Aquatic‐Terrestrial Insecticide Fluxes: Midges as Neonicotinoid Vectors

Abstract

Exposure of freshwater ecosystems to insecticides can negatively impact the development of emerging aquatic insects. These insects serve as an important nutritional subsidy for terrestrial insectivores. Changes in insect emergence phenology (i.e., emergence success and temporal pattern) or fluxes of insecticides retained by the emerging adults have the potential to negatively impact terrestrial food webs. These processes are influenced by contaminant toxicity, lipohilicity, or metabolic processes. The interplay between emergence phenology, contaminant retention through metamorphosis, and associated contaminant flux is not yet understood for current‐use insecticides. In a microcosm study, we evaluated the impacts of a 24‐h pulse exposure of one of three current‐use insecticides, namely pirimicarb, indoxacarb, and thiacloprid, at two environmentally realistic concentration levels on the larval development and emergence of the nonbiting midge Chironomus riparius. In addition, we measured insecticide concentrations in the larvae and adults using ultrahigh performance liquid chromatography coupled to tandem mass spectrometry by electrospray ionization. Exposure to pirimicarb delayed larval development and emergence, and exposure to indoxacarb reduced emergence success. The neonicotinoid thiacloprid had the greatest impact by reducing larval survival and emergence success. At the same time, thiacloprid was the only insecticide measured in the adults with average concentrations of 10.3 and 37.3 ng/g after exposure at 0.1 and 4 µg/L, respectively. In addition, an approximate 30% higher survival to emergence after exposure to 0.1 µg/L relative to a 4‐µg/L exposure resulted in a relatively higher flux of thiacloprid, from the aquatic to the terrestrial environment, at the lower exposure. Our experimental results help to explain the impacts of current‐use insecticides on aquatic–terrestrial subsidy coupling and indicate the potential for widespread dietary exposure of terrestrial insectivores preying on emerging aquatic insects to the neonicotinoid thiacloprid. Environ Toxicol Chem 2023;42:60–70. © 2022 The Authors. Environmental Toxicology and Chemistry published by Wiley Periodicals LLC on behalf of SETAC.

INTRODUCTION

Globally, freshwater ecosystems are at risk of contamination with a wide range of insecticides originating from agriculture (Ippolito et al., 2015; Stehle & Schulz, 2015). This includes both legacy insecticide classes that are no longer permitted for application as well as newer classes of current‐use insecticides (McKnight et al., 2015; Wolfram et al., 2018). Newer classes of current‐use insecticides were developed to have a reduced negative impact on the environment by being less persistent and having a more selective toxic mode of action compared with older classes (Carvalho, 2017). In the case of neonicotinoids and pyrethroids, selectivity was improved by lowering vertebrate toxicity while increasing invertebrate toxicity (Morrissey et al., 2015; Schulz et al., 2021). Alternatively, insecticides with a more selective toxicity (i.e., at the insect order level) have also been developed (Jeschke, 2016; Wing et al., 2000). Neonicotinoid insecticides, however, break away from the trend toward reduced environmental persistence and are found in freshwater ecosystems globally (Morrissey et al., 2015). The ubiquitous presence of insecticides in freshwater ecosystems, even for brief exposure periods during run‐off or spray drift events, has the potential to negatively impact linked aquatic–terrestrial food webs through emerging aquatic insects (Bundschuh et al., 2022; Kraus et al., 2021; Kraus, 2019; Schulz & Liess, 2001; Tooker & Pearsons, 2021).

Up to now aquatic–terrestrial contaminant transfer has been reported for several classes of contaminants, including metals (Chételat et al., 2008; Wesner et al., 2017), metal‐based nanoparticles (Bundschuh et al., 2019), pharmaceuticals (Previšić et al., 2021), halogenated organic pollutants (Liu et al., 2018), and some fungicides and herbicides (Roodt et al., 2022). In the terrestrial environment, these contaminants can enter the food web and potentially result in detrimental effects on consumers, such as prey avoidance or sublethal effects (Koch et al., 2020; Kraus et al., 2014, 2021; Richmond et al., 2018). Furthermore, exposure to insecticides during the development of emerging aquatic insects can impact the terrestrial recipient food web through changes in insect emergence phenology and productivity. Several studies have shown aquatic insect emergence to be either reduced or the temporal emergence pattern to be altered following larval insecticide exposure (Palmquist et al., 2008; Schulz & Liess, 2001; Tada & Hatakeyama, 2000). Both effects may lead to a de‐coupling of aquatic resource provision and terrestrial resource requirements (Marczak & Richardson, 2008). This can decrease the availability of a high‐quality food source supporting critical lifecycle stages in terrestrial consumers, such as breeding by insectivorous birds (Twining et al., 2018), or place additional pressure on species of high conservation value, such as insectivorous bats (Sirami et al., 2013; Stahlschmidt et al., 2012).

