COS 19-1 - Using Computer Vision to Understand the Effects of Flowering Stage and Temperature on Insect Visitation in an Apple Orchard

Monday, August 7, 2023

1:30 PM – 1:45 PM PDT

Location: E143

Presenting Author(s)

GM

Grace Melone

Graduate Student
University of Wisconsin-Madison, United States

Coauthor(s)

MS

Matthew Smith

Postdoctoral Fellow
University of Wisconsin-Madison, United States
OB

Olivia Bernaeur

Postdoctoral Fellow
University of Wisconsin-Madison, United States
JW

Julia Wiessing

Undergraduate Student
University of Wisconsin-Madison, United States
JC

James Crall

Assistant Professor
University of Wisconsin-Madison, United States

Abstract: Insect-mediated pollination provides a vital ecosystem service, but may be threatened by climate and land use change. In an era of increasing climate variability, understanding the role of temperature in pollinator activity and functional diversity is essential for maintaining food security and wild plant populations. Simplification of agricultural systems is decreasing structural heterogeneity, which reduces the thermal niches available to insect pollinators. To understand the effects of microclimate on pollinator diversity and activity, we sampled temperature and insect activity using camera traps in an apple orchard in South Central Wisconsin. We divided each tree into four microclimates based on cardinal direction and placed a camera and temperature probe within each microclimate with the camera focused on open flowers (n = 5 trees, 20 cameras in total). The cameras automatically recorded photos once every 6 seconds during daylight hours for 2 weeks and were independently powered with batteries. We classified insects into 5 main classes (e.g. flies, bees, ants, and hemipterans), and into further sub-classes for flies and bees. We found distinct differences in peak activity between different insect classes by temperature, day of experiment, and microclimate. Flies had a bimodal temperature-activity distribution, whereas bees had a peak in visitation at higher temperatures. Peak flower visitation differed for insect class by day of experiment, which corresponded to flower stage. Flies were present before and during bloom, and ants were most common after bloom. Unsurprisingly, bees were most frequently detected during peak bloom dates. With our microclimate data, we were also able to determine which thermal niche insects prefer at different air temperatures. When it was hot ( >35 ˚C), bees were more likely to be detected in colder microclimates, and at lower temperatures, bees selected for warmer microclimates. Flies were less likely to have a preference for warmer or colder microclimates. Our research demonstrates the value of using camera trapping to monitor insect diversity and activity along with environmental variables. The results of our study highlight the importance of developing thermal heterogeneity in crop systems to increase and conserve the breadth of insect functional diversity.