Search

In continuation of the annual national surveys of winter wheat, which began in 1970, samples from between 250 and 350 randomly selected wheat crops in England and Wales between 1999 to 2019 were visually assessed for disease symptoms during the milky ripe development stages (GS 73–75). Septoria tritici blotch was the most prevalent and severe foliar disease each year, although annual levels fluctuated considerably and there was no overall significant change over the two decades. Incidence of brown rust, yellow rust, take‐all, and barley yellow dwarf virus also showed no significant overall change during the survey period, whereas glume blotch, powdery mildew, eyespot, and sharp eyespot all showed significant decline. Fusarium ear blight has significantly increased in both incidence and severity, causing a serious epidemic in 2012, when 96% of crops were affected. Tan spot has been the third most prevalent foliar disease since 2009 although severity is still very low. Regional disease levels were consistent over the two decades, providing reliable baselines to measure changes in seasonal disease severity. There were significant changes in agronomic practice with a rise to predominance of minimum tillage over the use of ploughing, decreases in overall disease susceptibility of cultivars grown, a major increase in the use of oilseed rape in the rotation and a long‐term trend towards earlier sowing. Fungicide use increased considerably, with over 98% crops sprayed and an average of 3.5 applications made per crop each year since 2014. Implications of changes over the last 21 years are discussed. [ABSTRACT FROM AUTHOR] Copyright of Plant Pathology is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

The BREAM (Bystander and Resident Exposure Assessment Model) (Kennedy et al. in BREAM: A probabilistic bystander and resident exposure assessment model of spray drift from an agricultural boom sprayer. Comput Electron Agric 2012;88:63-71) for bystander and resident exposure to spray drift from boom sprayers has recently been incorporated into the European Food Safety Authority (EFSA) guidance for determining non-dietary exposures of humans to plant protection products. The component of BREAM, which relates airborne spray concentrations to bystander and resident dermal exposure, has been reviewed to identify whether it is possible to improve this and its description of variability captured in the model. Two approaches have been explored: a more rigorous statistical analysis of the empirical data and a semi-mechanistic model based on established studies combined with new data obtained in a wind tunnel. A statistical comparison between field data and model outputs was used to determine which approach gave the better prediction of exposures. The semi-mechanistic approach gave the better prediction of experimental data and resulted in a reduction in the proposed regulatory values for the 75th and 95th percentiles of the exposure distribution.

Background: Reducing the risks and impacts of pesticide use on human health and on the environment is one of the objectives of the European Commission Directive 2009/128/EC in the quest for a sustainable use of pesticides. This Directive, developed through European national plans such as Ecophyto plan in France, promotes the introduction of innovative cropping systems relying, for example, on integrated pest management. Risk assessment for human health of the overall pesticide use in these innovative systems is required before the introduction of those systems to avoid that an innovation becomes a new problem. Objectives: The objectives of this work were to assess and to compare (1) the human exposure to pesticides used in conventional and innovative cropping systems designed to reduce pesticide needs, and (2) the corresponding risks for human health. Methods: Humans (operator and residents) exposure to pesticides and risks for human health were assessed for each pesticide with the BROWSE model. Then, a method was proposed to represent the overall risk due to all pesticides used in one system. This study considers 3 conventional and 9 associated innovative cropping systems, and 116 plant protection products containing 89 different active substances (i.e. pesticides). Results: The modelling results obtained with BROWSE showed that innovative cropping systems such as low input or no herbicide systems would reduce the risk for human health in comparison to the corresponding conventional cropping systems. On the contrary, BROWSE showed that conservation tillage system would lead to unacceptable risks in the conditions of our study, because of a high number of pesticide applications, and especially of some herbicides. For residents, the dermal absorption was the main exposure route while ingestion was found to be negligible. For operators, inhalation was also a predominant route of exposure. In general, human exposure to pesticides and human health risks were found to be correlated to the treatment frequency index TFI (number of registered doses of pesticides used per hectare for one copping season), confirming the relationship between the reduction of pesticide use and the reduction of risks. Conclusions: Assessment with the BROWSE model helped to identify cropping systems with decreased risks from pesticides for human health and to propose some improvements to the cropping systems by identifying the pesticides that led to unacceptable risks. Keywords: BROWSE model, Human exposure, Plant protection products, Risk assessment, Cropping system

Mixtures of substances to which humans are exposed may lead to cumulative exposure and health effects. To study their effects, it is first necessary to identify a cumulative assessment group (CAG) of substances for risk assessment or hazard testing. Excluding substances from consideration before there is sufficient evidence may underestimate the risk. Conversely, including everything and treating the inevitable uncertainties using conservative assumptions is inefficient and may overestimate the risk, with an unknown level of protection. An efficient, transparent strategy is described to retain a large group, quantifying the uncertainty of group membership and other uncertainties. Iterative refinement of the CAG then focuses on adding information for the substances with high probability of contributing significantly to the risk. Probabilities can be estimated using expert opinion or derived from data on substance properties. An example is presented with 100 pesticides, in which the retain step identified a single substance to target refinement. Using an updated hazard characterisation for this substance reduced the mean exposure estimate from 0.43 to 0.28 μg kg-bw−1 day−1 and reduced the 99.99th percentile exposure from 24.9 to 5.1 μg kg-bw−1 day−1. Other retained substances contributed little to the risk estimates, even after accounting for uncertainty.

