Appropriate sampling methods and statistics can tell apart fraud from pesticide drift in organic farming
Pesticide residues in organic productsNon-use of synthetic pesticides is a major characteristic of organic farming, with the objectives of protecting (a) the environment, (b) consumer health, and (c) farm worker health. In consumer studies, “no chemical pesticides” is usually mentioned as one of the most important criteria for buying organic food1,2. These consumer expectations are mostly met in what is referred to in objective (b). Both European and U.S. sources consistently found the percentage of samples with residues of pesticides above the limit of quantification ( > LOQ) to be much lower in organic than in conventional food (Fig. 1a, see also Supplementary Fig. 1). This is especially true when it comes to fresh fruits and vegetables (Fig. 1c), which are known to be the most critical food groups in terms of pesticide residues3. It is elucidating, however, to not only look at the number of samples with an (unknown) level of residues > LOQ, but to quantify the residues found per sample. In many cases, more than one substance is found in a sample, therefore one meaningful indicator is the mean cumulative pesticide load per sample (MCPL, see Supplementary Table 1). This is represented in Fig. 1b for three out of the four datasets. The food authority CVUA (Chemisches- und Veterinäruntersuchungsamt) in Baden-Württemberg, Germany, has been comparing pesticide residues between organic and conventional food since 2002. In 2019, on average the residues in organic produce were more than 150 times lower than in the corresponding conventional products4 (Ratio Org./Conv., bottom of Fig. 1d). The USDA (U.S. Department of Agriculture) numbers tend to be higher than the European ones, both in percentages (Fig. 1a,c) and in MCPL (Fig. 1b). One reason for this is probably USDA’s risk-oriented sampling approach, in which some highly contaminated commodities are over-represented, as compared to their importance in most people’s diet (Supplementary Table 2, column C). If we correct this possible bias by assuming that every commodity would have been sampled with the same frequency, the MCPL across all commodities is cut by 40% (Supplementary Table 2, last row). Different LOQs and numbers of analytes covered by USDA on one hand, and different European laboratories on the other hand, also make comparison difficult.Figure 1Pesticide residues in conventional and organic food in tests conducted by four organisations: EFSA (European Food Safety Authority) collects official data from all EU member states3, CVUA from one federal state in Germany4, USDA from government laboratories across the U.S.5, while Eurofins is a commercial laboratory in Germany. Figures in brackets represent the number of samples. The legend is valid for (a), (b) and (c). In order to increase the number of samples (represented in brackets) and thus their representativeness, figures from several years were grouped together, as available from each organisation. Black bars symbolise standard errors across years. (a) Shows the percentage of samples with residues above the limit of quantification (LOQ), for all types of food ( available from two organisations only). (b) Represents the mean cumulative pesticide load (MCPL) for fruits and vegetables (available from three organisations). (c) Similar to (a), but for fresh fruits and vegetables only (CVUA uses “above 0.01 mg/kg” instead of LOQ, but this is identical for most substances). The same datasets were used for (b) and (c). (d) Multi-layer sieving model for residue testing at different points of the organic supply chain. The data above the white arrows are from the commercial laboratory Eurofins, and mostly represent the situation before products are released to the market, while the figures below the white arrows are from CVUA, representing the situation on the market (both wholesale and retail). Ratios from “before market” to “on market” are shown in the white arrows. In this process, the MCPL remains in the same range for conventional products (blue rectangle to the right), while it is reduced massively for organic products (green trapezium in the centre). As a result of this sieving mechanism, residues in samples from the market are 150 and more times lower in organic than in conventional produce (trapezium at the bottom). This shows that the process represented by the blue arrows works fairly well—which is not always the case for the investigation of the origin of such residues, symbolised by the yellow arrows.Full size imageOrganic businesses’ testing strategiesUnfortunately, the generally good news for consumers with respect to objective (b) does not always mean that objectives (a) and (c) are also met. With the steady growth of the organic market and globalisation of supply chains, integrity