Science of the Total Environment 463–464 (2013) 432–441

Contents lists available at ScienceDirect

Science of the Total Environment

j ourna l homepage: www.e lsev ie r .com/ locate /sc i totenv

Comparison of pesticide leaching potential to groundwater under EUFOCUS and site specific conditions

Herve Labite a, Nicholas M. Holden a, Karl G. Richards b, Gaelene Kramers a,b, Alina Premrov c,Catherine E. Coxon c, Enda Cummins a,⁎a UCD School of Biosystems Engineering, College of Engineering and Architecture, University College Dublin, Belfield, Dublin 4, Irelandb Teagasc, Johnstown Castle, Environmental Research Centre, Co Wexford, Irelandc Geology Department, School of Natural Sciences, Trinity College Dublin, Dublin 2, Ireland

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• Four locations were parameterised usingthe Pesticide Leaching Model PELMO.

• The EU FOCUS scenarios Okehamptonand Hamburg were simulated and com-pared to Irish site specific scenarios.

• The use of the EU FOCUS scenarios inIrish conditions provides a risk bufferin terms of certification of pesticides.

⁎ Corresponding author. Tel.: +353 1 7167395; fax: +E-mail address: enda.cummins@ucd.ie (E. Cummins

0048-9697/$ – see front matter © 2013 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.scitotenv.2013.06.050

a b s t r a c t

a r t i c l e i n f o

Article history:Received 23 April 2013Received in revised form 11 June 2013Accepted 11 June 2013Available online 2 July 2013

Editor: D. Barcelo

Keywords:PesticidePELMOEU FOCUS groundwater scenariosLeaching potential

The EU FOCUS scenarios are a set of nine standard scenarios based on a combination of crop, soil and weatherdata used throughout Europe to evaluate the leaching potential of pesticides to groundwater. In Ireland, twopredefined EU FOCUS scenarios (Okehampton and Hamburg) appear to be the most appropriate to Irish con-ditions. However, there is concern that these scenarios may not accurately represent Irish specific conditions,especially in terms of soil and climatic weather. Therefore, the objective of this study was to parameterise anumber of site specific locations in Ireland (represented by Oakpark, Clonroche, Rathangan and Elton seriessoils) and to compare simulated leachate levels at these locations to EU FOCUS scenarios using the PELMO(Pesticide Leaching Model) simulation model. The hydrological processes were validated using observeddata for soil tension and leachate. The appropriate EU FOCUS scenarios were then simulated for the givenlocations and compared to the parameterised scenario. All scenarios were run using the same version ofPELMO, therefore eliminating any software impacts. The models were run for 26 years using appropriatemeteorological data. The results showed significant difference between the parameterised model pesticideleaching and that resulting from the EU FOCUS scenarios, the latter overestimating site pesticide leachingfrom 42 to 99%. The results indicated a significant conservatism in using EU FOCUS scenarios to determinepotential pesticide concentration in the leachate under Irish specific conditions and ensure the desiredlevel of protection against pesticide contamination of national water resources.

© 2013 Elsevier B.V. All rights reserved.

353 1 7167415.).

rights reserved.

433H. Labite et al. / Science of the Total Environment 463–464 (2013) 432–441

1. Introduction

Pesticides are chemicals used in awide rangeof areas (e.g., agriculture,forestry and urban amenity) to control pests (e.g., weeds, birds, insectsand viruses) (CEC, 2007). In crop production, pesticides may be used toimprove crop yield, which may in return result in the improvement offood availability (Cooper and Dobson, 2007). Despite their importantrole, the use of pesticides can have negative effects on non-targetorganisms,water and soil compartments,while alsopotentially impactinghuman health (Pavlis et al., 2010). The negative influence of pesticide onwater quality and biota has been reported (Andreu and Picó, 2012; Blascoand Picó, 2009). A coordinated effort at global and regional level on theuse of pesticides is needed to ensure crop protection with the view toensuring food safety and quality while protecting the environment(Carvalho, 2006).

Several ranking tools have been developed to give a quick evaluationof the potential risks associated with pesticide use on agricultural landand various techniques are available including scoring approaches, de-cision trees and risk ratio methods (Labite et al., 2011; Wustenberghset al., 2012). At the European level, Directive 91/414/EEC stipulatesthat before EU countries licence plant protection products, associatedrisks (e.g., risk to human health, including workers, soil contamination,air, surface and groundwater contamination) must be evaluated (CEC,1994). The recommended approach is based on the ratio of the predict-ed environmental concentration and the appropriate toxicological datafor air, sediment, soil, surface water and groundwater (CEC, 1994). Forgroundwater, the FOCUS (i.e., Forum for the Coordination of PesticideFate Models and their Use) groundwater workgroup was set up to de-velop a set of standard scenarios that can be used to assess the leachingpotential of pesticide to groundwater (FOCUS, 2000, 2009). In FOCUSgroundwater, nine scenarios, (based on data from Châteaudun locatedin France, Hamburg in Germany, Jokioinen in Sweden, Kremsmünsterin Austria, Okehampton in United Kingdom, Piacenza in Italy, Porto inPortugal, Sevilla in Spain and Thiva in Greece) have been developedbased on a combination of crop, soil and weather data for these sites.In addition to these scenarios, four environmental fate models, PEARL,PRZM, PELMO and MACRO were selected within the FOCUS groundwa-ter framework. In thefirst release of the FOCUS groundwatermodels, allmodels, except MACRO were based on the convection dispersionprocess (i.e., chromatographic flow) in soils and parameterised for allthe nine scenarios. MACRO, which is a preferential flow model, wasonly parameterised for the Châteaudun scenario. In a preferential flowprocess (in contrast to the chromatographic flow), it is assumed thatwater and chemical transport occurs with by-pass in macro pores, andthis can introduce a bias in the risk assessment if the process is not

Parameterisation of Irishsite specific scenarios

Validation of Site specificscenario using PELMO

Comparison of Irish site specific and EU Focus scenarios

Simulation of pesticide flow to groundwater

Parameterisation ofHamburg and Okehampton

scenarios

Fig. 1. Schematic diagram of the method used to assess the risk of pesticide leaching.

considered. To date, efforts are being made to include the process ofpreferential flow in all FOCUS environmental fate models, althoughthis is at a preliminary stage (FOCUS, 2009). Most EU countries refereither to one or more of these scenarios when assessing the leachingpotential of pesticides. Some countries (e.g. France and Sweden) haveelaborated their own scenarios to reflect the agro environmental condi-tions of their countries (FROGS, 2011; FOCUS, 2009). In Ireland, theregistration of Plant Protection Products is based on using theOkehampton and Hamburg scenarios with PELMO, as these scenariosaremost appropriate for Irish conditions, but have limitations, includingthe relevance to Ireland of soil and weather conditions used in the sce-narios (Zhang and Moody, 2004). Recently, the work of Piwowarczyk(2013), who studied pesticide adsorption parameters in soils for severallocations, generated site sorption data which can be used to model andpredict the leaching potential of pesticides to groundwater. The objec-tive of this study is to parameterise and validate leachate flow from anumber of Irish specific locations using the PELMO pesticide leachingmodelling tool and compare pesticide concentration predictions fromthe EU FOCUS scenarios applied to the same sites.

