Environmental Toxicology and Chemistry, Vol. 33, No. 12, pp. 2775–2785, 2014 # 2014 SETAC Printed in the USA

CHRONIC AQUATIC EFFECT ASSESSMENT FOR THE FUNGICIDE AZOXYSTROBIN RENE P.A.

VAN

WIJNGAARDEN,y DICK J.M. BELGERS,y MAZHAR I. ZAFAR,zx ARRIENNE M. MATSER,y MARIE-CLAIRE BOERWINKEL,y and GERTIE H.P. ARTS*y

yAlterra, Wageningen University and Research Center, Wageningen, The Netherlands zDepartment of Aquatic Ecology and Water Quality Management, Wageningen University, Wageningen, The Netherlands xDepartment of Environmental Sciences, Faculty of Biological Sciences, Quaid-i-Azam University, Islamabad, Pakistan (Submitted 25 April 2014; Returned for Revision 27 May 2014; Accepted 2 September 2014)

Abstract: The present study examined the ecological effects of a range of chronic exposure concentrations of the fungicide azoxystrobin in freshwater experimental systems (1270-L outdoor microcosms). Intended and environmentally relevant test concentrations of azoxystrobin were 0 mg active ingredient (a.i.)/L, 0.33 mg a.i./L, 1 mg a.i./L, 3.3 mg a.i./L, 10 mg a.i./L, and 33 mg a.i./L, kept at constant values. Responses of freshwater populations and community parameters were studied. During the 42-d experimental period, the timeweighted average concentrations of azoxystrobin ranged from 93.5% to 99.3% of intended values. Zooplankton, especially copepods and the Daphnia longispina group, were the most sensitive groups. At the population level, a consistent no-observed-effect concentration (NOEC) of 1 mg a.i./L was calculated for Copepoda. The NOEC at the zooplankton community level was 10 mg azoxystrobin/L. The principle of the European Union pesticide directive is that lower-tier regulatory acceptable concentrations (RACs) are protective of highertier RACs. This was tested for chronic risks from azoxystrobin. With the exception of the microcosm community chronic RAC (highest tier), all other chronic RAC values were similar to each other (0.5–1 mg a.i./L). The new and stricter first-tier species requirements of the European Union pesticide regulation (1107/2009/EC) are not protective for the most sensitive populations in the microcosm study, when based on the higher tier population RAC. In comparison, the Water Framework Directive generates environmental quality standards that are 5 to 10 times lower than the derived chronic RACs. Environ Toxicol Chem 2014;33:2775–2785. # 2014 SETAC Keywords: Population

Community

Species sensitivity

Microcosm

Risk

population and community is limited [5]. Chronic microcosm and mesocosm studies are available for only a few pesticides: linuron, atrazin, carbendazim, chlorpyrifos, and 3,4-dichloroaniline [6–15]. In current regulatory risk assessment, risks from chronic exposure are mainly covered by applying an assessment factor (see discussion in Van den Brink et al. [8]). Effective scientific methods and experimental data to evaluate chronic risks of pesticides at the community level are not available, and the assessment factors applied have not been validated [4,15]. Hence, focused model ecosystem experiments are required to quantify the risks from chronic exposure to underpin risk assessment schemes and to evaluate the significance of effect assessments on the basis of laboratory toxicity tests and the use of a fixed assessment factor. In agricultural areas, multiple applications in space and time are common practice. Therefore, exposure to pesticides might be prolonged over a longer timeframe, leading to chronic exposures. Because information on the effects of long-term exposure on the structure and functioning of aquatic ecosystems is available for only a limited number of pesticides, and specifically fungicides [9,12,16], azoxystrobin was selected as a model compound. Azoxystrobin has a log octanol–water partition coefficient (KOW) of 2.5 [17] and a 50% dissipation time in water (DT50water of at least approximately 14 d [18–20]), indicating slow dissipation from the water phase and a slow decline over time, which makes it a suitable compound for a chronic study. Azoxystrobin is a systemic fungicide of the b-methoxyacrylate strobilurins group, which impedes the respiration process by inhibiting the passage of electrons from cytochrome-B to cytochrome-C in eukaryotic species [21]. A limited number of effect concentrations of azoxystrobin are known from the literature. The effects of a single application of a commercial formulation (YF9246) of azoxystrobin on freshwater

