Emerging threats and opportunities to managed bee species in European agricultural systems: a horizon scan

Managed bee species provide essential pollination services that contribute to food security worldwide. However, managed bees face a diverse array of threats and anticipating these, and potential opportunities to reduce risks, is essential for the sustainable management of pollination services. We conducted a horizon scanning exercise with 20 experts from across Europe to identify emerging threats and opportunities for managed bees in European agricultural systems. An initial 63 issues were identified, and this was shortlisted to 21 issues through the horizon scanning process. These ranged from local landscape-level management to geopolitical issues on a continental and global scale across seven broad themes—Pesticides & pollutants, Technology, Management practices, Predators & parasites, Environmental stressors, Crop modification, and Political & trade influences. While we conducted this horizon scan within a European context, the opportunities and threats identified will likely be relevant to other regions. A renewed research and policy focus, especially on the highest-ranking issues, is required to maximise the value of these opportunities and mitigate threats to maintain sustainable and healthy managed bee pollinators within agricultural systems.

Managed pollinators provide a wide range of benefits to society in terms of contributions to food security, farmer and beekeeper livelihoods, and social and cultural values 1 .Bees are important pollinators worldwide, with ~ 20,000 species; however, only 19 bee species are currently managed for crop pollination services 2 .In Europe, the main managed bee species are Apis mellifera, Bombus terrestris, and to a lesser extent, solitary bees such as those belonging to the genus Osmia 3 .Bees, along with other pollinators, face a range of threats including landscape modification, climate change, pests, pathogens, and agrochemicals [4][5][6] .While these issues are common across both wild and managed species, there may be other risks or opportunities that are specific to managed bees in a European agricultural context.Identifying these stressors or opportunities in a timely and effective manner can enable the development of effective policies and mitigation strategies across Europe (EU and national equivalents) to sustain healthy populations of managed bees.
Safeguarding European food security and promoting agricultural sustainability remains a prominent political ambition, driving the implementation of the European Green Deal and the Farm to Fork strategy 7,8 .Yet, current geopolitical instabilities and recovery from the worldwide COVID pandemic could potentially delay or even undermine many of the identified pathways to achieving these goals 9 .In hindsight, these issues might have been foreseeable, highlighting the importance of a forward scanning process to ensure policies are as preemptive as possible, rather than reactive.To make informed decisions, policymakers and practitioners need to anticipate the likely developments and their impact to understand and proactively develop preventative action plans.A systematic approach, such as routine horizon scanning, can provide the necessary insights to do this 10,11 , helping guide research priorities to generate actionable knowledge for policy and practice.
Managed pollinators are an important part of European food sustainability and are integral to the Farm to Fork strategy.To this end, we used a core expert group to horizon scan for potential threats and opportunities to managed bees in European agricultural systems over the next five to ten years.

Results
A summary for each of the 21 shortlisted issues follows (Fig. 1; Table 1).Issues are listed by whether they were identified as an opportunity, threat, or both.Issue rank order and broader theme are indicated in parentheses e.g., [4; Technology].

Opportunities
Greater availability of technology and automation to remotely monitor bee colony health [4; Technology]  The development of new techniques, to monitor and improve bee colony health status, based on artificial intelligence and deep learning has provided enormous recent advances in the field 12 .Advances include systems that track honey bees over hundreds of meters with high precision 13 , and new tools to monitor parameters such as duration and number of foraging trips (i.e., potential proxy for food flow) of individual solitary bees 14 .Furthermore, integration of disease and parasite prevalence with meteorological predictions and nectar flow information can provide the basis for important decision support tools for beekeepers, provided that the data is validated with appropriate field studies.A recent project attempted to integrate different types of data originating from diverse sources 15 , but further effort is required in this direction as currently data collection is highly unaggregated and diverse.Geographical information systems can also be used for supporting local and central authorities in decision-making processes relating to environmental planning 16 .The development of sensor technology, the spread of wireless infrastructures, and the increased ability to manage and model big data and provide predictions, could all represent an opportunity to interconnect apiaries across Europe and produce real-time predictions that could support decisions in the field.

