Pesticide Residues Remaining In Soils From Previous Growing Season.docx

Pesticide residues remaining in soils from previous growing season(s) - Can they accumulate in nontarget organisms and contaminate the food web? H I G H L I G H T S The fate and behavior of pesticides persisting from previous growing sea- son(s) were studied. Residues' potential for bioaccumulation in earthworms was generally low. • Residues' potential for bioconcentration in the lettuce plant was negligible. • Residues´ freely dissolved (i.e., soil porewater) concentrations were negli- gible. • Pesticide residues present at levels of ≤0.1 mg/kg in fields are not likely to present threats.

abstract Epoxiconazole, tebuconazole, flusilazole, prochloraz, pendimethalin, and the atrazine transformation product (2- hydroxyatrazine) have been found in Czech arable soils at high detection frequencies and/or concentrations. As they have been shown to persist from one growing season to following ones, the question arises of whether they can be taken up by non-target soil organisms and by subsequently planted crops. To reveal this, soils field- contaminated with pesticide residues were subjected to laboratory microcosm studies to measure i) dissipation rates, ii) accumulation in earthworms and lettuce, and iii) exposure by means of solid-phase microextraction (SPME). In parallel, tests with a freshly laboratory-contaminated soil were performed and rep- resented the worst case scenario to be compared with. It was observed that at the residual levels (≤0.1 mg/kg), the behavior of field aged and fresh residues was similar, except for bioaccumulation in earthworms that was sig- nificantly lower for aged residues than for fresh residues. Residues' potential for bioconcentration was generally low, i.e., below the maximum residue limits (MRLs) of lettuce. This is in line with SPME results showing low levels of exposure via soil porewater. It follows that these pesticide residues are not likely to pose significant threats to the soil environment, the food web and, consequently, human health if present in soils at levels of ≤0.1 mg/kg.

1. Introduction

The worldwide annual consumption of pesticides has amounted to 2.7 × 106 t in recent years (FAO, 2017) and may further increase due to the growing human population, climate change and increasing pest pressures (Network, 2016). It has been reported that pesticide residues undergo long-range transport and thus may appear in pristine areas like groundwater (Kodes and Rieder, 2005; Toccalino et al., 2014), montane forests (Daly et al., 2007), seawater and polar areas (Morris et al., 2016). In addition, many currently used pesticides (CUPs) have been reported as persistent, bioaccumulative, and toxic compounds and listed as po- tential candidates for substitution (European Commission, 2015) and/ or as priority pollutants (European Commission, 2015). This suggests that pesticides should be systematically monitored in the environment and their risks to non-target organisms should be carefully considered. In 2015, seventy-five Czech arable soils were analyzed for the pres- ence of 53 currently or recently used pesticides and 15 transformation products (Hvězdová et al., 2018). The most frequently found pesticides were triazine herbicides represented mainly by the transformation prod- uct of atrazine (banned for a decade now), i.e., 2hydroxyatrazine (39% of soils, levels up to 0.123 mg/kg), and conazole fungicides, namely epoxiconazole (48% of soils, up to 0.031 mg/kg), tebuconazole (36% of soils, up to 0.028 mg/kg), flusilazole (23% of soils, up to 0.019 mg/kg), and prochloraz (21% of soils, up to 0.028 mg/kg). Of the studied pesticides, pendimethalin was presented at the highest concentration (0.139 mg/kg) (Hvězdová et al., 2018). Although these findings might be alarming on their own (Hvězdová et al., 2018), they do not provide a complete picture of the real risks associated with the presence of pesticide residues in soils. What still remains unknown is: i) whether these chemical residues will further dissipate from soils to lower their contents, ii) what the actual ex- posure levels of these field-residues to non-targets are, and iii) whether the residues pose a significant potential for bioaccumulation in soil biota and