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.

Stress-induced loss of social resilience in honeybee colonies and its implications on fitness

Stressors may lead to a shift in the timing of life-history events of species, causing a mismatch with optimal environmental conditions, potentially reducing fitness. In honeybees, the timing of brood rearing and nest emergence in late winter/early spring is critical as colonies need to grow fast after winter to prepare for reproduction. However, the effects of stress on these life-history events in late winter/early spring and the possible consequences are not well understood. Therefore, we tested whether (i) honeybee colonies shift timing of brood rearing and nest emergence as response to stressors, and (ii) if there is a consequent loss of social resilience, reflected in colony fitness (survival, growth and reproduction). We monitored stressed (high load of the parasitic mite Varroa destructor or nutrition restricted) colonies and presumably non-stressed colonies from the beginning of 2020 till spring of 2021. We found that honeybee colonies do not shift the timing of brood rearing and nest emergence in spring as a coping mechanism to stressors. However, we show that there is loss of social resilience in stressed colonies, leading to reduced growth and reproduction. Our study contributes to better understanding the effects of stressors on social resilience in eusocial organisms.

B-GOOD: Giving Beekeeping Guidance by cOmputatiOnal-assisted Decision making

A key to healthy beekeeping is the Health Status Index (HIS) inspired by EFSA’s Healthy-B toolbox which we will make fully operational, with the active collaboration of beekeepers, by facilitating the coordinated and harmonised flow of data from various sources and by testing and validating each component thoroughly. We envisage a stepby-step expansion of participating apiaries, and will eventually cover all EU biogeographic regions. The key to a sustainable beekeeping is a better understanding of its socio-economics, particularly within local value chains, its relationship with bee health and the human-ecosystem equilibrium of the beekeeping sector and to implement these insights into the data processing and decision making. We will fully integrate socio-economic analyses, identify viable business models tailored to different contexts for European beekeeping and determine the carrying capacity of the landscape. In close cooperation with the EU Bee Partnership, an EU-wide bee health and management data platform and affiliated project website will be created to enable sharing of knowledge and learning between scientists and stakeholders within and outside the consortium. We will utilise and further expand the classification of the open source IT-application for digital beekeeping, BEEP, to streamline the flow of data related to beekeeping management, the beehive and its environment (landscape, agricultural practices, weather and climate) from various sources. The dynamic bee health and management data platform will allow us to identify correlative relationships among factors impacting the HSI, assess the risk of emerging pests and predators, and enable beekeepers to develop adaptive management strategies that account for local and EU-wide issues. Reinforcing and establishing, where necessary, new multi-actor networks of collaboration will engender a lasting learning and innovation system to ensure socialecological resilient and sustainable beekeeping.

Understanding social resilience in honeybee colonies

Honeybee colonies experience high losses, induced by several stressors that can result in the collapse of colonies. Experiments show what effects stressors, such as parasites, pathogens and pesticides, can have on individual honeybees as well as colonies. Although individuals may die, colonies do not always collapse from such disturbances. As a superorganism, the colony can maintain or return back to homeostasis through colony mechanisms. This capacity is defined as social resilience. When the colony faces a high stress load, this may lead to breakdown in mechanisms, loss in resilience and eventually colony collapse. Before social resilience can be measured in honeybees, we need to examine the mechanisms in colonies that allow recovery and maintenance after stressor exposure. Here, we discuss some of these mechanisms and how they affect the social resilience of honeybee colonies. Understanding social resilience in honeybees is essential to managing colony health and loss prevention.

Varroa destructor infestation impairs the improvement of landing performance in foraging honeybees

The parasitic mite Varroa destructor is an important contributor to the high losses of western honeybees. Forager bees from Varroa-infested colonies show reduced homing and flight capacity; it is not known whether flight manoeuvrability and related learning capability are also affected. Here, we test how honeybees from Varroa-infested and control colonies fly in an environment that is unfamiliar at the beginning of each experimental day. Using stereoscopic high-speed videography, we analysed 555 landing manoeuvres recorded during 12 days of approximately 5 h in length. From this, we quantified landing success as percentage of successful landings, and assessed how this changed over time. We found that the forager workforce of Varroa-infested colonies did not improve their landing success over time, while for control bees landing success improved with approximately 10% each hour. Analysis of the landing trajectories showed that control bees improved landing success by increasing the ratio between in-flight aerodynamic braking and braking at impact on the landing platform; bees from Varroa-infested colonies did not increase this ratio over time. The Varroa-induced detriment to this landing skill-learning capability might limit forager bees from Varroa-infested colonies to adapt to new or challenging conditions; this might consequently contribute to Varroa-induced mortality of honeybee colonies.

