Occurrence of pesticides in surface water bodies: a critical analysis of the Italian national pesticide survey programs

Abstract

In recent years, several monitoring programs have been established as a consequence of EU Directives which include requirements for monitoring the quality of water resources (drinking water, groundwater and surface water). Plant protection products (PPPs) are an obvious target for monitoring activities, since they are directly released into the environment. In Italy, the National Environmental Protection Agency (ISPRA ex APAT) has recently published two reports containing the results of the National Plans for Control of Environmental Effects of PPP. These documents contain the collection of monitoring data related to the presence of PPP residues in surface water and groundwater. The first objective of this work was to critically analyze the results of monitoring campaigns on pesticide residues in surface waters in Italy. In particular, the paper focuses on whether and how the Italian approach satisfies the requirements of the Water Framework Directive, whether they are representative of the whole national territory, and how the utilized approach is appropriate for the characterization of risk to surface water. Starting from this analysis, some considerations about the limits of environmental monitoring as a tool for managing risk were highlighted.

---

Environmental impact

The Water Framework Directive (WFD), entering into force in 2000, is a milestone in EU water policy. Considering the importance of this directive, it is interesting to analyse the response of Member States (MS) to its provisions, particularly on the development of appropriate monitoring programmes for chemicals in waters. According to the WFD, MS shall ensure monitoring programmes of water systems to establish a coherent and comprehensive overview of water status at the EU level. In this paper, monitoring data on pesticide residues detected in Italian surface water bodies were reviewed, with the aim to highlight whether and how the Italian approach satisfies the requirements of the WFD, and whether it may be considered representative of the whole national territory.

Introduction

The EU Directive 2000/60/EC (WFD: Water Framework Directive),¹ adopted in Italy with the Legislative Decree 152/2006,² established a framework for developing measures for the conservation, protection and improvement of the quality of water, a limited and vulnerable resource. The quality of water bodies deserves particular attention due to its importance in ecosystem survival as well as in the supply of water for humans. According to the WFD, water resources have to be assigned to River Basin Districts (RBD) which will be subject to River Basin Management Plans to determine how the quality objectives for all the water bodies within each river basin are to be achieved. According to such regulation, for surface waters, both the ecological and the chemical status have to be determined. Water bodies within RBDs should be monitored and the factors that adversely impact on their status should be identified; moreover, a program of action to prevent or reverse the adverse impact has to be planned.

To evaluate the chemical status of water bodies, the WFD introduced criteria for establishing a list of priority substances and priority hazardous substances, for which specific measures, such as quality standards and emission controls, has to be established to reduce or eliminate emissions, discharges and losses. A list of 33 priority substances was adopted in 2001, 13 of these can be classified as PPPs.³ Whilst the Community proposed measures for priority substances by the end of 2003, Member States had to prepare comprehensive programmes of measures within river basin management plans by 2009, which included the identification and monitoring of pesticides able to produce significant risk especially in sensitive zones (vulnerable areas) and, eventually, the prohibition of use of certain products in water catchment areas. For PPPs, these aspects have been further strengthened by the recent promulgation of the Directive on Sustainable Use of Pesticides⁴ which states that Member States should establish National Action Plans (NAPs) to reduce risks associated to pesticide uses. In this context, the development of appropriate monitoring programmes for risk management of water bodies assumes a crucial role. As a matter of fact, the measurement of chemical concentrations in water bodies represents the first step for the identification of an appropriate class of quality.⁵ Considering the PPPs not included in the list of priority substances, it has to be stated that there are still a lack of environmental quality standards, adopted at EU level, necessary to the classification of surface water bodies. For PPPs, however, there are a number of different aspects that hinder the planning of monitoring campaigns, such as the water body type, the huge number of active ingredients, the widespread use, crop distribution, the seasonality, and the peak pulses deriving from drift or runoff events.

To address these problematic aspects, in recent years, the use of models coupled with GIS systems have been proposed to predict the environmental concentration of PPPs in surface water systems.⁶ According to the WFD, standards for the protection of aquatic life have to be established from PNEC values (Predicted No Effect Concentration), derived from acute and chronic toxicological data on aquatic organisms as suggested by the EU Technical Guidance Documents (TGD).⁷ In Italy, regions are charged with water monitoring activities while the national government is empowered with supervision, coordination and regulation tasks. In this framework, the Italian Environmental Protection Agency (ISPRA ex APAT) has started monitoring campaigns to verify and quantify the occurrence of PPPs both in surface and groundwater systems and the results have been recently published (National Plan for Control of Environmental Effects of Plant Protection Products http://www.isprambiente.it) in two different documents.^8,9 To define the effect on surface water ecosystems, the monitored concentrations were compared to 0.1 μg L⁻¹, the standard limit of acceptable pesticide concentrations (for single PPPs) in drinking water. It has to be stated that the use of this value as a reference for PPP risk characterization in surface water systems can be somewhat misleading since this parameter is not related at all to the possible effects of a single pesticide to the aquatic communities. When an ecosystem is considered, a PNEC value has to be taken into account, as this parameter reflects the inherent toxicity of a substance.