2. PELMO description

The PELMO model is a one dimensional model which simulateswater and chemical movement in the unsaturated zone within andbelow the plant root zone (Ferrari et al., 2005; Klein et al., 1997). Itsimulates soil hydrology based on a “tipping bucket approach”where water will move to the next layer if the field capacity isexceeded while solute transport is based on convection–dispersionprocess (Klein et al., 1997; FOCUS, 2000). The PELMO model hasbeen tested by the model developers (Klein, 1997; Klein et al., 1997,2000) as well as third parties (Boesten, 2004; Dubus et al., 2003;Ferrari et al., 2003) to predict the environmental concentration ofpesticides. The main processes in themodel include watermovement,chemical transport, substance degradation, sorption, volatilisation,run off, soil erosion, soil temperature, plant uptake and substanceapplication (FOCUS, 2009). Klein (1997) noticed that predictedwater flow and pesticide transport were of the same order of magni-tude. The wide range of inputs (including soil, climate and pesticideparameters) and their importance in PELMO have been evaluated byKlein et al. (2000) and Dubus et al. (2003). Based on the lessonslearned from previous versions, the 4.4.3 version of themodel was de-veloped to improve the model predictions. In the 4.4.3 version ofPELMO (which is more user friendly), the volatilisation (from soil sur-face and plants) and sorption of pesticide to soil processes have beenupgraded and the preferential flow module has been included(FOCUS, 2011). Volatilisation from the soil surface was assumed tobe temperature dependant in the 4.4.3 version in contrast to the pre-vious versions. In addition, the plant volatilisation was refined by con-sidering volatilisation from leaves and photodegradation. With regardto pesticide sorption to soil, the pH dependency and kinetic sorptionmodules were made available if necessary to improve the sorptionprocess description. Recently, the concept of preferential flow hasbeen included in the PELMOmodel using three parameters: thresholddaily rainfall, fraction of excess rainfall, andmacro pore depth (FOCUS,2009). The process of macropore flow in PELMO is activated when thethreshold rainfall value is exceeded where a certain proportion ofrainfall (i.e., fraction of the excess rainfall) will be routed into macropores at a fixed depth. For this study, the simulations were performedwith the latest version of PELMO (i.e., version 4.4.3).

3. Modelling strategies

The overall modelling approach used in this study for assessingthe risk of pesticide leaching is detailed in Fig. 1. A wide range ofinputs was required (soil data, meteorological conditions, pesticideproperties and management practices) and were grouped into three

434 H. Labite et al. / Science of the Total Environment 463–464 (2013) 432–441

categories: site, climate and pesticide input parameters (FOCUS,2000, 2009). The overall strategy adopted was to parameterise thesite specific scenarios in PELMO, calibrate the hydrology for the se-lected sites followed by pesticide simulation for the sites. The simula-tion results were then compared to the EU FOCUS predictions (vizHamburg and Okehampton scenarios). The hydrology was calibratedby comparing the predicted quantity of percolation water to observedwater percolated and soil tension data. Firstly, the observed waterpercolation data were obtained by regular measurements fromundisturbed lysimeters containing four Irish grassland soils (i.e., OakPark, Clonroche, Rathangan and Elton soil series). The experimentswere carried out at Teagasc Johnstown Castle Environmental Re-search Centre from 3/8/2006 to 24/3/2008 and are detailed inKramers et al. (2012). Secondly, the soil tension measurementswere conducted at Oak Park Research Centre, Co. Carlow, at welldrained sites under cereal production. The soil tension measurementswere recorded at 0.3 and 0.6 m depth and the details of the experi-mental conditions were described in Premrov et al. (2010).

3.1. Site description

Based on diversity of soil, location and data availability, fourlocations were selected for simulation: Oak Park (County Carlow),Clonroche (County Wexford), Rathangan (County Wexford) and Elton(County Limerick). The soils selected have contrasting texturesthroughout their horizons with their key characteristics presented inTable 1. The depth of the soils is variable and a previous study highlight-ed differences in preferential flow as a result of earthworm activity inthese soils (Kramers et al., 2012). Preferential flow through deep earth-wormburrowswas observed in the Clonroche, Elton and in particular inthe Rathangan soil (burrows down to 1 m depth). Preferential flow inthe Oakpark soil is likely to be a combination of shallow earthwormburrows and unstable infiltration in this coarse-grained soil (Kramerset al., 2012). The predominant land uses for the Oakpark and Clonrochesoils is tillage and grassland for the Rathangan and Elton soils.

3.2. Simulation with PELMO

For each scenario, the model was parameterised using site specificsoil and climatic data while the pesticide parameters used were com-mon for all scenarios (EU FOCUS scenarios and site specific scenarios).

Table 1Physical properties of the soils selected (continued next page).

Scenarios Thicknessa Textureb Clayb (%) Siltb (%) Sandb (%) pH

Oak Park 20 Sandy loam 11 23 67 6.625 Sandy loam 12 20 68 7.8

Clonroche 20 Loam 17 39 44 6.525 Loam 25 37 38 7.345 Loam 14 41 45 7.1

Rathangan 20 Loam 19 37 44 630 Sandy clay loam 24 28 48 6.320 Clay loam 28 30 42 6.530 Clay loam 29 39 32 6.7

Elton 20 Loam 17 35 48 6.230 Silt loam 14 50 38 6.840 Loam 23 30 47 6.9

Sources:a Adjusted (to fit the model file) based on Kramers et al., 2012 and Brenan et al., 2010.b From Kramers et al., 2012 and Brenan et al., 2010.c PEDOSPHERE, 2011.d Organic matter divide by 1.72 as suggested by Schulte and Hopkins, 1996.e Kramers et al., 2012.f Brenan et al., 2010.