Bumblebees land remarkably well in red-blue greenhouse LED light conditions

Red-blue emitting LEDs have recently been introduced in greenhouses to optimise plant growth. However, this spectrum may negatively affect the performance of bumblebees used for pollination, because the visual system of bumblebees is more sensitive to green light than to red-blue light. We used high-speed stereoscopic videography to three-dimensionally track and compare landing manoeuvres of Bombus terrestris bumblebees in red-blue light and in regular, broad-spectrum white light. In both conditions, the landing approaches were interspersed by one or several hover phases, followed by leg extension and touchdown. The time between leg extension and touchdown was 25% (0.05 s) longer in red-blue light than in white light, caused by a more tortuous flight path in red-blue light. However, the total landing duration, specified as the time between the first hover phase and touchdown, did not differ between the light conditions. This suggests that the negative effects of red-blue light on the landing manoeuvre are confined to the final phase of the landing.This article has an associated First Person interview with the first author of the paper.

Short-term, but not long-term, increased daytime workload leads to decreased night-time energetics in a free-living song bird

Reproduction is energetically expensive and to obtain sufficient energy, animals can either alter their metabolic system over time to increase energy intake (increased-intake hypothesis) or reallocate energy from maintenance processes (compensation hypothesis). The first hypothesis predicts a positive relationship between basal metabolic rate (BMR) and energy expenditure (DEE) because of the higher energy demands of the metabolic system at rest. The second hypothesis predicts a trade-off between different body functions, with a reduction of the BMR as a way to compensate for increased daytime energetic expenditure. We experimentally manipulated the workload of wild pied flycatchers by adding or removing chicks when chicks were 2 and 11 days old. We then measured the feeding frequency (FF), DEE and BMR at day 11, allowing us to assess both short- and long-term effects of increased workload. The manipulation at day 2 caused an increase in FF when broods were enlarged, but no response in DEE or BMR, while the manipulation at day 11 caused an increase in FF, no change in DEE and a decrease in BMR in birds with more chicks. Our results suggest that pied flycatchers adjust their workload but that this does not lead to a higher BMR at night (no support for the increased-intake hypothesis). In the short term, we found that birds reallocate energy with a consequent reduction of BMR (evidence for the compensation hypothesis). Birds thus resort to short-term strategies to increase energy expenditure, which could explain why energy expenditure and hard work are not always correlated in birds.

Interaction between Varroa destructor and imidacloprid reduces flight capacity of honeybees

Current high losses of honeybees seriously threaten crop pollination. Whereas parasite exposure is acknowledged as an important cause of these losses, the role of insecticides is controversial. Parasites and neonicotinoid insecticides reduce homing success of foragers (e.g. by reduced orientation), but it is unknown whether they negatively affect flight capacity. We investigated how exposing colonies to the parasitic mite Varroa destructor and the neonicotinoid insecticide imidacloprid affect flight capacity of foragers. Flight distance, time and speed of foragers were measured in flight mills to assess the relative and interactive effects of high V. destructor load and a field-realistic, chronic sub-lethal dose of imidacloprid. Foragers from colonies exposed to high levels of V. destructor flew shorter distances, with a larger effect when also exposed to imidacloprid. Bee body mass partly explained our results as bees were heavier when exposed to these stressors, possibly due to an earlier onset of foraging. Our findings contribute to understanding of interacting stressors that can explain colony losses. Reduced flight capacity decreases the food-collecting ability of honeybees and may hamper the use of precocious foraging as a coping mechanism during colony (nutritional) stress. Ineffective coping mechanisms may lead to destructive cascading effects and subsequent colony collapse.

Acclimation responses to temperature vary with vertical stratification: implications for vulnerability of soil-dwelling species to extreme temperature events

The occurrence of summer heat waves is predicted to increase in amplitude and frequency in the near future, but the consequences of such extreme events are largely unknown, especially for belowground organisms. Soil organisms usually exhibit strong vertical stratification, resulting in more frequent exposure to extreme temperatures for surface-dwelling species than for soil-dwelling species. Therefore soil-dwelling species are expected to have poor acclimation responses to cope with temperature changes. We used five species of surface-dwelling and four species of soil-dwelling Collembola that habituate different depths in the soil. We tested for differences in tolerance to extreme temperatures after acclimation to warm and cold conditions. We also tested for differences in acclimation of the underlying physiology by looking at changes in membrane lipid composition. Chill coma recovery time, heat knockdown time and fatty acid profiles were determined after 1 week of acclimation to either 5 or 20 °C. Our results showed that surface-dwelling Collembola better maintained increased heat tolerance across acclimation temperatures, but no such response was found for cold tolerance. Concordantly, four of the five surface-dwelling Collembola showed up to fourfold changes in relative abundance of fatty acids after 1 week of acclimation, whereas none of the soil-dwelling species showed a significant adjustment in fatty acid composition. Strong physiological responses to temperature fluctuations may have become redundant in soil-dwelling species due to the relative thermal stability of their subterranean habitat. Based on the results of the four species studied, we expect that unless soil-dwelling species can temporarily retreat to avoid extreme temperatures, the predicted increase in heat waves under climatic change renders these soil-dwelling species more vulnerable to extinction than species with better physiological capabilities. Being able to act under a larger thermal range is probably costly and could reduce maximum performance at the optimal temperature.