In this paper, monitoring data on PPP residues detected in Italian surface water bodies, recently published by the Italian Environmental Protection Agency, will be critically analysed. In particular, this analysis will highlight whether and how the Italian approach satisfies the requirements of the WFD, whether it may be considered representative of the whole national territory, and if the risk assessment for surface water can be considered appropriate. Finally, considering the role that monitoring might assume in environmental risk management for chemicals, the pros and cons of this approach will be highlighted. The authors believe that such analysis, even if not requested by any regulatory body, might support people involved in establishing the compliance of Italy to the WFD.

Materials and methods

Monitoring and presence of pesticide residues in surface waters of the Italian territory.

The most recently published monitoring data on the presence of PPP residues in Italian surface water bodies (data refer to years 2005 and 2006) were analysed.^8,9 In the first part of both documents, the measured concentrations of PPP residues referred to the whole national territory and were reported as concentration percentiles (from 25th to 95th). However, in the 2007 APAT report⁸ (reporting data of 2005), all samples showing no measurable concentrations for a specific substance were excluded in the percentile calculations of concentration values. On the contrary, in 2008, the calculation of percentiles for monitoring data of 2006, was performed either considering just samples containing residues (as in the 2007 report) or considering the total number of analysed samples. In the latter case, samples with residue concentrations below the analytical detection limits (LOD: limit of detection) were taken into account: to each sample with no residue, a concentration value of half of the respective LOD was assigned. According to the reports a total number of PPPs and metabolites monitored in the Italian surface water network were 255 and 315 in 2005 and 2006, respectively. In 2005, 112 substance residues were found, while in 2006 there were 125. Herbicides accounted for the majority of the substance analysed and detected. In Table 1, a selection of monitored substances in surface water systems at national level is reported.

Table 1 PPPs and metabolites mainly monitored in Italian surface water systems (2005 and 2006): % of monitoring stations (% on the total of sampling station in which the pesticide was included in the list of substances to be monitored), % of presence in monitoring stations, % of presence on total analysed samples and percentiles values ([95th]) of detected water concentrations. Percentiles of 2006 were calculated with just samples containing residues as in 2005 (identified with *) or considering the total number of analysed samples (identified with **). Each sample with no residue is represented with a concentration value of half of the respective LOD (Limit of Detection).

Substance	2005			2006			Concentration percentiles
	% of monitoring stations	% presence in monitoring stations	% presence of total analysed samples	% of monitoring stations	% presence in monitoring stations	% presence of total analysed samples	[95th]/μg L^−1 2005	[95th]/μg L^−1 2006*	[95th]/μg L^−1 2006**
alachlor	75	7.5	1.3	82	9.6	2	2.9E^−01	3.9E^−01	5.0E^−02
atrazine	72	20.1	8.1	83	17.3	7	8.0E^−02	9.0E^−02	5.0E^−02
azinphos-C2H5	24	1.2	0.2	34	0.3	0.1	3.1E^−01	1.0E^−02	2.5E^−02
azinphos-CH3	28	11.9	4.4	38	12.0	6	4.5E^−01	2.7E^−01	2.5E^−02
benfluralin	16	10.1	1.4	24	0.7	0.1	3.0E^−02	2.0E^−02	2.5E^−02
chlorpyrifos	58	2.9	0.5	65	2.9	1	2.4E^−01	1.9E^−01	1.9E^−01
chlorpyrifos-CH3	47	1.0	0.2	54	1.3	0.2	6.4E^−02	1.4E^−01	5.0E^−02
desethyl-atrazine	58	17.3	9.4	70	18.0	7	8.0E^−02	1.1E^−01	5.0E^−02
desethyl-terbuthylazine	55	48.1	28.1	71	44.7	21	2.4E^−01	2.7E^−01	8.0E^−02
desisopropyl-atrazine	20	3.5	1.1	36	4.0	1	7.9E^−01	2.9E^−01	5.0E^−02
diazinon	38	3.9	0.7	44	0.8	0.2	2.4E^−01	3.3E^−01	2.4E^−01
dichlorvos	22	1.3	0.2	31	3.2	1	1.6E^−01	2.0E^+00	2.5E^−02
dimethoate	12	6.7	1.1	33	7.9	2	2.0E^+00	3.7E^−01	3.0E^−02
diuron	17	8.6	2.5	27	9.2	3	2.4E^−01	3.4E^−01	5.0E^−02
fenitrothion	23	0.0	0	38	1.2	0.2	<LOD	1.7E^−01	2.5E^−02
iprodione	21	3.6	1.1	33	0.5	0.1	7.4E^−02	5.0E^−02	5.0E^−02
linuron	36	1.1	0.3	55	2.4	0.4	3.1E^−01	1.6E^−01	5.0E^−02
malathion	41	1.2	0.3	40	1.8	0.3	1.6E^−01	3.0E^−01	5.0E^−02
metalaxil	31	8.2	1.9	55	3.7	1	2.2E^+00	6.9E^−01	5.0E^−02
methidathion	24	2.5	0.5	33	0.5	0.1	4.2E^−01	4.9E^−02	5.0E^−02
metolachlor	72	33.3	17.4	81	36.6	14	8.2E^−01	1.2E^+00	1.0E^−01
molinate	37	17.4	5.7	50	12.2	3	1.3E^+00	2.6E^+00	5.0E^−02
oxadiazon	47	16.6	9.3	63	17.4	7	7.0E^−01	1.3E^+00	5.0E^−02
parathion	36	0.5	0.1	43	0.6	0.1	1.1E^−02	5.5E^−01	2.5E^−02
paration-CH3	42	0.0	0	51	0.4	0.1	<LOD	4.9E^−02	5.0E^−02
pendimethalin	52	0.7	0.1	66	3.2	1	1.9E^−01	3.4E^−01	5.0E^−02
pirimifos-CH3	19	1.0	0.2	29	0.3	0.1	2.3E^−01	1.0E^−02	2.5E-^02
procymidon	38	23.3	10.8	47	14.3	6	3.8E^−01	2.2E^−01	5.0E^−02
propanil	20	12.5	2.1	34	3.9	1	8.3E^−01	2.7E ^+00	2.5E^−02
simazine	71	14.2	6.4	82	11.8	4	3.2E^−01	1.6E^−01	5.0E^−02
terbuthylazine	72	51.5	34.9	83	51.0	30	6.6E^−01	7.4E^−01	1.9E^−01
terrbumeton	18	1.1	0.1	27	0.3	0.1	5.0E^−02	8.0E^−02	5.0E^−02
trifluralin	56	3.8	0.7	61	1.5	0.2	5.5E^−02	1.9E^−01	5.0E^−02
vinclozoline	33	0.6	0.1	42	0.2	0	4.8E^−02	6.0E^−02	2.5E^−02

