• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Ecotoxicology >Analysing chemical-induced changes in macroinvertebrate communities in aquatic mesocosm experiments: a comparison of methods
【24h】

Analysing chemical-induced changes in macroinvertebrate communities in aquatic mesocosm experiments: a comparison of methods

机译：在水生中观实验中分析化学诱导的大型无脊椎动物群落中的变化：方法的比较

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Mesocosm experiments that study the ecological impact of chemicals are often analysed using the multivariate method 'Principal Response Curves' (PRCs). Recently, the extension of generalised linear models (GLMs) to multivariate data was introduced as a tool to analyse community data in ecology. Moreover, data aggregation techniques that can be analysed with univariate statistics have been proposed. The aim of this study was to compare their performance. We compiled macroinvertebrate abundance datasets of mesocosm experiments designed for studying the effect of various organic chemicals, mainly pesticides, and re-analysed them. GLMs for multivariate data and selected aggregated endpoints were compared to PRCs regarding their performance and potential to identify affected taxa. In addition, we analysed the inter-replicate variability encountered in the studies. Mesocosm experiments characterised by a higher taxa richness of the community and/or lower taxonomic resolution showed a greater inter-replicate variability, whereas variability decreased the more zero counts were encountered in the samples. GLMs for multivariate data performed equally well as PRCs regarding the community response. However, compared to first axis PRCs, GLMs provided a better indication of individual taxa responding to treatments, as separate models are fitted to each taxon. Data aggregation methods performed considerably poorer compared to PRCs. Multivariate community data, which are generated during mesocosm experiments, should be analysed using multivariate methods to reveal treatment-related community-level responses. GLMs for multivariate data are an alternative to the widely used PRCs.
机译：通常使用多元方法“主要响应曲线”（PRC）对研究化学物质生态影响的中观试验进行分析。最近，引入了将广义线性模型（GLM）扩展到多变量数据作为分析生态学中社区数据的工具。此外，已经提出了可以用单变量统计分析的数据聚合技术。这项研究的目的是比较它们的性能。我们编制了旨在研究各种有机化学物质（主要是农药）作用的中观实验的大型无脊椎动物丰度数据集，并对其进行了重新分析。将多变量数据和选定汇总端点的GLM与PRC进行比较，以了解其性能和识别受影响的分类单元的潜力。此外，我们分析了研究中重复发生的变异性。以群落中较高的分类单元丰富度和/或较低的分类学分辨率为特征的中观试验表明，重复样本之间的变异性较大，而样本中遇到的零计数越多，变异性就越低。在社区反应方面，用于多元数据的GLM与PRC表现一样好。但是，与第一轴PRC相比，GLM更好地表明了单个分类单元对处理的响应，因为每个分类单元都采用了单独的模型。与PRC相比，数据汇总方法的效果要差得多。应使用多元方法分析在中观实验期间产生的多元社区数据，以揭示与治疗相关的社区水平的反应。多元数据的GLM是广泛使用的PRC的替代方法。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号