International audience; A model and data toolbox is presented to assess risks from combined exposure to multiple chemicals using probabilistic methods. The Monte Carlo Risk Assessment (MCRA) toolbox, also known as the EuroMix toolbox, has more than 40 modules addressing all areas of risk assessment, and includes a data repository with data collected in the EuroMix project. This paper gives an introduction to the toolbox and illustrates its use with examples from the EuroMix project. The toolbox can be used for hazard identification, hazard characterisation, exposure assessment and risk characterisation. Examples for hazard identification are selection of substances relevant for a specific adverse outcome based on adverse outcome pathways and QSAR models. Examples for hazard characterisation are calculation of benchmark doses and relative potency factors with uncertainty from dose response data, and use of kinetic models to perform in vitro to in vivo extrapolation. Examples for exposure assessment are assessing cumulative exposure at external or internal level, where the latter option is needed when dietary and non-dietary routes have to be aggregated. Finally, risk characterisation is illustrated by calculation and display of the margin of exposure for single substances and for the cumulation, including uncertainties derived from exposure and hazard characterisation estimates.

Populations are exposed to mixtures of pesticides through their diet on a daily basis. The question of which substances should be assessed together remains a major challenge due to the complexity of the mixtures. In addition, the associated risk is difficult to characterise. The EuroMix project (European Test and Risk Assessment Strategies for Mixtures) has developed a strategy for mixture risk assessment. In particular, it has proposed a methodology that combines exposures and hazard information to identify relevant mixtures of chemicals belonging to any cumulative assessment group (CAG) to which the European population is exposed via food. For the purposes of this study, food consumption and pesticide residue data in food and drinking water were obtained from national surveys in nine European countries. Mixtures of pesticides were identified by a sparse non-negative matrix underestimation (SNMU) applied to the specific liver steatosis effect in children from 11 to 15 years of age, and in adults from 18 to 64 years of age in nine European countries. Exposures and mixtures of 144 pesticides were evaluated through four different scenarios: (1) chronic exposure with a merged concentration dataset in the adult population, (2) chronic exposure with country-specific concentration datasets in the adult population, (3) acute exposure with a merged concentration dataset in the adult population, and (4) chronic exposure with a merged concentration dataset in the paediatric population. The relative potency factors of each substance were calculated to express their potency relative to flusilazole, which was chosen as the reference compound. The selection of mixtures and the evaluation of exposures for each country were carried out using the Monte Carlo Risk Assessment (MCRA) software. Concerning chronic exposure, one mixture explained the largest proportion of the total variance for each country, while in acute exposure, several mixtures were often involved. The results showed that there were 15 main pesticides in the mixtures, with a high contribution of imazalil and dithiocarbamate. Since the concentrations provided by the different countries were merged in the scenario using merged concentration data, differences between countries result from differences in food consumption behaviours. These results support the approach that using merged concentration data to estimate exposures in Europe seems to be realistic, as foods are traded across European borders. The originality of the proposed approach was to start from a CAG and to integrate information from combined exposures to identify a refined list of mixtures with fewer components. As this approach was sensitive to the input data and required significant resources, efforts should continue regarding data collection and harmonisation among the different aspects within the pesticides regulatory framework, and to develop methods to group substances and mixtures to characterise the risk.

Risk assessments for pesticides and other plant protection products (PPPs) often use models to estimate the exposure of different population groups. These range from conservative, deterministic calculations to more complex simulation approaches. In a recent EU project (Browse) new models were developed including exposure scenarios for bystanders, residents, operators and workers. These were implemented alongside a software interface and included probability distributions to capture the variation of possible exposures within the modelled subpopulations. The software interface was designed to allow users to input either fixed (e.g. conservative) values, for various parameters, or to specify predefined probability distributions for those inputs. Some default choices are available based on internationally agreed defaults for risk assessments. The software can also be run using a batch mode and the outputs can be presented and exported in different ways to facilitate their use in subsequent studies. The probabilistic models for resident and bystander exposure are described. Orchard and arable cropping scenarios are included, for both long-term and short-term exposures. Mathematical details are presented alongside practical information for using the associated software interface for probabilistic analyses. Example outputs are shown and sources of simulation error are quantified.

Risk assessment of chemicals occurring in our diet is commonly performed for single chemicals without considering exposure to other chemicals. We performed a case study on risk assessment of combined dietary exposure to chemicals from different regulatory silos, i.e. pesticides (PPRs), persistent organic pollutants (POPs) and food additives (FAs). Chemicals were grouped into the cumulative assessment group (CAG) liver steatosis using a component-based approach. Based on literature, the CAG included 144 PPRs, 49 POPS and 7 FAs for which concentration data were available. For each silo, chronic combined dietary exposure was assessed for adults and children of nine European countries following the most commonly used exposure methodologies in Europe and by using a relative potency factor approach. For risk characterization, a Margin of Exposure (MOE) was calculated. To overarch the risk across silos, a normalised combined margin of exposure (nMOET) approach was proposed. This case study demonstrated that risk assessment of combined exposure to chemicals can be performed within regulatory silos. It also highlighted important differences in the conservatism of exposure scenarios, the derivation of point of departures and the subsequent acceptable MOEs between the silos. To overarch the risk despite these differences, a nMOET approach can be used.