---

Analytical techniques to determine pesticide residues in water have been developed in several laboratories belonging to a network of local environmental protection agencies (ARPA). These laboratories have participated to an intercomparison test to assess the quality and consistency of the obtained results.¹⁰ Nevertheless, the techniques still vary greatly among the laboratories and their LODs are often not comparable. The selection of Table 1 was made considering the percentage of the total number of monitoring stations in which a chemical was included in the local monitoring plans. Then, monitoring data for each of these substances can be theoretically considered as representative of the contamination of surface water systems of the entire national territory. Other substances have been included considering the percentage presence in the monitoring stations or in the analysed samples. On the contrary, the organochlorine insecticides were excluded from this selection (they were considered not relevant to the objectives of this work). Table 1 reports the total number of samples analysed for each substance and the number of positive samples (with measurable concentrations). More information can be found in the original documents.^8,9 Furthermore the 95th concentration percentiles of PPP residues in surface waters found in 2005 and 2006 are also reported

PNEC calculation

According to the Water Framework Directive (WFD) a PNEC (Predicted No Effect Concentration) should represent a real EQS (Environmental Quality Standard) that is, for a specific substance, the reference value to determine the chemical quality of surface water course. As a matter of fact a PNEC represents concentration at which a substance is expected to produce no adverse effects on aquatic ecosystems.

In this paper, PNECs were determined for all relevant pesticide residues reported in APAT⁸ and ISPRA⁹ documents. They were calculated in agreement with the procedures reported in the EU Technical Guidance Documents.⁷Table 2 shows the toxicity data (acute and chronic toxicity on algae, Daphnia and fish), and the safety factors used to derive PNEC values. The toxicity data were mainly obtained through The FOOTPRINT Pesticide Properties Database (PPDB: http://www.eu-footprint.org/it/ppdb.html). For desethyl terbuthylazine, toxicity data were gathered from Finizio.¹¹ For those substances which are included in the priority list of the WFD, the suggested EQS (Environmental Quality Standard) values were used. As a consequence of the lack of toxicity data, the two atrazine metabolites (desethyl-atrazine, desisopropyl-atrazine) were assigned the same EQS of the parent compound as a worst case approach, considering that metabolites are generally less toxic than the parent compound.

Table 2 Acute (LC₅₀), chronic (NOEL) ecotoxicological data on non target organisms (algae, Daphnia, Fish) and safety factors used for establishing PNEC values for the selected substances. For those substances which are included in the priority list of the WFD (in bold) the suggested EQS values were used.

