Written evidence submitted by Imperial College London (INS0012)
House of Commons Science, Innovation and Technology Committee inquiry ‘insect decline and UK food security’ submission
Explanatory note:
This response has been developed by Imperial colleagues from across the College’s expertise. A full list of contributors is included at the foot of the document.
1.1. How well do we understand declines in insect abundance?
Despite the integral ecological and economic significance of insects in wild and human-modified ecosystems, currently it is difficult to say where, when and to what extent species-specific population declines have occurred. Referring to the statement made in the committee’s announcement “Insect numbers are difficult to quantify, however, recent research suggests that in the UK flying insects have declined by 60% in the past 20 years”, we assume this is referring to the Technical Report provided by the Kent Wildlife Trust and Buglife. This report concludes that the number of insects killed on car number plates in 2004 versus 2019 (Kent area only) and 2021 has dropped dramatically. The results reveal an important and concerning trend, and align with declining insect biomass in Germany between 1989 and 2016. However such studies are unable to provide data on the resolution required to understand what taxonomic groups we are losing (or even gaining). This is not a direct criticism of the research – the work should be highly commended and applauded – but it does highlight where key knowledge gaps lay. For instance, from such study approaches we do not know if the decline is being driven by the loss of insect pests, which could reflect a beneficial trend.
Ultimately, the lack of key insight into how taxonomic or key functional groups (such as pollinators or natural enemies of crop pests) are responding means we are limited in our predictive power to how species assemblages will change under continuing changes in land-use, and environmental and climate conditions. Therefore, it is important to fund long-term and nationwide projects to sample insects using highly standardised approaches. Projects such as the EU Pollinator Monitoring Scheme (EU PoMS) Scheme and BIOSCAN are important flagship projects aiming to fill this gap. However, methodological developments are still needed to reduce lethality of sampling (i.e., assessing abundance without killing insects in the process) and reduce taxonomic biases (e.g., we have far more long-term data on butterflies and bumblebees than other insect groups) when collecting abundance data. This could be achieved through developments passive remote sensing technologies, such as acoustic devices with machine learning algorithms. Further, we need to ensure rigorous sampling efforts to avoid missing rare or hard to observe species, as this can make some species appear rarer than they actually are.
1.2. Benefits but also current limitations of citizen science
The number of Citizen Science (CS) projects are rapidly increasing primarily due to improved citizens’ access to digital technologies (e.g. smart phones and associated apps) for data recording, sharing, and verification. It enables scaling-up existing projects, while at the same time repositioning CS from its original understanding of being largely a voluntary scientific endeavour to becoming a tool for monitoring progress and engendering action in the pursuit of global sustainability agendas. The scale at which insects can be monitored would ideally be nationwide over many years. Some of the best examples focus on butterflies and bumblebees and include schemes such as the Big Butterfly Count and Bee Walk. For these specific insect groups, we have a better understanding of which species are expanding or contracting their distributions across UK. There is, however, still much room for improvement when considering the accuracy, resolution, and biases in CS data:
It is also important to improve engagement and develop educational tools. CS methods for large-scale data collection may risk limiting and constraining the types of scientific contributions and inquiries citizens can make, reducing them to data recorders. A recent project called X-Polli:Nation seeks to address this issue by focussing on young people as the next generation of actors to lead change. It brings together the different dimensions of CS to involve young people through primary school education to both leverage participation but also educate children on the insect pollinator crisis whilst learning about Sustainable Development Goals.
1.3. Peer-reviewed research points to general insect declines but are largely based on occupancy modelling
With the increasing access to public data repositories of insect observations, there have been important developments in so-called ‘occupancy modelling’ to better understand how changes in the distribution of species may be related to environmental change. But these models are primarily based on end-point metrics of whether a species (or taxonomic group) is present or absent from a given square grid (e.g., 10 x 10 km spatial grid of the UK). Efforts are being made to consider the number of observed recording as a proxy of abundance whilst accounting for sampling effort, but occupancy modelling is limited to what it can tell us about population dynamics and therefore how insect abundance has changed. It is also insufficient to understand how species and communities temporally respond and cope with newly emerging environments. Studies that have sought to use abundance data when available are typically focussed at the community as opposed to population level, and deal with ad-hoc occurrence records. Observational records or sampled specimens are typically also taken in an unstructured way - with biases in space, time, and detectability - rather than using standardised protocols.
