DRY–WET CYCLES INCREASE PESTICIDE RESIDUE RELEASE FROM SOIL

Soil drying and rewetting may alter the release and availability of aged pesticide residues in soils. A laboratory experiment was conducted to evaluate the influence of soil drying and wetting on the release of pesticide residues. Soil containing environmentally long-term aged (9–17 years) 14C-labeled residues of the herbicides ethidimuron (ETD) and methabenzthiazuron (MBT) and the fungicide anilazine (ANI) showed a significantly higher release of 14C activity in water extracts of previously dried soil compared to constantly moistened soil throughout all samples (ETD: p < 0.1, MBT and ANI: p < 0.01). The extracted 14C activity accounted for 44% (ETD), 15% (MBT), and 20% (ANI) of total residual 14C activity in the samples after 20 successive dry–wet cycles, in contrast to 15% (ETD), 5% (MBT), and 6% (ANI) in extracts of constantly moistened soils. In the dry–wet soils, the dissolved organic carbon (DOC) content correlated with the measured 14C activity in the aqueous liquids and indicated a potential association of DOC with the pesticide molecules. Liquid chromatography MS/MS analyses of the water extracts of dry–wet soils revealed ETD and MBT in detectable amounts, accounting for 1.83 and 0.01%, respectively, of total applied water-extractable parent compound per soil layer. These findings demonstrate a potential remobilization of environmentally aged pesticide residue fractions from soils due to abiotic stresses such as wet–dry cycles. Environ. Toxicol. Chem. 2012; 31: 1941–1947. © 2012 SETAC


INTRODUCTION
In agriculture, approximately 80% of all pesticides are used for crop protection, and between 1995 and 2007 pesticide use exceeded 5.0 billion pounds (2.3 million metric tons) annually worldwide [1][2][3][4][5]. Due to the large quantities of pesticides applied annually to fields, the receiving surface soils function as sinks for a number of chemicals and as buffers for deeper soil layers and aquifers. Surface soils exposed to pesticide applications are also subject to various abiotic influences such as repeated and strong drying and wetting cycles and mechanical disruption such as tillage, which influence the fate and disposition of pesticides in soils. Therefore, the impact of changing climatic conditions is also a subject of growing concern with regard to the distribution of chemical pollutants in the environment [6].
Soil organic carbon (SOC) plays an important role in soil fertility and pesticide fate. The effect of soil drying and wetting on the release of SOC has been described elsewhere [7], mainly with regard to soil aggregation, soil aggregate stability [8][9][10], and decomposition of SOC [7,8,[11][12][13][14]. A decline in SOC is a matter of growing concern [15], and it is likely to be exacerbated by intensified agricultural production and changing climatic conditions [16], with potentially adverse consequences on soil fertility, soil stability, and food production. Soil organic carbon has a substantial influence on organic pesticide sorption and retention, which influence the fate of these substances in the environment. As a consequence, it can be assumed that a decrease in SOC might also result in increased mobilization of pesticide residues in soils due to missing binding sites. Soil disaggregation, promoted by dry-wet cycles, results in increased release of SOC [9,10] and release of associated pesticide residues, as previously suggested [17]. In turn, released SOC might be subject to increased microbial mineralization [12,18]. In addition, microbially degradable carbon sources in soil may promote the degradation of pesticides cometabolically.
Experiments on the long-term fate of aged 14 C-labeled pesticides in soils are rarely conducted. Outdoor lysimeter studies using 14 C-labeled pesticides provide important information in this respect. The influence of dry-wet cycles of soil on the water extractability and release of long term-aged 14 C-labeled pesticide residues has not been systematically tested and will provide important information about the remobilization of pesticide residues under changing abiotic conditions. Therefore, the present study investigated the influence of successive dry-wet cycles under laboratory conditions on the release of 14 C-labeled residues of the thiadiazolylurea herbicide ethidimuron (ETD; 1-[5-ethylsulfonyl-1,3,4-thiadiazol-2-yl]-1,3-dimethylurea), the dimethylurea herbicide methabenzthiazuron (MBT; 1-[1,3-benzothiazol-2-yl]-1,3-dimethylurea), and the triazine fungicide anilazine (ANI; 4,6-dichloro-N-[2-chlorophenyl]-1,3,5-triazin-2-amine) ( Fig. 1) in soils aged under environmental conditions for 9 to 17 years. The following aims were addressed: (1) assessment of the influence of successive soil dry-wet cycles compared to permanently moistened soil on the water extractability of the aged 14 C-labeled pesticide residues, (2) quantitative and qualitative analyses of the pesticide residues in the soil water extracts, and (3) evaluation of the influence of dry-wet cycles on SOC and total nitrogen (TN) extractability.