3.2.1. Soil and climatic input data

3.2.1.1. EU FOCUS scenarios. The Okehampton and Hamburg scenarioswere parameterised in PELMO using FOCUS data. A number of climateparameters are required as inputs in PELMO: daily rainfall, evapotrans-piration, daily temperature at 2 pm, dailymean temperature, daily tem-perature difference and annual average relative humidity. A total of26 years weather data for Okehampton and Hamburg was set up inthe FOCUS scenarios and used for the Oakhampton and Hamburgscenarios (FOCUS, 2011).

3.2.1.2. Site-specific scenarios. Similarly to the FOCUS Okehampton andHamburg scenarios, daily rainfall, evapotranspiration, daily tempera-ture at 2 pm, daily mean temperature, daily temperature differenceand annual average relative humidity for a total of 26 years weatherdata of Kilkenny and Oak Park weather stations were used for theOak Park scenario. Rosslare and Johnstown weather stations datawere used to model Clonroche and Rathangan scenarios and for theElton scenario, Shannon Airport weather station data were used(Met Éireann, 2013).

The model was parameterised using the physical soil properties(i.e., pH, bulk density, organic carbon, sand, silt and clay content) ofOak Park, Clonroche, Rathangan and Elton soils (Table 1). Soil fieldcapacity can be calculated from the internal pedotransfer function ofPELMO based on the sand and clay content (Klein et al., 1997). Thevalue of 30% suggested by Kramers (2009) was used for initial soilwater content for all the sites. In the FOCUS groundwater framework,the results of pesticide leaching to groundwater are reported at a min-imum depth of 1 m. Where soil properties were not available for deeplayers, values of the layer above were repeated, consistent with stan-dard approaches (FOCUS, 2000).

The simulations in PELMO require crop input parameters and thisincludes emergence, maturity, senescence and harvest date, and rootdepth. Grassland was selected due to its importance in Ireland as itaccounted for more than 90% of the total area and received morethan 82% of the weight of active substances applied (DAFF, 2003);spring cereals were chosen to model a tillage scenario. Due to limitedstudies of Irish crop specific parameters (e.g., root depth, crop inter-ception and fate of pesticide on plant surface), the Okehampton cropparameters set up in FOCUS were selected to model Irish sites and adefault value of 0.5 was used to model plant uptake (FOCUS, 2000).In addition to soil properties, the scenario file in PELMO requires a

b Bulk densc (kg/L) Org Cd (%) Structuree Drainagef

1.49 2.9 Single grain Very good1.49 1.9 Fine granular1.29 4.9 Fine granular Good1.62 2.38 Coarse granular1.4 1.76 Fine angular blocky1.13 3.37 Fine subangular blocky Poor1.57 1.66 Moderate subangular blocky1.66 0.68 Coarse subangular blocky1.83 0.67 Very Coarse subangular blocky1.08 3.98 Fine subangular blocky Good1.52 1.98 Moderate subangular blocky1.49 1.13 Coarse angular blocky

Table2

Pesticidesorption

,deg

rada

tion

andvo

latilis

ationprop

erties

used

during

thesimulations

.

Pesticides

(IUPA

Cna

mes)

Koc

inL/kg

(Freun

dlichex

pone

nt:1/n)

aDT 5

0in

days

bHen

ry'sco

nstant

inPa

m3/m

olb

Vap

ourpressu

rein

Pab

Soilph

otolysis

rate

in1/da

y

Oak

Park

Clon

roch

eRa

than

gan

Elton

Grass

Tilla

geGrass

Tilla

geGrass

Tilla

geGrass

Tilla

ge

MCP

A(4

-chloro-o-tolyox

yaceticacid)

50 (0.91)

42 (0.91)

108

(0.90)

–49 (0

.95)

–52 (0

.97)

–24

5.50

×10

−5

(at25

°C)

0.00

04(at25

°C)

0.02

7c

Mecop

rop-P[(R)

-2-(4-ch

loro-o-tolylox

y)-propion

icacid]

44 (0.96)

30 (0.98)

––

––

––

85.70

×10

−05

(at25

°C)

0.00

023

(at25

°C)

0.02

d

Chlorothalon

il(tetrach

loroisop

htha

lonitrile

)97

8(0

.74)

–12

79(0

.83)

–17

22(0

.88)

–23

63(0

.84)

2149

(0.93)

222.50

×10

−02

(at25

°C)

0.07

6(at25

°C)

Not

cons

idered

e

(unlikelyto

occu

r)

(–)Not

available.

Sources:

aPiwow

arczyk

,201

3.b

FOOTP

RINTda

taba

se,2

006.

cEu

rope

anCo

mmission

,200

8.d

Europe

anCo

mmission

,200

3.e

Europe

anCo

mmission

,200

6.

435H. Labite et al. / Science of the Total Environment 463–464 (2013) 432–441

degradation factor and pesticide dispersion properties to characterisepesticide fate in the environment. The former is included in PELMO totake into account the influence of microbial activity in combinationwith soil type on pesticide degradation. In the absence of Irish sitespecific degradation parameters and due to the limited data availableto account for the depth dependency factor, 1, 0.5, 0.3, and 0.3 wereused for all soils and depths. In PELMO, a realistic description of the pes-ticide dispersion process in soil can be simulated with dispersion depthvalues varying from 2.5 to 5 cm (FOCUS, 2009). For the dispersiondepth, the default values of the thickness of soil compartments and dis-persion length of 2.5 and 5 cm (model default) were used by assumingconstant dispersion in the entire soil profile. Free drainagewas assumedfor all scenarios and additional default values for the simulations aredetailed in the Supplementary material.

After initialisation, site input parameter sensitivity was analysedby changing each by ±10% of their initial values while keeping allother parameters constant. In addition, the thickness of soil layers(i.e., calculation unit) was set at 5 cm, recognising its importance ininfluencing pesticide dispersion (FOCUS, 2000, 2009). Moreover,preferential flow in PELMO was simulated due to its occurrence inthe unsaturated zone and this can potentially increase dramaticallythe risk of pesticide leaching to groundwater, as highlighted in severalstudies (Kördel and Klein, 2006; Scorza Júnior et al., 2007; Vancloosteret al., 2003). The only published study on the use of preferential flowin PELMO was Jarvis et al. (2003a) who conducted preliminary tests ofthe preferential process in PELMO for two clay soils Lanna and Andelst.To model the preferential flow in these soils by using PELMO, the fol-lowing inputs were specified: macropore depth, daily rainfall, and frac-tion of the excess were set to 70 cm, 5 mm and 0.5, respectively forLanna sites and to 85 cm, 10 mm and 0.25, respectively for the Andelstsite. In this study, a default value of 50 cm, 2 mm and 0.3 were used forthe macropore depth, threshold daily rainfall and fraction of the excessrainfall, respectively to parameterise macropore flow in PELMO.

