- Email: [email protected]
Effect of grass hedges on runoff loss of soil surface-applied herbicide under simulated rainfall in Northern China
Agriculture, Ecosystems and Environment 253 (2018) 1–10
Contents lists available at ScienceDirect
Agriculture, Ecosystems and Environment journal homepage: www.elsevier.com/locate/agee
Research Paper
Effect of grass hedges on runoff loss of soil surface-applied herbicide under simulated rainfall in Northern China
MARK
⁎
Qinghai Wanga, , Cui Lia, Zhuo Panga, Haifeng Wena, Ruilun Zhenga, Jie Chena, Xueju Maa, ⁎⁎ Xiaoe Queb, a b
Beijing Research & Development Center for Grass and Environment, Beijing Academy of Agriculture and Forestry Sciences, Beijing, China Institute of Desertification Studies, Chinese Academy of Forestry, Beijing, China
A R T I C L E I N F O
A B S T R A C T
Keywords: Grass hedge Atrazine loss Surface runoff Sloping cropland
Pesticide loss triggered by runoff is one of the most important sources of non-point pollution. The effects of grass hedges (Melilotus albus and Pennisetum alopecuroides) on atrazine runoff under different rain intensities and slope gradients were evaluated. The plot-scale experiments were carried out on a maize (Zea mays) field on slopes with 15% and 20% gradients using simulated rainfall (rain intensity of 30 and 45 mm h−1). Atrazine residues were investigated in runoff water and soil taken from three depths (0–5, 5–10, and 10–15 cm) in the middle and base of the slope after runoff events. Total atrazine loss in runoff water ranged from 2.3% to 4.9% of that applied from plots without grass hedges. Grass hedges decreased atrazine loss by 37%-76% and surface runoff by 27%-72%, and Pennisetum showed better efficacy than Melilotus, especially under higher rain intensity. Atrazine loss showed a significant positive linear correlation with surface runoff volume. Grass hedges had a more significant effect on atrazine loss than rain intensity and slope gradient. But they functioned less effectively if used under intense rain or/and steep slope conditions. Atrazine residues remained in the surface 15 cm soil were higher for the plots with grass hedges than the control plots. These results suggest that grass hedges not only significantly reduced atrazine loss by reducing the surface runoff, but also reduced the amount of atrazine leaching to deeper soil layer. P. alopecuroides was a suitable grass-hedge species for controlling atrazine losses in northern China and similar regions. Other management practices or control measures should be integrated with grass hedges in strongly sloping cropland in high-rainfall areas to maintain pesticide losses at an acceptable level.
1. Introduction World pesticide usage at the producer level totaled nearly 2.7 × 109 kg annually in both 2011 and 2012 (Atwood and PaisleyJones, 2017). Given the magnitude of global pesticide use, widespread pesticide contamination of surface water and groundwater is expected (Rippy et al., 2017). During 1992-2011, the proportion of assessed streams with one or more pesticides that exceeded an aquatic life benchmark was 61% for agricultural streams in USA (Stone et al., 2014). In developing countries, contamination of ground and surface water with pesticides is more prevalent (Schwarzenbach et al., 2010). Because it is biologically active, water pollution caused by pesticides used in agriculture is a major concern worldwide (Juergens et al., 2015; Stehle and Schulz, 2015). It poses a significant health risk to humans and other animals (Hussain et al., 2009). Pesticides deposition in soil is one of the major pollution sources. These pesticides can enter surrounding surface waters mainly through diffuse sources, surface runoff ⁎
and erosion is one of the most important pathways of pesticide input in aquatic ecosystems among diffuse pollution (Vymazal and Březinová, 2015). Surface runoff water acted as the major mobile carrier of pesticide mobilization from agricultural land to ecosystem, because sediment amount is usually small compared to the runoff volume from a field (Reichenberger et al., 2007). Pesticides losses linked to runoff during the growing season have been reported to account for greater than 60% of the total loss (Potter et al., 2014). Herbicide concentrations in the runoff obtained during the first rainfall events after treatment, had a high probability of exceeding the ecotoxicological endpoint for algae (Vianello et al., 2005). Therefore, adequate control of runoff, and subsequent pesticide loss is very important. Among the various control measures, grass hedges, a special type of vegetative filter strips defined as narrow strips of stiff-stemmed grass used to control runoff and erosion (Vieira and Dabney, 2012), are commonly used in preventing soil and water loss on sloping croplands all over the world due to their low cost, eco-friendliness and efficiency (Xiao et al., 2011).
Corresponding author at: Beijing Research and Development Center for Grass and Environment, No. 9 Shuguang Garden Road of Haidian District, Beijing, 100097, China. Corresponding author. E-mail addresses: [email protected] (Q. Wang), [email protected] (X. Que).
⁎⁎
http://dx.doi.org/10.1016/j.agee.2017.10.024 Received 6 July 2017; Received in revised form 14 October 2017; Accepted 27 October 2017 0167-8809/ © 2017 Elsevier B.V. All rights reserved.
Agriculture, Ecosystems and Environment 253 (2018) 1–10
Q. Wang et al.
Fig. 1. Schematic diagram of experimental plot, T1 = Melilotus hedge, T2 = Pennisetum hedge, and CK = control.
This study also gains data for future use in predicting potential pesticide loss due to runoff and evaluating the risk to water quality.
Atrazine is a photosynthesis inhibiting triazine herbicide that has been used for approximately 50 years and has become one of the most heavily applied agricultural herbicides in the world (Jablonowski et al., 2011). Furthermore, it exhibits relatively high water solubility (32 mg L−1) and moderate environmental persistence (half-life in soil: 146 d) with adsorption coefficient (logKoc) of 1.97 (Bouldin et al., 2006), and is frequently detected in surface water and shallow groundwater (Bradley et al., 2017). It has become the subject of continuous concern due to its potential endocrine and carcinogenic activity (Jablonowski et al., 2011; Richter et al., 2016). Atrazine has been widely accepted as an important tool in controlling weeds in summer maize, a main crop in the predominant cropping system in Northern China where there is severe soil and water loss (Mao and Ren, 2004; Xiao et al., 2011). Atrazine application period is just in the rainy season, and frequent strong thunderstorms with such features as high intensity and short duration typically occur in this region, thus atrazine has huge potential of losses in surface runoff with water and sediment in this region. In terms of runoff-producing rainfall events after herbicide application, the first few storms are often responsible for the largest herbicide losses (Baker and Mickelson, 1994). These extreme events dominate herbicide transport and are critical for devising management practices or control measures, and monitoring total pesticide loss (Shipitalo and Owens, 2006). It is, therefore, of great significance and broadness to evaluate the effectiveness of grass hedges in reducing herbicide loss, as well as to quantify atrazine loss from sloping agricultural fields in runoff events, especially rainstorms shortly after its application. Melilotus (Melilotus albus) and Pennisetum (Pennisetum alopecuroides) are promising grass hedge candidates due to their native and non-invasive characteristics, tolerance to the climate extremes, as well as sufficient stem strength for blocking water flow, and provide an efficient way to prevent soil and water loss on sloping croplands (Wu et al., 2010; Xiao et al., 2011). However, little is known about their effectiveness on herbicide loss and their influences on herbicide distribution in soil at different rain intensities and slope gradients. Achieving such information may help to seek for suitable grass species to develop grass hedges for the temperate regions, and to explore the potential of grass hedges alone for reducing pesticide loss to an acceptable level under extreme conditions (such as heavy rainfall, steep slope). Consequently, the main objectives of this research were to (1) estimated runoff loss of atrazine for storms of high intensity and short duration occurring shortly (4 h) following application to soil surface, (2) determine the efficacy of the two types of hedges (Melilotus and Pennisetum) in decreasing atrazine loss and their influences on atrazine distribution in soil, and (3) analyze the relative importance of grass hedges, rain intensity and slope gradient for atrazine losses. This research extends previous work of Wu et al. (2010) and Xiao et al. (2011) by focusing on herbicide loss rather than water and soil loss, and quantifying the correlations between herbicide loss and surface runoff.
