【作者】 于传善；

【导师】 潘灿平；

【作者基本信息】 中国农业大学 ， 农产品安全， 2014， 博士

【摘要】 农药的科学管理需要大量的残留试验数据作为支持,但有限的人力、物力和财力决定了无法在每种农药/作物上都进行残留试验,并获得残留数据。小作物上农药登记不足,农药无限量标准可依的情况普遍存在。结合作物分类研究作物组或亚组的农药残留规律,并以此制定相关农药的作物组限量,是解决组内小作物上农药登记和限量标准缺失的有效方法。本论文对结球甘蓝上残留量的比例类推(proportionality),五种结球芸苔属蔬菜上六种典型农药的残留规律,五种结球芸苔属蔬菜的作物分类及结球芸苔属亚组的残留数据集应用进行了研究和评价。论文就几个方面的研究内容和主要结果如下：建立了五种结球芸苔属蔬菜上甲基托布津、多菌灵、甲霜灵、精吡氟禾草灵、毒死蜱和高效氯氟氰菊酯残留的QuEChERS提取,LC-MS/MS或GC-ECD分析的方法。在结球甘蓝、紫甘蓝、抱子甘蓝、球茎甘蓝和羽衣甘蓝五个基质的空白样品中进行三个水平的添加回收验证：0.01、0.05和0.1mg/kg,六种农药的平均回收率在76.9-117.4%之间,相对标准偏差在3.7-10.8%之间。六种农药在五种结球芸苔属蔬菜基质中的定量限(LOQ)均为0.01mg/kg。设计田间试验,考察了两种施药量条件(1倍和1.5倍GAP)下结球甘蓝上六种农药的残留水平,并以此进行了残留量的比例类推研究。通过不同种类农药的残留量自然对数线性回归分析可知：残留量的比例类推需要在有明显残留的情况下才可以应用；有明显残留量的农药残留汇总后的线性回归方程为y=0.803x+0.4928,相关系数为0.9346,斜率接近理论值1,由截距计算的残留倍数为1.6,接近理论值1.5。对各农药进行单独考察,非内吸性农药毒死蜱和高效氯氟氰菊酯残留量的比例关系明显,线性回归方程与理论方程较为一致,可以进行直接的残留量比例类推；内吸性农药多菌灵(含甲基托布津)和甲霜灵残留量的线性回归方程与理论方程在斜率上有一定的差异,但仍保持良好的线性关系。仓储条件下高低剂量施药试验结果与田间结球甘蓝的试验结果相似,说明环境因素对残留量的比例类推影响可以忽略。因此,可以通过残留量的比例类推实现不同GAP施药量产生的残留数据的整合利用。设计田间试验,对六种典型农药在五种结球芸苔属蔬菜上的残留规律进行了研究。田间消解动态试验结果表明：甲基托布津、多菌灵、甲霜灵、毒死蜱、高效氯氟氰菊酯和精吡氟禾草灵在五种结球芸苔属蔬菜中的半衰期分别在0.9-6.1天、1.3-5.0天、0.9-6.9天、0.8-4.0天、1.4-6.2天和0.3-3.3天之间,在所有的五种结球芸苔属蔬菜中,结球甘蓝上六种农药的半衰期几何平均数最长,为2.1-3.5天。多菌灵和高效氯氟氰菊酯的半衰期几何平均数最大。大多农药在球茎甘蓝上有最低的原始沉积量,而在羽衣甘蓝或抱子甘蓝上有最高的原始沉积量。根据7天安全间隔期下五种结球芸苔属蔬菜上的农药残留数据差异及作物形态的比较,对五种结球芸苔属蔬菜的作物分类进行了考察。羽衣甘蓝的残留量最高且差异明显,作物形态也不同,不应纳入结球芸苔属亚组；结球甘蓝、抱子甘蓝和紫甘蓝的残留量无明显差异且作物形态较类似,均应纳入结球芸苔属亚组；球茎甘蓝的残留量最低,无健康风险,可以纳入结球芸苔属亚组。因此,合理的结球芸苔属亚组应包括结球甘蓝、紫甘蓝、抱子甘蓝和球茎甘蓝,结球甘蓝应作为该亚组的代表作物。根据结球芸苔属亚组的蔬菜种类及每种蔬菜上的六种农药残留数据所形成的完整残留数据集对结球芸苔属亚组蔬菜的组限量进行了计算,OECD和NAFTA计算器的结果较为一致。对结球芸苔属亚组蔬菜的残留量进行了慢性和急性风险评估,慢性风险商在0-0.6%之间,急性风险商在0.4-54.1%之间,均低于100%,对消费者健康无慢性和急性风险。以完整数据集的组限量计算结果作为参照,评价了以代表作物结球甘蓝的残留数据集进行残留外推(residue extrapolation)的限量计算。结果表明,应用残留外推法的多菌灵总量、甲霜灵、高效氯氟氰菊酯和精吡氟禾草灵组限量值与基于完整数据集的组限量值无明显差异；而毒死蜱的外推组限量计算值0.06mg/kg则与参照0.4mg/kg差异较大。因此,在结球芸苔属亚组蔬菜上大多数农药的残留外推法应用具有可行性。对于非内吸性且具有较高残留量的农药,如毒死蜱,应在残留外推法应用的基础上,辅助组内作物残留试验进行验证,以保证组限量值可以反映真实残留情况,并保证消费者健康。更多还原