Contaminant transfer through metamorphosis can occur for those contaminants previously bioaccumulated in aquatic larval stages. Bioconcentration, which is defined as the bioaccumulation of contaminants through only passive absorption from the surrounding water, increases with increasing lipophilicity, that is, increasing log octanol/water partition coefficients (logK_ow), for mid‐polarity pesticides (logK_ow 2–5) in aquatic larvae and nymphs (Katagi & Tanaka, 2016). Once accumulated, retention of fungicides, herbicides, and other mid‐polarity contaminants through metamorphosis is generally negatively correlated with lipophilicity (Kraus et al., 2014; Roodt et al., 2022). For those contaminants which are retained by emerging insects, the contaminant flux is defined as the weight of contaminant transferred by emerging insects per unit area during a unit of time. Contaminant flux is therefore predicted to decrease with increasing toxicity of a contaminant to aquatic life stages due to decreased survival and emergence (Kraus, 2019). Fluxes of insecticides, including highly toxic neonicotinoids, have recently been reported from contaminated wetlands corresponding with reduced overall insect emergence (Kraus et al., 2021). However, there is a lack of information on the interplay of exposure concentration, toxicity, emergence phenology, and contaminant retention through metamorphosis on the insect‐mediated flux of current‐use insecticides.

Against this background, we conducted microcosm experiments in which we exposed 10‐day‐old larvae of the nonbiting midge, Chironomus riparius, to a 24‐h pulse of one of three current‐use insecticides at two environmentally realistic concentrations. All three insecticides, namely the selective carbamate pirimicarb, the oxadiazine pro‐insecticide indoxacarb, and the neonicotinoid thiacloprid, have seen large‐scale application in Europe in the last decade (Helbig, 2019). We measured endpoints related to insect development and emergence (i.e., larval mortality, emergence success, sex‐specific changes in development time, and adult body wt) as well as insecticide concentrations in the larvae and the adults as an indication of insecticide flux. Our hypotheses were based on the order of increasing insecticide logK_ow values (thiacloprid 1.26 < pirimicarb 1.7 ≪ indoxacarb 4.65) and increasing toxicity to the aquatic larvae (i.e., 28‐day no observed effect concentrations [NOECs]: pirimicarb ≫ indoxacarb > thiacloprid). We hypothesized that the bioconcentration of the insecticides in the aquatic larvae would increase with increasing insecticide lipophilicity. Retention of these accumulated insecticides across adult metamorphosis would decrease with increasing lipophilicity, similar to other contaminants within this logK_ow range (Kraus et al., 2014; Roodt et al., 2022). Finally, that overall insecticide flux would increase with exposure concentration and decrease with increasing toxicity to the larvae, that is, the highest concentration of pirimicarb would have the greatest flux and the highest concentration of thiacloprid would have the lowest.

MATERIALS AND METHODS

Chemicals and reagents

Analytical standards for all the pesticides were obtained from LGC Standards or HPC Standards. Solvents (liquid chromatography coupled to mass spectrometry [LC‐MS] grade) were purchased from Honeywell. All other chemical reagents (>99% purity) and clay were purchased from Carl Roth. Aquarium sand and peat were obtained from Schicker Mineral and Floragard, respectively.

