Over the past decades, both wild and domesticated pollinators have been in dramatic decline (Potts et al. 2010). Biesmeijer et al. (2006) reported parallel declines (pre- versus post 1980) in pollinators (mainly bees, but also hoverflies) and insect-pollinated plants in Britain and the Netherlands. In the US, dramatic widespread decline of bumblebees has been observed: the relative abundances of four previously common bumblebee species have declined by up to 96 %. Their surveyed geographic ranges have contracted by 23–87 %, some within the last 20 years (Cameron et al. 2011). Managed honeybees have also been found to be in decline both in the US (Allen-Wardel et al. 1998) and in Europe (Potts et al. 2010). In other parts of the world such as China and Argentina, the trend moves in the other direction as the number of managed hives is increasing at a pace higher than the number of colony collapses (Goulson et al. 2015). The decline of wild pollinators has since the 1990s forced orchard farmers in southwest China to employ human hand-pollinators, which are markedly less effective and economically unsustainable (Partap and Ya 2012) (with the exception of some fruit varieties where hand pollination has economic advantages, such as some pear varieties where male and female trees do not flower simultaneously). In Australia, strict quarantine is enforced in order to ensure that invasive species (whether competing bees, pathogens or parasites) do not disturb existing honeybee populations, and there is at this point no confirmed report of increases in colony collapses. In Africa, Egyptian beekeepers have reported colony collapses (UNEP 2010). However, few data exist on pollinators in these regions, particularly regarding wild species (Vanbergen and the Insect Pollinators Initiative 2013).
Increasing honeybee disorders, abnormally high winter colony losses, and reduced lifespan of honeybee queens has been observed over the past decades (Van der Sluijs et al. 2013a, Pisa et al. 2015). So far, research points in the direction that no single cause explains the global increase in winter colony losses. All viruses and other pathogens that have been linked to colony collapse are present year-round also in healthy colonies, which implies that the presence of these pathogens alone does not drive the collapse of colonies (Runckel et al. 2011). It seems more likely to be a combination of reciprocally enhancing causes (Goulson 2015; Vanbergen and the Insect Pollinators Initiative 2013). Among those, the large-scale introduction of systemic insecticides, in particular the so-called neonicotinoids has gained more weight (Van der Sluijs et al. 2015; EASAC 2015).
The global decline of unmanaged insect pollinators is part of a larger global catastrophic decline of insects and arthropods in general (Bijleveld van Lexmond et al. 2015). Evidence for decreased crop yields and species diversity due to decreased pollination has not yet been provided on a global scale, but is indicated through regional studies (mostly available for Europe and the US; see e.g., Chagnon et al. 2015). The root causes of the ongoing collapse of the Earth’s pollinators (and entomofauna at large) include the intensification of agriculture with its accompanying loss of natural habitats and loss of foraging and nesting resources, large scale use of agrochemicals such as insecticides, fungicides, herbicides, and fertilizers, nitrogen deposition, climate change, invasive species, spread of pathogens, the manifold increase in roads and motorised traffic, and the continent-wide nocturnal light pollution (Bijleveld van Lexmond et al. 2015; Vanbergen and the Insect Pollinators Initiative 2013). Within this larger context, pollinators deserve particular consideration due to their vital ecosystem function and critical role in ensuring local and global food security. In the following paragraphs, we further explore the key drivers.
Loss of Habitat, Foraging and Nesting Resources
Bees and other pollinators forage on flowers that provide them with pollen (proteins and lipids) and nectar (carbohydrates). Honeybee colonies typically consist of 40.000 bees and worker bees have a lifespan of roughly 4 weeks. This means that a continuous brood cycle from eggs to larvae to bees needs to be sustained over the entire foraging season while a food stock sufficient for overwintering of the colony also needs to be collected. Fresh pollen of sufficient nutritional quality and of sufficient quantity and diversity needs be available throughout the foraging season. This requires a tremendous and continuous influx of fresh pollen into the hive as the main protein resource. A single honeybee hive can visit more than 2 million flowers per day. Quantity and quality of pollen, nectar and water as well as temporal and spatial distribution of these foraging resources are essential for colony survival.