Fatty acid composition and extreme temperature tolerance following exposure to fluctuating temperatures in a soil arthropod

Ectotherms commonly adjust their lipid composition to ambient temperature to counteract detrimental thermal effects on lipid fluidity. However, the extent of lipid remodeling and the associated fitness consequences under continuous temperature fluctuations are not well-described. The objective of this study was to investigate the effect of repeated temperature fluctuations on fatty acid composition and thermal tolerance. We exposed the springtail Orchesella cincta to two constant temperatures of 5 and 20°C, and a continuously fluctuating treatment between 5 and 20°C every 2 days. Fatty acid composition differed significantly between constant low and high temperatures. As expected, animals were most cold tolerant in the low temperature treatment, while heat tolerance was highest under high temperature. Under fluctuating temperatures, fatty acid composition changed with temperature initially, but later in the experiment fatty acid composition stabilized and closely resembled that found under constant warm temperatures. Consistent with this, heat tolerance in the fluctuating temperature treatment was comparable to the constant warm treatment. Cold tolerance in the fluctuating temperature treatment was intermediate compared to animals acclimated to constant cold or warmth, despite the fact that fatty acid composition was adjusted to warm conditions. This unexpected finding suggests that in animals acclimated to fluctuating temperatures an additional underlying mechanism is involved in the cold shock response. Other aspects of homeoviscous adaptation may protect animals during extreme cold. This paper forms a next step to fully understand the functioning of ectotherms in more thermally variable environments.

A moderate change in temperature induces changes in fatty acid composition of storage and membrane lipids in a soil arthropod

A moderate change in ambient temperature can lead to vital physiological and biochemical adjustments in ectotherms, one of which is a change in fatty acid composition. When temperature decreases, the composition of membrane lipids (phospholipid fatty acids) is expected to become more unsaturated to be able to maintain homeoviscosity. Although different in function, storage lipids (triacylglycerol fatty acids) are expected to respond to temperature changes in a similar way. Age-specific differences, however, could influence this temperature response between different life stages. Here, we investigate if fatty acid composition of membrane and storage lipids responds similarly to temperature changes for two different life stages of Orchesella cincta. Juveniles and adults were cold acclimated (15 degrees C-->5 degrees C) for 28 days and then re-acclimated (5 degrees C-->15 degrees C) for another 28 days. We found adult membranes had a more unsaturated fatty acid composition than juveniles. Membrane lipids became more unsaturated during cold acclimation, and a reversed response occurred during warm acclimation. Membrane lipids, however, showed no warm acclimation, possibly due to the moderate temperature change. The ability to adjust storage lipid composition to moderate changes in ambient temperature may be an underestimated fitness component of temperature adaptation because fluidity of storage lipids permits accessibility of enzymes to energy reserves.

Maximized PUFA measurements improve insight in changes in fatty acid composition in response to temperature

A general mechanism underlying the response of ectotherms to environmental changes often involves changes in fatty acid composition. Theory predicts that a decrease in temperature causes an increase in unsaturation of fatty acids, with an important role for long-chain poly-unsaturated fatty acids (PUFAs). However, PUFAs are particularly unstable and susceptible to peroxidation, hence subtle differences in fatty acid composition can be challenging to detect. We determined the fatty acid composition in springtail (Collembola) in response to two temperatures (5 degrees C and 25 degrees C). First, we tested different sample preparation methods to maximize PUFAs. Treatments consisted of different solvents for primary lipid extraction, mixing with antioxidant, flushing with inert gas, and using different temperature exposures during saponification. Especially slow saponification at low temperature (90 min at 70 degrees C) in combination with replacement of headspace air with nitrogen during saponification and methylation maximized PUFAs for GC analysis. Applying these methods to measure thermal responses in fatty acid composition, the data showed that the (maximized) proportion of C(20) PUFAs increased at low acclimation temperature. However, C(18) PUFAs increased at high acclimation temperature, which is contrary to expectations. Our study illustrates that PUFA levels in lipids may often be underestimated and this may hamper a correct interpretation of differential responses of fatty acid composition.

Traffic mortality and the role of minor roads

Roads have large impacts on wildlife, as they form one of the principal causes of mortality, and disturbance and fragmentation of habitat. These impacts are mainly studied and mitigated on major roads. It is, however, a widespread misconception that most animals are killed on major roads. In this paper, we argue that minor roads have a larger impact on wildlife with respect to habitat destruction, noise load and traffic mortality. We use data on traffic related deaths in badgers (Meles meles) in The Netherlands to illustrate that traffic mortality is higher on minor roads. We ask for a more extensive investigation of the environmental impacts of minor roads. Moreover, we argue that the success of mitigation on roads drastically increases when both major and minor roads are integrated in the planning of traffic flows. Therefore, we propose a strategy based on the concept of a "traffic-calmed area". Traffic-calmed areas create opportunities for wildlife by decreasing limitations for animal movement. We ask for further studies to estimate what size traffic-calmed areas should be to maintain minimum viable animal populations.