INTRODUCTION

The Water Framework Directive (Regulation2000/60/EC) was launched by the European Commission in the year 2000 [1]. Its aim is to achieve a sound chemical and biological status for surface water in member states by 2015, and not later than 2027, after two 6-yr periods of derogation. Risk assessment in the context of the Water Framework Directive differs from that in the context of the market admission of plant protection products (Regulation 1107/2009/EC [2]). Differences include procedures for deriving regulatory acceptable concentrations (RACs; in the pesticide directive) and environmental quality standards (EQS; in the Water Framework Directive) and the way exposure estimates are linked to effect estimates [3,4]. Both the Water Framework Directive and Regulation 1107/2009/EC distinguish between risks attributable to short- and long-term exposure: microcosm and mesocosm experiments can be used to derive EQS and RAC concentrations, respectively. However, recovery of potentially sensitive organisms is not considered when deriving Water Framework Directive environmental standards, whereas this may be an option in the admission procedure according to Regulation 1107/2009/EC. Most current assessments of long-term pesticide effects rely heavily on individual-based effect metrics derived from relatively short duration studies in the lower tiers of Regulation 1107/2009/EC. As a consequence, current experiment-based knowledge of the effects of long-term exposure, especially chronic exposure, on the higher ecological levels of the All Supplemental Data may be found in the online version of this article. * Address correspondence to [email protected] Published online 8 September 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/etc.2739 2775

2776

Environ Toxicol Chem 33, 2014

microcosms resulted in a no-observed-ecologically-adverseeffect concentration (NOEAEC) of azoxystrobin in an outdoor pond microcosm of 10 mg active ingredient (a.i.)/L [19]. The available acute toxicity data have been analyzed for aquatic organisms and azoxystrobin, using the species sensitivity distribution approach [16]. On the basis of the species sensitivity distribution constructed with available acute toxicity data for fish, invertebrates, and primary producers, a median 5% hazard concentration (HC5) of 42 mg a.i./L was calculated in tests with the technical substance of azoxystrobin [16]. Because this HC5 is based on acute toxicity data, and because an NOEAEC of 10 mg a.i./L was derived from a single-application microcosm study, effects at lower concentrations can be expected under a long-term chronic exposure regime. Consequently, our experiment used chronic exposure concentrations below the acute median HC5, namely, 0 mg a.i./L, 0.33 mg a.i./L, 1 mg a.i./L, 3.3 mg a.i./L, 10 mg a.i./L, and 33 mg a.i./L. The highest concentration of 33 mg a.i./L was included because we expected to find clear effects at this highest chronic treatment concentration. This would facilitate a proper interpretation of the potential dose–effect relationships in the regression-designed experimental set-up of the present study. The experimental concentrations also include the range of azoxystrobin concentrations measured in Dutch water bodies in agricultural areas, where they can exceed 0.3 mg/L [22]. The aim of the present study was to evaluate whether first-tier chronic RAC values and higher-tier chronic species sensitivity distribution hazard concentrations for azoxystrobin are protective against chronic effects at the population and community level, as observed in microcosms.

