Bridging the Gap between Field Experiments and Machine Learning: The EC H2020 B-GOOD Project as a Case Study towards Automated Predictive Health Monitoring of Honey Bee Colonies.

Honey bee colonies have great societal and economic importance. The main challenge that beekeepers face is keeping bee colonies healthy under ever-changing environmental conditions. In the past two decades, beekeepers that manage colonies of Western honey bees (Apis mellifera) have become increasingly concerned by the presence of parasites and pathogens affecting the bees, the reduction in pollen and nectar availability, and the colonies' exposure to pesticides, among others. Hence, beekeepers need to know the health condition of their colonies and how to keep them alive and thriving, which creates a need for a new holistic data collection method to harmonize the flow of information from various sources that can be linked at the colony level for different health determinants, such as bee colony, environmental, socioeconomic, and genetic statuses. For this purpose, we have developed and implemented the B-GOOD (Giving Beekeeping Guidance by computational-assisted Decision Making) project as a case study to categorize the colony's health condition and find a Health Status Index (HSI). Using a 3-tier setup guided by work plans and standardized protocols, we have collected data from inside the colonies (amount of brood, disease load, honey harvest, etc.) and from their environment (floral resource availability). Most of the project's data was automatically collected by the BEEP Base Sensor System. This continuous stream of data served as the basis to determine and validate an algorithm to calculate the HSI using machine learning. In this article, we share our insights on this holistic methodology and also highlight the importance of using a standardized data language to increase the compatibility between different current and future studies. We argue that the combined management of big data will be an essential building block in the development of targeted guidance for beekeepers and for the future of sustainable beekeeping.

Floral resources,energetic value and pesticide residues in provisions collected by Osmia bicornis along a gradient of oilseed rape coverage.

Pollinators in agricultural landscapes are facing global decline and the main pressures include food scarcity and pesticide usage. Intensive agricultural landscapes may provide important food resources for wild pollinators via mass flowering crops. However, these are monofloral, short-term, and may contain pesticide residues. We explored how the landscape composition with a different proportion of oilseed rape (6-65%) around Osmia bicornis nests affects floral diversity, contamination with pesticides, and energetic value of provisions collected by this species of wild bees as food for their offspring. Altogether, the bees collected pollen from 28 plant taxa (6-15 per nest) and provisions were dominated by Brassica napus (6.0-54.2%, median 44.4%, 12 nests), Quercus sp. (1.2-19.4%, median 5.2%, 12 nests), Ranunculus sp. (0.4-42.7%, median 4.7%, 12 nests), Poaceae (1.2-59.9%, median 5.8%, 11 nests) and Acer sp. (0.6-42%, median 18.0%, 8 nests). Residues of 12 pesticides were found in provisions, with acetamiprid, azoxystrobin, boscalid, and dimethoate being the most frequently detected at concentrations up to 1.2, 198.4, 16.9 and 17.8 ng/g (median 0.3, 10.6, 11.3, 4.4 ng/g), respectively. Floral diversity and energetic value of provisions, but not the Pesticide Risk Index depended on landscape structure. Moreover, pollen diversity decreased, and energetic value increased with landscape diversity. Thus, even a structurally simple landscape may provide diverse food for O. bicornis if the nest is located close to a single but resource-diverse patch. Both B. napus and non-crop pollen were correlated with pesticide concentrations.

Effects of agricultural landscape structure, insecticide residues, and pollen diversity on the life-history traits of the red mason bee Osmia bicornis.

Agricultural landscapes have changed substantially in recent decades, shifting from the dominance of small fields (S) with diverse cropping systems toward large-scale monoculture (L), where landscape heterogeneity disappears. In this study, artificial nests of the red mason bee, Osmia bicornis, were placed in S and L landscape types on the perimeter of oilseed rape fields representing different oilseed rape coverages (ORC, % land cover). The local landscape structure around each nest was characterised within a 100, 200, 500, and 1000 m radius using ORC and 14 landscape characteristics, which were then reduced by non-metric multidimensional scaling (nMDS) to two axes: nMDS1 characterised the dataset primarily according to land fragmentation and the main crop, whereas nMDS2 captured the prevalence of more natural areas in the landscape. Pollen diversity and insecticide risk levels in the pollen provisions collected by the bees were analysed, and their dependence on the landscape structure was tested. Thereafter, the effects of pollen diversity, insecticide risk, and landscape structure on the life-history traits of bees and their sensitivity to topically applied Dursban 480 EC were determined. Pollen taxa richness in a single nest ranged from 3 to 12, and 34 pesticides were detected in the pollen at concentrations of up to 320 ng/g for desmedipham. The O. bicornis foraging range was relatively large, indicating that the landscape structure within a radius of ~1000 m around the nest is important for this species. Pollen diversity in the studied areas was of minor importance for bee performance, but the ORC or landscape structure significantly affected the life-history traits of the bees. Contamination of pollen with insecticides affected the bees by decreasing the mass of newly emerged adults but their sensitivity to Dursban 480 EC was not related to environmental variables.