Niche versus neutrality: a dynamical analysis.

Understanding the forces shaping ecological communities is of crucial importance for basic science and conservation. After 50 years in which ecological theory has focused on either stable communities driven by niche-based forces or nonstable "neutral" communities driven by demographic stochasticity, contemporary theories suggest that ecological communities are driven by the simultaneous effects of both types of mechanisms. Here we examine this paradigm using the longest available records for the dynamics of tropical trees and breeding birds. Applying a macroecological approach and fluctuation analysis techniques borrowed from statistical physics, we show that both stabilizing mechanisms and demographic stochasticity fail to play a dominant role in shaping assemblages over time. Rather, community dynamics in these two very different systems is predominantly driven by environmental stochasticity. Clearly, the current melding of niche and neutral theories cannot account for such dynamics. Our results highlight the need for a new theory of community dynamics integrating environmental stochasticity with weak stabilizing forces and suggest that such theory may better describe the dynamics of ecological communities than current neutral theories, deterministic niche-based theories, or recent hybrids.

Temporal fluctuation scaling in populations and communities.

Taylor's law, one of the most widely accepted generalizations in ecology, states that the variance of a population abundance time series scales as a power law of its mean. Here we reexamine this law and the empirical evidence presented in support of it. Specifically, we show that the exponent generally depends on the length of the time series, and its value reflects the combined effect of many underlying mechanisms. Moreover, sampling errors alone, when presented on a double logarithmic scale, are sufficient to produce an apparent power law. This raises questions regarding the usefulness of Taylor's law for understanding ecological processes. As an alternative approach, we focus on short-term fluctuations and derive a generic null model for the variance-to-mean ratio in population time series from a demographic model that incorporates the combined effects of demographic and environmental stochasticity. After comparing the predictions of the proposed null model with the fluctuations observed in empirical data sets, we suggest an alternative expression for fluctuation scaling in population time series. Analyzing population fluctuations as we have proposed here may provide new applied (e.g., estimation of species persistence times) and theoretical (e.g., the neutral theory of biodiversity) insights that can be derived from more generally available short-term monitoring data.