the food web. To answer, the most contaminated soils from Hvězdová et al. (2018) were subjected to laboratory microcosm studies in order to measure dis- sipation of the long-term residues, their uptake to the lettuce plant (Lactuca sativa) (Hamdi et al., 2015; Hwang et al., 2017) and earthworms (Eisenia andrei) (OECD, 2010 and 1984) and their exposure levels in pore- water by means of solid-phase microextraction (SPME) (Bending et al., 2007; Bondarenko et al., 2008; Ter Laak et al., 2006). In parallel, worst- case scenarios consisting of the exposure of biological endpoints and SPME fibers to laboratory-contaminated sand (for earthworms) and soil (for plants) were conducted. This allowed a comparison between the fate, partitioning and bioavailability of the freshly added

and of the aged (field) chemicals, which from the viewpoint of risk assessment repre- sented examples of the worstcase and the real-case scenarios, respec- tively. In general, the study increases our understanding of the threats related to previous pesticide use (e.g., via a systematic assessment of the behavior of 2-hydroxyatrazine) and to the current use of pesticides which the EU inspects for hazards (e.g., conazole fungicides). It also con- tributes to the quantification of acceptable levels of pesticide residues in soils that are not likely to pose risks to non-target species and the contam- ination of the food web in the following growing season(s). 2. Material and methods 2.1. Materials and chemicals Sand (N50% particles of 0.05–0.2 mm in size) was purchased from Filtrační Písky, s.r.o. (Hornbach, Chlum). Chemical standards of pesticides (epoxiconazole, tebuconazole, flusilazole, prochloraz, pendimethalin, and 2-hydroxyatrazine) were purchased from Pestanal® (Sigma Aldrich, Germany). The properties of the tested chemicals are summarized in Table S1. Acetonitrile and methanole (≥99.9% purity) was purchased from Chromasolv® (Sigma Aldrich, Germany). 2.2. Experimental soils and sand Soil sampling was performed within a large survey of 75 arable soils (Hvězdová et al., 2018) that were all screened for pesticide residues. The methodology of sampling and sample processing was described in de- tail in Hvězdová et al. (2018). Due to the sampling period in February and the expected last pesticide application in November, the pesticide residues in the sampled soils were at least 5 months old. Of the sampled soils, four model soils were used in the current study that contained the highest levels of conazole fungicides (FC01), pendimethalin (FC02), and of 2-hydroxyatrazine (FC03 and FC04). In addition to the above- mentioned compounds, FC02 contained also epoxiconazole, FC03 prochloraz and FC04 tebuconazole. Soil properties including the initial concentrations of chemicals are summarized in Table S2. Prior to the ex- periments, soils were mixed with N-P-K fertilizer (1 g per kg dry weight, dw) and moistened to 50% of the water holding capacity (WHC). Along with the four field soils, sand spiked at a concentration of 0.1 mg/kgdw (similar to the highest field levels) was used (CS – contam- inated sand) as a positive control (worstcase scenario) in the earth- worm bioaccumulation test. For the plant bioaccumulation test, sand was replaced with sandy soil since lettuce was unable to grow in sand. This control soil was contaminated at the level of 0.1 mg/kgdw and is re- ferred to as the laboratory-contaminated (LC) soil. The properties of both sand and soil along with the initial concentrations of tested chemicals are provided in Table S2. The fortification of control sand and soil was performed as follows: 1700 g of the dry sand/soil was moistened to 50% WHC and split into half. Then, 50 mL of an acetone stock solution containing 0.17 mg of epoxiconazole, tebuconazole, flusilazole, prochloraz, pendimethalin, and 2hydroxyatrazine each was added to one portion of the sand and soil, respectively. This sand/soil as well as the solvent-free equivalents was mixed thoroughly. The extent of solvent evaporation from the solvent-spiked matrices was measured as the differences between the weight losses of the solvent-spiked and the solvent-free variant. Com- parable weights between variants were gained after 2 h of mixing and indicated a complete evaporation of the carrying solvent. Then, spiked and non-spiked variants of sand and soil were mixed together, left three days in a fume hood to let the compounds associate with the solids, and then used in further experiments. 