Co-formulants in agrochemical formulations and managed bee health [5; Pesticides & Pollutants]
While co-formulants (i.e., ingredients added to active substances to produce the formulated product) are not expected to exert pesticidal impacts 17 , some were already shown to have lethal effects on honey bees in the early 1970s 18,19 and additional concerns have been raised recently 20,21 .Current regulatory requirements list acute and chronic toxicity studies for formulations, which includes the testing of co-formulants in the context of the entire formulation 22 .A recent study confirmed that this requirement is justified by showing that different formulations of a herbicide varied in toxicity to bumble bees, due to differences in co-formulants rather than the active ingredient 23 .However, not all formulations are tested, and for those that are, testing can be quite limited 21 .Reinforcing the systematic study of formulant and formulation toxicity is therefore a potential opportunity to improve managed bee health.For example, if future research shows that specific co-formulants have potential impacts on managed bees, these could be removed or replaced by less impactful ingredients reducing a potential risk to managed bee health.Finally, a more in-depth knowledge of co-formulant toxicity could help to inform risk management and product labelling, and training for use that reduces exposure.

Increase of varroa-resistant stocks of Apis mellifera [6a; Predators & Parasites]
The significant negative impact of varroa mites on honey bees is well-established and widely recognised 24,25 .Most beekeeping operations strongly rely on chemical treatments to control mite populations; however, these can cause negative side effects and may become ineffective 26 .An alternative approach is to selectively enhance heritable honey bee traits for resistance or tolerance to the mite through breeding programs or select for naturally surviving untreated colonies.A recent review 27 of studies on populations resistant or tolerant to varroa showed that in most cases, survival of both naturally and artificially selected populations is due to the expression of several traits (e.g., grooming, hygienic behaviour, varroa sensitive hygiene) that appear to collectively confer resilience to varroa infestation.Currently, around fifteen traits are recognised as regulatory traits that can be assessed in the field or in the lab 27 .However, a Europe-wide survey showed that despite huge demand, there is no well-established market for resistant stock in Europe, in part due to the increased cost of resistant stock and variable honey production benefits (i.e., resistant stock did not always produce more honey) 28 .and authorities 33 .Measures to achieve areas of 'high diversity' include implementing pollinator-friendly actions, such as the promotion of wild and cultivated flowers on large amounts of arable land 34,35 and improving the quality of existing habitats to better meet the needs of managed bees and other pollinators 36 .

Optimising diets of managed bees to develop better artificial diets and inform agri-environment schemes [11; Management Practices]
The nutritional requirements of managed bees today may not be sufficiently met due to landscapes being increasingly characterized by agriculturally intensive cropping and monocultures 37 .The differences between what bees require and what their environment can provide, has contributed to the decline in managed bee populations in some countries (e.g., USA) 38 , and raises the questions of whether and how managed bees should be provided with supplemental food when nutritional deficits occur.Studies show that access to floral, and pollen, resource diversity provides amino acids and lipids that can support overall development, tolerance to parasites and immune system activity of bees [39][40][41] .This knowledge could be used to improve artificial diets and inform agri-environment schemes by selecting appropriate floral resource combinations to support pollinators and could accompany ongoing actions under the EU Biodiversity Strategy.For example, pollen of Asteraceae plants, including sunflowers, have been shown to reduce parasitic infection in managed bee species 42 .However, solely relying on Asteraceae pollen might not be sufficient, as it has a low protein content 43 , but if included in a pollen mix it could help improve pollinator health.Developing tailored seed mixtures to meet bee nutritional and health requirements could be a great opportunity in the next few years.

Artificial intelligence for disease, weed and pest control to reduce pesticide use in agroecosystems [18b; Technology]
Artificial Intelligence (AI) is the use of digital data and technology to fulfill specific operations such as weeding (using robots that can recognize weeds and remove them), or sensor equipped sprayers that allow direct application of a herbicide on to weeds only (reducing the volume of products sprayed by more than 50% 44 ).It is estimated that one-third of global crop production is lost due to weed competition and another third due to pest and disease damage, with pesticides effective in combating these 45 .As early as the mid-1980s, AI for disease, weed  46 , and the first AI applications for crop production were developed 47 .The use of AI for disease and weed control is certainly expected to increase; however, even though AI solutions have already been used for over three decades in agriculture, their use to specifically reduce the risk to bees associated with pesticides is limited 48 .Nonetheless, it presents an opportunity to reduce potential risks to managed bee health.