Land use change and landscape changes have drastically reduced the number and diversity of wild flowers, especially in agricultural landscapes (Klein et al. 2007). Where hedgerows and field margins used to be rich in flowering plants (including flowering trees and shrubs) and a diversity of flowering weeds grew between crops (e.g., cornflower, Centaurea cyanus, in corn fields and redshank, Persicaria maculosa, in potato fields), modern large-scale agriculture with massive monoculture and massive use of herbicides has drastically reduced the availability and diversity of wild flowers. The use of genetically modified bulk crops also negatively impact pollinators in several ways. Herbicide tolerant GM crops go hand in hand with massive use of herbicides, which eliminates flowering weeds and weeds that act as host plants from the agricultural landscape. Insect resistant GM crops such as those that produce Bacillus thuringiensis (Bt) toxins may reduce insecticide use and could reduce the pressure on some pollinators (e.g., bees) but can harm other pollinators (e.g., some butterflies).
Depletion of floral resources is further reinforced by large-scale anthropogenic water and air pollution with reactive nitrogen. The present excess of reactive nitrogen has two main causes: agriculture and livestock production (fertilizer and manure), and the combustion of fossil fuels. Atmospheric reactive nitrogen deposits and accumulates in soils, including soils of nature areas. This accumulation of nitrogen in soils drives changes in species composition across the whole range of different ecosystem types. It alters the competitive interactions that lead to composition change and/or makes conditions unfavourable for some species (Bobbink et al. 2010). Nitrogen accumulation affects pollinators because it reduces the diversity and quantity of flowering plants in the landscape.
For solitary bees and other wild bees, micro-habitats that are suitable for nesting occur more in patchy and diverse landscapes than in monotonous landscapes. With massive changes over the past century in land use, landscapes, and agricultural practices, many nesting resources and micro-habitats have disappeared, which has contributed to global insect pollinator decline (Goulson et al. 2015).
Climate change is a further driver of habitat destruction. Climate and hydrology set the general conditions for the occurrence and thriving of wild species. With global warming, the Earth’s climate zones shift pole-wards. Pollinators and the flowering plant species on which they depend that cannot compensate for this or have limited abilities to keep up with the rate of change are at risk (Potts et al. 2010). Climate change also shifts the growing season, meaning that plants start flowering earlier in the season. As a consequence, gaps can occur in the required continuous availability of fresh pollen throughout the foraging season of honeybee colonies.
For managed honeybees in America and Europe, the introduction of the Varroa Destructor mite (an invasive species originally from Asia) and the global spread of bee pathogens that came with the globalisation of trade in bee queens, has caused major problems in beekeeping and bee health. For honeybee declines it is now widely held that the observed trends can be explained by combined stress from parasites, pesticides and lack of flowers (Goulson et al. 2015). Due to the complexity of these causes, there are few and uncertain numbers on how much of the decline is caused by the respective factors.
Large-Scale Prophylactic Use of Systemic Neonicotinoid Insecticides
Neonicotinoids are a new generation of insecticides introduced in the early 1990s (Maxim and Van der Sluijs 2013a). These chemicals have a substantially lower acute toxicity to humans, birds and mammals than the older insecticides which they replaced. However, evidence is mounting that these chemicals play a key role in bee disorders and pollinator decline observed over the past decades (EASAC 2015; Krupke and Long 2015; Rundlöf et al. 2015; Sanchez-Bayo 2014; Van der Sluijs et al. 2013a, 2015; Williams et al. 2015).
Neonicotinoids are widely applied as a coating to seeds of crops or as treatment of soil. The most widely used GM crops are also routinely coated with neonicotinoids. These neurotoxic agrochemicals act systemically: during growth the active substance is taken up by the roots and makes the whole plant toxic to insects for a long period. Unintendedly, neonicotinoids also end up in nectar and pollen, which are the food sources for bees (Botías et al. 2015; Maxim and Van der Sluijs 2013a).