MATERIALS AND METHODS

Experimental design

The experimental site involved 19 outdoor microcosms (diameter 1.8 m; depth 0.8 m; water volume 1270 L, 50-cm water column). The microcosms were located at the Sinderhoeve Experimental Station in Renkum, The Netherlands, and were lined with a water-tight nontoxic layer of black polyethylene. The microcosm facility was covered by a net to keep birds and leaf litter out. Each microcosm contained sediment that consisted of fine clay (8-cm layer) from a mesotrophic lake (dominated by the aquatic plants Elodea nuttallii and Chara sp.). The microcosms were filled with water from the water supply basin of the Sinderhoeve Experimental Station, which contains a mixture of groundwater and rainwater and houses a freshwater community. In the preparatory phase, 100 shoots of E. nuttallii were planted on 75% of the sediment surface of each microcosm. In addition to these introduced species, other macrophytes developed from diaspores in the sediment during the study (Eleocharis acicularis, Spirodela polyrhiza, Potamogeton berchtoldii, Potamogeton pectinatus, Elodea canadensis, Potamogeton crispus, and Ranunculus circinatus). Twenty-five percent of the sediment surface was reserved for the macrophyte bioassays. During the pretreatment period (3 mo), macroinvertebrates, phytoplankton, and zooplankton were collected from uncontaminated mesotrophic ditches at the Sinderhoeve Experimental Station and/or Veenkampen, an experimental field site of Wageningen University, Wageningen, The Netherlands, and introduced into the systems to develop a freshwater community characteristic of lentic, edge-offield surface water. The macroinvertebrates introduced comprised different taxonomic groups and represented various trophic levels. Dominant species were crustaceans (Asellus aquaticus, Gammarus pulex, Cladocera and Copepoda), insects (Cloeon dipterum,

R.P.A. van Wijngaarden et al.

Chaoborus sp., Plea minutissima, Chironomidae, Odonata, and Trichoptera), and the non-arthropods Hirudinea (Erpobdella sp.) and Gastropoda (Valvata sp.). To increase similarity among the microcosms, water was circulated among the 19 systems for a period of 2 wk in the pretreatment period. The microcosm study focused on long-term effects of chronic exposure and lasted for 42 d. This period covered at least 1 generation of the above taxa. Treatments were studied in triplicate, except for the control systems, which involved 4 microcosms. Treatments were randomly assigned to the microcosms. Test concentrations of azoxystrobin were kept at a constant value of 80% to 120% of intended nominal concentrations during the experimental period. One week before the first applications, all biological end points were sampled once to establish pretreatment conditions, followed by a treatment period of approximately 6 wk. Pesticide application, pesticide sampling, and analysis

Azoxystrobin was applied as the formulated product Amistar (250 g a.i./L soluble concentrate formulation, purity 99.5%, provided by Syngenta Crop Protection). The compound was introduced into microcosms in 6 different chronic treatment regimes: control, 0.33 mg a.i./L, 1 mg a.i./L, 3.3 mg a.i./L, 10 mg a.i./L, and 33 mg a.i./L. Extra dosing requirements were calculated from regular measurements of the actual concentrations in the microcosms. Actual concentrations were measured every 1 d to 3 d. Extra dosing of the microcosms was performed 7 times in the experimental period (on days 2, 9, 16, 20, 27, 32, and 37). The test substance was gently applied onto the water surface. The control microcosms were treated with water only. The water layer was mixed by means of a stainless steel rod to promote an even distribution of the test substance in the water compartment. Mixing was performed gently so as not to damage submerged macrophytes or disturb the sediment. Before application, a stock solution was prepared in tap water by diluting 4.36 g of AMISTAR (a.i. 250 g/mL; density 1.09 g/mL) in 997.4 mL of tap water in a 2.5-L bottle (borosilicate), resulting in a concentration of approximately 1 mg/mL (¼ stock solution 1); 399.5 g of stock solution 1 diluted with 1713.8 g of tap water resulted in stock solution 2, with a concentration of 0.1886 mg/mL. From this stock solution, 2.47 g, 7.45 g, 24 g, 44 g, 75.88 g, and 244 g were diluted with tap water to a final volume of 2 L. These dose solutions were applied to the microcosms, resulting in treatment concentrations of 0.1 mg a.i./L, 0.33 mg a.i./L, 1 mg a.i./L, 3.3 mg a.i./L, 10 mg a.i./L, and 33 mg a.i./L, respectively. The control systems received 2 L of tap water only. The azoxystrobin sampling method and the chemical analysis method used are described in Zafar et al. [19]. End points