3.2.2. Pesticide selection and input parametersThree commonly used pesticides MCPA, Mecoprop-P and Chloro-

thalonil used in the Irish agricultural sector were selected and due tothe availability of adsorption data for those chemicals (Piwowarczyk,2013). Out of 85 pesticides surveyed at national level, MCPA,Mecoprop-P and Chlorothalonil represented 40.87, 13.74 and 1.27%,respectively, of the total active substances used for grassland treatmentbased on the weight (DAFF, 2003). In addition, the work of Labite andCummins (2012) highlighted the first two compounds (i.e., MCPA andMecoprop-P) as potential chemicals of human health concern. The sorp-tion, degradation and volatilisation parameters used in the model aredescribed in Tables 2 and 3 and in the Supplementary material. Thepesticides were assumed to be applied twice on grassland (April 15thand July 15th) and once for tillage soils (May 1st) at the applicationrate described in Table 3 for every year over a period of 26 years. Theseapplication doses correspond to the recommended values by the Pesti-cide Control Service for Grass and spring cereals (Personal communica-tion, Irish Pesticide Control Service 2012). To allow comparison of eachpair of site specific scenarios and the Okehampton and Hamburg scenar-ios, the crop, pesticide parameters, management and application rateswere kept the same for each scenario so that the differences were dueto the soil and weather parameters only.

3.2.3. Model efficiencyModel efficiency (EF) was assessed (Eq. (1)) for PELMO with site

specific parameterisation and the same indicator used to assess sensi-tivity to input parameter choice: where Oi is the observed percolate(in mm); Ō the mean observed value (in mm); Pi the predictedpercolate (in mm). The maximum EF value of 1 indicates that thepredicted and observed values are equal and the model is perfectand a value less than −1 is regarded as unacceptably poor, i.e., themodel prediction should not be used for predictive purposes

Table 3Pesticide application dose.

Pesticides Multipleapplication(kg/ha)

Maxapplication(kg/ha)

Averageapplication(kg/ha)

Crop

MCPAa 1.65 (×2) 2.1 1.289 Grass, spring cerealsMecoprop-Pa 1.4 (×2) 1.98 0.840 Grass, spring cerealsChorothalonilb – – 1.2 –

(–) Not available.Sources:

a Personal communication, Irish Pesticide Control Service (2012).b Pesticide Survey (DAFF, 2003, 2004).

436 H. Labite et al. / Science of the Total Environment 463–464 (2013) 432–441

(Vanclooster et al., 2003; Loague and Green, 1991; Walker et al.,1995).

EF ¼ΣN

1Oi−O� �2−Σ

N

1Pi−Oið Þ2

!

ΣN

1Oi−O� �2 : ð1Þ

4. Results and discussion

4.1. Model efficiency, sensitivity and calibration

The model efficiency, sensitivity and calibration results for Oak Parkwere presented in Table 4. After parameterisation of the PELMOmodel,an EF value of−1.96, indicating unacceptable prediction was achieved(Table 4). Following sensitivity analysis the best EF was obtained by in-creasing the soil compartment depth from 2.5 to 5 cm and then thechange by ±10% of the root depth initial value by adjusting rootdepth. Each soil layer was therefore adjusted to 5 cm to achieve thebest prediction of leachate, which was, also consistent with the workof Akbar and Akbar (2012) who found that soil horizon (and layer)thickness was the most important input parameter.

Following the calibration, PELMO, was validated using lysimeterdata conducted from Johnstown Castle (Kramers et al., 2012) and fieldobservations Oak Park (Premrov et al., 2010), respectively. The predict-ed and observed cumulative daily percolate for Oak Park, Clonroche,

Table 4Model efficiency (EF) for the selected parameters.

Phases Description Value used EF

Parameterisation Initial water content in % 30 −1.96Mid season 1Late season 1Root depth in cm 45Evapotranspiration depth in cm 15Dispersion length in cm 5Thickness of layers in cm 2.5

Sensitivity analysis Initial water content + 10% 0.33 −1.96Initial water content − 10% 0.27 −1.96Mid season + 10% 1.1 −1.96Mid season − 10% 0.9 −1.96Late season + 10% 1.1 −1.96Late season − 10% 0.9 −1.96Root depth + 10% 49.5 0.96Root depth − 10% 40.5 0.96Evapotranspiration depth + 10% 16.5 −1.96Evapotranspiration depth − 10% 13.5 −1.96Dispersion length + 10% 5.5 −1.96Dispersion length − 10% 4.5 −1.96Thickness of layers(modification from 2.5 to 5 cm)

5 −0.39

Inclusion of macropore −2.01

Rathangan and Elton are presented in Fig. 2. PELMOwas able to predictthe total amount of percolation water for the sites although the modelslightly underestimatedwater percolated for Rathangan and Elton soils.

The lowest level of water percolation was noticed in Rathangan soilfor both, i.e., the experimental observations and the predictions. One ofthe reasons may be due to its high clay content (ranging 19 to 29%),especially in its deeper horizons. According to Brown et al. (1995) andCarter (2000), clay soils contain less coarse pores (in contrast to sandysoil) which are responsible for the slowwatermovement. The observedlow percolationmight also due to the presence of dead end pores at thesite as a low recovery of tracer was observed on this soil type (Kramerset al., 2012). The percolation simulated for all soils showed a plateau in2007 for all the sites and the probable explanation could be due to thedry conditions which occurred in the year 2007with the annual rainfallof 754 mm for Rathangan and Clonroche, 844 mm for Oak Park and forElton 922 mm (with the average annual long term rainfall of 912, 842and 982 mm, respectively). The mean annual temperature in 2007 forRathangan and Clonroche was 12.64 °C, for Oak Park 10.5 °C and forElton 11.1 °C (with the average annual long term temperature of 10.9,9.6 and 10.6 °C, respectively).

The results (Fig. 3) showed an inverse relationship between thepredicted percolate and observed soil tension at 0.3 and 0.6 mdepth. The peak values of the model predictions highlighted the po-tential occurrence of the leaching events. This confirms the model'sability to capture hydrological processes.

