Predicting Pesticide Environmental Risk in Intensive Agricultural Areas. II

Dec 15, 2008 - Corresponding author phone: +39 0264482741; fax: +39 0264482795; e-mail: [email protected]., † ... This data set has been used to...
2 downloads 10 Views 398KB Size
Download PDF
Environ. Sci. Technol. 2009, 43, 530–537

Predicting Pesticide Environmental Risk in Intensive Agricultural Areas. II: Screening Level Risk Assessment of Complex Mixtures in Surface Waters ROBERTO VERRO,† ANTONIO FINIZIO,† S T E F A N O T T O , ‡ A N D M A R C O V I G H I * ,† Department of Environmental Sciences, University of Milano Bicocca, Piazza della Scienza 1, 20126 Milano, Italy, and Institute of Agro-environmental and Forest Biology - CNR, Agripolis, Viale dell’Universita` 16, 35020 Legnaro (PD), Italy

Received July 4, 2008. Revised manuscript received October 21, 2008. Accepted November 3, 2008.

In a previous article, a procedure for assessing pesticide ecotoxicological risk for surface water was applied to all active ingredients in a pilot basin. This data set has been used to assess the composition of pesticide mixtures that are likely to be present in surface waters as a consequence of pesticide emissions from the crops grown within the basin (maize, soybean, sugar beet, and vineyard). Temporal evolution of the mixture composition has been evaluated as a function of the different contamination patterns (drift and runoff). Ecotoxicological risk has been assessed for the mixtures released by individual crops and from all the relevant crops cultivated in the basin. The different role of drift and runoff, as well as the temporal trends of exposure and risk are compared. Daphnia is the most affected among the three indicator organisms considered, particularly from drift mixtures after insecticide application on vineyard. The highest risk for algae occurs during runoff events in spring. In most risk events, one or a few chemicals are usually responsible for more than 80% of the toxic potency of the mixture. The CA model for predicting mixture response is assumed to be a reliable approach for assessing risk for ecologically relevant pesticide mixtures.

Introduction Ecosystems are usually exposed to a cocktail of chemicals rather than one individual substance. This is particularly apparent in surface waters where a multitude of potentially toxic substances enter the watercourse as a result of human activities throughout the drainage basin. In the case of plant protection products, several active ingredients may be applied to the same crop, and there are usually several crops types present in any agricultural basin. Thus, aquatic ecosystems may be exposed to complex pesticide mixtures, which are highly variable in both space and time. The European Water Framework Directive, WFD, dictates fundamental changes in the way in which water bodies should * Corresponding author phone: +39 0264482741; fax: +39 0264482795; e-mail: [email protected]. † University of Milano Bicocca. ‡ Institute of Agro-environmental and Forest Biology - CNR. 530

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 2, 2009

be managed. The Directive specifies the need for protecting water ecosystems as a whole and not only on the basis of water quality objectives, defined as safe concentrations of individual chemicals. To enable European water bodies to achieve a satisfactory ecological status it has been deemed necessary to develop tools for defining water quality objectives for chemical mixtures (1). Crop production is responsible for the release of complex mixtures of chemicals into the environment, which can vary throughout the season according to treatment practices. It can be assumed that in a realistic scenario the composition and relative toxicity of mixtures is a function of the crops that are contemporaneously present in the catchment (different pesticides applications). Furthermore, the composition of the mixtures will also be dependent on the physical-chemical properties and environmental persistence of the active ingredients (a.i.). Finally, the toxicity of the mixture will vary as a function of the nontarget organisms considered. A procedure for assessing the risk posed to aquatic ecosystems by pesticide mixtures has been proposed by Finizio et al. (2). However, to ensure land use management in areas of intensive agriculture, as well as fulfilling the requirements of the WFD, there is the need for site-specific tools, capable of assessing the variability of chemical risk in space and time. A GIS-based procedure for mapping pesticide risk for surface water was developed by Verro et al. (3). The procedure has been applied on all a.i. used in an intensive agricultural area in order to give an indication of the risk that individual pesticides pose to aquatic ecosystems (4). In the present paper, data produced by Verro et al. (4) on individual chemicals, have been used to assess the risk from mixtures derived from agriculture in a pilot area. Two different kinds of mixtures are taken into account, based upon their origin: Mixtures deriving from all the different a.i. applied on a specific crop; this approach allows comparison to be made by the risk posed to aquatic ecosystems by specific crops rather than as a function of individual chemicals. Furthermore this approach allows the environmental impact of specific crop production to be compared to other human activities (e.g., industrial typologies). Such comparisons can provide valuable data sets relevant for land use management in areas of intensive agriculture. Total mixtures deriving from all crops present in a given area; this approach allows the prediction of the total threat that agriculture poses to an aquatic community. Such data are useful for managing the quality of surface water bodies, in order to comply with the requirements of the WFD. To predict the toxicological response to a mixture of chemical substances, two different approaches have been developed: the concentration addition (CA), and independent action (IA) models (5). The two models are applicable to chemicals with the same mode of action or with different modes of action, respectively. Substances, for which concentration-additive mixture toxicity is assumed, are expected to act at the same molecular target site, with a common mechanism of action. The concept of independent action introduced by Bliss (6) assumes that different substances cause a common integral biological effect (e.g., death) through primary interaction with different molecular target sites. Therefore, the response of the mixture is calculated as a combination of effect and not as a sum of concentrations (7). 10.1021/es801858h CCC: $40.75

