Large-scale monitoring of effects of clothianidin-dressed oilseed rape seeds on pollinating insects in northern Germany: residues of clothianidin in pollen, nectar and honey

All bee species rely on pollen and nectar as food sources. In collecting these plant substrates, they serve as economically valuable pollinators of cultivated crop plants and contribute to ecosystem services (Klein et al. 2007). Not only honey bees (Apis spp.) but also bumble bees (Bombus spp.) and mason bees (Osmia spp.) are commercially reared for pollination services and can be more or less ifically targeted to crop monocultures. In addition, various species of bumble bees and mason bees (among many other bee species) are common and widespread in the wild and, to some extent, can also be attracted by crop monocultures. Since agricultural crop plants are commonly treated with plant protection products (PPPs) against insect pests, pollinating insects such as bees may potentially be affected by this treatment. Thus, PPPs contribute to the multiple and varying stressors bees are exposed to and which also include habitat loss, agricultural intensification, parasites and pathogens (Potts et al. 2010). In particular, systemically acting PPPs of the neonicotinoid class of compounds are often held responsible for losses of honey bee colonies and declines in the abundance of wild bees (Sánchez-Bayo, 2014; Goulson et al. 2015; Pisa et al. 2015, Rundlöf et al. 2015). As synthetic nicotine analogues, neonicotinoids affect the nicotinic acetylcholine receptor in the insect brain (for reviews, see: Tomizawa and Casida 2005; Jeschke et al. 2013). Neuroactive neonicotinoids are commonly used as seed dressings in a variety of crops including oilseed rape (OSR). The growing plant absorbs the insecticide, which is distributed to all plant tissues and substrates, including pollen and nectar (Elbert et al. 2008). Because of this systemic activity, neonicotinoids can be applied as a seed dressing or to soil at low rates, which reduces the need for foliar insecticide applications that are applied at much greater rates.

The European Union has temporarily suspended the use of three neonicotinoids (clothianidin, imidacloprid and thiamethoxam) for seed treatment, soil application and foliar treatment in bee attractive crops (European Commission 2013) to allow for in depth studies of their environmental effects. Many laboratory and semi-field studies have provided data about lethal and sublethal effects of neonicotinoids to bees under certain application regimes and at specific concentrations (Godfray et al. 2014, Godfray et al. 2015). However, from an environmental perspective they have been criticised (Cresswell and Thompson 2012; Guez 2013; Carreck and Ratnieks 2014; Godfray et al. 2014, Godfray et al. 2015), e.g., for not using field realistic doses or for subjecting bees exclusively to food spiked with neonicotinoids under laboratory conditions. Because of the basic toxicological principle of the relationship between dose and response, a central question is whether concentrations of neonicotinoid residues in pollen and nectar reach levels that are deleterious to bees under common agricultural practice and landscape conditions.

In 2013, a comprehensive monitoring project was initiated to examine potential effects of clothianidin seed dressing on pollinators under common agricultural practice. This large-scale field study aimed to investigate possible side effects of clothianidin-dressed OSR seeds at the landscape level on various pollinators under actual agricultural conditions (Schmuck and Lewis 2016). This project consisted of four different pollinator studies performed in the project area at the same time: a honey bee monitoring study (Rolke et al. 2016), a mason bee monitoring study (Peters et al. 2016), a bumble bee monitoring study (Sterk et al. 2016), and a residue analysis of pollen and nectar from foraging honey bees in tunnel tents placed over the test crop as well as of pollen, nectar and honey collected by free flying honey bees, bumble bees and mason bees (present study). The studies were conducted in Mecklenburg-West Pomerania (Mecklenburg-Vorpommern, Germany) at two circular study sites of approximately 65 km². Each site surrounded the investigated bee hives and nesting shelters. In autumn 2013, study fields were either drilled with clothianidin-free OSR seeds (reference site, R) or with clothianidin-dressed OSR seeds (Elado^®: 10 g clothianidin and 2 g β-cyfluthrin/kg seed (test site, T)).