4.2. Pesticide simulation: EU FOCUS versus Irish site specific scenarios

The 26 year simulation representing the site specific scenarios andFOCUS Okehampton and Hamburg scenarios, are presented in Table 5.The results showed that more chemical was predicted to leach withthe FOCUS Okehampton and Hamburg scenarios simulations com-pared to the site specific scenarios (Table 5). The simulated leachingpotential of pesticide using the Okehampton and Hamburg scenariosdiffered to that compared to the simulated Irish conditions with a dif-ference ranging from 42% (Rathangan grass versus Hamburg grassscenarios) to 99.6% (Clonroche grass versus Okehampton grass andOak Park tillage versus Okehampton tillage). The use of standard sce-narios may not appropriately allow the evaluation of the leaching po-tential of pesticides to groundwater as the environmental conditions(i.e., rainfall, temperature and soil conditions) are highly variable atlocal or regional level (van Alphen and Stoorvogel, 2002). In Swedenfor example, the national registration authorities of plant protectionproducts identified the most vulnerable areas based on the combina-tion of Swedish hydro-geological conditions, weather data and majorcrop types and use of pesticides in Swedish agriculture. A comparisonof the Swedish scenarios to the EU FOCUS Scenario Châteaudun(which is the only EU focus scenarios applied to the Swedish situa-tion) revealed that the FOCUS Châteaudun scenario underestimatedthe leaching potential of pesticides to groundwater for all the Swed-ish scenarios (Jarvis et al., 2003b; KEMI, 2010). This highlights the im-portance of comparing the EU FOCUS groundwater scenarios to sitespecific conditions, and several European countries (e.g., France, Swe-den, Czech Republic and Sweden) have defined their own scenarioswhich allow the protection of vulnerable areas (FOCUS, 2009). How-ever to date, several European countries use one or more of thepredefined EU FOCUS groundwater scenarios to assess the leachingpotential of pesticides to groundwater as their national procedurefor pesticide registration (FOCUS, 2009).

In this study, the difference of the Okehampton and Hamburg sce-narios to Irish scenarios may be explained due to the soil and climaticconditions. A comparison of organic carbon, clay, silt and sand con-tent and climatic conditions at the different sites is shown in Figs. 4and 5. The evaluation of the soils and climatic conditions revealedheterogeneous composition of organic carbon, clay, silt and sand con-tent throughout the profile of all scenarios. The Okehampton and

Predicted percolate.

0

100

200

300

400

500

600

700

800

900

08/06 12/06 05/07 09/07 02/08

cum

ula

tive

per

cola

te in

mm

Date (month/year)

0

100

200

300

400

500

600

700

800

900

08/06 12/06 05/07 09/07 02/08

Cu

mu

lato

ve p

erco

late

in m

m

Date (month/year)

0

100

200

300

400

500

600

700

800

900

08/06 12/06 05/07 09/07 02/08

Cu

mu

lati

ve P

erco

late

in m

m

Date (month/year)

0

100

200

300

400

500

600

700

800

900

08/06 12/06 03/07 07/07 11/07

Cu

mu

lati

ve p

erco

late

(m

m)

Date (month/year)

a c

b d

Observed percolate

Fig. 2. Percolate validation based on the lysimeter experiments for (a) Oak Park, (b) Clonroche, (c) Rathangan and (d) Elton.

Predicted percolate. Measured soil tension

0

50

100

150

200

250

300

350

400

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

01/07 05/07 09/07 02/08 06/08 10/08 03/09 So

il te

nsi

on

in n

egat

ive

hec

toP

asca

l (-h

Pa)

Per

cola

te in

cm

Month/ Year

at 0.6m. Average soil tension (at 0.3 and 0.6 m depth).

Measured soil tension at 0.3 m.

Fig. 3. Soil tension validation for Oak Park. [Soil tension data source: Premrov et al.,2010; Used with permission.]

437H. Labite et al. / Science of the Total Environment 463–464 (2013) 432–441

Hamburg scenarios were characterised by having less organic carbonthan the Irish scenarios, except in one occasion where the amount oforganic carbon in the B horizon of the Hamburg scenario is higherthan the one of Rathangan (Fig. 4). Among the Irish scenarios,Rathangan exhibited the highest pesticide leaching potential andClonroche the lowest based on the highest and average concentra-tion. This difference in terms of leaching potential could be causedby the organic carbon content which was relatively low irrespectiveof the horizons for Rathangan compared to Clonroche. Soil organiccarbon acts as an absorbent of a pesticide, therefore reducing thepotential to freely leach to the groundwater (Wauchope et al.,2002). In addition, another difference noticed was that the Rathangansoil texture was not homogenous and had a high clay content whichincreased with depth for all horizons. According to Flury (1994) andBrown et al. (1999), chemical transport occurs as a preferential flowin soil with heterogeneous texture, like the one of Rathangan. Thisis consistent with the work of Kramers et al. (2012) who noticedthat preferential flow is more likely to occur at this site. However,the simulation of preferential flow in PELMO is recent and there arelimited studies to verify the ability of the model to simulate thistype of chemical transport. Jarvis et al. (2003a) conducted prelimi-nary tests of the preferential process in PELMO for two clay soils.The authors showed that the results of macropore flow in PELMOwere promising as they were in line with the one of MACRO but re-quired further investigations. The pesticide leaching potential of theElton scenario is higher than the one of Oak Park scenario. Elton com-pared to the Oak Park soil, is deep with low organic carbon and highclay content in the lower horizons. An analysis of the inputs to theRathangan scenario (i.e., most vulnerable site) are presented in detailin Fig. 6. This highlights the importance of both pesticide and siteinputs on model predictions.

The long term annual rainfall and average annual average temper-ature of the EU focus scenarios Okehampton and Hamburg fluctuatedbetween 601 and 1401 mm and 5 and 12 °C, respectively which werecomparable to the Irish climatic conditions (Fig. 5a, b). However,effective rainfall (i.e., rainfall minus ET) is much lower for the Ham-burg scenario compared to all the Irish sites (Fig. 5c), while the effec-tive rainfall for the Okehampton scenario is applicable to just somesites (Elton and Oak Park). This highlights the need for caution in

Table 5Comparison of FOCUS scenarios to Irish site specific scenarios.

