Occupancy of Bombus affinis (Hymenoptera: Apidae) in Minnesota is highest in developed areas when standardized surveys are employed

Abstract

Mounting evidence of bumble bee declines and the listing of the rusty patched bumble bee (Bombus affinis Cresson) as federally endangered in the United States in 2017 and Canada in 2012 has stimulated an interest in monitoring and conservation. Understanding the influence of land use on occupancy patterns of imperiled species is crucial to successful recovery planning. Using detection data from community surveys, we assessed land use associations for 7 bumble bee species in Minnesota, USA, including B. affinis. We used multispecies occupancy models to assess the effect of 3 major land use types (developed, agricultural, and natural) within 0.5 and 1.5 km on occupancy of 7 Bombus (Hymenoptera: Apidae) species, while accounting for detection uncertainty. We found that B. affinis occupancy and detection were highest in developed landscapes and lowest in agricultural landscapes, representing an inverse relationship with the relative landcover ratios of these landscapes in Minnesota. Occupancy of 2 bumble bee species had strong positive associations with natural landscapes within 1.5 km and 2 species had strong negative associations with agricultural landscapes within 1.5 km. Our results suggest that best practices for imperiled Bombus monitoring and recovery planning depends upon the surrounding major land use patterns. We provide recommendations for urban planning to support B. affinis based on conservation success in the metropolitan areas of Minneapolis-St. Paul. We also encourage substantial survey effort be employed in agricultural and natural regions of the state historically occupied by B. affinis to determine the current occupancy state.

Advancing bee conservation in the US: gaps and opportunities in data collection and reporting

“…Occupancy models, when completed using best practices, provide a more accurate assessment of a species distribution because they consider imperfect detection (Graves et al, 2020; Guzman et al, 2021; Boone et al, 2023b, 2023a; Janousek et al, 2023; Otto et al, 2023). Their creation requires information about surveys where species were detected or not detected (i.e., absence data where the individual count is zero), and multiple visits at the same site.…”

“…This precludes performing some analyses that are critical for conservation decision-making, such as estimating species ranges with species distribution models (SDMs) or calculating extinction risk through population viability analyses (PVAs). For example, occupancy models, which can be powerful for detecting trends while accounting for some of these issues, have only recently begun to be developed for bee species (Graves et al, 2020; Otto et al, 2021, 2023; Boone et al, 2023b, 2023a), and the data needed to calculate these models are not often collected in routine field surveys. Recent research from other groups of insects (especially butterflies) highlights declines in broadly distributed, formerly abundant species (e.g.…”

“…These analyses helped to both document the extent of decline (Graves et al, 2020) and identify the primary causes, particularly the role of neonicotinoid insecticides (Janousek et al, 2023). In the case of B. affinis , fully standardized surveys, where effort is known and not inferred, have been conducted in recent years to support occupancy modeling (Boone et al, 2023b, 2023a; Otto et al, 2023). This work has helped to optimize detection probabilities and ultimately improved monitoring program design, which is essential for efficient and effective monitoring.…”

Advancing bee conservation in the US: gaps and opportunities in data collection and reporting

“…This precludes performing some analyses that are critical for conservation decision-making, such as estimating species ranges with species distribution models (SDMs) or calculating extinction risk through population viability analyses (PVAs). For example, occupancy models, which can be powerful for detecting trends while accounting for some of these issues, have only recently begun to be developed for bee species (Graves et al, 2020;Otto et al, 2021, Otto et al, 2023Boone et al, 2023a;Boone et al, 2023b), and the data needed to calculate these models are not often collected in routine field surveys. Recent research from other groups of insects (especially butterflies) highlights declines in broadly distributed, formerly abundant species (e.g., Wepprich et al, 2019;Forister et al, 2021;Van Deynze et al, 2022;Forister et al, 2023).…”

“…These analyses helped to both document the extent of decline (Graves et al, 2020) and identify the primary causes, particularly the role of neonicotinoid insecticides (Janousek et al, 2023). In the case of B. affinis, fully standardized surveys, where effort is known and not inferred, have been conducted in recent years to support occupancy modeling (Otto et al, 2023;Boone et al, 2023a;Boone et al, 2023b). This work has helped to optimize detection probabilities and ultimately improved monitoring program design, which is essential for efficient and effective monitoring.…”