The aim of the study part presented here was to provide reliable data on residue concentrations of clothianidin and its active metabolites thiazolylmethylurea (TZMU) and thiazolylnitroguanidine (TZNG) in OSR pollen, nectar and honey produced under common agricultural practice. Accordingly, honey bee and bumble bee hives as well as mason bee nesting blocks were installed at study locations within the reference and test sites. For one approach, bees were allowed to forage freely within their natural flight radius. Pollen collected by all three bee species was sampled and analysed. In addition, nectar collected by honey bee workers and, at the end of OSR flowering, honey from all honey bee colonies were sampled. Data from this field realistic scenario was then compared to data obtained from a worst-case scenario, in which honey bees were forced to forage on OSR only. This was achieved by using a semi-field tunnel tent approach.

The results of this study part serve as basic data for the understanding and interpretation of the results obtained in the bee monitoring studies (Rolke et al. 2016; Sterk et al. 2016; Peters et al. 2016) as well as providing the necessary background for a general discussion on field realistic exposure of bees to clothianidin and its metabolites.

Pollination by bees and insecticide treatments are essential components of modern agriculture. Unfortunately, potential bee toxicity of insecticides may cause problems for managed and wild bee populations (for recent reviews, see: Pisa et al. 2015; Goulson et al. 2015; Johnson 2015). Although the use of systemically acting neonicotinoid seed treatments is generally regarded a more ecologically sound alternative to foliar insecticide applications (Elbert et al. 2008), there is concern of pollinators be exposed to these chemical as they can be translocated into pollen and/or nectar when applied as seed treatments. However, potential negative effects of neonicotinoid seed-treatment of crops on pollinating insects critically depend on the concentrations to which they are exposed to by collecting pollen and nectar. Laboratory-based studies have clearly indicated adverse effects of neonicotinoids to bees at certain concentrations and under artificial exposure regimes (for reviews, see: Belzunces et al. 2012; Godfray et al. 2014; Godfray et al. 2015). However, no significant adverse effects have been detected in a number of field studies (Cutler and Scott-Dupree 2007, 2014; Blacquière et al. 2012; Pohorecka et al. 2012; Pilling et al. 2013; Cutler et al. 2014; Rolke et al. 2016; Sterk et al. 2016; Peters et al. 2016). This discrepancy maybe explained by pollinators experiencing lower PPP doses under field realistic conditions (Carreck and Ratnieks 2014). Therefore, data on neonicotinoid residues in pollen and nectar of neonicotinoid treated crops attractive to pollinating insects collected under common agricultural practice seem to be important for a realistic risk assessment.

This study presents data on residues of clothianidin and its active metabolites TZMU and TZNG in pollen, nectar and honey collected by bees in clothianidin seed-treated winter OSR. The OSR plants at all study fields developed relatively homogenously. Clothianidin, TZMU and TZNG were below LOD in almost all samples taken from the reference site, regardless of the sample matrix or collecting bee species. Thus, it can be assumed that honey bees, bumble bees and mason bees placed within the reference site exclusively foraged within this area free of clothianidin seed-treated OSR.

The species of honey bees, bumble bees and mason bees used in the present study are polylectic and thus collected pollen/nectar may become diluted by material from non-treated plants. Therefore, samples consisting exclusively of winter OSR pollen and nectar were analysed. This was achieved by forcing honey bees to collect pollen and nectar in tunnel tents. The resulting data provide information on the maximum possible concentrations when bees foraged on clothianidin-treated winter OSR only and thus are representative for a worst-case scenario. Average clothianidin concentrations of 1.7 ± 0.8 μg/kg in pollen and 1.3 ± 0.9 μg/kg in nectar were in the same range in both matrices. The maximal concentrations found in single samples were 3.5 µg/kg and 3.6 µg/kg in pollen and nectar, respectively. Thus, the concentrations were in the range of concentrations reported in previous field studies on OSR (Cutler and Scott-Dupree 2007; Cutler et al. 2014; Godfray et al. 2014; Godfray et al. 2015). Residues of TZMU and TZNG could not be detected or quantified in pollen and nectar samples from clothianidin seed-treated OSR even when collected in tunnel tents. This suggests that these two metabolites are only of minor importance for the risk assessment, at least for winter OSR in Central Europe.

