ABSTRACT
A model is presented which solves simultaneously for leaf-scale stomatal conductance, CO2 assimilation and the energy balance as a function of leaf position within canopies of well-watered vegetation. Fluxes and conductances were calculated separately for sunlit and shaded leaves. A linear dependence of photosynthetic capacity on leaf nitrogen content was assumed, while leaf nitrogen content and light intensity were assumed to decrease exponentially within canopies. Separate extinction coefficients were used for diffuse and direct beam radiation. An efficient Gaussian integration technique was used to compute fluxes and mean conductances for the canopy. The multilayer model synthesizes current knowledge of radiation penetration, leaf physiology and the physics of evaporation and provides insights into the response of whole canopies to multiple, interacting factors. The model was also used to explore sources of variation in the slopes of two simple parametric models (nitrogen- and light-use efficiency), and to set bounds on the magnitudes of the parameters.
For canopies low in total N, daily assimilation rates are ∼10% lower when leaf N is distributed uniformly than when the same total N is distributed according to the exponentially decreasing profile of absorbed radiation. However, gains are negligible for plants with high N concentrations. Canopy conductance, Gc should be calculated as Gc=Aσ(fslgsl+fshgsh), where Δ is leaf area index, fsi and fsh are the fractions of sunlit and shaded leaves at each level, and gsi and gsh are the corresponding stomatal conductances. Simple addition of conductances without this weighting causes errors in transpiration calculated using the ‘big-leaf’ version of the Penman-Monteith equation. Partitioning of available energy between sensible and latent heat is very responsive to the parameter describing the sensitivity of stomata to the atmospheric humidity deficit. This parameter also affects canopy conductance, but has a relatively small impact on canopy assimilation.
Simple parametric models are useful for extrapolating understanding from small to large scales, but the complexity of real ecosystems is thus subsumed in unexplained variations in parameter values. Simulations with the multilayer model show that both nitrogen- and radiation-use efficiencies depend on plant nutritional status and the diffuse component of incident radiation, causing a 2- to 3-fold variation in these efficiencies.