of the system is often at stake. Organic products mostly fetch higher prices, and therefore also attract fraud6,7. Since pesticide residues are easily detectable parameters, often indicating non-compliance with organic production rules, many organic businesses test each batch for such residues, before placing it on the market. Positive results should then lead to an investigation of the origin of the found residues: Did an organic farmer spray? Do the residues come from drift, from ubiquitous contamination, or from (avoidable or unavoidable) contamination during processing, transport, storage? Were organic and conventional products mixed at some point of the supply chain—or is somebody simply labelling conventional products as “organic”? The idea behind this is depicted in Fig. 1d. The filter process as such, and the exclusion of contaminated batches from the organic market, as represented by the blue and red arrows, often work well. Thus, there are remarkably lower average amounts of residues after undergoing this filtering process. Residues in organic produce reported from the market were reduced by 22 and 89 times in fruits and vegetables, respectively, compared to the levels reported by the commercial laboratory, which represent mostly pre-market samples, while the values for conventional samples remained in the same range. This shows that market actors often remove problematic batches by declaring them conventional. In Supplementary Tables 3 and 4 we provide further explanations why the datasets “before release to the market” and “on the market” in Fig. 1d are comparable.We do not have test results from a commercial laboratory in the U.S. that could be compared to Eurofins data. But, as opposed to the other sources of information, the USDA database identifies the country of origin of each sample. Anybody working in international organic certification would expect residues in imported food to be higher than in domestic products, because fraud is more widespread when the distance is bigger between producers on the one hand, and consumers and the competent authorities on the other. The U.S. data, however, suggest the opposite trend: Not only at the aggregate level, but also for most individual commodities, the MCPL is lower in imported than in domestic products (Supplementary Table 2, columns J and K). The reason is probably that samples are tested before signing purchase contracts, and products rejected or bought as conventional, if they do not comply with the expectations.This is good quality control practice—the problem is that the information about the “downgrading” of organic products to conventional is not always reaching the certification bodies (CBs), thus impeding the investigation of the origin of residues and the exclusion of fraudulent actors from the market (yellow arrows in Fig. 1d). It is in the nature of things that these processes are not publicly known and therefore cannot be quantified, but in Supplementary Fig. 2 we present anecdotic evidence, which also suggests that for some market actors the definition of “organic” is limited to “free of pesticide residues”.Certifiers’ testing strategiesThe two most important markets for organic food are the EU, where the “organic” label is legally governed by an EU Regulation, and the USA, where the corresponding rule is the National Organic Program (NOP). Although they have different approaches on how to deal with spray-drift and with residues (Supplementary Table 5), both regulations require CBs to take samples from at least 5% of their clients every year. A large amount of data is being generated through this mechanism, but the sampling procedures and interpretation of results often do not allow deriving clear results. A recent unpublished BSc thesis at the University of Kassel revealed that 80% of the samples by CBs in ten EU member countries are taken of final products, but only 20% from the field or during the production process. This suggests that not only for market actors, but also for many CBs, the purpose of sampling and testing is limited to ensuring that food sold on the market with an organic claim, is free of pesticide residues, without digging deeper to find the origin of contamination.The differentiation between active use and non-intentional contamination is difficult, if only final products are tested. Plant (mainly leaf) samples from the field have several advantages in this regard: (a) Often, there is a long time span between pesticide application and harvest. Because of dissipation of the residues, nothing or only traces may be found in the final product (Supplementary Table 6). Field samples can be taken during or shortly after a suspected pesticide application, so that the dissipation effect is reduced and residues are found even for substances with a short half-life. (b) Leaves have a surface/weight ratio between 10 