Neighborhood structure influences the convergence in light capture efficiency and carbon gain: an architectural approach for cloud forest shrubs

Although plant competition is recognized as a fundamental factor that limits survival and species coexistence, its relative importance on light capture efficiency and carbon gain is not well understood. Here, we propose a new framework to explain the effects of neighborhood structures and light availability on plant attributes and their effect on plant performance in two understory shade-tolerant species (Palicourea padifolia (Roem. & Schult.) C.M. Taylor & Lorence and Psychotria elata (Swartz)) within two successional stages of a cloud forest in Costa Rica. Features of plant neighborhood physical structure and light availability, estimated by hemispherical photographs, were used to characterize the plant competition. Plant architecture, leaf attributes and gas exchange parameters extracted from the light-response curve were used as functional plant attributes, while an index of light capture efficiency (silhouette to total area ratio, averaged over all viewing angles, STAR) and carbon gain were used as indicators of plant performance. This framework is based in a partial least square Path model, which suggests that changes in plant performance in both species were affected in two ways: (i) increasing size and decreasing distance of neighbors cause changes in plant architecture (higher crown density and greater leaf dispersion), which contribute to lower STAR and subsequently lower carbon gain; and (ii) reductions in light availability caused by the neighbors also decrease plant carbon gain. The effect of neighbors on STAR and carbon gain were similar for the two forests sites, which were at different stages of succession, suggesting that the architectural changes of the two understory species reflect functional convergence in response to plant competition. Because STAR and carbon gain are variables that depend on multiple plant attributes and environmental characteristics, we suggest that changes in these features can be used as a whole-plant response approach to detect environmental filtering in highly diverse tropical forest communities.