 2009 American Chemical Society

Published on Web 12/15/2008

TABLE 1. Application Scenarios of Plant Protection Products Used in the Meolo Basina date

chemicals

1 March 2004 15 March 2004 8 April 2004

chloridazon (H)- metamitron (H) glyphosate (H) terbuthylazine (H) terbuthylazine (H)- dimethenamide (H)- isoxaflutole (H)- metolachlor (H)flufenacet (H)- glyphosate (H) terbuthylazine (H) terbuthylazine (H)- glyphosate (H) terbuthylazine (H) chlorothalonil (F)- clethodim (H)- clopyralid (H)- ethofumesate (H)phenmedipham (H)- lenacil (H) flufenacet (H)- metolachlor (H)- metribuzin (H)- linuron (H) glufosinate-ammoniun (H)- cymoxanil (F)- dimethomorph (F)- folpet (F)fosetyl-Al (F)- mancozeb (F)- metalaxyl (F)- metiram (F) dicamba (H)- mesotrione (H)- nicosolfuron (H) deltamethrin (I)- propaquizafop (H)- quizalofop C2H5 (H)- triflusulfuron CH3 (H) dimethomorph (F)- fosetyl-Al (F) folpet (F)- mancozeb (F)- metiram (F)- cycloxydim (H)- imazamox (H)imazethapyr (H)- oxasulfuron (H)- quizalofop C2H5 (H)- thifensulfuron- CH3 (H)- propargite (I) penconazole azoxystrobin (F)- cyproconazole (F)- difenoconazole (F)- fenpropidin (F)propiconazole (F)- tetraconazole (F)- trifloxistrobin (F)- desmedipham (H)cyhalothrin (I) chlorpyrifos - fenitrothion - flufenoxuron penconazole (F)- cyprodinil (F)- fludioxonil (F)- pirimethanil (F)procymidone (F) folpet (F)- mancozeb (F)- metiram (F) chlorpyrifos (I)- flufenoxuron (I) azoxystrobin (F)- cyproconazole (F)- difenoconazole (F)- fenpropidin (F)propiconazole (F)- tetraconazole (F)- trifloxistrobin (F)- desmedipham (H)cyhalothrin (I) fenitrothion (I) folpet (F)- mancozeb (F)- metiram (F) azoxystrobin (F)- cyproconazole (F)- difenoconazole (F)- fenpropidin (F)propiconazole (F)- tetraconazole (F)- trifloxistrobin (F)- desmedipham (H)cyhalothrin (I) cyprodinil (F)- fludioxonil (F)- pirimethanil (F)- procymidone (F)

9 April 2004 18 April 2004 20 April 2004 24 April 2004 27 April 2004 11 May 2004 15 May 2004 18 May 2004 29 May 2004 30 May 2004 15 June 2004 20 June 2004 1 July 2004 2 July 2004 10 July 2004 15 July 2004 16 July 2004 20 July 2004 26 July 2004 4 August 2004 10 August 2004 22 August 2004 a

crop sb v m m m m; s m sb s v m sb v v; s v sb v v v v sb v v sb v

H ) herbicide; F ) fungicide; I ) insecticide; m ) maize; v ) vineyard; s ) soybean; sb ) sugar beet.