The primary issue is that standardised historic abundance data does not exist at the spatiotemporal and taxonomic resolution required. Moving forward we therefore need to setup nationwide networks of monitoring sites for all types of insect functional groups. Such data collecting schemes, however, must follow standardised protocols that represents a broad set of taxonomic groups across the insect phylogeny. With this type of data, we can then look to ground truth how well past archival data can be modelled to accurately predict and forecast population responses to environmental factors as well as understanding natural population dynamics and cyclical patterns.
1.4. Responses are rarely uniform across species
Revealing general rules on insect population responses to environmental stress factors, is no doubt important. Qualitatively, we can be confident that factors such as land-use change, pesticide use, emergent diseases, climate change are placing most beneficial insect populations at risk. Indeed, a growing number of studies show that even quite closely related species are responding quite differently to recent decades of environmental change. For example, for UK butterfly species (within the same genus) and bumblebee species, some species are seeing geographic distribution contractions and some expansions as is also being seen in bumblebees. Furthermore, we can even see differences between the different life stages of a single species in response to the same environmental variables, such as habitat loss or climate. A recent study on Arctic arthropods echoes this, showing complex nonlinear responses in abundance and diversity between different taxonomic groups to climate change. It is important, therefore, that research continues to investigate to what level of taxonomic or even life-stage resolution does data need to be collected if we are to understand general rules underpinning insect population declines and increases. This will enable up-scaling of decision making to support insect across large landscapes and wide geographic areas (e.g., nationally).
1.5. Understanding what insect functional diversity is being lost may be more important than describing species diversity loss
Currently, our ability to predict the consequences of insect declines on community function (for instance, such as pollination or pest predation in food webs) is surprisingly limited. We do know that increases in the diversity of species can lead to increased function, such as greater pollination and seed set of crop plants. We also know that manipulative studies removing whole species or functional guilds (e.g. removing a subset or all pollinators) can affect ecosystem services. But population declines rarely lead to sudden loss of species from a localised area but rather a slow and sometimes erratic decays. Therefore, understanding the functional redundancy in the system is important. We also need to conduct ambitious empirical studies that can look at large communities of insects and the interactions to understand real-world implications for food security. One approach to link population declines to changes in community function is to quantify the frequency distributions of functional traits (such as, morphology or key behaviours) of the species composing the community and how the functional diversity (the breadth of functional traits) of the community is changing. However, we currently understand little about how functional traits are being lost [or even gained]. For example, how has the rapid rate of environmental change over the past century selected for species or individuals with specific functional traits? Further, has homogenisation of modern landscapes from changing land use led to ecosystems becoming more composed of species with similar functional traits? If so, this could mean a loss of functional diversity, meaning agroecosystems may be less resilient to stressful events (like a drought) in the future.
1.6. Understanding how insect genetic diversity is being lost is important
Approaches such as occupancy modelling are limited to what they can say about demographic signatures which are important if we want to understand early warning signs of population crashes. A central problem here is that we have a limited understanding of how environmental change affects population size, structure, and genetic diversity of resident species. Addressing this is important, as decreased size and diversity can increase vulnerability to extreme events and lower population tipping point thresholds, and identifying when significant population reduction occurs could identify which anthropogenic actions contributed). Changes in genetic diversity may also represent either loss of adaptive potential, or possibly evidence of adaptation. The rate at which populations respond to newly emerging environments in the wild remains unclear but determining the levels of genetic diversity could explain why some species cope better than others (based on local species richness and sightings), resulting in so called ‘winners and losers’, which can alter ecological community. The current era of next generation sequencing enabling high-throughput and low costs sequencing of whole genomes or metabarcoding of whole communities means that we can feasibly look to understand genetic diversity across the insect tree of life. Furthermore, being able to apply population genomics to our museum specimens means that just a few individuals provide a good representation of the population demographic at the time of both historic and contemporary entomological collections.