Human exposure to pesticide mixtures can occur from the diet and other sources. Realistic exposure and risk assessments should include multiple sources and compounds and include the relative hazards of the different compounds. The EU-funded Euromix project is developing new web-based tools to facilitate these calculations. A case study is presented that exemplifies their use for a population of UK residents, including exposure from crop-spraying. A UK pesticide usage survey provided information on real pesticide combinations applied to crops of wheat, potatoes, sugar beet and dessert apples. This information was combined with outputs from two alternative simulation models of spray drift to estimate dermal, oral and inhalation exposures of residents. These non-dietary exposures were combined with dietary exposure estimates using the Monte Carlo Risk Assessment software to produce a distribution of aggregated and cumulative exposures. Compounds are weighted by relative potency to generate a measure of overall risk. Uncertainty quantification was also included in the distribution of exposures. These tools are flexible to allow diverse sources of exposure and can provide important information to decision-makers and help to prioritise testing of pesticide mixtures. Including non-dietary sources changed the prioritisation of pesticide mixtures, when compared to dietary exposure alone.

This article discusses the use of Bayesian methods in analysing data that evolve over time in sequential multilocation auditing. Using the New York food stamps program as a case study, it proposes a model that incorporates a nonparametric component for the error magnitudes (taints), a hierarchical model for overall error rates across counties and parameters controlling the variation of rates from one year to the next, including an overall trend in error rates. The article first provides an overview of the New York food stamps program, along with the auditing concepts and terminology, before introducing the Bayesian model. This model is used to examine a sample of individual awards of food stamps to see if the value awarded is correct according to the rules of the scheme. The model makes it possible to smooth estimation of error rates and error classes in small counties across counties and through time.

Uncertainty analysis is an important component of dietary exposure assessments in order to understand correctly the strength and limits of its results. Often, standard screening procedures are applied in a first step which results in conservative estimates. If through those screening procedures a potential exceedance of health-based guidance values is indicated, within the tiered approach more refined models are applied. However, the sources and types of uncertainties in deterministic and probabilistic models can vary or differ. A key objective of this work has been the mapping of different sources and types of uncertainties to better understand how to best use uncertainty analysis to generate more realistic comprehension of dietary exposure. In dietary exposure assessments, uncertainties can be introduced by knowledge gaps about the exposure scenario, parameter and the model itself. With this mapping, general and model-independent uncertainties have been identified and described, as well as those which can be introduced and influenced by the specific model during the tiered approach. This analysis identifies that there are general uncertainties common to point estimates (screening or deterministic methods) and probabilistic exposure assessment methods. To provide further clarity, general sources of uncertainty affecting many dietary exposure assessments should be separated from model-specific uncertainties.

In today’s changing climate, the development of robust, accurate and globally applicable models is imperative for a wider understanding of Earth’s terrestrial biosphere. Moreover, an understanding of the representation, sensitivity and coherence of such models are vital for the operationalisation of any physically based model. A Global Sensitivity Analysis (GSA) was conducted on the SimSphere land biosphere model in which a meta-modelling method adopting Bayesian theory was implemented. Initially, effects of assuming uniform probability distribution functions (PDFs) for the model inputs, when examining sensitivity of key quantities simulated by SimSphere at different output times, were examined. The development of topographic model input parameters (e.g., slope, aspect, and elevation) were derived within a Geographic Information System (GIS) before implementation within the model. The effect of time of the simulation on the sensitivity of previously examined outputs was also analysed. Results showed that simulated outputs were significantly influenced by changes in topographic input parameters, fractional vegetation cover, vegetation height and surface moisture availability in agreement with previous studies. Time of model output simulation had a significant influence on the absolute values of the output variance decomposition, but it did not seem to change the relative importance of each input parameter. Sensitivity Analysis (SA) results of the newly modelled outputs allowed identification of the most responsive model inputs and interactions. Our study presents an important step forward in SimSphere verification given the increasing interest in its use both as an independent modelling and educational tool. Furthermore, this study is very timely given on-going efforts towards the development of operational products based on the synergy of SimSphere with Earth Observation (EO) data. In this context, results also provide additional support for the potential applicability of the assimilation of spatial analysis data derived from GIS and EO data into an accurate modelling framework.

