Evidence Project Final Report - Defra, UK - Science...

EVID4 Evidence Project Final Report (Rev. 06/11) Page 1 of 24

General Enquiries on the form should be made to:

Defra, Procurements and Commercial Function (Evidence Procurement Team) E-mail: [email protected]

Evidence Project Final Report

Note

In line with the Freedom of Information Act 2000, Defra aims to place the results of its completed research projects in the public domain wherever possible. The Evidence Project Final Report is designed to capture the information on the results and outputs of Defra-funded research in a format that is easily publishable through the Defra website An Evidence Project Final Report must be completed for all projects.

This form is in Word format and the boxes may be expanded, as appropriate.

ACCESS TO INFORMATION

The information collected on this form will be stored electronically and may be sent to any part of Defra, or to individual researchers or organisations outside Defra for the purposes of reviewing the project. Defra may also disclose the information to any outside organisation acting as an agent authorised by Defra to process final research reports on its behalf. Defra intends to publish this form on its website, unless there are strong reasons not to, which fully comply with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.

Defra may be required to release information, including personal data and commercial information, on request under the Environmental Information Regulations or the Freedom of Information Act 2000. However, Defra will not permit any unwarranted breach of confidentiality or act in contravention of its obligations under the Data Protection Act 1998. Defra or its appointed agents may use the name, address or other details on your form to contact you in connection with occasional customer research aimed at improving the processes through which Defra works with its contractors.

Project identification

1. Defra Project code PS2249

2. Project title

Feasibility study for using modelling rather than monitoring to refine pesticide leaching assessments

3. Contractor organisation(s)

Environment Department University of York Heslington York YO10 5DD

54. Total Defra project costs £ 32923

(agreed fixed price)

5. Project: start date ................ 01/01/2013

end date ................. 31/12/2013

mailto:[email protected]


6. It is Defra’s intention to publish this form.

Please confirm your agreement to do so. ..................................................................................... YES X NO

(a) When preparing Evidence Project Final Reports contractors should bear in mind that Defra intends that they be made public. They should be written in a clear and concise manner and represent a full account of the research project which someone not closely associated with the project can follow.

Defra recognises that in a small minority of cases there may be information, such as intellectual property or commercially confidential data, used in or generated by the research project, which should not be disclosed. In these cases, such information should be detailed in a separate annex (not to be published) so that the Evidence Project Final Report can be placed in the public domain. Where it is impossible to complete the Final Report without including references to any sensitive or confidential data, the information should be included and section (b) completed. NB: only in exceptional circumstances will Defra expect contractors to give a "No" answer.

In all cases, reasons for withholding information must be fully in line with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.

(b) If you have answered NO, please explain why the Final report should not be released into public domain

Executive Summary

7. The executive summary must not exceed 2 sides in total of A4 and should be understandable to the intelligent non-scientist. It should cover the main objectives, methods and findings of the research, together with any other significant events and options for new work.

Mathematical modelling is used within pesticide registration procedures to determine the risk of pesticides and their metabolites leaching to groundwater. The regulatory assessment uses soil leaching models to predict annual average concentrations of pesticide leaching out of the soil at 1-m depth. An analysis by the British Geological Survey indicates that 74% of the land area of Great Britain has groundwater at >1-m depth and suggests the value of work to include modelling of pesticide fate in the deeper unsaturated zone. The aim of this project was to assess the feasibility of introducing a higher-tier of modelling with a more realistic endpoint as an additional step before proceeding to groundwater monitoring.

Initial work investigated how alternative modelling endpoints matched against the protection goal for groundwater as assumed by the European Food Safety Authority of “safeguarding the drinking water function of the groundwater”. It was decided to extend simulations below the soil root zone to include the full unsaturated zone with the prediction endpoint becoming the annual average concentration of pesticide reaching the surface of the groundwater body. Consideration was also given to including simulation of the groundwater body, but this was considered overly complex due to the groundwater body being an ill-defined endpoint; flow is primarily lateral within the saturated zone, so groundwater simulation would add great computational complexity as well as requiring explicit spatial treatment of source areas at the soil surface.

An existing model for nitrate transport through the unsaturated zone was modified to incorporate the pesticide fate processes of sorption, degradation and volatilisation. A modelling procedure was then undertaken comprising modelling of the soil root zone with the FOCUS-PEARL model and simulation of the unsaturated zone with the modified nitrate model to yield pesticide loads and concentrations at the surface of the groundwater body. The model accounted for spatial distribution of land use, properties of the unsaturated zone, and depth to groundwater.

Two case study areas of 900 km2 each were selected based on contrasting hydrogeological conditions

and the availability of pesticide monitoring data against which to evaluate model performance. The Doncaster study area overlies Triassic Sandstone and has a shallow water table with average depth of 4 m. The Stevenage study area overlies Chalk and has a wide range of depth to groundwater (0-111 m) with the average being 23 m. Four pesticides were selected for inclusion as case study examples based on widespread historic usage and availability of monitoring data for model evaluation. The selected compounds were atrazine, bentazone, isoproturon and mecoprop. FOCUS-PEARL modelling was based on the Okehampton scenario which is based on a broadly distributed sandy loam which is representative for the two study areas; modelling used generic pesticide properties and assumed usage of the four


compounds on all target crop which was selected as maize (atrazine), peas and beans (bentazone), and winter wheat (isoproturon and mecoprop). Loading of pesticide leaching through the soil was used as input to the unsaturated zone model with a correction to account for average annual recharge.

For three of the four test compounds (bentazone, isoproturon, mecoprop), concentrations of pesticide predicted to leach to 1-m depth based on FOCUS groundwater modelling were much larger than average concentrations from monitoring and also exceeded maximum concentrations in any sample from the monitoring data. In contrast, predictions that included the unsaturated zone to estimate concentrations at the groundwater surface were closer to average monitoring data and were exceeded in some cases by maxima from the monitoring. Concentrations of atrazine detected during monitoring were better represented by standard FOCUS modelling and this is attributed to much of this contamination deriving from non-agricultural uses that were not included within the modelling framework.

There were 12 monitoring locations within the Doncaster study area and here the model broadly matched spatial differences in vulnerability. The model indicated distinct variation in vulnerability across the Stevenage study area, with greater vulnerability predicted in the north-west and south-east of the 900-km

2

area. There were 37 monitoring locations for the Stevenage case study, and here measured data differed from the modelling in showing no consistent differences in vulnerability across the area. Discrepancies may arise due to rapid and/or lateral flow through fissures in the karstic Chalk. The north-west part of the Stevenage study area was predicted to be vulnerable to pesticide leaching, but there were no monitoring locations within the vulnerable zone to allow evaluation of this prediction.

Future development of the approach should consider four priorities: (i) inclusion of spatial distribution in soils and local weather into simulations for the soil zone; (ii) inclusion of a more sophisticated model of the unsaturated zone that is able to account for non-uniform recharge and hydrodynamic dispersion; (iii) detailed consideration of the appropriate scale for expressing variability in pesticide inputs at the surface and in environmental conditions controlling leaching to groundwater; and (iv) adoption of a pan-European approach consistent with the introduction of zonal authorisations for pesticides.


