Neonicotinoid insecticide mixtures: Evaluation of laboratory-based toxicity predictions under semi-controlled field conditions

Accepted Manuscript Neonicotinoid insecticide mixtures: Evaluation of laboratory-based toxicity predictions under semi-controlled field conditions E.M. Maloney, K. Liber, J.V. Headley, K.M. Peru, C.A. Morrissey PII:

S0269-7491(18)32197-3

DOI:

10.1016/j.envpol.2018.09.008

Reference:

ENPO 11550

To appear in:

Environmental Pollution

Received Date: 16 May 2018 Revised Date:

27 August 2018

Accepted Date: 2 September 2018

Please cite this article as: Maloney, E.M., Liber, K., Headley, J.V., Peru, K.M., Morrissey, C.A., Neonicotinoid insecticide mixtures: Evaluation of laboratory-based toxicity predictions under semicontrolled field conditions, Environmental Pollution (2018), doi: 10.1016/j.envpol.2018.09.008. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT

1 2 3

Title: Neonicotinoid insecticide mixtures: evaluation of laboratory-based toxicity predictions under semi-controlled field conditions Maloney, E.M.1, K. Liber1,2,3, J.V. Headley4, K.M. Peru4, and C.A. Morrissey*2,5

5

1

Toxicology Centre, University of Saskatchewan, Saskatoon, Saskatchewan, Canada

6

2

School of Environment and Sustainability, University of Saskatchewan, Saskatoon, Saskatchewan, Canada

7

3

Institute of Loess Plateau, Shanxi University, Taiyuan, Shanxi, P.R. China

8

4

Watershed Hydrology and Ecology Research Division, Water Science and Technology, Environment and Climate

9

Change Canada 5

Department of Biology, University of Saskatchewan, Saskatoon, Saskatchewan, Canada

11

SC

10

RI PT

4

*Corresponding author: [email protected]

13

Abstract

14 15

Neonicotinoid insecticide mixtures are frequently detected in aquatic environments in agricultural regions. Recent laboratory studies have indicated that neonicotinoid mixtures can elicit greater-than-additive toxicity in sensitive aquatic insects (e.g. Chironomus dilutus). However, this has yet to be validated under field conditions. In this study, we compared the chronic (28- and 56-d) toxicity of three neonicotinoids (imidacloprid, clothianidin, and thiamethoxam) and their mixtures to natural aquatic insect communities. Using experimental insitu enclosures (limnocorrals), we exposed wetland insects to single-compounds and binary mixtures at equitoxic concentrations (1 Toxic Unit under the principle of Concentration Addition). We assessed the composition of all emerged insect taxa and cumulative Chironomidae emergence and biomass over time. In all treated limnocorrals there were subtle shifts in community composition, with greater proportions of emerged Trichoptera and Odonata. Cumulative emergence and biomass increased over time and there was a significant interaction between time and treatment. At 28 days, cumulative Chironomidae emergence and biomass were not significantly different between neonicotinoid treatments and controls. However, cumulative emergences in imidacloprid, clothianidin, and clothianidin-thiamethoxam treatments were 42%, 20%, and 44% lower than predicted from applied doses. At 56 days, imidacloprid, clothianidin, and the clothianidin-thiamethoxam mixture elicited significant declines in cumulative emergence and biomass. However, contrary to laboratory predictions, greater-than-additive mixture toxicity was not observed under semi-controlled field settings. Furthermore, only clothianidin significantly shifted sex-ratios towards female-dominated populations. Results showed that the responses of natural Chironomidae populations to neonicotinoids and their mixtures cannot be adequately predicted from laboratory-derived single-species models, and that reductions in Chironomidae emergence and biomass can occur at neonicotinoid concentrations below some current water quality guidelines (albeit effects may have been attenuated by occasional

21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37

TE D

19 20

EP

18

AC C

16 17

M AN U

12

Maloney, 1

ACCEPTED MANUSCRIPT

38 39 40

overdosing). Therefore, these neonicotinoid guidelines should be reviewed and amended to ensure that Chironomidae and other sensitive aquatic insects inhabiting agricultural wetlands are adequately protected.

41

RI PT

Highlights:

Chronic toxicity of neonicotinoids and mixtures was evaluated for wetland insects. Contrary to predictions, mixtures were not more toxic than single neonicotinoids. Some neonicotinoids/mixtures had greater-than-predicted effects on emergence and biomass. Lab-derived mixture models did not adequately predict field-based mixture effects.

43 44 45 46 47

SC

42

48

Keywords:

50

Neonicotinoids, mixture toxicity, limnocorrals, aquatic insects, Chironomidae.

M AN U

49

51

Capsule: Lab-derived, single species models cannot adequately predict single compound and neonicotinoid mixture toxicity to natural Chironomidae populations under semi-controlled field conditions.

55

56

1

Introduction

TE D

52 53 54

Neonicotinoid insecticides are frequently detected in global surface waters (Anderson et

58

al., 2015; Morrissey et al., 2015). Widespread use and relatively high application rates have led

59

to the detection of neonicotinoids at concentrations ranging from <0.001 – 70.0 µg/L in marine

60

(Smith et al., 2012), freshwater lotic (e.g. wetlands and ponds) (Evelsizer and Skopec, 2016;

61

Lamers et al., 2011; Main et al., 2014; Schaafsma et al., 2015), and freshwater lentic (e.g. rivers

62

and streams) (Hladik and Kolpin, 2015; Sánchez-Bayo and Hyne, 2014) ecosystems across the

63

world. Due to a combination of their physicochemical characteristics (e.g. high water-solubility,

64

limited degradation in light-limited environments) (Lu et al., 2015) and current agricultural

65

application practices (e.g. field-rotation with different neonicotinoid-treated crops and

AC C

EP

57

Maloney, 2

ACCEPTED MANUSCRIPT

application of different compounds in the same watershed), neonicotinoids are commonly

67

detected as mixtures. For example, in a recent study of Canada’s Prairie Pothole Region (PPR),

68

an ecologically important area containing over a million shallow wetlands, binary or ternary

69

neonicotinoid mixtures were detected in 11 – 63% of wetlands sampled (n = 136 wetlands) at

70

cumulative concentrations ranging from 0.004 – 1.66 µg/L (Main, 2016). Other environmental

71

monitoring studies have reported similar findings, detecting neonicotinoid mixtures in

72

rivers/streams (Hladik and Kolpin, 2015), surface waters (Schaafsma et al., 2015; Smalling et al.,

73

2015), and groundwater (Giroux and Sarrasin, 2011) across North America.

SC

M AN U

74

RI PT

66

Despite the prevalence of neonicotinoid mixtures in aquatic environments, a relatively limited number of studies have focussed on characterizing the cumulative toxicity of

76

neonicotinoid mixtures to non-target organisms. As neonicotinoids all have a common

77

mechanism of action, eliciting neurotoxicity by binding to and continuously activating nicotinic

78

acetylcholine receptors (nAChR), it is typically assumed that neonicotinoid mixtures will elicit

79

directly-additive cumulative toxicities. Therefore, the risk of neonicotinoid mixtures to non-

80

target organisms is typically estimated using a predictive model based on the direct addition of

81

constituent concentrations (i.e. Concentration Addition). Furthermore, current environmental

82

regulations in Canada, the US, and Europe exclusively focus on single neonicotinoid compounds

83

(Canadian Council of Ministers of the Environment, 2007; Morrissey et al., 2015; Smit, 2014;

84

USEPA, 2017), and exclude potential cumulative effects of neonicotinoid mixtures. However,

85

recent studies have shown that the assumption of directly-additive toxicity may not be

86

adequately protective of non-target organisms. For example, we have previously reported that

87

binary and ternary mixtures of imidacloprid, clothianidin, and thiamethoxam can have greater

88

than additive effects in the freshwater midge Chironomus dilutus (C. dilutus) under acute (96 h)

AC C

EP

TE D

75

Maloney, 3

ACCEPTED MANUSCRIPT

and to a lesser extent, chronic (28 d), laboratory conditions (Maloney et al., 2017, 2018).

90

Similarly, deviation from the assumption of directly additive cumulative toxicity has been

91

reported for mixtures of other neonicotinoids (e.g. imidacloprid-thiacloprid) in laboratory assays

92

using water fleas (Daphnia magna) and roundworms (Caenorhabditis elegans) (Gomez-Eyles et

93

al., 2009; Loureiro et al., 2010; Pavlaki et al., 2011). However, the broader conclusions that can

94

be drawn from these neonicotinoid mixture toxicity assessments are relatively limited. This is

95

primarily because the cumulative toxicity of neonicotinoid insecticides appears to vary

96

depending on exposure time (e.g. chronic vs. acute), endpoint of interest (e.g. emergence, body

97

length, or mortality), and experimental organism (e.g. D. magna vs. C. dilutus). Furthermore,

98

most studies have been laboratory-based, focussing on the impacts of neonicotinoid mixtures on

99

single species under environmentally controlled conditions. None of the reported mixture effects

M AN U

SC

RI PT

89

have been validated under more field-realistic conditions. Therefore, we currently do not know

101

how neonicotinoid mixtures may impact the organisms (and populations/communities) that

102

typically inhabit contaminated aquatic environments.

TE D

100

Due to their distinct life-cycle traits (e.g. larvae and pupae are typically aquatic or semi-

104

aquatic and these stages often make up a majority of the life cycle), aquatic insects are likely to

105

be exposed to neonicotinoid mixtures receiving agricultural runoff in wetland environments like

106

the PPR. Furthermore, as they are often physiologically similar to target pests, freshwater insects

107

are likely to be sensitive to neonicotinoid mixture toxicity. Non-biting midges (Chironomidae)

108

have been shown to be particularly sensitive to neonicotinoids (Morrissey et al., 2015). This is of

109

concern because Chironomidae are ecologically important in aquatic environments. Indeed, as

110

the most ubiquitous and abundant insect family in freshwater ecosystems (Bataille and

111

Baldassarre, 1993), Chironomidae play vital roles as primary consumers and act as food sources

AC C

EP

103

Maloney, 4

ACCEPTED MANUSCRIPT

for a range of high trophic-level consumers including fish, water-fowl and other birds, predatory

113

insects, and bats (Benoit et al., 1997). Despite the high likelihood of their exposure to

114

neonicotinoid mixtures, their sensitivity to neonicotinoids, and their relative importance in

115

freshwater (and proximal terrestrial) ecosystems, to the best of our knowledge, the effects of

116

neonicotinoid mixtures on natural Chironomidae populations have yet to be evaluated under

117

field-realistic exposure conditions.

RI PT

112

In this study, we aimed to characterize the effects of chronic (28 and 56 d) exposure to

SC

118

three common neonicotinoid insecticides (e.g. imidacloprid (IMI), clothianidin (CLO), and

120

thiamethoxam (TMX)) and their binary mixtures (e.g. IMI-CLO, CLO-TMX, and IMI-TMX) to

121

natural Chironomidae populations under semi-controlled field conditions using limnocorrals (in-

122

situ shallow wetland enclosures). We compared the effects of single-compounds to the effects of

123

binary mixtures at concentrations estimated to elicit equivalent toxicity (1 Toxic Unit based on

124

the principle of Concentration Addition, derived from chronic (28 d) laboratory studies. This

125

design allowed us to determine if current risk assessment and regulatory approaches using single

126

compound EC50 values are adequately protective of natural Chironomidae populations under

127

semi-controlled field conditions or, as demonstrated in prior laboratory studies, if neonicotinoid

128

mixtures also display greater-than-additive cumulative toxicity.

