Structuring of ecological communities

点击翻译按钮获取中文摘要

Ecological communities are presumed to have non-random relationships in characteristics such as species composition. This premise is fundamental to most of our studies involving both basic (e.g., interspecific interactions such as competition) and applied science (restoration ecology), yet it is rarely tested formally. Early work (1970s & 1980s) provided significant new conceptual and methodological approaches to enable us to test whether or not ecological communities differ from random assemblages of species, specifically null models. During this earlier work most studies focused on the role of competition in leading to negative patterns of species co-occurrence when using these null models, yet some studies identified various other factors that may contribute to the same observed patterns (e.g. differential habitat preferences amongst species). Although some progress has been made in advancing some of these methods further, considerable questions remain unanswered or insufficiently addressed. Furthermore, advances in ecology have been hindered by the unique histories of species and sites considered, leading to variable outcomes. In order to broadly advance the generality of our understanding, it is important to consider the average and variability of the outcomes that we see across different groups of species (e.g. fish versus plants) and ecosystems (freshwater versus terrestrial). The proposed work will address several of these issues. Using extensive datasets from the literature plus a combination of field and experimental work, we will quantify and test the relative importance of habitat differences, experimental design and study scale in contributing to finding non-random community structure versus potential interspecific interactions. We will examine the relative roles of environmental disturbance (both natural and anthropogenic origins) in disrupting perceived community structure. Species dispersal has the potential to homogenize species composition across sites, leading to most species occurring everywhere and therefore may play a significant role in differences (or lack of differences) noted in null model studies. Therefore, considering extensive sets of data representing many different types of taxa from various ecosystems will allow us to partition the relative roles these various factors having in contributing to finding non-random community composition (or results not differing from random), determine generalization that can be drawn from these contrasts, and facilitate the development of relevant theory. Such an approach will characterize the generalities of these answers rather than being unduly influenced by the characteristics of individual datasets. We will also advance the types of null models available to the research community through the development of new approaches that explicitly incorporate differences among species in habitat selection – an area identified as an important one in this research field. From previous work on a large spatial analysis of crayfish communities, we identified dramatic declines in their abundance and species richness in many Ontario lakes, likely reflecting a suite of interacting environmental stressors. We will use a lengthy time series of population abundances for these species from a set of lakes and associated environmental data to better determine what factors are contributing to the decline and loss of these important components of freshwater food webs.