Development of field instruments for evaluation of genetic markers to be used in breeding for resilience [20; Technology]
Biotechnology is advancing at a fast pace 49 , and recent advances could help to facilitate efforts to identify and select molecular markers that indicate the presence of certain resilience traits in honey bees.For instance, causative genes and proteins associated with resistance or tolerance could be developed as marker-assisted selection (MAS) tools for improving breeding stock at a large scale 50,51 .In addition, DNA-based technologies have become more affordable over the last decades, so the financial aspects may not necessarily be prohibitive.Relatively cheap single nucleotide polymorphism (SNP)-based assays have already been developed for some traits linked to resilience 52 .Portable PCR tools are already in use 53 , and it is feasible to foresee portable genetic marker kits that could directly be used in the field and assist beekeepers in selecting colonies with traits linked to resilience (to parasites, to drought, to higher temperatures).However, this potential is offset by various issues including the differing suites of genes underlying resilience and sensitivity to stressors identified in different honey bee populations 25 .

Thermic vehicles and the hazardous pollutants they release will decrease in the coming years [21; Pesticides & Pollutants]
The opportunity arising from a shift from thermic to electric vehicles may be considered a relatively new issue.The global trend in electric vehicles suggests we will move from around a 5-10% market share in 2022 to a 25-50% share (depending upon region) by 2030 54 .The expectation is that the pressures on managed pollinators from pollutants from vehicles, in general, will be reduced, although it does not prevent all risks (e.g., turbulence and metals in dust) associated with road pollution 55 .Given the amount of land taken up by areas such as road verges (~ 270,000 km 2 ) 56 , a proportion of which would be visited by bees, this is not an insignificant change.
The situation is complex (e.g., environmental footprint of rare metal extraction) and hard to quantify, though qualitatively, the switch to electric vehicles would likely be an improvement.

Increasing threat of emerging predators and pathogens [1; Predators & Pathogens]
The spread of non-native and invasive species and the emergence of novel pathogens, variants of existing ones and shifting modes of transmission are a continuing threat to managed bee populations [57][58][59] .For example, a recent modelling study showed that the steady increase in alien species belonging to different taxa observed in the last fifty years will not slow down in the near future in all continents including Europe 60 .Europe may become a suitable niche for new (e.g., Vespa mandarinia 61 ) and spreading (e.g., Vespa orientalis [62][63][64] and Aethina tumida 65 ) species, thus adding to the pressure from current invasives (e.g., Vespa velutina 66 ).Furthermore, pathogen transfers between honey bees and invasive species have been found, underlining that impacts on honey bee populations may be direct (i.e., predation) and indirect (i.e., pathogen dynamic) 67 .Additionally, any potential shift in virus transmission mode (e.g., from faecal/food-oral to vector mediated) could pose a future threat to bees and apiculture 57,68 .Therefore, it is likely that both the number of invasive predators and the impact from pathogens will continue to grow in the next ten years increasing the burden posed to managed bees.

Extreme weather events [3; Environmental Stressors]
The impact of some extreme weather and climatic events on pollinator communities is well-characterized in the literature [69][70][71] .However, the significance of these events, including those that are less well-characterized (e.g., extreme frost events), and how such events might interact with other drivers of decline to exacerbate negative impacts on managed bee populations across Europe, is less well understood.The impact of extreme temperature and heatwaves are already emerging 72,73 , and there is further anecdotal evidence that the summer heatwaves of 2022 in France affected egg-laying in honey bees during Robinia pseudoacacia nectar flow and severe spring rainfall in Spain led to colony collapse due to lack of foraging resources (anecdotal communications gathered by horizon scan experts).Interactions between extreme climatic events and other drivers of decline are a significant threat in the foreseeable future.