129

2

130 131 132

2.1

133

(Pond 2) at the St. Denis National Wildlife Area (NWA), Saskatchewan, Canada (52°12’N/

134

106°5’W). The St. Denis NWA, currently under the management of Environment and Climate

135

Change Canada, is a protected area primarily used for research on the influences of agricultural

AC C

EP

TE D

M AN U

119

Materials and Methods

Study Site

Experiments were conducted in a single permanent (class V) prairie pothole wetland

Maloney, 5

ACCEPTED MANUSCRIPT

practices on waterfowl and wetland-dependent fauna habitat in the PPR. Pond 2 was selected due

137

to consistent inter-annual central water depth (1.0 – 1.3 m), water quality characteristics (e.g.

138

intermediary pH, conductivity, and dissolved oxygen), high insect secondary production, and

139

water inflow from ‘untreated’ agricultural land (i.e. unexposed to pesticides). Prior use of this

140

wetland for other semi-controlled field experiments provided historical reference values for

141

insect abundance, composition, peak emergence times, and water quality (Cavallaro et al., 2018).

142

Despite prior studies, neonicotinoid contamination of test sites was not of concern as the

143

limnocorrals were placed in different locations in Pond 2 than used in Cavallaro et al. (2018)

144

(e.g. at the periphery rather than the pond centre) and water quality testing prior to the initiation

145

of this study (~16 days of water column monitoring pre-dosing) indicated that there were no

146

traces of neonicotinoid contamination within the isolated water columns of the experimental

147

wetland. Sediment sampling was not carried out to avoid disturbance of the benthic community

148

prior to the exposure period.

149 150 151

2.2

152

depth: < 1.5 m) purchased from Curry Industries Ltd (Winnipeg, MB, Canada) were employed in

153

the experimental wetland. Limnocorral design was modified from Cavallaro et al. (2018).

154

Polyethylene sleeves were fastened to polyvinyl-encased Styrofoam floats (top), and the open

155

bottom of the sleeves were sealed with a heavy steel chain buried in the sediment (~15 cm

156

depth). Sediment seal was confirmed using an Aqua Scope IITM underwater viewer (Rickly

157

Hydrological Co., Inc., Columbus, OH, USA). Limnocorrals were fitted with custom built

158

aquatic insect emergence traps. These covered the entire open-air limnocorral surface and

159

featured a removable acrylic collection chamber leading to a polypropylene jar containing ~ 400

TE D

M AN U

SC

RI PT

136

Experimental Design

AC C

EP

Twenty-one custom-built limnocorrals (fixed size: 1.0 x 1.0 m diameter; adjustable

Maloney, 6

ACCEPTED MANUSCRIPT

mL of 70% ethanol for emerged insect collection. Due to variable weather patterns in the St.

161

Denis NWA region (e.g. strong winds and storms), wooden stakes secured to the corners via a

162

nylon rope-ABS plumber’s pipe ring (10 cm diameter) were used to anchor the limnocorral

163

floats, allowing them to rise and fall with the natural water level. Extra material in the

164

limnocorral walls allowed for small changes in water volume (i.e. expansion without disruption

165

of the sediment seal). Biological heterogeneity within the wetland and seasonal water-depth

166

variation was controlled for by randomizing treatments across three experimental blocks (3 x 7).

167

Limnocorrals were secured at water depths of approximately 0.7 m. Limnocorral volumes,

168

calculated from measured depths (i.e. L x W x H), ranged between 442. 3 – 793.8 L (mean:

169

605.4 ± 92.9 L) and were not significantly different among neonicotinoid treatments or

170

experimental blocks (one-way Analysis of Variance (ANOVA), p > 0.05). To avoid disturbing

171

either the limnocorral sleeves or the sediments after the limnocorrals were secured in the

172

experimental wetland, all experimental work (i.e. dosing, water sampling, and insect sampling)

173

was conducted from a canoe.

174

2.3 2.3.1

SC

M AN U

TE D

Neonicotinoids Experimental Compounds

EP

175 176 177

RI PT

160

Technical grade neonicotinoid insecticides were used as experimental compounds in this study. Imidacloprid (IMI) (98.8% pure: N-{1-[(6-chloropyridin-3-yl)methyl]-4,5-

179

dihydroimidazol-2-yl]nitramide) and clothianidin (CLO) (99.6% pure: 1-[(2-chloro-1,3-thiazol-

180

5-yl)methyl]-2-methyl-3-nitroguanidine) were acquired from Bayer CropScience (Kansas City,

181

MO, USA). Thiamethoxam (TMX) (98.8% pure: (NE)-N-[3-[(2-chloro-1,3-thiazol-5-yl)methyl]-

182

5-methyl-1,3,5-oxadiazinan-4-ylidene]nitramide) was acquired form Syngenta Crop Protection

183

LLC (Greensboro, NC, USA). Stock solutions were prepared by dissolving technical product in

184

reverse-osmosis water (Barnstead® DiamondTM NANOpure, 18 megaohm/cm, Barnstead

AC C

178

Maloney, 7

ACCEPTED MANUSCRIPT

International, Dubuque, IA, USA) and solutions were stored in amber glass bottles (4°C in the

186

dark) until experimental use.

187 188 189

2.3.2 Neonicotinoid Treatments

190

TMX) or binary mixtures (e.g. IMI-CLO, CLO-TMX, IMI-TMX) every 4 days for 56 days.

191

Each treatment was replicated three times (n = 3). Three experimental controls (untreated) were

192

also employed, yielding a total of 21 experimental limnocorrals. Neonicotinoid exposures began

193

in spring, following ice-off and complete sediment thaw, and continued into early summer (May

194

– July). Exposure lengths (e.g. 28 and 56 d) and conditions (e.g. continuous rather than pulse

195

exposure) were chosen to allow for direct comparison of prior laboratory studies (Maloney et al.,

196

2018) to our semi-controlled field study. Overall exposure length (56 d) was selected based on a

197

previous study by Cavallaro et al. (2018) that indicated that peak emergence of indigenous

198

Chironomidae in a Prairie wetland occurred within 50 days of ice-off and thus 56 days was

199

expected to capture the full life cycle. Therefore, responses of single-compound and binary

200

mixture treated limnocorrals were compared at 28 d of exposure (study midpoint) to emulate the

201

exposure lengths of prior laboratory studies and 56 d to evaluate effects on the full life cycle.

EP

TE D

M AN U

SC

Limnocorrals were treated with single neonicotinoid compounds (e.g. IMI, CLO, or

Neonicotinoid concentrations were selected based on toxicity thresholds (28-d EC50

AC C

202

RI PT

185

203

values for successful emergence) determined for a representative aquatic insect species

204

(Chironomus dilutus) in previous, laboratory studies (Maloney et al., 2018). To allow for direct

205

comparisons of effects, all treatments with single-compounds or binary neonicotinoid mixtures

206

were tested at equivalent toxic unit (TU) dose-ratios (1:1), and dose-levels (ΣTU = 1), yielding

207

three different binary mixtures and three single-compound exposures. Here a TU was defined as

Maloney, 8

ACCEPTED MANUSCRIPT

208

the concentration (c) of a compound (e.g. IMI, CLO, or TMX) divided by its toxicity threshold

209

(28-d EC50) (Equation 1): (1)

=

211

To determine if binary mixtures elicited greater-than-additive (synergistic) or less-than-additive

212

(antagonistic) effects, toxicological response (i.e. Chironomidae emergence) was compared

213

directly between limnocorrals treated with single compounds and theoretically equitoxic binary

214

mixtures. Under the principle of Concentration Addition, treatments with equivalent ΣTU should

215

yield equivalent toxicological responses (i.e. 50% of control response). Therefore, mixtures

216

eliciting significantly higher toxicological responses than the single compound exposures were

217

deemed ‘greater-than-additive’ and mixtures eliciting significantly lower toxicological responses

218

than the single compound exposures were deemed ‘less-than-additive’.

SC

M AN U

Neonicotinoid concentrations in limnocorrals were measured prior to (pre-dose) and 1 h

TE D

219

RI PT

210

after (post-dose) dosing events (Section 2.2.3). Between dosing periods, treatment-specific

221

dissipation rates were estimated for each neonicotinoid compound. Dissipation rates were

222

estimated by calculating the mean (± standard deviation) percent of active ingredient remaining

223

in each limnocorral 4 d after dosing. Treatment- and limnocorral-specific 4-d dissipation rates,

224

combined with estimated enclosure volumes, allowed us to estimate the concentrations of

225

neonicotinoids in each limnocorral and then adjust dosing accordingly to maintain constant

226

neonicotinoid exposures. Dosing solutions were prepared fresh for each treatment day by spiking

227

250 mL of culture water (carbon-filtered, bio-filtered Saskatoon municipal water, aerated in a

228

50-L Nalgene® carboy for >24 h prior to use) with concentrated neonicotinoid stock solutions to

229

achieve nominal test concentrations as follows: IMI (single compound = 0.50 µg/L; in binary

AC C

EP

220

Maloney, 9

ACCEPTED MANUSCRIPT

mixtures = 0.25 µg/L), CLO (single compound = 0.71 µg/L; in binary mixtures = 0.36), and

231

TMX (single compound = 8.91 µg/L; in binary mixtures = 4.46 µg/L). Dosing solutions were

232

then transferred into 250-mL amber bottles and transported in coolers to the study site (to limit

233

thermal and photo-degradation). During each dosing event, test solutions were poured directly

234

into limnocorrals, and then each limnocorral was gently stirred with a paddle to ensure mixing

235

and equivalent neonicotinoid distribution throughout. To control for potential disturbance-related

236

effects on aquatic insect communities, untreated controls were subjected to the same treatment as

237

the other limnocorrals (e.g. addition of 250 mL untreated culture water and stirring each dosing

238

day). Further mixing occurred through natural turnover within the pond, via wind and

239

temperature-guided water density changes.

240 241 242

2.3.3

243

neonicotinoid contamination (days -12, -8, -4, -2). Over the course of the study, water samples

244

were collected from each limnocorral and analyzed to measure actual concentrations of

245

neonicotinoid exposure. Pre- and post-dosing, water samples were collected from each

246

limnocorral (every 4 d over the course of the study). Pre-dose samples were collected

247

immediately before limnocorral dosing. Post-dose samples were collected 1-h after limnocorral

248

dosing (to allow limnocorrals to equilibrate after dosing/mixing events). Subsurface water

249

samples were collected grab-style in 250-mL amber bottles, from the centre of each limnocorral

250

at >30 cm depth below the surface and stored at 4°C in the dark until time of analysis (within 14

251

d).

252 253

M AN U

SC

RI PT

230

Water Sampling and Neonicotinoid Analysis

AC C

EP

TE D

Prior to the initiation of the study, water samples were collected from Pond 2 to assess for

All samples were analyzed at the National Hydrology Research Centre, Environment and Climate Change Canada, Saskatoon, as described in Main et al. (2014). Briefly, neonicotinoid Maloney, 10

ACCEPTED MANUSCRIPT

concentrations were quantified by solid-phase extraction (SPE) followed by high performance

255

liquid chromatography paired with tandem mass spectrometry (LC-MS/MS). Methods adapted

256

by Xie et al. (2011) allowed for simultaneous extraction and determination of IMI, CLO, and

257

TMX concentrations in aqueous samples. SPE was performed using OASIS® HLB cartridges

258

(Waters, Missisauga, ON, Canada). Internal standards (d4-imidacloprid, and d3-thiamethoxam)

259

were obtained from CDN Isotopes (Pointe-Claire, QC, Canada). LC-MS/MS was performed

260

using a Waters 2695 Alliance HPLC system (Waters Corp., Milford, MA, USA) equipped with a

261

Waters XTerra MS-C8 column, paired with a Micromass Quattro Premier triple quadrupole mass

262

spectrometer (Waters Corp., Milford, MA, USA) equipped with an electrospray ionization

263

interface (positive ion mode). Analytical standards obtained from Chem Service (West Chester

264

PA, USA) were used to create calibration curves and determine neonicotinoid recoveries. Limits

265

of quantification (LOQ) and recovery correction factors (RC %) were as follows: IMI (LOQ =

266

0.0018 – 0.0043 µg/L, (mean: 0.0028 µg/L); RC = 85.5 – 94.6 % (mean: 90.8 %)), CLO (0.0026

267

– 0.0056 µg/L, (mean: 0.0034 µg/L); RC = 65.2 – 74.7 % (mean: 69.8%)), and TMX (LOQ=

268

0.0042 – 0.0077 µg/L (mean: 0.0058 µg/L); RC = 79.2 – 100.9 % (mean: 88.3%)). Measured

269

neonicotinoid concentrations in limnocorrals were recovery corrected and averaged across

270

sample days. Mean measured concentrations were subsequently used in all statistical analyses.