Microcosm experiments

The nonbiting midge, C. riparius (Meigen), was used in three sequential microcosm experiments for exposure to pirimicarb, indoxacarb, or thiacloprid. In each case, freshly laid egg masses were collected from an in‐house laboratory culture. Egg masses (n = 9) were equally distributed between three aquaria (32 × 22 cm, 3.75 L) containing 2–3 cm of sediment, which was prepared in accordance with Organisation for Economic Co‐operation and Development (OECD) test guideline 219 (OECD, 2004), and SAM‐5S aqueous medium of depth approximately 5 cm (Borgmann, 1996). The aquaria were kept in a climate‐controlled chamber at 20 °C with 70% relative humidity and a 16:8‐h light:dark cycle. All aquaria were aerated during the experiment and water parameters, namely temperature (19–21 °C), pH (7–8), conductivity (550–700 µS/cm) and dissolved oxygen concentrations (5–7 ppm), were monitored during the experiments. After hatching, larvae were fed three times a week with finely ground TetraMin fish food. After 10 days of development, larvae were sieved from the sediment and randomly divided between three treatment levels (each with 225 larvae), the control, low, and high treatment levels. The concentrations of each treatment level were selected based on environmental monitoring data. Pirimicarb, indoxacarb, and thiacloprid have been widely applied in recent decades and have been reported in environmental monitoring of surface waters and sediments. Pirimicarb can be detected at low concentrations under base flow conditions in contaminated waterbodies (Struger et al., 2016) and has been measured at concentrations up to 10 µg/L after heavy rainfall triggered runoff events (Kreuger, 1998). Despite its higher lipophilicity, indoxacarb's high application rates in viticulture result in frequent detections in impacted ground water, with some concentrations exceeding 0.1 µg/L (Herrero‐Hernández et al., 2020). Average sediment concentrations of indoxacarb in vineyard‐adjacent streams have been reported to be 24 µg/kg (Bereswill et al., 2012). Within the widely detected neonicotinoid class, thiacloprid has been measured at concentrations in surface waters up to 4.5 µg/L (Stehle et al., 2018). In the present study, the control treatment contained only aqueous medium, the low treatment level had a nominal concentration of 0.1 µg/L for all the insecticides, and the high treatment level had a nominal concentration of 16 µg/L for pirimicarb and indoxacarb. A lower concentration of 4 µg/L was used as the high treatment level for thiacloprid on account of its higher toxicity. The NOECs for chronic 28‐day water exposure of C. riparius larvae to pirimicarb, indoxacarb, and thiacloprid are >1000, 1.8, and 0.2 µg/L, respectively (Lewis et al., 2016). The relative toxicity of each treatment level was calculated as the ratio of the tested concentration to the respective NOEC (Supporting Information, Table S1). The highest treatment level selected for thiacloprid is therefore considered to be approximately twice as toxic as the highest treatment level of indoxacarb and more than 10 000 times more toxic than the highest pirimicarb treatment level. The exposure period lasted for 24 h and took place in the absence of sediment or food. For this, three aquaria containing 2.5 L of aqueous medium were prepared at the relevant insecticide concentration and allowed to equilibrate for 1 h before larvae were added. The insecticides were prepared in aqueous medium from formulation products, namely, Pirimor (500 g/kg pirimicarb; Syngenta), Steward (300 g/kg indoxacarb; DuPont), and Calypso (480 g/L thiacloprid; Bayer). Concentrations of pirimicarb, indoxacarb, and thiacloprid in the medium were measured at the start of the experiment by ultrahigh performance liquid chromatography coupled to tandem mass spectrometry by electrospray ionization (UHPLC‐ESI‐MS/MS) to be approximately 87%, 104%, and 100% of the nominal concentrations, respectively. Details about the analysis are provided in Supporting Information, Table S2. We therefore use the nominal concentrations to refer to the treatment levels in the remaining text.