2.3. Dissipation The total contents of chemicals in soils and sand were measured at the beginning of the plant accumulation test (day 0 of the experiment) and on days 12, 40, and 90. The experiment was conducted in a glass- house under the conditions of a controlled temperature (15–22 °C) and air humidity (85%). Dissipation curves demonstrating the temporal changes in the total contents of pesticides applied to control soil/sand as well as of the field-aged pesticide residues during the plant accumula- tion test were constructed and used to derive pesticides' half-lives (see Section 2.8 (Data Evaluation and Statistical Analysis)). 2.4. Plant uptake

A portion equivalent to 1500 gdw of each soil was placed into boxes (20 × 30 cm) to which six seeds of Lactuca sativa lettuce (pre-cultivated on moistened cotton for 24 h) were added. Boxes were placed into a glasshouse where experiments ran under a controlled temperature (15–22 °C), air humidity (85%), and a photoperiod of 10 h light/14 h dark. Lettuce growth was checked and water content (loss controlled by box weighting) replenished every other day. On day 90, lettuce was sampled, washed in tap water, gently dried, lyophilized, and ana- lyzed for the content of target chemicals in leaves and roots separately as described in Section 2.7 (Sample Extraction and Analysis). 2.5. Solid phase microextraction (SPME) SPME fibers coated with polydimethylsiloxane (PDMS, thickness 30 μm, 13.55 ± 0.02 μL PDMS per meter of fiber) were purchased from Polymicro Technologies Inc. (USA) and cut into pieces 4 cm long. Prior to exposure, fibers were cleaned with methanol (2 × 24 h) and deionized water (2 × 24 h). Twelve fibers were buried into each soil used in the plant accumulation test. After 12, 40 and 90 days, 4 fibers were carefully removed from the soil, cleaned gently with a wet tissue to remove adherent soil particles, immersed in 20 mL of methanol amended with a surrogate standard (metolachlor, 500 ng per sample), and extracted by shaking for 48 h. Then, the volume of the extract was adjusted to 0.5 mL under nitrogen and analyzed as shown in the Supplementary Information (SI, Sample analysis and Table S3). 2.6. Earthworm uptake Bioaccumulation test was performed with the earthworm Eisenia andrei in 0.5 L glass jars. Ten adult (with a welldeveloped clitellum) earthworms (in triplicates) were weighted and placed in moistened (50% WHC) soil or sand equivalent to 150 gdw. Jars were covered with perforated lids (to enable aeration) and kept in the dark at 20 ± 2 °C. Every other day, water loss was checked by weighting the jars and, if necessary, the water content adjusted to the original level. The experi- ment was carried out for 14 days and for 21 days. The exposure period of 14 days was selected based on the results of our preliminary kinetic experiment (Svobodová et al., 2018) where the steady state of two cur- rently used pesticides in earthworms was reached within 14 days. This exposure period is also recommended by OECD Guideline 207 on the acute toxicity test with earthworms (OECD, 1984). The exposure period of 21 days followed the recommendation of OECD Guideline 317 on the bioaccumulation test with earthworms (OECD, 2010). At the end of ex- posure, earthworms were taken from soil and placed on a moistened fil- ter paper in a Petri dish for 24 h to empty their guts. After that, earthworms were rinsed with water, dried, weighted, lyophilized, weighted again and analyzed for the content of target chemicals as de- scribed in Section 2.7 (Sample Extraction and Analysis). 