Assuming that environmentally relevant mixtures have heterogeneous mechanisms of action, a two-stage approach (TSP) was introduced (8), combining the CA and IA models. Conceptually, the TSP approach is the best to assess pesticide mixtures that can be expected to be neither strictly similarly nor strictly dissimilarly acting. However, there are several limitations to the application of TSP and IA. Usually, the mode of action of pesticides is only well-known for the target organisms, whereas it remains largely unknown for the nontargets. Moreover, according to IA, to predict the contribution of individual toxicants to the toxicological potency of a mixture, it is necessary to obtain toxicological details from concentration-response curves (slope), data that are rarely reported in the literature. Both CA and IA approaches do not take into account possible interactions among chemicals that may produce synergistic or antagonistic effects. With the present knowledge, these effects cannot be predicted with relatively simple models and must be studied case by case. Therefore, they are not been considered in the described procedure. Taking into account that the mixture responses calculated using the CA model are usually higher than those calculated with the IA model, CA can be assumed as a conservative worst case (2, 7, 9-11). Moreover, it has been demonstrated that the ratio CA/IA for complex mixtures, even for those composed of a high number of chemicals, is generally relatively low and rarely exceeds 1 order of magnitude (2, 7). Thus, CA can be assumed as a “reasonable” worst case.

In this study, the composition of mixtures released from four different crops and the potential responses calculated with the CA model are described. Furthermore the temporal changes in the composition of the mixtures are reported according to application dates (drift) and each specific rain event (runoff) that occurred during the growing period. The potential risk posed by the mixtures calculated for the indicator organisms representative of the aquatic trophic chain (TUs for algae, Daphnia, and fish) and for the whole aquatic ecosystem (PRISW-1 values) are also reported.

Materials and Methods Description of the Site and Development of Scenarios. The study was conducted on the basin of the Meolo River (northeast Italy), a small resurgence river that flows into the Venice Lagoon. The average discharge of the watercourse has been estimated in the range of 2-3 m3 s-1. The Meolo’s basin covers an area of 2817 ha and is characterized by relatively flat land. A detailed description of the basin is reported by Bonzini et al. (12) and by Verro et al. (4). Information on application rates (times and amounts) for the 54 active ingredients (a.i.) applied to the four main crops (maize, soybean, sugar beet, and vineyard) in the 2004 growing season are also reported in Verro et al. (4). The pesticide application scenario is a key factor, especially for determining the composition of drift mixtures. Given the scarcity of data detailing the applications, namely location, date and mass, the application scenarios of all VOL. 43, NO. 2, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

531

FIGURE 1. Temporal trend of mixture toxicity (TUs > 0.001) derived from the four main crops present in the Meolo watershed. Numbers on the time axis represent days after March 1st, starting date of the procedure. spraying events were grouped into 25 different dates. The application dates were selected according to realistic application windows for all the pesticides used on specific crops in the studied area. Given this simplification, this may result in the reporting of unrealistic chemical mixtures being applied on a given day. However, such data sets will represent a worst-case scenario in the screening level risk assessment of the impact of agricultural activity on aquatic ecosystems. The application time sequence is reported in Table 1. Besides the 25 drift events, 21 rain events were taken into account for runoff during the considered period (from March to October), for a total of 46 emission events. The method for calculating predicted environmental concentrations (PECs) from drift and runoff processes for individual chemicals is described in detail by Verro et al. (4). Risk Characterization for Pesticide Mixtures. For each component of the mixture, toxic units have been calculated as the ratio between the PECs and a toxicological end point (EC50, LC50) for the three representative nontarget organisms of the aquatic environments (algae, Daphnia, and fish). The toxicological potency of mixtures was calculated using the CA model according to the following equation: n

TUm )

n

Ci

∑ TU )∑ EC i

i)1

i)1

(1)

x,i

where Ci is the concentration of the individual chemical “i”; ECx,i is the effect concentration (e.g., EC50) of the individual chemical “i”; TUi are the toxic units of the individual chemical “i”, i.e., the fraction of the ecotoxicological end-point produced by the chemical (TUi ) Ci/ECx,i); and TUm are the toxic units of the mixture. Application of the PRISW-1 Index to Pesticide Mixtures. In order to obtain data relating to the risk that pesticide mixtures pose in the Meolo River, the short-term pesticide risk index for the surface water system (PRISW-1) was applied 532

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 2, 2009

(13). The index is a scoring system that requires the calculation of TERs (toxicological exposure ratios) obtained by the ratio between acute toxicity (EC50 or LC50) data for selected organisms (algae, Daphnia, fish) and PECs in surface water. To each TER a score is assigned, which is combined with a weight related to the type of organism, according to the following equation: PRISW-1 ) (A × Wa) + (B × Wb) + (C × Wc)

(2)

where A, B, C are the scores assigned to the TERs for algae, Daphnia, and fish, respectively; and Wa, Wb, and Wc are the weights. More details on the index structure and on the rationale behind it are reported in the Supporting Information (SI) (see Table 1) and in the companion paper (13). Originally this index was developed for individual chemicals; however, considering the screening level risk assessment approach used in this study (CA model), PRISW-1 is applicable to pesticide mixtures. In order to give a representative score, the reciprocal of TUs (1/TU ) TER) was utilized.