Project Report to Defra

8. As a guide this report should be no longer than 20 sides of A4. This report is to provide Defra with details of the outputs of the research project for internal purposes; to meet the terms of the contract; and to allow Defra to publish details of the outputs to meet Environmental Information Regulation or Freedom of Information obligations. This short report to Defra does not preclude contractors from also seeking to publish a full, formal scientific report/paper in an appropriate scientific or other journal/publication. Indeed, Defra actively encourages such publications as part of the contract terms. The report to Defra should include:

the objectives as set out in the contract;

the extent to which the objectives set out in the contract have been met;

details of methods used and the results obtained, including statistical analysis (if appropriate);

a discussion of the results and their reliability;

the main implications of the findings;

possible future work; and

any action resulting from the research (e.g. IP, Knowledge Exchange).

1 INTRODUCTION

Since 2000, the regulatory assessment of pesticide leaching to groundwater has been based on the FOCUS Groundwater Scenarios (FOCUS, 2000). Subsequent work identified that four of these scenarios (Chateaudun, Hamburg, Kremsmunster, Okehampton) are relevant to UK conditions. The regulatory endpoint for pesticide leaching is both challenging and unrealistic in the majority of situations. This endpoint is currently defined as the concentration in soil pore water at 1-m depth should not exceed 0.1 µg/L for the parent compound and relevant metabolites and 10 µg/L for non-relevant metabolites. Recent changes to guidance on soil persistence (EFSA, 2010) will generally lead to larger concentrations of a particular active substance being simulated using the FOCUS GW scenarios (e.g. due to changes to assumptions on washoff). It is widely anticipated that a significant number of widely-used products will fail the FOCUS GW assessment during re-registration and that the requirement to address the concern will be a series of large, expensive monitoring programmes to assess presence of individual active ingredients in groundwaters across Europe. This places a large financial and manpower burden on industry and also on regulators who will need to review complex monitoring protocols and review the resulting studies. This increase in regulatory burden may be unreasonable without further analysis given that pesticide concentration at 1 m below a single field is a crude and precautionary surrogate for likely concentrations in groundwater.

An analysis by the British Geological Survey suggests that 74% of the British land mass has depth to groundwater >1 m (Figure 1). Overlay of these areas with the CEH Land Cover Map 2000 indicates that 70% of all arable and horticultural land in Great Britain has depth to groundwater >1 m. This suggests that there is merit in research into inclusion of modelling of pesticide fate in the deeper unsaturated zone.


Figure 1. Areas in Great Britain with depth to groundwater > 1m (source: BGS)

1.1 Aim and objectives

The aim of this project was to assess the feasibility of introducing a higher-tier of modelling with a more realistic endpoint as an additional step before proceeding to groundwater monitoring. It used pilot areas in the UK as a case study for the analysis, but the work is intended to be broadly applicable to the European agricultural area.

The specific objectives were to:

1. Assess possibilities for alternative prediction endpoints;

2. Select pilot areas as representative of UK conditions and assess the distribution of land use and groundwater depth within those areas;

3. Generate and evaluate case study assessments for four pesticides to demonstrate effect of endpoint and environmental conditions on PECgw; and

4. Report findings to a stakeholder workshop and propose options for regulatory development.

2 PREDICTION ENDPOINT

Figure 2 summarises the major components of water movement within the plant-soil-groundwater continuum. The FOCUS groundwater scenarios simplify this scheme in ignoring runoff or any sub-lateral flows within the soil-root zone; in addition, a simplified approach is adopted to any capillary rise of water and pesticide and this does not connect with the groundwater body. Thus the FOCUS groundwater approach reverts to an entirely 1-dimensional description of water and pesticide movement through soil. This simplification is conceptually easy to grasp within the upper 1 m of soil; for example, ignoring any surface runoff will maximise infiltration of rainfall into the soil profile and thus maximise the prediction of pesticide leaching. The assumption that there is no lateral component to water movement becomes less tenable as simulations penetrate deeper below the soil surface and completely breaks down within the groundwater body where flows are primarily lateral.

Figure 2. Simplified diagram of water movement within the plant-soil-groundwater continuum (source: www.helmholtz-muenchen.de)


A detailed analysis of the protection goal for groundwater assessments within the regulatory framework for pesticides is beyond the remit of this report. The latest scientific opinion of the EFSA PPR panel on the FOCUS groundwater report (FOCUS 2009) notes that “a specific protection goal has to be defined in order to be able to assess the risk assessment methodology” (EFSA, 2013). The EFSA opinion goes on to state that the PPR panel assumed in its work that the specific protection goal is “safeguarding the drinking water function of the groundwater”. This assumption forms the most appropriate baseline for considering options to refine the assessment endpoint to match the protection goal.

FOCUS (2009; p35) proposes that “a national protection goal upper limit should be 0.1 µg/L annual average

in ground water at the 90th percentile vulnerability taking into account both spatial variability for soil and

climatic conditions and temporal variability on a multi-year basis, in the agricultural use areas of the product.” FOCUS (2009; p56) further identifies “considering realistic depths to ground water” as one of the possible options for modelling using refined scenarios (Tier 2b). However, this possibility is not developed further within the report and there is some ambiguity over whether the refinement is intended to generate a more realistic bottom boundary condition but leave the assessment point at 1-m depth (FOCUS, 2009; p161) or move the assessment point towards the surface of the groundwater body.

The FOCUS proposal for national protection goals clearly establishes criteria on concentration threshold for parent substances (0.1 µg/L) and vulnerability threshold (90

th percentile across space and time) and also

the principle of assessing the concentration threshold against an annual average prediction. Two possibilities were considered initially within the current project to refine the regulatory endpoint at higher tiers of assessment. Each retains consistency with the FOCUS proposal for national protection goals:

(i) Modelling is extended below the soil root zone to include the full unsaturated zone with the prediction endpoint becoming the annual average concentration of pesticide reaching the surface of the groundwater body;

(ii) Modelling is extended as for (i) and also includes a simulation of the groundwater system, meaning that the prediction endpoint becomes the annual average concentration of pesticide within the groundwater body.

Endpoint (i) is conceptually very similar to the current endpoint of annual average pesticide concentration at 1-m depth. There is a trade-off between the additional realism introduced by considering the true depth to groundwater and the additional simplification introduced by assuming that all flow down to the surface of the ground water is vertical with no lateral element. A range of modelling approaches from simple attenuation factor models to fully mechanistic unsaturated zone models could be coupled to existing simulations of leaching through soil to generate the refined endpoint.

In contrast, endpoint (ii) presents the challenge of a rather ill-defined endpoint that conflicts with the current regulatory paradigm. For example, it would be necessary to define whether the annual average concentration was across the full groundwater body, some defined depth within the groundwater body, or an annual average arriving at a certain point (e.g. a groundwater abstraction point). Simulations would necessarily need to incorporate elements for the soil, unsaturated, and groundwater zones; any groundwater model would need to simulate lateral flow and to account for the full contributing source area at the soil surface. Such models exist for groundwater planning purposes, but they give little attention to the soil zone and there are no examples where these models have been used to simulate pesticide concentrations in groundwater.