271 272 273

2.4

274

dosing day using a hand-held Thermo ORION® dissolved oxygen meter, model 835 (Thermo

275

ORION, Beverly MA, USA). Due prior studies indicating their relative stability in the

276

limnocorrals over time (Cavallaro et al., 2018), all other water quality parameters were only

277

assessed at the beginning (d 0), middle (d 28), and end (d 56) of the study. Water samples (50

AC C

EP

TE D

M AN U

SC

RI PT

254

Water Quality

Dissolved oxygen (DO) and temperature were measured in each limnocorral on every

Maloney, 11

ACCEPTED MANUSCRIPT

mL) were removed from limnocorrals and analyzed in the laboratory for nitrate, phosphate, pH,

279

conductivity, total hardness, alkalinity, ammonia (NH3), and dissolved organic carbon (DOC)

280

(<12 h after sampling). Phosphate (PO4) and nitrate (NO3) were measured with an YSI EcoSense

281

photometer, model 9300 (YSI Inc., Yellow Springs, OH, USA); pH was measured with an

282

ORION® PerpHect LogR meter, model 170 (ORION Research, Beverly, MA, USA); hardness

283

and alkalinity were measured using a Hach Digital Titrator, model 16900 (Hach Company,

284

Loveland, CO, USA); ammonia was measured with a VWRTM SB301 sympHony ISE ammonia

285

meter (VWR International Lt, West Chester, PA, USA), paired with a Thermo ORION® 95-12

286

ammonia electrode (Thermo ORION, Beverly, MA, USA); and DOC was measured using a

287

Shimadzu Total Organic Carbon Analyzer (TOC-V), CPN model 5000 (Shimadzu Co., Kyoto,

288

Japan).

289 290 291

2.5

292

(every 4 d) over the course of the exposure period (56 d) to obtain cumulative abundance. To

293

determine a baseline level for limnocorral productivity, insects were also collected prior to the

294

dosing period (d -12, -8, -4, -2, and 0). Insects were collected by changing the entire collection

295

jar, and organisms were stored at 4°C (preserved in fresh 70% ethanol) until identified and

296

counted. All collected insects first were identified to order. Diptera were further identified to

297

family using dichotomous keys (Merritt et al., 2008). Adult Chironomidae were subsequently

298

sexed. Male chironomids were separated from females via their more slender abdomen, genital

299

appendages, and (typically) plumose antennae (Merritt et al., 2008). To determine the subfamily

300

composition of collected Chironomidae, a subset of samples from control treatments were

301

collected and sent to the Water Security Agency of Saskatchewan, Saskatoon, SK, Canada to be

M AN U

SC

RI PT

278

Insect Sampling, Identification, and Analysis

AC C

EP

TE D

Adult insects were collected from the polypropylene sample jars on each emergence trap

Maloney, 12

ACCEPTED MANUSCRIPT

further identified. Briefly, larval head capsules were removed, and soft tissue was dissolved in

303

10% KOH solution. Head capsules were then mounted on glass slides using Euparal mounting

304

medium, and dried for 48 h. Taxa-specific keys and literature were subsequently used to

305

accurately identify unique midge fauna (Diptera: Chironomidae) (Hirvenoja, 1973; Merritt et al.,

306

2008; Oliver, 1976; Oliver and Rousssel, 1983). A voucher series, linking the identified

307

Chironomidae to this research project, was deposited in the Water Security Agency of

308

Saskatchewan Invertebrate Voucher Collection, Saskatoon, SK, Canada. All remaining collected

309

Chironomidae were oven-dried (@ 60°C for 24-h) and weighed to evaluate total biomass for

310

each experimental day.

2.6

SC

M AN U

311 312 313 314

RI PT

302

Data Analysis

Water quality variables and neonicotinoid concentration data were averaged across time and then compared between treatments using one-way analysis of variance (ANOVA) analyses

316

paired with Tukey’s post hoc analysis for pairwise comparisons between treatments (95 % level

317

of confidence, α = 0.05). Initial (baseline) cumulative total insect and Chironomidae abundances

318

in limnocorrals were also compared across treatments using one-way ANOVAs paired with

319

Tukey’s post-hoc tests (95% level of confidence, α=0.05). As baseline insect abundances did not

320

significantly differ among treatments (one-way ANOVA, p > 0.05) (Table A4), baseline insect

321

abundance was not included as a variable in further analysis. Effects of three experimental

322

wetland blocks on abundance (total insect and Chironomidae) and biomass were tested using

323

two-way ANOVAs, using experimental block and neonicotinoid treatment as factors (95% level

324

of confidence, α = 0.05). Similarly, experimental blocks did not significantly affect cumulative

325

abundance (total insect or Chironomidae) or biomass (two-way ANOVA, pinteraction > 0.05), and

326

therefore block was not included as a variable in further analysis.

AC C

EP

TE D

315

Maloney, 13

ACCEPTED MANUSCRIPT

The mean, cumulative proportion of emerged insects in each order was compared

327

amongst treatments using chi-square tests (λ2, 95% level of confidence, α = 0.05). Prior to

329

statistical analysis, data (cumulative Chironomidae abundance and biomass) were transformed

330

[log(x +1)] to meet normality (Shapiro-Wilk, α = 0.05) and equality of variance (Brown-

331

Forsythe, α = 0.05) assumptions. To determine time-weighted treatment effects, cumulative

332

Chironomidae abundance and biomass were compared across neonicotinoid treated limnocorrals

333

over the course of the study (d 0 – 56) using a two-way repeated measures (RM) ANOVA (fixed

334

effect of time + treatment + time x treatment interaction) paired with the Holm-Sidak method for

335

post-hoc pairwise comparisons (95% level of confidence, α = 0.05). Cumulative Chironomidae

336

abundance and biomass were analysed at the two key time points, d 28 (study midpoint, allowing

337

for direct comparisons to prior 28-d laboratory-based tests) and d 56 (study culmination, aimed

338

to capture a more complete Chironomidae life cycle) using univariate statistical tests (i.e. one-

339

way ANOVAs paired with Tukey’s post hoc tests (95% level of confidence, α = 0.05), to

340

evaluate differences between neonicotinoid treatments. Significant differences in sex of emerged

341

Chironomidae at days 28 and 56 were assessed by comparing the mean proportion of emerged

342

males in treatment groups to experimental controls using z-tests (95% level of confidence, α =

343

0.05). All statistical analyses were performed using SigmaPlotTM Version 13.0 (Systat Software

344

Inc., San Jose, CA, USA).

345

AC C

EP

TE D

M AN U

SC

RI PT

328

346

3

Results

347 348 349

3.1

350

deviate among individual limnocorrals, across experimental blocks, or among neonicotinoid

Water Quality General water quality remained relatively consistent across time, and did not significantly

Maloney, 14

ACCEPTED MANUSCRIPT

treatments (One-way ANOVA tests, p>0.05) (Table A1). Mean (±SE) water quality parameters

352

during the dosing period were as follows: DO 5.2 (± 0.7) mg/L, temperature 20.5 (± 0.4) °C, pH

353

8.1 (± 0.0), conductivity 2934 (± 53) µS/cm3, total hardness 1708 (± 59) mg/L as CaCO3,

354

alkalinity 414 (± 26) mg/L as CaCO3, phosphate 1.3 (± 0.5) mg/L, nitrate 3.97 (± 0.89) mg/L,

355

ammonia 0.7 (± 0.3) mg/L, and DOC 32.0 (± 2.4) mg/L.

356 357 358

3.2

359

presented in Table 1. Over the course of the study, measured neonicotinoid concentrations

360

remained within 99.2 ± 28.7 %, 104.4 ± 30.5 %, and 102.8 ± 41 % of the target nominal doses,

361

respectively (Tables A2 and A3). Mean cumulative toxic units (ΣTU) ranged from 0.7 - 1.20

362

(target ΣTU = 1.0) and were not significantly different between treatments (one-way ANOVA, p

363

> 0.05). Occasional variation in neonicotinoid concentrations occurred between dosing days (e.g.

364

target nominal concentrations were exceeded on some dosing days, with maximum ΣTUs

365

ranging from 1.37 – 2.23 (Table A3)); however, this deviation from the target concentration was

366

relatively consistent across treatments. Despite experimental precautions, there were occasional

367

detections of low-level neonicotinoid contamination between treated limnocorrals (e.g. TMX

368

detection in IMI treatments in 3/42 analyzed samples) and in the experimental controls (Tables

369

A2 and A3). However, these were typically ΣTU < 0.01 (or < 5% of the lowest target dose), and

370

were included in the calculated ΣTU, thus it is unlikely that this significantly influenced

371

experimental results. Notably, in all TMX treated limnocorrals we observed slight degradation of

372

TMX into CLO. On average, 0.32 ± 0.43% of TMX was degraded into CLO (v/v) between 4 d

373

dosing periods.

Measured Neonicotinoid Concentrations

RI PT

351

AC C

EP

TE D

M AN U

SC

Mean measured IMI, CLO, and TMX concentrations in treated limnocorrals are

Maloney, 15

ACCEPTED MANUSCRIPT

374 375 376

3.3

Insect Diversity

377

collected from experimental limnocorrals are presented in Table 2. Across treatments, Diptera

378

were the dominant taxa, accounting for 92.2 ± 3.6 % of the total number of emerged insects over

379

the course of the study. Overall, Chironomidae were the most abundant insect family, accounting

380

for 91.4 ± 3.3 % of the total number of emerged insects across treatments. Within the

381

Chironomidae, there was relatively low subfamily diversity: 93.3 ± 3.9% were Chironominae,

382

4.3 ± 2.4% were Orthocladiinae, and 2.4 ± 1.2% were Tanypodinae. The remaining insect taxa

383

primarily consisted of Odonata (4.1 ± 6.1 %) and Trichoptera (2.2 ± 1.3 %). A limited number of

384

Hymenoptera (0.5 ± 0.3 %), Coleoptera (0.4 ± 0.4 %) and Ephemeroptera (0.0 ± 0.0 %) were

385

also collected.

M AN U

SC

RI PT

Mean (± standard deviation (SD)) cumulative proportions of emerged adult insects

Neonicotinoid and neonicotinoid mixture exposure subtly altered the insect community

387

composition. For most orders (i.e. Diptera, Hymenoptera, Coleoptera, and Ephemeroptera), the

388

mean proportion of emerged adult insects (d 56, cumulative) was not significantly different

389

between treated and untreated limnocorrals, or among single compounds and binary mixtures

390

(Chi-square test, p > 0.05). However, statistically-significant, treatment-specific differences were

391

found in the cumulative proportions of emerged Trichoptera (Chi-square test, χ2 = 25.0, p = <

392

0.001) and Odonata (Chi-square test, χ2 = 100.6, p < 0.001) over the course of the study (e.g. d

393

56 cumulative). In all treatments, there were increases in the mean proportions of emerged

394

Trichoptera relative to controls (e.g. 0.03 ± 0.1 % in controls vs. 1.1 ± 0.9 – 3.7 ± 5.3 % in

395

treatments), and in IMI, IMI-CLO, CLO, and CLO-TMX treatments there were statistically

396

significant increases in the mean proportions of emerged Odonata relative to controls (e.g. 1.2 ±

397

1.1 % in controls vs. 4.8 ± 1.6 % - 9.3 ± 14.7 % in treatments) However, due to the large amount

AC C

EP

TE D

386

Maloney, 16

ACCEPTED MANUSCRIPT

of variation amongst experimental replicates, the cumulative total abundances of Trichoptera and

399

Odonata in treated limnocorrals (Table A5) were not significantly different from controls (one-

400

way ANOVA, p > 0.05).