Soil and lysimeter history
All outdoor lysimeters had a surface area of 1 m 2 and consisted of an undisturbed soil column (soil depth 1.1 m) of an Orthic Luvisol (C org : 1.2%, sand: 6.4%, silt: 78.2%, clay: 15.4%; pH: 7.2) [19][20][21]. The time and amount of 14 C-pesticide applications on the lysimeter soils are presented in Table 1. The chemical properties of the parent pesticide compounds are summarized in Table 2.
Soil sampling and 14 C-residue detection A total of approximately 2 kg of each individual lysimeter soil was sampled randomly in April 2006 from the soil layer at 0 to 30 cm depth of the ETD, MBT, and ANI lysimeter, using a Humax stainless steel soil core sampler, 3 cm in diameter. Because no ploughing simulation was applied on the ETD soil within six years prior to sampling as was performed for MBT and ANI soils, ETD soil was subdivided into 10-cm layers and only the soil from 0 to 10 cm depth was used for this experiment. For further 14 C-residue calculation, estimated bulk soil densities of 1.5 g cm À3 for ETD soil and 1.3 g cm À3 for MBT and ANI soil were assumed. Soils were air-dried to a residual moisture content of 7 to 12% and then sieved (2 mm), homogenized, and stored in the dark at 38C. To detect residual 14 C activity in soils, samples of 50 g were dried at 1058C to accelerate the experimental process and ground in a mortar. Volatilization of pesticide residues during the drying process at 1058C can be excluded because results were compared with those obtained from soil samples that were air-dried only and found to be not different. The very low vapor pressure for these compounds (ETD: < 0.001 mPa; MBT: %590 nPa; ANI: 910 nPa; all at 208C [22]) supports this observation. Subsamples were oxidized in five parallels of 1 g (Biological Oxidizer OX500; R.J. Harvey Instrument). Evolving 14 CO 2 was trapped (Oxysolve C-400 scintillation cocktail; Zinser Analytik), and radioactivity was quantified using a liquid scintillation counter with internal quench correction (2500 TR, Tri-Carb; Packard Liquid Scintillation Analyzer).

Soil organic carbon and nitrogen analyses in soil samples
The SOC and TN were determined in triplicates of freezedried and homogenized soil samples prior to water-shaking extractions, as described previously [23]. Organic C was determined by radiofrequency heating of a 100-mg sample in flowing oxygen and subsequent infrared absorption by a Leco RC-412 multiphase carbon determinator. A Leco TCH 600 was applied to determine N 2 by thermal conductivity detection using 2-mg samples.

Desorption experiments
For each soil, two parallel experiments (A and B) were established. Triplicates of 10 g for each soil containing the aged 14 C-labeled and nonlabeled ETD, MBT, and ANI residues were either (A) directly mixed with distilled water (1 þ 2, w:w) or (B) first oven-dried at 458C until dryness before adding water (1 þ 2, w:w). All samples were simultaneously shaken for 1 h at 150 rpm on a horizontal shaker at room temperature (21 AE 28C) and subsequently centrifuged to separate the solids from the water phase (60 min, 3,000 rpm, equal to approximately 2,800 g; Allegra 6KR, GH-3.7 Horizontal Rotor; Beckmann Coulter). The resulting supernatants were filtered (0.45 mm, Porafil MV; Macherey-Nagel) to remove potential particulates from all water-extract samples prior to the detection of dissolved 14 C activity, dissolved organic carbon (DOC), dissolved TN, pH, and electrical conductivity. For setup A, all moistened soil samples were stored after centrifugation and removal of the supernatant in the dark at 38C, until all samples of setup B were again completely dried at 458C (approximate drying time 3-4 d), and the cycle could be repeated. Both setups A and B were subject to 20 successive water-shaking extractions.
Subsequent to the entire experiment, total 14 C recovery was calculated for all individual soils as a sum of total 14 C activity  extracted from each soil sample and the residual 14 C activity remaining in the water-extracted soil samples, as well as 14 C activity adsorbed to the filters.
Determination of water-extracted 14 C activity The amount of dissolved 14 C activity was determined in triplicates for each soil extract by mixing 1 ml of the liquid samples with 10 ml of scintillation cocktail (Instant Scint-Gel Plus; PerkinElmer), and radioactivity was detected by liquid scintillation counting. An external standard was used for quenching correction using 1 ml of distilled water with 10 ml of scintillation cocktail.

