Untangling the phylogenetic and spatial components of trait variation in ecological assemblages

特征是一组进化的特征，使一个物种能够利用其生态位并在给定的环境中生存。解释物种之间的特征如何以及为什么不同是生态学和进化生物学的一个基本问题。实现这一目标的根本任务是解开进化、环境和生态因素是如何驱动性状变异的。性状进化是由生物和非生物因素的相互作用驱动的，同时受到以往历史的限制，并在所有尺度上发挥作用。从生态学的角度来看，群落的组成和结构是由群落中存在的物种的生态和功能多样性决定的，而生态和功能多样性又取决于这些物种的特征。因此，群落内组织的进化和生态尺度是通过物种特征联系在一起的。为了了解群落内物种特征的变化，考虑物种分布和物种范围重叠的程度是很重要的。在一个地理区域内的群落中，物种特征之间的相似性和差异性将由物种范围重叠的程度，以及范围重叠区域内与外部相比，生态位和栖息地使用的相似性等因素决定。鉴于这一观察结果，该项目的第一个主要目标是建立比较模型，以解释物种在驱动性状变化中的共同分布。在探索性状变异的生态驱动因素时，一个明显的考虑因素是物种在生态上共存的程度。例如，两个物种可能有相似的栖息地需求，其结果是它们往往在同一个群落中一起出现。在这种情况下，当物种的范围重叠时，它们可能出现在同一个群落中。因此，出现在同一群落中的物种可能会共享共同的适应套件，并且特征会积极地共同变化。另外，作为生态相互作用（如竞争）的结果，对性状协方差的负面影响也是可能的：如果共存的物种只具有不同的性状，那么协方差就会减少。该项目的第二个主要目标是发展比较模型，将系统发育与大尺度分布和在其范围内的局部尺度发生模式联系起来。我们将使用我们开发的方法来检验两个假设：假设1：地理分布重叠的物种比地理分布不重叠的物种表现出更大的相似性。假设2：在相同的生态群落中共存的物种，在一定的范围重叠水平下，比那些没有重叠的物种表现出更大的分化。制定方法的战略如下。首先，我们将创建一个模型，根据定义良好的过程来模拟数据。其次，将开发一个统计框架，使模型能够适应数据并根据模拟输出进行测试。最后，我们将把我们的模型应用于真实的数据集，使用每个数据集测试我们的两个主要假设。为了应用这些方法，我们需要包含以下数据集：(i)在一系列生态群落中共同发生的物种特征；（ii）系统发育，并尽可能解决；（iii）物种地理分布范围图；此外，应用目标2中开发的方法：（iv）物种发生的生态措施。我们将把这些模型应用于包含这些元素的两个现有数据集——一个是巴西大西洋森林树木的数据集，另一个是哥伦比亚鸟类的数据集——以对驱动群落组成的因素产生新的见解，在每个案例中测试项目的两个主要假设。

Traits are the set of evolved characteristics that permit a species to exploit its niche and live in a given environment. Explaining how and why traits vary among species is a basic problem in ecology and evolutionary biology. Fundamental to achieving this is the task of disentangling how evolutionary, environmental and ecological factors drive trait variation. Trait evolution is driven by the interplay of biotic and abiotic factors, whilst constrained by previous history, and plays out at all scales. In ecological terms, the composition and structure of communities is determined by the ecological and functional diversity of species present in a community, which is dependent on the traits of those species. Thus, evolutionary and ecological scales of organisation within a community are linked through species traits. To understand the variation in traits of species within a community, accounting for species' distributions and the degree to which species' ranges overlap is important. Within the communities in a geographic region, the similarities and differences between species' traits will be shaped by factors such as the extent to which species' ranges overlap, as well as the similarity of niches and habitat use within areas of range overlap compared with outside. Given this observation, the first main objective of this project is to develop comparative models that account for species' co-distributions in driving trait variation.When exploring the ecological drivers of trait variation, an obvious consideration is the extent to which species co-occur ecologically. For example, two species may have similar habitat requirements, with the consequence that they tend to occur together in the same communities. In this case, it might be expected that when species' ranges overlap they occur in the same communities. Consequently, species that occur in the same communities may be expected to share common suites of adaptations, and traits would co-vary positively. Alternatively, as a consequence of ecological interactions (e.g. competition), negative effects on trait covariance would also be possible: if species that co-occur only possess divergent traits, then covariance will be reduced. The second main objective of this project is to develop comparative models that link phylogeny with both large-scale distributions and local-scale patterns of occurrence within their ranges.We will use the methods we develop to test two hypotheses:Hypothesis 1: species with overlapping geographic distributions show greater similarity than those with non-overlapping ones. Hypothesis 2: species that co-occur within the same ecological communities show greater divergence, for given levels of range overlap, than those that do not.The strategy for developing methods will be as follows. First, we will create a model that simulates data according to well-defined processes. Second, a statistical framework will be developed that allows the model to be fitted to data and tested against the simulation outputs. Finally, we will apply our models to real datasets, testing our two main hypotheses using each dataset.To apply the methods, we require datasets that contain the following: (i) traits of species co-occurring in a series of ecological communities; (ii) a phylogeny, resolved as well as possible; (iii) maps of species' geo-graphic ranges; and, in addition, to apply the methods developed in Objective 2: (iv) ecological measures of species' occurrences. We will apply the models to two existing datasets that include these elements--one on Brazilian Atlantic forest trees, the other on Colombian birds--to generate new insights into the factors that drive community composition, in each case testing the two major hypotheses of the project.