After completing the 24‐h exposure period, the larvae were rinsed with clean medium. Dead larvae were identified by a lack of response to a physical stimulation with tweezers, removed and counted. In the case of pirimicarb and indoxacarb, at least 98% of larvae were alive after the 24‐h insecticide exposure, regardless of insecticide or treatment level. In the case of thiacloprid, 88% and 75% of the larvae survived in the 0.1 and 4 µg/L treatment levels, respectively. Surviving larvae in each treatment level were randomly assigned to four replicate groups. A replicate group consisted of three samples: two larvae samples collected at different time points and the adult insect sample. One quarter of the larvae from each replicate group were frozen and stored at −80 °C directly after cleaning, in preparation for insecticide concentration determination. The remaining three quarters were transferred to smaller aquaria (three per replicate group, 11.5 cm diameter, 0.5 L) containing sediment and aqueous medium with depths of 2 and 5 cm, respectively. Each small aquarium contained 14 larvae in the case of pirimicarb and indoxacarb, providing at least 7.4 cm² surface area per larvae. The lower survival of larvae exposed to thiacloprid resulted in slightly smaller replicate groups. Each small aquarium contained 10 larvae in this case and the aquaria used to collect adult emergence were reduced to three replicates, instead of four, for the highest treatment level. The larvae were provided with 10–15 mg of fish food every second day during the post‐exposure period. Individuals in the less densely populated replicates for thiacloprid therefore had less competition for resources, which potentially aided their development, reflecting sequential effects after insecticide exposure that would occur in the field. After a 72‐h post‐exposure period, the larvae from one of the three aquaria per treatment replicate group were collected from the sediment, rinsed, counted, and frozen at −80 °C. The remaining two aquaria per treatment replicate group were covered with a 0.6‐mm polyester mesh to capture emerging adults. The emerging adults were collected daily, grouped by sex, counted, and frozen at −20 °C prior to insecticide concentration determination. Adult insects which died during emergence were excluded.

Sample preparation for pesticide analysis

Extraction of insecticides from the larvae and adults was performed by ultrasonically assisted solid–liquid extraction as previously described (Roodt et al., 2022). Briefly, midge larvae (n = 7–14) were pooled within each replicate and sampling time point. Similarly, adult midges (n = 7–19) were pooled within each replicate (consisting of two aquaria per replicate group) by sex. All samples were freeze dried and pulverized using a Tissuelyser with steel pellets (Retsch). The total dry weight of all samples was determined on a fine scale to an accuracy of 0.001 mg (Mettler‐Toledo). In the case of pirimicarb and indoxacarb, subsamples of midge larvae (7.0 ± 0.5 mg), midge females (20.0 ± 0.5 mg), and midge males (6.0 ± 0.5 mg) were prepared from each replicate. In the case of thiacloprid, subsamples of midge larvae (4.0 ± 0.5 mg), midge females (13.0 ± 0.5 mg), and males (4.5 ± 0.5 mg) were prepared from each replicate. Before solvent extraction was performed, 20 µl of a 50‐ng/ml solution containing the deuterated internal standards indoxacarb‐D3, pirimicarb‐D6 or thiacloprid‐D4 was added to the samples, which were subsequently allowed to stand for 30 min at room temperature. Samples were then extracted with acetonitrile (3 × 1.5 ml) and methanol (2 × 1.5 ml). The pooled extracts were evaporated to dryness under a gentle stream of nitrogen gas and redissolved in a water/methanol (70:30 v/v) solution containing 0.1% formic acid. Aqueous medium samples were diluted with methanol to achieve a 70% aqueous solution before being centrifuged at 16 000 rpm for 5 min. All samples were then filtered through 0.2 µm polytetrafluoroethylene syringe filters prior to UHPLC‐ESI‐MS/MS analysis.

Quantitative pesticide analysis

Insecticide concentrations were determined by UHPLC‐ESI‐MS/MS. Instrument parameters were set as previously described (Roodt et al., 2022). Details of the instrument parameters are provided in Supporting Information, Table S2. Matrix‐matched standards were prepared with concentrations of 0.005, 0.01, 0.05, 0.1, 0.5, 1.0, 5.0, and 10.0 ng/ml using insects collected from the in‐house culture. Two multiple reaction monitoring transitions were used for each analyte (Supporting Information, Table S3). Recoveries of internal standards were monitored for quality control, with recoveries between 70% and 120% being considered acceptable. Measurement limits of quantification (LOQs) and detection (LODs) are provided in Supporting Information, Table S4.