Accordingly, the project focused on extending the assessment endpoint for pesticide leaching from the base of the soil profile (1-m depth) to the surface of the groundwater body (variable depth according to hydrogeological situation). There is a precedent for this endpoint in the EU whereby Ireland includes the following clause within its regulatory criteria: “If it can be demonstrated that depth to ground water in areas of use is >1 m, then annual average concentration impacting on top of the ground water body must not exceed 0.1 μg/L.”

Setting the refined assessment endpoint at the groundwater surface raises a question over the spatial dimension of pesticide usage. The FOCUS groundwater scenarios predict concentrations at 1-m depth below a single treated field. This approach has certainty in meeting the EFSA (2013) assumed protection goal of “safeguarding the drinking water function of the groundwater”, but is likely to be precautionary because drinking water will never derive from a homogeneously treated area. As the simulation depth increases, it becomes increasingly logical to consider that water impacting on a groundwater surface is drawn from a wide area with a range of land uses. The case studies described below adopt a raster-based approach to this issue with cropping and pesticide use described at a resolution of 1-km

2 blocks. This

approach is adopted for illustration only and the issue of spatial aggregation of pesticide use is discussed within Section 4.


3 CASE STUDIES – METHODOLOGY

3.1 Case study locations

Two case study locations were selected to give contrasting underlying geology and a range in depth to groundwater (Figures 3 and 4). The Doncaster study area overlies Triassic Sandstone (Figure 3). Due to the outcrop and dip of the sandstone, a rectangular area 15 km wide and 60 km long was used for the case study. There are relatively few public supply/EA monitoring points in this formation, but high resolution shallow groundwater monitoring data to support model evaluation are available from studies published by the British Geological Survey (BGS). The water table is shallow in this area with an average depth of 4 m and there is a historical pesticide pollution problem (e.g. Lapworth et al., 2006).

A square 30 km by 30 km was used for the Stevenage study area (Figure 4). In this area there is a good network of monitoring points in the chalk, and there are several published studies on pesticide pollution in the UK Chalk (e.g. Chilton et al., 2005; Worrall et al., 2005; Lapworth and Gooddy, 2006). The area has a range in water level depths (0-111 m with an average value of 23 m), and to the east of the block is the start of the Palaeogene (clay) cover which is also interesting from a hydrogeology/contaminant transport perspective.

Four pesticides were selected for inclusion as case study examples. These were atrazine, bentazone, isoproturon and mecoprop. All four compounds have broad historical usage and monitoring data indicate potential for groundwater contamination under specific circumstances. Although two of these compounds are now prohibited from use (atrazine and isoproturon), precedence was given to the availability of high-quality monitoring data for use in model evaluation rather than to including only current-use compounds.

Figure 3. The Doncaster study area located over Permo-Triassic sandstone; crosses show locations of

public supply or Environment Agency monitoring points.


Figure 4. The Stevenage study area located primarily over chalk; crosses show locations of public supply

or Environment Agency monitoring points.

3.2 Overview of modelling procedures

The modelling procedure comprised three discrete steps. First, the FOCUS-PEARL model (v4.4.4) was used to generate a 20-year series of pesticide loadings leaching to the base of the soil layer for each of the study compounds. This loading from the soil zone was first corrected to account for intensity of cropping and hence of potential treatment area. The corrected loading from soil was then used as input to simulations for the unsaturated zone using a modified version of the nitrate time bomb model (Wang et al., 2012; Wang et al., 2013). The result of the unsaturated zone simulations was annual average concentrations of pesticide at the surface of the groundwater body for the 900 1x1-km

2 cells within each

case study area. Finally, these simulations were evaluated against available monitoring data. Figure 5 summarises the modelling scheme and provides details of the characteristics of the system included within simulations at each stage.


Figure 5. Overall schematic of the modelling procedure used to generate case studies

3.3 FOCUS modelling for transport to the base of the soil layer

The FOCUS groundwater scenarios were used for baseline soil modelling. This provides a direct comparison between a standard regulatory output and the result when unsaturated zone modelling is included as a refinement. Initially, all FOCUS groundwater scenarios were simulated. However, following review of results FOCUS Okehampton was selected as the modelling scenario because (i) the weather data and cropping parameters are from the UK, (ii) the soil is a broadly distributed sandy loam which is representative for the two study areas, (iii) all three target crops are considered within the scenario, and (iv) Okehampton is a relatively vulnerable scenario which increases the size and frequency of pesticide leaching from the base of the soil profile. Standard simulations comprising a 6-year warm up and 20-year prediction period were undertaken for application of atrazine to maize, bentazone to field beans, and isoproturon and mecoprop to winter wheat. Pesticide properties were compiled from available data sources (EU agreed endpoints and DARs, www.herts.ac.uk/ppdb) and are summarised in Table 1. Applications at the rates given in Table 1 were assumed to be made in every year of the simulation; crop rotation was ignored in this instance as the objective was to predict the leaching through the soil layer that would occur if the target pesticide was applied to the target crop somewhere within the case study area. A deviation from standard FOCUS practice was that crop interception was calculated by the model in the simulations (rather than applying the substance to soil and manually correcting for crop interception outside of the model run). The outputs fed into the unsaturated zone model were annual volume of water and mass of pesticide leaching to 1-m depth for each of 20 years of simulation.

It should be noted that no attempt was made to match actual use conditions for this proof-of-concept exercise. Thus, the actual distribution of soils within the case study areas was ignored to ensure that leached load from the soil layer was equal for all cropped areas. Similarly, weather data were the generic climate files for Okehampton rather than the true weather that would have occurred within the case study areas at the time pesticides were applied (e.g. throughout the 1980s, 1990s and 2000s). Finally, application rate was not adjusted to accommodate for any changes in use patterns over time (e.g. the removal of atrazine from the market in 1992).

http://www.herts.ac.uk/ppdb


Table 1. Pesticide parameters used for soil leaching simulations with FOCUS-PEARL