401 402

3.4 3.4.1

RI PT

398

Chironomidae Abundance and Biomass Time-weighted Treatment Effects

Cumulative abundance of emerged Chironomidae, the most abundant taxa, increased over

403

time (two-way RM ANOVA, F = 129.8, p < 0.001) (Fig. 1A). In the time-weighted analysis,

405

treatment was statistically insignificant (two-way RM ANOVA, F = 2.20, p = 0.11). However,

406

there was a significant interaction between treatment and exposure time across all treatments

407

(two-way RM ANOVA, F = 1.39, p = 0.03), indicating that although across time there were no

408

significant differences in cumulative Chironomidae emergence among neonicotinoid treatments,

409

time was a significant factor in treatment effects. Similarly, in all treatments, cumulative biomass

410

increased over the course of the study (Fig. 1B) (two-way RM ANOVA, F = 29.9, p < 0.001),

411

treatment was statistically insignificant in the time-weighted analysis (two-way RM ANOVA,

412

F= 2.379, p = 0.085), and there was a significant interaction between treatment and exposure

413

time (two-way RM ANOVA, F = 2.2, p < 0.001). Therefore, for cumulative biomass, across time

414

there were no significant differences in cumulative Chironomidae biomass among treatments,

415

and time was a significant factor in treatment effect. (Fig. 1B).

416

3.4.2

M AN U

TE D

EP

AC C

417

SC

404

Single Compounds

On d 28 (study midpoint), there were no statistically significant impacts of IMI, CLO,

418

and TMX on cumulative Chironomidae abundance or biomass (relative to the controls) (one-way

419

ANOVA, p > 0.05) (Fig. 1C, D). However, IMI and CLO treatments still reduced cumulative

420

abundance relative to controls by 8.0 ± 2.6 % (IMI) and 32.2 ± 6.5% (CLO) (vs. 50% expected)

421

(Fig. 1C). A similar trend occurred for mean cumulative Chironomidae biomass relative to Maloney, 17

ACCEPTED MANUSCRIPT

controls (13.8 ± 11.5 % (IMI), and 15.0 ± 6.6 % (CLO)) (Fig. 1D). At the termination of the

423

study (d 56), IMI and CLO treatments had elicited statistically significant declines in cumulative

424

Chironomidae abundance (Fig. 1E) (Tukey post-hoc test, n = 20, F = 5.1, p = 0.008). Indeed,

425

relative to the control, mean cumulative emergences were as follows: 10.0 ± 1.4 % (IMI) and

426

14.6 ± 5.5 % (CLO). Cumulative biomass was similarly affected. At the study termination (d 56),

427

IMI and CLO treatments had elicited statistically significant declines in the cumulative biomass

428

of emerged Chironomidae (Fig. 1F) (Tukey post-hoc test, n = 20, F = 3.2, p = 0.03), with mean

429

cumulative biomasses relative to controls as follows: 20.4 ± 6.0 % (IMI) and 18.6 ± 8.6 %

430

(CLO). However, the TMX treatment did not elicit a statistically significant decline in either

431

cumulative Chironomidae abundance or biomass relative to controls on d 56: 29.0 ± 7.9 %

432

(abundance) and 48.3 ± 6.8% (biomass) (Figs. 1E and 1F).

433 434 435

3.4.3

436

predictions. On d 28, there were (statistically insignificant) declines in mean cumulative

437

abundance (e.g. % mean emergence relative to controls (5.8 ± 1.7% vs. 50% expected) and mean

438

biomass relative to controls (2.5 ± 0.1 %) in the CLO-TMX treated limnocorrals (Fig. 1C and

439

1D). At the study culmination (d 56), there were statistically significant reductions in

440

Chironomidae emergence (one-way ANOVA, n = 20, F = 5.2, p = 0.008) and biomass (one-way

441

ANOVA, n = 20, F = 3.2, p = 0.03) in the CLO-TMX treated limnocorrals, relative to controls,

442

(Figs. 1E and 1F) with a mean relative cumulative emergence of 14.6 ± 12.7 % and a mean

443

relative cumulative biomass of 13.5 ± 9.1 %. However, partially due to the high variability in all

444

neonicotinoid (single compound and mixture) treatments, declines in emergence and biomass in

445

CLO-TMX treatments were not significantly different from single compound treatments (Tukey

M AN U

SC

RI PT

422

Binary Mixtures

AC C

EP

TE D

Binary neonicotinoid mixtures also elicited effects that deviated from lab-based

Maloney, 18

ACCEPTED MANUSCRIPT

post-hoc tests, p > 0.05). Therefore, although declines in cumulative abundance and biomass in

447

the CLO-TMX treated limnocorrals were greater-than-predicted, they could not be categorized as

448

greater-than-additive. Furthermore, due to high variation between treated limnocorrals, IMI-CLO

449

and IMI-TMX mixtures did not elicit statistically significant declines in cumulative emergence

450

or biomass relative to controls. There were no significant differences in cumulative

451

Chironomidae abundance or biomass between IMI-CLO and IMI-TMX treatments and the

452

controls on d 28 (Figs. 1C and 1D). Indeed, relative to controls, cumulative emergences were

453

69.5 ± 67.3 % (IMI-CLO) and 54.0 ± 86.0 % (IMI-TMX) (compared to the 50% expected), and

454

relative cumulative biomasses showed similar trends (mean (± SD) % relative to control: 72.1 ±

455

64.0 % (IMI-CLO) and 83.7 ± 134 % (IMI-TMX)) On d 56, cumulative emergence relative to

456

controls for these mixtures were 32.5 ± 18.6 % (IMI-CLO) and 37.7 ± 42.4 % (IMI-TMX) (Fig.

457

1E), and cumulative biomasses (relative to controls) were 52.3 ± 20.0 % (IMI-CLO) and 66.6 ±

458

54.8 % (IMI-TMX) (Fig. 1F). Similarly, toxicological responses in IMI-CLO and IMI-TMX

459

treated limnocorrals were not significantly different from single-compound treatments (Tukey

460

post-hoc tests: p > 0.05), thus the mixture effects had to be categorized as directly additive

461

(concentration addition).

462 463 464

3.5

465

on the sex of emerged Chironomidae at d 28 (study mid-point), and d 56 (Figs. 2A and 2B). We

466

found that in most neonicotinoid treatments, the cumulative proportions of male and female

467

chironomids were similar to controls. However, there were a few notable exceptions. Statistically

468

significant shifts toward female-dominated Chironomidae populations occurred at d 28 in CLO

469

(55.6% Female) (z = 4.29, p < 0.001), IMI-CLO (57.1% Female) (z = 3.25, p = 0.001), and IMI-

EP

TE D

M AN U

SC

RI PT

446

AC C

Chironomidae Sex Ratios

We also evaluated the impacts of single compound and neonicotinoid mixture exposure

Maloney, 19

ACCEPTED MANUSCRIPT

TMX treatments (62.8% Female) (z = 2.79, p = 0.005), compared to controls (43.2% Female)

471

(Fig. 2A). However, by the end of the study (d 56), with exception of the CLO treatment (51.2%

472

Female) compared to controls (43.2% Female) (z = 4.32, p < 0.001), sex ratios did not differ

473

significantly between other treatments and controls (Fig. 2B).

474

4

475 476 477 478

4.1 Effects of Neonicotinoid and Neonicotinoid Mixture Exposure on Aquatic Insect Communities

479

neonicotinoid mixture exposures could slightly shift the taxonomic composition of an aquatic

480

insect community. In all neonicotinoid-treated limnocorrals there were statistically significant

481

increases in the proportions of emerged Trichoptera, and in the IMI, IMI-CLO, CLO, and CLO-

482

TMX treated limnocorrals there were statistically significant increases in the proportions of

483

emerged Odonata. However, largely due to high within treatment variation, there were no

484

statistically significant increases in the absolute abundances of emerged Trichoptera and Odonata

485

in the treated limnocorrals. Interestingly, although Diptera (e.g. Chironomidae) are known to be

486

sensitive to neonicotinoids (Morrissey et al., 2015), the proportion of emerged dipterans were

487

similar among treatments. We suspect that this is may be, at least in part, due to the large amount

488

of variation among treatment replicates. Indeed, there was a visible trend of decreasing

489

proportions of dipterans in all neonicotinoid-treated limnocorrals relative to experimental

490

controls.

491

SC

Discussion

RI PT

470

AC C

EP

TE D

M AN U

In this study we found that both theoretically equitoxic single-compound and

Although many studies have focussed on the effects of anthropogenic disturbance on

492

wetland invertebrate communities, there is currently limited scientific consensus on what shifts

493

in community dynamics are likely to be ecologically significant (Batzer, 2013). In this study, the Maloney, 20

ACCEPTED MANUSCRIPT

proportional changes observed in aquatic insect taxa were relatively small, with emergence of

495

Odonata and Trichoptera in the neonicotinoid treated limnocorrals typically being 1 - 8 % higher

496

than in the controls (and Diptera being ~ 3 – 13 % lower than controls). Furthermore, these

497

Prairie wetland communities were highly dominated by a few dipteran taxa, and the abundances

498

of Odonata and Trichoptera were low in comparison. Similarly, a previous limnocorral study

499

found that neonicotinoid (IMI, CLO, TMX) exposure did not elicit changes in wetland insect

500

community composition (Cavallaro et al., 2018). Therefore, it is difficult to conclude if these

501

shifts in community composition are likely to be ecologically significant and we caution against

502

direct interpretation of these observed community shifts on an ecosystem level.

503 504 505

4.2

506

developed using a representative aquatic insect species (Chironomus dilutus), could adequately

507

predict the toxicity of three individual neonicotinoids and their binary mixtures to natural insect

508

populations, specifically Chironomidae, under semi-controlled field conditions. To accomplish

509

this, we exposed wetland limnocorrals containing natural Chironomidae populations to effect-

510

based neonicotinoid concentrations (derived from laboratory-based 28 d EC50 values) and

511

evaluated effects on cumulative emergence (as abundance) and biomass over time, and effects on

512

emergence, biomass and sex-ratios at two key time points, 28 d (comparative to the laboratory-

513

based studies) and 56 d (aimed at capturing a more complete Chironomidae life-cycle), relative

514

to controls. Although there were differences among individual limnocorrals (experimental

515

replicates), we found that the effects of chronic single-compound and binary mixture exposures

516

significantly deviated from what was predicted from previous laboratory studies using current

517

environmental risk assessment approaches.