and 118 cm2/g8, whereas for fruits this ratio is between 0.6 and 2.29, and for seeds between 2 and 10 cm2/g only10,11,12. Residues in leaves are therefore normally higher than in seeds, fruits or roots, which makes interpretation of test results easier. (c) Field sampling allows taking separate samples from centre and margin of the field, as explained below in more detail.Unfortunately, if CBs take field samples at all, they often take them only from field margins13,14 (“let’s see if there is a drift problem”). Positive results are then attributed to spray-drift, and farmers are required to establish buffers—without even considering the possibility of residues originating from an application by the organic farmer. Such procedures open the door for fraudulent use of pesticides by organic farmers.Other CBs have established so-called “action levels”, below which they consider the presence of residues in organic products to be the result of ubiquitous environmental contamination, with no need to investigate their origin13. While such thresholds may be necessary for specific cases (see below concerning the banana industry), using this approach as a general procedure disregards not only the spatial distribution, but also temporal dynamics of pesticides in plant tissue. As opposed to soil, half-lives in plant tissue exposed to UV radiation and weather, are relatively short for most modern pesticides15. A residue level of 0.02 mg/kg, used by some CBs as “action level”, is typically reached one to two months after the application of a pesticide, in some cases even after only five days (Supplementary Table 6).The time that has elapsed since an application, however, is unknown in most cases. Spraying records kept by conventional neighbours are normally not part of the inspection. In case of suspicious test results in samples from the organic farm, such records may sometimes be accessed as part of a follow-up investigation, but at that point the organic farmer may have asked the neighbour to manipulate the records. And if the organic farmer has sprayed, he or she obviously tries to hide this fact. This situation makes interpretation of low levels of residues found in samples from organic fields even more challenging, and increases the importance of being able to differentiate application from drift through other methods.Two forms of spray-driftOver the past decades, a distinction has been made between short distance primary spray-drift during the application, and long distance secondary spray-drift occurring after the application16. The latter was attributed to evaporation and considered to play a role only for pesticides with high vapour pressure17. On the one hand, recent studies have shown that evaporation and long-distance transport can already play a role during, not only after application18. On the other hand, long-distance transport has been found to be linked not only to evaporation. Pesticides adherent to dust from wind erosion can contaminate large areas19. In the present context, we use the terms short-range and long-range drift, instead of primary and secondary drift (Fig. 2).Figure 2Simplified model of short-range vs. long-range drift originating from air-blast spraying in a fruit orchard. The specific values for pesticide concentrations (mg/kg) expected for different downwind distances from the orchard can vary by a factor 10 or more, depending on the applied substance, dose, weather conditions, vegetation, etc., but the graph provides an approximate estimate of the ratios that can be expected. In the case presented here, pesticide concentration in fruit leaves immediately after the application is 15 mg/kg. In the area of short-range direct drift, deposit decreases exponentially, so that at 100 m distance, we can expect to find only 0.01 mg/kg. At further distances, deposits are often below this level.Full size imageLong-range driftLong-range drift is so far poorly understood, can lead to (normally very low) residues at distances as far as thousands of km19, and happens in the form of vapour or molecules adhering to dust. The main factors influencing long-range drift are vapour pressure of the pesticide, capacity of adherence to dust, incidence of wind erosion, and temperature inversion in the atmosphere17. Long-range pesticide drift has recently received more attention21,22,23,24,25. Examples have been used in the context of organic certification for supporting the argument of ubiquity of pesticides, linked to the assumption that low- or even medium-level residues in organic products are often derived from their omnipresence in the environment26,27.Cases from Brazil (endosulfan in soybeans), Montana (USA) and Saskatchewan (Canada) (glyphosate in khorasan wheat) and Germany (pendimethalin and prosulfocarb in different crops) have been quoted to demonstrate the ubiquity of pesticides27. None of