Centre d'expertise en analyse environnementale du Québec, ministère du Développement durable, de l'Environnement et de la Lutte contre les changements climatiques, 2700 rue Einstein, Québec GIP 3W8, Canada.
Université du Lorraine, LIEC, CNRS UMR 7360, Avenue du Général Delestraint, F-57070 Metz, France.
Sci Total Environ. 2019 Feb 1;649:233-246. doi: 10.1016/j.scitotenv.2018.08.267. Epub 2018 Aug 23.
This study aims to evaluate the anthropogenic pressure in the St. Lawrence River by assessing the relationships between chemical contamination of sediments and benthic community structure with the trait-based approach. Organic and inorganic contaminants as well as other sediment variables (sediment grain size, total organic carbon, nutrients, etc.) and benthic invertebrate assemblages were determined in 59 sites along the river. Biological and ecological traits of taxa were coded, taking into account regional climate and ecosystem conditions. The aims of this study were to (1) describe the relationships between traits and macroinvertebrate taxa and identify homogeneous clusters of taxa with the same combinations of functional traits, (2) describe spatial patterns in traits of macroinvertebrates in the St. Lawrence River, (3) link trait-based metrics and site groups to sediment quality and (4) define a trait-based strategy for diagnosing the ecological quality of the St. Lawrence River. Seven groups of taxa sharing similar trait-category attributes were defined. Moreover, four groups of sites were identified using the 'K-mean' non-hierarchical clustering approach. The 'IndVal' method enabled us to specifically defined trait categories corresponding to site groups on the basis of their indicator value. The relative abundances of taxa from five functional groups significantly varied among site groups. For example, some indicator traits such as multivoltine cycle, long life span, fixed clutches, tegumental respiration, asexual reproduction, and collector/gatherer feeding habit were associated to the most heavily polluted sites located in the Montreal harbour which showed the highest sediment concentrations in Pb, Zn and Cu. Three trait-based pressure-specific models were built, based on the random forest approach, for respectively (1) heavy metals, (2) BPCs and PAHs, and (3) TBTs occurring in the environment. These models could be applied to assess sediment quality using macroinvertebrate assemblages in a large Canadian river.
本研究旨在通过基于特征的方法评估圣劳伦斯河的人为压力，评估沉积物化学污染与底栖群落结构之间的关系。在该河流的 59 个地点测定了有机和无机污染物以及其他沉积物变量（沉积物粒径、总有机碳、营养物质等）和底栖无脊椎动物组合。考虑到区域气候和生态系统条件，对分类单元的生物和生态特征进行了编码。本研究的目的是：（1）描述特征与大型底栖无脊椎动物类群之间的关系，并确定具有相同功能特征组合的类群的同质聚类；（2）描述圣劳伦斯河中大型底栖无脊椎动物的特征空间模式；（3）将基于特征的指标与站点群联系起来，并与沉积物质量联系起来；（4）定义基于特征的策略来诊断圣劳伦斯河的生态质量。确定了 7 个具有相似特征类属属性的类群。此外，还使用“K-均值”非层次聚类方法确定了 4 个站点组。“IndVal”方法使我们能够根据其指示值，专门定义与站点组相对应的特征类别。五个功能组的分类单元相对丰度在站点组之间有显著差异。例如，一些指示特征，如多化周期、长寿命、固定卵堆、表皮呼吸、无性繁殖和收集者/觅食者摄食习性，与位于蒙特利尔港的污染最严重的站点有关，这些站点显示出 Pb、Zn 和 Cu 沉积物浓度最高。基于随机森林方法，分别建立了三个基于特征的特定压力模型，用于（1）重金属，（2）BPCs 和 PAHs，以及（3）环境中的 TBTs。这些模型可用于使用大型河流中的底栖无脊椎动物组合评估沉积物质量。
