A study showing the effects of land-use quality on the productivity of bumblebee colonies highlights the importance of resource availability across space and time in promoting survival over generations. See Letter p.547
The world's roughly 250 bumblebee species are valuable pollinators in agricultural and natural ecosystems, but population declines threaten to diminish their pollination services1,2. Possible causes of bumblebee declines include climate change, pesticides and pathogens1,3,4 — factors all ultimately connected to an overall deterioration of landscape quality owing to increasing urbanization and intensified farming. Much effort has been directed towards strategies to improve pollinator habitats in agricultural ecosystems5, but successful habitat restoration remains challenging because of uncertainty about how landscape quality influences bumblebee populations across their complex life cycle. Carvell et al.6 address this issue on page 547, tracking the effects of landscape characteristics on the survival of bumblebee families from one generation to the next.
Carvell et al. studied 3 common bumblebee species within 20 square kilometres of heterogeneous agricultural landscape in the United Kingdom, comprising farmland, grassland, woods and villages. This area is partially managed under a restoration scheme that involves planting wild flowers along the margins of crop fields. The authors used a remote-sensing technique to generate a map of land-use categories, which they then surveyed for flora and suitable nest sites to gauge each area's value for bumblebees. Examples of high-quality floral habitats include plots in which a high percentage of the land is covered by flower species that are visited by bumblebees throughout spring and summer. High-quality habitats for building nests on or under the ground include plant litter, thick grasses and burrows dug by mammals.
The usefulness of evaluating these measures of habitat quality stems from the complex life history of bumblebees. The insects follow an annual social life cycle, with colonies founded each spring by a queen that also performs nectar- and pollen-foraging duties. These duties are taken over by workers later in spring and summer, while the queen lays eggs that produce further workers. In autumn, egg production shifts to a new generation of queens and males. To understand population changes over generations of this complex life cycle, Carvell and colleagues took DNA samples from 537 queens and 2,101 worker bees throughout the 20-km2 landscape over 2 years.
One challenge in understanding how landscapes influence social insects such as bumblebees is that surveys of individuals do not necessarily provide accurate information about the number of colonies, which is important for predicting population trajectories. Carvell et al. used genetic markers to link the surveyed bees to their colonies; because bumblebees from the same colony are sisters, genetic relatedness can reveal which samples are taken from foragers of the same colony. By documenting the coordinates at which each bee was captured, it is then possible to triangulate the approximate location and foraging range of each nest, and to estimate nest density across the landscape. In addition, the authors go a step further than previous genetic studies that analysed bumblebee-colony density and foraging distance within single seasons, tracking relatedness from queen to worker to queen, both within generations and between them, and analysing the dispersal of queens across the landscape over the years.
Carvell and colleagues' primary finding is that the quality of the habitat that surrounds nest sites affects lineage survival between years (Fig. 1). Beneficial features include: mixed semi-natural vegetation (for example, field margins sown with flowers); spring flowers from trees and hedgerows that are good for queens early in the season; and summer flowers that are good for workers. By examining the spatial distributions of these resources around estimated colony locations, the authors quantify this finding, suggesting that having sufficient floral resources within about 1 km of each colony increases queen production and survival. A comparison of samples taken across seasons also clearly demonstrates the value of persistent floral resources for queens and workers throughout the colony life cycle, and of suitable hibernation and nesting sites to facilitate successful overwintering and founding of the next generation.
The results extend those from other studies7,8, which found that having semi-natural habitat and floral richness at similar spatial scales promotes colony growth, density and persistence within a season. No previous studies, to my knowledge, have tracked colony and queen survival within and between generations — a major component of year-to-year bumblebee population dynamics.
One unanswered question is whether remaining in the same patch of land over generations is necessarily ideal for a bumblebee colony. If high-quality restored habitat is nested within a poor-quality landscape that impedes queen emigration or immigration from other high-quality patches, this could promote inbreeding in the long term. Carvell et al. find some evidence that high levels of non-crop land cover can promote queen dispersal into the broader landscape, but this is a subject in need of more research.
Given the large geographical ranges in which individual bumblebee species are found and the potential for long-distance queen dispersal, even the 20-km2 landscape studied by Carvell and colleagues represents only a small fraction of each species' total population9. One could therefore imagine that a network of restored habitats across a broader agricultural landscape, each with a high density of season-long floral resources, would both maintain connectivity and boost local pollinator populations. But establishing the benefit of such a management scheme for pollinators and crops will require the non-trivial effort of expanding and replicating the current study in other landscapes.
Declines in bumblebee populations are increasingly recognized by both the public and policymakers, resulting in several notable conservation actions — including the recent proposed federal listing of a bumblebee species as endangered in the United States10 (although at the time of writing, this listing has been delayed). Carvell et al. have provided valuable information for landscape managers about restoration activities that could complement such conservation strategies by enhancing agricultural landscapes for wild pollinators. Their study focuses on land-use effects for just three species, but the detailed data set provides a framework that can be expanded to more species and larger areas. It could also be used to investigate the effects of pesticides, pathogens and other factors on multigenerational life-cycle dynamics, with the goal of identifying a realistic way to optimize land-use heterogeneity for bumblebee populations.