Scenarios Crop Pesticide Highest concentration(HC) in μg/L

Average concentration(AC) in μg/L

Solute outflow(SO) in kg/ha

Year breachthe DWSa

% Above sitespecific

HC AC SO

Oak Park Grass MCPA 6.34E-01 3.15E-02 1.85E-02 4Tillage MCPA 1.16E-01 1.06E-03 5.84E-04 17

Okehampton Grass MCPA 1.91E + 01 5.11E + 00 9.45E-01 2 96.7 99.4 98Tillage MCPA 1.83E + 00 2.64E-01 5.00E-02 3 93.7 99.6 98.8

Hamburg Grass MCPA 2.28E + 01 8.79E-01 5.00E-01 2 97.2 96.4 96.3Tillage MCPA 2.04E + 00 4.11E-02 2.30E-02 4 94.3 97.4 97.5

Clonroche Grass MCPA 2.33E-02 6.76E-04 6.08E-04 NoneOkehampton Grass MCPA 8.32E-01 1.67E-01 3.12E-02 3 97.2 99.6 98.1Hamburg Grass MCPA 9.66E-01 2.09E-02 1.25E-02 4 97.6 96.8 95.2Rathangan Grass MCPA 1.18E + 01 5.60E-01 4.26E-01 2Okehampton Grass MCPA 2.48E + 01 7.23E + 00 1.35E + 00 1 52.4 92.3 68.5Hamburg Grass MCPA 2.94E + 01 1.32E + 00 7.38E-01 1 59.9 57.5 42.2Elton Grass MCPA 4.94E + 00 2.98E-01 1.74E-01 1Okehampton Grass MCPA 2.41E + 01 7.21E + 00 1.35E + 00 1 79.5 95.9 87.1Hamburg Grass MCPA 2.86E + 01 1.31E + 00 7.34E-01 1 82.7 77.3 76.3Oak Park Grass Mecoprop-P 8.20E-02 7.29E-04 4.52E-04 None

Tillage Mecoprop-P 3.83E-01 3.70E-03 2.10E-03 10Okehampton Grass Mecoprop-P 5.99E-01 4.29E-02 1.29E-02 3 86.3 98.3 96.5Hamburg Tillage Mecoprop-P 1.33E + 00 6.68E-02 1.93E-02 3 71.2 94.5 89.1

Grass Mecoprop-P 6.29E-01 1.45E-02 8.65E-03 5 87.0 95.0 94.8Tillage Mecoprop-P 2.46E + 00 1.32E-02 8.28E-03 5 84.4 71.9 74.7

All scenariosb Grass Chlorothalonil 0 0 0 None 0 0 0

a Number of years (over a 26 years period) where pesticide concentration is predict the first time to breach the drinking water standard DWS (i.e., simulated pesticide concen-tration is equal or beyond 0.1 μg/L).

b Oak Park, Clonroche, Rathangan, Elton, Hamburg and Okehampton scenarios exhibited no leaching potential with Chlorothalonil.

438 H. Labite et al. / Science of the Total Environment 463–464 (2013) 432–441

applying these scenarios across different geographical locations.However, the leaching pattern of the Okehampton and Hamburg sce-narios compared to Irish site specific scenarios noticed in this study,

0 2 4 6

A

A-B

B

c

Organic carbon content (%)

Hor

izon

0 20 40 60

A

A-B

B

c

Silt content (%)

Hor

izon

Fig. 4. Organic carbon, clay silt and san

should be viewed as the resulting outcome of a number of interactionprocesses, combining pesticide properties and management, soil andweather conditions (CARTER, 2000).

0 10 20 30 40

A

A-B

B

c

Clay content (%)

Hor

izon

0 50 100 150

A

A-B

B

c

Sand content (%)

Hor

izon

d content in A, B and C horizons.

0

200

400

600

800

1000

1200

1400

1600

Pre

cip

itat

ion

(in

mm

/yea

r)

Scenarios

0

2

4

6

8

10

12

14

An

nu

al v

tem

per

atu

re (

in °

C)

Scenarios

0

100

200

300

400

500

600

700

800

Eff

ecti

ve r

ain

fall

in m

m/y

ear

Scenarios

Fig. 5. Analysis of climate data (yearly minimum and maximum values shown in bars). (a) precipitation, (b) annual average temperature and (c) effective rainfall.

439H. Labite et al. / Science of the Total Environment 463–464 (2013) 432–441

Three pesticides, MCPA, Mecoprop-P and Chlorothalonil were sim-ulated based on the availability of pesticide site specific data and theresults showed that all scenarios (EU FOCUS and Irish scenarios) hadno leaching of Chlorothalonil (Table 5). Chlorothalonil showed a highsorption potential in the four Irish sites with the soil sorption coeffi-cient ranging from 978 to 2363 L/kg and these results were in therange of the average values (of 300–7000 L/kg) published in theFOOTPRINT database (FOOTPRINT, 2006); this indicated that thischemical is not available in the soil solution as it is strongly sorbedby the soil particles. In contrast to Chlorothalonil, the results ofMCPA and Mecoprop-P indicated that these pesticides can leachwith the possibility of exceeding the drinking water standard underthe management practice described. The scenarios modelled in thisstudy assumed a maximum application rate (and hence may be pes-simistic, anecdotal evidence suggests most farmers may even onlyapply 70% of the recommended rate as a cost saving measure)based on the current maximum guideline of the Pesticide Control Ser-vice for the proposed use of MCPA and Mecoprop-P, consisting of twoapplication doses to grassland (2 × 1.65 kg/ha in grassland and2 × 1.4 kg/ha) and the maximum individual dose to spring cereals(2.1 kg/ha and 1.98 kg/ha). The simulated results highlight the needfor customised applications based on individual pesticide propertiesand site conditions. In this study, MCPA leached more than Mecoprop-Pin 4 out of 6 scenarios, based on the highest concentration. This findingis interesting as the sorption values of Mecoprop-P is lower than theone of MCPA and differed by 12.76 and 27.33% for grassland and tillage,respectively (Table 2). The difference between Mecoprop-P and MCPA islikely to be due to the fact that the degradation rate of Mecoprop-P ishigher and may reduce its leaching potential compared to MCPA. Thefindings of this study were based on site specific sorption data while the

degradation values where obtained from literature and therefore, fielddegradation studies in Irish conditions may help to reduce the uncer-tainties of the model predictions (Klein et al., 2000; Dubus et al., 2003).

4.3. Implications

According to the Council Directive 91/414/EEC, the determination ofthe predicted environmental concentration is a key step of the risk as-sessment for groundwater contamination by pesticides (CEC, 1994). Inthis study, the parameterisation and simulation of the EU Okehamptonand Hamburg scenarios and Irish site specific scenarios showed thatmore chemical was predicted to leach with the former scenarios(Table 5). This highlights the conservative nature of the EU FOCUSscenarios compared to Irish site specific scenarios and from an environ-mental point of view, will assist in reducing the risk of groundwatercontamination.

