Modelling functional trait acclimation for trees of different height in a forest light gradient: emergent patterns driven by carbon gain maximization.

Forest trees show large changes in functional traits as they develop from a sapling in the shaded understorey to an adult in the light-exposed canopy. The adaptive function of such changes remains poorly understood. The carbon gain hypothesis suggests that these changes should be adaptive (acclimation) and that they serve to maximize net vegetative or reproductive growth. We explore the carbon gain hypothesis using a mechanistic model that combines an above-ground plant structure, a biochemical photosynthesis model and a biophysical stomatal conductance model. Our simulations show how forest trees that maximize their carbon gain increase their total leaf area, sapwood area and leaf photosynthetic capacity with tree height and light intensity. In turn, they show how forest trees increased crown stomatal conductance and transpiration, and how the carbon budget was affected. These responses in functional traits to tree height (and light availability) largely differed from the responses exhibited by exposed trees. Forest and exposed trees nevertheless shared a number of emergent patterns: they showed a similar decrease in the average leaf water potential and intercellular CO(2) concentration with tree height, and kept almost constant values for the ratio of light absorption to electron transport capacity, the ratio of photosynthetic capacity to water supply capacity, and nitrogen partitioning between electron transport and carboxylation. While most of the predicted qualitative responses in individual traits are consistent with field or lab observations, the empirical support for capacity balances is scarce. We conclude that modelling functional trait optimization and carbon gain maximization from underlying physiological processes and trade-offs generates a set of predictions for functional trait acclimation and maintenance of capacity balances of trees of different height in a forest light gradient, but actual tests of the predicted patterns are still scarce.