2.7. Sample extraction and analysis Soil, sand, plant, and earthworm samples were analyzed for the total content of chemicals using the QuEChERS extraction method and commercial extraction kits (Agilent Technologies, USA). This method has been successfully used before for the extraction of the compounds studied (Anastassiades et al., 2003; Bruzzoniti et al., 2014; Lesueur et al., 2008; Sivaperumal et al., 2015; Yu et al., 2006). The extraction procedure was performed as follows: 5.0 ± 0.1 gdw of soil/sand, 0.55 ± 0.01 gdw of lettuce or 0.45–0.75 gdw of earthworms were shaken with 5 mL, 9.45 mL or 9.25–9.55 mL, re- spectively, of deionized water and 10 mL of acetonitrile amended with metolachlor as a surrogate standard (500 ng per sample). Sam- ples were further amended with 6.5 g of QuEChERS Extract Pouch (MgSO4, NaCl, HOC(COONa)(CH2COONa)2 • 2H2O, HOC(COOH) (CH2COONa)2 • 1.5H2O). The mixture was shaken by hand for 1 min and then sonicated for 15 min. After shaking, plant and earthworm samples were cleaned by dispersive solid phase micro- extraction (PSA – primary and secondary amine exchange material, MgSO4). Then, samples were centrifuged (5 min, 3000 rpm) and an aliquot of 1 mL of acetonitrile was taken for analysis using high- performance liquid chromatography tandem mass spectrometry (HPLC-MS/MS). Details on the analysis are provided in the Supple- mentary Information (SI, Sample analysis and Table S3). The extrac- tion efficiency was on average 97% ± 14% for soil samples, 79% ± 6.9% for SPME fibers, 88% ± 15% for lettuce samples, and 98% ±

6.0% for earthworm samples. All results were corrected for it. 2.8. Data evaluation and statistical analysis Dissipation of chemicals was modeled using non-linear regression where data were fitted with the one phase exponential decay equation: C = C0 • e(K.t) (GraphPad Prism 5, GraphPad Software, Inc., USA), where C is the concentration in soil (ng/gdw) on the respective sampling day (day), C0 is the initial soil concentration (ng/gdw), and K is the degradation rate constant (day−1). From the dissipation curves, the half-lives (DT50) of the tested compounds were derived. They provided a quantitative measure of compound persistency. The significance of the differences between test variants were assessed using the STATISTICA software (t-test, p = 0.05). Bioconcentration (BCF) and bio- accumulation factors (BAF) were calculated as the concentration in tis- sue (ng/gww) divided by the concentration in soils (ng/gdw) at the particular sampling timepoint. The threshold values (3.0 ng/gww lettuce roots, 1.3 ng/gww lettuce shoots, 1.5 μg/mL PDMS) provided in Tables 2 and 3 represent the low- est concentrations in the lettuce roots, lettuce shoots, and SPME fibers, respectively, that could be reliably measured by the HPLC-MS/MS method based on its limit of quantification (1 ng/mL of an extract) and the matrix weight extracted per sample. Accordingly, the threshold values for bioconcentration and bioaccumulation factors (1.5, 0.65 and 1.6, respectively) in Tables 2 and 4, respectively, were calculated as the threshold value for the concentration in tissue divided by the threshold value for the concentration in soil. It follows that if the thresh- old values are presented in table cells, it means that the compound was present in the soil above the threshold but did not accumulate in tissue to quantifiable levels. Empty cells indicate that the compound was not considered in the experiment since based on the previous screening (Hvězdová et al., 2018), its concentration in soil was below the thresh- old given by the limit of quantification. 3. Results and discussion 3.1. Concentration of compounds in soils and sand and their dissipation At the beginning of the dissipation test, the concentrations in spiked matrices were in the range of 0.065–0.095 mg/kg for the contaminated sand (CS) and of 0.064–0.102 mg/kg for the laboratory contaminated (LC) soil (Table S2). During 90 days, the content of chemicals in the lab- oratory contaminated sand decreased to 37%, 32%, 31%, 7%, 9%, and 51% of the initial concentrations for epoxiconazole, tebuconazole, flusilazole, prochloraz, pendimethalin and 2-HAT, respectively (Table S4, Fig. S1). In the laboratory contaminated soil, 63%, 63%, 61%, 42%, 14%, and 72% of the initial concentrations were found after 90 days for epoxiconazole, tebuconazole, flusilazole, prochloraz, pendimethalin and 2-HAT, respec- tively (Table S4, Fig. S1). Temporal changes in total concentrations were observed also for the field-contaminated soils, albeit to a lesser extent. For example, in FC01 that contained aged conazoles, the concentrations of epoxiconazole and flusilazole decreased to 75% and 78%, respectively, of tebuconazole to 60% and of prochloraz to 43% of the initial