Colin D. Brown

Sources of uncertainty in pesticide fate modelling

There is worldwide interest in the application of probabilistic approaches to pesticide fate models to account for uncertainty in exposure assessments. The first steps in conducting a probabilistic analysis of any system are: (i) to identify where the uncertainties come from; and (ii) to pinpoint those uncertainties that are likely to affect most of the predictions made. This article aims at addressing those two points within the context of exposure assessment for pesticides through a review of the different sources of uncertainty in pesticide fate modelling. The extensive listing of sources of uncertainty clearly demonstrates that pesticide fate modelling is laced with uncertainty. More importantly, the review suggests that the probabilistic approaches, which are typically being deployed to account for uncertainty in the pesticide fate modelling, such as Monte Carlo modelling, ignore a number of key sources of uncertainty, which are likely to have a significant effect on the prediction of environmental concentrations for pesticides (e.g. model error, modeller subjectivity). Future research should concentrate on quantifying the impact these uncertainties have on exposure assessments and on developing procedures that enable their integration within probabilistic assessments.

A European test of pesticide-leaching models: methodology and major recommendations

Testing of pesticide-leaching models is important in view of their increasing use in pesticide registration procedures in the European Union. This paper presents the methodology and major conclusions of a test of pesticide-leaching models. Twelve models simulating the vertical one-dimensional movement of water, solute, heat, and, in particular, pesticides, through the soil profile were used by 36 different modellers. The adopted modelling codes differ in terms of modelling concepts and modelling hypothesis. Modellers were affiliated to industry and to the scientific community as well. Four quality datasets were identified to perform the analysis. The dataset included field and lysimeter data, collected in the Netherlands, Germany, Italy and the UK. As well, non-structured as structured soils were available in the dataset. To elucidate the ability to model correctly water transport, solute transport, heat transport and pesticide transport in soils, a stepwise evaluation approach was followed. Splitting up the experimental dataset enabled us to quantify the calibration capability and the prediction capability of the models. The simulations were performed by different model users enabling us also to characterise output variability in terms of user dependent interpretation of the model input and parameters. Recommendations are formulated for improving the quality of modelling datasets, and the process description of water, solute, and heat transport in a pesticide-leaching model, plus the process description of pesticide fate. Application of the principles of good modelling practice (GMP) is briefly described

Predicting effects on aquatic organisms from fluctuating or pulsed exposure to pesticides

Exposure of aquatic nontarget organisms to pesticides almost always occurs as pulses or fluctuating concentrations. Extrapolation from laboratory to field thus depends on an understanding and ability to simulate effects resulting from these types of exposure. This paper reviews models that may be used to predict effects on aquatic organisms resulting from time-varying exposure to pesticides. We evaluate and compare the theoretical basis of these models and their applicability to the simulation of effects from fluctuating exposures. The many different models rest on only a few basic concepts with differing degrees of mechanistic character. Building on this critical review, we select the most appropriate models and propose modifications. Two process-based models, the threshold hazard model and the modified damage assessment model, represent the optimum descriptions that are available at present. They could facilitate a better understanding of the ecotoxicity of different compound and species combinations and even mixtures of noninteracting compounds. The possibility to model lethal and sublethal effects allows applications in risk assessment, standard setting, and ecological modeling.

Pesticides in rainfall in Europe.

Papers and published reports investigating the presence of pesticides in rainfall in Europe were reviewed. Approximately half of the compounds that were analysed for were detected. For those detected, most concentrations were below about 100 ng/l, but larger concentrations, up to a few thousand nanograms per litre, were detected occasionally at most monitoring sites. The most frequently detected compounds were lindane (gamma-HCH) and its isomer (alpha-HCH), which were detected on 90-100% of sampling occasions at most of the sites where they were monitored. For compounds developed more recently, detection was usually limited to the spraying season. A classification of pesticides according to their deposition pattern is proposed.

Development of a framework based on an ecosystem services approach for deriving specific protection goals for environmental risk assessment of pesticides

General protection goals for the environmental risk assessment (ERA) of plant protection products are stated in European legislation but specific protection goals (SPGs) are often not precisely defined. These are however crucial for designing appropriate risk assessment schemes. The process followed by the Panel on Plant Protection Products and their Residues (PPR) of the European Food Safety Authority (EFSA) as well as examples of resulting SPGs obtained so far for environmental risk assessment (ERA) of pesticides is presented. The ecosystem services approach was used as an overarching concept for the development of SPGs, which will likely facilitate communication with stakeholders in general and risk managers in particular. It is proposed to develop SPG options for 7 key drivers for ecosystem services (microbes, algae, non target plants (aquatic and terrestrial), aquatic invertebrates, terrestrial non target arthropods including honeybees, terrestrial non-arthropod invertebrates, and vertebrates), covering the ecosystem services that could potentially be affected by the use of pesticides. These SPGs need to be defined in 6 dimensions: biological entity, attribute, magnitude, temporal and geographical scale of the effect, and the degree of certainty that the specified level of effect will not be exceeded. In general, to ensure ecosystem services, taxa representative for the key drivers identified need to be protected at the population level. However, for some vertebrates and species that have a protection status in legislation, protection may be at the individual level. To protect the provisioning and supporting services provided by microbes it may be sufficient to protect them at the functional group level. To protect biodiversity impacts need to be assessed at least at the scale of the watershed/landscape.

