First audit of macroinvertebrate samples from an EU Water Framework Directive monitoring program: human error greatly lowers precision of assessment results
First audit of macroinvertebrate samples from an EU Water Framework Directive monitoring program: human error greatly lowers precision of assessment results
复制标题
DOI：
10.1899/09-183.1
复制
发表时间：
2010-12-01
影响因子：
--
通讯作者：
Sundermann, Andrea
中科院分区：
文献类型：
--
作者：
Haase, Peter;Pauls, Steffen U.;Sundermann, Andrea
无脊椎动物经常用于土壤和水生态系统的生物监测。由于无脊椎动物物种数量巨大，样品处理（分选和鉴定）是一项劳动密集型且往往困难的任务，容易出错。这些错误可能会使评估结果产生偏差，而环境管理人员往往利用这些结果来指导昂贵的恢复措施的资金决策。然而，许多淡水监测项目没有对评估结果进行质量控制。我们进行了第一次审计的官方欧洲淡水监测计划的基础上，414个大型无脊椎动物样本的溪流和河流在德国。这些样品是由7个不同商业环境实验室的工作人员使用欧洲联盟（欧盟）水框架指令协议收集的。我们对12%的样本进行了3个不同级别的审计：1）分类审计，2）识别审计，以及3）基于分类和识别的总体审计。分类审核显示，29%的标本和每5个分类单元（20.6%）被初级分析员忽略。分类的差异与分类单元的个体大小相关（r = 0.61，p < 0.001）。鉴定审计表明，>30%的分类群之间的主要分析师和审计员的结果不同。被认为难以识别的分类群并不比被认为容易识别的分类群更容易出错。初级分析员和审计员将34%的审计样品分配到3个评估模块（有机污染、酸化和一般降解）中的>= 1个的不同质量类别。对于16%的样本，这些变化导致了不同的最终生态评估。主要分析员和审计员之间如此高的差异率可能导致数百万欧元的无效分配。我们的研究结果清楚地表明，在淡水监测中需要适当的质量控制和审计。
Invertebrates are often used in biological monitoring of soil and water ecosystems. Because of the huge number of invertebrate species, sample processing (sorting and identification) is a labor-intensive and often difficult task that is prone to error. These errors can bias assessment results, which often are used by environmental managers to guide funding decisions for costly restoration measures. However, quality control of assessment results is not implemented in many freshwater monitoring programs. We conducted the first audit of an official European freshwater monitoring program based on 414 macroinvertebrate samples from streams and rivers in Germany. The samples were collected by personnel at 7 different commercial environmental laboratories using the European Union (EU) Water Framework Directive protocol. We audited 12% of all samples at 3 different levels: 1) a sorting audit, 2) an identification audit, and 3) a total audit based on both sorting and identification. The sorting audit revealed that 29% of the specimens and every 5(th) taxon (20.6%) had been overlooked by the primary analyst. Differences in sorting were correlated with taxon body size (r = 0.61, p < 0.001). The identification audit showed that >30% of taxa differed between the results of the primary analysts and auditors. Taxa considered difficult to identify were not more prone to error than were taxa considered easier to identify. Primary analysts and auditors assigned 34% of audited samples to different quality classes in >= 1 of 3 assessment modules (organic pollution, acidification, and general degradation). For 16% of the samples, these changes resulted in a different final ecological assessment. Such a high rate of differences between primary analysts and auditors could lead to ineffective allocation of several million Euros. Our results clearly illustrate the need for adequate quality control and auditing in freshwater monitoring.