EPPO BulletinVolume 33, Issue 2 p. 239-244 Free Access Environmental risk assessment scheme for plant protection products First published: 29 October 2003 https://doi.org/10.1046/j.1365-2338.2003.00641.xCitations: 3 European and Mediterranean Plant Protection OrganizationPP 3/13(1) Organisation Européenne et Méditerranéenne pour la Protection des Plantes AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat Chapter 12: Non-target terrestrial higher plants Specific scope This standard provides an assessment of risk presented by plant protection products to non-target terrestrial higher plants. Specific approval and amendment First approved in 2002-09. Introduction The subscheme in this chapter deals with the effects of plant protection products on non-target terrestrial higher plants. Non-target plants are defined as non-crop plants outside the treatment area. The assessment of the risk to crop plants is outside the scope of this subscheme; guidance on how to assess the phytotoxicity of plant protection products to crop species can be found in EPPO Standards PP/1/135(2) and PP/1/207(1). Risk assessment is required to help protect biodiversity both within the plant kingdom as a whole and in agro-ecosystems. Ideally, this subscheme should also provide an accurate categorization of the risk to non-crop plant species within the cropped area. However, for numerous reasons (which include lack of knowledge of weed threshold values, variation in composition and sensitivity of weed communities, the role of the seed bank, the impact of fertilizers and crop rotations), it is not currently possible to develop a readily useable subscheme for the ‘in crop’ scenarios. Since it is known that phytotoxic effects may occur for products other than herbicides, some level of testing of all plant protection products is necessary, unless it can be shown that exposure will not occur. The subscheme follows a sequential or tiered approach. Toxicity values are compared with predicted environmental concentrations to develop an exposure-to-toxicity ratio (ETR). The ETR is then compared with safety factors based on expert judgement or empirical derivation. Progression to the next tier is warranted if the safety margin is not met, or testing is stopped if the safety margin is met or exceeded. Notes 8 and 10 discuss safety margins, but expert judgement may be needed to decide suitable values for a given region. Research results (Fletcher, 1991) indicate that significant differences may exist among species in their response to plant protection products. While it would be ideal to develop dose–response data for a broad range of species of different genera and families, the cost of doing such studies under Good Laboratory Practice has to be considered. Furthermore, such extensive testing would be beyond that currently required in other areas of environmental risk assessment. Therefore, an alternative approach, using single-dose phytotoxicity screening data for terrestrial non-target plants (Aldridge et al., 1993; Boutin et al., 1993) at or above the maximum application rate on a wide range of species (e.g. at least six but preferably more), is recommended as a first tier to determine whether the plant protection product has phytotoxic properties. If phytotoxicity is observed, dose–response relationships for at least six species representing families for which significant herbicidal action has been found are recommended to quantify the level of effect using both soil and foliar exposure scenarios under tier 2. The subscheme contains decision points where expert judgement is specifically requested. This is important because, if the subscheme is to be useful throughout the EPPO region, it should incorporate sufficient flexibility to take into account regional differences in farming practices, climate, etc. It should not therefore be operated without the input of an ecotoxicologist and a plant protection expert. It is primarily designed to ensure consistency in the overall approach among different regulatory authorities. The subscheme can be used to assess both active substances and transformation products, if the necessary data is available. Figure 1 summarizes the subscheme in flow-chart format. The chart should not be used as a substitute for working through the subscheme itself. Figure 1Open in figure viewerPowerPoint Simplified diagram of the subscheme for the evaluation of the risk of plant protection products to higher plants. Risk assessment scheme Details of the products and its pattern of use 1 Take from Chapter 2 (Guidance on identifying aspects of environmental concern) the basic information on the product and its pattern of use: • mode of action (if known/understood) • uses (crop on which it is used, including, when relevant, growth stage) • formulation type • dosage and frequency • time of the year • vapour pressure • water solubilitygo to 2 Possibility of exposure Higher plants on non-target areas can be exposed to plant protection products via the air or by run-off. The major route of potential exposure is through drift (particulates that become air-borne during application). The other route through the air is the release of vapour from the applied product followed by gaseous transport and exposure, predominantly after application. These two routes of exposure are quantified in Chapter 3 (Air). Exposure by run-off is quantified in Chapter 4 (Soil). 