concentration (Table S4, Fig. S1). The highest decrease (i.e., 78%) in soil concentrations within the 90-day dissipation experiment was reported for pendimethalin in soil FC02 (Table S4, Fig. S1). Interestingly, the concentration of epoxiconazole in this soil decreased by only 25%. Finally, 2-HAT was found to be stable in soil FC03 while its concentration was reduced to 66% within 90 days in soil FC04 (Table S4, Fig. S1). The model-derived DT50 values ranged from 17 days for pendimethalin in control sand to 4537 days for 2-HAT in soil FC03 (Table 1). In some cases, the DT50 values could not be reliably deter- mined as indicated by a large confidence interval and a low goodness of fit (r). DT50 values derived for laboratory-spiked conazole fungicides in the control soil were similar to each other, i.e., 114 days for epoxiconazole, 113 days for tebuconazole, and 115 days for flusilazole. This was expected as these compounds have similar physico-chemical properties such as logKow values (3.7, 3.87 and 3.3, respectively). In contrast, prochloraz – though it also belongs to the group of conazole fungicides – dissipated from the control matrices twice as fast as compared to the other conazoles (Table 1). This may be related to the different

Table 1 Half-lives (DT50 values, day) of compounds in field-contaminated soils (FC01–04), laboratory-contaminated sand (CS) and laboratory-contaminated soil (LC) derived by fitting the dis- sipation data (n = 12) with non-linear regression (one phase exponential decay model). Values in brackets represent the 95% confidence interval (CI); inf means infinity; r characterizes the goodness of fit of the model used. substance groups where prochloraz is an imidazole while the other compounds are triazoles. In addition, in contrast to triazoles, prochloraz is very sensitive to photodegradation with DT50 of aqueous photolysis being only 1.5 days (Lewis et al., 2016). For field-contaminated soils, the DT50 values of aged conazole fungicides varied to a greater extent (see FC01 soil in Table 1) and tended to be higher in comparison to those determined for freshly spiked (nonaged) counterparts in the control soil. Nonetheless, the differences in the DT50 values between the freshly spiked and aged compounds were not significant (see the overlapping confidence intervals in Table 1) and corresponded to the literature values where, for example, tebuconazole half-lives varied from 49 (Bending et al., 2007) to 610 days (Strickland et al., 2004). DT50 values of pendimethalin did not differ between the FC02 soil and the LC soil, which suggests that the residence time did not signifi- cantly affect pendimethalin dissipation in soils. The measured DT50 values were in good agreement with the literature values, i.e., 11 days in Sondhia (2012), 12 and 13 days in Chopra et al. (2015), 20 days in Sondhia (2013) and from 24 to 34 days in Kočárek et al. (2016). The half-lives obtained for 2-HAT in soils FC04 (163 days) and LC (146 days) and in sand CS (95 days) were in accordance with those re- ported in the literature, i.e., 121 days in Winkelmann and Klaine (1991) and 164 days in Lewis (2016). In contrast, 2-HAT half-life in FC03 soil was very long (4537 days), which suggests that the extent of dissipation within 90 days was negligible in this soil. This likely resulted from the formation of bound residues in this soil within the 10-yearpost- banperiod. In Winkelmann and Klaine (1991) bound residue formation was observed in the case of radiolabeled 2-HAT and reached 28% during 190 days. The persistence of 2-HAT could have also been enhanced in this study due to the lower pH of the FC03 soil in comparison to the other tested soils as the formation and/or retention of hydroxy- metabolite residues was shown to be related to low soil pH values (Scherr et al., 2017). To sum up, the dissipation data clearly indicated that in terms of dissipation, differences exist between the various groups of chemicals (e.g., conazoles vs pendimethalin) as well as be- tween individual chemicals in different soils (e.g., 2-HAT in LC soil vs FC03 soil). In contrast, pesticides of the same group (i.e., triazoles) are likely to exhibit similar dissipation behavior that did not seem to be af- fected by the type of matrix (soil vs sand) or the