Data analysis and statistics

The sex‐specific effects of exposure to each insecticide on larval development were quantified by calculating the time taken for 50% of the total number of successfully emerged individuals to emerge in each replicate (EmT₅₀). Bioconcentration factors (BCFs) were calculated as the larval concentration after 24‐h exposure based on dry weight divided by the exposure concentration. Insect‐mediated insecticide flux was calculated as the product of the average insecticide concentrations and total average dry weights of successfully emerged adult insects divided by the respective EmT₅₀ when calculated for total emergence of both sexes. The pooled EmT₅₀ for both sexes was chosen as a proxy time dimension of the flux calculation in the present specific microcosm study. This was done to account for changes in the overall emergence pattern resulting from changes in the development duration and emergence success of female insects. Male midges develop faster than females (Day et al., 1994), and an accelerated female development can therefore narrow the emergence window of both sexes together, while increasing the daily insecticide flux. An increase in the development times of female insects would have the opposite effect, thus decreasing the daily flux. Equations for the calculation of the flux are provided in Supporting Information, Table S6. Treatment effects on adult sex‐specific EmT₅₀, sex‐specific biomass of emergent adults, and total emergence success were tested for using the Kruskal–Wallis H test followed by a post hoc Dunn's test with Bonferroni correction when treatment level effects were detected. The significance level, α, was set at 0.05 for all tests. Statistical analyses were performed in R Ver 4.0.3 (R Core Team, 2020). All average values reported in the discussion were calculated as an arithmetic mean.

RESULTS

Survival, development, and emergence phenology

The pulse exposure of 10‐day old larvae to pirimicarb had no effect on larval survival (Supporting Information, Table S5), but delayed emergence of the female insects by approximately 3–4 days regardless of exposure concentration (Figure 1). The males were not significantly affected, but a tendency toward a longer development time is apparent at the 16 µg/L treatment level. The overall emergence success was not affected by exposure to pirimicarb (Figure 2). Pulse exposure to indoxacarb did not have an effect on larval survival (Supporting Information, Table S5) and did not delay the emergence of either sex at any treatment level (Figure 1). However, total emergence success was significantly decreased by approximately 25% at the 16 µg/L treatment level, with a tendency toward lower emergence success in the 0.1 µg/L treatment level (Figure 2). Larvae exposed to thiacloprid were moribund after the 24‐h exposure period and showed convulsive and incoherent movements in response to physical stimulation. After the 72‐h post‐exposure period, the larvae in the 0.1 µg/L treatment level had recovered with no difference in survival relative to the control (Supporting Information, Table S5). Larval survival in the 4 µg/L treatment level was, however, significantly reduced by approximately 50% relative to the control. Exposure to 0.1 µg/L thiacloprid reduced the EmT₅₀ by approximately 1–2 days relative to the control, which was significant for the female insects (Figure 1). Exposure at 4 µg/L had no effect on the EmT₅₀ of male insects, and insufficient female insects emerged at this treatment level to allow for the calculation of the EmT₅₀. Emergence success in the control of the thiacloprid experiment was lower relative to the controls for the other two insecticides, but overall emergence success was, however, still >70% and is therefore considered valid based on criteria set out for laboratory studies (OECD, 2004). Thiacloprid exposure significantly reduced overall emergence success at both treatment levels relative to the control (Figure 2). None of the insecticides affected the sex‐specific average adult individual dry weight at any treatment level relative to the respective controls (Supporting Information, Figure S1).

Insecticide concentrations in larvae and adult insects

Directly after the 24‐h exposure period, pirimicarb concentrations were measurable above the LOQ in the larvae from the 16 µg/L treatment level, but not the 0.1 µg/L treatment level (Figure 3). The average (n = 4, ±standard deviation) concentration in the higher treatment level was 31.0 ± 8.1 ng/g, with an average BCF of 1.9 L/kg. This concentration decreased to 0.4 ± 0.2 ng/g over the 72‐h post‐exposure period, corresponding to an approximate 99% reduction. Average larval indoxacarb concentrations, directly after the 24‐h exposure period, were 415.0 ± 90.5 and 3837.9 ± 1142.1 ng/g, with average BCFs of 4150 and 240 L/kg in the 0.1 and 16 µg/L treatment levels, respectively. Over the 72‐h post‐exposure period, the larval concentrations decreased to below the LOD in the 0.1 µg/L treatment level, but were still measurable, at 224.7 ± 175.4, in the 16 µg/L treatment level. Thus, the concentration of indoxacarb decreased by approximately 94% over the 72‐h post‐exposure period. Neither pirimicarb nor indoxacarb was measured above their respective LODs in the emergent males or females, regardless of treatment level. Thiacloprid concentrations in larvae were 125.2 ± 18.3 and 287.2 ± 90.8 ng/g after 24 h of exposure to the 0.1‐ and 4‐µg/L treatment levels, respectively. The corresponding average BCFs were 1252 and 49 L/kg. These concentrations decreased by approximately 30%–50% over the 72‐h post‐exposure period. Furthermore, thiacloprid was measured in the adults at an average concentration approximately 20% of that measured in the larvae 72 h post‐exposure regardless of the insect sex or treatment level. The adult insects therefore transferred approximately 10–15% of the thiacloprid concentration measured in the larvae directly after exposure. Overall, the thiacloprid fluxes calculated for both sexes emerging from the microcosms were 18.6 ± 14.2 and 43.0 ± 12.4 pg/day for the 0.1 and 4 µg/L treatment levels, respectively. Parameters for the flux calculation are provided in Supporting Information, Table S6.