The need for improved tools to estimate the cumulative and aggregate exposure to compounds such as plant protection products (PPPs) is recognised in the EU Regulation 1107/2009. A new model has been developed to estimate the exposure within a population to single compounds or compounds within a Cumulative Action Group, considering dietary and non-dietary sources and multiple exposure routes. To test the model a field study was carried out in Italy with operators applying tebuconazole fungicides, with measurements of dermal exposure collected. Whole urine samples were collected and analysed to provide values for the absorbed dose of tebuconazole, with duplicate diet samples collected and analysed as a measure of dietary exposures. The model provided predicted values of exposure for combined dietary and non-dietary routes of exposures which were compared to the measured absorbed dose values based on urinary analysis. The model outputs provided mean daily exposure values of 1.77 (± 1.96) µg a.s./kg BW which are comparable to measured mean values from the biomonitoring field study of 1.73 (± 1.31) µg a.s./kg BW. To supplement the limited measurement data available, comparisons against other models were also made and found to be comparable.

A previous publication described methods for assessing and reporting uncertainty in dietary exposure assessments. This follow-up publication uses a case study to develop proposals for representing and communicating uncertainty to risk managers. The food ingredient aspartame is used as the case study in a simple deterministic model (the EFSA FAIM template) and with more sophisticated probabilistic exposure assessment software (FACET). Parameter and model uncertainties are identified for each modelling approach and tabulated. The relative importance of each source of uncertainty is then evaluated using a semi-quantitative scale and the results expressed using two different forms of graphical summary. The value of this approach in expressing uncertainties in a manner that is relevant to the exposure assessment and useful to risk managers is then discussed. It was observed that the majority of uncertainties are often associated with data sources rather than the model itself. However, differences in modelling methods can have the greatest impact on uncertainties overall, particularly when the underlying data are the same. It was concluded that improved methods for communicating uncertainties for risk management is the research area where the greatest amount of effort is suggested to be placed in future.

A Total Diet Study (TDS) consists of selecting, collecting and analysing commonly consumed foods to obtain concentration data of different chemical compounds in foods as eaten. A TDS food list summarises the most consumed foods and represents the dietary habits of the general population of the country under study. The work reported here investigated whether TDS food lists that were initially designed for the whole population of the country under study also sufficiently cover the dietary pattern of specific subpopulations that are extra vulnerable for certain contaminants. The work was performed using data of three European countries: the Czech Republic, France and the UK. Each national food consumption database was combined with the corresponding national TDS food list (containing 336, 212 and 119 food items for the Czech Republic, France and the UK, respectively). The data were aggregated on the highest level of hierarchy of FoodEx-1, a pan-European food classification system, including 20 main FoodEx-1 groups. For the group 'milk and dairy products', the coverage of the consumption by the food list was investigated for more refined subgroups. For each food group or subgroup and country, the average percentage of coverage of the diet by the national TDS food list was calculated for different subpopulations, including children versus adults, women versus men, vegetarians versus non-vegetarians, and women of child-bearing age versus older women. The average diet of the different subpopulations was sufficiently covered by the food list of the Czech Republic and France. For the UK the average coverage was low due to a different food-coding approach and because food lists were not derived directly from national food consumption data. At the level of the 20 main food groups, differences between the subpopulations with respect to the average coverage of consumption by the TDS food list were minimal. The differences were more pronounced when looking in detail at the coverage of the dairy consumption. TDS food lists based on the mean consumption of the general population are also applicable to study the chemical exposure of different subpopulations, e.g. children, women of child-bearing age and vegetarians. This lowers the effort when performing a TDS.

The EU project BRAFO proposed a framework for risk–benefit assessment of foods, or changes in diet, that present both potential risks and potential benefits to consumers ( Hoekstra et al., 2012a ). In higher tiers of the BRAFO framework, risks and benefits are integrated quantitatively to estimate net health impact measured in DALYs or QALYs (disability- or quality-adjusted life years). This paper describes a general model that was developed by a second EU project, Qalibra, to assist users in conducting these assessments. Its flexible design makes it applicable to a wide range of dietary questions involving different nutrients, contaminants and health effects. Account can be taken of variation between consumers in their diets and also other characteristics relevant to the estimation of risk and benefit, such as body weight, gender and age. Uncertainty in any input parameter may be quantified probabilistically, using probability distributions, or deterministically by repeating the assessment with alternative assumptions. Uncertainties that are not quantified should be evaluated qualitatively. Outputs produced by the model are illustrated using results from a simple assessment of fish consumption. More detailed case studies on oily fish and phytosterols are presented in companion papers. The model can be accessed as web-based software at www.qalibra.eu.

The fish ingredient N3-docosahexaenoic acid 22:6 n-3 (DHA) stimulates brain development. On the other hand methylmercury (MeHg) in fish disturbs the developing central nervous system. In this Context the IQ score in children is considered as an aggregate measure of in utero brain development. To determine the effect of DHA exposure on prenatal neurodevelopment the maternal DHA intake during pregnancy was compared with its epidemiologically observed effect on the IQ score of children. For MeHg the maternal intake was converted into its accumulation in the maternal body. The maternal body burden then was compared with its epidemiologically observed relationship with the IQ score. Taking the MeHg and DHA content of 33 fish species the net effect of these compounds on the IQ score was quantified. For most fish species the adverse effect of MeHg on the IQ score exceeded the beneficial effect of DHA. In the case of long-living predators a negative effect up to 10 points on the IQ score was found. The results of this study indicate that food interventions aiming at the beneficial effects of fish consumption should focus on fish species with a high DHA content, while avoiding fish species with a high MeHg content.

