Surface runoff and subsurface tile drain losses of neonicotinoids and companion herbicides at edge-of-field

Environmental Pollution xxx (2017) 1e10

Contents lists available at ScienceDirect

Environmental Pollution journal homepage: www.elsevier.com/locate/envpol

Surface runoff and subsurface tile drain losses of neonicotinoids and companion herbicides at edge-of-field*
tien a, *, Isabelle Giroux b, Georges The
riault a, Patrick Gagnon a, François Chre Julie Corriveau b a b

Agriculture and Agri-Food Canada, 2560 Hochelaga Blvd, Qu
ebec City, Qu
ebec, G1V 2J3, Canada Qu
ebec ministry of D
eveloppement Durable, de l'Environnement et de la Lutte contre les changements climatiques, Canada

a r t i c l e i n f o

a b s t r a c t

Article history: Received 9 August 2016 Received in revised form 1 February 2017 Accepted 1 February 2017 Available online xxx

With their application as seed coatings, the use of neonicotinoid insecticides increased dramatically during the last decade. They are now frequently detected in aquatic ecosystems at concentrations susceptible to harm aquatic invertebrates at individual and population levels. This study intent was to document surface runoff and subsurface tile drain losses of two common neonicotinoids (thiamethoxam and clothianidin) compared to those of companion herbicides (atrazine, glyphosate, S-metolachlor and mesotrione) at the edge of a 22.5-ha field under a corn-soybean rotation. A total of 14 surface runoff and tile drain discharge events were sampled over two years. Events and annual unit mass losses were computed using flow-weighted concentrations and total surface runoff and tile drain flow volumes. Detection frequencies close to 100% in edge-of-field surface runoff and tile drain water samples were observed for thiamethoxam and clothianidin even though only thiamethoxam had been applied in the first year. In 2014, thiamethoxam median concentrations in surface runoff and tile drain samples were respectively 0.46 and 0.16 mg/L, while respective maximum concentrations of 2.20 and 0.44 mg/L were measured in surface runoff and tile drain samples during the first post-seeding storm event. For clothianidin, median concentrations in surface runoff and tile drain samples were 0.02 and 0.01, mg/L, and respective maximum concentrations were 0.07 mg/L and 0.05 mg/L. Surface runoff and tile drain discharge were key transport mechanisms with similar contributions of 53 and 47% of measured mass losses, respectively. Even if thiamethoxam was applied at a relatively low rate and had a low mass exportation value (0.3%), the relative toxicity was one to two orders of magnitude higher than those of the other chemicals applied in 2014 and 2015. Companion herbicides, except glyphosate in tile drains, exceeded their water quality guideline during one sampling campaign after application but rapidly resumed below these limits. Crown Copyright © 2017 Published by Elsevier Ltd. All rights reserved.

Keywords: Neonicotinoid Runoff Subsurface Load Edge-of-field

1. Introduction Insect and weed chemical controls together with disease suppression are important components of integrated pest management strategies in commercial corn and soybean cropping systems. Drastic changes have recently occurred in pest insects management strategies. Use of broadcast organophosphate and

*

This paper has been recommended for acceptance by Eddy Y. Zeng. * Corresponding author.
tien), Isabelle.Giroux@ E-mail addresses: [email protected] (F. Chre
riault), mddelcc.gouv.qc.ca (I. Giroux), [email protected] (G. The [email protected] (P. Gagnon), [email protected] (J. Corriveau).

carbamate insecticides are declining while neonicotinoids are increasing since the introduction of imidacloprid in the 1990s and later with the use of the second generation neonicotinoid insecticides, thiamethoxam and clothianidin (Main et al., 2014; MDDELCC, 2016). The latter is also a degradation product of thiamethoxam (Nauen et al., 2003). The use of neonicotinoids increased dramatically during the last 10 years with their application as seed coatings. According to Baker and Stone (2014), about 4, 3 and 14 g/ ha of imidacloprid, thiamethoxam and clothianidin, respectively, were used in 2013 in selected watersheds of the Midwestern US corn belt. Douglas and Tooker (2015) have synthesized publicly available data in the U.S. and estimated that 34e44% of soybeans and 79e100% of corn hectares were treated in 2011. In 2012, it was estimated that applications of neonicotinoid covered about 11

http://dx.doi.org/10.1016/j.envpol.2017.02.002 0269-7491/Crown Copyright © 2017 Published by Elsevier Ltd. All rights reserved.


tien, F., et al., Surface runoff and subsurface tile drain losses of neonicotinoids and companion herbicides Please cite this article in press as: Chre at edge-of-field, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.02.002

2

F. Chr
etien et al. / Environmental Pollution xxx (2017) 1e10

million hectares with over than 216,000 kg of active ingredients in the Prairie cropland of Canada (Main et al., 2014). In Quebec (Canada), it is estimated that almost 100% of corn and 50% of soybeans seeds planted are coated with neonicotinoids which correspond to about 500,000 ha (MDDELCC, 2015). These usage rates coupled with long persistence and high water solubility have led to frequent detection of neonicotinoids in aquatic environments. For instance, in intensive corn and soybean producing areas, Giroux (2015) reported a mean detection frequency of thiamethoxam and clothianidin varying from 93% to 98% from 2012 to 2014 within four Quebec watersheds. In 2013 in southwestern Ontario, water samples collected in 5 corn-producing counties showed detection frequencies of 98.7% and 100% for thiamethoxam and clothianidin, respectively (Schaafsma et al., 2015). Hladik et al. (2014) reported mean detection frequencies at 9 sampling locations across Iowa in 2013 of 23, 47 and 75% for imidacloprid, thiamethoxam and clothianidin, respectively. In the Prairie Pothole region, Main et al. (2014) also reported neonicotinoid detection in 91% of the 136 wetlands sampled in spring 2013. Neonicotinoids, which act on insects by irreversibly binding to the nicotine acetylcholine receptors, have received increasing attention in the recent years due to their potential impacts on non-target organisms such as pollinators, predatory insects, earthworms and aquatic species (TFSP, 2015). Overall, it is recognized that neonicotinoids have the potential to cause negative effects at the individual and population levels of aquatic invertebrates at low concentrations (Pisa et al., 2014). In laboratory studies, negative relationships have been found between macroinvertebrate abundance and imidacloprid concentrations (Alexander et al., 2007; Mohr et al., 2012; Roessink et al., 2013; Stoughton et al., 2008). Effects on macroinvertebrates were observed at concentrations as low as 0,024 mg/L (Roessink et al., 2013). In a statistical analysis between macroinvertebrate abundance and imidacloprid concentration Van Dijk et al. (2013) showed a negative relationship. Indoor stream mesocosm studies had also demonstrated negative effect on the emergence of macroinvertebrates (Mohr et al., 2012) and that repeated pulses of imidacloprid also lead to massive drift of benthic macroinvertebrates (Beketov and Liess, 2008; Berghahn et al., 2012). Concurrently, chemical weeds control has continued to evolve in the last few decades primarily due to the introduction of new molecules, the development of herbicides-resistant crops and the apparition of herbicides-resistant weeds (Appleby, 2005). Chemicals introduced in the 1950s (e.g. atrazine) and the 1970s (e.g. glyphosate and metolachlor) continue to be widely used in NorthAmerica while newer products, including mesotrione which became available in the late 1990s, are increasing in popularity (Giroux, 2015). Aquatic ecosystems may be contaminated by pesticides, including neonicotinoids, through various routes such as atmospheric deposition, surface runoff, tile drain discharge and groundwater seepage losses (Anderson et al., 2013; Bonmatin et al., 2015; Smalling et al., 2015). As reported by Hladik et al. (2014), monitoring data for neonicotinoids transport to surface water are limited and more research on geographic occurrence, concentrations and transport mechanisms is needed. The intent of this study was to document surface runoff and tile drain losses of two commonly used neonicotinoids (thiamethoxam and clothianidin) and companion herbicides (atrazine, glyphosate, S-metolachlor and mesotrione) at the edge of a 22.5 ha field under corn and soybeans rotation. A particular attention was given to tile drain lines as they may contribute to a significant proportion of pollutant mass load exports in a cold and humid climate. To our knowledge, this study is the first conducted in the St. Lawrence lowlands providing flow-weighed event mean concentrations at