Results and Discussion Mixture Released from Different Crops. In Figure 1 the temporal trend of TUs released by the relevant crops present in the Meolo watershed are reported. Only mixtures showing TUs > 0.001 have been taken into account. A TU equalling 0.001 was assumed as a threshold of risk because 1000 is the application factor for calculating a PNEC from acute toxicity data proposed by the European Technical Guidance Document (TGD) on Risk Assessment (14). Studies on natural communities seem to confirm the validity of this threshold. Liess and Von der Ohe (15) observed significant effects on communities of aquatic invertebrates exposed to pesticides in the range 0.001 < TUs < 0.01, whereas no significant effects were observed in the range 0.0001 < TUs < 0.001. In another study (16) effects were observed even at slightly lower

pesticide exposure. However, in the present work, the threshold TUs > 0.001 has been preferred, being referred to an official European procedure. In Figure 1, drift and runoff are considered separately and the TUs refer to the three selected aquatic organisms. The toxicity of chemical mixtures on the different species varied significantly during the exposure period as a function of both the crop type and the risk event (drift or runoff) considered. For algae the TUs never exceeded a value of 1 (acute toxicity) throughout the considered period. Until the end of September TUs released from maize, soybean, and sugar beet (Figure 1) are in almost all cases close to or higher than 0.01. This is most apparent in the first phase of the growing period, which corresponds with the application of terbuthylazine and flufenacet on maize, (Figure 1 and Table 1) where TUs > 0.1 indicate a potential negative effect on algal communities. For these organisms maize and in a minor extent soybean pose a significant threat. With the exception of vineyards (Figure 1), the algal toxicity of mixtures produced by runoff is higher than those produced by drift. This is due to the physical-chemical properties of the compounds used (herbicides are often less hydrophobic than insecticides) and their application patterns (applied on soil and not sprayed on crops). Therefore it can be concluded that runoff is the predominant risk pathway for algae. Out of all the nontarget organisms investigated in this study, Daphnia is potentially the most affected. The risk for Daphnia starts in May as a result of the application of chlorotalonil on sugar beet (runoff event 30 April) and mancozeb and folpet on vineyards (drift events). In July the application of insecticides on vineyards (chlorpyrifos, fenitrothion, flufenoxuron drift events), result in maximum TUs which exceed the level of acute toxicity. Finally, in August, the risk posed to Daphnia appears relatively constant, remaining >0.1 TUs due to a.i. release from vineyards (Figure 1 and Table 1). The risk to fish appears to be lower in comparison to the other nontarget organisms (Figure 1). There are only two instances in the study period where the TUs for fish are >0.1 (mancozeb, drift vineyards, Figure 1 and Table 1). There are also a few episodes when TUs are between 0.1 and 0.01 (Table 2), corresponding to drift events from vineyards treatments. All the TU values, for individual chemicals and mixtures, are listed in SI Tables 2 and 3, respectively. To obtain a more complete picture of the risk posed by different crops on the aquatic ecosystem, the PRISW-1 index was applied (Figure 2), whereby runoff and drift were considered separately. Generally, mixtures released from vineyards resulted in the highest PRISW-1 scores, indicating that this crop poses a serious threat to the aquatic system. For a large part of the growing season (May-September) the PRISW-1 scores for vineyard are distributed between the high and very high classes of risk. The highest scores for vineyards correspond to drift events. However, it must be remembered that, according to the application scenario described by Verro et al. (4), it was assumed that all vineyards are treated the same day. Therefore drift events represent worst cases. All the other crops considered pose a relatively lower risk, with scores being generally distributed between the classes of negligible, low, and medium risk. The only exception was observed for runoff events for sugar beet during the first phase of the growing cycle, mainly due to the release of the fungicide chlorotalonil. Mixture Released from the Total Basin. In Figure 3 the TUs of pesticide mixtures at the basin level are reported. The mixtures are the result of the combined emission of a.i. from the different crops throughout the basin. The TU values, reported in Figure 3, were calculated by summing up the TUs deriving from individual crops in a particular risk event for both drift and runoff scenarios. These TU values (SI Table

