Pollinators are declining throughout the US and the cause of decline appears to be a variety of factors from climate change to seasonal desynchronization. Pollinators are an assemblage of different species ranging in size, functional group and behavior that interact with nectar resources. Can emerging technology and citizen science be used to detect pollinator assemblages based on these interactions? Our research question focused on detecting pollinator interactions defined as a contact between a nectar source by a pollinator. To detect interactions at multiple scales, we asked the following: (1) Can humans and technology equally detect interactions of pollinators of different sizes? (2) Can we detect most pollinator interactions within an observation framework; and (3) can we detect pollinators within a specific observation framework over time. To address these questions and pilot test our approach, we used GoPro Hero5 high resolution paired with human observers with a known species set in the Denver Butterfly Pavilion. At set positions in the pavilion, we paired human/GoPro observations for different durations and tested technology parameters such as focal length and photo bursts versus time lapse video to compare observation types and probability of detecting interactions.
Results/Conclusions
The preliminary pilot study at the Butterfly Pavilion was designed to set up protocols for summer field study in Yellowstone, Grand Tetons and North Cascades National Parks. Observation timing was determined based on a power analysis using interaction durations and video shutter speed from the pilot study. Results showed the need to adjust frame speed using high resolution video in time lapse bursts; however the high resolution videos were clear and captured interactions in great detail. Taxonomic details for fast moving bee species could be determined from preliminary data. In addition, the high resolution videos captured additional interaction details missed by human observers suggesting that paired observations of humans and cameras may reveal complex pollinator interactions across different size classes. The next phase of the study will take place in Yellowstone to test human versus GoPro detection abilities for long term pollinator studies with the added element of a typical wildlife camera trap focused within the observation frame as a third observation strategy. These results are a first step in setting up a pollinator monitoring project in a national park using non evasive GoPros to detect fine scale pollination interactions that can be replicated to determine long term trends.