M AN U

SC

RI PT

494

Effects of Neonicotinoid Exposures on Chironomidae Populations

AC C

EP

TE D

Importantly, this study focussed on determining if laboratory-derived toxicity models,

Maloney, 21

ACCEPTED MANUSCRIPT

518 519

4.2.1

Time-weighted Treatment Effects In this study, we specifically focussed on evaluating neonicotinoid and neonicotinoid

mixture toxicity at two key time points (d 28 and d 56). Therefore, this experiment was not

521

specifically designed to focus on time-weighted effects of neonicotinoid insecticides (single

522

compounds and binary mixtures) on Chironomidae populations. However, as a result of our

523

study design (i.e. insect collection every 4 d over the course of the study), we were able to

524

perform an analysis on the time-weighted effects of IMI, CLO, TMX, and their binary mixtures

525

on cumulative Chironomidae abundance (emergence) and biomass. As expected, we found that

526

Chironomidae emergence and biomass were mainly affected by time (with cumulative

527

emergence and biomass increasing over the study duration). However, treatment effects were

528

somewhat masked by within treatment variation. Indeed, when accounting for individual

529

limnocorral identity, exposure time and the time-treatment interaction, there were no statistically

530

significant differences in cumulative Chironomidae abundance or biomass among neonicotinoid

531

treatments and the experimental controls (i.e. treatment was insignificant). Importantly, we found

532

that the interaction between time and treatment was significant, with treatment effects being

533

dependent on exposure duration. This is significant, as it indicates that exposure time is critical

534

when considering the toxicity of neonicotinoid compounds and their mixtures to Chironomidae

535

and other similarly sensitive aquatic insect species. In fact, for neonicotinoids and their mixtures,

536

time-weighted analyses might be more important than concentration- or toxicological response-

537

based analyses in environmental risk assessment. Indeed, previous studies with lab-reared

538

Chironomidae have shown that both single compound and neonicotinoid mixture toxicity can

539

significantly vary depending on exposure time (Cavallaro et al., 2017; Maloney et al., 2018;

540

Maloney et al., 2017). Therefore, further studies should be carried out to better evaluate the time-

AC C

EP

TE D

M AN U

SC

RI PT

520

Maloney, 22

ACCEPTED MANUSCRIPT

weighted toxicological effects of neonicotinoids and their mixtures to determine if and how

542

exposure duration and life cycle stage should be incorporated into neonicotinoid risk assessment

543

for sensitive aquatic insect species.

544

4.2.2

545

Single Neonicotinoid Compounds

RI PT

541

The effects of individual neonicotinoids on natural invertebrate populations are typically estimated using laboratory-based toxicity tests with standard laboratory species (e.g. Daphnia

547

magna and Chironomus dilutus). However, in this study, we demonstrated that the laboratory-

548

derived toxicity thresholds for neonicotinoid insecticides (single-compounds) using C. dilutus

549

were not adequately predictive of the responses of natural Chironomidae populations. Indeed,

550

most of the single-compound exposures were much more toxic than hypothesized based on the

551

toxic unit approach. Applied neonicotinoid concentrations were chosen to elicit 50% decrease in

552

emergence of C. dilutus and other similarly sensitive species (e.g. ΣTU = 1, based on 28-d EC50

553

values). However, after 28 d of exposure, abundances of emerged Chironomidae in IMI and

554

CLO treated limnocorrals were ~ 42% and 20% lower than expected. Similarly, after 56 d of

555

exposure, abundances of emerged Chironomidae in IMI and CLO treated limnocorrals were

556

significantly impacted, with these treatments eliciting 71 – 90 % declines in emergence relative

557

to controls. Although some deviation is expected (due to sensitivity differences between lab-

558

reared and natural insect species), the magnitude of deviation in single-compound effect

559

observed in this study from laboratory-predicted toxicity was quite significant. Furthermore,

560

similar trends were observed with cumulative Chironomidae biomass, indicating that IMI and

561

CLO also significantly affected the size/weight of the emerged Chironomidae.

AC C

EP

TE D

M AN U

SC

546

Maloney, 23

ACCEPTED MANUSCRIPT

562 563

4.2.3

Binary Neonicotinoid Mixtures The effects of neonicotinoid mixtures are commonly estimated using the principle of

Concentration Addition (CA). That is, toxicity is predicted by directly summing neonicotinoid

565

concentrations (as toxic equivalents). In prior studies, mixtures of IMI, CLO, and TMX were

566

shown to elicit greater-than-additive toxicity in C. dilutus. Under acute (96 h) exposure

567

scenarios, all neonicotinoid mixtures were found to act synergistically, eliciting greater than

568

additive cumulative toxicity (e.g. IMI-CLO exposure resulted in ~56% greater mortality than

569

predicted by directly additive mixture models) (Maloney et al., 2017). Under chronic laboratory-

570

based exposure scenarios, cumulative effects were slightly diminished, with only IMI-TMX

571

mixtures eliciting greater-than-additive toxicity (up to a 10% decrease in successful emergence)

572

relative to what was predicted by CA (Maloney et al., 2018). However, in this study, the mixture

573

effects (biomass and abundance) deviated from what was predicted by these laboratory-based

574

toxicity tests. Contrary to laboratory predictions, there were no statistically significant

575

differences in toxicological response among single compound- and neonicotinoid mixture-treated

576

limnocorrals, and IMI-CLO, CLO-TMX, and IMI-TMX mixtures had to be categorized as

577

directly additive (i.e. they behave as predicted by CA). In addition, the relative responses to the

578

neonicotinoid mixture treatments varied from what was predicted from laboratory studies, with

579

the CLO-TMX mixture (and not IMI-CLO or IMI-TMX mixtures as previously predicted)

580

eliciting the most significant declines in Chironomidae emergence and biomass (e.g. ~44%

581

higher-than-predicted impact on emergence of field Chironomidae than predicted by laboratory

582

toxicity tests after 28 d of exposure).

AC C

EP

TE D

M AN U

SC

RI PT

564

Maloney, 24

ACCEPTED MANUSCRIPT

583

4.2.4

Chironomidae Sex Ratios Previous laboratory studies have also indicated that exposure to neonicotinoids and

585

neonicotinoid mixtures could potentially shift sex ratios toward male dominant populations

586

(Cavallaro et al., 2017; Maloney et al., 2018). Contrary to those laboratory-based predictions, in

587

this study we found that chronic neonicotinoid exposure did not significantly shift sex ratios in

588

natural Chironomidae populations. Exposure to some neonicotinoid treatments did elicit

589

statistically significant shifts in sex-ratios relative to the control. For example, after 28 d of

590

exposure there were shifts toward female-dominant populations in CLO, IMI-CLO, and IMI-

591

TMX treatments. However, for most treatments any statistically significant sex-ratio differences

592

that were apparent at d 28 became statistically insignificant by d 56. In fact, CLO was the only

593

neonicotinoid treatment that elicited a statistically significant sex-ratio shift over the full course

594

of the study, resulting in an increase in the overall proportion of emerged female Chironomidae

595

(e.g. an average 40.8% females in the controls vs. 51.2% females in the CLO treated

596

limnocorrals). Therefore, as with Chironomidae emergence and biomass, we found that the

597

effects of neonicotinoid exposure on Chironomidae sex-ratios significantly deviated from what

598

has been observed in laboratory tests.

599 600 601

4.3

602

from that predicted by laboratory-based studies. The reasons behind these deviations are unclear,

603

but there are several experimental factors that could have influenced the results. First,

604

neonicotinoid concentrations occasionally fell below or exceeded the target doses (Tables A2

605

and A3). One particular example of this was on d 8, where neonicotinoid concentrations were ~

606

2X higher than the target doses (e.g. d 8 measured ΣTUs ranged from 1.37 to 2.23, compared to

EP

TE D

M AN U

SC

RI PT

584

AC C

Deviations from Laboratory-Predicted Toxicity Under semi-controlled field conditions, single compound toxicity significantly deviated

Maloney, 25

ACCEPTED MANUSCRIPT

the target ΣTUs of 1.0). These concentration spikes could have resulted in increased internal

608

concentrations in the exposed organisms, which could have contributed to the greater-than-

609

predicted toxicity observed in some of the neonicotinoid treatments (Focks et al., 2018).

610

However, even these peak concentrations were much lower than concentrations likely to cause

611

lethality in the exposed invertebrates (e.g. IMI peak concentration = 0.94 µg/L, IMI 96 h LC50

612

for C. dilutus = 4.63 µg/L) and were relatively consistent across mixture and single-compound

613

treatments. Furthermore, although the limnocorrals were occasionally overdosed, they were also

614

occasionally under-dosed (Tables A2 and A3), with mean limnocorral concentrations remaining

615

within the target ΣTU over the course of the study (Table 1). Thus, internal concentrations in the

616

exposed organisms were likely also stabilized over the course of the study. Therefore, these

617

increased concentrations could have contributed to the observed deviations from predicted

618

toxicity, however there were likely other contributing factors that influenced the experimental

619

results. Another reason for this observed greater-than-predicted toxicity in the single compound

620

treatments could be that there are species-specific differences in neonicotinoid sensitivity among

621

the exposed Chironomidae. Although this has yet to be confirmed, previous studies have shown

622

that sensitivity to water quality (e.g. both natural and anthropogenic stressors) greatly varies

623

among different chironomid species (Odume and Muller, 2011; Orendt, 1999). Furthermore,

624

species-specific differences in physiology, reproduction (e.g. multivoltine vs. univoltine), life-

625

cycle length, and behaviour, and ecological preferences could result in different sensitivities to

626

neonicotinoid insecticides. Unfortunately, as we opted to quantify the biomass of the emerged

627

Chironomidae, we could not verify whether differences in species composition in the treated

628

limnocorrals significantly contributed to these differential effects. In addition, the doses applied

629

in this study were based on laboratory studies that used only one Chironomidae species (C.

AC C

EP

TE D

M AN U

SC

RI PT

607

Maloney, 26

ACCEPTED MANUSCRIPT

dilutus) in an environmentally controlled setting (Maloney et al., 2018). In this semi-controlled

631

field study, natural biotic and abiotic factors (i.e. predator stress, community dynamics, species-

632

specific differences in response) were present that could influence neonicotinoid toxicity. In fact,

633

although this experiment offers an initial bridge between laboratory- and field-based toxicity

634

studies, further studies are needed to determine how these biotic and abiotic factors could

635

influence toxicity of neonicotinoids to wetland insects like Chironomidae.

RI PT

630

Significantly, the response of natural Chironomidae populations to neonicotinoid

637

mixtures also deviated from that predicted from laboratory mixture studies. This was primarily

638

due to the high amount of variability among replicate limnocorrals, particularly in the mixture

639

treatments, which limited the amount of interpretation that could be made about mixture effects.

640

For example, for Chironomidae abundance, the coefficient of variance (CV) for the experimental

641

controls was 38.4% and for the single compound treatments CVs ranged from 13.7 to 29.3 %

642

(Table A6). However, for the mixtures variance was 2 – 4 times higher, with CVs ranging from

643

50.8 to 111.7% (Table A6). Therefore, it is possible that mixture effects do occur, but they were

644

masked by this significant variation. Although variance was not as significant for Chironomidae

645

biomass (Table A6), there was still a trend of higher variability in the mixtures (38.1 – 82.4 %)

646

when compared to the single compound treatments (14.0 – 46.3 %). Indeed, increased variability

647

can be an indicator of an ecological system experiencing stress (Forbes et al., 1995; Orlando and

648

Guillette, 2001), so it is possible that the neonicotinoid mixtures were negatively impacting these

649

Chironomidae populations. The lack of greater-than-additive mixture effects observed in this

650

study could also have been a factor of experimental design. Due to physical, practical and

651

financial constraints, the experimental design used in this study was kept relatively simple (i.e.

652

mixtures were only tested at one dose-level (ΣTU = 1.0) and one dose-ratio (1:1)). However, as

AC C

EP

TE D

M AN U

SC

636

Maloney, 27

ACCEPTED MANUSCRIPT

prior studies have demonstrated, mixtures tend to behave variably depending on composition and

654

cumulative concentration (Jonker et al., 2005; Maloney et al., 2017), and exposure time

655

(Maloney et al. 2018). Therefore, it is possible that the simplified exposure regime used here

656

failed to capture deviations from direct additivity, as seen in previous laboratory-based studies.

657

Further studies should thus focus on characterizing the toxicity of neonicotinoid mixtures under

658

field-realistic settings with more extensive exposure ranges and ratios (i.e. aim to better

659

characterize the toxicological response surface) and more replication to determine if synergistic

660

or antagonistic mixture effects do occur and if they pose a risk to natural Chironomidae

661

populations.

