Quantifying photosynthetic productivity from chlorophyll fluorescence: combining observations and modelling to improve understanding of the global carbon cycle

© 2018 Dr. Alexander John NortonPhotosynthesis is a fundamental ecological process of the Earth system. The amount of carbon that is ‘fixed’ by terrestrial plants during photosynthesis, termed gross primary production (GPP), constitutes the largest surface-to-atmosphere flux of carbon on Earth. This cellular scale process has global scale effects not only on the carbon cycle but through the regulation of water and surface energy balance whilst providing the primary energy source for natural and agricultural ecosystems. From regional to global scales estimates of GPP are highly uncertain. This thesis contributes to solving this challenge through observational and modelling studies. The focus is on using chlorophyll fluorescence, an observable signature of photosynthetic activity and a process that is intimately tied in with its biochemistry. This work spans spatial scales, providing insights from individual leaves to the global terrestrial biosphere. The underlying theme is to see how this indirect signature of photosynthesis can be used to interpret changes in GPP. The overarching aim is to develop a way to quantify global GPP using chlorophyll fluorescence data. Starting at the leaf scale this thesis investigates photosynthesis and energy dissipation in Eucalyptus leaves. Field data was collected in a mature forest canopy in south-eastern Australia. This included observations of the light-dependent reactions and light-independent reactions of photosynthesis using gas exchange and relative variations in chlorophyll fluorescence. This data is used to evaluate a leaf biochemical model. The model provides a reasonable representation of the observations although some weaknesses in the process formulation are identified. It is shown that the way regulatory feedback mechanisms and energy dissipation are formulated in the model substantially alters leaf scale chlorophyll fluorescence and its relationship to photosynthetic capacity. Model simulations suggest this sensitivity also propagates up to the canopy scale. This is integral to interpreting chlorophyll fluorescence observations and relating them to photosynthetic capacity at leaf and canopy scales. The thesis then investigates patterns of GPP over Australia using satellite observations of chlorophyll fluorescence alongside other satellite and ground-based observations. It is shown that temperate ecosystems around the east and south-east coasts have higher productivity on average while semi-arid ecosystems across the interior of the continent dominate the variability in productivity. The data also shows localised variability hot spots over ephemeral braided floodplains in the Lake Eyre basin, suggesting that during wet years the supply of water overland supports primary production. At the global scale, this thesis develops a framework to combine observations of chlorophyll with a terrestrial biosphere model. This applies Bayesian inference theory to improve prior knowledge on uncertain model parameters using chlorophyll fluorescence observations and propagate this improvement to model estimates of GPP, in essence an optimisation. To achieve this, a process-based model for chlorophyll fluorescence is implemented in the terrestrial biosphere model which can then be compared directly with observed chlorophyll fluorescence. This model provides a process-based relationship between chlorophyll fluorescence and GPP based on current theory which can account for non-linear effects such as non-photochemical quenching, chlorophyll content and canopy structure. Following the optimisation using satellite chlorophyll fluorescence observations global GPP is estimated for 2015 to be 137 ± 6 billion tonnes of carbon. This is an increase of about 7% on the models prior estimate. The increase is driven by increased photosynthetic light-use efficiency and reduced sunlight absorption. The spatial distribution of this GPP estimate provides a better match with independent estimates that are upscaled from site-based observations. It is also shown that using one year of chlorophyll fluorescence data can reduce the uncertainty in GPP that stems from uncertain model parameters by about 70%. However, with the current model theory there is limited information from chlorophyll fluorescence data on photosynthetic capacity, a key model parameter. Improvements are still needed in how vegetation cover (leaf area index) is modelled. Additionally, the model is limited in its ability to match high chlorophyll fluorescence values. Opportunities for improvements are discussed. As atmospheric carbon dioxide levels continue to rise there is an increased need to understand and quantify terrestrial carbon fluxes. This thesis provides an insight into how chlorophyll fluorescence observations can be used across spatial scales to improve knowledge on the photosynthetic carbon flux