2. Materials and methods 2.1. Study site The study was carried out at the National Experimental Station for Precision Agriculture (116°26′ E, 40°10′ N) in Beijing, China. The experimental station is located in North China and is characterized by a continental, semi-humid, and monsoon climate. The mean precipitation is 640 mm y−1, 80% of which occurs between June and August. The mean annual temperature is 11.5 °C. Rotation of winter wheat and summer maize dominates agricultural activities in the region. The soil is brown fluvo aquic soil in type and loamy clay in texture (41% sand, 24% silt and 35% clay) with a pH of 7.6. The bulk density is 1.37 g cm−3, The soil organic matter content and the total nitrogen and phosphorus are 14.0, 2.46 and 0.63 g kg−1. 2.2. Experimental design and herbicide application The experiment was based on three independent variables: grass hedges (Melilotus hedges, Pennisetum hedges, and without grass hedges as control), slope gradients (15% and 20%), and rainfall intensities (30 and 45 mm h−1). Eighteen 1.5 m by 12 m plots (3 hedge levels × 2 slope gradients × 3 replications) were established. Plots were separated by 2-mm-wide sheet metal frames driven approximately 30 cm into the soil in two sides of the plots (but not the top and bottom). The grass hedges were located at the middle and bottom positions of selected test plots. They were established in June 2012 in two parallel rows perpendicular to the slope direction (Fig. 1). The grass hedges were thick enough (densities of Melilotus and Pennisetum was 684 and 125 stems m−2) when the experiments were conducted. Summer maize was sown manually in all plots of every treatment after the harvest of winter wheat without tillage in both years (27 June 2013 and 20 June 2014, respectively). Afterwards, atrazine (38% Suspension concentrates) was applied on the soil surface using a hand sprayer at recommended field dosage [1282.5 g ha−1 (ai)]. All sprays were applied on the day that the summer maize was sown. 2.3. Rainfall Simulation and Runoff Collection Rainfall simulation has been used extensively as a cost-effective method to evaluate the effects of managements on infiltration, surface runoff, soil erosion, and contaminant transport in overland flow (Humphry et al., 2002; Sharpley and Kleinman, 2003). It exhibits advantages in expediting data collection, as well as creating controlled and reproducible artificial rainfall for comparison among locations (Humphry et al., 2002). In this study, precipitation was simulated using 2
Agriculture, Ecosystems and Environment 253 (2018) 1–10
Q. Wang et al.
Fig. 2. Surface runoff and effectiveness of grass hedges in reducing runoff under different slope gradients and rainfall intensities: (A) runoff under rainfall intensity of 30 mm h−1; (B) runoff under rainfall intensity of 45 mm h−1; (C) effectiveness of grass hedges in reducing runoff under rainfall intensity of 30 mm h−1; and (D) effectiveness of grass hedges in reducing runoff under rainfall intensity of 45 mm h−1. *and ** indicate significance of the t-test between the slope gradient of 15% and 20% within control, Melilotus or Pennisetum hedges at P < 0.05, and P < 0.01, respectively. Different letters indicate significant differences between the control, Melilotus or Pennisetum hedges at P < 0.05 under the same slope gradient according to Duncan’s test. Error bars indicate SE of the mean (n = 3).
Fig. 3. The atrazine concentration in runoff water under the rain intensities of 30 mm h−1 (A) and 45 mm h−1 (B); there were no significant treatment effects. Error bars indicate SE of the mean (n = 3).
and 0.22 MPa. The rainfall coefficient of uniformity measured according to the Christiansen method averaged 86% (Xiao et al., 2010). Generally speaking, moderate to extreme rains contribute 68.2% of the precipitation amount in Northern China (Gong et al., 2004). Compared
a rainfall simulator which can sprinkle water at intensities of 10 to 100 mm h−1. Three types of nozzles were used in the simulator (PROS17, PROS-15 and microsprinkler with flow rate of 102 L h−1, Hunter Company in USA), and the working pressures ranged between 0.16 MPa 3
Agriculture, Ecosystems and Environment 253 (2018) 1–10
Q. Wang et al.
Fig. 4. Atrazine loss and effectiveness of grass hedges in reducing atrazine loss under different slope gradients and rainfall intensities: (A) atrazine loss under rainfall intensity of 30 mm h−1; (B) atrazine loss under rainfall intensity of 45 mm h−1; (C) proportion of the amount atrazine loss to application under rainfall intensity of 30 mm h−1; (D) proportion of the amount atrazine loss to application under rainfall intensity of 45 mm h−1. *and ** indicate significance of the t-statistic between the slope gradient of 15% and 20% for the control, or Melilotus or Pennisetum hedges at the 0.05, and 0.01 levels, respectively. Different letters indicate significant differences between the control, Melilotus or Pennisetum hedges at P < 0.05 under the same slope gradient according to Duncan’s test. Error bars indicate SE of the mean (n = 3).
through a paper filter. Filtered runoff water was stored at −20 °C for later analysis for atrazine residue.
to rainfall events of medium intensity and medium duration, rainfalls with high intensity and short duration cause the greater proportion of runoff and soil loss (Wei et al., 2007). In study area during natural rainfall events, annual maximum 1 h precipitation intensity fell in the range 30–50 mm h−1 in last decades (Zhong et al., 2013). For those reasons, the intensities employed for this study were 30 and 45 mm h−1, representing a rang for strong thunderstorms of this region. To avoid the effect of atrazine retained in the soil on the later rainfall intensity test, these two intensities were designed in 2013 and 2014, respectively. In the second studied year (2014), in order to check for the presence of atrazine residues from the last year, soil was sampled on the day before herbicide application in the upper 5 cm layer (the most critical segment for runoff and strongly influenced the concentrations of chemicals in runoff water (Milan et al., 2013)). The determined data showed that the soil samples contained no detectable atrazine (< 0.2 μg kg−1) from the previous year prior to the application of herbicide. Heavy rain can transport soil surface-applied atrazine to the deep soil beyond the weed germination zone and such reduce atrazine performance (Walter and Lyle, 1962). The effective atrazine performance required an interval of at least 4 hours between herbicide application and heavy rain in agriculture practice. The simulated rainfall was applied 4 h after herbicide application for a duration of 1 h. Before rainfall simulation, soil water contents in 0–15 cm layer on the upper, middle, and bottom slope position of plots in 2013 were 20.3%, 21.0% and 22.6%, respectively; and in 2014 were 20.1%, 20.5% and 22.5%, respectively. Soil water content at same slope position did not differ significantly between 2013 and 2014. The surface runoff was collected in containers at the lower edge of each plot, and volumes were measured. A 1 L sample of total runoff with sediment was obtained from each container after thorough agitation, and was filtered immediately
2.4. Soil sampling Soil in each plot was sampled after the end of the runoff. Two sampling positions, representing middle and bottom slope positions, were located within the grass hedges. The corresponding sampling positions were selected in the control plots. Composite samples of five cores were taken randomly at depths of 0-5, 5-10 and 10–15 cm at each slope position using a soil sampling drill (2.0 cm diameter, 10.0 cm depth) in each plot. After collection, the composite soil samples were freeze-dried and kept at −20 °C until analysis. They were stored for no more than 7 d. 2.5. Atrazine analysis 2.5.1. Atrazine extraction from water sample Samples of 2 mL filtered water were extracted using 0.2 g C18 cartridge under vortex oscillating for 2 min, then separated at 4 °C in a centrifuge at 10000 rpm for 5 min. 1 mL supernatant from the extracts was evaporated with N2, and 1 mL acetonitrile/water mixture (1:1, v/v) was added immediately. The solution passed through 0.22 μm microporous filter before the analysis using ultra-high performance liquid chromatography coupled with tandem mass spectrometry (UPLC-MS/ MS) (Waters, USA), according to Wille et al. (2011). 2.5.2. Atrazine extraction from soil sample Soil samples were extracted 2 times. A 5 g sample was put into a 50mL polystyrene centrifuge tube, subsequently 1 g anhydrous magnesium sulfate and 10 ml acetonitrile were added. After being shaken for 4