Increasing numbers of inexperienced beekeepers [6b; Management Practices]
Beekeeper experience is a key factor in determining responses to honey bee health issues 74 , and an increase in the number of inexperienced beekeepers has been identified as an emerging threat to bee health.Several studies at a pan-European level have found that beekeeper background and apicultural practices are major drivers of honey bee colony losses 75,76 .Inexperienced beekeepers with small apiaries experience up to double the winter mortality rate compared to experienced beekeepers, possibly due to inadequate disease control 77 .Sick colonies can also favour the spread of pathogens within Apis mellifera due to typical honey bee behaviour (robbing, swarming) and possibly also across other bee species 78 .
Vol:.( 1234567890) Micro-plastics (MPs) (plastics < 5 mm, including nano-plastics which are < 0.1 μm) have been identified as an emerging threat in terrestrial systems 79 .MPs are readily absorbed into plants from the soil 80 , and bee bodies through contaminated food under laboratory conditions 81 ; they can also absorb pollutants such as pesticides acting as a source and sink of environmental contaminants 82 .MPs can increase honey bee mortality (albeit only at high concentrations 83 ), decrease feeding rate and body weight 84 , change the diversity of gut biota and gene expression related to oxidative damage, detoxification, and immunity, and increase worker susceptibility to antibiotics 82 .MPs likely interact with other environmental stressors, and co-occurrences are highly likely in agricultural landscapes; for example, honey bees showed higher mortality to viral infection when exposed to MPs 85 .More research to monitor MPs (e.g., http:// www.insig nia-bee.eu) is needed to generalise exposure patterns, i.e., across food webs (nectar and pollen), between bee species and in different landscape contexts, to provide essential information for their monitoring and management 82,86 .Given MPs are already ubiquitous in the environment 87 and are poorly understood in the context of managed bees 88 there is the potential for them to be a significant threat to managed bees.

Direct and indirect effects of biopesticides on bees [13; Pesticides & Pollutants]
Biopesticides include a broad range of products, including natural (or nature identical) chemical substances, plant or animal extracts, pheromones or semiochemicals, untransformed inorganic pesticides and microorganisms (e.g., bacteria, viruses, or fungi).A recent update in the EU Regulations has clarified the data requirements and approval criteria for a subcategory of biopesticides (microorganisms) 89 , yet concerns remain around the risk assessment of biopesticides in general.In the case of semiochemicals, inorganics and nature-identical chemicals that are usually the sole active component in a formulation, risk assessments are well established.However, for complex mixtures or microorganisms that typically exert activity as an organism plus secondary active metabolites, testing methods are still evolving and, in some instances, may not be developed enough to provide clear results 90,91 .Without new standardized testing methods to address potential non-intentional effects of biopesticide active substances and their formulations on managed bees, biopesticides could represent a significant threat.

Increase of migratory beekeeping [16b; Management Practices]
More frequent droughts and severe heat waves will likely contribute to an increase in migratory beekeeping, with increases expected in terms of the proportion of hives relocated and the distance travelled.Additionally, European policies provide subsidies for migratory beekeeping, as a means of providing ecosystem services to marginal areas 92 .Recent studies, however, suggest that migratory beekeeping leads to increased disease risk 93 (although see Bartlett et al. 94 ), genetic introgression 95,96 and may affect local pollinator biodiversity 97 .Given the importance of locally adapted genotypes in Europe 98 and the threats posed by disease, increases in migratory beekeeping could have a high negative impact on European honey bee health.

Cutting pollinators out of food production [16a; Crop Modification]
Excluding pollinators from food production continues to be a threat to the sustainability of managed bee populations, through plant breeding and cultivation practices.For example, methods to promote parthenocarpy (fruit set in the absence of fertilisation), such as genetic modification, hormone application and selective breeding, may reduce the need for pollinators in many horticultural crops 99 .Whilst reducing our dependence on pollinators may allow growers to extend their growing seasons, it could remove our imperative to utilise them 10 .This may have unintended consequences for commercial beekeepers and apiaries, to ultimately affect the pollination of non-parthenocarpic pollinator-dependent crops such as seed and nut crops and wild plants.