Free flying honey bees, bumble bees and mason bees potentially had access to flowering plants other than OSR within the study sites (Heimbachet et al. 2016). Indeed, palynological analyses showed that the mean (±SD) percentage of OSR in pollen samples collected by honey bees at the test site ranged from 12.8 ± 13.8 % to 91.4 ± 2.7 % depending on study location and sampling date (Rolke et al. 2016). Similarly, the percentage of OSR in pollen sampled by bumble bees varied from 16 to 95 % (Sterk et al. 2016). Mason bees provisioned their brood cells with pollen consisting of 10.6 ± 6.8 % and 21.4 ± 13.2 % OSR pollen (Peters et al. 2016). OSR thus became the major pollen source, at least for honey bees and bumble bees during the course of our study (Rolke et al. 2016, Sterk et al. 2016). The percentage of OSR was lower than 100 % in almost all pollen samples taken from free flying bees (note that bumble bee-collected and mason bee-collected pollen samples for palynology and residue analysis were taken on different days). Due to dilution with pollen from other untreated plants, the concentrations of clothianidin have to be lower compared to those samples from the worst-case (semi-field) conditions. Indeed, average concentrations of clothianidin in honey bee-collected, bumble bee-collected and mason bee-collected pollen from the test site remain below LOQ (1.0 µg/kg).

This result is in accordance with other field studies. In a study conducted in Canada, on two experimental fields of spring OSR treated with PROSPER 8 FL and PONCHO 600 FS (delivering clothianidin at 4.0 g/kg seeds), for example, the majority of samples (>75 %) had no detectable level of clothianidin residues and the maximum concentration detected in pollen samples was 2.6 µg/kg (Cutler and Scott-Dupree 2007). In a more recent study, the same authors reported an average concentration of 0.8 ± 0.5 µg/kg in pollen samples collected by honey bees in PROSPER FX^®-treated OSR (Cutler et al. 2014). This would fall below the LOQ when applying our method of residue analysis. Pohorecka et al. (2012) reported similar values for spring OSR treated with MODESTO 480 FS (4.9 g clothianidin/kg seeds) in a honey bee study conducted in Poland. Here, on average 0.6 ± 0.6 µg/kg (mean ± SD) and 2.2 ± 1.3 µg/kg (mean ± SD) clothianidin were detected in pollen loads and bee bread (processed pollen), respectively. In a German bee monitoring study, clothianidin was not detected in any of the 215 samples of honey bee bread collected from 2005–2007, although they included samples from colonies with high OSR input (Genersch et al. 2010).

Similar to residues in pollen, average concentrations of clothianidin in nectar samples taken from free flying honey bees were below the LOQ. Thus, the clothianidin residue levels remained slightly below the level that had been shown as worst-case scenario in tunnel tents. This indicates that free flying honey bees did not use OSR as an exclusive nectar source, but visited other plants for nectar foraging, as well. Since honey bees are polylectic, this is to be expected. The mixing of nectar from clothianidin seed-treated OSR and nectar from alternative non-treated plants explains the reduced clothianidin concentration found under realistic conditions. Nevertheless, OSR appears to be the major nectar source for honey bees, as has been shown by the palynological analysis of spring honey samples (Rolke et al. 2016). For test site colonies, the average percentage of OSR pollen in spring honey samples was 79.6 % ± 7.48 % (mean ± SD) and 77.9 % ± 8.93 % (mean ± SD) for edge and distant locations, respectively (Rolke et al. 2016). Thus, in the present study clothianidin concentrations in nectar samples were slightly lower than residue concentrations reported by Pohorecka et al. (2012), who analysed nectar samples from clothianidin seed-dressed spring OSR in Poland. These authors found 2.6 ± 4.0 µg/kg and 1.3 µg/kg depending on whether nectar was collected directly from rape flowers or nectar flow from combs was used, respectively (Pohorecka et al. 2012). In the Canadian study, Cutler et al. (2014) did not detect clothianidin residues in nectar collected by bees exposed to seed-treated OSR.

Bee larvae (and often adults) of eusocial bee species usually consume nectar in a more or less processed state. In honey bees, specialised workers process the collected nectar by adding glandular secretions and greatly reduce the water content to less than 20 % (Winston 1987). This modification and concentration leads to the production of honey. We analysed honey samples taken from colonies of free flying honey bees at the end of OSR flowering. The average clothianidin concentration was 1.35 ± 0.48 µg/kg in honey samples from test site colonies. Although mixing with nectar from plants other than OSR may have occurred (Rolke et al. 2016), clothianidin concentrations increased during the concentrating process of converting nectar into honey. However, clothianidin concentrations remained below LOQ in 23 % of the samples and the average concentration found was only slightly above LOQ (1.0 µg/kg). In general, clothianidin residues in honey stayed below the maximum residue level specified by the European Union (10 µg/kg, European Commission 2015). Clothianidin was not detected in honey samples in the Canadian study performed by Cutler et al. (2014), while in the Polish study slightly higher concentrations of clothianidin (3.4 ± 1.0 µg/kg) were found (Pohorecka et al. 2012).

