Vieira Danilo Cândido, Fonseca Gustavo
Centro de Estudos do Mar, Universidade Federal do Paraná, Caixa Postal 50.002, Pontal do Paraná, PR, 83255-000, Brazil.
Universidade Federal de São Paulo, Av. Dona Ana Costa, 95, Santos, SP, CEP 11060-001, Brazil.
Oecologia. 2019 Jul;190(3):535-545. doi: 10.1007/s00442-019-04441-w. Epub 2019 Jun 20.
The purpose of this framework is to identify the relative importance of selection and dispersion processes in structuring ecological communities. Using a "pattern-oriented modelling" approach, it consists of five steps: (a) aggregate information from the empirical community and its environment, (b) simulate communities under different degrees of dispersal and selection, (c) select the best set of simulations into a composite model using the environmental boundary (EB) and niche breadth (NB) of each observed species, (d) validate the composite model by comparing expected and observed results from three additional community patterns and (e) classify each observed species along the selection/non-selection continuum. A free-living marine nematodes data set from a coastal bay was used as empirical example. A total of 20 parameterizations were applied varying selection and dispersion levels. In the absence of selection, species from high-dispersal parameter sets showed maximum EBs and NBs, whilst selection parameter sets generated species with narrower EB and NB values. EB and NB values declined with decreasing dispersal. The composite model encompassed 96% of the 194 nematode species and predicted all the three patterns evaluated without further calibration, i.e., they are independent: (1) abundance-rank distribution, the assemblage structures along both the (2) spatial and (3) environmental gradients. Non-selection and selection parameter sets accounted for 34% and 85% of the observed species, respectively. The main advantage of this approach is that empirical niche measurements are placed in the context of model-generated expectations, enabling a deeper understanding of community assembly processes and how they vary from species to species.
该框架的目的是确定选择和扩散过程在构建生态群落中的相对重要性。它采用“面向模式建模”方法，包括五个步骤：(a) 汇总来自经验群落及其环境的信息；(b) 在不同扩散和选择程度下模拟群落；(c) 使用每个观察物种的环境边界 (EB) 和生态位宽度 (NB)，从一组最佳模拟中选择构建复合模型；(d) 通过比较另外三种群落模式的预期结果和观察结果来验证复合模型；(e) 沿着选择/非选择连续体对每个观察物种进行分类。以一个沿海海湾的自由生活海洋线虫数据集作为实证示例。总共应用了20种参数设置，改变选择和扩散水平。在没有选择的情况下，来自高扩散参数集的物种显示出最大的EB和NB值，而选择参数集产生的物种具有较窄的EB和NB值。EB和NB值随着扩散的减少而下降。复合模型涵盖了194种线虫中的96%，并在无需进一步校准的情况下预测了所有三种评估模式，即它们是独立的：(1) 丰度-秩分布，以及沿 (2) 空间和 (3) 环境梯度的组合结构。非选择和选择参数集分别占观察物种的34%和85%。这种方法的主要优点是，将经验性生态位测量置于模型生成的预期背景下，从而能够更深入地理解群落组装过程以及它们如何因物种而异。