Carbon and nitrogen analysis of water extracts
All aqueous samples were analyzed for total DOC and dissolved TN content, using a Shimadzu Total Organic Carbon/Nitrogen Analyzer (TOC-5050A, ASI-5000A Auto Sampler).
Liquid chromatography atmospheric pressure chemical ionization-tandem mass spectrometry analysis. Liquid chromatography-atmospheric pressure chemical ionization-tandem mass spectrometry analysis (LC-APCI-MS/MS) was applied in accordance with our method described elsewhere [24]. Briefly, an Agilent 1100 series HPLC coupled with a Thermo Electron TSQ Quantum triple quadrupole mass spectrometer was used. Liquid chromatographic separations were carried out with a Phenomenex Synergi 4m Polar RP18 column, 150 Â 3.0 mm I.D. A 1 mM ammonium acetate þ 0.1% formic acid (A)/acetonitrile þ 0.1% formic acid (B) gradient was applied for ETD and MBT, while ANI and ANI metabolites were separated with 5 mM ammonium acetate (A) and acetonitrile (B). The mass spectrometer was operated in the positive atmospheric pressure chemical ionization (APCI(þ)) mode for the detection of ETD and MBT and in the negative atmospheric pressure chemical ionization (APCI(À)) mode in case of ANI and ANI metabolites. Multiple reaction monitoring was used to quantify all analytes and standards.
Calibration and quantification. Quantification was carried out using the internal standard method as stated previously [24]. Analysis of MBT/ETD was performed with IPU as an internal standard and that of ANI and ANI metabolites with d 5 -2hydroxyatrazine.

Electrical conductivity and pH analyses of water extracts
To determine the potential loss of soil minerals as a result of the successive dry-wet cycles, the electrical conductivity of previously filtered water extracts was measured using a WTW device (Cond 340i/SET, WTW). The pH of all filtered water extracts was monitored using a Metler Toledo pH meter (PTB01 ATEX 2166X, Type 1120X).

Statistical analysis
For statistical analysis, the independent two-sample t test was applied to determine the significance of differences between mean values.

Soil analyses
Even after long-term environmental aging (9-17 years) (Table 1), a major fraction of residual 14 C activity was still present in the upper soil layer of 0 to 10 cm depth for ETD soil and 0 to 30 cm depth for MBT and ANI soil, accounting for 18.7% (ETD soil), 34.8% (MBT soil), and 43.2% (ANI soil) of total initially applied 14 C activity (Table 3). Considering additionally the detected residual 14 C activity in the ETD soil layers at 10 to 20 and 20 to 30 cm depths, the residual detected 14 C activity accounted for 16.4 and 12.2% (data not shown), respectively. In this case, the total residual 14 C activity accounted for 47.3% of initially applied 14 C activity in the entire ETD soil layer at a depth of 0 to 30 cm. As presented in another environmental long-term study using 14 C-labeled atrazine [23,25,26], most of the residual 14 C activity was located in the upper 10-cm soil layer and is most likely associated with the higher organic carbon content in this soil layer [26][27][28].
The SOC and TN contents in the soils at the date of sampling accounted for 1.14 AE 0.06% and 0.15 AE 0.00%, respectively. Because all soil monoliths were taken at the same field plot, minor differences in SOC content can be attributed to potential differences in the soil management of the lysimeters [24].
An association of ETD, MBT, and ANI with SOC has previously been demonstrated [21,29]. Even after long-term environmental aging, SOC plays a major role in pesticide residue retention in the upper soil layers, where the SOC content is generally higher than in deeper layers.
A decrease in SOC content could be related to intensified agricultural production and climatic changes leading to a potential increase of SOC microbial turnover; however, current scientific data on SOC decline in soils are rather inconsistent with regard to this issue [30][31][32][33][34]. A decrease in SOC could be facilitated by soil drying and rewetting due to a subsequent increase in microbial SOC turnover [35]. Reemtsma et al. [17] suggested that the contaminant release from soils Pesticide residues released from soil increased by dry-wet cycles Environ. Toxicol. Chem. 31,2012 might be influenced by microbial activity over time, altering the long-term function of soils as a sink or source of organic contaminants.
Considering a potential decline in SOC in agricultural soils, it can be assumed that the buffering function of the soils for pesticides and their metabolites could be reduced. This might result in greater remobilization of pesticides in soils due to altered SOC content.