Agriculture, Ecosystems and Environment 253 (2018) 1–10
Q. Wang et al.
3. Results
Table 1 Effects of experimental factors on atrazine residue in soil. Factors
r g sg sp sd r*g r*sg r*sp r*sd g*sg g*sp g*sd sg*sp sg*sd sp*sd r*g*sd r*sg*sp r*sg*sd r*sp*sd g*sg*sp sg*sp*sd g*sg*sd g*sp*sd r*g*sg*sp r*sg*sp*sd r*g*sg*sd r*g*sp*sd g*sg*sp*sd r*g*sg*sp*sd
Degree of freedom
1 2 1 1 2 2 1 1 2 2 2 4 1 2 2 4 1 2 2 2 2 4 4 2 2 4 4 4 4
3.1. Surface runoff
Atrazine residue in soil Sum of Squares
F
P
337320.590 1138303.909 540444.091 137029.571 3182346.965 133592.200 5479.289 19126.766 57911.676 11036.823 12257.606 163651.400 23781.671 115852.067 33948.281 74414.867 623.832 5509.649 28520.222 42709.709 9688.345 30007.996 39133.264 4980.972 4424.210 7752.228 40401.601 30957.858 7236.321
113.08 190.80 181.17 45.94 533.41 22.39 1.84 6.41 9.71 1.85 2.05 13.72 7.97 19.42 5.69 6.24 0.21 0.92 4.78 7.16 1.62 2.51 3.28 0.83 0.74 0.65 3.39 2.59 0.61
< 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 0.1774 0.0124 0.0001 0.1610 0.1319 < 0.0001 0.0054 < 0.0001 0.0042 0.0001 0.6481 0.3995 0.0098 0.0011 0.2007 0.0441 0.0132 0.4360 0.4782 0.6280 0.0111 0.0389 0.6586
Obvious changes in surface runoff were observed under different grass hedge, rain intensity and slope gradient treatments (Fig. 2). The total amount of surface runoff under the rain intensity of 30 mm h−1 was markedly smaller than that under 45 mm h−1 (P < 0.001) (Fig. 2A, B). Grass hedges reduced surface runoff in the range of 27% to 72% compared to the control (Fig. 2C, D). The runoff in all plots increased with slope gradients, and this difference was significant for the control and Melilotus at higher rain intensity (45 mm h−1) (P = 0.005, 0.002, respectively) (Fig. 2B), but was not significant under at 30 mm h−1 rain intensity (P = 0.090) (Fig. 2A). Regardless of the rain intensity and slope angle, plots with grass hedges released pronouncedly less runoff compared to the control (P < 0.001). Moreover, the highest surface runoff occurred in the control plots, and the lower in the plots with Pennisetum hedge. The difference in surface runoff was significant between the three treatments for the 45 mm h−1 rainfall intensity (P < 0.001). However, for the 30 mm h−1 rainfall intensity, there was no significant difference between Pennisetum and Melilotus treatments (P = 0.197). 3.2. Atrazine in surface runoff Under the conditions of same rain intensity, the grass hedge and slope gradient did not lead to significant differences of atrazine concentration in runoff (P = 0.7307, 0.5260, respectively), but atrazine concentrations in the runoff were significantly higher at the intensity of 30 mm h−1 than at 45 mm h−1 (P < 0.001) (Fig. 3). The control plots had the highest amount of atrazine loss, followed by plots with Melilotus and Pennisetum hedges (Fig. 4A, B). For the control, Melilotus, and Pennisetum hedges, atrazine losses at 45 mm h−1 intensity with 20% slope gradient were highest at 63, 40, and 26 mg ha−1, respectively; atrazine losses at 30 mm h−1 intensity with 15% slope gradient were lowest at 30, 13, and 7 mg ha−1, respectively. Other combinations of slope and rainfall were intermediate. Atrazine loss rate was 0.6% to 4.9% depending on the grass hedge, rain intensity and slope gradient. Of the applied atrazine 0.6%-3.1% was lost in runoff water from plots protected by grass hedges, compared with 2.3%-4.9% when grass hedges were not present (Fig. 4C, D). Atrazine loss increased with slope increase for both planted and unplanted grass hedges. At the rain intensity of 45 mm h−1, the amount and rate of atrazine loss from control, Melilotus and Pennisetum plots with the slope gradient of 20% were all significantly higher than that with the slope gradient of 15% (P = 0.025, 0.005 and 0.041, respectively). For the same slope, the amount and rate of atrazine loss from control, Melilotus and Pennisetum plots at the rain intensity of 45 mm h−1 were significantly higher than that at the rain intensity of 30 mm h−1 (for 15% slope, P = 0.020, 0.039 and 0.019, respectively; for 20% slope, P = 0.028, 0.005 and 0.040, respectively). The reduction of atrazine loss compared to the controls was 56% and 45% in the Melilotus treatment and of 76% and 67% in the Pennisetum treatment, respectively for 15% and 20% slope gradient. There was a significant difference in atrazine loss between plots with and without grass hedges (P < 0.001), and Pennisetum was significantly more effective in reducing atrazine losses than Melilotus (P < 0.001), except at 30 mm h−1 rain intensity with 20% slope. At the rain intensity of 45 mm h−1, Melilotus and Pennisetum hedges exhibited a significant decrease in efficiencies at 20% slope compared with the case at 15% slope (P = 0.043 and 0.044, respectively).
Note: r = rainfall intensity, g = grass hedge, sg = slope gradient, sp = slope position, sd = soil depth.
30 min on a horizontal shaker, the mixture was separated in a centrifuge at 10000 rpm for 5 min at 4 °C. Repeated extraction was carried out twice, and the supernatants were combined. Then the soil final extract was evaporated to dryness at 40 °C, and the residue was made up to a volume of 5 mL with acetonitrile (Mei et al., 2011). The following extraction, purification and analysis procedures were the same as for the water sample. Soil recovery for atrazine was 95.2%, and the detection limits were 0.05 μg L−1 for atrazine in water and 0.2 μg kg−1 in soil. There was a small quantity of soil in runoff obtained from each plot, but not enough for chemical analysis. Therefore, atrazine loss with sediment could not be included in the calculation.
2.6. Data analysis Atrazine loss rate was expressed as the percentage of applied atrazine lost. Data were analyzed using the Mixed Procedure (SAS Ins., Chicago, USA) according to the method of Piepho et al. (Piepho et al., 2003). Probability of 0.05 or less was considered significant, and the results were expressed as mean ± standard error. SPSS optimal scaling (SPSS Inc., Chicago, USA) was performed to determine the comparative importance of the grass hedge (defined as unordered categorical variable), rain intensity and slope gradient to the surface runoff and atrazine loss. Optimal scaling is a useful method in the analysis of nominal and ordinal variables, and detailed information about this method can be found in Young et al. (1976).
3.3. Atrazine in soil After application atrazine residue levels in the top 15 cm soil were significantly influenced by the rain intensity, grass hedge, slope 5
Agriculture, Ecosystems and Environment 253 (2018) 1–10
Q. Wang et al.
Fig. 5. Atrazine residue in soil of under rainfall intensity of 30 mm h−1 (A) in the control plots under the slope gradient of 15%; (B) in the plots with Melilotus hedge under the slope gradient of 15%; (C) in the plots with Pennisetum hedge under the slope gradient of 15%; (D) in the control plots under the slope gradient of 20%; (E) in the plots with Melilotus hedge under the slope gradient of 20%; and (F) in the plots with Pennisetum hedge under the slope gradient of 20%. *and ** indicate significance of the t-statistic between the slope gradient of 15% and 20% for the control, or Melilotus or Pennisetum hedges at the 0.05, and 0.01 levels, respectively. Different letters indicate significant differences between the control, Melilotus or Pennisetum hedges at P < 0.05 under the same slope gradient according to Duncan’s test. Error bars indicate SE of the mean (n = 3).
degraded under the above rainfall conditions.
gradient, slope position and soil depth (Table 1, Figs. 5 and 6). Soil atrazine residue at 30 mm h−1 rain intensity was significantly higher than that at 45 mm h−1 (P < 0.001). Soil in the plots with grass hedges contained notably higher atrazine residues compared to the plots without grass hedges (P < 0.001), and residue levels in soils under Pennisetum hedge were significantly higher than those under Melilotus hedge (P = 0.021). The soils with 15% gradient or at bottom slope position maintained significantly higher residue level in comparison with the plots with the slope gradient of 20% or at middle slope position (P < 0.001). Atrazine residue levels showed a decreasing trend with soil depth, and differences were significant between all soil layers (P < 0.001). The rain intensity, grass hedge, slope gradient and slope position had significant interaction effects with soil depth on atrazine residue content (P = 0.003, < 0.001, < 0.001, and 0.032, respectively). Under conditions of heavier rain intensity or steeper slope, atrazine residues were significantly lower at 0–5 cm depths (P = 0.043, 0.017, respectively), but there was no significant difference at 5-10 (P = 0.623, 0.106, respectively) and 10–15 cm depths (P = 0.281, 0.247, respectively). Conversely, soil atrazine residues at 5-10 and 10–15 cm depths within grass hedges were higher than those in plots without grass hedges (P = 0.016, 0.019, respectively), while the significant difference was not found in soil layer of 0–5 cm (P = 0.272). As can be seen in Table 2, atrazine in the soil at 0–15 cm depth was higher for the plots with grass hedges than the control plots. Particularly for the plots with Pennisetum hedge, 68%-89% of applied atrazine still remained in the top 15 cm soil after rainfall events at an intensity of 30 mm h−1 for 1 h. The unaccounted-for atrazine in the mass balance analysis mainly included the quantity leaching to deeper soil layers (below 15 cm depth) and the quantity degraded. It was lower in grass hedge treatments than the control. Only 10%-31% of applied atrazine in Pennisetum hedge treatments was leached to lower depths and
3.4. Regression between atrazine loss and related factors Using the optimal scaling in SPSS, we evaluated the comparative importance of the grass hedge, rain intensity and slope gradient to the runoff and atrazine loss (Table 3). Their contributions followed decreasing effect on runoff were rain intensity, grass hedge and slope gradient; on atrazine loss were grass hedge, rain intensity and slope gradient. The analysis result confirmed the importance of grass hedges in the management of atrazine loss from sloping fields, also revealed a significant influence of the grass hedge and rain intensity on runoff and atrazine loss (P < 0.001, P = 0.009, respectively). The slope gradient was significantly related to runoff (P = 0.005) but not to atrazine loss (P = 0.057). Additionally, a strong positive linear relationship existed between the atrazine loss amount and runoff volume (Fig. 7A), but no statistically significant relationship existed between atrazine loss and atrazine concentration in runoff (Fig. 7B). 4. Discussion Rainfall and grass hedges were dominant factors influencing atrazine loss. The importance of the slope gradient was less. So the slope gradient was not discussed in a separate section. 4.1. Effects of the rainfall on atrazine loss The good linear correlation between atrazine loss and surface runoff indicated that greater loss of atrazine caused by higher rainfall intensity and/or steep slope mainly resulting from runoff increase. The linear relationship was consistent with the published studies in both 6
