Pendimethalin and oxyfluorfen degradation under two irrigation conditions over four years application

Abstract

A four-year field study was conducted to determine the effect of pluviometric conditions on pendimethalin and oxyfluorfen soil dynamics. Adsorption, dissipation and soil movement were studied in a sandy loam soil from 2003 to 2007. Pendimethalin and oxyfluorfen were applied every year on August at 1.33 and 0.75 kg ha⁻¹, respectively. Herbicide soil concentrations were determined at 0, 10, 20, 40, 90 and 340 days after application (DAA), under two pluviometric regimens, natural rainfall and irrigated (30 mm every 15 days during the first 90 DAA). More than 74% of the herbicide applied was detected at the top 2.5 cm layer for both herbicides, and none was detected at 10 cm or deeper. Pendimethalin soil half-life ranged from 10.5 to 31.5 days, and was affected mainly by the time interval between application and the first rain event. Pendimethalin soil residues at 90 DAA fluctuated from 2.5 to 13.8% of the initial amount applied, and it decreased to 2.4 and 8.6% at 340 DAA. Oxyfluorfen was more persistent than pendimethalin as indicated by its soil half-life which ranged from 34.3 to 52.3 days, affected primarily by the rain amount at the first rainfall after application. Oxyfluorfen soil residues at 90 DAA ranged from 16.7 to 34.8% and it decreased to 3.3 and 17.9% at 340 DAA. Based on half-life values, herbicide soil residues after one year, and soil depth reached by the herbicides, we conclude that both herbicides should be considered as low risk to contaminate groundwater. However, herbicide concentration at the top 2.5 cm layer should be considered in cases where runoff or soil erosion could occur, because of the potential for surface water contamination.

Keywords:

• Herbicides
• adsorption
• dissipation
• leaching
• vineyards

Introduction

Pendimethalin [N-(1-ethylpropyl)-2,6-dinitro-3,4-xylidine] and oxyfluorfen [2-chloro-α, α, α-trifluoro-p-tolyl 3-ethoxy-4-nitrophenyl ether] are pre- and pre- and post-emergence herbicides, respectively, that are used in vegetable crops, soybean, cotton, fruit orchards and vineyards. Both herbicides are lipophilic, with a LogK _ow of 5.18 and 4.38 for pendimethalin and oxyfluorfen, respectively.^{[ 1 , 2 ]} This chemical property is associated with a strong organic soil adsorption that results in limited soil mobility.^{[ 3 , 4 , 5 , 6 , 7 ]}

Vineyard production systems have been searching for new herbicides with low environmental risk to replace current compounds that have a higher risk of groundwater contamination.^{[ 8 , 9 , 10 , 11 , 12 , 13 ]} Therefore, pendimethalin and oxyfluorfen could be viable alternatives. However, they are considered to have moderate to high soil persistence,^{[ 1 , 2 , 7 , 14 ]} which increases their potential risk to contaminate the environment.

Considering, that herbicide leaching and dissipation depend on climatic conditions and soil characteristics^{[ 5 , 7 , 15 , 16 , 17 , 18 ]} the main objective of this study was to determine and compare soil persistence and leaching of pendimethalin and oxyfluorfen under two pluviometric conditions in a vineyard under the climatic conditions of The Casablanca Valley, Chile.

Materials and methods

Study conditions and herbicide treatments

From August 2003 to August 2007, pendimethalin at 1.33 kg ha⁻¹ and oxyfluorfen at 0.75 kg ha⁻¹ were applied to bare soil on August 20th of each year using a backpack sprayer calibrated to deliver 200 L ha⁻¹ at 0.35 mPa. The experimental area was 1.4 ha and was located in an 8 year old commercial vineyard in The Casablanca Valley, Chile (Latitude 33°16′ S and Longitude 71°23′ W). The sub-surface water table oscillated from 0.6 m to 3.0 m deep during the growing season (between August and March). The average soil temperature and annual rainfall are presented in Figure 1.

Fig. 1 Climatic parameters. a) Rainfall distribution during the study period. b) Average, maximum and minimum monthly soil temperature from 0 to 15 cm depth.