the edge-of-field as well as surface and subsurface partitioning of neonicotinoid losses. 2. Materials and methods 2.1. Site description The experimental site was established on a 22.5-ha field under corn and soybean rotation located in Saint-Samuel, Quebec, Canada (460201400 N - 72120 2000 W) (Fig. 1). Environment Canada station No. 7022160 located at 30 km from the site shows a historical annual temperature average of 6.4  C, varying from 10.2  C in January to 20.9  C in July, on average. The mean growing season length from 1979 to 2008 is 204 days based on a temperature above 5.5  C. The average annual precipitation is 1114 mm out of which 242 cm fall as snow, and 598 mm fall as rain during the growing season
te
o Que
bec, 2015). No supplemental irrigation was used on (Agrome the experimental site. The field is composed of three soil series, namely the Des Sault,
vrard loams, issued from deep marine depositions to Courval and Le superficial lagoon layers. According to the Canadian System of Soil Classification, the three soil series belong to the orthic humic gleysol (Humaquept) sub-group of the gleysolic order. All three
et al., soils are characterized as imperfect to poor drained (Rompre 1984). The average slope of the field is 0.25% and is artificially drained through a network of nine shallow ditches and thirteen tile drain lines flowing into two collectors. Hay was produced for four years without use of insecticides before it was burned with glyphosate in June 2013 and seeded with soybean after conventional tillage (CT) with a deep moldboard plow in half of the field and directly drilled through crop residues in the no-till (NT) second half. Successive crops were also grown at a proportion of 50% CT and 50% NT. In 2014, the first monitored year, corn was planted on May 20th with thiamethoxam coated seeds corresponding to 116 g/ha of active ingredients (a.i.). Chemical weeds control was achieved with commercial mixtures of atrazine, glyphosate, S-metolachlor and mesotrione at respective total concentrations of 285, 1717, 1050, and 105 g a.i./ha. The applications of the herbicide mixtures were made on June 9th, 2014 in the CT section and on June 17th in the NT section. Corn silage was harvested on October 1st, 2014. In 2015, soybean was partially planted on May 27th and completed on June 2nd with fungicides treated seeds (Baccilus Subtilis MBI600, Sedaxane, Fludioxonil, MetalaxylM and S-isomer). Weeds were controlled using glyphosate and chlorimuron ethyl at concentrations of 1334 and 9 g a.i./ha, respectively. No neonicotinoid treated soybean seeds were used. Soybean was harvested on October 25th, 2015. The two fungicides Baccillus Subtilis and sedaxane, and the herbicide chlorimuron ethyl were not monitored in this study. 2.2. Flow monitoring and sampling Flow monitoring and water quality sampling was implemented between May 21st, 2014 and December 1st, 2015. Winter and spring snowmelt events could not be monitored due mainly to issues with back flooding and freezing of sampling equipment. Surface monitoring was performed at the outlet of four shallow ditches (Fig. 1) using for each outlet a 2H flume (Plasti-Fab, Tualatin, OR, USA) equipped with an ISCO 720 Flow Module mounted in a stilling well to control a programmed ISCO 6712 Autosampler (Teledyne ISCO, Lincoln, NE, USA) to collect flow-weighted water samples. Surface runoff volumes triggering the collection of sub-samples were periodically adjusted to account for differences in runoff coefficient during the growing and the non-growing seasons. Flowweighted samples, collected at the four surface runoff monitoring


tien, F., et al., Surface runoff and subsurface tile drain losses of neonicotinoids and companion herbicides Please cite this article in press as: Chre at edge-of-field, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.02.002

F. Chr
etien et al. / Environmental Pollution xxx (2017) 1e10

3

Fig. 1. Study area and instrumentation. Map retrieved from Government of Quebec (2016).

locations, were then pooled to generate composite runoff samples. Accordingly, tile drain monitoring and samples collection was achieved using the same protocol at the outlets of the two tile drain collectors (Fig. 1) using Palmer-Bowlus flumes (eight and ten inches) (Plasti-Fab, Tualatin, OR, USA) mounted in a manhole. Surface runoff and tile drainage water samples were collected from a total of fourteen storm events during the 2014e2015 period. Six surface runoff and six tile drain discharge events were sampled in 2014 and seven surface runoff and seven tile drain discharge events were sampled in 2015 for a total of twenty-six water samples analyzed. No tile discharge samples were collected during the storm event on July 27, 2014 because there was no tile discharge. Likewise, no surface runoff samples were collected during the storm event on August 13, 2014 because the triggering volume was not attained. Event and annual (from planting to the establishment of snow cover) unit mass losses were computed using flowweighted concentrations and total surface runoff and tile drain volumes. Annual unit mass losses were individually computed for 2014 and 2015 since winter and snowmelt events were not sampled. During the snow free period for the non-sampled storm events occurring when the triggering volume was not attained, a mean concentration calculated from the previous and the following storm event concentrations was used.

2.3. Analytical methods All laboratory analyses were performed at the Centre d'expertise
bec (CEAEQ), which hosts the en analyse environnementale du Que
veloppement durable, laboratory of the Provincial Ministry of De Environnement et Lutte contre les changements climatiques (MDDELCC) in Quebec City, Canada. For all surface runoff and tile drain composite samples, three analyses were performed (CEAEQ, 2016a, b, c). For samples intended for analysis of atrazine and S-metolachlor, water samples (500 ml) were extracted with 300 ml of dichloromethane. The solvent was evaporated to a final volume of 500 ml under an argon gas jet. Pesticides in the extracts were separated on a DB-5-MS column and analyzed using a gas chromatography coupled with a mass spectrometer (Agilent 7890A-5975C). For the analysis of glyphosate and its breakdown product, aminomethyl-phosphonic acid (AMPA), 80 ml of the samples were acidified to a pH of 1. After 15 min, the samples were neutralized and mixed with Fmoc-Cl and a borate buffer. After 60 min, the Fmoc-Cl excess was eliminated with a dichloromethane extraction. The samples were then passed on a solid phase extraction cartridge (Oasis HLB), followed by an elution with 5 ml of basic methanol. The extracts were injected in a liquid phase chromatograph and


tien, F., et al., Surface runoff and subsurface tile drain losses of neonicotinoids and companion herbicides Please cite this article in press as: Chre at edge-of-field, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.02.002

