Investigating the role of water availability on the physiology, leaf growth and productivity of a temperate grassland under elevated carbon dioxide

Grassland ecosystems are sensitive to both the amount and temporal distribution of precipitation. Predicting how grasslands will respond to projected changes in precipitation and rising atmospheric concentration of carbon dioxide ([CO2]) remains a key uncertainty. The focus of this thesis is how the timing and quantity of precipitation affects the physiological and growth responses of perennial ryegrass (Lolium perenne), a significant agricultural species, to Free-Air CO2 Enrichment (FACE), using the TasFACE2 experiment in SE Tasmania, Australia. A randomised split plot design was utilised, comprising two [CO2] treatments (ambient and elevated to 550 μmol mol-1 ) with each plot split into five equal sectors, facilitating the application of five watering treatments. I executed two simultaneous experiments. First, I evaluated the influence of irrigation frequency on the CO2-response using three irrigation intervals in which sectors were watered to field capacity every three, five or ten days. The second experiment was designed to assess the effects of increased spring precipitation; here, the three-day irrigation treatment was used as the control, receiving water up to field capacity, while the remaining two sectors received 25% and 50% increases in water during spring (i.e. 1 September to 31 November). Findings of the experiments demonstrate that irrigation frequency affected both the physiological and growth impacts of e[CO2]. I found that immediately after irrigation, mean net photosynthetic rate was 21.2% higher under e[CO2] compared to ambient plots. Photosynthetic rate remained high relative to ambient plots as days since water increased, reaching 31.3% higher under e[CO2], suggesting that the photosynthetic benefit increased substantially with decreasing frequency of watering. Mean stomatal conductance was significantly reduced in ryegrass grown under e[CO2] immediately after irrigation compared with ambient plots. However, the degree to which stomatal conductance was suppressed under e[CO2] declined as days since last irrigation event increased, and 7 – 10 days after an irrigation event there was no difference in stomatal conductance between [CO2]treatments. These results indicate the importance of timing of water supply, either through irrigation or precipitation, in the potential for the carbon fertilisation effect, with implications for understanding future rainfall changes.I found that e[CO2] enhanced belowground biomass production, potentially altering the resilience of perennial ryegrass to water availability. I also found that the irrigation treatments employed to alter temporal and seasonal water availability influenced growth responses of perennial ryegrass under e[CO2]. I demonstrated that SWC variability plays a role in driving brief changes in biomass, however water savings strategies, or a lack of water deficit, allowed fairly consistent biomass allocation across irrigation treatments. The changes in soil moisture influenced the carbon fertilisation effect (CFE) on total cumulative biomass, with plots irrigated every five days experiencing the highest CFE. Importantly, more frequent and less frequent irrigation regimes caused a decline in CFE, opposing the carbon centric view that assumes higher CFE under drier conditions. I also showed that increased precipitation in spring did not enhance growth at this site under e[CO2], either because the site was not equipped to utilise the extra water or precipitation in seasons other than spring contributed to the site’s ability to respond to e[CO2]. Additionally, I investigated potential legacy effects of previous irrigation patterns by altering the irrigation treatments in the second year such that all sectors received only 50% of the water required to bring the soil to field capacity at a frequency of once every 10 days. My results show that a legacy of precipitation frequency influenced the [CO2]-impact on aboveground biomass production but did not noticeably affect belowground or total biomass production. There were no legacy effects of increased spring water on the [CO2]-effect on any biomass measurements. However, a strong positive correlation was found between last year's total biomass production and this year's total biomass production in the study of precipitation frequency legacy for both ambient and elevated [CO2]. Under control plots, a strong positive correlation was found between last year's and this year's aboveground biomass production, but there was no discernible relationship in FACE plots indicating that e[CO2] eliminated the influence of previous years aboveground biomass production. Similar patterns were seen in the study of the increased spring water legacy, with last year's aboveground biomass production only influencing this year's aboveground biomass production in control plots. However, there was no correlation between last year's belowground or total biomass production and this year's corresponding biomass production. This indicates that the legacy of increased spring water eliminated the impact of last year's biomass on this year's biomass production in both control and FACE plots. These observations add to our understanding of how biomass production responds to fluctuations in precipitation in the previous year under e[CO2]. Lastly, I investigated leaf elongation rates and maximum leaf length as a function of altered temporal and seasonal precipitation. I found that e[CO2] caused reductions in both of these metrics in some months, while other months showed no differences among [CO2] treatments. I found that despite some soil moisture variability, daily leaf elongation rates were maintained in less frequently watered treatments. A 50% reduction in irrigation did not influence leaf elongation rates or maximum leaf length, thus perennial ryegrass prioritised aboveground growth above root growth. My findings present novel insights into the interaction between [CO2] and the provision of water, and legacies of previous irrigation patterns, and their effects on growth and physiology. The mechanistic underpinnings of the patterns observed should improve our ability to simulate grassland biomass production now and in the future.