2 Is the exposure of non-target terrestrial higher plants possible (see Note 1)? If no classify as negligible risk, go to 17 If yes go to 3 3 To which group does the active substance belong (see Note 2)? Herbicide or plant growth regulator go to 7 Other go to 4 4 Obtain screening data (efficacy data can also be used) from at least six vascular plant species from six different families including both mono- and dicotyledons (see Note 3).go to 5 5 Do the results in one or more of the screening or efficacy tests with vascular plant species show ≥ 50% phytotoxic effects at or above the maximum recommended application rate (MRR) (see Notes 4 and 5)? If no classify as low risk, go to 16 If yes go to 6 6 Obtain concentration/response tests on the affected species (see Note 6). If MRR/EC50 is < 1 for all affected species classify as low risk, go to 16 If MRR/EC50 is ≥ 1 for one or more affected species go to 7 7 Obtain concentration/response tests on at least six species representing families for which significant herbicidal action has been found (see Note 7) and determine the EC50 value for each of the tests.go to 8 8 Using the EC50 values for six or more different species, generate the fifth percentile of the log-logistic or the log-normal distribution according to Aldenberg & Slob (1993) or Wagner & Løkke (1991) or the final acute value (triangular distribution) according to Stephan et al. (1985) depending on the most likely shape of the toxicity distribution curve (see Note 8) and use this calculated toxicity value in the risk assessment. go to 9 Drift events 9 Obtain from Chapter 3 (Air) information about the drift concentration of the active substance that can be expected at distances of 1 and 5 m from the treated area, or 3 and 7 m for orchards (see Note 9).go to 10 10 Calculate the exposure over toxicity ratio. If PEC(1 m distance)/calculated toxicity value < 1 classify as low risk for drift events, go to 11 If PEC(5 m distance)/calculated toxicity value < 1 classify as medium risk for drift events, go to 11 If PEC(5 m distance)/calculated toxicity value > 1 classify as high risk for drift events, go to 11 Gaseous transport events For most products, gaseous transport events are not likely to be significant, compared with drift events. However, under certain conditions, they may be the major routes of contamination (for instance, compounds with a high volatility). In some cases, indications of possible effects of gaseous transport events can be found on the application label (e.g. ‘do not use this product near a certain type of crop’). Atmospheric deposition consists of dry and wet deposition. The dry deposition flux of gaseous substances is proportional to the dry deposition velocity. At a low dry deposition velocity, the total deposition flux is mainly determined by wet deposition. The wet deposition flux of a substance depends on its solubility in rain and the precipitation intensity. For a substance that is highly soluble, wet deposition is more important than for a substance that is poorly soluble. 11 Obtain from Chapter 3 (Air) information about the concentration of the active substance that can be expected at a certain location. If PEC(gaseous transport) ≤ PEC(drift) at 1 m distance go to 12 If PEC(gaseous transport) > PEC(drift) at 1 m distance go to 13 12 No further risk assessment for gaseous transport events is necessary, because the outcome of this risk assessment will always be lower than that based on drift events.go to 14 13 Calculate the exposure over toxicity ratio (see Note 10). If PEC(gaseous transport)/calculated toxicity value is < 1 classify as low risk for gaseous transport, go to 14 If PEC(gaseous transport)/calculated toxicity value is > 1 and < 10 classify as medium risk for gaseous transport, go to 14 If PEC(gaseous transport)/calculated toxicity value is > 10 classify as high risk for gaseous transport, go to 14 Run-off events For most products, run-off events are not likely to be significant compared with drift events. However, under certain conditions, they may be the major routes of contamination (for instance, compounds used on paved areas and/or in areas with steep slopes). Whether a run-off event will occur depends on the location (shape of the landscape) and the weather conditions. Guidance can be found in the recommendations of the FOCUS group (standard scenarios for certain regions of Europe). 14 Is the uptake of the active substance mainly via the roots (in general soil-applied compounds/formulations) If yes go to 15 If no classify as low risk for run-off events, go to 16 15 Obtain from Chapter 4 (Soil) information about the concentration of the active substance that can be expected at the scenario location (see Note 10). If PEC(run-off scenario)/calculated toxicity value is < 1 classify as low risk for run-off events, go to 16 If PEC(run-off scenario)/calculated toxicity value is > 1 and < 10 classify as medium risk for run-off events, go to 16 If PEC(run-off scenario)/calculated toxicity value is > 10 classify as high risk for run-off events, go to 16 Analysis of uncertainty After completing the risk assessment based on data reflecting normal use of the product (e.g. average data for drift), it is necessary to consider whether errors in measurement, or variations in conditions of use, or the outcome of field studies, might alter the conclusions. This is appropriate for products initially categorized as medium or low risk to higher plants, to detect cases in which risk might be higher in practice. It is also appropriate for products initially categorized as high risk to consider the possibility that several of the assumptions may have been overconservative (see also Note 11). 