New models have been developed, with the aim of improving the estimate of exposure of residents and bystanders to agricultural pesticides for regulatory purposes. These are part of a larger suite of models also covering operators and workers. The population that is modelled for residents and bystanders relates to people (both adults and children) who have no association with the application (i.e. not occupational exposure) but are adjacent to the treated area during and/or after the application process. The scenarios that the models aim to describe are based on consideration of both best practice and of real practice, as shown in surveys and from expert knowledge obtained in stakeholder consultations. The work has focused on three causes of exposure identified as having potential for improvement: boom sprayers, orchard sprayers and vapour emissions. An overview of the models is given, and a description of model input values and proposed defaults. The main causes of uncertainty in the models are also discussed. There are a number of benefits of the BROWSE model over current models of bystander and resident exposure, which includes the incorporation of mitigation measures for reducing exposure and the use of probabilistic modelling to avoid an over-conservative approach. It is expected that the levels of exposure that the BROWSE model predicts will, in some cases, be higher than those predicted by the current UK regulatory model. This is largely because the modelled scenarios have been updated to account for current practice and current scientific knowledge.

Computer simulation models are widely used in the study of real-world processes, and those that are associated with Earth observation (EO) are increasingly complex. To correctly integrate these models with observation data, it is important to understand the relative importance of individual parameters. For example, identifying the key driving parameters could dramatically reduce the costs and efforts in terms of data acquisition. The freely available software tool Gaussian Emulation for Sensitivity Analysis (GEM-SA) is designed to handle complex models and efficiently explore large parameter spaces, without requiring many thousands of simulation runs. This chapter provides an overview of GEM-SA functionalities for employing sensitivity analysis, and the practical use of the tool is illustrated using the SimSphere land biosphere model, which is a model that is used synergistically with EO data for deriving spatiotemporal estimates of key parameters characterizing Earth's water cycle. In this example, it is shown how using only 400 model simulation runs, the most important model inputs can be robustly identified and how each of the main effect relationships can be mapped. Such information can be of key importance to the SimSphere user's community where this model is used either as a stand-Ealone tool or synergistically with EO data.

A new suite of models has been developed for assessing the exposure of bystanders, residents, operators and workers to pesticides used in agricultural applications. The aim of these ‘BROWSE’ models was to improve regulatory exposure assessment by including recent data, and changes in current knowledge and application practice. The new models for bystander and resident exposure focused on spray drift from boom and orchard sprayers, and vapour emissions from treated crops. The structure of the resident and bystander models is reported elsewhere, together with a description and discussion of model inputs. This paper describes model outputs, model validation where experimental data are available, a sensitivity analysis for some model components and compares predictions with other European regulatory models, including the recently developed European Food Safety Authority (EFSA) calculator. Because BROWSE models can simulate a range of scenarios, there is a wide range of possible outputs. When using recommended default inputs, the resident and bystander models predict higher exposures than existing regulatory models. This is because the model incorporates more exposure routes and defaults are based on a reasonable worst case scenario. However, the probabilistic nature of the BROWSE models results in lower predicted exposures than the new EFSA calculator. Validation of sections of the model suggests that it is not overly conservative. Sensitivity analysis of the vapour exposure component showed that the interaction between temperature and vapour pressure, which is not taken account of in other models, can result in lower exposures for higher vapour pressures in warmer climates.

Complex simulation models are available to predict the level of exposure to bystanders and residents after a crop spraying event. In this paper we consider a particle-tracking spray drift model whose input parameters define particular scenarios of interest. Model outputs based on fixed values for these inputs ignore natural random variation and therefore give no indication of realistic variation in exposures, nor do they quantify the probability of rare extreme exposures. We describe a probabilistic modelling framework that allows the effect of variability in the input parameters to be quantified. An efficient statistical method for approximating the spray drift model is used, by creating a statistical emulator. An additional statistical model is then used to link airborne spray outputs to bystander exposures based on measured data. Uncertainty and variability are quantified in this model component. Validation of our approach is considered in two stages: first the accuracy of the emulator is assessed, as a surrogate for the true spray model. Secondly, the overall probabilistic outputs are compared with corresponding field measurements. Results are presented for a selection of typical exposure risk scenarios for bystanders and residents, illustrating the potential to generate a richer source of information for decision-makers. Sensitivity analysis results suggest strategies to reduce risk, such as minimising boom height.

We consider the use of emulator technology as an alternative method to second-order Monte Carlo (2DMC) in the uncertainty analysis for a percentile from the output of a stochastic model. 2DMC is a technique that uses repeated sampling in order to make inferences on the uncertainty and variability in a model output. The conventional 2DMC approach can often be highly computational, making methods for uncertainty and sensitivity analysis unfeasible. We explore the adequacy and efficiency of the emulation approach, and we find that emulation provides a viable alternative in this situation. We demonstrate these methods using two different examples of different input dimensions, including an application that considers contamination in pre-pasteurised milk.