DISCUSSION

Effects of insecticides on larval development and emergence

All three insecticides had an effect on at least one aspect of larval development and emergence. Pirimicarb caused a delay in the emergence of female midges, but had no effect on emergence success. The elimination of accumulated pirimicarb by the larvae post‐exposure may incur additional energetic costs, likely resulting in an extended development time (Monteiro et al., 2019). Similar to the present study, extended development times have been reported for target pest species of aphids when exposed to sublethal concentrations of pirimicarb (Xiao et al., 2015). Contrastingly, indoxacarb did not delay emergence, but reduced emergence success. Studies investigating the chronic exposure of C. riparius larvae to indoxacarb reported delays in development without reducing emergence success (Monteiro et al., 2019). However, reduction of successful metamorphosis, as was observed in the present study, is known for some terrestrial pest species after exposure to indoxacarb (Gamil et al., 2011; Saryazdi et al., 2012). The toxic mode of action of indoxacarb is related to the rate of metabolization of the parent compound to its toxic metabolite (Wing et al., 2000). Concentrations of indoxacarb still present in the larvae at the onset of metamorphosis and their conversion to its active toxic metabolite during a more vulnerable life stage may therefore explain the reduced emergence success in the highest, 16‐µg/L, treatment level.

Thiacloprid reduced larval survival 72 h post‐exposure, decreased overall emergence success, and accelerated the emergence of female midges. A delayed larval mortality after exposure, such as occurred in the 4‐µg/L treatment level, has also been observed in a variety of freshwater macroinvertebrates exhibiting a range of sensitivities (Beketov & Liess, 2008). A microcosm study in which C. riparius larvae were exposed to a related neonicotinoid insecticide, imidacloprid, during development found reduced emergence success at similar concentrations to the ones used in the present study, although no effect on the timing of emergence was observed (Chandran et al., 2018). In the present study, the accelerated development and earlier emergence observed in only the lowest, 0.1‐µg/L, treatment level could be interpreted as a nonmonotonic, potentially hormetic, response when considering that there was no reduction in larval density and no negative impact on the average dry weights of the emerging adults relative to the control (Steinberg, 2012). Similar accelerated development and emergence has been reported for C. riparius, as well as more complex insect communities, after exposure to low concentrations of pyrethroids (Goedkoop et al., 2010; Rogers et al., 2016). These authors hypothesized that stimulation of behavioural or biochemical processes by one or more of the pyrethroid stereoisomers at a low concentration resulted in the observed accelerated development. An accelerated larval development may, however, have negative consequences (e.g., decreased longevity) for the adult life stage which were not measured in the present study (Metcalfe & Monaghan, 2001).

Bioconcentration of insecticides in larvae

Indoxacarb was the most bioconcentrated in the larvae and pirimicarb the least, thus correlating positively with their lipophilicities. This result is based on the assumption that an equilibrium between uptake and elimination was achieved over the exposure period. In this context, similar BCFs have been reported for pirimicarb in Daphnia magna (BCFs 31–50) and for indoxacarb in the zebra fish Danio rerio (BCFs 1080–1752; Kusk, 1996; Y. Li et al., 2021). Thiacloprid did not fit the predicted trend and, despite having the lowest lipophilicity of the insecticides, was more accumulative than pirimicarb but less than indoxacarb. Similar underestimation of bioconcentration potential based on the logK_ow has been reported for the neonicotinoid imidacloprid in an aquatic oligochaete Lumbriculus variegatus (Contardo‐jara & Gessner, 2020). These authors reported much lower 24‐h BCFs (between 20 and 70 for exposure at 0.1 µg/L) than the ones determined in the present study, but the authors reported an increasing BCF with increasing exposure time and a maximum value was not established. Underestimation of the bioconcentration potential of neonicotinoids in amphipods is also well established in the literature and may be attributed to binding of the insecticide to biomolecules and the exoskeleton of the organisms (Chen & Kuo, 2018; Lauper et al., 2022; H. Li et al., 2021). In the present study, the lower BCFs of indoxacarb and thiacloprid at the higher treatment level can result from the saturation of uptake mechanisms or changes in the physiological condition of the larvae which reduce uptake, for example reduced respiration rate and immobilization (Mackay & Fraser, 2000). In addition, surfactants which are present in the formulation products as chemical adjuvants can also reduce the bioaccumulation of mid‐polarity pollutants in benthic organisms (Garcia‐galan et al., 2017). Because the exact composition of the adjuvants used in each formulation is not publicly available, a comparison is not possible and the reported BCF values in the present study may therefore be specific to the formulation products used.