The sampling and measurement techniques of the end points are briefly summarized below. For a more detailed description of the methods and sampling frequencies, see Table 1. Water quality parameters

Dissolved oxygen, pH, electric conductivity, and temperature were measured in each microcosm to detect possible changes in community metabolism. Alkalinity and concentrations of ammonia, nitrate, nitrite, total nitrogen, orthophosphate, and total phosphate were determined in all microcosms before the treatments started (day 5) and at the end (day 42) of the experiment. See Zafar et al. [19] for a detailed description of the methods used.

Chronic effect assessment of azoxystrobin

Environ Toxicol Chem 33, 2014

Table 1. Summary of sampling days and methods used for the sampling of the investigated endpoints in the microcosms Compartment/Community

Sampling days

Referencea

Water quality parametersb Dosing of water

5, 2, 9, 16, 23, 32, 42 1, 2, 9, 16, 20, 27, 23, 37

[19] Present study [19] [19] [19] [19] [19] [19]

c

Nutrients Zooplankton sampling Phytoplankton sampling Phytoplankton chlorophyll-a Periphytic chlorophyll-a Maycrophyte species composition Myriophyllum spicatum Macroinvertebrates Pebble baskets Litter bags Decomposition

5, 5, 5, 5,

2, 2, 2, 2,

0, 42 9, 16, 23, 32, 9, 16, 23, 32, 9, 16, 23, 32, 9, 16, 23, 32, 1, 14, 42

42 42 42 42

3, 14, 42

[19]

7, 3, 10, 17, 42 7, 3, 10, 17, 42 7, 3, 10, 17, 42

[23] [24] [24]

a

For a detailed description of methods, see references. Dissolved oxygen, pH, alkalinity, electric conductivity, and temperature. Ammonia, nitrate, nitrite, total nitrogen, orthophosphate, and total phosphate.

b c

2777

surface where E. nuttallii had been planted (75% of the surface area of each microcosm). In addition, a Myriophyllum spicatum bioassay was performed [19]. Macroinvertebrates

Artificial substrates, consisting of litter bags (see Decomposition section) and pebble baskets, were used to monitor effects of azoxystrobin on the benthic macroinvertebrate assemblage. Two pebble baskets and 2 litter bags were placed on tiles on the sediment in each microcosm 2 wk before treatment initiation to allow colonization by macroinvertebrates (for a detailed description of methods, see Brock et al. [23]). The pebble baskets were gently retrieved using a net. The litter bags were collected by hand. The substrates were then rinsed in a container to collect the invertebrates. Live macroinvertebrates were identified and counted, and then released back into their original microcosms. The organisms were identified to the lowest practical taxonomic level. Abundance data of macroinvertebrates from the pebble baskets and litter bags from each microcosm were pooled before data analysis. Decomposition

Zooplankton and phytoplankton sampling and identification

Zooplankton and phytoplankton were sampled simultaneously using a Perspex (polymethyl methacrylate) tube. See Zafar et al. [19] for a detailed description of the method. Cladocerans, copepods, and ostracods (macrozooplankton) were counted using a stereo microscope (magnification 25). Rotifers and copepod nauplii (microzooplankton) were quantified and identified with an inverted microscope (magnification 100), using a subsample of known volume. Rotifers and cladocerans were identified to the lowest practical taxonomic level (i.e., genus or species level), whereas copepods were identified to the suborder by classifying them as calanoids or cyclopoids. A distinction was also made between nauplii and the more mature stages of the copepods. Phytoplankton species composition was studied by counting the number of organisms in a known volume, the species being identified to the lowest practical taxonomic level. Taxa and numbers of cells were based on a maximum of 200 observations, consisting of a series of 20 to 40 counting fields of a single cuvette under an inverted microscope (magnification 400). Zooplankton and phytoplankton data were expressed as numbers of individuals per liter. Chlorophyll-a