3) are the most ecologically relevant, as they represent the mixtures which the ecosystem is exposed. To compare the contribution of different crops to the total risk, all risk events where TUs > 0.01 at basin level (considering all crops) are reported in Table 2 along with the crop type that made the highest contribution (%) to the overall toxic potency of the mixture. Maize is potentially the most toxicologically relevant crop for algae. From mid April to late July the a.i. applied on maize represented the largest contribution to the total mixture predicted in the river. Soybean became toxicologically relevant from mid May, whereas sugar beet only became significant in early August. Emissions from vineyards were only of relevance for algae in a limited number of drift events through the study period, corresponding to the application of fungicides. The mixtures released from vineyards pose the greatest toxicological risk to both Daphnia and fish from mid May until the end of the observed period. Prior to May, fungicide applications to sugar beet represent some risk to Daphnia and fish, whereas maize and soybean are shown not to represent a substantial threat to these components of the aquatic community. Further details on the contribution of individual chemicals to the overall mixture potency are given in SI Table 4. The chemicals responsible for the greatest contribution (%) to the toxic threat to the three nontarget organisms are also reported for both drift and runoff events. The following considerations are possible. (a) In most risk events, just one or a few chemicals are usually responsible for the threat posed by mixtures to the aquatic ecosystem. In almost all cases, the greatest contribution (usually more than 80%) to the toxic potency of a mixture is represented by no more than three chemicals. In 23% of the cases, only one compound was observed to be the major cause of the overall toxicity of the mixture. This finding is particularly evident when considering drift events and to a lesser degree runoff, especially in the first period of pesticide application. These results are also in agreement with data on pesticide mixture reported in previous papers (2, 10, 17) where it was observed that, in most cases, only a few active ingredients dominate the toxicity of mixtures. These results are extremely important from a risk management perspective as they indicate that, by applying risk mitigation measures to a small number of active ingredients, a substantial reduction of the adverse effects posed by mixtures to aquatic systems could be achieved. (b) Of a total of 25 drift risk events (DRE) only nine result in TUs > 0.01, posing a potential risk to at least one of the three nontarget organisms considered. Of these DRE only a few chemicals are responsible for the major risk: the herbicide flufenacet, in the first part of the season, (from April to May); the fungicides mancozeb and folpet as a consequence of repeated applications from middle May to August; and the insecticides flufenoxuron, chlorpyrifos, and fenitrothion in July. During this period, the TUs of the mixtures in some cases exceed the level of acute toxicity for Daphnia (TUs > 1). As previously mentioned, it must be remembered that risk events are worst cases. In contrast, in all the 21 runoff risk events (RRE) the TUs are greater than 0.01 for one or more of the aquatic organisms. Due to the different environmental properties and application patterns of individual chemicals, the composition and the relevant chemicals in runoff mixtures are generally different in comparison to that of drift mixtures. (c) Algae are always subject to the toxicological stress posed by pesticide mixtures and this is most apparent during runoff events (see also Figure 3). In fact in all RRE the TUs for algae are always >0.01; furthermore, between April 30 and early June, the TUs always exceed 0.1. The chemicals VOL. 43, NO. 2, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

533

534

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 2, 2009

1.6 × 10-1

1.9 × 10-1

1.2 × 10-1 2.7 × 10-1

1.1 × 10-1 3.2 × 10-1 1.1 × 10-1

7.2 × 10-1 4.3 × 10-1 3.2 × 10-1 2.1 × 10-1 1.9 × 10-1 5.5 × 10-1 2.5 × 10-1

TUall crops > 0.1

6.9 × 10-2 1.8 × 10-2 4.1 × 10-2 3.0 × 10-2 5.3 × 10-2 5.6 × 10-2 1.4 × 10-2

1.0 × 10-2 7.2 × 10-2

7.0 × 10-2 5.6 × 10-2

3.4 × 10-2

0.01

m., s. (46%, 40%) s. 60, m21., b. 17 v. (100%) all crops b., m., s. (42%, 28%, 28%) b., m., s. (38%, 32%, 28%) b. 51, m. 30, s. 15 b., m., s. (40%, 31%, 24%) b., m., s. (40%, 32%, 23%) b., m., s. (45%, 27%, 23%)