16 Review the data that led to high-, medium- or low-risk category and check whether the conclusion is correct If yes, confirm assessment go to 17 If no, obtain more information as needed go to 2 Risk management 17 In most cases, products classified as negligible risk do not require any particular risk management measures. Products in the low-risk category are not likely to affect non-target plants when used according to good plant protection practice up to the highest recommended application rate. Products in the medium-risk category may have some effect on some non-target plants adjacent to treated areas. They should be labelled accordingly. The range of plant species at risk should be considered before a risk phrase (e.g. ‘avoid use within a distance of x m from sensitive areas’) appears warranted. Products in the high-risk category are likely to have a negative impact on non-target plants. They should be labelled accordingly and, where necessary, the range of plant species at risk should be considered when the appropriate risk reduction measures (e.g. ‘do not use within a distance of x m from sensitive areas’) are added to the label. Sensitive areas can be defined as areas containing the species at risk and/or particular non-target areas such as hedgerows, forest margins or wetlands. Measures to prevent drift or run-off may lead to lower exposure of non-target higher plants outside the field of application. This can be achieved by using low-emission application techniques such as low-drift nozzles or directed applications, or by creating buffer zones. Alternative formulation types and/or application techniques (e.g. as granules rather than to foliage) should be considered. Formulation characteristics and mode of application will determine the size of the buffer zones. Another way of preventing exposure of non-target higher plants to plant protection products is to insert barriers, e.g. artificial or natural wind shields. However, where windbreaks have been inserted for drift interception purposes, they should not be regarded as sensitive areas of particular concern, as discussed above. A boom-sprayer application may, in some cases, be replaced by a row-sprayer application (which is possible within the seed or plant furrow), or a seed treatment might be substituted in other cases. A change in the type of formulation may also be considered, e.g. a seed treatment or a coarse granulate instead of a spray. Lower application rates or less frequent use represent other possible risk management tools. Explanatory notes Note 1 Negligible risk If the nature of the product and its use are such that exposure of higher plants in the non-target area will most likely not occur (e.g. substances used for wound protection, for stored products, substances used in glasshouses, rodenticides, substances used for seed treatment or by stem injection), no further consideration of the effects on plants is necessary. Note 2 Type of plant protection products Herbicides and plant growth regulators are designed and/or selected to affect higher plants in one way or another. Although screening data is available for these compounds, steps 4–6 are omitted and concentration/response data is necessary to evaluate the potential risk of these compounds on non-target higher plants. In most countries, products are classified according to their major use. Where a product is also used as a herbicide or plant growth regulator, it should be treated as such even if the main use is another (i.e. insecticide/fungicide). For active substances with a specific mode of action or a specific type of application, which would not be detected in the two standard tests [vegetative vigour test and seedling emergence test conducted according to OECD (1984), EPA (1985), Holst (1986) or Holst & Ellwagner (1982)], specifically designed methods using expert knowledge should be employed (this may be needed, for example, for soil sterilants working mainly via the gaseous phase). Note 3 Screening tests for phytotoxicity Data should be provided on all vascular plant species that have been tested during the screening process. The tested concentration should be equal to or higher than the maximum recommended application rate (MRR). Screening data and efficacy data allow one to assess whether phytotoxic properties of the test substance can be expected. They are usually conducted prior to the registration process and not according to GLP. For assessments of this data to be useful, it should include a certain minimum set of information (for details, see OECD guideline 208). Screening data will usually be obtained with some kind of standard formulation. This is regarded to be acceptable, as such tests are generally conducted so as not to miss any significant activity (e.g. by adding significant amounts of effective additives). They may therefore be regarded as at least as sensitive as a worst case. Note 4 Why 50% effect is enough? In higher-plant testing, quantitative end-points like biomass, plant height, number of leaves, percent chlorosis, percent ground cover, etc. are nearly always measured. In such tests, the reaction of a plant population or any subpopulation is of interest and testing is therefore usually done with more than one plant so that even a qualitative end-point like germination becomes a quantitative end-point and