Neonicotinoids are persistent in soil and water, remain in the environment for a long time, and spread quickly through surface water (Bonmatin et al. 2015). Through systemic uptake it also contaminates wild flowers (Botías et al. 2015). In the Netherlands levels of imidacloprid (the most widely used neonicotinoid, introduced in 1994 as an insecticide to coat sunflower seeds with before planting) far in excess of what is considered safe for aquatic ecosystems have been measured continually in the surface water since 2004 (Van Dijk et al. 2013), 1000 to 25,000 times the Maximum Permissible Concentration in the Netherlands of 13 ng l−1. Van Dijk et al. (2013) found that high levels of imidacloprid in surface water consistently correlate to low aquatic insect abundance. Several non-bee insect pollinators have an aquatic larval stage and are thus affected by this. Many insect pollinators, including honeybees and bumblebees, forage on surface water and can thus be exposed to neonicotinoid residues. For honeybees, imidacloprid is more than 7,000 times more toxic than the insecticide DDT (acute toxicity). Furthermore, it gradually becomes lethal to insects as a result of prolonged exposure to extremely low levels (chronic toxicity) and has behaviour-disturbing effects on almost all non-target insect species. In low dose, it disturbs flight behaviour, navigation, brood development and impairs individual and social grooming. Synergistic effects with other agrochemicals have been found (Van der Sluijs et al. 2013a; Pisa et al. 2015).
After their introduction to the market in the early 1990s by Bayer Cropscience, neonicotinoid use grew rapidly to occupy more than a quarter of the world market of insecticides within less than 15 years (Jeschke and Nauen 2008; Jeschke et al. 2011; Simon-Delso et al. 2015; Van der Sluijs et al. 2013a). By 2010, imidacloprid was registered as an insecticide in more than 120 countries (Maxim and Van der Sluijs 2013a). Neonicotinoids are now the most commonly used and fastest growing type of insecticide in the world. In Europe neonicotinoids are authorised for hundreds of crops (Simon-Delso et al. 2015).
The role of neonicotinoids in worldwide honeybee disorders has led to strong controversies (Maxim and Van der Sluijs 2013a). Major declines of honeybee colonies have been reported in France since 1994. Some years later other parts of the world experienced similar sudden colony losses. Colony collapse disorder soon became a global phenomenon, which coincides with the booming world-wide use of neonicotinoids. During the French honeybee crisis in the 1990s beekeepers and scientists involved in public research suspected the neonicotinoid imidacloprid. Representatives from Bayer Cropscience (the producer of imidacloprid) and the French Food Safety Authority (AFSSA) denied a causal relationship between imidacloprid and the honeybee decline, and pointed at many other factors that could also be to blame. A French expert committee (CST 2003) concluded that imidacloprid was indeed likely to be implied in the bee losses and the Minister of Agriculture banned its use in sunflower and maize seed-dressing as a precaution (Maxim and Van der Sluijs 2007, 2010, 2013a). Since then, an ever increasing number of studies with contradictory conclusions appeared, and two camps emerged: (mainly) industry scientists pointing at the Varroa Destructor mite as the main cause and academic scientists pointing at a complex of causes that synergistically reinforce one another, with neonicotinoids as a key factor. The controversy has led to a vehement societal conflict between beekeepers, environmental NGO’s, industry and regulatory agencies (including EFSA and US-EPA) around the globe, with a call for a global ban on neonicotinoids and court cases in many countries. Some countries, such as Italy and Slovenia, decided early on to completely ban the use of neonicotinoids for seed dressing, even while the science was still inconclusive and contested. The EU followed in 2013 with a 2-year moratorium on the use of 3 neonicotinoids in crops attractive to bees (Maxim and Van der Sluijs 2013a; EASAC 2015; the ban is currently being reviewed by EFSA scientists). Fipronil, another systemic neurotoxic insecticide implied in bee disorders and also widely used, was considered for inclusion in the EU moratorium, but was instead subjected to restrictions (EFSA 2015; Pisa et al. 2015).