Nanotechnology-based pesticides (NBPs) [2; Pesticides & Pollutants]
Nanotechnology can modify a pesticide's solubility, stability, and efficacy to improve crop protection 100 .However, this process changes NBPs' environmental fate and behaviour, and this emerging technology has outpaced our understanding of how NBPs may affect pollinators 101,102 .NBPs may be an opportunity for managed bees as their stability and controlled-release mechanisms increase efficiency to reduce the chemical required on crops 103 .Only one study has explored the effect of NBPs on pollinators, showing that a pyrethrum extract in a nanocarrier was safer than a traditional pyrethrum extract 104 .However, like traditional pesticides, NBPs may threaten managed bees and other non-target organisms through toxicity, yet virtually no data exist to test this 105 .Indeed, the structure of NBPs, which is similar to pollen, means that bees are adapted to collect and move NBPs, resulting in exposure, and no studies have explored bees' exposure to NBPs 101 .NBPs are rapidly evolving, poorly understood, and likely to substantially impact managed bees in agricultural landscapes.
Changing farm practice and timing of the demand for managed bees [10; Management Practices] Among the EU Green Deal strategic policies, the development of Sustainable Food Systems foresees a significant change in food production schemes and practices 8 , which may either pose an opportunity or a threat depending on the context and the practices recommended or adopted.Opportunities may exist through fulfilling global strategic moves to diverse crop production, less dependence on global markets and increased connection to local production sources, and more sustainable approaches taken with respect to the use of water and energy resources or the use of land 16 .For example, recent research has highlighted the potential benefits of crop diversification for pollinators while keeping crop yield stable 106 , although crop diversity also drives the frequency and intensity of pesticide use 107 .Refining effective agricultural best-practices, such as selecting optimal seed-mixes for floral www.nature.com/scientificreports/strips, may also increase the benefits for pollinators and offer further opportunities 108 .These practices would operate alongside changes triggered by adaptations to climate change, which the policies are trying to tackle.In this context, changes that may negatively impact managed bees will be observed in crop availability, growing and flowering seasons, with concomitant impacts on the need for managed pollinators in space and time to meet crop pollination demands, and honey production.

Strengthening trade and biosecurity measures in Europe to better protect local managed bee populations, managed bee breeding and trade [12; Political & Trade Influences]
The lack of limitations on the trade and movement of managed bees has benefitted disease spread and has been causing genetic erosion of local bee populations 93,109,110 , ultimately resulting in the loss of traits involved in bee resilience.Currently, bees fall under several regulations at European level for importations [111][112][113] , and only honey bee queens and bumble bees are permitted to enter the EU, subject to health requirements.Health requirements include checking for signs of small hive beetle (Aethina tumida), mite (Tropilaelaps spp.and Varroa spp.) and bacterial (Paenibacillus larvae) infestations, however there are no regulations regarding other pathogens or trade magnitude 114 .To prevent genetic erosion of local bee populations, subspecies of bees need to be included in regulations.This is particularly pertinent given genotype-environment interactions are described as underlying the complex relationships between local populations of honey bees, landscape, infection, and parasites (particularly Varroa spp., viruses and Nosema spp.).Furthermore, regulations for solitary bee trade should also be introduced.Without these changes the threat to managed bee populations will continue, however, there is an opportunity for EU legislators to include genetic diversity protection of managed bees in the CAP strategy and more specifically in the National Apiculture Programmes.In this way, trade and biosecurity measures can contribute to the protection of local managed bee populations from genetic introgression, as well as from the spread of diseases.

Impact of war in Ukraine on the EU Common Agricultural Policy, food prices and agroecological transitions [14a; Political & Trade Influences]
The Russian invasion of Ukraine has significantly affected the import and export of crops and grains that impact food security.In response, the European Commission 115 has presented a range of short-term and medium-term actions to enhance global food security and to support farmers.Impacts of the war in Ukraine on the agricultural policy of Europe may be both a threat and an opportunity for managed bees.For example, the recent decision to allow the tillage of fallow lands to palliate food shortages due to the conflict may lead to a reduction in the uptake of agri-environment type measures (e.g., wildflower strips) that benefit bees.However, if alternative crops which are mass flowering, such as clover or sunflower, are planted then at least for the flowering period there could be a benefit for bees 116 .