Parameter Unit Atrazine Bentazone Isoproturon Mecoprop

Application Target crop Maize Field beans Winter wheat W wheat

Application rate kg/ha 2,000 1,440 2,000 1,500

Application date 8 May 17 Apr 24 Oct 8 Nov

General Molecular mass g/mol 215.68 240.3 206.3 214.6

Saturated vapour pressure Pa 39,000 0.00017 6.30E-06 0.0016

Temp. for vapour pressure °C 25 20 20 20

Molar enthalpy for vaporisation kJ/mol 95 95 95 95

Solubility in water mg/L 35 570 70.2 620

Temp. for sol. in water °C 20 20 20 20

Molar enthalpy of dissolution kJ/mol 27 27 27 27

Freundlich sorption pH dependency no no no no

Kom ( = Koc/1.72) L/g 101.16 18.02 60.47 18.02

Temp. for Kom °C 25 20 20 20

Molar enthalpy of sorption kJ/mol 0 0 0 0

Reference concentration in liquid phase mg/L 1 1 1 1

Freundlich sorption exponent 0.9 0.9 0.9 0.9

Desorption rate coefficient /d 0 0 0 0

Factor relationgCofFreNeq and CofFreEql 0 0 0 0

Transformation Half-life d 66 45 22 8.2

Temp. for half-life °C 20 20 20 20

Optimum moisture conditions pF 2 or wetter no yes no no

Liquid content in incubation experiment kg/kg 1 1 1 1

exponent for the effect of liquid 0.7 0.7 0.7 0.7

Molar activation energy kJ/mol 65.4 65.4 65.4 65.4

Diffusion Reference temp. for diffusion °C 20 20 20 20

Reference diffusion coefficient in water m2/d 4.3E-05 4.3E-05 4.30E-05 4.3E-05

Reference diffusion coefficient in air m2/d 0.43 0.43 0.43 0.43

Crop Wash-off factor /m 0.0001 0.0001 0.0001 0.0001

Canopy process option

Lumped Lumped Lumped Lumped

Half-life at crop surface d 1000000 1000000 1000000 1000000

Coefficient for uptake by plant - 0.5 0.5 0.5 0.5

3.4 Modelling of the unsaturated zone

3.4.1 The nitrate time-bomb model

The section below briefly summarises the nitrate time-bomb model which was used as a basis for the pesticide modelling in this project. It can take decades for nitrate transport in groundwater systems, and a national scale nitrate time-bomb model was developed to simulate the nitrate storage and long lag-time in unsaturated zones (Wang et al., 2013); this is the first step in predicting the future long-term evolution of nitrate contamination in groundwater. This model simulates the distribution of nitrate arriving at the water table which depends on three functions:

the nitrate input at the land surface (the temporally varying but spatially uniform leaching of nitrate from the base of the soil);

the rate of travel of nitrate through the unsaturated zone (spatially dependent on variations in hydrolithological characteristics); and

the thickness of the unsaturated zone (the distance between the groundwater surface and the base of the soil layer)

The unsaturated zone thickness and nitrate velocity are combined to estimate the spatial distribution of


nitrate travel time in the unsaturated zone and from this the input year for nitrate reaching the water table at any defined time. The nitrate input function over time can then be used to estimate the concentration reaching the water table at any point and defined time. The long-term average groundwater heads were derived using hydrogeological groundwater level contours, observed boreholes and river water levels (Wang et al., 2012). Nitrate transport speed in the unsaturated zone was obtained from a literature review and expert input from hydrogeologists. The nitrate input function was derived from literature review and has been verified against BGS historic pore water nitrate concentrations from nearly 300 boreholes.

The nitrate time-bomb model is based on the following assumptions:

nitrate input/loading is from the base of the soil;

movement is through the matrix only in dual-porosity strata;

the mass of nitrate in the unsaturated zone is preserved, except where the bedrock is overlain by low permeability superficial deposits;

nitrate moves vertically from the land surface to the water table;

nitrate moves at a constant velocity through the unsaturated zone; and

there is no hydrodynamic dispersion of nitrate in the unsaturated zone.

The model has been successfully applied in predicting the arrival of peak nitrate concentrations at the water table in Great Britain (Wang et al., 2012), and in investigating nitrate storage and lag-time in the unsaturated zone of Permo-Triassic Sandstones in the Eden Valley, UK.

3.4.2 Modifications to the model to incorporate pesticide fate

Pesticide loadings to groundwater are determined by soil, chemical, management, hydrogeological and climatic factors. Depth to the water table, rate of pesticide travel and the residence time within the unsaturated zone will clearly affect the risk of groundwater pesticide pollution; however, the most important modification to model structure to enable pesticide simulations is to incorporate pesticide attenuation processes such as sorption, volatilisation and degradation.

Leonard and Knisel (1988) summarised numerical methods for simulating pesticide behaviour during leaching to groundwater, highlighting particularly those methods that considered degradation and volatilisation as well as convective flow. Rao and Davidson (1985) incorporated pesticide degradation functions into travel time expressions to yield an attenuation factor approach to represent pesticide sorption, volatility and degradation processes. This method was adopted here in modifying the nitrate time-bomb model for use in simulation of pesticides.

The amount of pesticide reaching the water table ( 2M ; µg) can be expressed by the equation below:

2/1

02

693.0)1(exp

tq

FC

FC

KAC

FC

KfBDLMM hoc

oc (1)

where oM is the mass of pesticide loading from the base of the soil layer in each modelling cell (µg); L is

the distance between the base of the soil layer and the water table (m); BD is the bulk density in the

unsaturated zone (kg/m3); ocf is the fraction of organic carbon in the unsaturated zone (-); ocK is the

organic carbon pesticide partitioning coefficient (m3/kg); FC is the field capacity (its value for aquifers can

be calculated using porosity and specific yield for unconfined aquifers); AC is the air-filled porosity; q is

the amount of recharge (m); hK is the dimensionless Henry’s law constant; and 2/1t is pesticide

degradation half-life (year).

In Equation 1, FC

KfBD oc

oc , FC

KAC h and 2/1t represents the pesticide processes of sorption, volatility

and degradation respectively in the unsaturated zone.

The modified model for use with pesticides was developed by integrating equation 1 into the nitrate time-bomb model; experiments were then carried out to verify the model. It uses datasets comprising:

Annual pesticide leaching rate from soil layer (output from FOCUS-PEARL)

Annual recharge

Geological units (GIS layer)

Long-term average groundwater heads (GIS layer)

Percentage crop coverage (GIS layer)

Historic land use change (optional)

The code produces the following results:


A time series of annual average pesticide concentrations reaching the water table for each geological unit

Spatial distribution of annual average pesticide concentration at the water table (GIS layer)

Spatial distribution of annual mass of pesticide reaching the water table (GIS layer)

Pesticide travel time in the unsaturated zone (GIS layer)

Spatial distribution for the mass of pesticide attenuated within the unsaturated zone (GIS layer)

The maximum amount of pesticide in each grid during the simulation period (GIS layer)

3.4.3 Modelling procedure

The modified model was applied to simulate leaching of the four pesticides in the two case study areas. The model was applied in predictive mode, with no attempt to calibrate the model against the results of monitoring. First, each study area was discretised into 100 m by 100 m grids by clipping and re-sampling the nitrate travel time, geological units and long-term average groundwater heads. It was assumed that recharge to groundwater is homogenous over time; its value was calculated by matching the calculated travel time with the existing nitrate travel time data used in the study of Wang et al. (2012). Since nitrate was treated as a conservative contaminant in the unsaturated zone, this calculation assumed that there is

no sorption or volatility (i.e. ocK and hK =0). The calculated recharge values in Doncaster and Stevenage

were 147.0 and 215.9 mm p.a.