The number of residue measurements in an individual field trial, carried out to provide data for a pesticide registration for a particular crop, is generally too small to estimate upper tails of the residue distribution for that crop with any certainty. We present a new method, using extreme value theory, which pools information from various field trials, with different crop and pesticide combinations, to provide a common model for the upper tails of residue distributions generally. The method can be used to improve the estimation of high quantiles of a particular residue distribution. It provides a flexible alternative to the direct fitting of a distribution to each individual dataset, and does not require strong distributional assumptions. By using a hierarchical Bayesian model, our method also accounts for parameter uncertainty. The method is applied to a range of supervised trials containing residues on individual items (e.g. on individual apples), and the results illustrate the variation in tail properties amongst all commodities and pesticides. The outputs could be used to select conservative high percentile residue levels as part of a deterministic risk assessment, taking account of the variability between crops and pesticides and also the uncertainty due to relatively small datasets.

The CRASH computer model simulates the effect of a vehicle colliding against different barrier types. If it accurately represents real vehicle crashworthiness, the computer model can be of great value in various aspects of vehicle design, such as the setting of timing of air bag releases. The goal of this study is to address the problem of validating the computer model for such design goals, based on utilizing computer model runs and experimental data from real crashes. This task is complicated by the fact that (i) the output of this model consists of smooth functional data, and (ii) certain types of collision have very limited data. We address problem (i) by extending existing Gaussian process-based methodology developed for models that produce real-valued output, and resort to Bayesian hierarchical modeling to attack problem (ii). Additionally, we show how to formally test if the computer model reproduces reality. Supplemental materials for the article are available online.

SummaryA crucial issue in the current global warming debate is the effect of vegetation and soils on carbon dioxide (CO2) concentrations in the atmosphere. Vegetation can extract CO2 through photosynthesis, but respiration, decay of soil organic matter and disturbance effects such as fire return it to the atmosphere. The balance of these processes is the net carbon flux. To estimate the biospheric carbon flux for England and Wales, we address the statistical problem of inference for the sum of multiple outputs from a complex deterministic computer code whose input parameters are uncertain. The code is a process model which simulates the carbon dynamics of vegetation and soils, including the amount of carbon that is stored as a result of photosynthesis and the amount that is returned to the atmosphere through respiration. The aggregation of outputs corresponding to multiple sites and types of vegetation in a region gives an estimate of the total carbon flux for that region over a period of time. Expert prior opinions are elicited for marginal uncertainty about the relevant input parameters and for correlations of inputs between sites. A Gaussian process model is used to build emulators of the multiple code outputs and Bayesian uncertainty analysis is then used to propagate uncertainty in the input parameters through to uncertainty on the aggregated output. Numerical results are presented for England and Wales in the year 2000. It is estimated that vegetation and soils in England and Wales constituted a net sink of 7.55 Mt C (1 Mt C = 1012 g of carbon) in 2000, with standard deviation 0.56 Mt C resulting from the sources of uncertainty that are considered.

We develop a statistical approach for characterizing uncertainty in predictions that are made with the aid of a computer simulation model. Typically, the computer simulation code models a physical system and requires a set of inputs---some known and specified, others unknown. A limited amount of field data from the true physical system is available to inform us about the unknown inputs and also to inform us about the uncertainty that is associated with a simulation-based prediction. The approach given here allows for the following: uncertainty regarding model inputs (i.e., calibration); accounting for uncertainty due to limitations on the number of simulations that can be carried out; discrepancy between the simulation code and the actual physical system; uncertainty in the observation process that yields the actual field data on the true physical system. The resulting analysis yields predictions and their associated uncertainties while accounting for multiple sources of uncertainty. We use a Bayesian formulation and rely on Gaussian process models to model unknown functions of the model inputs. The estimation is carried out using a Markov chain Monte Carlo method. This methodology is applied to two examples: a charged particle accelerator and a spot welding process.

Pesticide risk assessment is hampered by worst-case assumptions leading to overly pessimistic assessments. On the other hand, cumulative health effects of similar pesticides are often not taken into account. This paper describes models and a web-based software system developed in the European research project ACROPOLIS. The models are appropriate for both acute and chronic exposure assessments of single compounds and of multiple compounds in cumulative assessment groups. The software system MCRA (Monte Carlo Risk Assessment) is available for stakeholders in pesticide risk assessment at mcra.rivm.nl. We describe the MCRA implementation of the methods as advised in the 2012 EFSA Guidance on probabilistic modelling, as well as more refined methods developed in the ACROPOLIS project. The emphasis is on cumulative assessments. Two approaches, sample-based and compound-based, are contrasted. It is shown that additional data on agricultural use of pesticides may give more realistic risk assessments. Examples are given of model and software validation of acute and chronic assessments, using both simulated data and comparisons against the previous release of MCRA and against the standard software DEEM-FCID used by the Environmental Protection Agency in the USA. It is shown that the EFSA Guidance pessimistic model may not always give an appropriate modelling of exposure.

