Predicting forest dynamics and biomass production efficiency based on optimality principles

Carbon (C) allocation refers to the processes by which plants distribute assimilated C among growth, storage, and respiration. While most ecosystem and land surface models explicitly represent C allocation, its treatment in many models remains rudimentary, reflecting a lack of consensus and limiting their ability to capture the processes governing C partitioning. A long-standing theory explains C allocation as maximizing growth, with foliage and below-ground investments balancing light, water and nutrients availability. However, the large C investment in tree stems does not contribute to primary production but reflects an evolutionary strategy to maximize light capture and competitive ability. Biomass production efficiency (BPE) quantifies the efficiency of assimilated C that is converted into structural growth. It reflects the balance between C gain by photosynthesis and C losses, principally autotrophic respiration (R_a). However, the controls on BPE remain poorly constrained, and even the sign of its response to growth temperature is unclear. Here we develop robust semi-empirical models of C allocation of forest dynamics, maximum tree height (H_m) and BPE in order to explore how C partitioning is influenced by the availability of different resources. We hypothesize that the demands of foliage production, and concomitant below-ground production to support that foliage, are satisfied with highest priority; and that any excess C (the net C profit, P_n) is allocated to stems in such a way as to maximize height growth, as a strategy for competitive fitness. Under this framework, the average diameter growth of a tree, and H_m, in an even-aged forest are shown to be proportional to P_n. We further show that BPE is shown to decrease with growth temperature (T_g), stand age, soil C:N ratio, pH and sand content, while increasing with mean temperature of the coldest month—resolving a contradiction in the literature, about its apparent response to mean annual temperature—and to be greater for deciduous than evergreen woody plants. These findings contribute to an optimality-based theoretical framework for improved process-based C allocation modelling in forest ecosystem models.