662

4.4

SC

M AN U

663

RI PT

653

Relevance to Derivation of Global Water Quality Regulations The neonicotinoid concentrations employed in this study were specifically chosen to

evaluate if laboratory-based predictions of mixture toxicity would adequately translate to semi-

665

controlled field conditions. Therefore, the neonicotinoid doses applied were effect-based

666

(designed to be equitoxic if the principle of Concentration Addition held (e.g. ΣTU = 1)), and

667

higher than concentrations typically observed in the field. Indeed, in aquatic environments, IMI,

668

CLO, TMX and their mixtures have been detected at cumulative toxic units (ΣTU) of 0.09

669

(agricultural surface waters) (Schaafsma et al., 2015; Smalling et al., 2015), 0.36 (rivers and

670

streams) (Hladik and Kolpin, 2015), 4.38 (wetlands) (Main et al., 2016), and 32.0 (groundwater)

671

(Giroux and Sarrasin, 2011). However, some of the concentrations employed here were either

672

equivalent to or lower than current neonicotinoid water quality guidelines. For example, in this

673

study there were significant reductions in Chironomidae emergence and biomass in limnocorrals

674

treated with CLO at concentrations (0.34 µg/L, mixtures; 0.73 µg/L, single-compound)

675

substantially lower than the current United States Environmental Protection Agency (US EPA)

AC C

EP

TE D

664

Maloney, 28

ACCEPTED MANUSCRIPT

aquatic life benchmark (1.10 µg/L) (USEPA, 2017). Similarly, in the TMX-treated limnocorrals,

677

concentrations applied ranged from 5.27 (CLO-TMX and IMI-TMX mixtures) to 9.31 (single-

678

compound) µg/L, which are much lower than the current US EPA aquatic life benchmark of 17.5

679

µg/L (USEPA, 2017). Therefore, under current US water quality criteria (the only specific

680

criteria for CLO and TMX), Chironomidae abundance and biomass could potentially be affected

681

in aquatic environments chronically contaminated with CLO, TMX, or CLO-TMX mixtures.

682

Regulations for IMI are more stringent, thus the concentrations applied here were higher than

683

most chronic water quality benchmarks (Smit, 2014; USEPA, 2017). However, IMI

684

concentrations used in the IMI-CLO and IMI-TMX mixtures (0.23 – 0.25 µg/L) were equivalent

685

to current Canadian and European water quality guidelines (0.23 and 0.20 µg/L, respectively)

686

(Canadian Council of Ministers of the Environment, 2007; European Food Safety Authority,

687

2006). Therefore, the effects on Chironomidae emergence observed in this study could

688

potentially occur in wetlands chronically contaminated with mixtures of neonicotinoid

689

insecticides where IMI is above 0.23 µg/L.

SC

M AN U

TE D

690

RI PT

676

To be adequately protective of natural Chironomidae populations and similarly sensitive taxa (e.g. Ephemeroptera), the current aquatic regulatory values should be modified to account

692

for enhanced neonicotinoid toxicity observed under field conditions. The US EPA and the

693

National Institute for Public Health and the Environment (RIVM, Netherlands) have set chronic

694

aquatic life benchmarks for IMI of 0.01 and 0.0083 µg/L, respectively (Smit, 2014; USEPA,

695

2017). According to this study (and previous laboratory toxicity tests), these benchmarks would

696

be protective for natural Chironomidae populations and similarly sensitive aquatic insects.

697

Therefore, other regulatory bodies (e.g. Environment and Climate Change Canada and the

698

European Food Safety Authority) should consider employing similar regulatory guidelines for

AC C

EP

691

Maloney, 29

ACCEPTED MANUSCRIPT

IMI concentrations in aquatic systems. In previous studies, CLO and IMI have been shown to be

700

approximately equitoxic to a sensitive aquatic insect species, C. dilutus (Cavallaro et al., 2017;

701

Maloney et al., 2018, 2017). In this study, we found that IMI and CLO were also approximately

702

equitoxic to natural Chironomidae populations. Thus, to be adequately protectively of

703

Chironomidae species, water quality guidelines for IMI and CLO should be set at similar levels.

704

Therefore, we recommend CLO chronic toxicity benchmark should be in the range of 0.0083 -

705

0.01 µg/L (Smit, 2014; USEPA, 2017). For TMX, under both laboratory and field conditions

706

TMX appears to be ~ 10 X less toxic to Chironomidae than IMI or CLO. Therefore, we

707

recommend a chronic toxicity benchmark for TMX below 0.1 µg/L in freshwater environments.

708

Although this is likely to be protective of Chironomidae, it does not account for other more

709

sensitive species (e.g. Ephemeroptera (Van den Brink et al., 2016)). Therefore, further studies

710

with other sensitive aquatic invertebrates should be considered when deriving and modifying

711

current water quality benchmarks. Furthermore, our work stresses the importance of having

712

complementary field-based studies to better characterize toxic responses of neonicotinoids and

713

their mixtures.

SC

M AN U

TE D

EP

714

RI PT

699

Acknowledgements

716

We would like to acknowledge R. Brailsford, S. Schiffer, A. Gerhart, M. Cavallaro, L. D’Silva,

717

K. Bianchini, A. Moate, and S. Hanson (Toxicology Centre, University of Saskatchewan) for

718

their assistance with field experiments. We would also like to thank J. Fehr (Environment and

719

Climate Change Canada) for chemical analysis, and I. Phillips (Water Security Agency) for

720

assistance with invertebrate identification. Pure, technical grade neonicotinoids used in the field

721

experiments were donated for research by Bayer CropScience (IMI and CLO) and Syngenta

AC C

715

Maloney, 30

ACCEPTED MANUSCRIPT

722

Crop Protection LLC (TMX). No external parties influenced the experimental design, objectives,

723

or results of this study.

724

Funding: This work was supported by Environment and Climate Change Canada, Fisheries and

726

Oceans Canada (National Contaminants Advisory Group), G00016699; the Natural Sciences and

727

Engineering Research Council of Canada, STPG-P 430138-12. Support for E. Maloney was

728

through graduate student funding from the Toxicology Centre, University of Saskatchewan

729

(Saskatoon, Canada), and the Natural Sciences and Engineering Council Canadian Graduate

730

Scholarship (NSERC CGS-D) CGS-D.

M AN U

AC C

EP

TE D

731

SC

RI PT

725

Maloney, 31

ACCEPTED MANUSCRIPT

References

733 734 735

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, 409–422. https://doi.org/10.1016/j.scitotenv.2014.09.090

736 737

Bataille, K.J., Baldassarre, G.A., 1993. Distribution and abundance of aquatic macroinvertebrates following drought in three prairie pothole wetlands. Wetlands 13, 260–269.

738 739

Batzer, D.P., 2013. The seemingly intractable ecological responses of invertebrates in North American Wetlands: A review. Wetlands 33, 1–15. https://doi.org/10.1007/s13157-012-0360-2

740 741 742

Benoit, D.A., Sibley, P.K., Juenemann, J.L., Ankley, G.T., 1997. Chironomus tentans life-cycle test: Design and evaluation for use in assessing toxicity of contaminated sediments. Environ. Toxicol. Chem. 16, 1165–1176. https://doi.org/10.1002/etc.5620160611

743 744

Canadian Council of Ministers of the Environment, 2007. Canadian Water Quality Guidelines for the Protection of Aquatic Life - Imidacloprid. Can. Environ. Qual. Guidel. 1–8.

745 746 747

Cavallaro, M.C., Liber, K., Headley, J. V., Peru, K.M., Morrissey, C.A., 2018. Community-level and phenological responses of emerging aquatic insects exposed to three neonicotinoid insecticides: an in situ wetland limnocorral approach. Environ. Toxicol. Chem. in press.

748 749 750

Cavallaro, M.C., Morrissey, C.A., Headley, J. V., Peru, K.M., Liber, K., 2017. Comparative chronic toxicity of imidacloprid, clothianidin, and thiamethoxam to Chironomus dilutus and estimation of toxic equivalency factors. Environ. Toxicol. Chem. 36, 372–382. https://doi.org/10.1002/etc.3536

751 752

European Food Safety Authority, 2006. Conclusion regarding the peer review of the pesticide risk assessment of the active substance 2007, 1–67.

753 754

Evelsizer, V., Skopec, M., 2016. Pesticides, Including Neonicotinoids, in Drained Wetlands of Iowa’s Prairie Pothole Region. Wetlands 1–12. https://doi.org/10.1007/s13157-016-0796-x

755 756 757

Focks, A., Belgers, D., Boerwinkel, M.-C., Buijse, L., Roessink, I., den Brink, P.J., 2018. Calibration and validation of toxicokinetic-toxicodynamic models for three neonicotinoids and some aquatic macroinvertebrates. Ecotoxicology. https://doi.org/10.1007/s10646-018-1940-6

758 759 760

Forbes, V.E., Moller, V., Depledge, M.H., 1995. Intrapopulation Variability in Sublethal Response to Heavy Metal Stress in Sexual and Asexual Gastropod Populations. Funct. Ecol. 9, 477–484. https://doi.org/10.2307/2390012

761 762

Giroux, I., Sarrasin, B., 2011. Pesticides et nitrates dans l’eau souterraine près de cultures de pommes de terre Échantillonnage dans quelques régions du Québec en 2008 et 2009. https://doi.org/ISBN 978-2-550-61396-1

763 764 765

Gomez-Eyles, J.L., Svendsen, C., Lister, L., Martin, H., Hodson, M.E., Spurgeon, D.J., 2009. Measuring and modelling mixture toxicity of imidacloprid and thiacloprid on Caenorhabditis elegans and Eisenia fetida. Ecotoxicol. Environ. Saf. 72, 71–79. https://doi.org/10.1016/j.ecoenv.2008.07.006

766 767

Hirvenoja, M., 1973. Revision der Gattung Cricotopus van der Wulp und ihrer Verwandten (Diptera: Chironomidae). Ann. Zool. Fennici 10, 1–363.

768 769

Hladik, M., Kolpin, D., 2015. First national-scale occurrence of neonicotinoid insecticides in streams across the U.S.A. Environ. Chem. 12–20. https://doi.org/10.1071/EN15061

770 771 772

Jonker, M.J., Svendsen, C., Bedaux, J.J.M., Bongers, M., Kammenga, J.E., 2005. Significance testing of synergistic/antagonistic, dose level-dependent, or dose ratio-dependent effects in mixture dose-response analysis. Environ. Toxicol. Chem. 24, 2701–2713. https://doi.org/10.1897/04-431r.1

773 774 775

Lamers, M., Anyusheva, M., La, N., Nguyen, V.V., Streck, T., 2011. Pesticide Pollution in Surface- and Groundwater by Paddy Rice Cultivation: A Case Study from Northern Vietnam. CLEAN - Soil, Air, Water 39, 356–361. https://doi.org/10.1002/clen.201000268

AC C

EP

TE D

M AN U

SC

RI PT

732

Maloney, 32

ACCEPTED MANUSCRIPT

Loureiro, S., Svendsen, C., Ferreira, A.L.G., Pinheiro, C., Ribeiro, F., Soares, A.M.V.M., 2010. Toxicity of three binary mixtures to daphnia magna: Comparing chemical modes of action and deviations from conceptual models. Environ. Toxicol. Chem. 29, 1716–1726. https://doi.org/10.1002/etc.198

779 780 781

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, 188–192. https://doi.org/10.1021/acs.estlett.5b00136

782 783

Main, A.R., 2016. Spatiotemporal Patterns of Distribution and Drivers of Neonicotinoid fate in Canadian Prairie Pothole Wetlands. University of Saskatchewan.