Previously to herbicide applications, 20 soil samples from the experimental area were obtained using a grid sampling pattern, and were geo-referenced with a differential global positional system (DGPS). Soil samples were taken at four depths (0–15; 15–30; 30–60 and 60–90 cm), and characterized according to their soil physicochemical properties (Table 1). Four homogeneous zones were identified using cluster analysis based on soil properties. Vegetation was removed from two 3-m² sections in the inter-row in the each zone to avoid debris interference. One section received the natural rainfall while the other section was augmented with irrigation equivalent with 30 mm every 15 days during the 90 days after application (DAA).

Table 1 Soil physicochemical properties in the experimental area at 0–15 cm soil depth and herbicide soil adsorption coefficient. ^†

Density (g cm^−3)	1.544 ± 0.137
pH	6.100 ± 0.198
CEC (meq 100 g^−1)	14.884 ± 2.801
OC (%)	1.284 ± 0.266
Sand (%)	44.014 ± 6.478
Silt (%)	35.196 ± 4.476
Clay (%)	20.790 ± 2.370
Field Capacity (%)	0.264 ± 0.006
Pendimethalin soil adsorption (K d ) (mL g^−1) ^‡	39.922 ± 5.921
Oxyfluorfen soil adsorption (K d ) (mL g^−1) ^‡	13.731 ± 3.953
CEC: Cationic exchange.
OC: Organic carbon.

Herbicide soil sampling

Soil samples were collected using a steel soil sampler (soil core 0.06 m-diameter) from: 0–5; 5–10; 10–15; and 15–30 cm at 0, 10, 20, 40, 90 and 340 DAA in 2003, avoiding to contaminate deeper soil layers with surface soil. Based on the limited leaching of these two herbicides during the 2003, soil samples were collected from: 0–2.5, 2.5–5, 5–10 and 10–15 cm depth at 0, 10, 20, 40, 90 and 340 DAA during 2004 to 2007 seasons. Each soil sample was packed in a plastic bag, kept in a cooler at 4°C and carried to the laboratory to be frozen at −18°C until analysis.

Herbicide soil quantification

Herbicides were extracted from 20 g of oven-dried soil samples (8 hours at 30°C) and shaken at 300 rpm for 90 min with methanol (40 mL). The extracts were filtered and concentrated to dryness in a rotary evaporator and re-suspended in methanol (1.5 mL), and transferred to a glass vial and analyzed using high-pressure liquid chromatography with a diode-array detector (HPLC-DAD). The HPLC unit was equipped with a LiChrocart 125-4 Lichrospher® 100 RP-18 5 μ m column (125 mm-length). The liquid phase used was acetonitrile and ammonium acetate 10 mmol. The acetonitrile gradient was: 0 at 10 min 30%; 10–11 min 70%; and 11–13 min 30%. The column temperature was 30°C and flow rate 1 mL min⁻¹. The injection volume was 30 μ L. The detector (Hitachi model Elite LaChrom L-2450) was set at 240 nm and 220 nm for pendimethalin and oxyfluorfen, respectively (Fig. 2).

Fig. 2 High performance liquid chromatography (HPLC) chromatograms for Pendimethalin (a) and Oxyfluorfen (b) at 0, 90 and 340 days after herbicide application (DAA).

Herbicide recovery from spiked soil (concentrations: 0; 0.5; 1; 5 and 10 μ g g⁻¹) were 104 ± 9 and 91 ± 11%, with a detection limit of 0.006 ± 0.002 and 0.004 ± 0.002 μ g g⁻¹ for pendimethalin and oxyfluorfen, respectively. For all soil sample herbicides quantification included blanks, spiked and no applied soil samples.

Adsorption coefficient determination

Six milliliters of aqueous 0.01 M CaCl₂ solutions of pendimethalin and oxyfluorfen, in a concentration of 0.3 and 0.1 μ g mL⁻¹, respectively, were added to 3 g air-dried soil. These soil suspensions were shaken end-over-end for 24 h at 200 rpm at room temperature, and then centrifuged for 15 min at 4000 RFC. One milliliter of each supernatant was filtered through a 0.45 μ m fiberglass filter and directly quantified using HPLC-DAD as mentioned. The adsorbed herbicide was calculated as the difference between the amount in the initial solution and that remaining in solution after centrifugation. The adsorption coefficient K _d (mL g⁻¹) was determined using Equation 1 where Cs (μ g g⁻¹) is the sorbed herbicide and Ce (μ g mL⁻¹) is the herbicide concentration in the solution after 24 h of shaking.