F. Chr
etien et al. / Environmental Pollution xxx (2017) 1e10

4

separated on a column XBridge BEH C18, 50 mm X 2,1 mm, 3, 5 mm. The quantitation was done with a tandem mass spectrometer Quattro Ultima PT from Micromass. For the analysis of thiamethoxam, clothianidin and mesotrione 0,5 ml of the samples was directly injected on a C18 column (Aqua C18, 30 mm  2 mm, 3 mm) and analyzed by liquid chromatography coupled with a tandem mass spectrometer Sciex API 5000. For these analyses, quantification was achieved comparing the peaks of the samples with those of known standard concentration calibration curves. The detection limits for the targeted pesticides are 0,001 mg/l for thiamethoxam, 0,002 mg/l for clothianidin, 0,01 mg/l for atrazine, Smetolachlor and mesotrione and 0,04 mg/l for glyphosate. For each sequence of analysis, several blanks, external control samples prepared by the CEAEQ with certified material, extraction and injection surrogates have been achieved as laboratory quality control. No field blank have been collected for this specific project but a set of field blanks for the two years of the project were available for each type of analyses performed by the ministry as a part of the annual MDDELCC pesticide monitoring program. 2.4. Pesticide leaching potentials The Groundwater Ubiquity Score (GUS) was used to compare the leaching potentials of the two insecticides and the four herbicides. GUS is a methodology to estimate the potential of pesticides to contaminate groundwater (Gustafson, 1989). Briefly, it is related to the pesticide's persistence in soil and to its binding ability to the organic fraction of soil. The higher the GUS value, the higher the potential for pesticides to move toward groundwater. 2.5. Relative toxicity In this work, the relative toxicity of a chemical was arbitrarily defined as the value of annual flow-weighted mean concentration of the chemical divided by its chronic water quality guideline value described in Giroux (2015). Only mesotrione did not have a fixed or interim guideline value and hence no relative toxicity for mesotrione was calculated. The relative toxicity value was used as a mean of comparing pesticides based on observed mobility (annual mass load exportation weighted by annual discharge volume) and possible impacts on aquatic life (chronic water quality guideline). 3. Results and discussion 3.1. Detection frequencies Fourteen storm events were monitored during the 2014e2015 period. Despite the fact that thiamethoxam was only used in 2014 as corn seed coating, it was detected in all surface runoff and tile drain water samples of 2014 and 2015. The insecticide clothianidin,

which is also a degradation product of thiamethoxam, had not been used but was detected in 92.5% of the samples. The herbicides Smetolachlor, atrazine, glyphosate, and mesotrione were detected in 96.3%, 92.6%, 88.5% and 70.4% of the samples, respectively. The breakdown products AMPA, deethyl-atrazine and deisopropylatrazine were detected in 77.8%, 66.7% and 14.8% of the samples. The fungicide metalaxyl was detected in 7.4% of the samples. The detection of thiamethoxam and clothianidin 527 days after seeding was coherent with persistence values, quantified as halflife values (DT50), reported in different matrices such as soil ~ a et al., (Hilton et al., 2016; Jones et al., 2014) or wastewater (Pen 2011). Moreover, a review by Goulson (2013) reported DT50 values in field studies for thiamethoxam and clothianidin ranging from 7 to 109 days and 277 to 1386 days, respectively. Half-life values reported in the Pesticide Properties DataBase of the University of Hertfordshire (PPDB, 2016) are presented in Table 1 as an arbitrary mean of comparison between pesticides monitored in this study. Degradation of neonicotinoids in soils is influenced by local conditions and is affected by soil type, ultraviolet radiation, moisture, temperature and pH (Bonmatin et al., 2015). As demonstrated in other cold regions for other pesticides (Laitinen et al., 2009; Stenrød et al., 2008), the northern latitude of the study site and hence lower radiation and cold temperature may have contributed to the prolonged detection frequencies throughout the sampling campaign. In fact, Main et al. (2016) detected neonicotinoids in snowmelt suggesting long term persistence. It is unclear however how the soil moisture content affected the degradation of monitored neonicotinoids by microorganisms since local soil conditions varied from optimal moisture content to saturated and dry conditions. The organic matter (OM) content (total carbon percentages at 0e5 cm and 5e20 cm of 2.24% and 1.90%, respectively) together with the important 0e25 cm silt fraction of the three soil series
vrard-40% and Courval-34%) (Rompre
et al., (Des Saults-44%, Le 1984) may also have contributed to retain thiamethoxam and clothianidin on sorption sites throughout the monitoring period. In a study looking at the displacement of imidacloprid in soils with varying OM and sand contents, Selim et al. (2010) showed that retention of this neonicotinoid was mainly attributed to hydrophilic bonding between soil OM phenolic hydroxyl and carboxylic acidic groups and imidacloprid functional groups. In addition to adsorption-desorption phenomena, mainly governed by desorption kinetic in weakly sorbed chemicals (Gouy et al., 1999), high detection frequencies in collected surface and subsurface samples can be related to high solubility in water and leaching potential values of thiamethoxam and clothianidin (Table 1). These neonicotinoids were most likely mobilized during successive storm events and exported in surface runoff water and tile drain discharge after leaching by preferential flow associated to cracks and macropores, and flow through non-structured soils. Exportation routes are discussed in more details in the section “Exportation rates and routes”.

Table 1 Properties of the active ingredients analyzed. Active ingredient

GUS leaching potential indexa

Half-life in soil (day)a

Solubility in water at 20  C (mg/L)a

Organic carbon distribution coefficient (Koc; mL/g)a

Octanol-water partition coefficient at pH 7, 20  C (Kow: log Pow) a

Henry's law constant at 20  C (dimensionless)a

Toxicity threshold (mg/L)b

Clothianidin Thiamethoxam Atrazine Glyphosate Mesotrione S-metolachlor

4.91 4.69 3.20 0.25 2.80 1.91

545 50 75 15 32 15

340 4100 35 10500 160 480

123 56.2 100 1424 122 e

0.905 0.13 2.7 3.2 0.11 3.05

8.44E-15 1.93E-13 1.20E-07 6.60E-19 4.99E-09 8.98E-07

0.0083 0.0083 1.8 65 e 7.8

a b

Source: PPDB (2016). Source: Giroux (2015). Used for chronic effects on aquatic life: For glyphosate, the toxicity of the adjuvant is also accounted for; - No values were provided for mesotrione.


tien, F., et al., Surface runoff and subsurface tile drain losses of neonicotinoids and companion herbicides Please cite this article in press as: Chre at edge-of-field, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.02.002

F. Chr
etien et al. / Environmental Pollution xxx (2017) 1e10

3.2. Neonicotinoid concentrations Neonicotinoids are known to be toxic to a wide range of nontarget organisms, including aquatic invertebrates, at levels often below expected environmental concentrations (Pisa et al., 2014). The Pest Management Regulatory Agency of Health Canada has recently proposed to remove imidacloprid and to re-examine the use of thiamethoxam and clothianidin (PMRA, 2016a, b). For instance, while Morrissey et al. (2015) recommended ecological thresholds for neonicotinoid on aquatic invertebrate communities below 0.2 and 0.035 mg/L for acute and chronic effects, respectively, large proportions of globally collected water samples exceeded these targets (81% exceeded the acute effect threshold and 74% the chronic effect threshold). Due to a lack of comprehensive environmental monitoring programs for neonicotinoids and toxicity tests at different scales for a broad range of aquatic organisms, no clear ecological thresholds exist for thiamethoxam and clothianidin (Anderson et al., 2015). The Canadian Council of Ministers of the Environment (CCME, 2007) has proposed an interim guideline of 0.23 mg/L for imidacloprid. Based on recent studies and reviews (Mineau, 2013; Morrissey et al., 2015; Smit, 2014), the Government of Quebec recommended a long-term chronic ecological threshold of 0.0083 mg/L for individual and cumulative sum of imidacloprid, thiamethoxam and clothianidin concentrations in stream water (Giroux, 2015). This guideline set by the Government of Quebec is consistent with the 0,009 mg/L provisional regulatory acceptable concentration (RAC) set by the European Food Safety Authority for the concentration of imidacloprid in stream water (EFSA, 2014). Fig. 2 shows flow-weighted concentrations of thiamethoxam and