aging periods (e.g., freshly added tebuconazole in LC soil vs aged tebuconazole in FC 04 soil, both showing similar reduction in their contents after 90 days). The data on field soils also suggests that even the aged pesti- cide residues tended in most cases to further dissipate from soils. (Plant uptake section). Uptake to lettuce was considered separately for leaves and roots where the concentrations in both matrices measured on day 90 of the experiment are presented in Table 2. In the field contaminated soils, only epoxiconazole and pendimethalin were taken up to reach quantifi- able levels. The measured concentration reached 5.8 ng/gww for epoxiconazole and 70.2 ng/gww for pendimethalin. Concentrations of the other compounds in lettuce roots were below the limit of quantifi- cation though the compounds were present in soils the lettuce was ex- posed to. In shoots, none of tested chemicals were present at quantifiable levels, which suggest that the aged chemicals' potential for biouptake and bioconcentration was generally low. In the roots of lettuce exposed to the LC soil containing freshly added chemicals, epoxiconazole (22.1 ng/gww), tebuconazole (17.3 ng/gww), flusilazole (18.8 ng/gww), as well as pendimethalin (72.2 ng/gww) were found. Uptake by shoots was reported only for pendimethalin (1.5 ng/gww). A comparison between root and shoot concentrations suggests that the tested compounds had a tendency to accumulate in the lipid-rich roots with a limited potential for being translocated to shoots. The translocation from roots to shoots is generally considered to be restricted and less likely with increasing hydrophobicity of the compound (Felizeter et al., 2012). Low concentrations of the tested chemicals in lettuce observed in this study may also arise from com- pound degradation upon its uptake by plant. As reported by Im et al. (2016), the half-life of flusilazole in lettuce is only 4 days. Given that the characteristics of flusilazole are similar to other conazoles (Table S1), we may assume that all conazoles could be rapidly elimi- nated from lettuce. This in combination with the extensive dissipation of some compounds (e.g., prochloraz) in the studied soils can explain the observed low amounts of conazoles in the lettuce plant. To provide bioconcentration factors (BCFs), the concentrations in shoots and roots were normalized to the soil residual concentration measured on day 90 (Table 2). The highest BCFs (roots) were observed for pendimethalin,

i.e., 53 and 69 in the FC02 and the LC soil, respec- tively. These BCF-roots values were one-order of magnitude higher than those calculated for the other tested chemicals. BCFs of epoxiconazole in FC02 soil (5.1) were comparable to that in the labora- tory contaminated soil LC (5.2). BCFs for tebuconazole and flusilazole in the later soil were 4.1 and 3.7, respectively. Following the results, it appears that aged chemicals (except for pendimethalin) tended to be less biologically available for uptake by let- tuce in comparison to their freshly added counterparts. In neither case, including pendimethalin, were the concentrations in the upper edible plants found to exceed the maximum residue levels for pendimethalin (0.1 mg/kgww in EU, 4 mg/kgww in USA and 0.1 mg/kgww in China – see Table S5). This suggests that contamination of the food web is un- likely, and also that the quality of crops is unlikely to be affected by pes- ticide residues that have persisted into the following growing season. 3.2. SPME uptake Uptake to SPME fibers was expressed as the concentration of a com- pound (μg) in PDMS coating (mL) for each sampling timepoint (i.e., 12, 40 and 90 days). PDMS-SPME concentrations are presented in Table 3 Table 2. concentrations of the target chemicals in the shoots and roots (wet weight) of lettuce Lactuca sativa measured on day 90 and the corresponding bioconcentration factors (BCFs) calculated as the concentration in shoots or roots (ng/gww) divided by the residual soil concentration on day 90 (ng/gdw). and were used as a measure of exposure via soil porewater. SPME con- centrations ranged from b1.5 to 99.6 μg/mL in control sand (CS). In some cases, e.g. in soil FC 01, the concentrations in SPME fibers were shown to drop below the LOQs at the prolonged