Statistical data analysis

Herbicide field dissipation was fitted to a first order equation (Equation 2), using nonlinear regression analysis PROC NLIN (SAS®), where C (mg kg⁻¹) is the soil herbicide concentration at time t (days); C _○ is the initial soil concentration (mg kg⁻¹); k is the dissipation rate (days⁻¹). The goodness-of-fit of the model was calculated according to Schabenberger.^{[ 19 ]} The DT ₅₀ values were estimated from the Equation 3. The use of superior order equation did not show any improvement in the correlation values (data not show).

The amount of herbicides found in soil samples and the K _d values were analyzed using PROC UNIVARIATE (SAS®). Correlations and regressions between DT ₅₀, soil residues and rainfall regimens were estimated using PROC CORR and PROC REG with stepwise selection model (SAS®).

Results and discussion

Pendimethalin and oxyfluorfen had been considered as a low potential of leaching herbicides,^{[ 4 , 5 , 6 , 7 , 20 , 21 ]} due to their high K _ow coefficients, which result in a strong soil adsorption, which agrees with our results in this study, where pendimethalin and oxyfluorfen showed relatively high K _d values in a soil with low organic matter content (1.28% OC)(Table 1). However, Kulshrestha et al.^{[ 22 ]} detected pendimethalin at 90 cm deep after five years of applications. They concluded that pendimethalin leached because of high soil temperature (from 25 to 35°C) and wetting and drying cycles that would have influenced processes of adsorption/desorption. In contrast, in this study pendimethalin did not leach below 5 cm and oxyfluorfen was detected at 10 cm deep only in 2006. Thus, more than 74% of both herbicides were found at the top 2.5 cm at 90 and 340 days after their applications. Under the same soil and climatic conditions, simazine and flumioxazin reach 90 and 30 cm of soil depth, respectively.^{[ 23 , 24 ]} Some reports had showed a positive effect of irrigation or rainfall on pendimethalin and oxyfluorfen soil movement.^{[ 5 , 25 , 26 , 27 ]} Results from Anke-Gowda et al.^{[ 26 ]} showed a reduction of 10% in pendimethalin soil mobility (cm d⁻¹) when the soil moisture was at 50% of field capacity. In the present study, rainfall conditions did not affect soil moisture content (Fig. 3), which could explain the lack of effect of rainfall conditions on herbicides soil movement. Calderon et al.^{[ 28 ]} did not detect trifluralin below 12 cm soil depth after applying 50 mm of simulated rain in a soil with similar organic carbon content that the one in this field study, and found more than 80% of trifluralin at the superficial soil layer (0–4 cm) at 83 DDA.

Fig. 3 Variation in soil moisture (% of oven dry soil) between natural rainfall and irrigated conditions from 0 to 15 cm depth. (a) 2003, (b) 2004, (c) 2005 and (d) 2006. Error bars correspond to standard deviations.

Pendimethalin half life values ranged from 10.5 to 31.5 days (Table 2; Fig. 4), being closer to the ones reported for golf fairway area conditions, which varied from 12 to 23 days,^{[ 29 , 30 ]} but smaller than that reported in the literature under agricultural conditions, where DT ₅₀ values ranged from 23 to 62 days.^{[ 18 , 31 , 32 , 33 , 34 , 35 ]} Similarly, Alister et al.^{[ 23 ]} found that simazine dissipation, under the climatic and soil conditions of the Casablanca Valley (Chile) was faster than the expected for this herbicide.

Table 2 Soil dissipation half-life (DT ₅₀) and rate (k) during four seasons for Pendimethalin and Oxyfluorfen under natural rainfall and irrigated conditions.