Substance	LC50 Fish/μg L^−1	EC50Daphnia/μg L^−1	EC50 Algae/μg L^−1	NOEC Fish/μg L^−1	NOEC Daphnia/μg L^−1	NOEC Algae/μg L^−1	Application factor	PNEC or EQS/μg L^−1
a EQS of the parent compound atrazine was considered (worst case approach).
alachlor	—	—	—	—	—	—	—	1.0E ^−01
atrazine	—	—	—	—	—	—	—	1.0E ^−01
azinphos-C2H5	8.0E^+01	3.2E^+00	1.3E^+03	—	—	—	1000	3.2E^−03
azinphos-CH3	2.0E^+01	1.1E^+00	7.2E^+03	1.7E^−01	4.0E^−01	—	50	3.4E^−03
benfluralin	8.1E ^+01	>100000	>100000	1.9E^+00	1.6E^+01	—	50	3.8E^−02
chlorpyrifos	—	—	—	—	—	—	—	1.0E ^−01
chlorpyrifos-CH3	4.1E ^+02	6.0E^−01	5.7E^+02	5.0E^+00	1.0E^−02	—	50	2.0E^−04
desethyl-atrazine†	—	—	—	—	—	—		1.0E^−01
desethyl-terbuthylazine	1.1E^+06	2.8E^+06	1.4E ^+05	—	1.3E^+06	—	100	1.4E ^+03
desisopropyl-atrazine ^a	—	—	—	—	—	—	—	1.0E^−01
diazinon	3.1E ^+03	1.0E^+00	6.4E ^+03	7.0E^+02	5.6E^−01	1.0E^+04	10	5.6E^−02
dichlorvos	5.5E^+02	1.9E^−01	5.3E^+04	1.1E^+02	—	4.7E ^+03	50	3.8E^−03
dimethoate	3.0E^+04	2.0E^+03	9.0E^+04	4.0E^+02	4.0E^+01	3.2E^+04	10	4.0E^+00
diuron	—	—	—	—	—	—	—	1.0E ^−01
fenitrothion	1.3E^+03	8.6E ^+00	1.3E^+03	1.7E^+01	8.8E^−02	1.0E^+02	10	8.8E^−03
iprodione	3.7E ^+03	6.6E ^+02	1.8E ^+03	2.6E^+02	1.7E^+02	3.2E^+03	10	1.7E^+01
linuron	3.2E^+03	3.1E ^+02	1.6E^+01	1.0E^+02	1.8E ^+02	1.0E^+01	10	1.0E^+00
malathion	2.2E^+01	7.0E^−01	1.3E^+04	2.1E ^+01	6.0E^−02	—	50	1.2E^−03
metalaxil	1.0E^+05	2.8E^+04	3.3E^+04	—	—	1.0E^+04	1000	1.0E^+01
methidathion	1.0E^+01	6.4E ^+00	2.2E^+04	—	—	—	1000	6.4E^−03
metolachlor	3.9E^+03	2.4E^+04	5.7E^+04	—	7.1E^+02	—	100	7.1E^+00
molinate	1.6E^+04	1.5E^+04	5.0E^+02	3.9E^+02	3.8E ^+02	—	50	7.6E ^+00
oxadiazon	1.2E^+03	2.4E^+03	4.0E^+00	7.5E^+00	3.0E^+01	—	50	8.0E^−02
parathion	1.5E^+03	2.5E^+00	5.0E^+02	—	—	—	1000	2.5E^−03
parathion-CH3	2.7E ^+03	7.3E ^+00	3.0E^+03	—	—	—	1000	7.3E^−03
pendimethalin	1.4E ^+02	2.8E^+02	6.0E^+00	6.0E^+00	1.5E^+01	3.0E^+00	10	3.0E^−01
pirimifos-CH3	4.0E^+02	2.1E^−01	1.0E^+03	2.3E ^+01	8.0E^−02	—	50	1.6E^−03
procymidon	7.2E^+03	1.8E ^+03	2.6E^+03	4.8E ^+02	9.9E^+02	—	50	9.6E ^+00
propanil	2.3E ^+03	4.8E ^+03	5.0E^+01	—	—	—	1000	5.0E^−02
simazine	9.0E^+04	1.1E^+03	4.0E^+01	7.0E^+02	2.5E^+03	6.0E^+02		1.0E^−01
terbuthylazine	3.8E ^+03	2.1E ^+04	1.6E^+01	9.0E^+01	1.9E^+01	—	50	3.2E^−01
terbumeton	1.4E ^+04	4.0E^+04	9.0E^+00	—	—	—	1000	9.0E^−03
trifluralin	—	—	—	—	—	—	—	1.0E ^−01
vinclozoline	2.8E^+03	3.7E ^+03	1.0E^+03	—	—	—	1000	1.0E^+00

---

Risk characterization

The APAT and ISPRA documents should describe the status of water resource contamination in Italy. In both reports, the risk for surface water systems was characterised comparing the percentiles of monitored water concentrations with 0.1 μg L⁻¹ (quality standard for drinking water).

In this analysis, a different approach was used, based on the ratio between the 95th percentile of measured water concentrations and PNEC. This ratio may be lower, higher or equal to 1. The 95th percentile is the value below which 95% of the observations occur. In other words, for a specific substance, the 95% of water samples are characterized by a residue concentration below the 95th percentiles. As a consequence, a ratio [95th]/PNEC lower than 1 indicates no risk for almost all water bodies analysed. On the contrary if the [95th]/PNEC ratio is equal or higher than 1, the substance may be considered “of concern” therefore requiring further action (risk management).

Results

In Table 3, some details about the characteristics of the Italian monitoring plans on PPP residues in surface water systems (2005 and 2006) are reported. Further information can be gathered from APAT and ISPRA reports.^8,9

Table 3 Number of monitoring stations, analysed samples, analysed and detected substances in Italian surface water bodies in 2005 and 2006 (in brackets the percentage of monitoring points in which at least one PPPs residue was detected)

2005				2006
No. monitoring stations	No. analysed samples	No. analysed subst.	No. detected subst.	No. monitoring stations	No. analysed samples	No. analysed subst.	No. detected subst.
1032 (47%)	6361	255	112	1123 (57%)	6968	315	131

---

The inclusion of a specific substance in the monitoring list greatly varied depending on the geographical areas considered. In fact, in Italy, monitoring activities are organised on a regional basis; therefore, the inclusion of a substance in the monitoring list depends only on the decisions of the regional authorities responsible for the monitoring activity.

It has to be stated that the agriculture also varies a lot across the territory. In general, agriculture cannot be defined “regional”, high regional variations in crop distributions occur throughout Italy. This “regional” crop variability is strongly related to differences in climate which greatly influence the type of pest to be controlled (and the type of PPP to be used). Knowledge of the regional characteristics of the territory and of the type of crops are therefore key parameters for a correct planning of a national reliable monitoring. As stated before, the type of active substance analysed varied according to region. In 2005, 47% of the monitoring stations showed the presence of at least one contaminant and a total of 112 substances were detected.

