Drummond Francis A, Ballman Elissa S, Eitzer Brian D, Du Clos Brianne, Dill James
School of Biology and Ecology, University of Maine, Deering, Orono, ME.
Cooperative Extension, University of Maine, Orono, ME.
Environ Entomol. 2018 Apr 5;47(2):378-387. doi: 10.1093/ee/nvy023.
In 2015, we conducted a statewide assessment of honey bee exposure to pesticides with assistance of volunteer beekeepers. Pollen trapping was conducted at 32 sites in the spring, summer, and early fall. Apiary locations ranged from unmanaged natural landscapes to managed agricultural or urban landscapes. Pollen samples at each site were aggregated over the collection dates and chemical residue analysis was conducted on each pollen sample for 190 pesticides and metabolites using HPLC/MS. Twenty-five different residues were detected for an average of 2.9 detections per site. Detections were dominated by fungicides, but risk, calculated as: ppb residue concentration/LD50, was mostly due to insecticides. Beekeeper perceived land-use in the vicinity of each apiary was associated with significant differences in the number of detections and residue concentrations, agricultural landscapes greater than nonagricultural. However, there was no significant difference in oral or contact risk quotients due to land-use type. The landscape composition surrounding apiaries, derived with GIS, determined pesticide exposure for honey bees when total detections, log pesticide residue concentration, and log contact risk quotients were used as measures. Partial least squares explained 43.9% of the variance in pesticide exposure due to landscape composition. The best predictors describing pesticide exposure were: area (ha) of blueberry, coniferous forest, and urban/developed land cover types. Maine is the most forested state in the United States (as determined by % land area forested, 93%) and a negative exponential decay was observed between land area in conifer forest and the number of pesticide detections per apiary.
2015年，我们在志愿养蜂人的协助下，对全州范围内蜜蜂接触农药的情况进行了评估。在春季、夏季和初秋时节，于32个地点进行了花粉采集。养蜂场的位置范围涵盖了从未管理的自然景观到管理的农业或城市景观。每个地点的花粉样本在采集日期内进行汇总，并使用高效液相色谱/质谱联用仪（HPLC/MS）对每个花粉样本进行190种农药及其代谢物的化学残留分析。共检测到25种不同的残留物，每个地点平均检测到2.9种。检测结果中杀菌剂占主导，但以“ppb残留浓度/LD50”计算的风险大多归因于杀虫剂。养蜂人对每个养蜂场附近土地利用情况的认知与检测数量和残留浓度的显著差异相关，农业景观中的差异大于非农业景观。然而，由于土地利用类型的不同，口服或接触风险商数并无显著差异。当使用总检测量、农药残留浓度对数和接触风险商数对数作为衡量指标时，利用地理信息系统（GIS）得出的养蜂场周围景观构成决定了蜜蜂接触农药的情况。偏最小二乘法解释了因景观构成导致的农药接触差异的43.9%。描述农药接触情况的最佳预测指标是：蓝莓、针叶林以及城市/已开发土地覆盖类型的面积（公顷）。缅因州是美国森林覆盖率最高的州（根据森林覆盖土地面积百分比确定，为93%），并且观察到针叶林面积与每个养蜂场农药检测数量之间呈负指数衰减关系。