is expressed as percent germination. In concentration/response tests studying the effect of plant protection products on plants, an EC50 value should be determined for each species by non-linear regression, and the 95% confidence interval should be calculated. As non-lethal end-points are determined for the calculation of the EC50 value, the EC50 value is proposed as the basis for further assessment steps and not a NOEC value. Furthermore, a 50% reduction of plant biomass production in an early stage, or of seedling emergence, can largely be compensated by a plant population. The EC50 value is also proposed for technical reasons. To estimate a NOEC value on the basis of a concentration/response curve means to determine an EC5 or EC10 value. This is extremely difficult because of the high variability within plant populations, particularly if non-crop species are used. Note 5 False positives To exclude false-positive outcomes, additional tests may be carried out if 10%, or at most 20%, of the species in the screening tests show phytotoxic responses (e.g. when the tested concentrations are above the maximum recommended application rates). Note 6 Definitive tests A definitive test is a test conducted with a range of concentrations of the product, preferably in a geometric progression, but covering the EC50 values for the test species selected (e.g. OECD 208). For post-emergence herbicides and other plant protection products, the vegetative vigour test should be used (unless the mode of action otherwise indicates a specific test, for example, growth regulators influencing flowering, or some inhibitors of cell division causing flower sterility). For pre-emergence herbicides, the seedling emergence test is more applicable. The definitive tests should be conducted using an appropriate widely used formulation. Note 7 Number of species A minimum number of six species, representing families for which significant herbicidal activity has been claimed, should be tested. Statistical methods should be applied to calculate the fifth percentile of the available toxicity data (see Note 8). These methods apply uncertainty factors that depend on the sample size; the safety factors decrease with increasing sample size. Note 8 Species sensitivity distributions In contrast to the situation with many other species of concern (e.g. fish, daphnids, birds, etc.), the risk assessment for higher plants exposed to phytotoxic compounds can be based on toxicity data for six or more species. This starting point makes it possible to use a statistical method to calculate the toxicity value for the risk assessment, instead of applying a fixed uncertainty factor of 10 to the lowest available toxicity value. The statistical methods available assume that the toxicity data follow a certain type of distribution: Stephan et al. (1985) Triangular distribution Wagner & Løkke (1991) Log-normal distribution Aldenberg & Slob (1993) Log-logistic distribution For the method of Stephan et al. (1985), data is required for at least eight species. The final acute toxicity value calculated by this method is based only on the lowest three toxicity figures. The other two methods calculate the fifth percentile of the log-normal or log-logistic distribution for all available toxicity data. The methods of Wagner & Løkke (1991), van Straalen & Denneman (1989) and Aldenberg & Slob (1993) estimate, on the basis of NOEC data, the hazardous concentration for 5% of the species (the fifth percentile of the log-normal or log-logistic distribution of all NOECs). This method can also be used to calculate the fifth percentile of the log-logistic distribution of all EC50 values: 5th percentile = 10(AVG−E*STD) in which: AVG = the mean of the log10 transformed EC50 values STD = the standard deviation of the log10 transformed EC50 values E = extrapolation factor dependent on sample size (Table 1). Table 1. Extrapolation factors for the median and one-sided left confidence limits for the log-logistic and log-normal distribution (after Aldenberg & Slob, 1993; Aldenberg & Jaworska, 2000) Sample size Median estimate Log-logistic Log-normal 6 1.81 1.75 7 1.78 1.73 8 1.76 1.72 9 1.75 1.71 10 1.73 1.70 11 1.72 1.70 12 1.72 1.69 13 1.71 1.69 14 1.70 1.68 15 1.70 1.68 All three methods are included in the ETX program of Aldenberg (1993) and, in addition, a goodness-of-fit test, based on the Kolmogorov–Smirnov test statistics according to d’Agostina & Stephens (1986), is provided for testing the log-normal or log-logistic distribution. Note 9 Non-target area The aim of the subscheme is to protect higher plants in the non-target area from unacceptable phytotoxic effects of plant protection products under realistic conditions. The non-target area is defined as the area outside the treated area. The type of vegetation and plant species composition in these areas is extremely diverse and cannot be easily classified. Such areas include hedgerows, shelter belts, windbreaks, wetland, wood lots and grassland, which are widespread terrestrial plant habitats in a landscape. Hardly any of these habitats border on a treated area directly. Usually, there is a more or less wide strip between the treated area and these habitats. The flora of these transitional zones (field margins) is often a typical permanent grassland vegetation, strongly influenced by the cropping system of the adjacent fields. This influence is not restricted to the entry of plant protection products but also to the entry of