784 785 786

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. https://doi.org/10.1371/journal.pone.0092821

787 788 789

Maloney, E.M., Morrissey, C.A., Headley, J.V., Peru, K.M., Liber, K., 2018. Chronic exposure to imidacloprid, clothianidin, and thiamethoxam mixtures can exert greater than additive toxicity in Chironomus dilutus. Ecotoxicol. Environ. Saf. in press.

790 791 792

Maloney, E.M., Morrissey, C.A., Headley, J.V., Peru, K.M., Liber, K., 2018. Can chronic exposure to imidacloprid, clothianidin, and thiamethoxam mixtures exert greater than additive toxicity in Chironomus dilutus? Ecotoxicol. Environ. Saf. 156. https://doi.org/10.1016/j.ecoenv.2018.03.003

793 794 795

Maloney, E.M., Morrissey, C.A., Headley, J.V., Peru, K.M., Liber, K., 2017. Cumulative toxicity of neonicotinoid insecticide mixtures to Chironomus dilutus under acute exposure scenarios. Environ. Toxicol. Chem. 36. https://doi.org/10.1002/etc.3878

796 797 798

Maloney, E.M., Morrissey, C.A., Headley, J. V., Peru, K.M., Liber, K., 2017. Cumulative toxicity of neonicotinoid insecticide mixtures to Chironomus dilutus under acute exposure scenarios. Environ. Toxicol. Chem. https://doi.org/10.1002/etc.3878

799 800

Merritt, R.W., Cummins, K.W., Berg, M.B., 2008. An Introduction to the Aquatic Insects of North America, 4th ed. Kendall Hunt Publishing Company, Dubuque, IA.

801 802 803

Morrissey, C.A., Mineau, P., Devries, J.H., Sanchez-Bayo, F., Liess, M., Cavallaro, M.C., Liber, K., 2015. Neonicotinoid contamination of global surface waters and associated risk to aquatic invertebrates: A review. Environ. Int. 74, 291–303. https://doi.org/10.1016/j.envint.2014.10.024

804 805

Odume, O.N., Muller, W.J., 2011. Diversity and structure of Chironomidae communities in relation to water quality differences in the Swartkops River. Phys. Chem. Earth 36, 929–938. https://doi.org/10.1016/j.pce.2011.07.063

806 807

Oliver, D.R., 1976. Chironomidae (Diptera) of Char Lake, Cornwallis Island, NWT with descriptions of two new species. Can. Entomol. 108, 1053–1064.

808 809

Oliver, D.R., Rousssel, M.E., 1983. The Genera of Larval Midges of Canada. Diptera: Chironomidae. Ottawa, Ontario, Canada.

810 811

Orendt, C., 1999. Chironomids as bioindicators in acidified streams: a contribution to the acidity tolerance of chironomid species with a classification in sensitivity classes. Int. Rev. Hydrobiol. 84, 439–449.

812 813

Orlando, E.F., Guillette, L.J., 2001. A re-examination of variation associated with envrionmentally stressed organisms. Hum. Reprod. Update 7, 265–272.

814 815 816

Pavlaki, M.D., Pereira, R., Loureiro, S., Soares, A.M.V.M., 2011. Ecotoxicology and Environmental Safety Effects of binary mixtures on the life traits of Daphnia magna. Ecotoxicol. Environ. Saf. 74, 99–110. https://doi.org/10.1016/j.ecoenv.2010.07.010

817 818 819

Sánchez-Bayo, F., Hyne, R. V., 2014. Detection and analysis of neonicotinoids in river waters – Development of a passive sampler for three commonly used insecticides. Chemosphere 99, 143–151. https://doi.org/10.1016/J.CHEMOSPHERE.2013.10.051

AC C

EP

TE D

M AN U

SC

RI PT

776 777 778

Maloney, 33

ACCEPTED MANUSCRIPT

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, 1–21. https://doi.org/10.1371/journal.pone.0118139

823 824 825

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, 80–90. https://doi.org/10.1016/j.scitotenv.2014.08.114

826 827

Smit, C.E., 2014. Water quality standards for imidacloprid. Proposal for an update according to the Water Water quality standards for imidacloprid 94 pp.

828 829 830

Smith, R., Middlebrook, R., Turner, R., Huggins, R., Vardy, S., Warne, M., 2012. Large-scale pesticide monitoring across Great Barrier Reef catchments – Paddock to Reef Integrated Monitoring, Modelling and Reporting Program. Mar. Pollut. Bull. 65, 117–127. https://doi.org/10.1016/J.MARPOLBUL.2011.08.010

831 832 833

USEPA, 2017. Aquatic Life Benchmarks for Pesticide Registration [WWW Document]. URL www.epa.gov/pesticide-science-and-assessing-pesticide-risks/aquatic-life-benchmarks-pesticide-registration (accessed 7.25.17).

834 835 836

Van den Brink, P.J., Van Smeden, J.M., Bekele, R.S., Dierick, W., De Gelder, D.M., Noteboom, M., Roessink, I., 2016. Acute and chronic toxicity of neonicotinoids to nymphs of a mayfly species and some notes on seasonal differences. Environ. Toxicol. Chem. 35, 128–133. https://doi.org/10.1002/etc.3152

837 838 839

Xie, W., Han, C., Qian, Y., Ding, H., Chen, X., Xi, J., 2011. Determination of neonicotinoid pesticides residues in agricultural samples by solid-phase extraction combined with liquid chromatography–tandem mass spectrometry. J. Chromatogr. A 1218, 4426–4433. https://doi.org/10.1016/J.CHROMA.2011.05.026

M AN U

SC

RI PT

820 821 822

AC C

EP

TE D

840

Maloney, 34

ACCEPTED MANUSCRIPT

Appendices

Treatment

Dissolved Oxygen (mg/L)

Temperature (°C)

Ammonia (mg/L)

pH

Conductivity (µS/cm3)

Control

4.5 ± 0.7

19.7 ± 0.8

0.4 ± 0.1

8.1 ± 0.0

2937 ± 120

IMI

4.7 ± 0.8

20.3 ± 0.2

0.6 ± 0.0

8.1 ± 0.0

IMI-CLO

6.2 ± 3.0

20.5 ± 0.1

1.4 ± 0.7

8.2 ± 0.1

CLO

5.2 ± 1.2

20.6 ± 0.1

0.7 ± 0.0a

8.1 ± 0.1

CLO-TMX

5.1 ± 1.4

20.7 ± 0.2

0.6 ± 0.2

8.2 ± 0.2

TMX

4.8 ± 0.7

20.9 ± 0.2

0.6 ± 0.1

IMI-TMX

6.1 ± 1.6

20.8 ± 0.9

0.6 ± 0.1

1

Alkalinity (mg/L as CaCO3)

Nitrate (mg/L)

Phosphate (mg/L)

Dissolved Organic Carbon (mg/L)

1773 ± 87

419 ± 9

3.5 ± 0.4

1.1 ± 0.4

31.5 ± 1.6

2935 ± 86

1771 ± 82

405 ± 20

3.7 ± 0.7

1.7 ± 0.6

31.6 ± 2.3

2959 ± 125

1741 ± 73

411 ± 39

3.8 ± 0.6

1.3 ± 0.7

34.5 ± 5.3

2899 ± 90

1709 ± 21

413 ± 15

3.0 ± 0.3

0.7 ± 0.5

31.9 ± 1.6

3021 ± 85

1677 ± 50

393 ± 36

4.7 ± 0.1

0.8 ± 0.4

30.9 ± 1.1

TE D

M AN U

SC

Total Hardness (mg/L as CaCO3)

2850 ± 23

1609 ± 57

392 ± 10

4.0 ± 0.2

1.0 ± 0.2

31.2 ± 1.2

8.1 ± 0.0

2936 ± 58

1678 ± 27

469 ± 72

4.9 ± 0.4

2.2 ± 1.6

32.5 ± 1.6

EP

8.0 ± 0.1

Significantly different from control (One-way ANOVA, p < 0.05).

AC C

a

RI PT

Table A1. Mean (± standard deviation) water quality parameters in control and neonicotinoid treated limnocorrals over the 56-d exposure period.

Maloney, 1

ACCEPTED MANUSCRIPT

Treatment (µg/L) IMI

IMI-CLO

CLO

(0.50 µg/L)

(0.25 µg/L, 0.35 µg/L)

(0.70 µg/L)

CLO-TMX

TMX

IMI-TMX

(0.35 µg/L, 4.45 µg/L)

(8.91 µg/L)

(0.25 µg/L, 4.45 µg/L)

CLO

TMX

IMI

CLO

TMX

IMI

CLO

TMX

IMI

CLO

TMX

IMI

CLO

TMX

IMI

CLO

TMX

IMI

CLO

TMX

0

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

0.04

-

-

0.02

4

-

-

<0.01

0.46

-

-

0.25

0.30

-

-

0.53

<0.01

-

0.30

5.64

-

0.02

8.41

0.22

<0.01

4.72

8

-

-

-

0.94

-

-

0.43

0.65

0.14

-

1.28

0.21

-

0.66

11.33

-

-

17.00

0.41

0.02

9.74

12

-

-

-

0.73

-

-

0.32

0.48

0.01

-

0.97

0.04

-

0.49

9.41

-

0.04

16.31

0.33

0.03

8.92

16

-

-

0.002

0.61

-

0.03

0.27

0.34

0.02

-

0.83

0.12

-

0.43

8.38

-

0.07

12.78

0.24

0.03

7.18

20

-

-

-

0.54

-

-

0.26

0.33

-

-

0.76

-

-

0.43

6.07

-

0.07

10.60

0.24

0.03

5.78

24

<0.01

-

-

0.49

-

0.07

0.20

0.25

-

-

0.63

-

-

0.33

4.78

-

-

8.17

0.18

0.03

4.89

28

<0.01

-

-

0.51

-

<0.01

0.15

0.19

-

-

0.71

-

-

0.38

3.68

<0.01

0.07

7.48

0.16

0.04

3.29

32

-

-

-

0.42

-

-

0.17

0.19

0.02

-

0.51

-

-

0.33

3.89

-

-

6.50

0.15

0.03

2.86

36

-

-

-

0.52

-

-

0.26

0.41

-

-

0.84

-

-

0.45

3.50

-

-

5.72

0.25

0.02

3.17

40

-

-

<0.01

0.48

-

44

<0.01

-

-

0.45

-

48

-

-

-

0.37

-

52

-

-

<0.01

0.39

56

-

-

-

0.32

AC C

TE D

M AN U

IMI

EP

SC

Control

Day

RI PT

Table A2. Mean measured neonicotinoid concentrations (µg/L) in limnocorrals on each dosing day (post-dose) over the 56-d exposure period. Nominal doses of neonicotinoids were as follows: IMI (single compound = 0.50 µg/L; in binary mixtures = 0.25 µg/L), CLO (single compound = 0.71 µg/L; in binary mixtures = 0.36), and TMX (single compound = 8.91 µg/L; in binary mixtures = 4.46 µg/L).

-

0.27

0.39

-

-

0.74

0.03

-

0.37

3.71

-

-

6.15

0.29

0.02

3.91

-

0.25

0.37

-

-

0.65

-

-

0.36

4.47

-

-

6.50

0.25

0.02

3.81

-

0.23

0.28

-

-

0.56

-

-

0.28

4.55

-

-

7.10

0.24

0.03

4.27

-

-

0.21

0.30

-

-

0.60

-

-

0.34

6.81

-

-

10.68

0.20

0.04

5.99

-

-

0.17

0.29

-

-

0.59

-

-

0.33

4.60

-

0.07

6.87

0.20

0.03

3.83

* All unreported concentrations were lower than the limits of quantification (LOQ): IMI: 0.0028 ± 0.0006 µg/L, CLO: 0.0034 ± 0.0008 µg/L, TMX: 0.0058 ± 0.0009 µg/L.

