetation. The results show viable options to improve the leaf gas exchange and LAI submodels that are used by terrestrial vegetation models. Overall, this thesis suggests ways for future terrestrial vegetation models to address these gaps for more realistic predictions under changing climate and rising Ca.

Date of Award	2018
Original language	English