Simulated effects of canopy structural complexity on forest productivity

Late-successional unmanaged forests (or old-growth forests) are important terrestrial carbon sinks. Their structurally complex features have been hypothesized to strongly affect forest productivity. However, the mechanisms through which structural complexity influences forest productivity remain unresolved. Here, we report a set of idealised simulations in boreal, sub-boreal, cool-temperate, warm-temperate and tropical forests using a dynamic model of atmosphere and vegetation interactions (Multilayered Integrated Numerical Model of Surface Physics-Growing Plants Interaction, MINoSGI). We aimed to elucidate the effects of tree-size and canopy structural complexity on stand-scale gross primary productivity (GPP) of old-growth forests over 300-year time series. We assumed the tree-size structures of mature old-growth forests as the initial conditions of the simulations. We focused on three different developmental phases of the forests: 29-48 years (phase I), 77-96 years (phase II) and 277-296 years (phase III) from the start of the simulations. Phase I corresponds to an old-growth forest because of our settings of the initial conditions of the simulations, and phase III to a long-term resultant forest without recruitment or mortality. We show that, for a given biome-specific leaf area index (LAI) and individual foliage shape (eta), the greatest difference in GPP (Delta GPP) between forests with the most and least structurally complex canopies was 263 g C/m(-2) year(-1) in phase I and II when old-growth forests with the most complex canopy structures sustained their complexity, equivalent to almost 13 % of the annual GPP. By contrast, the maximum difference in Delta GPP was reduced to 153 g C/m(-2) year(-1) in phase III, equivalent to 8 % of the annual GPP total because the complex canopy structures of old-growth forests without recruitment or mortality approached less multi-layered ones similar to those of less complex and managed forests. Our simulation results indicate that greater tree-size and canopy structural complexity is associated with a broader vertical distribution of foliage, supporting the localisation of leaves in multiple layers. This increases the efficacy of light usage, particularly for shaded leaves receiving only diffuse light within the canopy. As a consequence, for all forest biomes, an increase in in-canopy diffuse light increases GPP in forests with complex canopy structures due to an enhancement of light use efficiency (LUE) of shaded leaves. Overall, our results quantify one of the mechanisms underlying the effects of canopy structural complexity on forest productivity. Moreover, differences in foliage shape between forest biomes can add to uncertainties when predicting GPP based on simulation models. We, therefore, suggest that precise values of the biome-specific factors associated with canopy structure should be employed into simulation models of global carbon budgets.