Maloney, 2

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

2

Maloney, 3

ACCEPTED MANUSCRIPT

Table A3. Mean measured neonicotinoid concentrations expressed as toxic units (TU) in limnocorrals at each dosing day over the 56d exposure period. Nominal TU of neonicotinoids were as follows: IMI (single compound = 1; in binary mixtures = 0.5), CLO (single compound = 1; in binary mixtures = 0.5), and TMX (single compound = 1; in binary mixtures = 0.5).

Day

Control

IMI

IMI-CLO

CLO

(0 TU)

(1.0 TU)

(1.0 TU)

(1.0 TU)

RI PT

Treatment (TU)

CLO-TMX

TMX

IMI-TMX

(1.0 TU)

(1.0 TU)

(1.0 TU)

CLO

TMX

IMI

CLO

TMX

IMI

CLO

TMX

IMI

CLO

TMX

IMI

CLO

TMX

IMI

CLO

TMX

IMI

CLO

TMX

0

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

<0.01

-

-

<0.01

4

-

-

<0.01

0.93

-

-

0.49

0.42

-

-

0.75

<0.01

-

0.43

0.63

-

0.03

0.94

0.43

0.01

0.53

8

-

-

-

1.89

-

-

0.86

0.92

0.02

-

1.80

0.02

-

0.93

1.3

-

-

1.91

0.82

0.03

1.09

12

-

-

-

1.46

-

-

0.65

0.67

<0.01

-

1.37

<0.01

-

0.68

1.06

-

0.06

1.83

0.66

0.04

1.00

16

-

-

<0.01

1.21

-

<0.01

0.54

0.48

<0.01

-

1.17

0.01

-

0.61

0.94

-

0.10

1.43

0.47

0.04

0.81

20

-

-

-

1.08

-

-

0.52

0.46

-

-

1.07

-

-

0.61

0.68

-

0.01

1.19

0.47

0.05

0.65

24

<0.01

-

-

0.98

-

<0.01

0.40

0.35

-

-

0.88

-

-

0.47

0.54

-

-

0.92

0.36

0.04

0.55

28

<0.01

-

-

1.03

-

<0.01

0.30

0.27

-

-

1.0

-

-

0.53

0.41

<0.01

0.10

0.84

0.31

0.05

0.37

32

-

-

-

0.84

-

-

0.35

0.27

<0.01

-

0.72

-

-

0.47

0.44

-

-

0.73

0.29

0.04

0.32

36

-

-

-

1.03

-

-

0.52

0.58

-

-

1.19

-

-

0.63

0.39

-

-

0.64

0.50

0.03

0.36

40

-

-

<0.01

0.97

-

-

0.55

0.54

-

-

1.04

<0.01

-

0.51

0.42

-

-

0.69

0.46

0.02

0.44

44

<0.01

-

-

0.90

-

-

48

-

-

-

0.73

-

-

52

-

-

<0.01

0.78

56

-

-

-

0.63

M AN U

TE D

EP

SC

IMI

0.52

-

-

0.92

-

-

0.51

0.50

-

-

0.73

0.50

0.03

0.43

0.46

0.39

-

-

0.80

-

-

0.39

0.51

-

-

0.80

0.49

0.04

0.48

AC C

0.50

-

-

0.41

0.42

-

-

0.85

-

-

0.48

0.77

-

-

1.20

0.40

0.06

0.67

-

-

0.34

0.32

-

-

0.84

-

-

0.46

0.52

-

0.09

0.77

0.40

0.04

0.43

* All unreported toxic units were measured at concentrations lower than the limits of quantification (LOQ): IMI: 0.0028 ± 0.0006 µg/L, CLO: 0.0034 ± 0.0008 µg/L, TMX: 0.0058 ± 0.0009 µg/L.

3

Maloney, 4

ACCEPTED MANUSCRIPT

Table A4. Mean (± standard deviation) cumulative total insect and Chironomidae abundance from experimental limnocorrals prior to neonicotinoid treatment (n = 3 days of sampling over 8 days prior to neonicotinoid treatment).

Significantly different from control (one-way ANOVA, p < 0.05)

M AN U

a

Chironomidae Abundance 22 ± 25 6±5 31 ± 26 43 ± 42 22 ± 5 8 ± 12 18 ± 16

RI PT

Total Insect Abundance 29 ± 24 10 ± 7 36 ± 25 47 ± 42 31 ± 3 12 ± 14 23 ± 17

SC

Treatment Control IMI IMI-CLO CLO CLO-TMX TMX IMI-TMX

AC C

EP

TE D

4

Maloney, 5

ACCEPTED MANUSCRIPT

CLO CLO-TMX TMX IMI-TMX

284 ± 109 284 ± 258 579 ± 156 765 ± 853

282 ± 103 284 ± 249 572 ± 156 735 ± 823

Hymenoptera

Odonata

Coleoptera

Ephemeroptera

6±3 4±1 8±6

2±0 1±1 1±2

3±2 0±1 2±3

0±1 0±0 0±0

8±5 7±6 9±4 12 ± 4

2±1 2±3 3±1 2±2

2±2 3±1 1±1 2±1

0±1 0±0 0±1 0±0

20 ± 12 24 ± 38 25 ± 17

16 ± 10 14 ± 19 12 ± 11 13 ± 11

Significantly different from the control (One-way ANOVA, p < 0.05).

AC C

EP

TE D

*

5

Trichoptera

SC

Control IMI IMI-CLO

Diptera Total Chironomidae 1972 ± 773 1962 ± 782 201 ± 31 197 ± 28 653 ± 364 638 ± 364

M AN U

Treatment

RI PT

Table A5. Mean (± standard deviation) abundances of collected insect taxa in untreated limnocorrals (controls) compared to limnocorrals treated with single compounds or binary neonicotinoid mixtures for 56 days (cumulative), (n = 3 limnocorrals/treatment).

Maloney, 6

ACCEPTED MANUSCRIPT

Table A6. Coefficients of variation (%) for cumulative abundance and biomass of emerged Chironomidae in neonicotinoid treated limnocorrals after 56 days of exposure (n = 3 replicates/treatment). Coefficient of Variation for Abundance (%) 39.9 14.2 57.1 36.6 87.8 27.4 111.9

AC C

EP

TE D

M AN U

SC

Control IMI IMI-CLO CLO CLO-TMX TMX IMI-TMX

Coefficient of Variation for Biomass (% CV) 52.4 29.7 38.1 46.3 67.4 14.0 82.4

RI PT

Treatment

Maloney, 7

ACCEPTED MANUSCRIPT

Tables

CLO

TMX

<0.01 0.52 ± 0.16 0.25 ± 0.08 <0.01 0.23 ± 0.07

0.34 ± 0.12 0.73 ± 0.21 0.39 ± 0.10 0.02 ± 0.03 0.03 ± 0.00

<0.01 0.01 ± 0.03 0.01 ± 0.04 0.03 ± 0.06 5.80 ± 2.40 9.31 ± 3.70 5.27 ± 2.19

IMI

SC

IMI

Toxic Units (TU)(1)

<0.01 1.03 ± 0.32 0.49 ± 0.14 <0.01 0.47 ± 0.14

TE D

CON IMI IMI-CLO CLO CLO-TMX TMX IMI-TMX

Measured Neonicotinoid Concentrations (µg/L)

CLO

TMX

0.47 ± 0.17 1.03 ± 0.30 0.55 ± 0.14 0.03 ± 0.04 0.04 ± 0.01

<0.01 <0.01 <0.01 <0.01 0.65 ± 0.28 1.04 ± 0.42 0.58 ± 0.24

M AN U

Treatment

RI PT

Table 1. Mean (± SD) measured neonicotinoid concentrations (µg/L) and toxic units (TU) in single compound and binary neonicotinoid mixture treated limnocorrals over the course of the study period (56 days, 14 measurements) (n = 3 limnocorrals/treatment). Cumulative Toxic Units (ΣTU)

<0.01 1.03 ± 0.32 0.97 ± 0.32 1.03 ± 0.30 1.20 ± 0.38 1.09 ± 0.44 1.10 ± 0.37

AC C

EP

* All unreported concentrations were lower than the limits of quantification (LOQ): IMI: 0.0028 ± 0.0006 µg/L, CLO: 0.0034 ± 0.0008 µg/L, TMX: 0.0058 ± 0.0009 µg/L. (1) Toxic units were calculated from laboratory-based 28-day EC50 values (emergence) previously characterized for Chironomus dilutus (Maloney et al., 2018).

ACCEPTED MANUSCRIPT

Table 2. Mean (± SD) proportions (%) of collected invertebrate taxa in untreated limnocorrals (control) compared to limnocorrals treated with single compounds or binary neonicotinoid mixtures for 56 days (cumulative), (n = 3 limnocorrals/treatment).

1.2 ± 1.1 9.3 ± 14.7 * 4.9 ± 3.9 * 4.8 ± 1.6 * 6.4 ± 9.6 * 2.2 ± 1.7 1.9 ± 0.6 4.1 ± 6.1

M AN U

SC

0.1 ± 0.1 0.6 ± 0.3 0.9 ± 0.1 1.9 ± 0.8 5.3 ± 0.9 1.1 ± 0.5 3.6 ± 0.6 0.5 ± 0.3

Odonata

RI PT

Hymenoptera

Significantly different from the control (Chi-square test, p < 0.05).

TE D

*

EP

Control IMI IMI-CLO CLO CLO-TMX TMX IMI-TMX Average

Diptera Trichoptera Total Chironomidae 96.7 ± 3.6 96.0 ± 4.4 0.3 ± 0.0 87.7 ± 14.7 86.1 ± 14.1 2.0 ± 0.6 * 93.7 ± 3.5 91.3 ± 5.2 1.1 ± 0.9 * 90.7 ± 2.7 90.4 ± 3.6 2.9 ± 1.9 * 87.5 ± 9.8 92.3 ± 22.5 3.7 ± 5.3 * 93.4 ± 4.4 94.3 ± 2.45 1.5 ± 1.0 * 95.4 ± 2.2 89.3 ± 6.2 3.6 ± 3.7 * 92.2 ± 3.6 91.4 ± 3.3 2.2 ± 1.3

AC C

Treatment

Coleoptera 0.1 ± 0.0 0.2 ± 0.3 0.2 ± 0.3 0.6 ± 0.7 1.3 ± 0.9 0.2 ± 0.2 0.6 ± 0.4 0.4 ± 0.4

Ephemeroptera 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 0.1 ± 0.2 0.0 ± 0.0 0.1 ± 0.1 0.0 ± 0.0 0.0 ± 0.0

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Figures

Figure 1. Chronic (56-d) exposures of Chironomidae populations to single compounds (imidacloprid (IMI), clothianidin (CLO), and thiamethoxam (TMX)) and binary neonicotinoid mixtures in experimental limnocorrals, compared to untreated controls: mean (± SD) cumulative emergence (A) and biomass (B) over time; mean (± SD) cumulative

emergence (C) and biomass (D) at day 28 (study mid-point); and mean (± SD) cumulative emergence (E) and biomass (F) at day 56 (study cessation). (n = 3 limnocorrals/treatment, data.). * Significantly different from untreated control (one-way ANOVA, p < 0.05). *** Dashed line (B) indicates predicted effect (Σ Toxic Unit = 1), based on 28-d EC50 values (emergence) derived from laboratory studies using the aquatic insect Chironomus dilutus as an experimental organism.

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 2. Sex of emerged Chironomidae (mean, %) after exposure to single neonicotinoids and their binary mixtures in experimental limnocorrals for (A) 28 and (B) 56 days. (n = 3 limnocorrals/treatment).

AC C

EP

TE D

* Significantly different from control (z-test, p < 0.05). ** Sex ratios for specific treatments (i.e. d 28 IMI and CLO, and d 56 CLO) are not exactly at 100% of total, due to variation of emerged proportions of males/females within treatment.