Phytoplankton chlorophyll-a was sampled in parallel with the phytoplankton and zooplankton sampling. See Zafar et al. [19] for a detailed description of the method. Glass slides (7.6 cm  2.6 cm) were used as artificial substrates for sampling periphyton. They were allowed to be colonized for 14 d prior to sampling. The slides were positioned vertically in a frame at a depth of approximately 10 cm below the water surface in the center of each microcosm. On each sampling day, 8 slides/microcosm were collected and scraped visually clean with razor blades, and the removed periphyton was collected in tap water. The chlorophyll-a content of the periphyton solution was analyzed as described in Zafar et al. [19]. Macrophyte cover, biomass, and bioassay

Development of macrophyte species composition and macrophyte species cover/abundance was monitored 3 times by assessing macrophyte cover/abundance over the sediment

Decomposition of particulate organic matter was studied using leaf litter bags containing Populus x canadensis (hybrid black poplar) leaves. The litter bag technique is described in Brock et al. [24]. In each microcosm, 2 litter bags were placed on tiles on the sediment surface, in an almost upright position, for a 2-wk incubation period. At the end of each incubation period, the 2 litter bags were gently retrieved from the microcosms and emptied into a white tray to separate particulate organic matter from adhering sediment particles and macroinvertebrates by rinsing with tap water. Macroinvertebrates were included in the overall macroinvertebrate sampling of the microcosm. The leaf material was dried in aluminum foil at a temperature of 105 8C. After 24 h, dry weight was determined. The decomposition over a 2-wk period was expressed as % organic material remaining. Data analysis Univariate analysis. Effective concentration, 50% and 10% (EC50 and EC10, respectively) values for the bioassays with M. spicatum were calculated using logistic regression. A Poisson distribution of the data was assumed [8]. In the case of biomass, the 100% effect was defined as a living biomass of 0 g (see Maltby et al. [8] for a visual representation). The model was programmed in GenStat for Windows [25]. Prior to univariate and multivariate analyses, abundance data were transformed to Ln(Ax þ 1), where x is the abundance value, taking the lowest abundance value higher than 0 (A ¼ 2/[lowest abundance value higher than zero]; see Van den Brink et al. [9] for the rationale). The aim of this transformation was to downweight high abundance values and to approximate a normal distribution of the data. The macroinvertebrate data were transformed to Ln(2x þ 1), the zooplankton data to Ln(10x þ 1), and the phytoplankton data to Ln(x þ 1) before statistical analysis. All other variables were tested as untransformed values. Noobserved-effect concentrations (NOECs) were calculated at the parameter or taxon level using the Williams test [26]. The analyses were performed with the Community Analysis computer program [27]. The analysis resulted in an overview of NOECs for the data analyzed for each sampling week. If the end point had been measured frequently (i.e., more than 3 times after the first application), effects were only considered when they were consistent, occurring on at least 2 consecutive sampling dates (see Effect class approach section, based on De Jong et al. [28]).

2778

Environ Toxicol Chem 33, 2014

R.P.A. van Wijngaarden et al.

Multivariate analysis. At the community level, the effects of azoxystrobin on macroinvertebrates and zooplankton were analyzed with the principal response curves (PRC) method, using the CANOCO software package [29]. The PRC method is based on the redundancy analysis ordination technique, the constrained form of principal component analysis [30]. In the CANOCO computer program, redundancy analysis is accompanied by Monte Carlo permutation to assess the statistical significance of the effects of the treatments on the species composition of the samples. The significance of the PRC diagram, in terms of treatment variance displayed, was tested by Monte Carlo permutation of microcosms (i.e., by permuting entire time series of microcosms in the partial redundancy analysis from which the PRC is derived), using an F-type test statistic based on the eigenvalue of the component [29]. To assess the significance of treatment effects for each sampling date, all treatments were also tested against the controls and against each other by means of Monte Carlo permutation tests.