1.7. Accurate forecasting relies on historic baseline data and understanding lag-effects from threats
The UK has some of the best data on historic insect observations across the world, yet we have only partially exploited such a resource effectively. Despite the success of the Technical Report provided by the Kent Wildlife Trust and Buglife (described in section 1.1), nuance can be lost when modelling population trends based on two single timestamps (e.g., 2004 vs 2021). Factors such as weather conditions can cause significant variation in in abundance between consecutive years, so a single year may not be representative of straddling years. Whilst there is strong evidence of the decline of insects over the past century or more, care must be taken to interpret this evidence carefully accounting for historical context. For instance, moth biomass over a 50-year period has both increased and decreased across Britain, showing that longer term trends may be non-linear and thus relying on shorter term linear results may be unreliable if wanting to forecast to the end of this century. Thus, deeper baseline data is needed for future modelling to improve the accuracy of predictions and projections, and to better understand uncertainties of any models. To deepen this baseline data, we should continue to exploit the abundance of data in repositories such as Global Biodiversity Information Facility (GBIF), but crucially couple this with what is currently underexploited public and private museum collection specimens and associated metadata.
An additional challenge when relying on recently collected data (i.e., post-1970) to model changes in insect abundance, is that the effects of environmental change on insect abundance may only appear after longer time horizons. This “extinction debt” means the number of effectively threatened species could be underestimated. Further, the existence of lag effects and extinction debts make it difficult to pinpoint exactly which environmental stressor (pesticide use or climate change, for example) contributed most to a change in insect abundance. Here, again, data from museum specimens and long-term monitoring schemes – especially when collected at high spatiotemporal resolution – can provide vital historical insight into the specific factors that place insects at risk.
1.8. Quantifying multi-stressor interactions on insects will provide a more accurate framework for understanding and projecting insect responses
Our ability to predict and project when and where beneficial insects are at risk remains limited. This is primarily down to our poor understanding of how co-occurring stressors interact to affect individual insects and how this translates to population impacts. Even for stressors that individuals frequently and simultaneously experience in the field such as pesticides and daily/seasonal changes in temperature, we have only scratched the surface on quantifying how these interact impact individual. It is therefore unclear whether we are under- or over-estimating the risks of key global-change factors, undermining our ability to map how ecosystem services (like crop pollination) will be impacted. Furthermore, if we are to effectively develop safeguarding strategies, we must investigate insect responses to environmental change spanning multiple biological levels (from changes in genetics to their roles within ecosystems) under single study frameworks. Designing projects to gain a holistic understanding of responses is urgently needed, as it will provide a mechanistic understanding of dynamic responses of organisms in the wild. Critically, this will revolutionise i) predictions of spatiotemporal responses (across which landscapes or climate regions and over what periods of the day or season), ii) application of findings across taxa, and iii) projections of responses across large geographical scales, iii) understanding of impacts on crop health.
2.1 Growing body of evidence indicates pesticides are a key contributor to beneficial insect declines
The threat of pesticides to beneficial insects is a global issue. While acute exposure to field realistic concentrations of contemporary pesticides is typically non-lethal, exposure can impair behaviours underpinning organism fitness and service provision. Quantifying such ‘sublethal’ effects on organismal function and behaviour is thus considered a more reliable metric to assess pesticide impacts on natural populations. Work on the effects neonicotinoids on non-target insects, such as pollinators and primarily bees, has been conducted by many groups across the world on different end-point responses. But here we list work conducted by the Gill group and collaborators on bumblebees (an important wild pollinator representative, and a growing lab model system) which spans effects from the molecular level through to whole colonies, showing that sublethal exposure to neonicotinoids can affect:
Such sublethal effect have the potential to scale-up to country-wide population effects, which is supported in the research by Woodcock et al. This study, however, revealed that pesticide effects can substantially differences between countries highlighting the importance of understanding how different nationwide stressors interact (see section 1.8, 2.2 and 2.3). Furthermore, pesticide toxicity assessments have been done on a relatively small number of insect species and focused primarily on insect pests or a handful of insect pollinators (particularly honeybees). However, it is possible that species in different positions across the insect phylogeny may respond differently. Thus, with respect to understanding non-target effects on pollinators, moving from beyond honeybee-centric pesticide risk assessments to protect all pollinators is likely to be important.
2.2 Understanding context dependency of pesticide risk is key
Reported toxicity levels on insect sublethal endpoints has shown pronounced variation between studies – even when testing the same species and pesticide(s). Determining why such variation exists is critical to improving our understanding of pesticide risk to beneficial organisms, explaining past trends, informing safeguarding and policy decisions, and using pesticides sustainably. A likely explanation for such conflicting results is that confounding factors have not been appropriately accounted for, in which pesticides have interacted with other stressors to alter toxicity. To reveal the ‘true-risk’ a pesticide poses, we must therefore consider how the environmental context at the time of exposure influences an organism’s response.