Agriculture, Ecosystems and Environment 253 (2018) 1–10
Q. Wang et al.
Fig. 6. Atrazine residue in soil of under rainfall intensity of 45 mm h−1 (A) in the control plots under the slope gradient of 15%; (B) in the plots with Melilotus hedge under the slope gradient of 15%; (C) in the plots with Pennisetum hedge under the slope gradient of 15%; (D) in the control plots under the slope gradient of 20%; (E) in the plots with Melilotus hedge under the slope gradient of 20%; and (F) in the plots with Pennisetum hedge under the slope gradient of 20%. *and ** indicate significance of the t-statistic between the slope gradient of 15% and 20% for the control, or Melilotus or Pennisetum hedges at the 0.05, and 0.01 levels, respectively. Different letters indicate significant differences between the control, Melilotus or Pennisetum hedges at P < 0.05 under the same slope gradient according to Duncan’s test. Error bars indicate SE of the mean (n = 3).
Table 2 Atrazine mass balance in the grass hedge system. Atrazine amount (g ha−1) −1
Rainfall intensity (mm h
)
30
Slope gradient Control
Surface runoff Soil layer
Surface runoff Soil layer
Surface runoff Soil layer
Unaccounted
30
45
20%
15%
20%
15%
20%
15%
20%
0–5 cm 5–10 cm 10–15 cm Sum
29.7 391.8 168.3 95.4 655.5 597.3
36.6 253.3 113.9 82.3 449.6 796.3
48.8 363.9 160.2 79.5 603.6 630.1
62.6 242.5 100.9 67.0 410.5 809.4
2.3 30.5 13.1 7.4 51.1 46.6
2.9 19.8 8.9 6.4 35.1 62.1
3.8 28.4 12.5 6.2 47.1 49.1
4.9 18.9 7.9 5.2 32.0 63.1
0–5 cm 5–10 cm 10–15 cm Sum
13.0 355.6 224.4 129.6 709.6 559.9
20.0 270.5 212.2 140.6 623.2 639.2
23.7 312.9 218.4 114.4 645.8 613.0
39.7 237.6 174.9 135.6 548.0 694.7
1.0 27.7 17.5 10.1 55.3 43.7
1.6 21.1 16.5 11.0 48.6* 49.8*
1.9 24.4 17.0 8.9 50.4 47.8
3.1 18.5 13.6 10.6 42.7* 54.2*
0–5 cm 5–10 cm 10–15 cm Sum
7.2 600.9 261.6 283.5 1146.0 129.3
12.0 488.4 200.4 179.1 867.9 402.6
12.6 344.8 250.5 212.2 807.5 462.4
26.1 299.4 201.3 199.9 700.5 555.9
0.6 46.9 20.4 22.1 89.4* 10.1*
0.9 38.1 15.6 14.0 67.7* 31.4*
1.0 26.9 19.5 16.5 63.0* 36.1*
2.0 23.3 15.7 15.6 54.6* 43.3*
Unaccounted Pennisetum
45
15%
Unaccounted Melilotus
Percentage to total applied atrazine (%)
Note: In the calculation of atrazine mass balance, soil residue in each layer at middle slope position represents the average residue level (the atrazine residue in the soil at different slope position is different). * means significant differences between grass hedge treatments and the control (P < 0.05).
conventional-tillage and strip-tillage fields (Potter et al., 2014). The atrazine loss rate caused by the single rainfall event with an intensity of 45 mm h−1 in our study approached the highest annual loss (4.7%) in the 9-year period investigation of Shipitalo and Owens (2006). They
observed that most of this loss (78%) was also the result of a single runoff event that began 2 d after herbicide application. Their another long-term study further confirmed that yearly loss was only 0.014% of the applied atrazine when the first runoff occurred 51 days after 7
Agriculture, Ecosystems and Environment 253 (2018) 1–10
Q. Wang et al.
Table 3 The contribution of grass hedge, rain intensity and slope gradient to surface runoff, atrazine loss based on optimal scaling regression analysis. Factor
Surface runoff Atrazine loss
Model ANOVA
Grass hedge
Rain intensity
Slope gradient
R2
p
Standardized coefficient
p
Importance
Standardized coefficient
p
Importance
Standardized coefficient
p
Importance
0.992
0.008
0.446
0.001
0.201
0.869
< 0.001
0.761
0.193
0.005
0.038
0.957
< 0.001
0.800
< 0.001
0.669
0.483
0.009
0.244
0.290
0.057
0.088
Fig. 7. Relationships of atrazine loss to overland flow (A) and atrazine concentration in runoff (B).
effectively reduced exported herbicide load, including dissolved-phase and sediment bound transport (Lin et al., 2011; Lafrance et al., 2013). Furthermore, the contour buffers planted in cultivated fields decreased the loss amount of pollutants mainly by inhibiting their mobilization (Dosskey, 2001). Pennisetum exhibited significantly higher effectiveness than Melilotus, particularly at the higher rain intensity. Previous report confirmed that the role of plants in controlling water and soil loss was possibly related to their root characteristics (Baets et al., 2007), and plants with denser stems and richer roots proved to be more effective (Xiao et al., 2010). The difference of efficacy between Melilotus and Pennisetum hedges may be explained by differences in their root systems: Pennisetum has a fibrous root system with more fine roots (diameter < 1 mm), a larger root surface area, and faster root growth (Cheng et al., 2008); while Melilotus has a taproot system with a smaller root surface area and less fine roots in the topsoil layer. The main mechanisms of grass hedges for reducing water and soil loss has been reported: 1) grass roots may increase soil porosity, thus forming water channels for infiltration; 2) dense and stiff stems of grass hedges may catch surface soil particles, thus enabling increased sediment deposition (Gilley et al., 2000; Blanco-Canqui et al., 2004). Consequently, grass hedges control transport of herbicides lost from fields through reducing water and sediment carriers of atrazine. The result of mass balance analysis indicated that grass hedge, especially Pennisetum, intercepted more amount of atrazine to retain in the topsoil and hence held promise for protecting ground water, which may be related to the increase of soil organic matter content facilitating strong sorption of the herbicides in grass hedge system. From these results, we concluded that grass hedges demonstrated their abilities to decrease atrazine export from agricultural lands mainly through reducing the runoff amount, as well as showed their potential for reducing the hazard of atrazine leaching. But the efficiency for decreasing the herbicide loss was limited by the slope gradient. Therefore, although grass hedges action alone maintained a certain protective function, other control measures (e.g. water harvesting, buffer strips on the edges of fields) may be
atrazine application (Shipitalo and Owens, 2003). This data implies that the rainfall intensity had greater effects on the total atrazine loss, and a few extreme rainfall events close to application time were responsible for most of the herbicide loss. These findings further support the conclusion of Potter et al. (2015) that the runoff close to pesticide application time was the determinant source in total pesticide loss. High rain intensity increased total atrazine runoff loss, but the mean atrazine concentrations in runoff were higher under lower rain intensity. This was mainly attributed to enough runoff volume provided by the higher rainfall intensity to reduce the concentration (Pote et al., 1994). As found in the present study, under 45 mm h−1 the runoff was nearly 5 times higher than under 30 mm h−1, while the former caused less than 2 times as much atrazine loss as the latter. The significant variation of atrazine residue at 0–5 cm soil depth under different rain intensities and slope gradients suggested that this soil layer was the key soil segment for herbicide loss after rainfall events. The deeper soil layer of 10–15 cm contained significantly lower atrazine with intensive rainfall. A study by Ouyang et al. (2016) reported an opposite observation that the herbicide leaching load increased under the highest rain intensity. The difference may have resulted from no surface runoff being generation in their soil column study. These results imply that the rain intensity strongly changed physical redistribution of residues by affecting the occurrence and volume of runoff. Therefore, in farming practice, soil application of atrazine should be likely to avoid favorable conditions for surface runoff formation, such as saturated soil and heavy rainfall.