v. (100%)

m. (92%) m. (89%) m. (91%) m. (91%) m. (89%) m. (90%) m. (88%) s. (100%) v. (100%) m., s. (56%, 40%) s., m. (70%, 29%) m., s. (67%, 30%) m., s. (64%, 33%) v. (96%) s., m. (43%, 56%)

crop % contribution to the total TUs

7.4 × 10-1 1.0 × 10-1 5.5 × 10-1 1.1 × 10-1 2.8 × 10-1 1.4 × 10-1 3.1 × 10-1 3.9 × 10-1

63 1.2 × 10-1 2.0 42

TUall crops > 0.1

9.7 × 10-2

7.2 × 10-2

7.3 × 10-2

v. (100%) v. (100%) v. (100%) v. (100%) v. (100%) v. (100%) v. (100%) v. (100%) v. (100%) v. (100%) v. (100%) v. (100%) v. (100%) v. (100%)

v. (99%)

v. (100%)

b. (97%)

1.6 × 10-2 6.8 × 10-2

b. (97%) b. (96%)

crop % contribution to the total TUs

1.8 × 10-2 1.3 × 10-2

0.01

Daphnia

1.2 × 10-1

1.4 × 10-1

TUall crops > 0.1

v. (100%) v. (100%) v. (100%)

6.3 × 10-2

v. (91%)

1.0 × 10-2

1.2 × 10-2

v. (92%) v. (100%) v. (92%)

1.2 × 10-2

6.2 × 10-2

v. (99%)

v. (100%)

b. (90%) b. (91%)

b. (88%) b. (90%) b. (89%)

crop % contribution to the total TUs

8.4 × 10-2

7.8 × 10-2

2.5 × 10-2 1.3 × 10-2

2.7 × 10-2 2.0 × 10-2 1.3 × 10-2

0.01

fish

a Crops (m ) maize; s ) soybean; v ) vineyard; b ) sugar beet) giving the major contribution to the total mixture are reported. In bold TU > 1, representing potential acute toxicity, are reported.

30 April (r) 4 May (r) 5 May. (r) 6 May. (r) 7 May. (r) 8 May. (r) 9 May. (r) 11 May (d) 15 May (d) 22 May (r) 3 June (r) 12 June (r) 13 June (r) 15 June (d) 25 June (r) 02 July (d) 15 July (d) 16 July (d) 26 July (d) 27 July (r) 3 August (r) 4 August (d) 7 August (r) 13 August (r) 21 August (r) 31 August (r) 14 September (r) 16 Setember (r) 24 September (r)

date

algae

TABLE 2. Emission Events (d = drift; r = runoff) That Result in ΣTU > 0.01 for the Total Chemical Mixture on at Least One Non Target Organism (Algae, Daphnia, and Fish)a

FIGURE 2. PRISW-1 scores for mixtures deriving from the main crops in the Meolo River watershed. Numbers as in Figure 1. that pose the greatest risk to algae are the herbicides terbuthylazine and to a lesser extent imazamox. For Daphnia, two different risk maxima corresponding to RRE can be observed. The first episode is observed between the end of April until the middle of May (TUs ∼ 0.01) and is a result of the application of chlorotalonil on sugar beet. The second major RRE occurs between late July and September and is characterized by higher toxicity (TUs ∼ 0.1) due to insecticide applications (fenitrothion, chlorpyrifos, and flufenoxuron) on vineyards. Fish seem to be less sensitive to pesticide mixtures in comparison to the other two nontarget organisms. The toxicologically relevant chemicals for fish are similar to Daphnia; however, the levels of risk are significantly lower. Risk from Mixtures and Individual Chemicals at Ecosystem Level. In Figure 4, the PRISW-1 values corresponding to the mixtures found in the whole-basin are reported. According to a risk classification (4, 13), the PRISW-1 values in the 46 emission events (drift and runoff) can be listed as follows: • 4 negligible (100% drift); • 7 low (100% drift); • 17 medium (47% drift and 53% runoff); • 12 high (33% drift and 67% runoff); • 6 very high (33% drift and 67% runoff). While the highest risk values (PRISW-1 ) 75) correspond to two drift events, runoff events produce more frequent conditions of concern. According to the risk classification, of the 21 runoff events observed in this study, 43% represent a medium risk, 38% a high risk, and 19% a very high level of risk. From this assessment, the role of different crops and the increasing complexity of the mixture can be evaluated. In many cases, the calculated PRISW-1 values for mixtures are close to those of the most potent a.i. present, especially during the spring and early summer. In contrast, particularly for