Insecticide retention through metamorphosis

Thiacloprid was the most toxic insecticide to the larvae and was also the insecticide best retained through development and metamorphosis. This contradicted the hypothesized trend, in which pirimicarb would have the greatest flux from the aquatic to terrestrial ecosystem because of its low lipophilicity and low toxicity to the larvae. This hypothesis was based on what has been previously reported for the retention of a range of polycyclic aromatic hydrocarbons, fungicides, and herbicides, with similar lipophilicity and low toxicity, by emerging aquatic insects (Kraus et al., 2014; Roodt et al., 2022).

In the present study, the relatively rapid and complete elimination of pirimicarb and indoxacarb in the larvae resulted in no measurable concentrations remaining in the adult insects after metamorphosis. Elimination of thiacloprid by the larvae was, however, much slower, which resulted in higher concentrations being present in the larvae at the onset of metamorphosis, generally 6–7 days post‐exposure based on the median EmT₅₀ values. When considering only passive uptake of insecticides from the surrounding water, metabolism by relevant enzymes is an important factor determining the resulting concentrations in aquatic organisms (Katagi, 2010). Insecticide toxicity, in turn, is related to the concentrations present in the organism in combination with the toxic mode of action (Katagi & Tanaka, 2016). The observed rates of elimination therefore correlated with the insecticides' relative toxicities (i.e., 28‐day NOECs pirimicarb ≫ indoxacarb > thiacloprid). These results indicate a positive correlation between insecticide toxicity and the potential for aquatic–terrestrial transfer due to the rate of insecticide metabolism.

Thiacloprid flux from water to land

The relative thiacloprid flux was approximately 17 times higher from the 0.1‐µg/L than the 4‐µg/L treatment level when considering the exposure concentration (Supporting Information, Figure S2). More specifically, the average calculated flux of thiacloprid from the 0.1‐µg/L treatment level was approximately 40% of the flux from the 4‐µg/L treatment level, despite two orders of magnitude difference in the exposure concentrations. This result implies that relatively significant fluxes of thiacloprid may occur even at lower aqueous‐phase exposure concentrations than were used in the present study. Our results may also be relevant for other neonicotinoids. In this context, similar concentrations (2.7–47.2 ng/g) of the neonicotinoids clothianidin and imidacloprid have been reported in adult Diptera emerging from contaminated wetlands, although the exact aquatic exposure concentrations were not reported (Kraus et al., 2021). Neonicotinoids are, however, frequently detected at low concentrations as mixtures in surface waters (Schmidt et al., 2022). A compilation of neonicotinoid monitoring data from 29 studies across nine countries found a geometric mean concentration of 0.13 µg/L in water which occurred frequently and long term (Morrissey et al., 2015) and exceeds the lower thiacloprid treatment level used in the present study. Moreover, in temperate climates emerging Chironomidae (Diptera) have a very wide emergence period, discontinuing only during the coldest part of winter (Raitif et al., 2018). In addition, the proportion of Chironomidae in the aquatic insect community positively correlates with anthropogenic disturbances associated with agriculture, whereas abundances of other more sensitive taxa decrease (Raitif et al., 2018; Stenroth et al., 2015). Our results therefore indicate the potential for a widespread and near‐permanent aquatic–terrestrial flux of neonicotinoid insecticides from impacted freshwater ecosystems mediated by emerging midges.