4.2. Effects of grass hedges on atrazine loss Both Melilotus and Pennisetum grass hedges significantly reduce surface runoff and consequent atrazine loss, and their impacts on the control of atrazine loss is more important than rain intensity and slope gradient. As reported in other pesticide runoff management studies under simulated or natural rainfall conditions, grass filter strips 8
Agriculture, Ecosystems and Environment 253 (2018) 1–10
Q. Wang et al.
Baets, S., Poesen, J., Knapen, A., Barberá, G.G., Navarro, J.A., 2007. Root characteristics of representative Mediterranean plant species and their erosion-reducing potential during concentrated runoff. Plant Soil 294, 169–183. Baker, J.L., Mickelson, S.K., 1994. Application technology and best management practices for minimizing herbicide runoff. Weed Technol. 8, 862–869. Blanco-Canqui, H., Gantzer, C.J., Anderson, S.H., Alberts, E.E., 2004. Grass barriers for reduced concentrated flow induced soil and nutrient loss. Soil Sci. Soc. Am. J. 68, 1963–1972. Bouldin, J.L., Farris, J.L., Moore, M.T., Smith, S., 2006. Hydroponic uptake of atrazine and lambda-cyhalothrin in Juncus effusus and Ludwigia peploides. Chemosphere 65, 1049–1057. Bradley, P.M., Journey, C.A., Romanok, K.M., Barber, L.B., Buxton, H.T., Foreman, W.T., Furlong, E.T., Glassmeyer, S.T., Hladik, M.L., Iwanowicz, L.R., Jones, D.K., Kolpin, D.W., Kuivila, K.M., Loftin, K.A., Mills, M.A., Meyer, M.T., Orlando, J.L., Reilly, T.J., Smalling, K.L., Villeneuve, D.L., 2017. Expanded target-chemical analysis reveals extensive mixed-organic-contaminant exposure in U.S. streams. Environ. Sci. Technol. 51, 4792–4802. Cheng, X., Chen, W., Gu, B., Liu, X., Chen, F., Chen, Z., Zhou, X., Li, Y., Huang, H., Chen, Y., 2008. Morphology, ecology, and contaminant removal efficiency of eight wetland plants with differing root systems. Hydrobiologia 623, 77–85. Devlin, D.L., Regehr, D.L., Barnes, P.L., 2000. Managing to minimize atrazine runoff. Kansas State University Agricultural Experiment Station and Cooperative Extension Service MF2208. Kansas State University, Manhattan, KS. Dosskey, G.M., 2001. Toward quantifying water pollution abatement in response to installing buffers on crop land. Environ. Manage. 28, 577–598. Gilley, J.E., Eghball, B., Kramer, L.A., Mooreman, T.B., 2000. Narrow grass hedge effects on runoff and soil loss. J. Soil Water Conserv. 55, 190–196. Gong, D., Shi, P., Wang, J., 2004. Daily precipitation changes in the semi-arid region over northern China. J. Arid Environ. 59, 771–784. Humphry, J.B., Daniel, T.C., Edwards, D.R., Sharpley, A.N., 2002. A portable rainfall simulator for plot-scale runoff studies. Appl. Eng. Agric. 18, 199–204. Hussain, S., Siddique, T., Arshad, M., Saleem, M., 2009. Bioremediation and phytoremediation of pesticides: recent advances. Crit. Rev. Env. Sci. Tec. 39, 843–907. Jablonowski, N.D., Ffer, A.S., Burauel, P., 2011. Still present after all these years: persistence plus potential toxicity raise questions about the use of atrazine. Environ. Sci. Pollut. Res. 18, 328–331. Juergens, M.D., Chaemfa, C., Hughes, D., Johnson, A.C., Jones, K.C., 2015. PCB and organochlorine pesticide burden in eels in the lower Thames River (UK). Chemosphere 118, 103–111. Lafrance, P., Caron, E., Bernard, C., 2013. Impact of grass filter strips length on exported dissolved masses of metolachlor, atrazine and deethylatrazine: a four-season study under natural rain conditions. Soil Use Manage. 29, 87–97. Lin, C., Lerch, R.N., Goyne, K.W., Garrett, H.E., 2011. Reducing herbicides and veterinary antibiotics losses from agroecosystems using vegetative buffers. J. Environ. Qual. 40, 791–799. Mao, M., Ren, L., 2004. Simulating nonequilibrium transport of atrazine through saturated soil. Ground Water 42, 500–508. Mei, M., Du, Z., Cen, Y., 2011. QuEChERS-ultra-performance Liquid Chromatography Tandem Mass Spectrometry for Determination of Five Currently Used Herbicides. Chinese J. Anal. Chem. 39, 1659–1664. Milan, M., Vidotto, F., Piano, S., Negre, M., Ferrero, A., 2013. Buffer strip effect on terbuthylazine, desethyl-terbuthylazine and S-metolachlor runoff from maize fields in Northern Italy. Environ. Technol. 34, 71–80. Ouyang, W., Huang, W., Wei, P., Hao, F., Yu, Y., 2016. Optimization of typical diffuse herbicide pollution control by soil amendment configurations under four levels of rainfall intensities. J. Environ. Manage. 175, 1–8. Piepho, H.P., Büchse, A., Emrich, K., 2003. A Hitchhiker's Guide to Mixed Models for Randomized Experiments. J. Agron. Crop Sci. 189, 310–322. Pote, D.H., Daniel, T.C., Edwards, D.R., Mattice, J.D., Wickliff, D.B., 1994. Effect of drying and rainfall intensity on cyromazine loss from surface-applied caged-layer manure. J. Environ. Qual. 23, 101–104. Potter, T.L., Bosch, D.D., Strickland, T.C., 2014. Comparative assessment of herbicide and fungicide runoff risk: A case study for peanut production in the Southern Atlantic Coastal Plain (USA). Sci. Total Environ. 490, 1–10. Potter, T.L., Bosch, D.D., Strickland, T.C., 2015. Tillage impact on herbicide loss by surface runoff and lateral subsurface flow. Sci. Total Environ. 530–531, 357–366. Reichenberger, S., Bach, M., Skitschak, A., Frede, H., 2007. Mitigation strategies to reduce pesticide inputs into ground- and surface water and their effectiveness: a review. Sci. Total Environ. 384, 1–35. Richter, C.A., Papoulias, D.M., Whyte, J.J., Tillitt, D.E., 2016. Evaluation of potential mechanisms of atrazine-induced reproductive impairment in fathead minnow (Pimephales promelas) and Japanese medaka (Oryzias latipes). Environ. Toxicol. Chem. 35, 2230–2238. Ries, J.B., Seeger, M., Iserloh, T., Wistorf, S., Fister, W., 2009. Calibration of simulated rainfall characteristics for the study of soil erosion on agricultural land. Soil Till. Res. 106, 109–116. Rippy, M.A., Deletic, A., Black, J., Aryal, R., Lampard, J., Tang, J.Y., McCarthy, D., Kolotelo, P., Sidhu, J., Gernjak, W., 2017. Pesticide occurrence and spatio-temporal variability in urban run-off across Australia. Water Res. 115, 245–255. Schwarzenbach, R.P., Egli, T., Hofstetter, T.B., von Gunten, U., Wehrli, B., 2010. Global water pollution and human health. Annu. Rev. Environ. Resour. 35, 109–136. Sharpley, A., Kleinman, P., 2003. Effect of Rainfall simulator and plot scale on overland flow and phosphorus transport. J. Environ. Qual. 32, 2172–2179. Shipitalo, M.J., Owens, L.B., 2003. Atrazine, deethylatrazine, and deisopropylatrazine in surface runoff from conservation tilled watersheds. Environ. Sci. Technol. 37, 944–950.
needed to achieve sufficient protection under the steep slope conditions (particularly in frequent intense rain regions). The present finding may be helpful in controlling pesticide losses from sloping cropland. 4.3. Limitation and prospective In present study, the loss of atrazine adsorbed on soil particulates in runoff was not taken into consideration, which somewhat influenced accurate evaluation of atrazine loss. But the underestimate is unlikely to be marked, because only 1.0-2.1 g sediment was obtained after being filtered and dried from 1 L total runoff. Research typically found that most of the atrazine transported was associated with the water fraction because of the greater amounts of water lost as compared with soil, the total amounts lost in water were about 7 and 10 times greater than in the soil (White et al., 1967). Further, Devlin et al. (2000) confirmed that more than 90 percent of atrazine loss occurred in runoff water. Our study focused on atrazine due to its worldwide use with a long application history, relative long-term persistence, and potential threat to human and ecosystem health. So the findings are limited to the loss of pesticides with similar properties as atrazine. Pesticide loss problems are usually associated with natural conditions of weather. Though simulated rainfall has been widely used to conduct research on runoff generation and soil erosion on agricultural land, there are performance limitations due to its inability to simulate all aspects of natural storms, particularly in terms of drop fall velocity and rainfall distribution on the plot (Humphry et al., 2002; Ries et al., 2009). Therefore, under natural rainfall conditions, more comprehensive studies are needed to better estimate the loss of pesticides with widely different properties within the grass hedge system. 5. Conclusions The current plot study evaluated the effectiveness of grass hedges in reducing surface runoff and accompanying atrazine loss, and compared the importance of the grass hedge, rainfall intensity and slope gradient to atrazine loss. The main conclusions are as follows: (1) Atrazine loss in a single rainfall induced surface runoff was less than 5% of applied amount under the experimental conditions. The intense rainfall and steep slope increased atrazine loss. (2) Both Pennisetum and Melilotus hedges were efficient in controlling atrazine loss in Northern China. Pennisetum provided more effective protection, and can be preferentially selected when designing grass hedges in similar climatic regions. Under heavy rainfall and steep slope conditions, other engineering measures incorporating with grass hedges should be devised for sufficient reducing pesticide runoff. (3) The grass hedge and rain intensity were significant factor for atrazine loss, and had more important effects than slope gradient. Further researches are necessary to fully evaluate the efficiency of grass hedges in reducing pesticide runoff under natural rainfall and broader scales conditions, and a comprehensive analysis of protective mechanisms of grass hedges against runoff losses of atrazine are needed. Acknowledgments The research was supported by the National Natural Science Foundation (31370540) and National Key Technologies R & D Program (2012BAD14B02) of China. References Atwood, D., Paisley-Jones, C., 2017. Pesticides industry sales and usage: 2008-2012 market estimates. US Environmental Protection Agency, Washington, DC.