5

clothianidin in surface runoff and tile drain discharge of the 14 sampled storm events. Recall that only thiamethoxam was applied as seed coatings in 2014. The maximum concentrations of thiamethoxam reached respectively 2.2 and 0.44 mg/L in surface runoff and tile drain samples during the first post-seeding event of 2014. The maximum concentration of 0.07 mg/L of clothianidin occurred during the surface runoff event of June 12th, 2015 while the maximum subsurface concentration of 0.05 mg/L occurred on June 17th, 2014. The individual storm concentration values (Fig. 2) are coherent with transformation and mobilization processes between thiamethoxam and clothianidin. The results clearly show that most of the transformation of thiamethoxam occurred shortly after the 2014 seeding period with subsequent mobilization in the first three major events. The highest concentration of clothianidin in surface runoff, measured immediately after the 2015 seeding period, points to two potential sources, namely a probable degradation of thiamethoxam used in 2014 into clothianidin and its subsequent mobilization following the soil disturbance. In addition, dust particles from nearby fields could also be partly responsible for this surface runoff concentration spike. Dust clouds near planting equipment is now recognized as a key route of exposure for nontarget organisms with air contamination levels of up to 30 mg/m3 (Bonmatin et al., 2015). Furthermore, a recent study conducted in Ontario where agricultural practices and the environment are comparable to those of this study, strongly suggested that neonicotinoid residues adsorbed on surface soil dust may be a significant source of contamination transportable by wind erosion to neighboring land (Limay-Rios et al., 2016). However, use of seed lubricant and legislation to enforce their usage would most likely mitigate

Fig. 2. Thiamethoxam and clothianidin concentrations for the 14 sampled events. There was no tile drain discharge on July 27, 2014 and no samples were collected for surface runoff on August 13, 2014. Note: A different concentration scale (y-axis) was used for 2015 to facilitate the assessment of the temporal dynamics of the neonicotinoid concentrations.


tien, F., et al., Surface runoff and subsurface tile drain losses of neonicotinoids and companion herbicides Please cite this article in press as: Chre at edge-of-field, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.02.002

F. Chr
etien et al. / Environmental Pollution xxx (2017) 1e10

6

dust related risks (Canada, 2015a). More studies on dust transport and deposition would be required to conclude on the relative importance of clothianidin deposition from dust as compared to the degradation of applied thiamethoxam. Moreover, two thirds of total clothianidin exported load was lost in 2014, which indicate that peak concentrations may occur during the seeding period regardless of the application rates of thiamethoxam but total mass loss was related to degradation of applied parent material. As one might expect, all of these flow-weighted concentrations measured at the edge-of-field, before dilution in the receiving stream, exceeded the chronic threshold value of 0.0083 mg/L for instream concentration set by the Government of Quebec with the exception of tile drain concentrations during the second year. Concentration values measured at the edge-of-field are hereafter compared to those of two “in-stream” (Giroux, 2015; Hladik et al., 2014) and two “in-field” studies (Samson-Robert et al., 2014; Schaafsma et al., 2015). The observed edge-of-field median concentration of 0.215 mg/L for thiamethoxam in 2014 is one order of magnitude larger than in-stream measurements in intensively cropped watersheds with corn and soybean. The 2014 median value from Giroux (2015) was 0.024 mg/L, for four small watersheds varying from 78 to 313 km2, while the median value observed by Hladik et al. (2014) was 0.007 mg/L for nine larger watersheds with drainage areas of 521e836,000 km2. Interestingly in the present study and for both clothianidin and thiamethoxam, the median concentrations for the non-application years are of the same order of magnitude as in-stream measurements (0.016 and 0.010 mg/L for clothianidin and 0.006 mg/L for thiamethoxam). The thiamethoxam median concentration during the application year (2014) is also coherent with in-field studies from Samson-Robert et al. (2014) and Schaafsma et al. (2015) who respectively reported median values of 0.475 and 0.685 mg/L for thiamethoxam sampled in puddles of water within fields sown with neonicotinoid-coated seeds. These results, obtained in standing water where evaporation and photolytic breakdown can occur (Lu et al., 2015), are two to three times larger than the median value (0.215 mg/L) coming from surface runoff and tile drain discharge. 3.3. Herbicides concentrations Four herbicides, namely atrazine, glyphosate, mesotrione and S-metolachlor, were monitored essentially for comparison purposes with the neonicotinoids used in this study. Giroux (2015) has proposed long-term chronic ecological threshold for atrazine (1.8 mg/L), glyphosate (65 mg a.i./L) and S-metolachlor (interim guideline, 7.8 mg/L) for the protection of aquatic life. According to the Pesticide Action Network, there are currently no water quality standards or criteria established for mesotrione by the U.S. or Canadian governments (Kegley et al., 2016). Maximum concentrations of 59, 130 and 130 mg/L were measured for atrazine, glyphosate and S-metolachlor, respectively, in surface runoff during the third post-seeding storm event (58e3.5 mm-mm/hr) on June 24th, 2014. The maximum concentration of 4.1 mg/L for mesotrione was measured during the preceding storm event (21e4.0 mm-mm/hr) on June 18th, 2014. In tile drain samples, maximum concentrations were all measured on June 24th, 2014 where concentrations of atrazine, glyphosate, S-metolachlor and mesotrione reached 21, 25, 26 and 4.8 mg/L, respectively. With the exception of glyphosate in tile drains which constantly remained below its long term exposure guideline, maximum concentrations in surface runoff and tile drain samples exceeded briefly the guidelines in June 24th, 2014 but rapidly resumed to concentrations less of two orders of magnitude and consequently well below water quality guidelines for most of the remaining eleven storm events monitored in 2014 and 2015.

