Transpiration as the leak in the carbon factory: a model of self-optimising vegetation

[Truncated abstract] In the most common (hydrological) viewpoint, vegetation is a "water pump". However, while vegetation does move vast quantities of water, it does not behave like a simple pump - it is rather more like a "carbon factory" where transpiration becomes an inevitable cost in the process of carbon assimilation. In this respect, the trade-off between water loss and carbon gain emerges as a fundamental constraint on vegetation function. While vegetation does experience other trade-offs (e.g. nutrient acquisition vs. nutrient retention in nutrient-poor environments), the link between water and carbon provides a convenient way to assemble a hypothetical optimal vegetation, even without detailed knowledge of the local site. The present study introduces a model of hypothetical optimal vegetation based on the assumption that natural vegetation has co-evolved with its environment over a long period of time and that natural selection has led to a species composition that is most suited for the given environmental conditions. The trade-off between water loss and carbon gain is formulated in terms of the costs associated with the maintenance of roots, water transport tissues and foliage, and the benefits related to the exchange of water for CO2 with the atmosphere, driven by photosynthesis ... This was done for two contrasting months and the results were tested against the observations. In the second stage, water availability was modelled using a physically-based catchment water balance model and vegetation was optimised dynamically to maximise its Net Carbon Profit over a period of 30 years. This included a dynamic optimisation of roots and the foliage properties of trees and grasses. The modelled canopy properties and fluxes matched the available measurements surprisingly well, given that no calibration was undertaken and only coarse climate and catchment data were used. Furthermore, no knowledge about vegetation on the site is needed to run the model. Finally, the model was applied to a variety of catchments and vegetation types around Australia, which span a wide range of climate properties, ranging from arid to humid. The potential effect of atmospheric increase in CO2 concentrations was then investigated by optimising the vegetation for different historic levels of atmospheric CO2, which has led to interesting insights and surprising outcomes. These results represent a significant advance in our ability to model or predict evapo-transpiration rates in catchments with natural vegetation, in that hydrologist can treat vegetation response in a more realistic way in their models than has been possible in the past.