UBCO researchers have determined that flower gardens intentionally planted beside fruit crops can double crop yield if the timing and blossoms appeal to the targeted bumble bees. Photo credit: Dr. Amanda Liczner.
The term “if you build it, they will come” has taken on a whole new meaning when it comes to creating flower gardens to attract specific pollinators like wild bumble bees.
UBC Okanagan researchers Drs. Rebecca Tyson and Bruno Carturan, both with the Irving K. Barber Faculty of Science, wanted to investigate whether flower gardens—specifically placed beside a crop to attract and support insect pollinators—actually benefit both the bees and crop production.
“Planting wildflower patches near crop fields is considered a potentially effective strategy to support both the abundance and diversity of pollinators and the services they provide,” says Dr. Carturan. “But these management strategies can be costly and not always effective in enhancing crop yield.”
Planting supplemental gardens can lead to larger and healthier wild bee populations, which should be good for crop pollination. However, field studies show contradictory results—while some indicate an increase in crop yield directly related to pollination services, others show no discernible effect.
“While the plan makes sense on paper, it can create a conundrum,” he says. “With more bees in the landscape, there is the potential for greater pollination of crop flowers. But bees can prefer different flowers, guided by nectar sugar content, flower shape and pollen nutrient composition. Consequently, the presence of wildflower patches beside a berry crop could divert bees from pollinating the crop.”
Curious about this distraction phenomenon, Dr. Carturan set out to understand the interplay between the relative timing of crop and wildflower bloom, as well as the quantity, quality and relative attractiveness of the flowers.
For this study, he focused specifically on blueberry crops, an emblematic agricultural product in BC, and bumble bees, which are known for their superior efficiency in pollinating blueberry flowers compared to honey bees. He wanted to vary the size of the crop area, the size of the planned garden relative to the crop and the relative nutritional quality and bloom time of both the crop and additional flowers.
“Creating a field study large enough to properly test how bumble bee pollination services respond to changes in all of these different parameters would be quite a challenge,” he explains. “So, we chose a mathematical modelling approach which required two steps that involved a lot of reading and thinking.”
The first is to design a model that aligns with the goals of the project and realistically captures the key ecological processes at play in the ecosystem. The second is to find proper values for the model parameters.
Dr. Tyson explains the model is fairly complex and they ran thousands of simulations—each characterized by a unique combination of wildflower patch size, nutritional quality of the blossoms and blooming period—before they were able to predict blueberry crop yield.
“Such an extensive sampling design, attainable only through simulation, offers a comprehensive picture of the interacting processes and trade-offs within the system,” she says.
The net result of those simulations determined that providing highly nutritious wildflower resources before the crop blooms can more than double the crop yield. Conversely, providing wildflower resources at the same time as crop bloom can reduce the yield by up to 50 per cent.
“The main result of our virtual experiment clearly shows that the most beneficial strategy is to generate a temporal spillover effect by providing a continuous supply of resources to the bees and avoiding too much competition between the wildflowers and the crop flowers,” she explains. “This keeps the bees well fed during the early foraging season when the colonies are growing, and it prevents a potential distraction effect during crop bloom.”
The researchers hope to refine the model to implement additional aspects of the ecosystem by, for instance, modelling several different bumble bee species rather than just one “average” species. The ultimate goal is to calibrate the model with locally relevant empirical data to help inform planting strategies on a real farm.
However, they advise caution while interpreting these results as they pertain to a virtual system, not actual bees and blueberry crops.
The research appears in Ecological Modelling.