In 2006, there was an increasing effort of the Italian Authorities to improve monitoring activities, with a total of 6968 samples collected in 1123 monitoring stations with a total of 315 substances included in the monitoring lists. As a result, an increased percentage of contaminated sites (57%) and detected substances (131) were observed. According to Table 1, the majority of the monitored substances belong to the category of herbicides. In fact, in both years alachlor, atrazine (and relative metabolites), terbuthylazine (and its metabolite desethyl-terbuthylazine), simazine, and metolachlor are present in almost all regional monitoring programs. Other herbicides such as pendimethalin and oxadiazion were also frequently monitored. Among the insecticides, just chlorpyriphos (58% in 2005 and 61% in 2006) and chlorpyriphos-CH₃ were frequently included in monitoring campaigns, since chlorpyriphos has to be included in all monitoring programs as required by the WFD. Other insecticides such as the organophosphorous compounds (diazinon, azinphos-CH₃, azinphos-C₂H₅, fenitrothion, etc) were less frequently monitored. Finally, fungicides are in general scarcely included in the regional monitoring, due to the Italian historical concern on herbicides. Table 4 reports the ratios between the percentiles of concentrations of pesticide residues, detected in 2006 in the Italian surface waters ([95th] calculated with all results of the analysis), and their relative PNEC. These ratios are compared with those reported in the ISPRA report where the characterization of risk for surface water systems was performed considering the quality standard for drinking water (0.1 μg L⁻¹). All those compounds whose EQS is equal to 0.1 μg L⁻¹ were excluded from the table. The cases where the ratios ([95th]/PNEC suggest a possible risk (risk ratio higher than 1) are highlighted in bold. Conceptual differences between the two different approach will be addressed in the discussion section. Fig. 1 shows the ratios between concentration residues [95th percentiles] and PNECs (or EQS) for the selected compounds. These ratios refer to 2005 and 2006 monitoring campaigns. For 2006 both methods used for the calculation of percentiles were considered. In this way, it was possible to make consistent comparisons with 2005 using the concentration percentiles data calculated by excluding all samples containing no residues (first method of calculation).

Fig. 1 Risk ratios between concentration residues [95th percentiles] and PNECs (or EQS) for selected PPPs residues in surface water in Italy for 2005 and 2006 (for 2006 both methods used for the calculation of percentiles are reported: * percentile with exclusion of no residue data; ** percentile with inclusion of no residue data).

Table 4 95th percentiles measured concentration versusPNEC compared with 95th percentiles measured concentration versus 0.1 μg L⁻¹ (limit for drinking water). The potential risk is highlighted in bold

Substances	[95th]/PNEC	[95th]/0.1μg L^−1	Substances	[95 ^th]/PNEC	[95th]/0.1μg L^−1
azinphos-C2H5	7.8E ^+00	2.5E^−01	metolachlor	1.4E^−02	1.0E^+00
azinphos-CH3	7.4E^+00	2.5E^−01	molinate	6.6E^−03	5.0E^−01
benfluralin	6.6E^−01	2.5E^−01	oxadiazon	6.3E^−01	5.0E^−01
chlorpyrifos-CH3	2.5E^+02	5.0E^−01	parathion	1.0E^+01	2.5E^−01
desethyl-terbuthylazine	8.0E^−04	8.0E^−01	parathion-CH3	6.8E ^+00	5.0E^−01
desisopropyl-atrazine	5.0E^−01	5.0E^−01	pendimethalin	1.7E^−01	5.0E^−01
diazinon	4.3E ^+00	2.4E^+00	pirimifos-CH3	1.6E^+01	2.5E^−01
dichlorvos	6.6E ^+00	2.5E^−01	procymidone	5.2E^−03	5.0E^−01
dimethoate	7.5E^−03	3.0E^−01	propanil	5.0E^−01	2.5E^−01
fenitrothion	2.8E^+00	2.5E^−01	simazine	5.0E^−01	5.0E^−01
iprodione	2.9E^−03	5.0E^−01	terbuthylazine	5.9E^−01	1.9E^+00
malathion	4.2E^+01	5.0E^−01	terbumeton	5,6E ^+00	5,0E^−01
metalaxyl	5.0E^−03	5.0E^−01	vinclozolin	2,5E^−02	2,5E^−01
methidathion	3.9E^+00	2.5E^−01

---

This provided a general trend of risk variability over time for surface water systems at the national level. Furthermore, for 2006, it was also possible to make a comparison between the level of risk obtained as a function of the method of percentile calculation. Both aspects will be further analysed in the discussion section.

Discussion

The WFD established a framework for EU common policy in the water field. Particularly, for surface water systems the objective is to achieve a good ecological and chemical status for all surface water bodies present in EU territories by 2015. In this context, one of the main technical tools to support the enforced actions is represented by monitoring activities. As a matter of fact, the assessment of chemical status is dependent on achieving environmental quality standards for a number of priority pollutants (33 molecules listed in Annex X of WFD, 13 of which can be classified as PPPs).