runoff events in late summer, the total PRISW-1 values of mixtures are substantially higher than those of the most potent a.i. These mixtures are composed of chemicals from different classes (herbicides, fungicides, insecticides) capable of posing a risk to all components of the aquatic community. Value of the Approach and Relevance for Management Purposes. The proposed approach allows for the risk assessment of environmentally relevant pesticide mixtures that are likely to occur in surface waters in areas of intensive agriculture. The data generated are of importance for the development of land management practices which ensure compliance to international regulations such as the European Water Framework Directive with regard to water quality. The information generated in this study is of value at different stages of the environmental assessment procedure: • assessing the potential threat that mixtures pose to the aquatic community and the needs for mitigation measures; • comparing the potential impact of different crops in an agricultural basin and their roles during the productive season; • assessing the potential relevance of individual chemicals in the mixtures in order evaluate priorities and to optimize more environmental friendly agronomic practices. The assessment of the different role of drift and runoff, as well as the relevance of different chemicals in the two release pathways, provide valuable information for land management purposes and on the need for different mitigation measures to be applied. Finally, when evaluating the described results, the simplifications made in the protocol must be considered. As outlined in the scenario description, the need for grouping chemical releases over a selected number of application dates was necessary in order to simplify drift scenarios. However, in doing so this will have resulted in the VOL. 43, NO. 2, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

535

FIGURE 3. Temporal toxicity trend (TUs > 0.001) of mixtures at the basin level. Numbers as in Figure 1.

FIGURE 4. PRISW-1 scores for the total mixture in the 46 emission events. Dots on the lines represent the PRISW-1 scores of the most dangerous individual component of the mixture. Numbers as in Figure 1. characterization of multicomponent drift mixtures. As it has been demonstrated that only a limited number of chemicals are responsible for the majority of the relative toxicity of mixtures, particularly in drift events (Figure 4 and SI Table 3), the risk events identified may be assumed as “reasonable” worst cases scenarios. Another approximation is the use of the CA model for calculating the effect of mixtures. In addition to the reasons already highlighted there are a number of further considerations that support the suitability of this model for the described application: • The high prevalence of a few chemicals in determining the toxic potency of a mixture is one of the key factors affecting the difference between the CA and IA models (10, 11, 18). The results generated in this study demonstrate that even in mixtures with a high number 536

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 2, 2009

of components (up to 54 in the Meolo basin mixtures), a few chemicals (usually no more than three) are responsible for >80% of the toxicity. This observation therefore renders differences between CA and IA predictions very small. • With respect to individual nontarget organisms, the most toxic components of the mixtures often have the same mode of action (e.g., organophosphorus insecticides on Daphnia). In contrast, for those chemicals with different, or unknown, modes of action (e.g., herbicides on Daphnia) the toxic risk posed was found to be either low or negligible. • For the assessment of the risk that chemicals pose to ecosystems (PRISW-1), individual TERs derived from the application of CA are summed and relate to the overall risk to the aquatic community (effects on algae,

Daphnia, and fish) taking into account the different characteristics of the chemicals. It follows that the CA model can be assumed as a reliable approach for assessing risk for ecologically relevant pesticide mixtures.

(8)

Acknowledgments This research was supported by the Italian Ministry of University and Research (Project: GIS-based assessment of the ecotoxicological risk deriving by pesticide use, COFIN 2002) and by the European Commission (FP6 Contract No. 506675, ALARM and Contract No. 003956, NoMiracle). We acknowledge Mr. Oddino Bin for supporting field data collection.

(9) (10)

(11)