these case studies, however, provides solid evidence for the assumption that long-distance transport of pesticides leads to residues in organic food above the level of, say, 0.01 to 0.03 mg/kg. The problem of the herbicides pendimethalin and prosulfocarb being subject to long-distance drift because of their high vapour pressure, has been known for a long time28, but this phenomenon cannot be extrapolated to other substances. Even for these herbicides, there is no evidence that residues at larger distances could be above the indicated levels. Across 15 vegetation samples from nature reserves in Germany, on average, 0.009 mg/kg pendimethalin and 0.004 mg/kg prosulfocarb were found29. Exceptions may exist, e.g., when pesticide applications are followed by heavy wind erosion, as seems to be the case in some of the North American wheat growing areas, where glyphosate is used for cereal desiccation shortly before harvest.In a survey in Switzerland30, neonicotinoid residues were found in 93% of plant samples from organic farms (as compared to 100% of samples from conventional farms), thus supporting the ubiquity suspicion. But there were substantial quantitative differences between organic and conventional farms (Fig. 3). The average sum of neonicotinoid residues in plant and soil samples from organic farms was lower by a factor of 11 than that of plant samples from conventional farms. For soil samples, this factor was as high as 71. Even the highest value for one single substance (imidacloprid) found in organic plants (2.13 µg/kg = 0.00213 mg/kg) would be below the limit of quantification (LOQ) used for this substance in most screenings (0.01 mg/kg).Figure 3Maximum and average residues of neonicotinoid insecticides in soil and plant samples from organic farms, integrated crop production (“IP Suisse”: this program involves reduced pesticide application) and conventional farms in Switzerland. The figures represent the sums of acetamiprid, chlothianidin, imidacloprid, thiacloprid and thiamethoxan. Figures in brackets represent standard errors.(Data from Humann-Guilleminot et al.30).Full size imageIn a study in Germany29, the MCPL in natural vegetation in five reference areas (average distance from arable fields >3 km) was 0.003 mg/kg, and in 15 nature conservation areas (average distance from arable fields 143 m) it was 0.006 mg/kg, but in three buffer zones (average distance 54 m) it was 5.4 mg/kg. To make figures comparable with other data in this article, we have subtracted the concentration of non-agricultural pesticides from the total amounts, and divided the numbers by a factor five, because the residues in this study refer to dry matter, while all the others use fresh matter. Although 5.4 mg/kg at 54 m distance is a disturbingly high value, the survey confirms that concentrations at larger distances do not exceed the “traces” level. The intention of this article is not to put in doubt the environmental damage caused by such traces. What we try to show is that the “ubiquity” argument may sometimes be hiding cases of fraudulent pesticide use by organic farmers.Short-range driftAs opposed to long-range drift, short-range drift is well understood, has its impact mainly in a range from 1 m up to a maximum of 1,000 m (for aerial spraying), happens in the form of droplets, and is not substance specific. The main factors influencing this form of drift are droplet size, windspeed, and height of the boom (nozzles) above soil17,19,31,32,33. The fact that long-range drift is poorly understood and leads to low concentrations of certain substances over wide areas, should not stop certification bodies (CBs) from using the available knowledge about short-range drift as a tool for assessing farmers’ compliance with organic production rules. The dynamics of short-range spray-drift have been widely studied in the context of preventing liability problems due to herbicide damage, contamination of water bodies and natural habitats, and direct risks for human settlements19,31,32,33,34,35,36. Pesticide deposit decreases exponentially with increasing distance from the field on which the substance is applied. With a tractor boom sprayer, deposit at 25 m distance is expected to be only 1% of that in the target field. While distances are greater for air-blast or aerial spraying, the basic principle of exponential decrease is the same (Fig. 2 and Supplementary Fig. 3).ObjectivesThe objectives of our study are: (I) to demonstrate that appropriate field sampling methods can differentiate the effects of fraudulent pesticide application by the organic farmer, from the results of both short-range and long-range spray-drift, and (II) for the specific case of aerial fungicide spraying in the banana industry, identify appropriate variables, which allow us to interpret the test results correctly for the purpose of this differentiation. More