9
Agriculture, Ecosystems and Environment 253 (2018) 1–10
Q. Wang et al.
Wille, K., Kiebooms, J.A.L., Claessens, M., Rappé, K., Vanden Bussche, J., Noppe, H., Van Praet, N., De Wulf, E., Van Caeter, P., Janssen, C.R., De Brabander, H.F., Vanhaecke, L., 2011. Development of analytical strategies using U-HPLC-MS/MS and LC-ToF-MS for the quantification of micropollutants in marine organisms. Anal. Bioanal. Chem. 400, 1459–1472. Wu, J.Y., Huang, D., Teng, W.J., Sardo, V.I., 2010. Grass hedges to reduce overland flow and soil erosion. Agron. Sustain. Dev. 30, 481–485. Xiao, B., Wang, Q., Wang, H., Dai, Q., Wu, J., 2011. The effects of narrow grass hedges on soil and water loss on sloping lands with alfalfa (Medicago sativa L.) in Northern China. Geoderma 167–168, 91–102. Xiao, B., Wang, Q., Wu, J., Huang, C., Yu, D., 2010. Protective function of narrow grass hedges on soil and water loss on sloping croplands in Northern China. Agr. Ecosyst. Environ. 139, 653–664. Young, F.W., de Leeuw, J., Takane, Y., 1976. Regression with qualitative and quantitative variables: An alternating least squares method with optimal scaling features. Psychometrika 41, 505–529. Zhong, Y., Jia, Y., Li, Z., 2013. Spatial and temporal changes of maximum 1 h precipitation intensity in Beijing region in last 53 years. J. China Hydrol. 33, 32–37 (in Chinese with English abstract).
Shipitalo, M.J., Owens, L.B., 2006. Tillage system, application rate, and extreme event effects on herbicide losses in surface runoff. J. Environ. Qual. 35, 2186–2194. Stehle, S., Schulz, R., 2015. Pesticide authorization in the EU-environment unprotected? Environ. Sci. Pollut. Res. 22, 19632–19647. Stone, W.W., Gilliom, R.J., Ryberg, K.R., 2014. Pesticides in US streams and rivers: occurrence and trends during 1992-2011. Environ. Sci. Technol. 48, 11025–11030. Vianello, M., Vischetti, C., Scarponi, L., Zanin, G., 2005. Herbicide losses in runoff events from a field with a low slope: role of a vegetative filter strip. Chemosphere 61, 717–725. Vieira, D.A.N., Dabney, S.M., 2012. Two-dimensional flow patterns near contour grass hedges. Hydrol. Process. 26, 2225–2234. Vymazal, J., Březinová, T., 2015. The use of constructed wetlands for removal of pesticides from agricultural runoff and drainage: a review. Environ. Int. 75, 11–20. Walter, E.S., Lyle, A.D., 1962. Effects of environment upon herbicides applied preemergence. Weeds 10, 304–307. Wei, W., Chen, L., Fu, B., Huang, Z., Wu, D., Gui, L., 2007. The effect of land uses and rainfall regimes on runoff and soil erosion in the semi-arid loess hilly area, China. J. Hydrol. 335, 247–258. White, A.W., Barnett, A.P., Wright, B.G., Holladay, J.H., 1967. Atrazine losses from fallow land caused by runoff and erosion. Environ. Sci. Technol. 1, 740–744.
10
Contents lists available at ScienceDirect
Agriculture, Ecosystems and Environment journal homepage: www.elsevier.com/locate/agee
Research Paper
Effect of grass hedges on runoff loss of soil surface-applied herbicide under simulated rainfall in Northern China
MARK
⁎
Qinghai Wanga, , Cui Lia, Zhuo Panga, Haifeng Wena, Ruilun Zhenga, Jie Chena, Xueju Maa, ⁎⁎ Xiaoe Queb, a b
Beijing Research & Development Center for Grass and Environment, Beijing Academy of Agriculture and Forestry Sciences, Beijing, China Institute of Desertification Studies, Chinese Academy of Forestry, Beijing, China
A R T I C L E I N F O
A B S T R A C T
Keywords: Grass hedge Atrazine loss Surface runoff Sloping cropland
Pesticide loss triggered by runoff is one of the most important sources of non-point pollution. The effects of grass hedges (Melilotus albus and Pennisetum alopecuroides) on atrazine runoff under different rain intensities and slope gradients were evaluated. The plot-scale experiments were carried out on a maize (Zea mays) field on slopes with 15% and 20% gradients using simulated rainfall (rain intensity of 30 and 45 mm h−1). Atrazine residues were investigated in runoff water and soil taken from three depths (0–5, 5–10, and 10–15 cm) in the middle and base of the slope after runoff events. Total atrazine loss in runoff water ranged from 2.3% to 4.9% of that applied from plots without grass hedges. Grass hedges decreased atrazine loss by 37%-76% and surface runoff by 27%-72%, and Pennisetum showed better efficacy than Melilotus, especially under higher rain intensity. Atrazine loss showed a significant positive linear correlation with surface runoff volume. Grass hedges had a more significant effect on atrazine loss than rain intensity and slope gradient. But they functioned less effectively if used under intense rain or/and steep slope conditions. Atrazine residues remained in the surface 15 cm soil were higher for the plots with grass hedges than the control plots. These results suggest that grass hedges not only significantly reduced atrazine loss by reducing the surface runoff, but also reduced the amount of atrazine leaching to deeper soil layer. P. alopecuroides was a suitable grass-hedge species for controlling atrazine losses in northern China and similar regions. Other management practices or control measures should be integrated with grass hedges in strongly sloping cropland in high-rainfall areas to maintain pesticide losses at an acceptable level.
1. Introduction World pesticide usage at the producer level totaled nearly 2.7 × 109 kg annually in both 2011 and 2012 (Atwood and PaisleyJones, 2017). Given the magnitude of global pesticide use, widespread pesticide contamination of surface water and groundwater is expected (Rippy et al., 2017). During 1992-2011, the proportion of assessed streams with one or more pesticides that exceeded an aquatic life benchmark was 61% for agricultural streams in USA (Stone et al., 2014). In developing countries, contamination of ground and surface water with pesticides is more prevalent (Schwarzenbach et al., 2010). Because it is biologically active, water pollution caused by pesticides used in agriculture is a major concern worldwide (Juergens et al., 2015; Stehle and Schulz, 2015). It poses a significant health risk to humans and other animals (Hussain et al., 2009). Pesticides deposition in soil is one of the major pollution sources. These pesticides can enter surrounding surface waters mainly through diffuse sources, surface runoff ⁎
and erosion is one of the most important pathways of pesticide input in aquatic ecosystems among diffuse pollution (Vymazal and Březinová, 2015). Surface runoff water acted as the major mobile carrier of pesticide mobilization from agricultural land to ecosystem, because sediment amount is usually small compared to the runoff volume from a field (Reichenberger et al., 2007). Pesticides losses linked to runoff during the growing season have been reported to account for greater than 60% of the total loss (Potter et al., 2014). Herbicide concentrations in the runoff obtained during the first rainfall events after treatment, had a high probability of exceeding the ecotoxicological endpoint for algae (Vianello et al., 2005). Therefore, adequate control of runoff, and subsequent pesticide loss is very important. Among the various control measures, grass hedges, a special type of vegetative filter strips defined as narrow strips of stiff-stemmed grass used to control runoff and erosion (Vieira and Dabney, 2012), are commonly used in preventing soil and water loss on sloping croplands all over the world due to their low cost, eco-friendliness and efficiency (Xiao et al., 2011).