Exposures to plant protection products (PPPs) are assessed using risk analysis methods to protect public health. Traditionally, single sources, such as food or individual occupational sources, have been addressed. In reality, individuals can be exposed simultaneously to multiple sources. Improved regulation therefore requires the development of new tools for estimating the population distribution of exposures aggregated within an individual. A new aggregate model is described, which allows individual users to include as much, or as little, information as is available or relevant for their particular scenario. Depending on the inputs provided by the user, the outputs can range from simple deterministic values through to probabilistic analyses including characterisations of variability and uncertainty. Exposures can be calculated for multiple compounds, routes and sources of exposure. The aggregate model links to the cumulative dietary exposure model developed in parallel and is implemented in the web-based software tool MCRA. Case studies are presented to illustrate the potential of this model, with inputs drawn from existing European data sources and models. These cover exposures to UK arable spray operators, Italian vineyard spray operators, Netherlands users of a consumer spray and UK bystanders/residents. The model could also be adapted to handle non-PPP compounds.

A Total Diet Study (TDS) consists of selecting, collecting and preparing commonly consumed foods purchased at retail level and analysing them for harmful and/or beneficial chemical substances. A food classification system is needed to link food consumption data with the contaminant concentration data obtained in the TDS for the exposure assessment. In this study a comparison was made between the use of a national food classification systems and the use of FoodEx-1, developed and recommended by the European Food Safety Authority (EFSA). The work was performed using data of six European countries: Belgium, Czech Republic, France, The Netherlands, Spain and the UK. For each population, exposure to contaminant A (organic compounds) and/or contaminant B (inorganic compound) was assessed by the Monte Carlo Risk Assessment (MCRA) software using the national classification system and FoodEx-1 for food consumption data and for TDS laboratory results. Minimal differences between both approaches were observed. This observation applied for both contaminant A and contaminant B. In general risk assessment will be similar for both approaches; however, this is not guaranteed. FoodEx-1 proved to be a valuable hierarchic classification system in order to harmonise exposure assessment based on existing TDS results throughout Europe.

Risk assessments for human exposures to plant protection products (PPPs) have traditionally focussed on single routes of exposure and single compounds. Extensions to estimate aggregate (multi-source) and cumulative (multi-compound) exposure from PPPs present many new challenges and additional uncertainties that should be addressed as part of risk analysis and decision-making. A general approach is outlined for identifying and classifying the relevant uncertainties and variabilities. The implementation of uncertainty analysis within the MCRA software, developed as part of the EU-funded ACROPOLIS project to address some of these uncertainties, is demonstrated. An example is presented for dietary and non-dietary exposures to the triazole class of compounds. This demonstrates the chaining of models, linking variability and uncertainty generated from an external model for bystander exposure with variability and uncertainty in MCRA dietary exposure assessments. A new method is also presented for combining pesticide usage survey information with limited residue monitoring data, to address non-detect uncertainty. The results show that incorporating usage information reduces uncertainty in parameters of the residue distribution but that in this case quantifying uncertainty is not a priority, at least for UK grown crops. A general discussion of alternative approaches to treat uncertainty, either quantitatively or qualitatively, is included.

We consider Bayesian techniques for estimating distribution functions and quantile functions from an intractable posterior density function f(·), which are efficient in the sense that relatively few evaluations of f(·) are required. This Bayesian quadrature approach is therefore appropriate when each function evaluation is very expensive. A Gaussian process model is assumed for the prior distribution of f(·), from which we can derive the posterior for f(·) and its integrals. Efficiency is achieved by incorporating prior information about the density function. In particular, we assume f(·) is reasonably well approximated by a normal density. The Bayesian model-based approach also allows us to select optimal evaluation points, and we use this to find an optimal set for estimating univariate distribution functions. If we have the additional information that f(·) integrates to 1, then we show how this information can also be modeled, and can produce more accurate results. The method is demonstrated by integrating a range of density functions with known distribution functions.

We consider an efficient Bayesian approach to estimating integration-based posterior summaries from a separate Bayesian application. In Bayesian quadrature we model an intractable posterior density function f(·) as a Gaussian process, using an approximating function g(·), and find a posterior distribution for the integral of f(·), conditional on a few evaluations of f (·) at selected design points. Bayesian quadrature using normal g (·) is called Bayes-Hermite quadrature. We extend this theory by allowing g(·) to be chosen from two wider classes of functions. One is a family of skew densities and the other is the family of finite mixtures of normal densities. For the family of skew densities we describe an iterative updating procedure to select the most suitable approximation and apply the method to two simulated posterior density functions.

The European Food Safety Authority (EFSA) asked the Panel on Plant Protection Products and their Residues to provide guidance on methodology for performing probabilistic dietary exposure assessment of single or multiple active substances, as a potential additional tool to supplement or complement the standard deterministic methodologies which are currently used in the EU for conducting dietary exposure assessments for pesticides. Specific guidance is provided for basic assessments but not for refined assessments, where specialised expertise is required to select methods appropriate to the assessment in hand. The guidance includes probabilistic methods for quantifying some of the major sources of variability and uncertainty affecting dietary exposure to pesticides. Other important sources of variability and uncertainty might be quantified probabilistically in refined assessments but are addressed more simply in basic assessments by conducting alternative model runs with optimistic and pessimistic assumptions. Guidance is provided on problem formulation, including definition of appropriate scenarios for acute and chronic exposure assessment in the differing contexts of approval of new substances, MRL setting, authorisation of products, evaluation of residues found above the MRL, and annual reviews of residue monitoring data.