exposure period(s). The temporal changes reflected depletion of the porewater occurring as a result of compound dissipation from soil (Fig. S1 and Table S4). Regarding the field contaminated soils, uptake by fibers was signifi- cant only in a few cases, such as for 2-HAT (5.0 μg/mL) in FC 04 soil on day 90, for pendimethalin (3.6–5.8 μg/mL) in FC02 soil, and for conazoles (1.8– 2.8 μg/mL) in FC 01 soil on day 12. In the control soil (LC), concentrations of compounds in SPME fibers on day 12 ranged from b1.5 to 7.5 μg/mL for pendimethalin. In comparison to the control soil, concentrations in SPME fibers exposed to the control sand (CS) were one order of magnitude higher, except for 2-HAT which showed similar SPME concentrations irrespective of the matrix. To provide a measure of compound partitioning between the freely dissolved and the solids-bound fractions, SPME concentrations were normalized to the residual concentration in solids at the respective time point to give the KSPME-solids values (Table 3). As hypothesized, the highest KSPME-solids values were observed for compounds in the control sand where chemicals are mostly present as freely dissolved in the aqueous phase and biologically available. The KSPME-soil values ranged from 2.4 to 631. Unfortunately, it could not be concluded whether there were trends in K values across compounds, soils or the aging periods as limited data were collected for the field-aged compounds. 3.3. Earthworm uptake BAF values of chemicals measured on day 14 and on day 21 in the control (contaminated sand) did not statistically (p N 0.05) differ from each other. Therefore, we may assume that the incubation period of 14 days was sufficient for the tested chemicals to reach equilibrium con- centrations in earthworms. BAFs (day 14) in control sand varied from low (1.9 for 2-HAT) to substantial (695 for pendimethalin) and are pre- sented in Table 4. Of the studied chemicals, 2-HAT showed the lowest potential for bioaccumulation in earthworms. Bioaccumulation of 2- HAT was low even for the weakest sorbent tested, i.e., the contaminated sand (Table 4). This finding is consistent with the low hydrophobicity of Table 3 Concentrations of compounds in the PDMS-SPME fiber (μg/mL of PDMS) exposed to field-contaminated soils (FC01–04), laboratory-contaminated sand (CS) and laboratory-contaminated soil (LC) and sampled on days 12, 40, and 90 (n = 3) and corresponding partition coefficients between PDMS-SPME fiber and soil (KPDMS-soil). 2-HAT is 2-hydroxyatrazine.

Table 4 Bioaccumulation factors (Cearthworm/Csoil residual, (ng/gwet weight)/(ng/gdry weight), mean ± the standard deviation) of tested chemicals in field-contaminated soils (FC01–04) and laboratory- contaminated sand (CS) following the incubation period of 14 days or of 21 days. 2-HAT (Table S1) that does not favor bioaccumulation in hydrophobic lipids. As such, 2-HAT is not expected to pose risks to terrestrial biota in terms of bioaccumulation in lipids. In contrast to 2-HAT, epoxiconazole, tebuconazole, and pendimethalin accumulated in earth- worms to significant extents with BAF values (21 days) of up to 695, 557, and 210, respectively. The trend in bioaccumulation was in line with the compounds' hydrophobicity (Table S1). Prochloraz and flusilazole showed moderate tendencies for bioaccumulation if freshly spiked to sand (BAF values of 23 and 20, respectively) while bioaccumu- lation of their aged pesticides was negligible. With the exception of pendimethalin in soil FC02, no other com- pound was shown to significantly bioaccumulate in earthworm tissues from the field-contaminated soils. When comparing the BAF values of aged pendimethalin for the field-contaminated FC02 soil (i.e., 2.1 on both day 14 and 21) with BAFs for the freshly-spiked control sand (250 on day 14 and 695 on day 21), it was observed that aged pendimethalin is about two orders of magnitude less available then its counterpart freshly added to sand. The relatively low organic carbon (OC) of 2.0% in the FC02 is unlikely to explain this two orders of magni- tude difference in bioaccumulation, suggesting that other mechanisms act against bioaccumulation. Typically, reduced bioaccumulation in soils is reported to be a result of compound sequestration both in time and space (Alexander et al., 1997; Northcott and Jones, 2001). Sequestration mechanisms are hypothesized to include reversible sorption as well as diffusion into remote pores of the soil matrix (Pignatello and Xing, 1995). Examples from the literature suggest that reduced bioaccu- mulation in earthworms resulting from compound sequestration was reported also for pesticides studied here such as pendimethalin (Belden et al., 2003) and epoxiconazole (Nélieu et al., 2016). The extent of BAF reduction due to sequestration was reported by Belden et al. (2003), who found that BAF decreased from 1.9 to 0.5 upon 340 days of pendimethalin aging in soil. 3.4. Partitioning versus uptake to biota and implications When considering the behavior of individual contaminants (or groups) in the field contaminated soils and in the freshly spiked sand/ soil, several interesting trends can be derived following the data on dis- sipation, bioconcentration, bioaccumulation, and SPME. For example, conazole pesticides aged in field contaminated soils did not accumulate in the tested biomatrices (with the only exception of epoxiconazole in soil FC02). The SPME results gained for conazole pesticides in the con- trol soil (LC) appeared in line with the root uptake data. Both methods detected the presence of epoxiconazole, tebuconazole and flusilazole at comparable levels to each other. Prochloraz (probably due to its fast dissipation from soils) was not found in the fibers nor in roots. Similar- ities were found also between SPME and Eisenia andrei results, both de- tecting the presence of quantifiable amounts of conazoles on day 12 and 14, respectively. As for pendimethalin, SPME seems to indicate well its potential for root uptake in both field contaminated (FC02) and labora- tory contaminated (LC) soils as well as for earthworm uptake. Combin- ing the data on 2-HAT, it follows that 2-HAT's availability for uptake by fibers or biological endpoints was generally limited. This was likely due to its strong sorption to control soil and slow release from field soils. Taking the results together, it appears that pesticide residues present into the following growing season(s) at levels of ≤0.1 mg/kg are not ex- pected to pose significant risks to the environment. Based on the find- ings reported here, this implies that the frequent occurrence of residues at such levels (Hvězdová et al., 2018) should not necessarily provoke follow-up actions, such as restricted use or replacement by new “less harmful” compounds, as the actual risks may be acceptable. Nevertheless, since the combination of factors that influences the be- havior of residues is limitless, making the fate of the residues difficult to foresee, both monitoring of pesticide occurrence and bioavailability testing should advantageously be employed within post-application as- sessment of pesticide risks and when revising the marketed active in- gredients in pesticide products. This should ensure that the risks associated with pesticide occurrence in the soil environment are realis- tically considered in the long-term and that these risks are well balanced with the benefits that these pesticides bring.

4. Conclusion According to this study, it seems that the presence of pesticide resi- dues and of 2-hydroxyatrazine studied here is less problematic than as- sumed on the basis of their wide and frequent occurrence in agricultural soils (Hvězdová et al., 2018). This is mostly due to the limited bioavail- ability of the potentially hazardous compounds, resulting in low poten- tial for bioaccumulation in non-target biota such as earthworms and the lettuce plant. In this study, no obvious evidence was found that persis- tence of chemical residues into the following season(s) is associated with significant environmental risks. On contrary, it can be expected that the concentrations of pesticide residues are likely to further de- crease a result of natural dissipation processes in soil and thus, reduced exposure to soil biota and crops is anticipated. Another factor likely diminishing the risks of pesticide residue is their ability to become se- questered in soil over time, where as a result the compounds become even less biologically available to higher organisms and plants than at the time of their application.