Analyses of water extracts
Liquid chromatography MS/MS analyses. Even after longterm environmental aging, liquid chromatography MS/MS analyses revealed the parent compounds ETD (nine years after application, Table 1) and MBT (12 years after application, Table 1) in water extracts of dry-wet soils, as presented in Figures 2 and 3, right y axis. The parent compound ETD was found in the first five water extracts of dry-wet soils and in the first two water extracts of constantly moistened soils, albeit in much smaller quantities (Fig. 2, right y axis), totaling 15.1 and 2.8 mg kg À1 of the water-extractable parent compound ETD, respectively ( Table 3). As shown in Table 3, these values correspond to 1.83 and 0.34%, respectively, of the initially applied parent compound.
In the case of MBT, the parent compound was detected in the first four water extracts of dry-wet soil samples, totaling 0.23 mg kg À1 water-extractable parent compound MBT (Table 3). No MBT was detected in water extracts of constantly moistened soils (Fig. 3, right y axis). The detected amount of water-extractable MBT equals 0.01% of the initially applied parent compound.
This observation must be attributed to a combination of a generally higher water solubility of ETD (Table 2; 3.0 g ETD vs 0.059 g MBT L À1 at 208C), a low adsorption affinity to soils, as well as a much higher environmental persistence in soils compared to MBT [22,24,36,37]. Although ETD is much more water-soluble, its biodegradability is much lower than that of MBT, resulting in proposed environmental half-lives of 162 to 2,059 d compared to approximately 24 to 230 d for MBT [21,37,38]. Furthermore, MBT was found to be biodegradable [38][39][40], giving an additional explanation for the overall smaller MBT residue fractions in the water extracts. In contrast, the microbial ETD degradation in soil was found to be very limited [21].
In the case of ANI, neither the parent compound nor dihydroxy-ANI as its main metabolite was detected in the water extracts. This fact must be attributed to the very low water solubility of ANI (Table 2; 0.008 g L À1 at 208C), its high adsorption affinity to the soil matrix, and the extended environmental aging time of 17 years (Table 1) as a driving factor for the formation of soil-bound residues [22,29,41,42].
As shown in a previous study, drying and wetting soils containing phenanthrene and di(2-ethylhexyl)-phthalate reduced the biodegradability, extractability, and uptake by earthworms [43]. In our case, however, the parent pesticide compounds ETD and MBT were detected in the water phase after long-term aging and thus could be more accessible for biodegradation or uptake. Generally, the increased release of aged pesticide residues from dried and rewetted soils could be due to disruption of soil aggregates, facilitating the release of entrapped pesticide molecules, which are normally excluded from remobilization under moistened conditions. These observations might also be valid for a number of other pesticide residues in soils, which could be remobilized by environmentally relevant dry-wet cycles. These data help to assess the chemical nature of long-term aged pesticide residues in soils and their remobilization potential. Biological and toxicological effects such as endocrine disrupting activity, as discussed for small concentrations of atrazine [44], are so far not published for the pesticide compounds described herein. Therefore, further research is needed to evaluate the biological or toxicological relevance of our findings.
Water-extracted residual 14 C activity and DOC/TN determination. The results showed a significant increase in desorbed 14 C activity in water extracts of all soils after drying, accounting for 44% (ETD), 15% (MBT), and 20% (ANI) of residual 14 C activity in soil samples after 20 dry-wet cycles compared to water extracts of constantly moistened soils (Figs. 2-4, left  14 C activity and the detected parent compound were significantly higher in dry-wet versus constantly moistened soil water extracts ( p < 0.001, except cycle 20: p < 0.01 and cycles 8 and 14: p < 0.1). Standard deviation n ¼ 9. y axis). The amounts of water-extracted 14 C activity from constantly moistened soil remained significantly lower at 16% (ETD), 5% (MBT), and 6% (ANI) of residual 14 C activity in soil samples after 20 cycles. This observation clearly indicates that release of residual pesticide 14 C activity is strongly influenced by soil drying and rewetting, irrespective of the chemical pesticide class. However, the waterextractable amount of residual pesticide 14 C activity, and therefore the pesticide or its metabolites, is variable, depending on the specific water solubility of the individual compound. As shown in Figure 5a, c, and e, the DOC content was significantly higher in all water extracts obtained from soils exposed to dry-wet cycles compared to those from constantly moistened soils. This is in accordance with previous observations [7]. After 20 water-extraction cycles, the DOC content accounted for 11 AE 1% versus 5 AE 0.3% (ETD soil), 8 AE 0.4% versus 4 AE 0.3% (MBT soil), and 10 AE 1% versus 5 AE 0.2% (ANI soil) of the total organic soil carbon detected as DOC in dry-wet versus constantly moistened soil water extracts. The DOC content correlated positively with the measured 14 C activity in the aqueous liquids, and the correlation was more pronounced in dry-wet soil water extracts than in water extracts of constantly moistened soils (r ¼ 0.80-0.91 vs r ¼ 0.41-0.70).
