农药残留分析不确定度的评价及其应用

本文讨论了农药残留分析的主要误差来源、分析结果的不确定度、减少不确定度的可能措施以及分析结果可置信度的实际应用。为使分析结果具有可比性, 在取样和样本制备时应采用国际共同认可的FAO法规方法。残留数据不确定度的主要来源是残留量在田间施药点之间的可变性,占80%~100%;其次是施药点田间残留量分布的可变性, 占30%~40%;再次是样本实验室制备(占20%~30%)和分析方法不确定度(占10%~20%)。 从某一田间取样并分析的植物样本, 其残留结果的组合不确定度包括取样、样本制备和分析, 至少在33%~44%之间。平行样本分析结果的评价应考虑分析的组合不确定度。在一定可置信范围评价某一农产品是否超出法定最高残留限量(MRL)时需要分析较大数量的样本。在实际工作中如何经济而可靠地分析并得出结论,本文从统计学角度提出了解决方案。

Estimation of Uncertainty of Pesticide Residue Analysis and Its Use

The objectives of this paper are to identify the major sources of errors in pesticide residue data, estimate the inevitable uncertainty (variability) of the results, recommend actions to reduce the uncertainty, and discuss the practical consequences of inevitable variation of the results. In performing sampling and sample preparation the internationally agreed upon procedures and definitions should be carried out and followed if comparable results are sought. The field-to-field variability of residues is the largest source of uncertainty (80%-100%) of residue data, generally followed by the within-field (30%-40%), sampling process (20%-30%) and analytical (10%-20%) variability. The combined uncertainty of residue data, comprising the uncertainty of sampling, sample processing and analysis, is estimated to be a minimum of 30%-50% for plant samples taken from a single lot. The combined uncertainty should be taken into account in deciding on the significance of differences of residue content of replicate samples in dispute cases. The certification of the compliance of a lot with Maximum Residue Limits (MRL) requires the analysis of a number of samples. Practical approaches are recommended for performing such tasks economically.