Tracking pollinating bees is a huge business in the world of agriculture. It's estimated that farms in the U.S. pay more than $656 million each year to rent more than 80 billion bees, which are set loose in fields of almonds and other crops to pollinate the plants.

Keeping track of those bees can be expensive, and being alerted to when their pollinator services are lacking can help save a lot of money.

"Causes of pollinator decline are complex and include diminishing flower resources, habitat loss, climate change, increased disease incidence and exposure to pesticides, so pinpointing the driving forces remains a challenge," said Candace Galen, professor of biological science in the University of Missouri College of Arts & Science. "For more than 100 years, scientists have used sonic vibrations to monitor birds, bats, frogs and insects. We wanted to test the potential for remote monitoring programs that use acoustics to track bee flight activities."

Webster University biology professor Nicole Miller-Struttmann believes she has a solution that could help lower those costs and head off pollination deficits by keeping track of all those bees. In a paper published in the scientific journal PLOS ONE, Miller-Struttmann and her colleagues tracked the activity of bumblebees using microphones strategically placed in Colorado meadows and reported great success in being able to predict the bee activity and pollination services.

"Tracking these dynamic populations is costly, and the current methods used to track them are time consuming and often lethal," Miller-Struttmann said. "We used inexpensive sound equipment to monitor for buzzing sounds created by bees as they fly. We then developed a computer algorithm that rapidly identifies and quantifies bee flight activity. We believe that our method could be a much more cost- and time-efficient method for monitoring bee activity."

Using microphones and iPad minis, the researchers recorded bees in three different alpine meadows on Pennsylvania Mountain in Colorado. At the same time, they visually recorded the number of bees foraging in the area. The computer algorithm was used to extrapolated how many "buzzing" sounds were recorded that matched the frequency range of bumblebees. They compared the two counts to gauge accuracy. The acoustic counts were remarkably accurate and highly correlated with the visual counts.

In phase two, the researchers tested to see if the amount of buzzing caught by the microphones accurately predicted if the bees were pollinating plants in an area or just flying near the microphones. The researchers set up two types of plants that bees pollinate, but they made it so one set could only attract bees and could not be pollinated.

Using the recorded sound from each set of plants, their system successfully predicted which flowers would set fruit and which did not, indicating that the method could be used to track bee activity and also detect the services bees provide to the plants they visit.

"We believe that this could potentially be a more cost-effective method of monitoring bees while also providing real-time data to conservation managers and farmers," Miller-Struttmann said. "The technology for acoustic data collection and processing is nimble, low cost and suited to remote locations."

"Eavesdropping on the acoustic signatures of bee flights tells the story of bee activity and pollination services," Galen said. "Farmers may be able to use the exact methods to monitor pollination of their orchards and vegetable crops and head off pollination deficits. Finally, global 'citizen scientists' could get involved, monitoring bees in their back yards."

Currently, using the algorithms developed in this study, the team is developing a smartphone app that could record buzz activity as well as document the bees photographically. Future studies could determine whether bees detect competitors by sound and whether flowers have chemical responses to bee buzzes, Galen said.