The overall observation of the increased water extractability of residual pesticide 14 C activity and DOC from dry-wet soils compared to constantly moistened soils must be attributed to a physical disruption of soil aggregates, hosting residual pesticide molecules and potentially associated DOC in micropores. A physical disruption of soil aggregates due to dry-wet cycles resulting in extended release of DOC has also been discussed previously [7,8]. Throughout 20 successive dry-wet cycles, continuous disaggregation of soil particles resulted in higher release of residual pesticide 14 C activity and DOC. Each drywet cycle changed the stability of aggregates. Air-dried aggregates are stable, and an increase in water potential decreases the stability, causing a new equilibrium to be established [10]. This new equilibrium generally results in a reduced amount of protected SOC and an increased amount of rather unprotected DOC. Residues of organic pollutants are part of this cycling.
At this point, the binding mechanisms and the nature of DOC-associated pesticide residues remain unknown and need further investigation. However, the wet-dry cycles induced the release of SOC. As such, the pesticides went into solution as well, where they might be subject to degradation.
The release of TN was highly pronounced in the first extracts of both dry-wet and constantly moistened soil water extracts in all cases (Fig. 5b,d,f). It is likely that water-dissolvable inorganic nitrogen salts, such as potassium and sodium nitrates, were dissolved preferentially in the first water extracts. Throughout 20 successive dry-wet cycles, the measured TN content remained significantly higher in dry-wet versus constantly moistened soil water extracts. In accordance with the DOC contents in dry-wet soil water extracts, it can also be assumed that the higher TN values in these soil water extracts was related to increased mobilization of organic nitrogen fractions. This assumption can be supported by the small but relatively higher correlation between the TN and DOC contents in dry-wet soil water extracts compared to water extracts of constantly moistened soils for all samples (r ¼ 0.14-0.23 vs r ¼ À0.07-0.05).
Electrical conductivity and pH. The differences between the electrical conductivity in dry-wet and constantly moistened soil water extracts were small but statistically significant ( p ¼ 0.01-0.1).
The highest values were obtained in the first water extracts, accounting for 656 mS cm À1 in dry-wet versus 613 mS cm À1 in moist soil water extract for ETD soil, 625 versus 632 mS cm À1 for MBT soil, and 917 versus 913 mS cm À1 for ANI soil. In the second extract obtained after water-shaking extraction, electrical conductivity decreased significantly by a factor of four to five in all samples, leveling off to 20 to 38 mS cm À1 in water extracts of constantly moistened soils versus 40 to 45 mS cm À1 in dry-wet water extracts. This finding indicates that dry-wet cycles in soils moderately trigger the release of ions and micronutrients. The pH value remained constant for all water extracts, with an average of 6.6 AE 0.3, indicating that successive dry-wet or wet-wet cycles had no influence on the pH value in water extracts.
The overall results demonstrate long-term binding and sequestration mechanisms of organic pesticides in soils. The data suggest that environmentally relevant dry-wet cycles may Fig. 3. Water-extracted 14 C activity (left y axis, percentage of total residual 14 C activity) and liquid chromatography-MS/MS-detected compound (right y axis, in mg kg À1 soil) in water extracts of MBT soil. Values for 14 C activity and the detected parent compound were significantly higher in dry-wet versus constantly moistened soil water extracts ( p < 0.001, except cycle 1: p < 0.01). Standard deviation n ¼ 9. Fig. 4. Water-extracted 14 C activity (percentage of total residual 14 C activity) in water extracts of anilazine (ANI) soil. Values for 14 C activity were significantly higher in dry-wet versus constantly moistened soil water extracts ( p < 0.001, except cycles 1 and 10: p < 0.01). Standard deviation n ¼ 9.
change interfacial soil properties intensively. This may result in increased remobilization and release of aged pesticides or pesticide metabolites in surface soils. However, the experimental setup was designed to describe a maximum possible remobilization potential of the aged residual pesticide fractions present in our experimental soils.