2.3 Will climate change alter the risk of pesticides to pest and beneficial insects?
Environmental temperature universally governs the “pace of life” by determining rates of chemical and physical processes. Yet, environmental temperature is also highly variable and globally increasing under climate change. For insects, maintaining internal temperature is challenging under changing external temperatures, affecting organism function at different levels of organisation, including gene function, energy metabolism, pesticide breakdown, and behavioural performance. Yet, how temperature modulates pesticide toxicity on sublethal endpoints of exposed non-target insects has been largely overlooked. Furthermore, how such effects at lower levels of biological organisation translate to detectable effects on higher whole-organismal responses is little understood.
A recent study from the Gill group investigated how changing temperature can alter the risk of a neonicotinoid and sulfoximine, independently, to six different behaviours of bumblebees. Four of the behaviours – responsiveness, likelihood of movement, walking rate, and food consumption rate – were affected by the neonicotinoid more strongly at the lower temperature, suggesting cold snaps could increase pesticide toxicity on behaviours important for nest duties. However, a key behaviour – how far the bees could fly – was most strongly affected by the neonicotinoid at the highest temperature. The drop-off in flight performance at the highest temperature suggests a ‘tipping point’ has been reached in the bees’ ability to tolerate the combined temperature and pesticide exposure. The seeming cliff-edge effect happened over the span of just three degrees, which changes our perception of pesticide risk dynamics given such temperature changes can commonly occur over the space of a day. This work can help to inform the right concentrations and application times of pesticides across different climatic regions of the world to help safeguard pollinators, such as bees.
2.4 How do impacts on insect populations (such as pollinators) translate to large scale deficits in service delivery?
The pollination service provided by pollinating insects is essential for ecosystem functioning and food production, and in the absence of pollinators yields of many crops are substantially limited by ‘pollination deficits’. This is where flowers receive insufficient pollen to product fruits or seeds. Therefore, any impacts on insect pollinators (as described above) have the potential to expand this deficit with huge implications for the economy and human welfare. Yet, to date, understanding how pesticide induced impairment on beneficial insect behaviours, such as pollinators, can affect crop yield has been little attempted. One of the first studies to do this was by Stanley et al. Their findings showed that whilst have neonicotinoid exposed bees can translate to changes in crop seed set, bees exposed to a low concentration improved seed set on average (but not-significantly), and so whilst bees exposed to a high concentration did correlate with lower seed set this was only when compared to the low concentration treatment and not the control.
As our dependence on pollinating insects increases (e.g., there is growing demand for pollinator-dependent crops such as orchard fruits and oilseed), pollination deficits are increasingly likely to impose yield limits and increase costs of food production. Yet, we currently lack methods to effectively identify and track pollination deficits and predict the associated yield outcomes. Therefore, the extent to which the UK food system is pollinator-dependent is limited, and farmers of pollinator dependent crops are unable to mitigate deficits such as through targeted introduction of managed pollinators. Current exploratory research [Gill group] has identified the potential for remote sensing technology to allow monitoring of pollination deficits at the field level. However, further investment will be necessary to elucidate the pollinator dependence of the UK food system, and model predicted outcomes of pollination deficits under changing pollinator population scenarios.
3.1 How do we define a reservoir population and what spatiotemporal scales should be considered?
It is critical we understand what size of habitat areas are required to support sustainable insect populations. This will vary between taxa and thus we need to take a community-wide approach. Whilst field flower margins may provide support, it is not always clear as to whether such schemes are merely attracting insects or capable of retaining healthy populations. Attracting insect pollinators or natural enemies to these habitats can provide pollinator and pest control benefits in the short term, but without nesting sites or habitats that last for the necessary seasonal periods, stable populations may not be able to survive. Larger, whole landscape management schemes are therefore likely to be required to help improve connectivity between habitats and across farmland. This will aid in providing a variety of landscape needed for supporting different life stages of insects (particularly for insects that have different larval and adult phases - such as caterpillars and butterflies from the same species having different host plants) and create corridors for connecting populations. In doing so this will better account for scale dependency across different taxa.