Furthermore, in Annex VIII of the WFD, an indicative list of major pollutants (for instance PPPs) that should be included in monitoring activities is reported. The inclusion should be planned and their pressure (pollutants discharged in significant quantities) in the river basin district should be considered. For these substances, Member States should set environmental quality standards according to procedures suggested by the EU Technical Guidance Documents.

Different monitoring programmes for surface water bodies in a river basin are requested by the WFD:

• Surveillance monitoring to provide an assessment of the overall surface water status within a catchment or sub-catchment in a river basin district and to monitor long term changes. Minimum frequencies of surveillance monitoring are defined in WFD Annex V (1 month for a list of priority substances, 3 months for other substances).

• Operational monitoring, established to follow a targeted approach to assess the ecological and chemical status of those water bodies that have been identified as being at risk of failing the environmental objectives. For the operational monitoring, Member States establish a monitoring frequency to ensure sufficient data to outline a reliable assessment of the chemical status and to assess the effectiveness of measures undertaken.

• In addition, investigative monitoring has to be carried out where the reason for failing the good status is unknown or to determine the magnitude and impacts of accidental pollution.

On the basis of this legislative context, a discussion on the fulfilment of the WFD requirements by the Italian monitoring activities on pesticides can be performed.

The transposition of the WFD in Italy has followed a long and complex path which culminated in the Legislative Decree 152/2006 (3 years after the deadline established in Art. 24 of the WFD) in which 8 river basin districts were identified. However, in the first part of both the analysed documents, monitoring data were aggregated at a national level and expressed as percentiles of concentrations. On the contrary, in the second part of both reports, monitoring data (not shown in this work) were reported at a regional level. However, it should be considered that, according to the WFD, monitoring programs should be aggregated considering the River Basin Districts. Consequently, also considering the Annex II of the EU Commission report of programmes for monitoring of water status (COM(2009)156),¹² it seems that in Italy the WFD monitoring programmes were not yet set up and that the reported information more realistically reflects common existing monitoring practice. This is evident if we consider that in the reports there is no differentiation between surveillance and operational monitoring programs. As previously reported, the majority of monitored pesticides belong to the class of herbicides. The question to be raised is if there is any real reason to focus almost exclusively on this category of substances, or whether the common practice influences their inclusion in the list of substances to be monitored. Herbicides are widely used in Italy; according to the National Institute of Statistics¹³ a total of 25900 tons of commercial formulations containing herbicides were sold in Italy in 2008. However, it has to be underlined that these quantities are comparable to the amounts of insecticides sold (22200 tons) and far lower than fungicides (79700 tons). Most herbicides show good solubility in water, indicating an affinity of such compounds for this compartment. On the other hand, there are several widely sprayed insecticides and fungicides showing the same water solubility properties that are not included in the list of monitoring programs. Furthermore, herbicides are in general sprayed on soil; this reduces the drift processes, often responsible for the contamination of surface water bodies. On the contrary, drift processes are present with insecticides and fungicides that are typically sprayed on crops. On this basis, no scientific reason seems to be the basis of the inclusion of almost just herbicides in the monitoring lists (decisions on substances to be monitored are taken at a regional level). The reasons might probably be found in the historical concern of the presence of herbicide residues in groundwater systems in Italy: about 20 years ago, there was the case of an extensive atrazine pollution of drinking water in Northern Italy that was of high concern among citizens and authorities.¹⁴ This fact focused the surveillance attention on the extent of the presence of herbicides residues in water resources. Consequently, the analytical techniques for these compounds were also fully developed and became a common activity for the regional analytical laboratories; herbicides were extensively included in the regional monitoring lists. Furthermore, until today, a well defined system of control of pesticide usage in Italy is still missing. The previously mentioned ISTAT data are information aggregated by categories of pesticides and refer to the sales of pesticides across the whole country. As a matter of fact, there is still a lack of information about the real use of pesticides on a local basis. This obviously limits the possibility to identify pesticides used within individual river catchments, which have to be included in the list of substances to be monitored.

On this basis it can be surely concluded that:

• It is difficult to evaluate whether currently unreported compounds are really not present in Italian rivers or if they are not monitored due to several reasons. The identification of which substances have to be monitored depends on the availability of information on their real use in the territory and on the analytical competence of the laboratory.

• It is difficult to define whether the most frequently detected compounds are really the most commonly occurring in the Italian surface water systems or if they are simply the most frequently investigated.

• In our opinion, it seems that surface water monitoring campaigns, in Italy, were not designed to determine the magnitude and frequency of contamination events and their potential impact on water ecosystems. It rather seems that they were mainly implemented to obtain a sort of instantaneous picture of how far the water quality is from the drinking water limit, even if no surface water bodies specifically assigned to drinking water abstraction are included in the monitoring survey.