Potential impacts on terrestrial consumers

Overall, thiacloprid had the highest potential to negatively impact terrestrial consumers at environmentally realistic pulse‐exposure concentrations. The resulting decrease in larval densities and emergence success has the potential to place significant pressure on higher trophic levels through reduced food availability (Tooker & Pearsons, 2021). In addition, reductions in adult aquatic insect abundance also reduce the availability of essential dietary polyunsaturated fatty acids which are not substituted by consumption of terrestrial insects (Hixson et al., 2015; Martin‐creuzburg et al., 2017; Twining et al., 2016). The presence of the neonicotinoid imidacloprid in surface waters has already been linked to declining insectivorous bird populations in Europe (Hallmann et al., 2014). In addition, shifts to earlier or later emergence of aquatic insects, as found in the present study after exposure to pirimicarb or thiacloprid, has the potential to decouple aquatic subsidy availability from terrestrial insectivore life history (Marczak & Richardson, 2008), although this is likely more relevant in the case of emerging insects with longer lifecycles and fewer generations per season.

The retention of thiacloprid by adult insect results in the potential dietary exposure of terrestrial predators to thiacloprid, among other neonicotinoids, via consumption of contaminated midges (Kraus et al., 2021). This may be relevant for a wide range of predators, including birds, bats, lizards, and spiders, which obtain a large proportion of their energy requirements through consumption of emerging aquatic insects (Baxter et al., 2005). Furthermore, this may be especially relevant for predators which specialize on emerging aquatic insects as a food source, for example some species of riparian spiders (Wieczorek et al., 2015). In addition, the emerging adult insects themselves may be negatively impacted by retained neonicotinoid insecticides. The neurotoxic mode of action of neonicotinoids has been linked to a range of sublethal effects in nontarget insects, such as vision loss, reduced immune response to pathogens, and behavioural effects (Pisa et al., 2021; Tasman et al., 2021). Retention of these compounds may therefore have negative impacts on the fitness and longevity of the successfully emerged adults, with potential for further cascading impacts on terrestrial consumers through further reduced food availability in addition to dietary insecticide exposure. The impacts of dietary exposure to neonicotinoids on vertebrate predators are challenging to study at the landscape scale due to the confounding effects of multiple stressors. However, laboratory investigations have revealed sublethal effects on development, behaviour, immune function, and reproductive success for a wide variety of terrestrial insectivores, including insectivorous birds and bats (Gibbons et al., 2015; Pisa et al., 2021; Wu et al., 2020).

CONCLUSION AND IMPLICATIONS

The present study offers a detailed laboratory‐scale investigation of aquatic–terrestrial insecticide fluxes propagated by emerging midges after a brief exposure to an environmentally realistic concentration. The results highlight the interplay between insecticide effects on insect emergence, the rate of elimination of the insecticide by the aquatic larvae, and insecticide retention by emerged adult insects. Exposure of the larvae to the most toxic (lowest NOEC) insecticide, thiacloprid, which was also the slowest to be eliminated, resulted in its transfer to the adults. Furthermore, relatively high fluxes were measured at the lower exposure concentration (0.1 µg/L) relative to the higher (4 µg/L) due to increased emergence success. Our results thus imply the potential for significant fluxes to take place even at lower exposure concentrations, potentially negatively affecting terrestrial insectivores hunting in adjacent terrestrial ecosystems. Our study also adds to the growing body of literature describing an underestimation of the potential for neonicotinoid bioaccumulation and persistence in food webs relative to their lipophilicities (Tooker & Pearsons, 2021). The extent and impacts of the aquatic–terrestrial transport of these insecticides by emerging aquatic insects at the field‐scale is a potentially important topic of future research.

Supporting Information

The Supporting Information is available on the Wiley Online Library at https:/10.1002/etc.5495.

Acknowledgment

The present study was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – Grant No. 326210499/GRK2360. Open Access funding enabled and organized by Projekt DEAL.

Conflict of Interest

The authors declare no conflict of interest.

Author Contributions Statement

Alexis P. Roodt: Conceptualization; Investigation; Formal analysis; Writing—original draft; Writing—review & editing. Sonja Schaufelberger: Investigation; Writing—review & editing. Ralf Schulz: Supervision; Writing—review & editing.

Data Availability Statement

Data not available in the Supporting Information are available from the corresponding author (roodt@uni-landau.de).

REFERENCES