A systems approach to understanding the impacts of sublethal doses of neonicotinoids on bumblebee and honeybees

新烟碱类杀虫剂是目前世界上使用最广泛的杀虫剂之一，通常用于保护油菜和向日葵等开花作物免受食草动物的侵害。在英国，2010年有127万公顷的作物使用了新烟碱类杀虫剂，其中噻虫胺、吡虫啉和噻虫嗪是最常用的杀虫剂。大多数被用作种子敷料，作物在生长过程中吸收这种化合物，从而免受草食性害虫的侵害。新烟碱类也被用作软性水果作物的叶面喷雾剂，并作为喷雾剂或颗粒剂广泛出售，用于花园中的花卉、水果和蔬菜作物。由于它们是全身性化学物质，在处理过的作物的花蜜和花粉中发现了低浓度的新烟碱类，因此它们被传粉者消耗。它们也是水溶性和持久性的，因此在处理过的作物附近的树篱/田野边缘野花中发现了可检测的数量。越来越多的证据表明，蜜蜂暴露于新烟碱类杀虫剂会产生亚致死效应，如导航能力下降，而这在实验室分析或将蜂箱紧靠处理过的开花作物的现场试验中是检测不到的。只有当蜜蜂在远处觅食时，它们才会很明显，就像自然界中发生的那样。首席PI最近的一项研究表明，将大黄蜂巢穴暴露于已知存在于经种子处理的油菜花蜜和花粉中的吡虫啉水平，会导致新蜂后的产量减少85% （Whitehorn et al. 2012）。很大一部分野生大黄蜂巢穴和人工管理的蜂群很可能在经过处理的开花作物的觅食范围内，因此，这类杀虫剂很可能是引起广泛讨论的野生蜜蜂数量下降的主要原因。迫切需要了解在野外条件下新烟碱类杀虫剂对大黄蜂行为和种群水平性能的亚致死影响。该项目将量化在英国耕地作物和野花中发现的新烟碱的水平，并将这些数据输入花卉资源分布的景观地图（由一个平行项目BB/J014753/1制作，从2013年开始）。我们将量化蜜蜂和大黄蜂群体在农田自然觅食时暴露于新烟碱类杀虫剂的剂量，然后再评估这种暴露对大黄蜂群体表现的可能影响。暴露和影响将通过实证研究和建模方法进行检查。奥斯本博士和贝彻博士开发了一个蜂群的综合模型，该模型可以预测在真实或模拟的环境中觅食者的行为和蜂群的表现。在BB/J014753/1项目中，我们将为大黄蜂群体开发该模型的一个版本。这些模型将使我们能够预测蜜蜂和大黄蜂工蜂将在哪里觅食，从而预测它们的暴露，并将其与放置在景观中的真实巢穴所接收的暴露进行比较。然后，我们将使用洛桑研究所独特的谐波雷达设备，检查暴露于实际现场水平的新烟碱是否会改变或损害觅食行为。这些数据将用于开发蜜蜂和大黄蜂的子模型，其中包括新烟碱类在预测蜂群成功方面的影响，该模型将在实际的大黄蜂群体中进行实地试验，并可作为用户的风险评估工具。最后，数据和模型结果将使我们能够制定管理建议，以尽量减少新烟碱类杀虫剂对管理蜜蜂和野生大黄蜂种群的影响。

Neonicotinoid insecticides are now among the most widely used pesticides in the world, and are routinely used to protect a range of flowering crops such as oilseed rape and sunflowers against herbivores. In the UK, 1.27 million ha of crops were treated with neonicotinoids in 2010, with clothianidin, imidacloprid and thiamethoxam being the most frequently used types. Most are used as a seed-dressing, with the crop plant absorbing the compound as it grows, and hence becoming protected against herbivorous pests. Neonicotinoids are also used as foliar sprays on soft fruit crops, and are widely sold as sprays or granules for garden use on flowers, fruit and vegetable crops. As they are systemic chemicals, low concentrations of neonicotinoids are found in nectar and pollen of treated crops, and hence they are consumed by pollinators. They are also water soluble and persistent, so that detectable quantities are found in hedgerow / field margin wildflowers near treated crops. Evidence is mounting that exposure of bees to neonicotinoids produces sublethal effects such as reduced navigational ability that would not be detected in laboratory assays, or in field trials where hives are placed immediately adjacent to a treated flowering crop. They are only apparent when bees are foraging at distance across the landscape, as occurs in nature. A recent study by the lead PI demonstrated that exposure of bumblebee nests to levels of imidacloprid which are known to occur in the nectar and pollen of seed-treated oilseed rape resulted in an 85% reduction in output of new queens (Whitehorn et al. 2012). A substantial proportion of both wild bumblebee nests and managed honeybee colonies are likely to be within foraging range of a treated, flowering crop, so it seems likely that this class of insecticide may be contributing substantially to the much-discussed declines in wild bee populations. There is an urgent need to understand the sublethal impacts of neonicotinoids on behaviour and colony-level performance of bumblebees under field conditions. This project will quantify the levels of neonicotinoids found in crops and wildflowers in UK arable farmland, and feed these into landscape maps of the distribution of floral resources (produced in a parallel project, BB/J014753/1, starting 2013). We will quantify the doses of neonicotinoids that honeybee and bumblebee colonies are exposed to when naturally foraging in farmland, before going on to assess the likely impacts of this exposure on bumblebee colony performance. Exposure and impacts will be examined by both empirical studies and modelling approaches. Drs Osborne and Becher have developed an integrative model of a honeybee colony, which predicts forager behaviour and colony performance when placed into a real or simulated landscape. In project BB/J014753/1 we will develop a version of this model for bumblebee colonies. These models will enable us to predict where both honeybee and bumblebee workers will forage, and thus predict their exposure, which will be compared to that received by real nests placed in the landscape. We will then examine whether exposure to neonicotinoids at realistic field levels alters or impairs foraging behaviour using Rothamsted's unique harmonic radar facility. These data will be used to develop a sub-model for both honeybees and bumblebees which includes impacts of neonicotinoids in predictions of colony success, which will be tested with real colonies of bumblebees in field experiments, and can be used as a risk assessment tool by users. Finally, the data and model results will enable us to develop management recommendations to minimise the impacts of neonicotinoids on both managed honeybees and wild bumblebee populations.