3.4. Exportation rates and routes Exportation rates were calculated as the percentage ratio of the pesticide annual mass loss from surface runoff and tile drain discharge over its annual applied mass. The exportation rates for thiamethoxam and companion herbicides are shown by the dots in Fig. 3. Over 0.3% (0.386 g/ha) of applied thiamethoxam (116 g/ha) was exported from the experimental site in 2014 as thiamethoxam and clothianidin (GUS values of 4.69 and 4.91 respectively). In comparison, atrazine, which also has a high GUS index value of 3.20, had an exportation rate of 4.7% and was broadcast at a rate of 285 g/ha which was nearly twice the rate of thiamethoxam. On the other hand, glyphosate, which is known to adsorb strongly to soil particles and to have a low GUS index of 0.25, was lost at a proportion of 1.7% even though it was applied twice in 2014 for a total application load of 1717 g/ha, about ten times more than thiamethoxam. The herbicides S-metolachlor and mesotrione with respective GUS values of 1.91 and 2.80 and high and low application dose (1050 and 105 g/ha) exhibited exportation rates of 2.4% and 1.4%, respectively. Exportation rates of neonicotinoids, when compared to those of herbicides, indicate that low dosage application combined with soil incorporation tend to limit the percentage of mass applied lost to aquatic ecosystems. However, due to the high toxicity of neonicotinoids, impacts on aquatic ecosystems are susceptible to occur. See section “Relative toxicity” for a detailed comparison of pesticides relative toxicity based on mass losses and toxicity values. In addition to dust deposition on water bodies, highly soluble pesticides such as thiamethoxam and clothianidin are susceptible to reach waterways through snowmelt runoff, storm event runoff, tile drain discharge and exfiltration from contaminated groundwater (Bonmatin et al., 2015; Hladik et al., 2014; Main et al., 2016). Dust deposition and exfiltration processes could not be monitored in this work. Recent studies have shown that a toxic dust cloud can form around drilling equipment, but new seed lubricant should limit the degree to which these clouds affect waterways (Bonmatin et al., 2015; Girolami et al., 2013). Studies on groundwater contamination have also demonstrated that neonicotinoids are increasingly being detected in monitored wells. In a potato producing area of Quebec, Giroux (2003) measured imidacloprid maximum and median concentrations of 6.4 and 0.0385 mg/L respectively and the USEPA (2008) reported level ranging from 0.2 to 7 mg/L in New York State. More studies on atmospheric deposition and exfiltration from contaminated groundwater would be required to estimate the contribution of these transport mechanisms on waterways contamination. Table 2 shows that both surface runoff and tile drain discharge are important exportation pathways with 53% of measured neonicotinoid load coming from surface runoff and 47% from tile drains. Moreover, most of the exported loads (>85%) of pesticides assessed in this study were observed during the first storm events of the post-seeding period. The high water solubility for neonicotinoids and the transitory field condition (bare soils before the crop establishes a protective canopy) for all pesticides are certainly critical features during this period. In addition to the high water solubility of thiamethoxam and clothianidin (Table 1), leaching through preferential flows in cracks and macropores as well as through the less structured horizon of the sandy loam soils could explain the large proportion of subsurface neonicotinoids losses. Stone and Wilson (2006), showed that preferential flow could represent 11e51% of total storm drain volume thus bypassing the soil matrix and reaching tile drain faster than expected by traditional displacement theory. For instance, 29% of glyphosate (Table 2) was exported by tile drains during the study period which also indicate that transport of adsorbed pesticides through preferential flow was


tien, F., et al., Surface runoff and subsurface tile drain losses of neonicotinoids and companion herbicides Please cite this article in press as: Chre at edge-of-field, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.02.002

F. Chr
etien et al. / Environmental Pollution xxx (2017) 1e10

7

Fig. 3. Barplots of relative toxicity (left y-axis) for the two neonicotinoids and three herbicides (no relative toxicity was calculated for mesotrione). The relative toxicity is defined as the ratio of the annual flow-weighted mean concentration (from planting to the establishment of snow cover) over the threshold concentration for chronic effects on aquatic life (see Table 1). Dots illustrate exportation rates of pesticides applied during the year (right y-axis). Note: A different concentration scales (left y-axis) was used for the 2015 values to display the difference between the two groups of pesticides. The single neonicotinoid exportation rate in 2014 combines both the thiamethoxam and clothianidin data.

an important transport mechanism. This is also coherent with the observed rapid mobilization of neonicotinoids and companion herbicides as illustrated by the high loads exported during the first three storm events following the 2014 seeding and spraying period. Fig. 4 indicates that 86% of the total 2014 and 2015 neonicotinoid load exportation was attained after the first three post seeding events of 2014 with similar values for companion herbicides. Other studies on herbicides have shown that rainfalls characteristics, including timing, depth and intensity, following applications have a significant impact on mass losses (Caron et al., 2010; Kladivko et al.,

1999; Norgaard et al., 2014; Shipitalo and Owens, 2003). Kladivko et al. (1999) demonstrated that 55 to >90% of the pesticides exported from their project site through tile drains occurred during the first storm event after application. In 2014 in this study, the first three rainfall events following seeding had respective depth and mean intensity of 29e2.2, 21e4.0 and 58e3.5 mm-mm/hr. During these events, 35, 11 and 45% of thiamethoxam losses were observed which is coherent with observations made with herbicides (Caron et al., 2010; Norgaard et al., 2014; Shipitalo and Owens, 2003). As per other pesticides, the importance of neonicotinoid exported

Table 2 Total load and mean concentration in surface runoff and tile drain discharge at the edge of field during four periods for the six active ingredients analyzed. Period

S-metolachlor Mesotrione Glyphosate Total herbicides Thiametoxam Clothianidin Total neonicotinoids Source Water Atrazine volume Conc Load Conc Load Conc Load Conc Load Conc Load Conc Load Conc Load Conc Load 3 (m /ha) (mg/L) (g/ha) (mg/L) (g/ha) (mg/L) (g/ha) (mg/L) (g/ha) (mg/L) (g/ha) (mg/L) (g/ha) (mg/L) (g/ha) (mg/L) (g/ha)

Post-seeding 2014 Runoff 182 (MayeJune) Drain 474 Total 655 Summer-fall 2014 Runoff 130 (JulyeNov) Drain 647 Total 777 Post-seeding 2015 Runoff 85 (MayeJune) Drain 857 Total 943 Summer-fall 2015 Runoff 229 (JulyeNov) Drain 1320 Total 1549

48.9 8.35 14.7 0.42 0.99 0.94 0.13 0.06 0.07 0.09 0.05 0.05

8.88 3.96 12.8 0.06 0.64 0.70 0.01 0.05 0.06 0.02 0.06 0.08

105 10.5 25.3 1.62 1.30 1.33 0.77 0.09 0.12 0.29 0.07 0.09

19.1 4.97 24.1 0.21 0.84 1.05 0.07 0.07 0.14 0.07 0.10 0.16

1.76 1.99 1.96 0.15 0.26 0.25 0.12 0.01 0.01 0.01 0.01 0.01

0.32 0.94 1.26 0.02 0.17 0.19 0.01 0.01 0.02 0.002 0.01 0.01

119 12.8 29.4 2.11 1.65 1.69 17.9 2.85 3.54 2.20 0.14 0.30

21.7 6.05 27.7 0.27 1.07 1.34 1.53 2.45 3.97 0.50 0.19 0.69

275 33.6 71.3 4.30 4.19 4.20 18.9 3.01 3.74 2.60 0.27 0.45

50.0 15.9 66.0 0.56 2.71 3.27 1.61 2.58 4.20 0.59 0.36 0.95

1.09 0.31 0.43 0.03 0.02 0.02 0.06 0.007 0.009 0.004 0.003 0.003

0.20 0.15 0.34 0.004 0.01 0.02 0.005 0.006 0.01 0.001 0.004 0.005

0.04 0.02 0.02 0.01 0.01 0.01 0.05 0.005 0.007 0.01 0.003 0.003

0.007 0.009 0.02 0.001 0.007 0.008 0.004 0.004 0.009 0.002 0.004 0.006

1.13 0.33 0.45 0.04 0.03 0.03 0.11 0.01 0.02 0.01 0.01 0.01

0.20 0.15 0.36 0.005 0.02 0.03 0.01 0.01 0.02 0.003 0.01 0.01


tien, F., et al., Surface runoff and subsurface tile drain losses of neonicotinoids and companion herbicides Please cite this article in press as: Chre at edge-of-field, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.02.002

F. Chr
etien et al. / Environmental Pollution xxx (2017) 1e10

8

Fig. 4. Cumulative total load (solid lines) and cumulative tile drain load (dotted line) for thiamethoxam and clothianidin in 2014 and 2015. The difference between solid and dotted lines represents the cumulative surface runoff load. The first date in each graphic indicates the day when hydrological data collection started and the second date indicates the day when the first water samples were collected. The 2015 graphic depicts the assumption that neonics concentrations were constant between the first and second dates. Note: Cummulative loads were reset to zero in 2015 since winter and snowmelt events were not monitored. A different concentration scale was used for 2015 to display the temporal dynamics of the neonicotinoid cumulative loads.

