Agri-environment schemes, like flower fields, have been implemented in the EU to counteract the dramatic decline of farmland biodiversity. Farmers in Lower Saxony, Germany, may receive payments for three flower field types: annual, perennial (five years old), and mixed flower fields composed of yearly alternating annual and biannual parts. We assessed the effectiveness of these flower field types in providing bumblebee foraging habitat compared to control cereal fields. We sampled bumblebees with transect walks and assessed the richness of exploited pollen plants using DNA meta-barcoding and direct observations.

All flower field types enhanced bumblebee abundance and species richness compared to control fields but attracted mostly three generalist species. Although we expected highest benefits from the more heterogeneous mixed flower fields, abundance was highest in annual, only intermediate in mixed, and lowest in perennial flower fields. Bumblebee species richness did not differ between flower field types.

Overall, the proportion of sown plants in pollen loads was surprisingly low (< 50%). Bombus pascuorum, but not B. terrestris agg., exploited 10% of the sown plant species in perennial, 36% in annual and 45% in mixed flower fields, respectively. Compared to direct observations, pollen samples revealed 4.5 times more visited plant species and thus assessed floral resource use more reliably. Plant species richness in pollen loads decreased with local flowering plant species richness and increased with proportion of annual crops in the landscape, potentially due to the exploitation of more diverse and scattered resources, including flowering crops, in homogenized landscapes to fulfil dietary requirements.

Our results indicate that under the current management, both annual and mixed flower fields provide the most attractive food resources, while perennial flower fields offered the poorest foraging habitats. Conclusively, flower fields seem important but resources from the surrounding landscape are still needed to sustain bumblebees in agricultural landscapes.

欧盟已经实施了诸如花田之类的农业环境计划，以应对农田生物多样性的急剧下降。德国下萨克森州的农民可能会收到三种花田类型的付款：一年生，多年生（五岁）以及由每年交替的一年生和半年生组成的混合花田。与对照谷物田地相比，我们评估了这些花田类型在提供大黄蜂觅食栖息地方面的有效性。我们通过样带行走对大黄蜂进行采样，并使用DNA元条形码和直接观察评估了被开发花粉植物的丰富度。

与对照田相比，所有花田类型均提高了大黄蜂的丰度和物种丰富度，但主要吸引了三种通才种。尽管我们期望从更多种类的混合花田中获得最大收益，但丰度在一年中最高，在混合花田中仅为中度，而在多年生花田中最低。大黄蜂物种丰富度在花田类型之间没有差异。

总体而言，播种植物在花粉中的比例令人惊讶地低（<50％）。Bombus pascuorum，但不是B. terrestris agg。，利用所述播种植物物种的10％在多年生，36％的年和45％在混合花田，分别。与直接观察相比，花粉样品显示出探访植物物种多出4.5倍，因此更可靠地评估了花卉资源的利用。花粉负荷中的植物物种丰富度随当地开花植物物种丰富度而降低，并随景观中一年生作物的比例增加而增加，这可能是由于在均质化景观中开发了更加多样化和分散的资源（包括开花作物）以满足饮食需求。

我们的结果表明，在目前的管理下，一年生和混合花田均提供最诱人的粮食资源，而多年生花田则提供最差的觅食生境。总而言之，花田似乎很重要，但是仍然需要来自周围景观的资源来维持农业景观中的大黄蜂。