This paper presents the benefit–risk assessment of adding plant sterols to margarine as an illustration of the QALIBRA method and software. With the QALIBRA tool health effects, risks as well as benefits are expressed in a common metric (DALY) which allows quantitative balancing of benefits and risks of food intake. The QALIBRA software can handle uncertainties in a probabilistic simulation. This simple case study illustrates the data need and assumptions that go into a quantitative benefit–risk assessment. The assessment shows that the benefits of plant sterols added to margarine outweigh the risks, if any.

We propose new models for dealing with various sources of variability and uncertainty that influence risk assessments for dietary exposure. The uncertain or random variables involved can interact in complex ways, and the focus is on methodology for integrating their effects and on assessing the relative importance of including different uncertainty model components in the calculation of dietary exposures to contaminants, such as pesticide residues. The combined effect is reflected in the final inferences about the population of residues and subsequent exposure assessments. In particular, we show how measurement uncertainty can have a significant impact on results and discuss novel statistical options for modeling this uncertainty. The effect of measurement error is often ignored, perhaps due to the laboratory process conforming to the relevant international standards, for example, or is treated in an ad hoc way. These issues are common to many dietary risk analysis problems, and the methods could be applied to any food and chemical of interest. An example is presented using data on carbendazim in apples and consumption surveys of toddlers.

Human exposure to a specific pesticide or other chemical can occur from a combination of food and drink products. Probabilistic risk assessments are used to quantify the distribution of mean total daily exposures in the population, from the available data on residues and consumptions. We present a new statistical method for estimating this distribution, based on dietary survey data for multiple food types and residue monitoring data. The model allows for between-food correlations in both frequency and amounts of consumption. Three case studies are presented based on consumption data for UK children, considering the distribution of daily intakes of pyrimethanil, captan and chlorpyrifos aggregated over 4, 6 and 10 food types, respectively. We compared three alternative approaches, each using a Bayesian approach to quantify uncertainty: (i) a multivariate model that explicitly includes correlation parameters; (ii) separate independent parametric models for individual food types and (iii) a single parametric model applied to intakes aggregated directly from the data. The results demonstrate the importance of accounting for correlations between foods, using model (i) or (iii), for example, but also show that model (iii) can produce very different results when the aggregated intakes distribution is bimodal. The influence of residue uncertainty is also demonstrated.

We present a detailed study into the effects that uncertainty and variability have on the levels of dietary exposure of young children to carbendazim in a range of apple products, including raw apples, apple juice and processed apple foods. It is estimated that exposure is greater than the provisional EU acute reference dose of 0.02 mg/kg-bw/day on 0.29% of person days, although there is large uncertainty in the true number. The contribution from apple juice consumption appears to be significant. The issues presented are common to many dietary risk analysis problems, and the methods could be applied to any food group and chemical of interest.

Background The quantification of uncertainty and variability is a key component of quantitative risk analysis. Recent advances in Bayesian statistics make it ideal for integrating multiple sources of information, of different types and quality, and providing a realistic estimate of the combined uncertainty in the final risk estimates. Methods We present two case studies related to foodborne microbial risks. In the first, we combine models to describe the sequence of events resulting in illness from consumption of milk contaminated with VTEC O157. We used Monte Carlo simulation to propagate uncertainty in some of the inputs to computer models describing the farm and pasteurisation process. Resulting simulated contamination levels were then assigned to consumption events from a dietary survey. Finally we accounted for uncertainty in the dose-response relationship and uncertainty due to limited incidence data to derive uncertainty about yearly incidences of illness in young children. Options for altering the risk were considered by running the model with different hypothetical policy-driven exposure scenarios. In the second case study we illustrate an efficient Bayesian sensitivity analysis for identifying the most important parameters of a complex computer code that simulated VTEC O157 prevalence within a managed dairy herd. This was carried out in 2 stages, first to screen out the unimportant inputs, then to perform a more detailed analysis on the remaining inputs. The method works by building a Bayesian statistical approximation to the computer code using a number of known code input/output pairs (training runs). Results We estimated that the expected total number of children aged 1.5-4.5 who become ill due to VTEC O157 in milk is 8.6 per year, with 95% uncertainty interval (0,11.5). The most extreme policy we considered was banning on-farm pasteurisation of milk, which reduced the estimate to 6.4 with 95% interval (0,11). In the second case study the effective number of inputs was reduced from 30 to 7 in the screening stage, and just 2 inputs were found to explain 82.8% of the output variance. A combined total of 500 runs of the computer code were used. Conclusion These case studies illustrate the use of Bayesian statistics to perform detailed uncertainty and sensitivity analyses, integrating multiple information sources in a way that is both rigorous and efficient.