【Abstract】 Scientific regulation of pesticides requires a large amount of residue data, while limited resources could be used in supervised residue trials for each pesticide/crop combination, in which residue levels are investigated. Therefore, the situations that few pesticides are registered on minor crops and lack of Maximum Residue Limits (MRLs) are widely observed. It would be a useful way that pesticide registrations and MRLs setting could be carried out for crop group by utilizing the combined residue dataset of crop group or the representative commodity. In this thesis, residue behaviors of six pesticides on brassica vegetables were investigated. Commodity group was classified and different types of residue datasets of the Head Brassicas subgroup were evaluated in MRLs setting for the subgroup. Specified results are summarized as follows:A QuEChERS-LC-MS/MS and GC-ECD method for simultaneously analysis of thiophanate-mehtyl, carbendazim, metalaxyl, Fluazifop-P-butyl, chlorpyrifos and lambda-cyhalothim residues on five brassica vegetables (cabbage, red cabbage, kohlrabi, kale and brussels sprout) was developed and validated. At fortification levels of0.01,0.05and0.1mg/kg, it was shown that recoveries ranged from76.9%to117.4%with RSDs of3.7-10.8%(n=5) in all samples. The LOQs of six pesticides in five matrixes were all0.01mg/kg.Pesticide residues in cabbage at the application rates of both GAP and1.5*GAP were investigated. The result showed that a proportionality existed in residues from different application rates on condition that the residues were at measurable levels. Equation of linear regression of natural logarithms of residues was y=0.803x+0.4928, with a correlation coefficient of0.9346. The slope was close to1(theoretical value) and the calculated multiple from the concept (1.6) was close to1.5(theoretical value). Proportionality was obvious in chlorpyrifos and lambda-cyhalothim residues rather than thiophanate-methyl and metalaxyl residues, but linear relationships performed well in all residues. Results from field and storage were similar and it confirmed that environment did not affect the proportionality of residues in side-by-side trials. This result would be useful in utilizing residue data from different application rates of pesticides.Residue behaviors of six pesticides in five brassica vegetables were investigated. Half-lives of thiophanate-methyl, carbendazim, metalaxyl, chlorpyrifos, lambda-cyhalothrin and fluazifop-P-butyl in five brassica vegetables were0.9-6.1days,1.3-5.0days,0.9-6.9days,0.8-4.0days,1.4-6.2days and0.3-3.3days. Pesticides had the longest geo-mean half-life of2.1-3.5days in cabbage. Carbendazim and lambda-cyhalothim had the slowest dissipation in all five brassica vegetables. The lowest initial deposits of most pesticides were found on kohlrabi and the highest initial deposits were on kale or brussels sprouts.Commodity group was classified according to the results of comparison of residue data on different commodities and the morphology of five brassica vegetables. With the highest residue and different morphology, kale should not be classified into Head Brassicas subgroup. Cabbage, red cabbage and brussels sprouts could be classified into Head Brassicas as they had similar residue data and morphology. Though residues in kohlrabi were at low levels, it could be classified into Head Brassicas because it would not present health risk to customers. Therefore, an appropriate subgroup of Head Brassicas should include cabbage, red cabbage, kohlrabi and brussels sprouts. Cabbage should be selected as the representative commodity as it had the highest residues and largest production and consumption.Combined datasets of5pesticide residues in the subgroup of Head Brassicas were used in MRLs calculations with both NAFTA and OECD MRL calculators. The results of two calculators were similar. Acute and chronic risk assessments were carried out and the chronic and acute risk quotients were0-0.6%and0.4-54.1%, respectively. That both results were below100%indicated that no acute and chronic risk would be presented. Estimated MRLs based on the residue data of the representative commodity cabbage with residue extrapolation were evaluated by comparing to those from the residue datasets of the whole subgroup. The results showed that there were no difference between the two kinds of estimated MRLs of calculated carbendazim (sum of thiophanate-methyl and carbendaizm), metalaxyl, lambda-cyhalothirn and fluazifop-P-butyl based on datasets of representative commodity and the whole subgroup. Significant lower estimated MRL of chlorpyrifos was found in residue extrapolation based on the representative commodity:a MRL of0.06ing/kg compared to0.4mg/kg. This finding indicated that extra residue trials on crops within the commodity group should be carried out to comfirm the rational use when residue extrapolation was applied on pesticides like chlorpyrifos which might have high residue levels. While residue extrapolation in the other pesticides could be applied directly in the subgroup of Head brassicas.更多还原