The average annual recharge to the surface of the groundwater body (147.0 or 215.9 mm p.a.) is frequently smaller than the recharge leaving the soil profile (e.g. 165-627 mm p.a. for maize grown in the Okehampton scenario). Conceptually, the difference was attributed to lateral movement of water at the base of the soil layer and it was assumed that this water had no further interaction with the unsaturated zone or linkage to the groundwater body. Further, it was assumed that pesticide reaching the base of the soil profile could be divided proportionate to flow between that lost to lateral flow and that moving vertically from the base of the soil through the unsaturated zone. Mathematically, the estimated annual recharge was used to correct the pesticide leaching load from FOCUS-PEARL using:

ttV WRMM (2)

where MV is the pesticide moving vertically out of the soil profile and entering the unsaturated zone (µg), R is the annual recharge to groundwater (mm), and Wt and Mt are the total amount of water (mm) and pesticide (µg), respectively, predicted by FOCUS-PEARL to reach the base of the soil profile. To exemplify the approach, if FOCUS-PEARL predicted annual leaching to 1-m depth of 294 mm water carrying 10 mg of pesticide for the Doncaster case study (147 mm p.a. groundwater recharge), then 147 mm of water carrying 5 mg of pesticide would enter the unsaturated zone for that year and the other 147 mm of water carrying the remaining 5 mg of pesticide would be assumed to be lost to lateral transport processes. The process was repeated for all 20 years of FOCUS-PEARL simulations to give 20 annual inputs of pesticide into the unsaturated zone model. Where leachate from FOCUS-PEARL was less than the mean annual recharge to groundwater, all of the leached pesticide load was assumed to enter the unsaturated zone.

Next, pesticide loading from soil was corrected for intensity of cultivation of the target crop. This was done using the 5x5 km cropping data from 1995 embedded within the SEISMIC database. The 1995 dataset was selected because this was considered a representative date for pesticide usage at the surface that could result in residues reaching groundwater during the monitoring period of 1992-2013. It was considered beyond the scope of this proof-of-concept project to incorporate changes in land use patterns over time. Pesticide loadings into the unsaturated zone were scaled by assuming that all target crop area was treated at the maximum rate and all other areas were untreated:

CropMM V 0 (3)

where M0 is the initial loading of pesticide into the unsaturated zone (defined in Equation 1) and Crop is the proportion of total land area within each 1-km

2 block that comprises the target crop.

The parameters in Equation 1 were selected from literature review. For example, Table 2 shows the

fractions of organic carbon ( ocf ) for major UK aquifers collected from the work of Milne and Kinniburgh

(2007). Hydrological parameters for the two unsaturated zone systems are listed in Table 3. Details of the data underpinning these parameter sets are given by Wang et al. (2012; 2013).


Table 2. Fraction of organic carbon (%) in selected UK aquifer material from Milne and Kinniburgh (2007)

Group code Group name Min

25th percentile Median Mean

75th percentile Max n

Quaternary Superficial Deposits (undifferentiated)

0.46 0.46 0.46 0.46 0.46 0.46 3

Palaeogene Bracklesham Group 0.41 0.45 0.49 0.49 0.53 0.57 2

Thames Group 0.16 0.32 0.42 0.428 0.56 0.6 9

Barton Group, Bracklesham Group and Bagshot Formation (undifferentiated)

0.05 0.05 0.05 0.05 0.05 0.05 1

Lambeth Group 0.05 0.108 0.145 0.516 0.3 5.01 14

Other Palaeogene Rocks 0.04 0.06 0.155 1.02 0.508 7.56 10

Cretaceous Chalk Group 0.04 0.05 0.06 0.129 0.09 1.6 64

Triassic Sherwood Sandstone Group 0.01 0.0225 0.05 0.0907 0.0575 0.89 30

Mercia Mudstone Group 0.01 0.05 0.05 0.0884 0.07 0.58 32

Other New Red Sandstone Supergroup Rocks

0.04 0.05 0.05 0.072 0.05 0.17 5

Table 3. Aquifer properties used for the two case study areas

Parameter Unit Sandstone in

Doncaster Chalk in

Stevenage

Rock bulk density ( BD ) kg/m3 2000 1850

Fraction of organic carbon ( ocf ) - 0.0000907 0.00129

Field capacity ( FC ) - 0.147 0.296

Air-filled porosity ( AC ) - 0.297 0.314

Whilst there are relatively few studies on pesticide behaviour within unsaturated zone materials, those studies that are available have often investigated one or more of the pesticides simulated within this exercise. A review of this literature was undertaken showing, for example, half-lives in the unsaturated zone of 1200-1500 days for atrazine (McMahon et al., 1992; Gaus & Casteele, 2004), 120-1400 days for mecoprop (Heron and Christensen, 1992; Gaus & Casteele, 2004), 80-400 days for isoproturon (Johnson et al., 2000) and 1500 days for bentazone (Gaus & Casteele, 2004). The conclusions from this short review exercise were:

(i) There is no consistent difference between Koc values measured in soils and the unsaturated zone. Hence, the Koc concept was assumed to be valid and soil Koc values were directly transposed to describe sorption in the unsaturated zone.

(ii) Degradation half-lives in the unsaturated zone showed a distribution of values for all four pesticides, with half-lives significantly slower in the unsaturated zone. Overall, a half-life in the unsaturated zone that was 10 times longer than that measured in topsoil under field conditions described the available data relatively well (hence a compound with topsoil half-life of 14 days would show an unsaturated zone half-life of 140 days). There is large uncertainty about rates of pesticide degradation within the unsaturated zone, so all simulations were run using a half-life set at 10 times that in topsoil and the simulations were repeated with half-life set at 100 times that in topsoil.

Table 4 shows the parameters used as the basis for describing pesticide behaviour in the unsaturated zone. All four compounds have negligible volatility and this aspect was not simulated following initial tests with isoproturon.


Table 4. Parameters for properties of the four pesticides

Parameter Unit Atrazine Bentazone Isoproturon Mecoprop

Organic carbon pesticide

partitioning coefficient ( ocK ) m

3/kg 0.100 0.0553 0.122 0.0470

Pesticide degradation half-life

in topsoil ( 2/1t ) year 0.0795 0.0384 0.0630 0.0225

3.5 Model evaluation

Monitoring data for the four selected pesticides were obtained from the Environment Agency from the two chosen case study areas. Table 5 provides summary statistics on number of monitoring locations, number of samples, and frequency and concentrations of positive detections.

Table 5. Summary of monitoring data for both study area and all study compounds

Summary data Atrazine Bentazone Isoproturon Mecoprop

Doncaster

Number of monitoring locations 12 12 12 12

Number of monitoring locations with positive detections 4 2 0 1

Relative detection rate (% of monitoring locations) 33.3 16.7 0.0 8.3

Total number of samples 118 126 107 127

Number of samples with positive detections 28.0 4.0 0.0 9.0

Relative detection rate (% of samples) 23.7 3.2 0.0 7.1

Average concentration (µg/L) 0.013 0.0003 0 0.011

Maximum concentration (µg/L) 0.218 0.0141 0 0.305

Stevenage

Number of monitoring locations 37 37 37 37

Number of monitoring locations with positive detections 37 6 15 12

Relative detection rate (% of monitoring locations) 100.0 16.2 40.5 32.4

Total number of samples 2091 776 1869 832

Number of samples with positive detections 2091 776 1869 832

Relative detection rate (% of samples) 100.0 100.0 100.0 100.0

Average concentration (µg/L) 0.033 0.0004 0.0087 0.0023

Maximum concentration (µg/L) 0.5 0.06 1.35 0.59

The pesticide monitoring data used for model evaluation are point samples that are discrete in space and time. Modelling undertaken within the project was proof-of-concept and pesticide application data input into the FOCUS-PEARL soil modelling were maximum label rates rather than actual values varying over time. Similarly, FOCUS weather data were used for simulations at this pilot stage rather than true weather for the case study locations that pertained over the time of pesticide applications into the system. For all of these reasons, it was not possible to make a direct, quantitative comparison between pesticide monitoring data and model simulations. Rather, the monitoring data were used to evaluate the overall plausibility of the model simulations within the context of measured observations. The following questions were investigated:

(i) Is the model able to differentiate locations in the study areas where pesticides have been detected in groundwater from those with no detections?