5. Conclusion

A number of site specific scenarios were parameterised and vali-dated with percolation and soil tension in this paper. The aim ofthis study was to compare simulate leachate levels at site specific sce-narios to the EU FOCUS scenarios applied to the same sites. Theparameterisation and simulation of the EU Okehampton and Hamburgscenarios and Irish site specific scenarios showed that the formeroverestimated the leaching potential. From this modelling exercise, itcan be concluded that the FOCUS scenarios are more conservativethan the site-specific scenarios, and therefore providing a risk bufferin terms of certification of products and hence might be regarded aspositive from an environmental point of view. This indicates that the

Fig. 6. Sensitivity analysis (±10% from baseline model) of the model inputs for Rathangan scenario (i.e., most vulnerable scenario).

440 H. Labite et al. / Science of the Total Environment 463–464 (2013) 432–441

use of EU FOCUS scenarios under Irish specific conditions ensures theprotection of Irish groundwater from pesticides. Among the four sitesevaluated, Clonroche was identified as the least vulnerable area whileRathangan had the highest vulnerability to pesticide loss. Additionaldata on pesticide transport and site specific field studies (in particular,pesticide soil degradation studies) will provide greater insight into therisk of groundwater contamination by pesticides in Ireland.

Acknowledgement

This work was funded by the Irish Department of Agriculture, Fisher-ies and Food (DAFF) under the Research Stimulus Fund. Climatic datawaspurchased from the Irish National Meteorological Service, Met Éireann.

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.scitotenv.2013.06.050.

References

Akbar TA, Akbar RA. Pesticide health risk mapping and sensitivity analysis of parame-ters in groundwater vulnerability assessment. CLEAN-Soil, Air, Water 2012.http://dx.doi.org/10.1002/clen.201200232. [Accepted article].

Andreu V, Picó Y. Determination of currently used pesticides in biota. Anal BioanalChem 2012;404:2659–81.

Blasco C, Picó Y. Prospects for combining chemical and biological methods for integratedenvironmental assessment. Trends Anal Chem 2009;28(6). http://dx.doi.org/10.1016/j.trac.2009.04.01.

Boesten JJTI. Influence of dispersion length on leaching calculated with PEARL, PELMOand PRZM for FOCUS groundwater scenarios. Pest Manag Sci 2004;60:971–80.

Brenan FP, O'Flaherty V, Kramers G, Grant J, Richards KG. Long-termpersistence and leachingof Escherichia coli in temperate maritime soils. Appl Environ Microbiol 2010;6(5):1449–55.

Brown CD, Carter AD, Hollis JM. Soils and pesticide mobility. In: Roberts TR, Kearney P,editors. Environmental behaviour of agrochemicals, Progress in Pesticide Biochemistryand Toxicology. Chichester: John Wiley and Sons; 1995. p. 131–84.

Brown CD, Marshall VL, Deas A, Carter AD, Arnold D, Jones RL. Investigation into the effect oftillage on solute movement through a heavy clay soil. II. Interpretation using aradio-scanning technique, dye tracing and modelling. Soil Use Manag 1999;15:94–100.

Carter AD. Herbicide movement in soils: principles, pathways and processes. Weed Res2000;40(1):113–22.

Carvalho FP. Agriculture, pesticides, food security and food safety. Environ Sci Policy2006;9(7–8):685–92.

CEC. EU Policy for a sustainable use of pesticides. The story behind the strategy. EuropeanCommission; 20071–28 [Online retrieved at: http://ec.europa.eu. (Accessed on 15/10/2012)].

CEC (Commission of the European Communities). Council Directive 94/43/EC.Establishing Annex VI to Directive 91/414/EEC.O J L 227 1/10/1994; 19941–31.

Cooper J, Dobson H. The benefits of pesticides to mankind and the environment. CropProt 2007;26(9):1337–48.

DAFF. Pesticide usage survey. Grassland and fodder crops. Online retrieved at:http://www.pcs.agriculture.gov.ie, 2003. [Accessed on 15/10/2012].

DAFF. Pesticide usage survey. Arable crops. Online retrieved at:http://www.pcs.agriculture.gov.ie, 2004. [Accessed date: 15/10/2012].

Dubus IG, Brown CD, Beulke S. Sensitivity analyses for four pesticide leaching models.Pest Manag Sci 2003;59:962–82.

European Commission. Review report for the active substance mecoprop-P. DirectorateE — Food Safety: plant health, animal health and welfare, international questions.E1 — Plant health; 2003 [Online retrieved at: http://ec.europa.eu/food/plant/protection/evaluation/existactive/list1-49_en.pdf. (Accessed date: 26/05/13)].

European Commission. Review report for the active substance chlorothalonil. Produc-tion and distribution chain Unit D.3 — Chemicals, contaminants and pesticides. Onlineretrieved at:http://ec.europa.eu/food/plant/protection/evaluation/existactive/list_chlorothalonil.pdf, 2006. [Accessed date: 26/05/13].

European Commission. Review report for the active substance MCPA. Directorate D —

Food Safety: Production and distribution chain Unit D.3 — Chemicals, contami-nants and pesticides; 2008 [Online retrieved at http://ec.europa.eu/food/plant/protection/evaluation/existactive/list_mcpa.pdf. (Accessed date: 26/05/13)].

Ferrari F, Trevisan M, Capri E. Predicting and measuring environmental concentrationof pesticides in air after soil application. J Environ Qual 2003;32:1623–33.

Ferrari F, Klein M, Capri E, Capri E, Trevisan M. Prediction of pesticide volatilizationwith PELMO 3.31. Chemosphere 2005;60:705–13.

Flury M. Pesticide transport through unsaturated field soils: preferential flow. SwissFederal Institute of Technology; 1994. p. 1–293. [ETH Zurich].

FOCUS. FOCUS groundwater scenarios in the EU review of active substances. Report ofthe FOCUS Groundwater Scenarios Workgroup 2000. EC Document ReferenceSanco/321/2000 rev.2; 2000. p. 1–202.

FOCUS. Assessing potential for movement of active substances and their metabolites toground water in the EU. Report of the FOCUS Ground Water Work Group. ECDocument Reference Sanco/13144/2010 version 1; 2009. p. 1–604.