loads is expected to vary from year-to-year based on climatic conditions following applications and crops. For instance, after the first three storm events of 2014 under corn production, 95% of glyphosate was lost while after the first three storm events of 2015 under soybean production it reached 85%. These values correspond to 29.0 g/ha (1.7% of mass applied) in 2014 and 4.7 g/ha (0.3% of mass applied). In addition to crops, these significant differences can be explained by the important variations in surface runoff and tile drain discharges from the first three events that totalized 66 mm in 2014 but only 28 mm in 2015. 3.5. Relative toxicity In intensively cropped watersheds, substantial edge-of-field concentrations of thiamethoxam and clothianidin diluting in stream water may still remain at level of concentrations susceptible to be harmful for aquatic organisms by exceeding the chronic ecological threshold value (Giroux, 2015). Neonicotinoids are known to be hazardous for aquatic ecosystems (EFSA, 2014; Smit, 2014) at concentrations lower than other chemicals commonly used in integrated pest management, such as herbicides. For instance, after review of imidacloprid impacts on aquatic ecosystems, the Canadian Pest Management Regulatory Agency proposed to prohibit its usage in agricultural production and to review the usage of thiamethoxam and clothianidin (PMRA, 2016a, b). Fig. 3 compares the relative toxicity of thiamethoxam and clothianidin to those of herbicides applied during the study period. From the chemicals applied in 2014, thiamethoxam showed a

relative toxicity value of 95 followed by atrazine (18), S-metolachlor (9) and glyphosate (1). In 2015, only glyphosate was applied and interestingly, residual levels of both neonicotinoids gave similar relative toxicity values (~3) followed by glyphosate, atrazine and Smetolachlor with relatively low values (~0.1). Even if thiamethoxam was applied at a comparatively low rate (116 g/ha) and had a low mass exportation value (0.3%), its relative toxicity was about one to two orders of magnitude higher than other applied chemicals for both years. For instance, atrazine which is a known carcinogen was applied at a rate of 285 g/ha and showed a relative toxicity value of 18 in 2014 compared to 95 for thiamethoxam. Moreover, glyphosate which has recently been upgrade to a probable carcinogen (Guyton et al., 2015) was applied in 2014 and 2015 at respective rates of 1717 and 1334 g/ha and gave corresponding relative toxicity values of 1 and 0.1. However, Health Canada has re-evaluated glyphosate in 2015 and concluded that glyphosate presented no harm to human health when managed in conformity with the current regulations (Canada, 2015b). 4. Conclusion This study has shown over two years detection frequencies close to 100% in edge-of-field surface runoff and tile drain water samples for both thiamethoxam and clothianidin although thiamethoxam had been applied during the first year only. The observed edge-offield concentrations often exceeded the Government of Quebec recommended threshold for chronic effect on aquatic life (0.0083 mg/L) and the provisional chronic guideline of the European


tien, F., et al., Surface runoff and subsurface tile drain losses of neonicotinoids and companion herbicides Please cite this article in press as: Chre at edge-of-field, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.02.002

F. Chr
etien et al. / Environmental Pollution xxx (2017) 1e10

Food Safety Authority for in-stream individual and cumulative sum of imidacloprid, thiamethoxam and clothianidin concentrations (0.009 mg/L). The median concentration of thiamethoxam (0.215 mg/ L) during the application year 2014 was in the same order of magnitude of the in-field observations in puddles of standing water from other studies (Samson-Robert et al., 2014; Schaafsma et al., 2015) and one order of magnitude larger than in-stream measurements from drainage basins varying from 78 to 836,000 km2 (Giroux, 2015; Hladik et al., 2014). This study has also shown that both surface runoff and tile discharge were key transport mechanisms with similar contributions of 53 and 47% of total mass loads, respectively. In coherence with other studies on herbicides, timing of rainfalls and precipitation characteristics greatly affected transport rates with the majority of neonicotinoids and companion herbicides losses occurring within the first three events following applications. Even if mass losses and exportation rates of neonicotinoids were smaller due to low dosage and incorporation as coated seeds, their relative toxicity for aquatic invertebrates was still higher in both monitored years when compared to companion herbicides. Due to their high detection frequencies, concentrations, exportation rates and relative toxicity values, neonicotinoids are expected to adversely impact aquatic invertebrates (PMRA, 2016a, b). Since invertebrates are at the base of the food web, more studies are required to quantify the acute and chronic impacts of neonicotinoid treatments on the different trophic levels. Moreover, recent agronomic studies have shown that neonicotinoids do not systematically increased crop yields in corn and soybean production (MDDELCC, 2015; Myers and Hill, 2014). In light of these observations and as proposed in the Quebec Pesticide Strategy 2015e2018 (MDDELCC, 2015), it is recommended that neonicotinoids, including coated seeds, should not be used systematically when no infestation by seedling insect pests has been observed. The Quebec Pesticides Strategy aims to reduce the use of neonicotinoids by requiring that the application of neonicotinoids for agricultural purposes be justified by an agronomist. This will encourage best practices when using highest-risk pesticides. In light of the observations made in this study, it appears that mitigation strategy targeting only dust and surface runoff control would not be sufficient. In addition to dust control, such as the usage of seed flow lubricants, best management practices that tackle both surface and subsurface transport mechanisms of neonicotinoids should be developed. Finally, more studies on other transport pathways, such as atmospheric deposition and exfiltration from contaminated groundwater, should be conducted to estimate the relative contribution of these transport mechanisms on a mass balance approach. Acknowledgement This research project was co-funded by Agriculture and Agri
bec ministry of De
veloppement durable, EnviFood Canada, Que ronnement et de la lutte contre les changements climatiques (MDDELCC) and Ferme Bergeroy S.E.N.C. We would like to thank Nadia Goussard (AAFC), Marie-Claire Grenon from the laboratory
bec Centre d'expertise en analyse environnementale du Que (CEAEQ), Jean-Thomas Denault from MDDELCC, Claude Bergeron,
Bergeron for their valuable Guylaine Bergeron and Rene contributions. References
te
o Que
bec, 2015. “Atlas agroclimatique du Que
bec.”. 〈http://www. Agrome agrometeo.org/index.php/atlas〉 (Feb. 03, 2017). Alexander, A.C., Culp, J.M., Liber, K., Cessna, A.J., 2007. Effects of insecticide exposure on feeding inhibition in mayflies and oligochaetes. Environ. Toxicol. Chem. 26, 1726e1732.