(ii) Is the model able to rank the relative frequency with which the different pesticides are detected in groundwater?

(iii) How do model simulations for average concentrations at the surface of groundwater compare with concentrations of pesticides detected in monitoring programmes?


4 CASE STUDIES - RESULTS AND DISCUSSION

4.1 Model simulation results

The model produces a time-series of the annual average concentration and loading of pesticide reaching the surface of the groundwater body for every 1-km

2 block within the case study area (900 blocks for each

area). Figures 6 and 7 give results for the four pesticides in the Doncaster and Stevenage areas, respectively. Results for each pesticide comprise (i) intensity of land use under the target crop (maize for atrazine, peas and beans for bentazone, and wheat for isoproturon and mecoprop), (ii) distribution of travel time to the surface of the groundwater body in years, and (iii) the maximum annual average concentration of the pesticide for each block within the area. As the third chart presents the maximum for each block, the simulation year in which that maximum occurs may vary between blocks (i.e. a block with shorter travel time to groundwater will see its maximum concentration earlier in the simulation than a block with longer travel time to groundwater).

4.1.1 Doncaster

Intensity of land use under the target crop was up to a maximum of ca. 2, 9 and 50% for maize, peas and beans, and wheat, respectively (Figure 6, left charts). These intensities will effectively dilute the loading of pesticide into the unsaturated zone across a 1-km

2 block to account for the proportion of land that is

treated.

The spatial pattern of travel time to groundwater was almost identical for the four pesticides (Figure 6, central charts) as this spatial distribution is controlled by depth to groundwater and hydrogeological characteristics of the unsaturated zone. The range in travel times varied with pesticide. There were some blocks with zero travel time in the north-east of the Doncaster area; this result arises because groundwater is found directly at the base of the soil profile. Longer travel times to groundwater occurred predominantly in the south-west of the area with the longest travel times predicted to range from 84 years for mecoprop to 134 years for isoproturon.

The spatial pattern in maximum annual average concentrations reaching groundwater is a function of the overlying land use, travel time to groundwater and properties of the pesticide. The largest annual average concentrations of atrazine occurred in the east of the area where an area of maize cultivation overlay hydrogeology leading to very short travel times to groundwater (Figure 6, right charts). Bentazone, isoproturon and mecoprop all had more dispersed patterns in their maximum concentrations, but the maximum annual average concentration tended to be larger towards the north and east of the area and smaller in the south-west where travel times were longer. The absolute maximum annual average

concentration in any of the 900 1-km2 blocks was 0.015, 1.11, 3.94 and 0.75 µg/L for atrazine, bentazone,

isoproturon and mecoprop, respectively. All four compounds had locations with insignificant concentrations predicted to reach the surface of groundwater. The proportion of blocks with maximum annual concentration <0.01 µg/L was 98.1, 49.7, 42.9 and 59.5% for the four compounds, respectively.

4.1.2 Stevenage

There is a great range in depth to groundwater in the Stevenage case study area. Maximum travel times to groundwater were predicted to be longer (several hundred years) although there were also large areas with much shorter travel times (Figure 7). Shortest travel times were predicted to occur in the north-west and south-east corners of the area. These locations also had relatively intense arable land use, so patterns of maximum annual average pesticide concentrations reaching groundwater in Stevenage were more discretised than for Doncaster. The absolute maximum annual average concentration in any of the 900 1-

km2 blocks was 0.011, 0.93, 2.46 and 0.47 µg/L for atrazine, bentazone, isoproturon and mecoprop,

respectively. These values are all slightly smaller than for the Doncaster area. There were far more locations predicted to show negligible leaching in the Stevenage area; the proportion of blocks with maximum annual average concentration <0.01 µg/L was 99.1, 86.5, 84.2 and 88.9% for the four compounds, respectively.


Atrazine

Bentazone

Isoproturon

Mecoprop

Figure 6. Distribution of associated crop, travel time to groundwater and maximum annual pesticide concentration for the four study compounds in the Doncaster area


Atrazine

Bentazone

Isoproturon

Mecoprop

Figure 7. Distribution of associated crop, travel time to groundwater and maximum annual pesticide concentration for the four study compounds in the Stevenage area


4.2 Model Evaluation

4.2.1 Spatial distribution of detections

Model simulations for maximum annual concentration reaching groundwater (Figures 6 and 7) can be compared with spatial pattern of positive detections from groundwater monitoring (Figures 8 and 9) for a qualitative assessment of degree of match. The Doncaster case study area has rather few monitoring locations (12 in total) and these are all located in the north of the area. Figure 8 shows that detections of atrazine, bentazone and mecoprop are concentrated into the north-west corner of the area. This location is identified by the model as one where short travel times to groundwater coincide with relatively dense arable land use and thus potential for pesticide use. Of the remaining monitoring locations within the study area, those in a western position and just to the north of Doncaster show no detections; this agrees with model predictions which suggest little potential for leaching to groundwater due to long travel times. Two monitoring locations near Goole in the east of the study area also showed no pesticide detections during monitoring. The model does not agree in this instance as this location is predicted to be vulnerable, particularly to leaching of bentazone, isoproturon and mecoprop (Figure 5). There is no obvious reason for this discrepancy, but caution should be exercised in over-interpretation from just two monitoring points in this location.

Figure 8. The Doncaster study area (full area left, detail right) showing monitoring sites and the number of compounds detected on any sampling occasion at each; zero compounds detected (white), one compound (yellow), two compounds (orange), three compounds (blue).

There is far greater intensity of groundwater monitoring sites within the Stevenage area (37 in total). It is notable that none of these locations lie within the part of the study area that is predicted by the model to be most vulnerable (i.e. the north-west). In this instance, the spatial match between model predictions and monitoring data is rather poor. The model suggests a clear distinction between areas with low risk of pesticide leaching in the centre of the study area and small areas with higher risk in the south-east corner (signalling a potential use of the model in identifying vulnerable locations for future monitoring programmes). This distinction is not apparent in the monitoring data with all monitoring locations having at least one pesticide detected at some point (Figure 9). This pattern in the monitoring data is also reflected in a relative vulnerability map reported by Worrall and Besien (2005) and shown here as Figure 10. Possible explanations for the discrepancy include that (i) the model does not account explicitly for flow through fractures in the Chalk, and (ii) flow within the groundwater will be lateral and very rapid in places due to the karstic nature of the Chalk. The overlying Paleogene cover will have a strong control on the karstic


development of the Chalk as well as the degree of vertical transport and hence groundwater vulnerability. This means that differences in vulnerability due to unsaturated zone thickness/travel times are not reflected in the groundwater sampling programme.