The latter aspect seems to be confirmed by the approach used by ISPRA in the risk characterization. In fact, as pointed out in the Materials and Methods section of both reports, the characterisation of risk was based on the comparison between the percentile of water concentration ([95th]) and 0.1 μg L⁻¹. However, as already stated, the use of the quality standard for drinking water as a cut off criteria for surface waters can provide a misleading representation and perception of the risk. In fact, this parameter cannot be considered appropriate for the evaluation of the effects of pesticides on the aquatic communities, as each pesticide is characterised by its own inherent toxicity for different non target organisms. This is clearly evident when the procedures for setting environmental quality standards (based on the calculation of PNECs) suggested by the WFD are considered. From the results reported in Table 2, the differences in risk characterisation, depending on the approach used, are quite evident. In general, the risk for surface water seems to be higher in the case of insecticides when the PNEC approach is used. On the contrary, the criterion of the 0.1 μg L⁻¹ cut-off indicates herbicides as the major driver of risk for surface water. This choice of approach, other than strongly influencing the identification of substances to be monitored, has implications of a technical nature. In fact, the objective of the monitoring itself has to be fully defined prior of any development of sampling strategy, to obtain a proper and correct monitoring campaign. For example, it is generally assumed that agricultural spray drift is one of the main sources of contamination of surface waters. Subsequent contamination can occur via drain flow, lateral seepage, leaching, overland flow and atmospheric deposition. Only in intensively instrumented catchment-based studies could all these factors be identified and quantified with identification of the contamination leading process.

The Italian monitoring campaigns, on the contrary, are not designed to determine the magnitude and frequency of contamination events and their potential impact on aquatic ecosystems but just the level of contamination with respect to 0.1 μg L⁻¹.

Anyway, whichever cut off criteria (0.1 μg L⁻¹ or PNEC) is used, other critical issues has to be discussed to understand how well the results of the Italian monitoring campaigns of PPPs could contribute to the knowledge on the quality status of surface water systems in Italy:

• First, it has to be understood how the results of both monitoring campaigns are representative of the contamination of surface water bodies at a national level. As previously described in the first part of both reports, the monitored water concentrations of PPP residues are expressed in percentiles. Therefore, they should theoretically reflect the status of contamination of surface water bodies for the overall national territory. However, a deeper analysis of both reports show that most of the data underlying the calculations of concentration percentiles were derived from monitoring activities carried out mainly in a few regions of northern Italy (data not shown). For instance, the Piemonte Region alone accounts for about 22% of the total number of sampling points and 23% of total samples analysed in 2006. On the other hand, some regions did not provide any monitoring data.

• Second, it has be analysed how the monitoring data may vary over time. This aspect is extremely important in identifying substances which could pose a risk to aquatic organisms, in order to implement the most adequate strategies of risk mitigation. Variations of residue concentration for a specific substance, in fact, may result in deep changes in the associated risk for the aquatic ecosystems as a consequence of changes in the effective exposure during the time. In Fig. 1, a comparison between the risk ratios log([95th]/PNEC) for the two considered years is reported. From the figure it appears that in many cases there is a constant risk trend between the two years. This is particularly true for herbicides, whereas for some fungicides (cycloate and metalaxyl) and for several insecticides (dimethoate, pirimiphos-CH₃, methidathion, fenitrothion, parathion-CH₃, dichlorvos) there is a high variability in aquatic risk, even more than one order of magnitude in some cases. These differences can be probably attributed to different processes like, as previously mentioned, the different pattern of surface water contamination between different categories of PPPs. For example, drift is significantly different between insecticides and fungicides. Frequently, after these phenomena, high concentrations of PPP residues in surface waters for short periods of time (particularly in small catchments) can be detected. Moreover, herbicides generally show a higher persistence in water rather than insecticides. Therefore, for substances such as insecticides or fungicides, the adopted sampling strategy (sampling frequency and period) becomes crucial for obtaining reliable results. This aspect is extremely complex and delicate and is related to the nature of environmental monitoring. As a matter of fact, a measured concentration in the environment offers a picture of what is happening (in terms of concentration levels) during the sampling, in a context where several dynamic processes are acting contemporaneously.

• Finally, the last aspect to be considered is the choice of the method used for the calculation of the percentile concentrations which may greatly affect the evaluation of risk. This is evident from the analysis of Fig. 1, where a comparison of the risk obtained using the two different methodologies used in 2006 for percentile calculation is reported. Obviously, the inclusion in the calculation of samples with residue concentrations below the LOD significantly lowers the level of risk obtained. This approach seems to be more appropriate from a scientifically point of view. However, another critical issue has to be highlighted. In many cases, in ISPRA documents, several percentile concentrations were calculated on the basis of 1 or very few positive samples. For instance, chlorpyrifos-CH₃ residues (the PPPs showing the highest risk in Fig. 1) were detected just in 0.2% of the total analysed samples (8 positive samples from a total of 3827 analysed). Therefore, the calculated percentile depends almost exclusively on the analytical detection limit (LOD) since, as already stated, samples showing no residues were identified with a default value of concentration equal to half of the LOD.

All these aspects strongly limit any identification of potentially hazardous substances on aquatic ecosystems.

Conclusion

The WFD provisions regarding monitoring of chemical substances in surface waters are laid down and aim at establishing a coherent and comprehensive overview of water quality status within each river basin district at the EU level. However, monitoring programmes without an appropriate background knowledge (for example, use and emissions sources of chemicals in a specific territory) and programme design, suffer limitations that can reduce their usefulness as a tool to support risk management decisions. This is particularly true for PPPs: these substances, in fact, are characterized by widespread use over the territory, a high seasonality and peak pulses deriving from drift or runoff events.

