Unraveling Amazon tree community assembly using Maximum Information Entropy: a quantitative analysis of tropical forest ecology

Abstract

Drivers of species distributions and their predictions have been a long-standing search in ecology, with approaches varying from deterministic to neutral (i.e. stochastic) and almost everything in between (e.g. near neutral, continuum or emergent-neutral1,2 ). Most models are based on prior assumptions of processes that drive community dynamics. Te Maximum Entropy Formalism (hereafer called MEF) makes no such, potentially unjustifed, a-priori assumptions in generating predictions of species abundance distributions, as such it is a use ful construct to infer processes driving community dynamics given the constraints imposed by prior knowledge (e.g. functional traits or summed regional abundances)3 . Quantifying the relative importance of these distinct constraints can thus provide additional answers to understand the complexity of community dynamics (see Supporting Materials SM: boxes S1–S3). Tis is especially so because, although many diferent tests are available that link variation in taxon abundances to (1) trait variation, (2) taxon turnover between habitats or environ ments and (3) the distance decay of similarities between samples, none quantify the importance of these relative to each other. Te MEF as applied here, however, is capable of and designed to do exactly this by decomposing variation to separate information explained by each of these aspects in a four-step model (Fig. 1 and Box S2). Its application to an unprecedented large tree inventory database on genus level taxonomy consisting of>2,000 1-ha plots distributed over Amazonia4 and a genus trait database of 13 key functional traits representing global axes of plant strategies5 allows us to advance the study of Amazonian tree community dynamics from a new cross-disciplinary perspective.