Supporting Information Available Tables S1-S4. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Vighi, M.; Altenburger, R.; Arrhenius, Å; Backhaus, T.; Bo¨deker, W.; Blanck, H.; Consolaro, F.; Faust, M.; Finizio, A.; Froehner, K.; Gramatica, P.; Grimme, L. H.; Gro¨nvall, F.; Hamer, V.; Scholze, M.; Vighi, M.; Walter, H. Water quality objectives for mixtures of toxic chemicals: problems and perspectives. Ecotoxicol. Environ. Saf. 2003, 53, 139–150. (2) Finizio, A.; Villa, S.; Vighi, M. Predicting pesticide mixtures load in surface waters from a given crop. Agric. Ecosyst. Environ. 2005, 111, 111–118. (3) Verro, R.; Calliera, M.; Maffioli, G.; Auteri, D.; Sala, S.; Finizio, A.; Vighi, M. GIS based system for surface water risk assessment of agricultural chemicals. 1. Methodological approach. Environ. Sci. Technol. 2002, 36, 1532–1538. (4) Verro, R.; Finizio, A.; Otto, S.; Vighi, M. Predicting pesticide environmental risk in intensive agricultural areas. I: Screening level risk assessment of individual chemicals in surface waters. Environ. Sci. Technol. 2009, 2, 522–529. (5) Greco, W.; Unkelbach, H. D.; Po¨ch, G.; Suhnel, J.; Kundi, M.; Bodeker, W. Consensus on concepts and terminology for combined-action assessment: the Saariselka¨ agreement. Arch. Complex Environ. Stud. 1992, 4, 65–72. (6) Bliss, C. I. The toxicity of poisons applied jointly. Ann. Appl. Biol. 1939, 26, 585–615. (7) Faust, M.; Altenburger, R.; Backhaus, T.; Blanck, H.; Bo¨deker, W.; Gramatica, P.; Hamer, V.; Scholze, M.; Vighi, M.; Grimme,

(12)

(13) (14)

(15) (16)

(17)

(18)

L. H. Joint algal toxicity of 16 dissimilarly acting chemicals is predictable by the concept of independent action. Aquat. Toxicol 2003, 63, 43–63. Junghans, M.; Backhaus, T.; Faust, M.; Meyer, M.; Scholze, M.; Grimme, L. H.; Predicting the joint algal toxicity og chemical mixtures using a mechanism-based two stage prediction (TSP). In: Studies on combination effects of environmentally relevant toxicants., PhD Thesis, University of Bremen, Germany, 2004, pp 78-108. Drescher, K.; Boedeker, W. Assessment of the combined effects of substances: The relationship between concentration addition and independent action. Biometrics 1995, 51, 716–730. Boedeker, W.; Drescher, K.; Altenburger, R.; Faust, M.; Grimme, L. H. Combined effects of toxicants: The need and soundness of assessment approaches in ecotoxicology. Sci. Total Environ. 1993, 931–938. Junghans, M.; Backhaus, T.; Faust, M.; Scholze, M.; Grimme, L. H. Application and validation of approaches for the predictive hazard assessment of realistic pesticide mixtures. Aquat. Toxicol. 2006, 76, 93–110. Bonzini, S.; Verro, R.; Otto, S.; Lazzaro, L.; Finizio, A.; Vighi, M. Experimental validation of a GIS-based procedure for predicting pesticide exposure in surface water. Environ. Sci. Technol. 2006, 40, 7561–7569. Finizio, A.; Calliera, M.; Vighi, M. Rating systems for pesticide risk classification on different ecosystems. Ecotoxicol. Environ. Saf. 2001, 49, 262–274. Technical Guidance Document (TGD) on Risk Assessment of Chemical Substances, EUR 20418 EN/2, 2nd ed.; European Commission, European Chemical Bureau, Joint Research Centre: Luxembourg, 2003. Liess, M.; von der Ohe, P. C. Analyzing effects of pesticides on invertebrate communities in streams. Environ. Toxicol. Chem./ SETAC 2005, 24, 954–65. Scha¨fer, R. B.; Caquet, T.; Siimes, K.; Mueller, R.; Lagadic, L.; Liess, M. Effects of pesticides on community structure and ecosystem functions in agricultural streams of three biogeographical regions in Europe. Sci. Total Environ. 2007, 382, 272– 285. Henning-De Jong, I.; Van Zelm, R.; Huijbregts, M. A. J.; De Zwart, D.; van der Linden, T. M. A.; Wintersen, A.; Posthuma, L.; van de Meent, D. Ranking of Agricultural Pesticides in the RhineMeuse-Scheldt basin based on toxic pressure in marine ecosystems. Environ. Toxicol. Chem./SETAC 2008, 27, 737–745. Backhaus, T.; Faust, M.; Scholze, M.; Gramatica, P.; Vighi, M.; Grimme, L. H. Joint algal toxicity of phenylurea herbicides is equally predictable by concentration addition and independent action. Environ. Toxicol. Chem./SETAC 2004, 23, 258–264.

ES801858H

VOL. 43, NO. 2, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

537