A general recommendation is that a strong and very careful effort has to be devoted to the basic questions of ‘what’ ‘when’ and ‘how’, to monitor. Furthermore, in planning monitoring activities, the protection subjects (drinking water, aquatic ecosystems) should be clearly identified to organize a proper monitoring plan design. For instance, the timing and/or the frequency of sampling should be different if related to the monitoring of drinking water or to the evaluation of impacts on non target organisms. Finally, the reports of monitoring should be organized considering River Basin Districts (RBD) as required by the legislative framework (WFD). All these controversial aspects have been evidenced by the analysis of the Italian National Monitoring Plans on PPPs. Probably, these programs suffered from the delay in the implementation of the WFD in Italy and the reported information reflects existing monitoring practices. Furthermore, this work has highlighted other weakness of the reports directly related to the previously mentioned basic questions (what, when and how). In Italy there is still a lack of knowledge about the real use of pesticides at national/regional/catchment levels. The criteria used for selecting the substances to be included in the monitoring lists are probably biased by other aspects such as the historical concerns related to the case of intensive herbicide pollution of groundwater systems in Northern Italy and by the availability of proper analytical methods. In addition, from this analysis, it seems that monitoring activities in Italy are focused mainly on drinking water, even if, in agreement with the WFD, aquatic communities should be the key target for surface water monitoring programs. In this paper, we highlighted that the perception of which substances might present a risk for surface water can be completely different according to the cut-off criteria identified.

In conclusion, monitoring programmes may even be counterproductive when used uncritically. In some recent reviews of scientific papers^15,16 different authors concluded by stating that monitoring plans are often inadequate (in planning objectives, in setting experimental) and, consequently, may lead to wrong conclusions, dangerous not just due to their inability to measure a significant change in the system, but also and mainly because they can create the illusion that something useful was done, while in reality, they should be repeated with a better planning, with higher quality standards and with additional costs.

References

1. European Commission, 2000. Directive 2000/60/EC of the European Parliament and of the Council of 23 October 2000 establishing a framework for Community action in the field of water policy. OJ L 327, 22.12.2000, p. 1–73.
2. Decreto Legislativo 3 aprile 2006, n. 152 “Norme in materia ambientale”. Gazzetta Ufficiale n. 88 del 14 aprile 2006-Supplemento Ordinario n. 96.
3. European Commission, 2001. Decision No 2455/2001/EC of the European Parliament and of the Council establishing the list of priority substances in the field of water policy and amending Directive 2000/60/EC. OJ L 331, 15.12.2001, p. 1.
4. European Commission, 2009. Directive 2009/128/EC of the European Parliament and of the Council of 21 October 2009 establishing a framework for Community action to achieve the sustainable use of pesticides OJ L 309 24.11.2009, p. 71–86.
5. European Commission, 2009. Commission staff working document accompanying the report from the Commission to the European Parliament and the Council in accordance with article 18.3 of the Water Framework Directive 2000/60/EC on programmes for monitoring of water status. [COM82009)156], pp86.
6. S. Bonzini, R. Verro, S. Otto, L. Lazzaro, A. Finizio and M. Vighi, Environ. Sci. Technol., 2006, 40, 7561–7569.
7. European Commission, 2003. Technical Guidance Document (TGD) on Risk Assessment of Chemical Substances (2nd edition). European Commission, European Chemical Bureau, Joint Research Centre, EUR 20418 EN/2.
8. APAT, 2007. Piano nazionale di controllo degli effetti ambientali dei prodotti fitosanitari. Residui di prodotti fitosanitari nelle acque. Rapporto annuale dati 2005. APAT, 12/2007, pp96.
9. ISPRA, 2008. Piano nazionale di controllo degli effetti ambientali dei prodotti fitosanitari. Residui di prodotti fitosanitari nelle acque. Rapporto annuale dati 2006. ISPRA, 12/2008 ISBN 978-88-448-0373-5.
10. S. Balzamo, S. Barbizzi, M. Belli, D. Centioli, P. de Zorzi, R. Mufato, G. Sartori, G. Stocchero, 2006. Rapporto conclusivo “interconfronto APAT-IC009 Fitofarmaci in acqua- (intercomparison APAT-IC009: pesticides in water). Rapporto APAT-IC009, 2006, pp 51.
11. A. Finizio, 2008. Il D.Lgs 152/06 e l'identificazione delle aree vulnerabili per gli agrofarmaci: spunti di riflessione per lo sviluppo di linee guida condivise a livello nazionale. Report GEMMA (Gruppo Esperti Misure di Mitigazione Agroambientali Agrofarmaci), pp56. (downloadable at http://www.gemmagrofarmaci.org).
12. European Commission, 2009. Report from the Commission to the European Parliament and the Council in accordance with article 18.3 of the Water Framework Directive 2000/60/EC on programmes for monitoring of water status.[SEC(2009)415], pp 6.
13. ISTAT, 2008. La distribuzione per uso agricolo dei prodotti fitosanitari. Istituto nazionale di Statistica, Roma.
14. E. Funari, A. L. Brambilla, I. Camoni, A. Canuti, A. Cavallaro, S. Chierici, G. Cialella, G. Donati, A. Jaforte and L. Prandi, Biomed. Environ. Sci, 1988, 1, 350–5.
15. N. G. Yoccoz, J. D. Nichols and T. Boulinier, Trends Ecol. Evol., 2001, 16(8), 446–453.
16. H. J. Byron, J. R. Treweek, W. R. Sheate and S. Thompson, Journal of Environmental Planning and Management, 2000, 43, 71–97.

---