species sensitivity distribution. All toxicity values had been generated under chronic conditions, following protocols. As for Daphnia magna, 3 toxicity values were available that referred to the same end point, so the geometric mean of these toxicity values was included in the distribution. This approach follows the procedure described in the European Food Safety guideline [15]. The method developed by Aldenberg and Jaworska [36], as incorporated in the ETX software [37], was used to generate the sensitivity distribution. Model fit was evaluated using the Anderson–Darling goodness-of-fit test for normality. Nondeterminate (greater than or lower than) values were not used in the log-normal regression. The HC5 and 50% hazard concentration (HC50) were derived from the species sensitivity distribution and compared with threshold values resulting from the microcosm study.

Effect class approach

The 42-d time-weighted average concentrations of azoxystrobin are presented in Table 2. During the 42-d experimental period, the analytically determined time-weighted average concentrations ranged from 93.5% to 99.3% of the intended values. At time 0 (immediately after the first application), the percentage of intended concentrations calculated from the measured concentrations in the dosing solutions ranged from 104.1% to 125.4%. Figure 1 shows the measured concentrations of azoxystrobin (mg a.i./L) over time for all the treatment regimes.

The NOECs obtained from the univariate and multivariate analyses were further analyzed with regard to statistical artefacts and biological significance. In the first instance, effects were considered consistent when they showed statistically significant deviations from the controls pointing in the same direction for at least 2 consecutive sampling points, or when a single sampling date showed effects shortly after the initiation of the treatment. Statistically significant deviations were further evaluated in relation to the magnitude of effects, whether counts were unevenly distributed over the samples, and whether there was a treatment-related concentration response or a clear causal relation with community interactions or timing. The observed effects of the azoxystrobin treatments were summarized by assigning the end points studied to end point categories and assigning effects to effect classes based on De Jong et al. [28]. Using this method, we defined the following effect classes based on the duration of the present study: effect class 1, no effects observed (evaluation of the statistical information played an important role in deciding whether results showed causal relations with treatments); effect class 2, slight effects (effects observed only in individual samplings, especially shortly after treatment initiation); effect class 3, clear short-term effects 2) for nauplii, cyclopoids, and D. longispina group in the PRC diagram indicate that the abundances of these taxa showed the best correlation with the community response. Calanoid copepods also had a positive species score, which was just above 1. The aforementioned taxa showed a strong overall treatment-related decline. In contrast, several other taxa showed an overall treatment-related increase, indicated by a negative species score (bk < 1); this included the rotifers Synchaeta sp. and Cephalodella gibba (Figure 2). Of the species with the highest positive species scores in Figure 2, the most pronounced treatment-related effects were observed for copepod nauplii, followed by cycloploids, D. longispina group, and calanoids (Figure 3 and Table 3). Effects on all of these 4 taxa became apparent immediately after the azoxystrobin application. For all 4 taxa, statistically significant treatment-related effects on abundance were observed on consecutive sampling days at the highest treatment concentration of 33 mg a.i./L (Figure 3). This treatment differed from the control over the entire experimental period. No recovery was observed at

Table 3. Consistent (bold font) no-observed-effect concentrations (NOECs) for azoxystrobin (mg a.i./L)a Days after initiation chronic treatment

Physiochemical Water chemistry Community metabolism pH Functional Decomposition Zooplankton Community (PRC) Copepoda Nauplii Calanoida Cyclopoida Cladocera Daphnia longispina group Rotifera Synchaeta sp. Cephalodella gibba Phytoplankton Community (PRC) Chlorophyll-a Cosmarium moniliferum Cosmarium crenulatum Cosmarium turpinii Ankyra sp. Chroococales 2–5 mm Periphyton Chlorophyll-a Macrophytes Community (PRC) Chara globularis Elodea nuttallii Macroinvertebrates Community (PRC) Population level