Corresponding author at: Beijing Research and Development Center for Grass and Environment, No. 9 Shuguang Garden Road of Haidian District, Beijing, 100097, China. Corresponding author. E-mail addresses: [email protected] (Q. Wang), [email protected] (X. Que).
⁎⁎
http://dx.doi.org/10.1016/j.agee.2017.10.024 Received 6 July 2017; Received in revised form 14 October 2017; Accepted 27 October 2017 0167-8809/ © 2017 Elsevier B.V. All rights reserved.
Agriculture, Ecosystems and Environment 253 (2018) 1–10
Q. Wang et al.
Fig. 1. Schematic diagram of experimental plot, T1 = Melilotus hedge, T2 = Pennisetum hedge, and CK = control.
This study also gains data for future use in predicting potential pesticide loss due to runoff and evaluating the risk to water quality.
Atrazine is a photosynthesis inhibiting triazine herbicide that has been used for approximately 50 years and has become one of the most heavily applied agricultural herbicides in the world (Jablonowski et al., 2011). Furthermore, it exhibits relatively high water solubility (32 mg L−1) and moderate environmental persistence (half-life in soil: 146 d) with adsorption coefficient (logKoc) of 1.97 (Bouldin et al., 2006), and is frequently detected in surface water and shallow groundwater (Bradley et al., 2017). It has become the subject of continuous concern due to its potential endocrine and carcinogenic activity (Jablonowski et al., 2011; Richter et al., 2016). Atrazine has been widely accepted as an important tool in controlling weeds in summer maize, a main crop in the predominant cropping system in Northern China where there is severe soil and water loss (Mao and Ren, 2004; Xiao et al., 2011). Atrazine application period is just in the rainy season, and frequent strong thunderstorms with such features as high intensity and short duration typically occur in this region, thus atrazine has huge potential of losses in surface runoff with water and sediment in this region. In terms of runoff-producing rainfall events after herbicide application, the first few storms are often responsible for the largest herbicide losses (Baker and Mickelson, 1994). These extreme events dominate herbicide transport and are critical for devising management practices or control measures, and monitoring total pesticide loss (Shipitalo and Owens, 2006). It is, therefore, of great significance and broadness to evaluate the effectiveness of grass hedges in reducing herbicide loss, as well as to quantify atrazine loss from sloping agricultural fields in runoff events, especially rainstorms shortly after its application. Melilotus (Melilotus albus) and Pennisetum (Pennisetum alopecuroides) are promising grass hedge candidates due to their native and non-invasive characteristics, tolerance to the climate extremes, as well as sufficient stem strength for blocking water flow, and provide an efficient way to prevent soil and water loss on sloping croplands (Wu et al., 2010; Xiao et al., 2011). However, little is known about their effectiveness on herbicide loss and their influences on herbicide distribution in soil at different rain intensities and slope gradients. Achieving such information may help to seek for suitable grass species to develop grass hedges for the temperate regions, and to explore the potential of grass hedges alone for reducing pesticide loss to an acceptable level under extreme conditions (such as heavy rainfall, steep slope). Consequently, the main objectives of this research were to (1) estimated runoff loss of atrazine for storms of high intensity and short duration occurring shortly (4 h) following application to soil surface, (2) determine the efficacy of the two types of hedges (Melilotus and Pennisetum) in decreasing atrazine loss and their influences on atrazine distribution in soil, and (3) analyze the relative importance of grass hedges, rain intensity and slope gradient for atrazine losses. This research extends previous work of Wu et al. (2010) and Xiao et al. (2011) by focusing on herbicide loss rather than water and soil loss, and quantifying the correlations between herbicide loss and surface runoff.
2. Materials and methods 2.1. Study site The study was carried out at the National Experimental Station for Precision Agriculture (116°26′ E, 40°10′ N) in Beijing, China. The experimental station is located in North China and is characterized by a continental, semi-humid, and monsoon climate. The mean precipitation is 640 mm y−1, 80% of which occurs between June and August. The mean annual temperature is 11.5 °C. Rotation of winter wheat and summer maize dominates agricultural activities in the region. The soil is brown fluvo aquic soil in type and loamy clay in texture (41% sand, 24% silt and 35% clay) with a pH of 7.6. The bulk density is 1.37 g cm−3, The soil organic matter content and the total nitrogen and phosphorus are 14.0, 2.46 and 0.63 g kg−1. 2.2. Experimental design and herbicide application The experiment was based on three independent variables: grass hedges (Melilotus hedges, Pennisetum hedges, and without grass hedges as control), slope gradients (15% and 20%), and rainfall intensities (30 and 45 mm h−1). Eighteen 1.5 m by 12 m plots (3 hedge levels × 2 slope gradients × 3 replications) were established. Plots were separated by 2-mm-wide sheet metal frames driven approximately 30 cm into the soil in two sides of the plots (but not the top and bottom). The grass hedges were located at the middle and bottom positions of selected test plots. They were established in June 2012 in two parallel rows perpendicular to the slope direction (Fig. 1). The grass hedges were thick enough (densities of Melilotus and Pennisetum was 684 and 125 stems m−2) when the experiments were conducted. Summer maize was sown manually in all plots of every treatment after the harvest of winter wheat without tillage in both years (27 June 2013 and 20 June 2014, respectively). Afterwards, atrazine (38% Suspension concentrates) was applied on the soil surface using a hand sprayer at recommended field dosage [1282.5 g ha−1 (ai)]. All sprays were applied on the day that the summer maize was sown. 2.3. Rainfall Simulation and Runoff Collection Rainfall simulation has been used extensively as a cost-effective method to evaluate the effects of managements on infiltration, surface runoff, soil erosion, and contaminant transport in overland flow (Humphry et al., 2002; Sharpley and Kleinman, 2003). It exhibits advantages in expediting data collection, as well as creating controlled and reproducible artificial rainfall for comparison among locations (Humphry et al., 2002). In this study, precipitation was simulated using 2
Agriculture, Ecosystems and Environment 253 (2018) 1–10
Q. Wang et al.
Fig. 2. Surface runoff and effectiveness of grass hedges in reducing runoff under different slope gradients and rainfall intensities: (A) runoff under rainfall intensity of 30 mm h−1; (B) runoff under rainfall intensity of 45 mm h−1; (C) effectiveness of grass hedges in reducing runoff under rainfall intensity of 30 mm h−1; and (D) effectiveness of grass hedges in reducing runoff under rainfall intensity of 45 mm h−1. *and ** indicate significance of the t-test between the slope gradient of 15% and 20% within control, Melilotus or Pennisetum hedges at P < 0.05, and P < 0.01, respectively. Different letters indicate significant differences between the control, Melilotus or Pennisetum hedges at P < 0.05 under the same slope gradient according to Duncan’s test. Error bars indicate SE of the mean (n = 3).
Fig. 3. The atrazine concentration in runoff water under the rain intensities of 30 mm h−1 (A) and 45 mm h−1 (B); there were no significant treatment effects. Error bars indicate SE of the mean (n = 3).
and 0.22 MPa. The rainfall coefficient of uniformity measured according to the Christiansen method averaged 86% (Xiao et al., 2010). Generally speaking, moderate to extreme rains contribute 68.2% of the precipitation amount in Northern China (Gong et al., 2004). Compared
a rainfall simulator which can sprinkle water at intensities of 10 to 100 mm h−1. Three types of nozzles were used in the simulator (PROS17, PROS-15 and microsprinkler with flow rate of 102 L h−1, Hunter Company in USA), and the working pressures ranged between 0.16 MPa 3
Agriculture, Ecosystems and Environment 253 (2018) 1–10
Q. Wang et al.
Fig. 4. Atrazine loss and effectiveness of grass hedges in reducing atrazine loss under different slope gradients and rainfall intensities: (A) atrazine loss under rainfall intensity of 30 mm h−1; (B) atrazine loss under rainfall intensity of 45 mm h−1; (C) proportion of the amount atrazine loss to application under rainfall intensity of 30 mm h−1; (D) proportion of the amount atrazine loss to application under rainfall intensity of 45 mm h−1. *and ** indicate significance of the t-statistic between the slope gradient of 15% and 20% for the control, or Melilotus or Pennisetum hedges at the 0.05, and 0.01 levels, respectively. Different letters indicate significant differences between the control, Melilotus or Pennisetum hedges at P < 0.05 under the same slope gradient according to Duncan’s test. Error bars indicate SE of the mean (n = 3).
through a paper filter. Filtered runoff water was stored at −20 °C for later analysis for atrazine residue.