Figure 9. The Stevenage study area showing monitoring sites and the number of compounds detected on any sampling occasion at each; zero compounds detected (white), one compound (yellow), two compounds (orange), three compounds (blue), four compounds (green).

Figure 10. Results of vulnerability calculations based on measured monitoring data and reported by Worrall and Besien (2005). Vulnerability is a relative index based on the likelihood of detecting pesticides at a specific location relative to the average likelihood across all monitoring locations. The Stevenage case study area is shown by the bold square.


4.2.2 Ranking of frequency of detection

Table 5 summarises the relative frequency with which pesticides were detected during monitoring in the two case study areas. Stevenage is clearly more vulnerable than Doncaster based on the monitoring data that are available. This is not well matched by the model which predicts roughly equal vulnerability in the two locations. There are relatively small differences in intensity of land use for the two locations, so it is most likely that the additional vulnerability in the Stevenage area results from fissure flow bypassing the Chalk matrix and that this process is not sufficiently well represented by the average travel times used as input to the model. A further possible factor is differences in overlying soils within the two areas. There is a much greater predominance of silty and clayey soils in the Stevenage area; some of these soils (Evesham. Hanslope) will be drained under arable cultivation with limited potential for bypass flow to depth, but soils of the Upton association are shallow silty soils overlying Chalk which are likely to be relatively vulnerable to pesticide leaching. Characteristics of the overlying soil layer were not incorporated into this proof-of-concept exercise.

In Stevenage, the frequency of detections is ranked atrazine > isoproturon > mecoprop > bentazone with relative detection rates of 100, 40.5, 32.4 and 16.2% of samples, respectively. The pesticide rank for Doncaster is similar with the exception of isoproturon. The rank order for Doncaster is atrazine > mecoprop > bentazone >isoproturon with relative detection rates of 23.7, 7.1, 3.2 and 0% of samples, respectively. The model predicts that bentazone, isoproturon and mecoprop have fairly similar behaviour both in terms of

number of grid cells with negligible leaching (concentration <0.01 µg/L; Section 4.1.1 and 4.1.2) and the

maximum concentration detected for the three compounds (Table 6). Overall, leaching is ranked by the model as bentazone > isoproturon > mecoprop. Atrazine is predicted to have the lowest vulnerability to leaching to groundwater in both areas; this is partially a reflection of soil leaching results whereby the PECgw value for atrazine is an order of magnitude smaller than those for the other compounds (Table 6). Much of the contamination of UK groundwaters by atrazine was attributed to non-agricultural usage including on railways and hard surface areas (e.g. Chilton et al., 2005). As the model does not include this usage and the hydrogeologic conditions do not match those within the model, the mismatch between predicted and actual frequency of detection for atrazine should be treated with considerable caution.

Table 6. Comparison between FOCUS PEARL simulations, modelling to the surface of the groundwater body and groundwater monitoring data for the two study areas and four pesticides. FOCUS-PEARL modelling values are the 80

th-percentile of 20 annual average values at 1-m depth. PEC

at the groundwater surface is the 80th-percentile of 20 annual values averaged across the full

study area. All concentrations are in µg/L.

Compound FOCUS 80th

percentile

PEC at groundwater surface based on

Monitoring data

DT50*10 DT50*100 Average a Max

Doncaster (sandstone)b

Atrazine 0.14 0.0003 0.0005 0.013 (4) 0.218

Bentazone 9.7 0.11 0.21 0.0003 (2) 0.0141

Isoproturon 2.1 0.08 0.26 0 (0) 0

Mecoprop 1.2 0.0003 0.10 0.011 (1) 0.305

Stevenage (chalk)c

Atrazine 0.14 0.00004 0.00005 0.033 (37) 0.5

Bentazone 9.7 0.012 0.019 0.0004 (6) 0.06

Isoproturon 2.1 0.009 0.016 0.0087 (15) 1.35

Mecoprop 1.2 0.006 0.008 0.0023 (12) 0.59

a The number in brackets represents the number of monitoring points where the compound was detected (concentration > detection

limit); the average is for all monitoring results, not just for the positive detections

b Area contains 12 monitoring points

c Area contains 37 monitoring points

4.2.3 Average concentration in groundwater

Figures 11 and 12 compare measured concentrations from groundwater monitoring with the concentration predicted by the model to reach the top of the groundwater body. Monitoring data in these figures are the average of all monitoring results for the study area in each year with data, thus providing a time series of


how concentrations change over time. A single predicted annual average concentration is shown for each pesticide and area; this is selected as the single largest annual average concentration in any year simulated.

Figure 11. Measured and predicted concentration of the study compounds in Doncaster. ______

Average measured annual concentrations from 12 monitoring points

- - - - Maximum predicted concentration

Figure 12. Measured and predicted concentration of the study compounds in Stevenage. ______

Average measured annual concentrations from 37 monitoring points - - - - Maximum predicted concentration


Atrazine concentrations detected within monitoring programmes were largest at the start of the period for both case study areas and have progressively declined over time. The model significantly underestimates the concentrations detected as described in Section 4.2.2. Average concentrations of bentazone and isoproturon within groundwater in Doncaster are extremely small and here the model prediction for maximum annual concentration is much larger than average detections. Bentazone and isoproturon have been detected at several sampling locations in the Stevenage area. The model prediction is larger than average measurements by ca. 200 and 50%, respectively. Finally, the maximum annual concentration predicted by the model falls within the range of average concentrations of mecoprop. There is a single year that exceeds the model prediction for Stevenage; the rapid increase in average mecoprop concentration in Doncaster arises from a single sampling point with consistent contamination (up to 0.3 µg/L) and is almost certainly associated with a point-source of pollution.

4.2.4 Comparison between FOCUS modelling and groundwater monitoring data

Table 6 shows that the 80th percentile annual average values from standard FOCUS modelling are much

larger than average concentrations of the four compounds in groundwater detected during monitoring campaigns. The closest match is for atrazine (measured average 4 to 10 times smaller than 80

th percentile

modelling value), but this contamination is known to result largely from non-agricultural uses.

In the Doncaster area, maximum concentrations of bentazone and isoproturon in any groundwater sample are much smaller (e.g. ca. three orders of magnitude for bentazone) than the 80

th percentile value derived

from FOCUS modelling. The maximum mecoprop concentration in Doncaster is almost certainly associated with point-source pollution and does not provide a reasonable point of comparison.

Maximum concentrations of pesticide detected in groundwater are generally larger in the Stevenage area. Here, the 80

th percentile modelling value is less than a factor of two larger than the measured maximum for

isoproturon and mecoprop, but is more than two orders of magnitude larger than the measured maximum for bentazone.

It should be noted that the FOCUS modelling did not incorporate any higher-tier refinements (e.g. availability of additional sorption and/or degradation studies, incorporation of time-dependent sorption), so provides a larger PECgw estimate than may be present in regulatory dossiers.