3.2 How should landscapes be managed / organised to support stable insect populations?
Land uses, such as intensive farming, can negatively impact insect biodiversity. Farm subsidies have to date largely promoted wildlife on farms through a ‘land sharing’ approach – lower-intensity farming combined with wildlife-friendly practices providing many small areas of habitat to support insect such as pollinators. There are several UK agri-environment schemes (giving farmers financial incentives to implement environmentally beneficial practises), including the Sustainable Faming Incentive, Countryside Stewardship, England Woodland Creation Offer, Nature for Climate Peatland Grant Scheme, and Farming in Protected Landscapes programme. However, ‘land sparing’ - high intensity farming to maximise yields, with the remaining area used for biodiversity conservation farming on smaller areas of land – could be more cost-effective and beneficial for biodiversity. Indeed, land sparing generally has shown better outcomes for biodiversity than land sharing, both abroad and in the UK, even when imperfectly implemented. Further research is needed to see how these two approaches can support better taxonomic and functional insect diversity. For example, suitable insect monitoring schemes alongside the planned Landscape Recovery programme could achieve this. This may also support larger-scale and longer-term landscape restoration projects that can benefit organisms up the food chain (such as insectivorous birds) and enabling more stable reservoir populations for insect pest control.
3.3 Empowering farmers for monitoring and landscape management to support beneficial insects
Farmers already carry out extensive monitoring regimes to optimise their production, however, they do not see the benefits of beneficial insect monitoring. They have been shown to have the propensity to include pollinators in their monitoring regimes if the benefits are made clear to them, or as part of a wider scheme, however, current methods are too time consuming to implement. Indeed, current recommended pollinator monitoring strategies are limited to field surveys and pan-trapping (UN-FAO guidelines: https://www.fao.org/3/i1929e/i1929e00.htm) which are time consuming, require expert knowledge, and provide limited insight to true functional pollinator population health. It appears important, therefore, to provide better guidance: i) on which crops most benefit from insect supporting schemes, with potential economic values attached; ii) what habitats can support what insects and which functional groups (e.g., pollinators, natural enemy of pests, are most needed, decomposers, etc). Breeze et al (2021) highlight the importance of cost-benefit analysis to show that national monitoring schemes can be cost effective, and farmers need to have incentives. Schemes must therefore be coupled with technological advancements, whereby development of in-hand tools can help farmers to easily and more automatedly monitor insect population responses and the services they deliver. For long-term sustainable farmers adoption of insect supporting schemes, it must be made clear to farmers that efforts are working (requires a metric of success) and that it is increasing profits with decreased inputs.
3.4 Moving towards self-sufficient farming requires nature-based solutions
The UK Food Recruit report 2021 [Theme 2: UK Food Supply] discusses the levels of self-sufficiency in the UK agricultural industry. It mentions that whilst we are roughly self-sufficient in grains, for horticultural products (fruits and vegetables) we rely heavily on importation with the production in the UK declining. Oilseeds, orchard fruits, and legumes are pollinator-dependent crops, and with 21st century projections suggesting increased mean temperatures alongside higher demands for meat-free protein sources and healthier diets (UN SDGs G3&12), there is an increasing dependency on insect pollinators. It is therefore of integral importance that we can support future farming by safeguarding and supporting healthy wild insect pollinator populations. Studies have shown that reliance on managed pollinators such as honeybees is not always effective and can place financial loads on farmers. Moreover, by managing wild areas that can help support insect pollinators (as reservoirs, see section 3.1) we can also harbour other beneficial insects such as natural enemies of crop pests. This nature-based solution can increase yield through increased pollination whilst lowering reliance on and costs of pesticides which will also help safeguard the pollinators.
27 April 2023
Contributors:
Dr Richard Gill
Senior Lecturer, Department of Life Sciences and leader of a research group investigating the threats to insect pollinators (primarily bees), and the implications for population responses and functional roles across landscapes.
Aoife Cantwell-Jones
3rd year PhD student who studies how population dynamics and functional diversity can determine plant-pollinator interactions.
2nd year PhD student who uses museum specimens to determine morphological responses of bees and butterflies to a century of environmental change.
2nd year PhD student who is studying how we can improve food production by better assessing plant pollination deficits across agricultural landscapes to inform pollinator management.
Jacqui James
Masters student investigating the affects of climate change on pollinator populations and the implications for host plant health.
--
Further information:
This submission was collated by the Imperial Policy Forum, working in partnership with Imperial College London’s Department of Life Sciences based in Silwood Park.