9

Anderson, J.C., Dubetz, C., Palace, V.P., 2015. Neonicotinoids in the Canadian aquatic environment: a literature review on current use products with a focus on fate, exposure, and biological effects. Sci. Total Environ. 505, 409e422. Anderson, T.A., Salice, C.J., Erickson, R.A., McMurry, S.T., Cox, S.B., Smith, L.M., 2013. Effects of landuse and precipitation on pesticides and water quality in playa lakes of the southern high plains. Chemosphere 92, 84e90. Appleby, A.P., 2005. A history of weed control in the United States and Canada - a sequel. Weed Sci. 53, 762e768. Baker, N.T., Stone, W.W., 2014. Annual agricultural pesticide use for midwest stream-quality assessment, 2012-13. U. S. Geol. Surv. Data Ser. 863. Beketov, M.A., Liess, M., 2008. Potential of 11 pesticides to initiate downstream drift of stream macroinvertebrates. Archives Environ. Contam. Toxicol. 55, 247e253. Berghahn, R., Mohr, S., Hübner, V., Schmiediche, R., Schmiedling, I., Svetich-Will, E., Schmidt, R., 2012. Effects of repeated insecticide pulses on macroinvertebrate drift in indoor stream mesocosms. Aquat. Toxicol. 122e123, 56e66. Bonmatin, J.-M., Giorio, C., Girolami, V., Goulson, D., Kreutzweiser, D., Krupke, C., Liess, M., Long, E., Marzaro, M., Mitchell, E., 2015. Environmental fate and exposure; neonicotinoids and fipronil. Environ. Sci. Pollut. Res. 22, 35e67. Canada, H., 2015a. Pollinator Protection and Responsible Use of Insecticide Treated Seed. http://www.hc-sc.gc.ca/cps-spc/pubs/pest/_fact-fiche/pollinatorprotection-pollinisateurs/treated_seed-semences_traitees-eng.php, 2. Canada, H., 2015b. Proposed Re-evaluation Decision PRVD2015-01, Glyphosate. http://www.hc-sc.gc.ca/cps-spc/pest/part/consultations/_prvd2015-01/ prvd2015-01-eng.php. Caron, E., Lafrance, P., Auclair, J.C., Duchemin, M., 2010. Impact of grass and grass with poplar buffer strips on atrazine and metolachlor losses in surface runoff and subsurface infiltration from agricultural plots. J. Environ. Qual. 39, 617e629. CCME, 2007. Canadian water quality guidelines for the protection of aquatic life: imidacloprid. In: Canadian environmental Quality Guidelines, 1999. Canadian Council of Ministers of the Environment, Winnipeg, Canada, p. 8.
termination des pesticides de type organophosphore
, triazine, CEAEQ, 2016a. De
e substitue
e, phtalimide et pyre
thrinoide dans l’eau, les sols et carbamate, ure
diments : extraction liquide-liquide; dosage par chromatographie en les se
e a  un spectrome tre de masse, MA.400 e Pest. 1.0. Centre phase gazeuse couple
bec, Ministe re du d'expertise en analyse environnementale du Que
veloppement durable, de l’environnement et de la Lutte contre les changede
bec (Current edition available on demand). ments climatiques, Que
termination des pesticides flumetsulam, rimsulfuron, imaze
thaCEAEQ, 2016b. De
sotrione: dosage pyr, nicosulfuron, imazapyr, sulfosulfuron, clothianidine et me
e a  un spectrome tre de masse en par chromatographie en phase liquide couple tandem (MS/MS), MA.403 e FRIN 1.2. Centre d'expertise en analyse environ
bec, Ministe re du De
veloppement durable, de l’Environnenementale du Que
bec (Current ment et de la Lutte contre les Changements climatiques, Que edition available on demand).
termination du glyphosate, de l’AMPA et du glufosinate dans les CEAEQ, 2016c. De
ge
taux et le sol : dosage par chromatographie en phase liqeaux, les tissus ve
e avec un spectrome tre de masse en tandem et de
rivation au Fmoc, uide couple MA.400 e Glyphosate. Centre d'expertise en analyse environnementale du
bec, Ministe re du De
veloppement durable, de l’environnement et de la Que
bec (Current edition available on Lutte contre les changements climatiques, Que demand). Douglas, M.R., Tooker, J.F., 2015. Large-scale deployment of seed treatments has driven rapid increase in use of neonicotinoid insecticides and preemptive pest management in U.S. Field crops. Environ. Sci. Technol. 49, 5088e5097. EFSA, 2014. Conclusion on the peer review of the pesticide risk assessment for aquatic organisms for active substance imidacloprid. In: Authority, E.F.S. (EFSA Journal). Girolami, V., Marzaro, M., Vivan, L., Mazzon, L., Giorio, C., Marton, D., Tapparo, A., 2013. Aerial powdering of bees inside mobile cages and the extent of neonicotinoid cloud surrounding corn drillers. J. Appl. Entomol. 137, 35e44. Giroux, I., 2003. Contamination de l’eau souterraine par les pesticides et les nitrates
gions en culture de pommes de terre. Campagne d’e
chantillonnage dans les re re de l'Environnement. Direction de suivi de l'e
tat de 1999-2000-2001. Ministe
bec, Que
bec. de l'environnement. Gouvernement du Que
sence de pesticides dans l’eau au Que
bec : Portrait et tendances Giroux, I., 2015. Pre re du De
veloppement dans les zones de maïs et de soya e 2011  a 2014. Ministe durable, de l’Environnement et de la Lutte contre les changements climatiques.
tat de l’environnement. Bibliothe que et Archives Direction du suivi de l’e
bec, Que
bec, Canada, p. 47. þ 45 ann. nationales du Que Goulson, D., 2013. An overview of the environmental risks posed by neonicotinoid insecticides. J. Appl. Ecol. 50, 977e987. Gouy, V., Dur, J.C., Calvet, R., Belamie, R., Chaplain, V., 1999. Influence of adsorptiondesorption phenomena on pesticide run-off from soil using simulated rainfall. Pesticide Sci. 55, 175e182. Gustafson, D.I., 1989. Groundwater ubiquity score: a simple method for assessing pesticide leachability. Environ. Toxicol. Chem. 8, 339e357. Guyton, K.Z., Loomis, D., Grosse, Y., El Ghissassi, F., Benbrahim-Tallaa, L., Guha, N., Scoccianti, C., Mattock, H., Straif, K., 2015. Carcinogenicity of tetrachlorvinphos, parathion, malathion, diazinon, and glyphosate. Lancet Oncol. 16, 490e491. Hilton, M.J., Jarvis, T.D., Ricketts, D.C., 2016. The degradation rate of thiamethoxam in European field studies. Pest Manag. Sci. 72, 388e397. Hladik, M.L., Kolpin, D.W., Kuivila, K.M., 2014. Widespread occurrence of neonicotinoid insecticides in streams in a high corn and soybean producing region, USA. Environ. Pollut. 193, 189e196. Jones, A., Harrington, P., Turnbull, G., 2014. Neonicotinoid concentrations in arable


tien, F., et al., Surface runoff and subsurface tile drain losses of neonicotinoids and companion herbicides Please cite this article in press as: Chre at edge-of-field, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.02.002