2–3

9–10

16

23

32

42

33 33 33

33 33 33

33 33 33

33 33 33

33 33 3.3

33 33 0.33

33

33

33

33

33

33

10

10

10

10

10

10

2

3.3 3.3 3.3

10 1.0 10

10 1.0 3.3

3.3 1.0 10

3.3 1.0 3.3

3.3 1.0 10

3A 3B 3C

1.0

3.3

10

10

10

10

3D

33(þ) 33(þ)

33(þ) 33(þ)

10(þ) 33(þ)

33(þ) 33(þ)

33(þ) 33(þ)

33(þ) 33(þ)

33 33 33 33 10 33 10

33 33 0.33 3.3(þ) 10 10(þ) 33

33 33 0.33 3.3(þ) 33 10(þ) 10

33 33 0.33 3.3(þ) 33 10(þ) 10

33 33 — — — — —

33 33 33 33 33 33 10

33

33

33

33

33

33

33 33 10

33 33 33

33 1 10

33 33 33

33 33 33

33 10 10

33 N/A

33 N/A

33 N/A

33 N/A

33 N/A

33 N/A

Concentrations > no-observed-effect concentration showed significant reductions or increases (þ). a.i. ¼ active ingredient; PRC ¼ principal response curves; — ¼ not sampled; N/A ¼ numbers were too low for evaluation. a

Figure

4A 4B 4E 4C 4D

2780

Environ Toxicol Chem 33, 2014

R.P.A. van Wijngaarden et al.

Nauplii

4

3 Cyclopoida Daphnia group longispina

2

1

bk

1 0.5

0

C dt

0 -0.5

-1

-1 -1.5

-2

-2 -5

0

5

10

15

20

25

30

35

40

45

0.33 μg /L

1 μg /L

3.3 μg /L

10 μg /L

33 μg /L

Euchlanis dilatata Squanella rostrum Ascomorpha. sp Keratella quadrata Scardium longicaudum Mylina ventralis Lepadella patella Cephalodella gibba Synchaeta. sp

Days post start applicaon Control

Calanoida Lecane luna group Anureopsis fissa Hexarthra. sp Chydorus sphaericus Alona sp Graptoleberis testudinaria Rhinoglena. sp

-3

Figure 2. Principal response curves resulting from analysis of the zooplankton dataset, indicating the effects of different azoxystrobin treatment levels. Sixteen percent of all variance could be attributed to the sampling date (displayed on the horizontal axis). Thirty-one percent of all variance could be attributed to treatment level, 31% of which is displayed on the vertical axis. The lines represent the development of the treatments in time. The species weight (bk) can be interpreted as the affinity of a taxon with the principal response curves (cdt). Taxa with a species weight between 0.25 and 0.25 are not shown. A Monte Carlo permutation test indicated that the diagram displays a significant amount of the variance explained by the treatment (p ¼ 0.002).

this treatment concentration (Figure 3). Clear effects also occurred at the 10 mg a.i./L treatment concentration, but these were less pronounced (nauplii, cyclopoids, D. longispina group) and were mostly followed by recovery for the cyclopoids and the D. longispina group. For nauplii, cyclopoids, and the D. longispina group, the above effects result in NOEC values of 3.3 mg a.i./L and 10 mg a.i./L in most cases (Table 3). Calanoids showed statistically significant reductions at concentrations of 3.3 mg a.i./L, 10 mg a.i./L, and 33 mg a.i./L (Table 3). For calanoids, no recovery was observed at the 3.3 mg a.i./L, 10 mg a.i./L, and 33 mg a.i./L treatment concentrations, resulting in NOEC values of 1 mg a.i./L (Table 3). Note, however, that densities of calanoids were low in all microcosms, including controls (generally