Accessibility of European pesticide exposure datasets [14b; Pesticides & Pollutants]
Researchers, particularly ecotoxicologists, need precise information on pesticide use in the landscape.While the EU Pesticides Database 117 provides information such as active substances used in plant protection products or Maximum Residue Levels (MRLs) in food products, it does not provide information on spatial and temporal patterns of use of commercial products across Europe.There are two main sources of information for pesticide use at European level: the Common Agriculture Policy (CAP) dataset and data produced to comply with the regulations on statistics on pesticides 118 .Currently, these datasets are not readily accessible to the public.Although attempts to address these issues in the regulatory framework are underway (e.g., through the requirement for records of pesticide use to be kept by farmers 118 ), data from the different European countries are not aggregated in a single database and efforts still need to be made to standardise data collection and collation across Member States.

Prime editing and genetically modified crops in Europe [18a; Crop Modification]
The EU currently has extensive limits on the use and development of GM crops.Member States are seeking new regulatory frameworks to make EU research institutions competitive at an international level 119 .Along with base editing, prime editing is a relatively new genomic technique based on the CRISPR-Cas9 system 120 .This presents an opportunity, as the first prime edited plant species could be commercially available in 2023 121 joining a number of genetically modified (GM) crops already utilised worldwide 122 .While pest-resistant crops benefit non-target organisms due to reductions in insecticide use 123,124 , herbicide-resistant crops favour the use of herbicides around valuable crops.This extensive use of herbicides eliminates non-cultivated plants around crop fields that are known to be beneficial to pollinators 125,126 .Impacts of other GM crop types, such as abiotic stress-resistant, disease-tolerant, and nutritionally improved crops, have not yet been assessed on managed bees but could pose both a threat and an opportunity.

Concluding remarks
Through the horizon scanning process 21 issues with the potential to impact managed bees in European agricultural systems were prioritised, from an initial 63.These fell under seven broader themes (Fig. A consistent point raised across multiple issues under the theme of Pesticides & pollutants was a current dearth of knowledge on the impact on managed bee populations.Examples include the threat posed by microplastic accumulation and its movement through the food chain, whether the fast-paced emergence of nanotechnologybased pesticides will provide threats or opportunities, or the benefits in transitioning from thermic to electric vehicles.For microplastics, current EU-funded research projects (e.g., www.insig nia-bee.eu) are beginning to quantify their impact on various aspects of managed bee health, and with EU policies in place set to ban all single Vol:.(1234567890 127 , these results will be best placed to inform future monitoring activities.There was also a recognition of the need to support EU pesticide use and risk reduction policies, through recommendations on how to reduce risks from co-formulants and microorganisms used as biopesticides and providing standardised data on the spatial and temporal use of commercial pesticide products across the Member States.Three opportunities prioritised in this scan fell under the theme of Technology.These ranged from remotely monitoring bee health and evaluating genetic markers in the field to the use of artificial intelligence in reducing pesticide use in agriculture.Rapid advancements in biotechnology and available tools are facilitating in-field monitoring and evaluation capabilities, however rapid adoption is key for these tools to be effective in beekeeper practices in real life.
Two issues were prioritised under the theme of Crop modification.The key aspect for both of these issues, which included cutting pollinators out of food production through a shift towards parthenocarpic crops and the uncertainty surrounding newer genomic techniques such as prime editing, is the lack of assessment on the impact on managed bees.
The threat to managed bees from extreme weather events was the only issue to fall under the theme of Environmental stressors.The impacts of well-characterised events, such as heat waves and drought, are already impacting bees and beekeeping practices.However, the potential threat to managed bees from interactions between extreme weather events (including less well characterised events such as frosts) and other stressors (e.g., pesticides and parasites) was recognised as a high priority area for research and should be considered in future policy outlooks.
Several issues resulting from changes to various Management practices were raised through this horizon scan process.Two key opportunities to support managed bee diets were highlighted, these included research-driven bee diet optimisation with the potential to lead to the creation of tailored seed mixes to meet nutritional requirements.These could then be utilised for implementing diverse on-farm floral resources, which has gained further policy support under the sustainability goals of the European Green Deal.In contrast, increases in both inexperienced beekeepers and migratory beekeeping practices were recognised as emerging threats with the potential to impact on managed bee health through higher disease prevalence and genetic introgression.Lastly, uncertainty around the impact of changing farm practices on managed bees was recognised, with both opportunities and threats foreseeable dependent upon the context of the situation and the practices adopted.
The continually changing threat from invasive predators and emerging pathogens across Europe was the most highly ranked issue in this horizon scan and was one of two issues to come under the theme of Predators & parasites.The second was the opportunity around the development of Varroa resistant stocks, with the next few years recognised as a potential turning point for this issue.
Finally, two issues were raised that fell under the theme of Political and trade influence.The European Commission response to recent geopolitical developments, such as the war on Ukraine, was raised here, particularly the uncertainty around the impact on managed bees of short-and medium-term actions aimed at supporting farmers and food security that may negate bee beneficial practices.Alongside the uncertainty of rapid policy changes in response to ongoing geopolitical issues was a recognition of the need to strengthen trade regulations to better protect managed bee populations.
Given the accelerating pace of technology, trajectory for current policy development and geopolitical crises we highlight the need to repeat this exercise in 5 years' time.

