An integrated model for predicting bumblebee population success and pollination services in agro-ecosystems

现在人们普遍认识到，野生和人工授粉都提供了一种重要的生态系统服务，对维持自然和半自然生态系统中的生物多样性是必不可少的，对世界上约75%的作物物种的作物生产是必要的。最近的估计表明，英国农作物授粉的价值为4.4亿英镑。在英国和整个温带北半球，主要的农作物传粉者是蜜蜂和熊蜂；蜜蜂种群受到健康问题的威胁，而许多熊蜂物种已经经历了范围内的下降，可能主要是由于栖息地的丧失。因此，迫切需要制定农场管理战略，既维持健康的传粉者数量，又维持作物产量。对于大黄蜂来说，我们缺乏对大黄蜂巢穴生长和存活之间的关系、它们所依赖的花卉资源的时空分布以及它们提供的授粉服务的空间分布的详细了解。我们无法预测一处景观可以支撑多少个大黄蜂巢穴，或者它们可以为多少农作物授粉。我们不知道一个健康的鸟巢需要多少花朵才能支撑，也不知道捕食者和寄生虫在移走蜜蜂和整个鸟巢方面有多重要。该项目建议使用系统方法来回答这些问题，将模拟模型与一系列田间和遗传研究相结合，以提供对农业生态系统中大黄蜂种群控制因素的新见解。我们将使用基于地理信息系统的地图量化10个代表性景观(英格兰5个，苏格兰5个)中花蜜(糖)和花粉(蛋白质)的空间和时间分布。在实验室研究中，我们将量化巢穴健康生长所需的资源(蛋白质和糖)。使用申请者开发的新型摄像系统，我们将监测野生大黄蜂的巢穴，以确定蜂巢和单只蜜蜂的捕食率。这些数据将被输入到模型中，这些模型将被开发来模拟播种到我们的十个景观中的真实大黄蜂巢穴数量的增长，模型还将模拟个体工蚁(代理人)从每个巢穴中觅食的过程，估计他们可能利用的食物块。殖民地将通过资源开发在景观中相互作用，并受到现实的捕食和寄生比率的影响。该模型将通过测量放置在这些景观中的实验巢的实际群体增长率，以及通过测量在这些景观中利用特定饲料斑块的不同大黄蜂物种的巢数(使用遗传技术来识别姐妹)来验证。一旦得到验证，这些模型将被用来预测为田间豆类提供作物授粉服务的空间分布，并通过实地蜜蜂访问率的测量来验证这些预测。这些模型还将被用来模拟在景观中添加花粉和花蜜带的花卉资源的后果，使我们能够量化可以支持多少额外的蜂巢，以及多少额外的作物授粉。该项目将建立在BBSRC资助的Rothamsted项目中目前正在开发的蜜蜂模型的基础上，并与之形成对比。总体而言，这个项目将第一次为我们提供机械联系和定量、有效的预测，将大黄蜂种群、野花、蜜蜂捕食者和作物授粉联系起来。

It is now widely recognized that both wild and managed pollinators provide a vital ecosystem service, essential for the maintenance of biodiversity in natural and semi-natural ecosystems, and necessary for crop production in ~75% of the world's crop species. Recent estimates suggest that crop pollination in the UK is worth £440 million. The main crop pollinators in the UK and throughout the temperate northern hemisphere are honeybees and bumblebees; honeybee populations are threatened by health issues, while many bumblebee species have undergone range declines, probably primarily due to habitat loss. There is thus an urgent need to develop farm management strategies which sustain both healthy populations of pollinators and crop yields. For bumblebees, we lack a detailed understanding of the relationships between the growth and survival of bumblebee nests, the spatial and temporal distribution of the floral resources on which they depend, and the spatial distribution of pollination services they provide. We cannot predict how many bumblebee nests a landscape can support, or how many crop plants they can pollinate. We do not know how many flowers are needed to support a single healthy nest, or how important predators and parasites are in removing bees and whole nests. This project proposes to use a systems approach to answer these questions, combining simulation models with a range of field and genetic studies to provide a new insight into factors governing bumblebee populations in agroecosystems. We will quantify the spatial and temporal distribution of pollen (protein) and nectar (sugar) provided by flowers in 10 representative landscapes (5 in England, 5 in Scotland) using GIS-based maps. In lab studies we will quantify the resources (protein and sugar) needed for healthy growth of a nest. Using novel camera systems developed by the applicants we will monitor wild bumblebee nests to establish rates of predation of nests and of individual bees. These data will feed in to models which will be developed to simulate the growth of realistic numbers of bumblebee nests seeded into our ten landscapes, with models also simulating the foraging of individual workers (agents) from each nest, estimating which food patches they are likely to exploit. Colonies will interact in the landscape through exploitation of resources, and be subject to realistic rates of predation and parasitism. The model will be validated by measuring actual colony growth rates of experimental nests placed into these landscapes, and by measuring the number of nests of different bumblebee species exploiting particular patches of forage within these landscapes (quantified using genetic techniques to identify sisters). Once validated, the models will then be used to predict the spatial distribution of crop pollination services for field beans, and these predictions tested by field measurements of bee visitation rates. The models will also be used to simulate the consequences of adding floral resources to the landscape in terms of pollen and nectar strips, enabling us to quantify how many extra bee nests, and how much extra crop pollination, can be supported. The project will build on and contrast with honeybee models currently being developed in a BBSRC-funded project at Rothamsted. Overall this project will, for the first time, provide us with mechanistic links and quantitative, validated predictions linking bumblebee populations, wildflowers, bee predators and crop pollination.