10

F. Chr
etien et al. / Environmental Pollution xxx (2017) 1e10

soils after seed treatment applications in preceding years. Pest Manag. Sci. 70, 1780e1784. Kegley, S.E., Hill, B.R., Orme, S., Choi, A.H., 2016. PAN Pesticide Database. Pesticide Action Network, North America (Oakland, CA) (2016). http://www. pesticideinfo.org. Kladivko, E., Grochulska, J., Turco, R., Van Scoyoc, G., Eigel, J., 1999. Pesticide and nitrate transport into subsurface tile drains of different spacings. J. Environ. Qual. 28, 997e1004. €mo €, S., Nikunen, U., Jauhiainen, L., Siimes, K., Turtola, E., 2009. Laitinen, P., Ra Glyphosate and phosphorus leaching and residues in boreal sandy soil. Plant Soil 323, 267e283. Limay-Rios, V., Forero, L.G., Xue, Y., Smith, J., Baute, T., Schaafsma, A., 2016. Neonicotinoid insecticide residues in soil dust and associated parent soil in fields with a history of seed treatment use on crops in southwestern Ontario. Environ. Toxicol. Chem. 35, 303e310. Lu, Z., Challis, J.K., Wong, C.S., 2015. Quantum yields for direct photolysis of neonicotinoid insecticides in water: implications for exposure to nontarget aquatic organisms. Environ. Sci. Technol. Lett. 2, 188e192. Main, A.R., Headley, J.V., Peru, K.M., Michel, N.L., Cessna, A.J., Morrissey, C.A., 2014. Widespread use and frequent detection of neonicotinoid insecticides in wetlands of Canada's Prairie Pothole region. PloS one 9, e92821. Main, A.R., Michel, N.L., Cavallaro, M.C., Headley, J.V., Peru, K.M., Morrissey, C.A., 2016. Snowmelt transport of neonicotinoid insecticides to Canadian Prairie wetlands. Agri. Ecosyst. Environ. 215, 76e84.
bec pesticides strategy 2015-2018. d.l.E.e.d.l.L.c.l.c.c. In: DiMDDELCC, 2015. Que
tat de l’environnement. Ministe re du De
veloppement rection du suivi de l’e durable, p. 24.
bec 2014. http://www. MDDELCC, 2016. Bilan des ventes de pesticides au Que mddelcc.gouv.qc.ca/pesticides/bilan/index.htm. Mineau, P.P.C., 2013. The Impact of the Nation's Most Widely Used Insecticides on Birds. American Bird Conservancy. Mohr, S., Berghahn, R., Schmiediche, R., Hübner, V., Loth, S., Feibicke, M., Mailahn, W., Wogram, J., 2012. Macroinvertebrate community response to repeated short-term pulses of the insecticide imidacloprid. Aquat. Toxicol. 110e111, 25e36. Morrissey, C.A.M Pierre, Devries, James H., Sanchez-Bayo, Francisco, Liess, Matthias, Cavallaro, Michael C., Liber, Karsten, 2015. Neonicotinoid contamination of global surface waters and associated risk to aquatic invertebrates: a review. Environ. Int. 74, 291e303. Myers, C., Hill, E., 2014. Benefits of neonicotinoid seed treatments to soybean production. United States Environmental Protection Agency, Washington, DC, USA. Nauen, R., Ebbinghaus-Kintscher, U., Salgado, V.L., Kaussmann, M., 2003. Thiamethoxam is a neonicotinoid precursor converted to clothianidin in insects and plants. Pesticide Biochem. Physiol. 76, 55e69.
, T.P.A., Olsen, P., Rosenbom, A.E., de Jonge, L.W., 2014. Norgaard, T., Moldrup, P., Ferre Leaching of glyphosate and aminomethylphosphonic acid from an agricultural field over a twelve-year period. Vadose Zone J. 13. ~ a, A., Rodríguez-Lie
bana, J.A., Mingorance, M.D., 2011. Persistence of two Pen neonicotinoid insecticides in wastewater, and in aqueous solutions of surfactants and dissolved organic matter. Chemosphere 84, 464e470. Pisa, L., Amaral-Rogers, V., Belzunces, L., Bonmatin, J., Downs, C., Goulson, D., Kreutzweiser, D., Krupke, C., Liess, M., McField, M., 2014. Effects of

neonicotinoids and fipronil on non-target invertebrates. Environ. Sci. Pollut. Res. 22, 68e102. PMRA, 2016a. Pest Management Regulatory Agency - Health Canada. Consultation on Imidacloprid, Proposed Re-evaluation Decision PRVD2016e20. http://www. hc-sc.gc.ca/cps-spc/pest/part/consultations/_prvd2016-20/index-eng.php. PMRA, 2016b. Pest Management Regulatory Agency - Health Canada. Initiation of Special Reviews: Potential Environmental Risk to Aquatic Invertebrates Related to the Use of Clothianidin and Thiamethoxam. Re-evaluation Note. REV2016e17. http://www.hc-sc.gc.ca/cps-spc/pest/index-eng.php, 4. PPDB, 2016. Pesticide Properties DataBase. University of Hertfordshire. Accessed 20 July 2016. http://sitem.herts.ac.uk/aeru/ppdb/en/index.htm. Roessink, I., Merga, L.B., Zweers, H.J., Van den Brink, P.J., 2013. The neonicotinoid imidacloprid shows high chronic toxicity to mayfly nymphs. Environ. Toxicol. Chem. 32, 1096e1100.

, M., Laflamme, G., Ouellet, L., Carrier, D., Dube
, J.-C., Page
, F., 1984. Etude Rompre
dologique du comte
d'Arthabaska. Ministe re de l'Agriculture, des Pe ^cheries et pe
bec, Direction de la recherche agricole. De
po ^t le
gal de l'Alimentation du Que que nationale du Que
bec, ISBN 2-5511er trimestre 1984, vol. 89. Bibliothe 06164-4. Samson-Robert, O., Labrie, G., Chagnon, M., Fournier, V., 2014. Neonicotinoidcontaminated puddles of water represent a risk of intoxication for honey bees. PloS one 9, e108443. Schaafsma, A., Limay-Rios, V., Baute, T., Smith, J., Xue, Y., 2015. Neonicotinoid insecticide residues in surface water and soil associated with commercial maize (corn) fields in southwestern ontario. PloS one 10, e0118139. Selim, H., Jeong, C.Y., Elbana, T.A., 2010. Transport of imidacloprid in soils: miscible displacement experiments. Soil Sci. 175, 375e381. Shipitalo, M.J., Owens, L.B., 2003. Atrazine, deethylatrazine, and deisopropylatrazine in surface runoff from conservation tilled watersheds. Environ. Sci. Technol. 37, 944e950. Smalling, K.L., Reeves, R., Muths, E., Vandever, M., Battaglin, W.A., Hladik, M.L., Pierce, C.L., 2015. Pesticide concentrations in frog tissue and wetland habitats in a landscape dominated by agriculture. Sci. Total Environ. 502, 80e90. Smit, C., 2014. Water quality standards for imidacloprid: proposal for an update according to the Water Framework directive. RIVM Lett. Rep. 270006001. Stenrød, M., Perceval, J., Benoit, P., Almvik, M., Bolli, R.I., Eklo, O.M., Sveistrup, T.E., Kværner, J., 2008. Cold climatic conditions: effects on bioavailability and leaching of the mobile pesticide metribuzin in a silt loam soil in Norway. Cold Regions Sci. Technol. 53, 4e15. Stone, W.W., Wilson, J.T., 2006. Preferential flow estimates to an agricultural tile drain with implications for glyphosate transport. J. Environ. Qual. 35, 1825e1835. Stoughton, S.J., Liber, K., Culp, J., Cessna, A., 2008. Acute and chronic toxicity of imidacloprid to the aquatic invertebrates Chironomus tentans and hyalella azteca under constant- and pulse-exposure conditions. Archives Environ. Contam. Toxicol. 54, 662e673. TFSP, 2015. The task force on systemic pesticides. Worldw. Integr. Assess. impacts Syst. pesticides Biodivers. Ecosyst. 175. www.tfsp.info/assets/WIA_2015.pdf. USEPA, 2008. Imidacloprid summary document registration review: initial docket. December 2008. In: Agency, U.E.P.. Van Dijk, T.C., Van Staalduinen, M.A., Van der Sluijs, J.P., 2013. Macro-invertebrate decline in surface water polluted with imidacloprid. PloS one 8.


tien, F., et al., Surface runoff and subsurface tile drain losses of neonicotinoids and companion herbicides Please cite this article in press as: Chre at edge-of-field, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.02.002