fertilizer, to soil disturbance by machinery, etc. Therefore, the impact of the cropping system is more significant next to the treated area than further away. As the risk assessment should be done under realistic conditions, the classification of plant protection products into risk categories is made according to the distance to the treated area. Up to a distance from 1 m to the treated area, a greater impact can be tolerated than at a greater distance. For orchards, a field margin of 3 m is assumed and for all other crops one of 1 m. Risk assessment is carried out at 3 m and at 7 m for orchards and at 1 m and 5 m for the other crop types. Note 10 Safety factors No scientific reason can be given for the safety factor of 10. This value is chosen because it offers roughly the same level of protection as provided in the drift scenario (differences in percentage of drift at 1 m and 5 m). Note 11 Analysis of uncertainty Where, on the basis of laboratory toxicity tests and reasonable worst-case exposure assumptions, a high risk is identified, the following risk refinement options should be considered. It may be possible to consider more realistic exposure scenarios, i.e. the use of PECs which are not based on 95th percentile estimates of exposure. This area is currently the subject of international discussion. Dose justification should be critically reviewed. Toxicity values may be refined by further testing, which will be conducted under more environmentally relevant conditions. It may be possible to perform these tests under glasshouse or semi-field conditions. The use of techniques which better simulate drift (where this is the major route of exposure) should be considered. Application at less sensitive growth stages should be considered, if appropriate to the plant protection product and its intended use pattern. The use of test plants which have been grown outside should be considered. The importance of the seed bank for the species of concern should be assessed in relation to the proposed use. It is often difficult to measure the effects of plant protection products on natural communities in the field because of the inherently high variability in natural populations (Marrs & Frost, 1997). These authors propose a semi-field (microcosm) approach, which may provide a suitable method for studying the impact of repeated applications. Field studies should only be used after consideration of other refinement options such as studies in a glasshouse or on small field plots. References Aldenberg T (1993) ETX 1 3a A Program to Calculate Confidence Limits for Hazardous Concentrations Based on Small Samples of Toxicity Data. RIVM Report 719102015 . National Institute of Public Health and Environment, Bilthoven (NL). Aldenberg T & Jaworska JS (2000) Uncertainty of the hazardous concentration and fraction affected for normal species sensitivity distributions. Ecotoxicology and Environmental Safety 46, 1– 18. Aldenberg T & Slob W (1993) Confidence limits for hazardous concentrations based on logistically distributed NOEC toxicity data. Ecotoxicology and Environmental Safety 25, 48– 63. Aldridge CA, Boutin C & Peterson HG (1993) Guidelines for testing effects on non-target plants. In: Brighton Crop Protection Conference – Weeds – 1993, pp. 145– 150. BCPC, Farnham (GB). Boutin C, Freemark KE & Keddy CJ (1993) Proposed Guidelines for Registration of Chemical Pesticides: Non-Target Plant Testing and Evaluation. Technical report Series no. 145. Canadian Wildlife Service, Ottawa (CA). D’Agostina RB & Stephens MA (1986) Goodness-of-Fit Techniques, Vol. 68. Statistics. Textbooks and Monographs. Marcel Dekker, New York (US). EPA (1985) Toxic substances control act test guidelines: environmental effects testing guidelines. Seed germination/root elongation toxicity test. US Federal Register 50 (188), 39389– 39391. Fletcher JS (1991) Assessment of published literature concerning pesticide influence on non-target plants. In: Plant Tier Testing: a Workshop to Non-target Plant Testing in Subdivision J Pesticides Guidelines (Ed. by JS Fletcher & H Ratsch), pp. 6– 15. Environmental Protection Agency, Corvalis (US). Holst RW (1986) Hazard Evaluation Division, Standard Evaluation Procedure, Non-target Plants: Seed Germination/Seedling Emergence – Tiers 1 and 2. Report no. EPA 540/9-86-132. Environmental Protection Agency, Washington (US). Holst RW & Ellwagner TC (1982) Pesticide Assessment Guidelines, Subdivision J, Hazard Evaluation: Non-target Plants. Environmental Protection Agency, Washington (US). Marrs RH & Frost AJ (1997) A microcosm approach to the detection of the effects of herbicide spray drift in plant communities. Journal of Environmental Management 50, 369– 388. OECD (1984) Guidelines for Testing of Chemicals no. 208: Terrestrial Plants Growth Test. OECD, Paris (FR) (a revision will be available in the near future). Stephan CE, Mount DI, Hansen DJ, Genrile JH, Chapman GA & Brungs WA (1985) Guidelines for Deriving Numerical National Water Quality Criteria for the Protection of Aquatic Organisms and Their Uses. Environmental Protection Agency, Washington (US). Van Straalen NM & Denneman CAJ (1989) Ecotoxicological evaluation of soil quality criteria. Ecotoxicology and Environmental Safety 18, 241– 251. Wagner C & Løkke H (1991) Estimation of ecotoxicological protection levels from NOEC toxicity data. Water Research 25, 1237– 1242. Citing Literature Volume33, Issue2August 2003Pages 239-244 FiguresReferencesRelatedInformation