In summary, the concentration of clothianidin in plant substrates collected by honey bees, bumble bees and mason bees under field conditions was found to be below 1.0 µg/kg (LOQ) in the majority of cases. Various studies have reported similar or slightly higher clothianidin concentrations in substrates originating from bees foraging near clothianidin seed-treated OSR fields (Cutler and Scott-Dupree 2007; Pohorecka et al. 2012; Pilling et al. 2013; Cutler et al. 2014). However, a recent study conducted by Rundlöf et al. (2015) in Sweden has shown considerably higher residual concentrations of clothianidin in pollen and nectar collected by honey bees and bumble bees in Elado^® seed-dressed spring OSR. These authors reported average clothianidin concentrations of 13.9 ± 1.8 µg/kg (mean ± SEM) and 10.3 ± 1.3 µg/kg (mean ± SEM) for honey bee-collected pollen and nectar, respectively, and 5.4 ± 1.4 µg/kg for bumble bee-collected nectar (Rundlöf et al., 2015). In comparison to our study, the Rundlöf et al. (2015) study clearly differed in at least two parameters. Firstly, spring OSR was planted in contrast to winter OSR in the present study. In the nectar (and pollen) samples from winter OSR fields, lower levels of neonicotinoid residues are usually found in comparison to spring OSR. This has, for example, been shown by Pohorecka et al. (2012) who found the concentration of thiamethoxam to be on average 1.6 µg/kg (in pollen < LOD) and 8.7 µg/kg (in pollen 5.6 µg/kg) in forage (nectar and honey) from winter OSR and spring OSR, respectively. These differences are most likely due to the longer period between planting neonicotinoid seed-dressed winter OSR and rape flowering (>6 months). During this time, partial degradation of neonicotinoids may occur. In contrast, only about 2 months usually pass from planting to flowering of spring OSR. Secondly, the seed density in the Rundlöf et al. (2015) study (7.7 kg OSR seeds/ha) was about twice as high as in the present study (3.4 ± 1.1 kg OSR seeds/ha in test fields; Heimbach et al. 2016). Together, these factors result in a considerably higher clothianidin load per ha and may have led to the higher residue values found in plant substrates in the Swedish study. In addition, the numbers of pollen and nectar samples analysed in the Rundlöf et al. (2015) study appear to be small (n = 5 honey bees per field for pollen sampling, n = 3–5 honey bees and bumble bees per field for nectar sampling). Due to high variation of residue concentrations in different parts of the same field, this may lead to a reduced discriminatory power. Nevertheless, the design of the present monitoring study followed the agricultural practice typical for winter OSR in the study area. Winter OSR is economically more important than spring OSR, e.g., the proportion of the cultivated area of winter OSR vs. spring OSR in the European Union in 2014 was 93.5 % and 99.7 % in Germany in 2015 (EUROSTAT 2016).

In conclusion, the residual concentrations of clothianidin and its active metabolites TZNG and TZMU in analysed samples of pollen, nectar and honey were clearly lower than the reported no observable adverse effect concentration for honey bees of 20 µg/kg, derived from feeding experiments using spiked diets (Schmuck and Keppler 2003). The clothianidin concentrations were also lower than those used in most laboratory-based studies that have reported adverse effects of clothianidin on bees (Godfray et al. 2014; Godfray et al. 2015; Thompson and Miles 2015). Accordingly, no detrimental effects on reproduction and health parameters on honey bees, bumble bees and mason bees exposed to clothianidin seed-treated OSR could be observed within the conditions of the present monitoring studies (Rolke et al. 2016; Sterk et al. 2016; Peters et al. 2016). Thus, it can be concluded that clothianidin seed-treated winter OSR, when used as directed, provides a favourable margin of safety for pollinating bees.