Methods
We followed a horizon scanning approach based on a modified Delphi technique and previous horizon scans 10,11 .A core group of 20 experts from nine European countries undertook the scanning exercise.Participants were members of a wider consortium collaborating on the EU-funded project, PoshBee-Pan-European Assessment, Monitoring and Mitigation of Stressors on the Health of Bees (http:// www.poshb ee.eu).Experts were affiliated with research institutes, universities, government and non-government organisations and industry.In this scan, we consider both policy and practice contexts, and issues in the EU, the UK, Switzerland, and Norway.
Each expert was encouraged to consult with their networks to collect up to 5 potential horizon issues.The aim was to identify poorly known issues that could have a substantial positive or negative impact on managed bees (e.g., Apis mellifera, Bombus spp., Osmia spp.) in European agricultural systems over the next 10 years.
Initial submissions that dealt with similar issues were grouped together by topic area and direction of impact (threat or opportunity), to be scored collectively.A list of 63 issues, including references, was compiled, and sent out to the core expert group to complete a first round of anonymous scoring (Table 1).Issues were scored from 1 (well known, unlikely to have a substantial impact on pollinators) to 100 (poorly known, likely to have a substantial impact on pollinators) following the methods adopted by Brown et al. 10 .From this first round of scoring, we produced a ranked list of issues for each participant and then calculated the median rank for each horizon issue (Table 1).The 20 top ranking issues, along with comments and references, were kept as a reasonable number which could be assessed in depth in the next stages of the process.After this initial scoring participants were given the opportunity to retain any issues they felt strongly should have been included.One issue was retained by this process, therefore there were 21 issues in total (Fig. 1; highlighted in Table 1).
Based on their established domain knowledge two experts were assigned to each of the 21 issues to play the role of cynic and to further investigate their novelty, likelihood of emergence, and whether the impact on managed pollinators would be a threat, opportunity, or potentially both.Experts were not assigned to issues they had originally proposed.Experts wrote a short report on their assigned issues that included a summary of the current knowledge and evidence for why it was likely, or not, to be a significant threat or opportunity over the next decade.These reports were then compiled and shared with the group (authorship of individual reports was not revealed to the group) prior to the workshop discussion.To reduce biases due to reader fatigue the order of these short reports in the compiled document was reversed for half the participants.www.nature.com/scientificreports/An online workshop, with 16 experts in attendance, was held in July 2022.Each of the 21 issues was discussed, and following each discussion, experts privately re-scored the issue between 1 and 100, as previously described.The four experts unable to attend the workshop were sent detailed accounts of the discussions that took place and were asked to re-score each issue after reading these accounts.