We present a model that scales from the physiological and structural traits of individual trees competing for light and nitrogen across a gradient of soil nitrogen to their community-level consequences. The model predicts the most competitive (i.e., the evolutionarily stable strategy [ESS]) allocations to foliage, wood, and fine roots for canopy and understory stages of trees growing in old-growth forests. The ESS allocations, revealed as analytical functions of commonly measured physiological parameters, depend not on simple root-shoot relations but rather on diminishing returns of carbon investment that ensure any alternate strategy will underperform an ESS in monoculture because of the competitive environment that the ESS creates. As such, ESS allocations do not maximize nitrogen-limited growth rates in monoculture, highlighting the underappreciated idea that the most competitive strategy is not necessarily the "best," but rather that which creates conditions in which all others are "worse." Data from 152 stands support the model's surprising prediction that the dominant structural trade-off is between fine roots and wood, not foliage, suggesting the "root-shoot" trade-off is more precisely a "root-stem" trade-off for long-lived trees. Assuming other resources are abundant, the model predicts that forests are limited by both nitrogen and light, or nearly so.

Understanding tropical forest dynamics and planning for their sustainable management require efficient, yet accurate, predictions of the joint dynamics of hundreds of tree species. With increasing information on tropical tree life histories, our predictive understanding is no longer limited by species data but by the ability of existing models to make use of it. Using a demographic forest model, we show that the basal area and compositional changes during forest succession in a neotropical forest can be accurately predicted by representing tropical tree diversity (hundreds of species) with only five functional groups spanning two essential trade-offs—the growth-survival and stature-recruitment trade-offs. This data-driven modeling framework substantially improves our ability to predict consequences of anthropogenic impacts on tropical forests.

Abstract. The long-term and large-scale dynamics of ecosystems are in large part determined by the performances of individual plants in competition with one another for light, water, and nutrients. Woody biomass, a pool of carbon (C) larger than 50% of atmospheric CO2, exists because of height-structured competition for light. However, most of the current Earth system models that predict climate change and C cycle feedbacks lack both a mechanistic formulation for height-structured competition for light and an explicit scaling from individual plants to the globe. In this study, we incorporate height-structured competition for light, competition for water, and explicit scaling from individuals to ecosystems into the land model version 3 (LM3) currently used in the Earth system models developed by the Geophysical Fluid Dynamics Laboratory (GFDL). The height-structured formulation is based on the perfect plasticity approximation (PPA), which has been shown to accurately scale from individual-level plant competition for light, water, and nutrients to the dynamics of whole communities. Because of the tractability of the PPA, the coupled LM3-PPA model is able to include a large number of phenomena across a range of spatial and temporal scales and still retain computational tractability, as well as close linkages to mathematically tractable forms of the model. We test a range of predictions against data from temperate broadleaved forests in the northern USA. The results show the model predictions agree with diurnal and annual C fluxes, growth rates of individual trees in the canopy and understory, tree size distributions, and species-level population dynamics during succession. We also show how the competitively optimal allocation strategy – the strategy that can competitively exclude all others – shifts as a function of the atmospheric CO2 concentration. This strategy is referred to as an evolutionarily stable strategy (ESS) in the ecological literature and is typically not the same as a productivity- or growth-maximizing strategy. Model simulations predict that C sinks caused by CO2 fertilization in forests limited by light and water will be down-regulated if allocation tracks changes in the competitive optimum. The implementation of the model in this paper is for temperate broadleaved forest trees, but the formulation of the model is general. It can be expanded to include other growth forms and physiologies simply by altering parameter values.

The dependence of forest productivity and community composition on rainfall is the result of complex interactions at multiple scales, from the physiology of carbon gain and water loss to competition among individuals and species. In an effort to understand the role of these multiscale interactions in the dependence of forest structure on rainfall, we build a tractable model of individual plant competition for water and light. With game-theoretic analyses, we predict the dominant plant allocation strategy, forest productivity, and carbon storage. We find that the amount and timing of rainfall are critical to forest structure. Comparing two forests that differ only in the total time plants spend in water saturation, the model predicts that the wetter forest has fewer fine roots, more leaves, and more woody biomass than the drier forest. In contrast, if two forests differ only in the amount of water available during water limitation, the model predicts that the wetter forest has more fine roots than the drier forest and equivalent leaves and woody biomass. The difference in these responses to increases in water availability has significant implications for potential carbon sinks with rising atmospheric CO2. We predict that enhanced productivity from increased leaf-level water-use efficiency during water limitation will be allocated to fine roots if plants respond competitively, producing only a small and short-lived carbon sink.

Almost all models of plant resource limitation are grounded in either one or both of two simple conceptual models: Liebig's Minimum Hypothesis (LMH), the idea that plants are limited by the resource in shortest supply, and the Multiple Limitation Hypothesis (MLH), the idea that plants should adjust to their environment so that all essential resources are equally limiting. Despite the differences in their predictions, experiments have so far failed to discriminate between them. In a simple factorial nitrogen and water addition experiment in a Minnesota grassland, we observed shifts in allocation that, as in previous studies, are not all explained by a single theory. We found that leaf biomass responded positively to nitrogen additions but did not respond to water additions. We found that fine-root biomass increased in response to water additions, but only at low nitrogen levels, and that fine-root biomass decreased in response to nitrogen additions, but only at high water levels. To understand these responses we built a physiologically based model of plant competition for water, nitrogen, and space to predict plant allocation to fine roots and leaves. Critically, we include in our model the inherent variability of soil moisture and treat light, water, and nitrogen as resources with distinct mechanistic roles. Experimental results showed that plants were nitrogen and water limited. The model explains the experimental results, under conditions of co-limitation, as follows. Foliage increases with nitrogen additions but not water additions because leaf construction is constrained by nitrogen uptake. When water is added, plants spend a larger fraction of the growing season limited by light (and effectively nitrogen) than by water. Thus, water additions cause fine-root biomass to increase because of the increased importance of nitrogen limitation. The response of fine-root biomass to water additions decreases with nitrogen additions because these additions reduce nitrogen limitation. In general, our results are explained by sequential resource limitation. The rate of carbon assimilation may be limited by a single resource at any one moment, but the identity of the limiting resource(s) changes throughout the growing season.