5 REGULATORY IMPLICATIONS AND RECOMMENDATIONS FOR FUTURE WORK

This project has demonstrated the use of a refined endpoint that may have application in protecting groundwater within regulatory risk assessment procedures for pesticides. The primary refinement of simulating annual average pesticide concentrations at the groundwater surface fits with the protection goal as assumed by EFSA (2009) of “safeguarding the drinking water function of the groundwater”; indeed, it can be argued that the refined endpoint is more in line with this protection goal than the established endpoint of annual average concentration leaching to 1-m depth, at least in areas with groundwater that is substantially below the plant root zone.

The approach applied in this project both extended the depth for leaching assessment to the surface of the groundwater and included a spatial aggregation whereby water leaching from beneath treated fields was mixed with water from untreated areas. Groundwater used for abstraction will draw water from a relatively large area and include a strong lateral transport component. It thus seems intuitive that some element of spatial mixing/aggregation is appropriate. The scale of mixing within the current project was largely operational, being based on the resolution of the pre-existing nitrate model (100 x 100 m grid cells).

Evaluation of simulations against monitoring data for the two case study areas gave mixed results. On the one hand, predicted concentrations were a better match against monitoring data than the standard FOCUS modelling except where groundwater contamination was known to arise from point sources and or non-agricultural use of pesticides. Set against this, there were clear differences between model simulations and monitoring data, particularly in the Stevenage area where the model suggested clear spatial differences in vulnerability that were not shown in the monitoring results. These differences arise in part from the relative simplicity of the modelling approach; they indicate the need for refinement of the modelling framework to better capture conditions of use and groundwater flow processes within the case study areas.

There are four main areas identified for future work:

(i) Whilst FOCUS-PEARL is an established, mechanistic model for pesticide leaching through soil, future work would need to link actual soil and weather conditions at the surface to the underlying hydrogeology. The FOCUS-PEARL simulations used here provided a useful benchmark against the existing regulatory approach, but consideration of true soil and weather conditions and how they vary in space and time should strengthen the realism of concentrations predicted to reach


groundwater.

(ii) The nitrate time-bomb model used for unsaturated zone simulations has a simplified flow regime which is based on average pore water velocity and ignores hydrodynamic dispersion; further research should investigate use of more refined modelling approaches for the unsaturated zone. The limited evaluation of simulations against monitoring data undertaken here suggests that transport of pesticide via fissure flow through chalk bedrock is a particular issue for consideration.

(iii) The issue of scale of simulation becomes increasingly important as the prediction endpoint moves from pesticide concentration at the base of the soil profile to pesticide concentration at the groundwater surface. A more comprehensive analysis is required to investigate this issue in detail and to recommend how best to capture the macro-scale heterogeneity in pesticide use at the surface that will impact on any residues leaching to groundwater.

(iv) Pesticide regulation operates at the European level with some possibilities for differentiation of assessment within the three assessment zones. Further development of groundwater modelling approaches should take place within this pan-European context. It would be helpful to include a wider range of hydrogeological case studies and to draw in partners from several Member States. A particular opportunity may be to engage the agrochemical industry into any further development of the concept as there are current monitoring programmes which could be deployed in anonymised form for validating the approach.


References to published material

9. This section should be used to record links (hypertext links where possible) or references to other published material generated by, or relating to this project.

Chilton, P. J., Stuart, M. E., Gooddy, D. C., Williams, R. J., & Johnson, A. C. (2005). Pesticide fate and behaviour in the UK Chalk aquifer, and implications for groundwater quality. Quarterly journal of engineering geology and hydrogeology, 38(1), 65-81.

EFSA Panel on Plant Protection Products (2010). Guidance for evaluating laboratory and field dissipation studies to obtain DegT50 values of plant protection products in soil. EFSA Journal 8(12):1936 [67 pp].

EFSA Panel on Plant Protection Products and their Residues (2013). ; Scientific Opinion on the report of the FOCUS groundwater working group (FOCUS, 2009): assessment of higher tiers. EFSA Journal 11(6):3291 [25pp].

FOCUS (2000). FOCUS groundwater scenarios in the EU review of active substances. Report of the FOCUS Groundwater Scenarios Workgroup, EC Document Reference SANCO/321/2000 rev.2, 202pp.

FOCUS (2009) “Assessing potential for movement of active substances and their metabolites to ground water in the EU” Report of the FOCUS Ground Water Work Group, EC Document Reference Sanco/13144/2010 version 1, 604 pp.

Gaus, I, Van de Casteele, K (2004). Assessing the contamination risk of five pesticides in a phreatic aquifer based on microcosm experiments and transport modelling at Sint-Jansteen (Zeeland, the Netherlands). Netherlands Journal of Geosciences-geologie En Mijnbouw 83(2):101–112.

Heron, G. and Christensen, T.H. (1992). Degradation of the herbicide Mecoprop in an aerobic aquifer determined by laboratory batch studies. Chemosphere, 24:547-557.

Johnson AC, White C and Bhardwaj CL (2000). Potential for isoproturon, atrazine and mecoprop to be degraded within a chalk aquifer. Journal of Contaminant Hydrology 44:1-18.

Lapworth, D. J., Gooddy, D. C. (2006). Source and persistence of pesticides in a semi-confined chalk aquifer of southeast England. Environmental Pollution, 144(3), 1031-1044.

Leonard, R.A., Knisel, W.G., 1988. Evaluating groundwater contamination potential from herbicide use. Weed Technology, 2(2), 207-216.

McMahon, P.G., Chapelle, F.H., Jagucki, M.L. (1992). Atrazine mineralization potential of alluvial-aquifer sediments under aerobic conditions. Environmental Science & Technology 26:1556–1559.

Milne, C.J., Kinniburgh, D.G., 2007. Geochemical properties of aquifers and other geological formations in the UK. British Geological Survey Commissioned Report, CR/06/216N Environment Agency Science Report SC030110/SR. 246pp.

Rao, R.S.C., Davidson, J.M., 1985. Indices for ranking the potential for pesticide contamination of groundwater. Soil Crop Sci. Soc. FL Proc. 44, 1-8.

Wang, L., Stuart, M. E., Bloomfield, J. P., Butcher, A. S., Gooddy, D. C., McKenzie, A. A., and Williams, A. T., 2012. Prediction of the arrival of peak nitrate concentrations at the water table at the regional scale in Great Britain. Hydrological Processes, 26(2), 226-239.

Wang, L., Butcher, A.S., Stuart, M.E., Gooddy, D.C., Bloomfield, J.P., 2013. The nitrate time bomb: a numerical way to investigate nitrate storage and lag time in the unsaturated zone. Environmental Geochemistry and Health 35(5), 667-681.

Worrall, F., & Besien, T. (2005). The vulnerability of groundwater to pesticide contamination estimated directly from observations of presence or absence in wells. Journal of Hydrology, 303(1), 92-107.

Date post:	26-May-2018
Category:	Documents
Upload:	phamkhuong
View:	219 times
Download:	0 times

Evidence Project Final Report - Defra, UK - Science...

Documents