One element of the controversy concerns the chronic toxicity model. The regulatory science assumes that there is a dose below which no harmful effects will occur. However, an early study by Suchail et al. (2001) found chronic effects on bees at unexpectedly low dose, 1000× lower than in acute toxicity studies. The study was heavily attacked by industry scientists (see Maxim and Van der Sluijs 2013a, b). Recently Tennekes (2010) concluded that the toxicity of neonicotinoids is reinforced by exposure time and postulated a chronic toxicity model based on Haber’s law, originally developed to characterise the toxicity of neurotoxic chemical warfare gases, which states that the prolonged exposure to diminishing concentrations of toxins over time produces a constant (irreversible) toxic effect (ct = constant). In other words: toxicity of such chemicals is reinforced by exposure time: the longer the exposure time, the lower the daily dose required to produce a chronic lethal effect. A lower daily dose thus means a longer time to the mortal effect. Only a (chronic) dose where the time to mortality would exceed the natural lifespan of the insect could be considered non-lethal, but it can still produce sub-lethal harmful effects. The regulatory tests prescribed for chronic toxicity in amongst others the EU and North America are limited to a 10 day chronic toxicity test, which is excessively short in view of Haber’s law and the life span of pollinators and thus does not protect pollinators. Tennekes’s findings are consistent with Suchail et al.’s empirical findings. Tennekes’s paper was criticised by Bayer scientists (Maus and Nauen 2011) but later corroborated by empirical data (Tennekes and Sanchez-Bayo 2011, 2013). While it is still contested by industry scientists, the European Food Safety Authority (EFSA) Panel on Plant Protection Products and their Residues (EFSA 2013) adopted the new chronic toxicity model as valid in their recent Scientific Opinion on the science behind the development of a risk assessment of Plant Protection Products on bees. They did not, however, change the prescribed duration of the chronic toxicity test for bees, which is still 10 days, while the life span of a healthy overwintering worker honeybee can be up to 6 months and the lifetime of a healthy honeybee queen up to 4 years.
Another item of dissent stems from contradicting findings of lab studies and field studies: The present day European protocols for the authorisation of plant protection products give more weight to findings from field studies than to lab studies and semi field studies (Maxim and Van der Sluijs 2013b). This reflects the severe influence of the industry on the regulation (see also Boone et al. 2014). While many lab studies and semi field studies found harmful effects of neonicotinoids to bees at field realistic dose (Maxim and Van der Sluijs 2013a), most of the published field studies did not confirm these effects. Field studies have so far mostly been industry sponsored, and these studies have been heavily criticised for lacking the statistical power to prove absence of effects (e.g., Van der Sluijs et al. 2013a). Many flaws in experimental set-up of field studies used for authorisation have been pointed out (Goulson 2015). In 2010, for example, the US Environmental Protection Agency reclassified the Cutler and Scott-Dupree (2007) field study (on the basis of which the neonicotinoid clothianidin has been authorised in the US, Canada and Europe) as “invalid” because of the severe shortcomings identified in the test set-up, such as too short distance between case and control fields, and too small scale and follow-up time of the experiment. Given that bees forage in a 3 to 9 km radius around the hive, it is almost impossible to design a reproducible field test with sufficient statistical power. Still many countries base the market authorisation of neonicotinoids on the findings of flawed field studies, because the only criterion for inclusion or exclusion is whether the study has a Good Laboratory Practice (GLP) certificate, which is a quality system promulgated by the OECD. GLP has been criticised because it does not address quality of experimental set up, nor does it address the statistical power of the experiment (Maxim and Van der Sluijs 2013b). Regulations such as those found in Italy and Slovenia, as well as the EU moratorium on 3 neonicotinoids in 2013–present, constitute constructive steps, but are far from enough to counteract the local and global trends of pollinator decline.