FOCUS. FOCUS PELMO 4.4.3. Online retrieved at:http://viso.ei.jrc.it/focus/gw/index.html, 2011. [Accessed on 15/10/2012].

FOOTPRINT. The FOOTPRINT pesticide properties database. Database collated by theUniversity of Hertfordshire as part of the EU-funded FOOTPRINT Project(FP6-SSP-022704). Online retrieved at:www.eu-footprint.org.ppdb.html, 2006.[Accessed on 15/10/2012].

FROGS. French refinement of groundwater scenarios. Report of the UIPP EnvironmentalMethodology Working Group version 2.0; 2011. p. 1–314.

Jarvis N, Boesten JJTI, Hendriks RFA, Klein M, Larsbo M, Roulier S, et al. Incorporatingmacropore flow into FOCUS PEC models. In: Del Re AAM, Capri E, Padovani L,Trevisan M, editors. Pesticides in air, plant, soil and water system. Proceedings ofthe XII International Symposium on Pesticide Chemistry, Piacenza, Italy;2003a. p. 963–72.

Jarvis N, Hanze K, Larsbo M, Stenemo F, Persson L, Roulier S, et al. Scenario developmentand parameterization for pesticide exposure assessments for Swedish groundwater.Swedish University of Agricultural Sciences. Department of Soil Sciences. Division of En-vironmental Physics; 2003b [Report 4. ISSN 1651–7210].

KEMI (Kemikalieinspektionen, Swdish Chemical Agency). Reason for modelling theSwedish groundwater scenarios with the MACRO model for registration of plant

441H. Labite et al. / Science of the Total Environment 463–464 (2013) 432–441

protection products in Sweden. Online retrieved at:www.kemi.se, 2010. [Accessed on20/10/2012].

Klein M. Statistical distribution of pesticide concentrations in leachate results of aMonte-Carlo analysis performed with PELMO. Chemosphere 1997;35(112):319–89.

Klein M, Müller M, Dust M, Görlitz G, Gottesbüren B, Hassink J, et al. Leachingmodel PELMOusing lysimeter studies performed for registration. Chemosphere 1997;35(11):2563–87.

Klein M, Hosang J, Schäfer H, Erzgräber B, Resseler H. Comparing and evaluating pesti-cide leaching models. Results of simulations with PELMO. Agric Water Manag2000;44:263–81.

Kördel W, Klein M. Prediction of leaching and groundwater contamination bypesticides. Pure Appl Chem 2006;78(5):1081–90.

Kramers G. Preferential flow in Irish grassland soils. Ireland: School of Agriculture, FoodScience and Veterinary Medicine University College Dublin; 2009 [PhD thesis].

Kramers G, Holden NM, Brennan F, Green S, Richards KG. Water content and soiltype effects on accelerated leaching after slurry application. Vadose Zone J 2012.http://dx.doi.org/10.2136/vzj2011.0059.

Labite H, Cummins E. A quantitative approach for ranking human health risks frompesticides in groundwater. Hum Ecol Risk Assess 2012;18(6):1156–85.http://dx.doi.org/10.1080/10807039.2012.722797.

Labite H, Butler F, Cummins E. A review and evaluation of plant protection productranking tools used in agriculture. Hum Ecol Risk Assess 2011;17(2):1–28.

Loague K, Green RE. Statistical and graphical methods for evaluating solute transportmodels: overview and application. J Contam Hydrol 1991;7:51–73.

Met Éireann (the Irish National Meteorological Service). Climate data and products.Online retrieved at:http://www.met.ie/climate/climate-data-information.asp,2013. [Accessed from 1/1/2011 to 24/02/2013].

Pavlis M, Cummins E, McDonnell K. Groundwater vulnerability assessment of plantprotection products: a review. Hum Ecol Risk Assess 2010;16(3):621–50.

PEDOSPHERE. Bulk density calculator work table (US). Online retrieved athttp://pedosphere.ca/resources/bulkdensity/worktable_us.cfm, 2011. [Accessed on10/10/2012].

Piwowarczyk A. Assessment of the groundwater vulnerability to pesticides inputs from Irishagriculture: sorption isotherms and leaching of selected pesticides in representativeagricultural soils. Biosystems Engineering. University College Dublin; 2013 [PhD thesis].

Premrov A, Schulte RPO, Coxon CE, Hackett R, Richards KG. Predicting soil moistureconditions for arable free draining soils in Ireland under spring cereal crop produc-tion. Ir J Agr Food Res 2010;49:99–113.

Schulte EE, Hopkins BG. Estimation of organic matter by weight loss-on ignition. In:Magdoff FR, editor. Soil organic matter: analysis and interpretation. SSSA Spec.Publ.Madison, WI: SSSA; 1996. p. 21–31.

Scorza Júnior RP, Jarvis NJ, Boesten JJ, van der Zee SE, Roulier S. Testing MACRO (version5.1) for pesticide leaching in a Dutch clay soil. Pest Manag Sci 2007;63:1011–25.

van Alphen BJ, Stoorvogel JJ. Effects of soil variability and weather conditions onpesticide leaching. A farm level evaluation. J Environ Qual 2002;31:797–805.

Vanclooster M, Piñeros Garcet JD, Boesten JJTI, Leistra M, Smelt J, Jarvis N, et al. Effectiveapproaches for Assessing the Predicted Environmental Concentrations of Pesti-cides: a proposal supporting the harmonised registration of pesticides in Europe.APECOP project. Final report; 2003. p. 1–158.

Walker A, Calvet R, Del Re AAM, PestemerW, Hollis JM. Evaluation and improvement ofmathematical models of pesticide mobility in soils and assessment of theirpotential to predict contamination of water systems. Berlin-Dahlem: BiologischenBudensanstalt fur Land- und Forstwirtschaft3-8263-3069-2; 1995. p. 115. [Finalreport. ISSN 0067–5859].

Wauchope RD, Yeh S, Linders JBHJ, Klaskowski R, Tanaka K, Rubin B, et al. Pesticide soilsorption parameters: theory, measurement, uses, limitations and reliability. PestManag Sci 2002;58:419–45.

Wustenberghs H, Delcour I, D'Haene K, Lauwers L, Marchand F, Steurbaut W, et al. A dualindicator set to help farms achieve more sustainable crop protection. Pest Manag Sci2012;68(8):1130–40.

Zhang C, Moody A. Relevance to Ireland of EU FOCUS scenarios for assessing pesticideleaching to groundwater. Kildare Street, Dublin 1, Ireland: Department of Agriculture,Fisheries and Food; 2004.