		Pendimethalin			Oxyfluorfen
Season	Condition ^†	DT50 days	k days ^−1	R^2	DT50 days	k days ^−1	R^2
2003–2004	Natural rainfall (39 mm)	10.453 (± 0.842)	0.069 (± 0.009)	0.917	40.069 (± 3.729)	0.018 (± 0.003)	0.867
	Irrigated (39 + 180 mm)	15.868 (± 1.082)	0.045 (± 0.005)	0.936	52.312 (± 6.279)	0.014 (± 0.003)	0.769
2004–2005	Natural rainfall (78 mm)	13.664 (± 01.325)	0.053 (± 0.008)	0.882	34.311 (± 3.215)	0.021 (± 0.003)	0.869
	Irrigated (78 + 180 mm)	17.366 (± 1.657)	0.042 (± 0.006)	0.892	36.417 (± 2.780)	0.019 (± 0.003)	0.913
2005–2006	Natural rainfall (156 mm)	25.645 (± 2.521)	0.029 (± 0.005)	0.894	36.858 (± 2.822)	0.019 (± 0.003)	0.924
	Irrigated (156 + 180 mm)	31.511 (± 3.162)	0.023 (± 0.004)	0.892	42.690 (± 3.436)	0.017 (± 0.002)	0.917
2006–2007	Natural rainfall (87 mm)	24.557 (± 3.074)	0.031 (± 0.006)	0.808	38.614 (± 2.493)	0.018 (± 0.002)	0.928
	Irrigated (87 + 180 mm)	21.635 (± 2.126)	0.034 (± 0.006)	0.880	46.193 (± 4.053)	0.016 (± 0.002)	0.864
DT50: Soil dissipation half-life (days).
k: Dissipation rate.
R2: Coefficient of determination.

Fig. 4 Pendimethalin (a) and Oxyfluorfen (b) dissipation curves under natural rainfall and irrigated conditions. Solid lines correspond to a first order model Eq. (2). Errors bars correspond to the standard error of the means.

In general, irrigation produced a slight increase in DT ₅₀ values, except in the 2006–2007 season (Table 2). After analyzing the effect of rainfall on DT ₅₀ (time between the herbicide application and first rainfall event, rain amount at the first event, number of rainfall events and total rainfall during 90 DAA) it was determined that the interval between herbicide application and first rainfall event and water amount at the first rainfall event showed a significant correlations with pendimethalin DT ₅₀ values (R² = 0.29; p = 0.006 and R² = 0.28; p = 0.04, respectively). In the case of oxyfluorfen only the amount of rainfall at the first event showed a significant correlation with the DT ₅₀ (R² = 0.2; p = 0.02). On the other hand, Ismail and Kalithasan,^{[ 36 ]} and Tsiropoulus and Lolas,^{[ 35 ]} reported an accelerated pendimetahlin and oxyfluorfen dissipation because of an increase in the amount of simulated rainfall after their application. A similar behavior was observed for trifluralin.^{[ 28 ]} The soil water content in the experimental plots of this study could explain these contradictory results. Even though the extra water added to the soil (irrigated) corresponded to a significant amount, it was not sufficient to modify the soil moisture. Thus, in both conditions the soil was always close to field capacity during the first 40 DAA in the four study seasons (Table 1; Fig. 3).

The slight increase on DT ₅₀ values when rainfall closely followed herbicide application could be explained because of an increase in soil incorporation that could avoid herbicides photo-decomposition and volatilization.^{[ 2 , 5 , 37 , 36 , 37 , 38 , 39 , 40 , 41 ]} However, dissipation models with more than one order were fitted to the field data and did not show any improvement in the R ², indicating that photo-decomposition and volatilization were not statistically significant under the experimental study conditions.