Adsorption of Ionisable Pesticides in Soils

Understanding the fate of a pesticide in soil is fundamental to the accurate assessment of its environmental behaviour and vital in ensuring the safe use of new and existing products. Ionisable pesticides comprise a significant proportion of both existing and new active substances registered for use in agriculture worldwide. This group of pesticides includes chemicals that are frequently found in groundwater and surface waters in many different countries. Despite this, approaches to predict the influence of soil properties on the behaviour of ionisable pesticides in soils are poorly developed. Current regulatory assessments frequently default to methods developed for nonionic chemicals, although it is evident that ionisable compounds do not often react like neutral molecules. This review presents the state of knowledge on the adsorption of ionisable pesticides in soils. It first introduces the issues concerning adsorption and the characteristics of this particular kind of chemical. The mechanisms postulated for the adsorption of ionisable pesticides are then described: these are hydrophobic partitioning, ionic exchange, charge transfer, ligand exchange, cation or water bridging, and the formation of bound residues. Relatively little experimental evidence is available, and we are still unable to determine the quantitative contribution of each process in a particular situation. Knowledge is still lacking concerning phenomena occurring at the surfaces of soil particles. Measurements do not allow determination of the operative pH at the surface of soil particles or in microenvironments, and the influence of ionic strength or competition effects is difficult to assess. Subsequently, the review focuses on the influence of soil properties on adsorption and on potential to predict the behaviour of ionisable pesticides in soils. Unlike hydrophobic compounds, adsorption of ionisable pesticides is highly sensitive to variation in pH. This relationship mainly derives from the different proportion of ionic and neutral forms of the pesticide present at each pH level but also from the presence of surfaces with pH-dependent charges in soils. Soil organic matter generally promotes adsorption, although a negative influence has sometimes been reported. Clay and oxides can also play a significant role in some cases. So far, no modelling approach has been applied successfully to a range of ionisable pesticides to predict their adsorption in soils. The standardization of experimental settings and the application of approaches specific to a particular class of pesticide or different type of soil might be necessary to describe the complexity of interactions among ionisable molecules. Degradation of ionisable pesticides is influenced by soil pH in a particular way that relates to changes in sorption, changes in composition and activity of the microbial community, and to shifts in the balance between different degradative mechanisms.

LogD: lipophilicity for ionisable compounds.

The octanol/water partition coefficient (Kow) for organic compounds is widely used in predictive environmental studies. A significant proportion of contaminants of surface and ground water are ionisable (e.g. many pesticides, pharmaceuticals, metabolites). Such compounds may be partially ionised dependent on the pH. Since the neutral and ionic species exhibit different polarities, the Kow value of ionisable pesticides is pH dependent. It is therefore essential to determine Kow values over the full range of pH that occurs in the environment in order to get appropriate predictors. Numerous methods are available to measure lipophilicity but only a few are appropriate for ionisable pesticides (e.g. pH metric and filter probe methods). Parameters such as pH and ionic strength need to be carefully controlled when working with ionisable compounds. Variation of these factors probably explains why literature can yield Kow values that differ by more than one order of magnitude for some compounds. In this article, Kow values obtained for six acidic pesticides with three different methods are compared as well (data from the literature, measured by pH metric method and calculated with five computer programs). The values used in predictive regression equations needs to be either measured with a suitable method or selected from the literature with great care.

Evaluation of methods to derive pesticide degradation parameters for regulatory modelling

Abstract. Models to simulate the fate of pesticides in the environment are frequently used for risk assessments within the registration process. An adequate description of pesticide degradation in soil is important to provide input for these models. Often, DT50 values (time required for 50% dissipation of the initial concentration) are used as model input, but there is no widely agreed methodology to derive DT50 values from experimental data. DT50 values are often obtained by fitting first-order kinetics to observed degradation patterns. The result depends on the handling of pesticide data (e.g. logarithmic transformation) and initial concentrations (variable or fixed). Kinetics other than first-order may be more suitable to describe the decline of measured concentrations, but the derived DT50 values are then not appropriate as input for many simulation models. Field or laboratory DT50 values can be used for modelling and this has consequences for model parameterisation. Degradation parameters derived from static laboratory experiments may not be applicable to pesticide behaviour under flow conditions in the field. Several methods to simulate the fate of metabolites and to evaluate experimental data are available. The methodology used to derive model input parameters must be consistent with the approach used within the simulation model.

Pesticide transport via sub-surface drains in Europe.

Transport of pesticides from point of application via sub-surface drains can contribute significantly to contamination of surface waters. Results of 23 field drainage experiments undertaken at sites across Europe were collated and analysed by residual maximum likelihood. Both maximum concentration of pesticide in drainflow (n = 167) and seasonal loss of pesticide to drains (n = 97) were significantly related to strength of pesticide sorption to soil, half-life of the pesticide in soil, the interval between application and first drainflow and the clay content of the soil. The statistical models accounted for 71% of the variability in both maximum concentration and seasonal load. Next, the dataset was used to evaluate the current methodology for assessment of aquatic exposure used in pesticide registration in Europe. Simulations for seven compounds with contrasting properties showed a good correspondence with field measurements. Finally, the review examines management approaches to reduce pesticide transport via sub-surface drains. Despite a large amount of work in this area, there are few dependable mitigation options other than to change application rate or timing or to restrict use of a compound in the most vulnerable situations.

Toxicodynamic assumptions in ecotoxicological hazard models.

Existing toxicokinetic and toxicodynamic models and dynamic formulations of popular ecotoxicological concepts (e.g., the critical body residue concept) are examined. Their underlying assumptions about speed of recovery and thresholds are clarified, and a rigorous mathematical treatment shows that they can all be placed within a unifying framework. Such analysis aids in the selection of appropriate ecotoxicological models.