Pendimethalin and oxyfluorfen soil residues at 340 DAA were similar, independently of the season and rainfall regimens (Table 3). In a bioassay study Ismail and Kalithasan^{[ 36 ]} found a decrease in pendimethalin half-life when soil temperature was higher than 25°C. In the present study, soil temperature after 40 DAA increased from 10.9°C at the end of September to over 22°C in November in each season (Fig. 1b). Favorable soil temperature and moisture near field capacity (Fig. 3), could result in very good conditions for pendimethalin degradation.^{[ 42 , 43 , 44 , 45 ]} No soil enrichment was detected in contrast to Kulshrestha et al.^{[ 22 ]}

Table 3 Pendimethalin and Oxyfluorfen soil residues (0–10 soil depth) at 90 and 340 days after herbicide applications. Values are expressed as percentage of the initial amount applied in each season (four replication averages ± standard error of the mean).

		Pendimethalin		Oxyfluorfen
Season	Condition ^†	90 DAA	340 DAA	90 DAA	340 DAA
2003–2004	Natural rainfall (39 mm)	10.342 (± 2.404)	3.077 (± 0.774)	24.635 (± 3.873)	4.589 (± 0.420)
	Irrigated (39 + 180 mm)	8.800 (± 2.717)	2.412 (± 0.181)	34.822 (± 5.028)	3.706 (± 0.811)
2004–2005	Natural rainfall (78 mm)	7.506 (± 1.314)	4.273 (± 1.797)	21.797 (± 2.907)	9.679 (± 2.146)
	Irrigated (78 + 180 mm)	2.511 (± 0.858)	4.119 (± 0.802)	16.755 (± 1.845)	3.376 (± 1.326)
2005–2006	Natural rainfall (156 mm)	13.838 (± 2.597)	6.114 (± 0.171)	16.608 (± 1.845)	17.909 (± 2.830)
	Irrigated (156 + 180 mm)	9.812 (± 3.253)	4.468 (± 0.939)	18.886 (± 3.596)	16.672 (± 3.052)
2006–2007	Natural rainfall (87 mm)	13.685 (± 2.640)	8.642 (± 1.782)	23.104 (± 3.279)	8.140 (± 0.603)
	Irrigated (87 + 180 mm)	9.986 (± 3.627)	8.123 (± 2.134)	29.005 (± 3.283)	15.834 (± 2.618)
DAA: Days after Herbicidc application.

Oxyfluorfen DT ₅₀ values were greater (34.3 to 52.3 days) compared to pendimethalin, but lower than those reported in the literature (30 to 160 days).^{[ 1 , 2 , 5 , 7 ]} It seems like soil temperature (below 15°C; Fig. 1b) and moisture content (Fig. 3) during the first 40 DAA, which were not limiting for pendimethalin degradation, could have had an effect on oxyfluorfen microbiological degradation.^{[ 7 ]} However, after 40 DAA soil temperature increased over 22°C (Fig. 1b) and soil moisture remained approximately at 50% of field capacity (Fig. 3). These factors were not limiting for oxyfluorfen further degradation, which could explain its low concentration found at 340 DAA. The exception was the 2005–2006 seasons in which soil residues were similar at 90 and 340 DAA (Table 3). This difference could be related with a soil average temperature decrease in 2°C during December as compared to the other seasons (Fig. 1b).

Conclusions

According to the results obtained in this four-year field study, we conclude that the time interval between oxyfluorfen and pendimethalin applications and the first rain event, and the amount of rainfall of that event would determine herbicide soil dynamics.

The smaller DT ₅₀ values obtained for both herbicides, compared to the ones published in the literature, are related to soil temperature and moisture indicates that pendimethalin and oxyfluorfen have short soil persistence. In addition, their very limited soil leaching because of their high organic matter adsorption, make them very low risk herbicides to contaminate groundwater. Nevertheless, the high herbicide concentration in the first 2.5 cm soil layer is an important factor to consider in orchards and vineyards planted in sloppy soils, exposed to run-off because of rainfall water and wind soil erosion.

Acknowledgments

The authors wish to thank the FONDECYT (Chilean Fund for Science and Technology) for funding Project Number 1030990, and PhD Alejandro Pérez-Jones for their kind revision of the manuscript.

Notes

^†Values are means of 20 samples ± SD.

^‡Values are means of 4 repetitions ± SD.

^†Conditions correspond at the natural rainfall and irrigated during the 90 days after herbicide application.

^†Conditions correspond at the natural rainfall and irrigated during the 90 days after herbicide application.