to rainfall events of medium intensity and medium duration, rainfalls with high intensity and short duration cause the greater proportion of runoff and soil loss (Wei et al., 2007). In study area during natural rainfall events, annual maximum 1 h precipitation intensity fell in the range 30–50 mm h−1 in last decades (Zhong et al., 2013). For those reasons, the intensities employed for this study were 30 and 45 mm h−1, representing a rang for strong thunderstorms of this region. To avoid the effect of atrazine retained in the soil on the later rainfall intensity test, these two intensities were designed in 2013 and 2014, respectively. In the second studied year (2014), in order to check for the presence of atrazine residues from the last year, soil was sampled on the day before herbicide application in the upper 5 cm layer (the most critical segment for runoff and strongly influenced the concentrations of chemicals in runoff water (Milan et al., 2013)). The determined data showed that the soil samples contained no detectable atrazine (< 0.2 μg kg−1) from the previous year prior to the application of herbicide. Heavy rain can transport soil surface-applied atrazine to the deep soil beyond the weed germination zone and such reduce atrazine performance (Walter and Lyle, 1962). The effective atrazine performance required an interval of at least 4 hours between herbicide application and heavy rain in agriculture practice. The simulated rainfall was applied 4 h after herbicide application for a duration of 1 h. Before rainfall simulation, soil water contents in 0–15 cm layer on the upper, middle, and bottom slope position of plots in 2013 were 20.3%, 21.0% and 22.6%, respectively; and in 2014 were 20.1%, 20.5% and 22.5%, respectively. Soil water content at same slope position did not differ significantly between 2013 and 2014. The surface runoff was collected in containers at the lower edge of each plot, and volumes were measured. A 1 L sample of total runoff with sediment was obtained from each container after thorough agitation, and was filtered immediately
2.4. Soil sampling Soil in each plot was sampled after the end of the runoff. Two sampling positions, representing middle and bottom slope positions, were located within the grass hedges. The corresponding sampling positions were selected in the control plots. Composite samples of five cores were taken randomly at depths of 0-5, 5-10 and 10–15 cm at each slope position using a soil sampling drill (2.0 cm diameter, 10.0 cm depth) in each plot. After collection, the composite soil samples were freeze-dried and kept at −20 °C until analysis. They were stored for no more than 7 d. 2.5. Atrazine analysis 2.5.1. Atrazine extraction from water sample Samples of 2 mL filtered water were extracted using 0.2 g C18 cartridge under vortex oscillating for 2 min, then separated at 4 °C in a centrifuge at 10000 rpm for 5 min. 1 mL supernatant from the extracts was evaporated with N2, and 1 mL acetonitrile/water mixture (1:1, v/v) was added immediately. The solution passed through 0.22 μm microporous filter before the analysis using ultra-high performance liquid chromatography coupled with tandem mass spectrometry (UPLC-MS/ MS) (Waters, USA), according to Wille et al. (2011). 2.5.2. Atrazine extraction from soil sample Soil samples were extracted 2 times. A 5 g sample was put into a 50mL polystyrene centrifuge tube, subsequently 1 g anhydrous magnesium sulfate and 10 ml acetonitrile were added. After being shaken for 4
Agriculture, Ecosystems and Environment 253 (2018) 1–10
Q. Wang et al.
3. Results
Table 1 Effects of experimental factors on atrazine residue in soil. Factors
r g sg sp sd r*g r*sg r*sp r*sd g*sg g*sp g*sd sg*sp sg*sd sp*sd r*g*sd r*sg*sp r*sg*sd r*sp*sd g*sg*sp sg*sp*sd g*sg*sd g*sp*sd r*g*sg*sp r*sg*sp*sd r*g*sg*sd r*g*sp*sd g*sg*sp*sd r*g*sg*sp*sd
Degree of freedom
1 2 1 1 2 2 1 1 2 2 2 4 1 2 2 4 1 2 2 2 2 4 4 2 2 4 4 4 4
3.1. Surface runoff
Atrazine residue in soil Sum of Squares
F
P
337320.590 1138303.909 540444.091 137029.571 3182346.965 133592.200 5479.289 19126.766 57911.676 11036.823 12257.606 163651.400 23781.671 115852.067 33948.281 74414.867 623.832 5509.649 28520.222 42709.709 9688.345 30007.996 39133.264 4980.972 4424.210 7752.228 40401.601 30957.858 7236.321
113.08 190.80 181.17 45.94 533.41 22.39 1.84 6.41 9.71 1.85 2.05 13.72 7.97 19.42 5.69 6.24 0.21 0.92 4.78 7.16 1.62 2.51 3.28 0.83 0.74 0.65 3.39 2.59 0.61
< 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 0.1774 0.0124 0.0001 0.1610 0.1319 < 0.0001 0.0054 < 0.0001 0.0042 0.0001 0.6481 0.3995 0.0098 0.0011 0.2007 0.0441 0.0132 0.4360 0.4782 0.6280 0.0111 0.0389 0.6586
Obvious changes in surface runoff were observed under different grass hedge, rain intensity and slope gradient treatments (Fig. 2). The total amount of surface runoff under the rain intensity of 30 mm h−1 was markedly smaller than that under 45 mm h−1 (P < 0.001) (Fig. 2A, B). Grass hedges reduced surface runoff in the range of 27% to 72% compared to the control (Fig. 2C, D). The runoff in all plots increased with slope gradients, and this difference was significant for the control and Melilotus at higher rain intensity (45 mm h−1) (P = 0.005, 0.002, respectively) (Fig. 2B), but was not significant under at 30 mm h−1 rain intensity (P = 0.090) (Fig. 2A). Regardless of the rain intensity and slope angle, plots with grass hedges released pronouncedly less runoff compared to the control (P < 0.001). Moreover, the highest surface runoff occurred in the control plots, and the lower in the plots with Pennisetum hedge. The difference in surface runoff was significant between the three treatments for the 45 mm h−1 rainfall intensity (P < 0.001). However, for the 30 mm h−1 rainfall intensity, there was no significant difference between Pennisetum and Melilotus treatments (P = 0.197). 3.2. Atrazine in surface runoff Under the conditions of same rain intensity, the grass hedge and slope gradient did not lead to significant differences of atrazine concentration in runoff (P = 0.7307, 0.5260, respectively), but atrazine concentrations in the runoff were significantly higher at the intensity of 30 mm h−1 than at 45 mm h−1 (P < 0.001) (Fig. 3). The control plots had the highest amount of atrazine loss, followed by plots with Melilotus and Pennisetum hedges (Fig. 4A, B). For the control, Melilotus, and Pennisetum hedges, atrazine losses at 45 mm h−1 intensity with 20% slope gradient were highest at 63, 40, and 26 mg ha−1, respectively; atrazine losses at 30 mm h−1 intensity with 15% slope gradient were lowest at 30, 13, and 7 mg ha−1, respectively. Other combinations of slope and rainfall were intermediate. Atrazine loss rate was 0.6% to 4.9% depending on the grass hedge, rain intensity and slope gradient. Of the applied atrazine 0.6%-3.1% was lost in runoff water from plots protected by grass hedges, compared with 2.3%-4.9% when grass hedges were not present (Fig. 4C, D). Atrazine loss increased with slope increase for both planted and unplanted grass hedges. At the rain intensity of 45 mm h−1, the amount and rate of atrazine loss from control, Melilotus and Pennisetum plots with the slope gradient of 20% were all significantly higher than that with the slope gradient of 15% (P = 0.025, 0.005 and 0.041, respectively). For the same slope, the amount and rate of atrazine loss from control, Melilotus and Pennisetum plots at the rain intensity of 45 mm h−1 were significantly higher than that at the rain intensity of 30 mm h−1 (for 15% slope, P = 0.020, 0.039 and 0.019, respectively; for 20% slope, P = 0.028, 0.005 and 0.040, respectively). The reduction of atrazine loss compared to the controls was 56% and 45% in the Melilotus treatment and of 76% and 67% in the Pennisetum treatment, respectively for 15% and 20% slope gradient. There was a significant difference in atrazine loss between plots with and without grass hedges (P < 0.001), and Pennisetum was significantly more effective in reducing atrazine losses than Melilotus (P < 0.001), except at 30 mm h−1 rain intensity with 20% slope. At the rain intensity of 45 mm h−1, Melilotus and Pennisetum hedges exhibited a significant decrease in efficiencies at 20% slope compared with the case at 15% slope (P = 0.043 and 0.044, respectively).
Note: r = rainfall intensity, g = grass hedge, sg = slope gradient, sp = slope position, sd = soil depth.
30 min on a horizontal shaker, the mixture was separated in a centrifuge at 10000 rpm for 5 min at 4 °C. Repeated extraction was carried out twice, and the supernatants were combined. Then the soil final extract was evaporated to dryness at 40 °C, and the residue was made up to a volume of 5 mL with acetonitrile (Mei et al., 2011). The following extraction, purification and analysis procedures were the same as for the water sample. Soil recovery for atrazine was 95.2%, and the detection limits were 0.05 μg L−1 for atrazine in water and 0.2 μg kg−1 in soil. There was a small quantity of soil in runoff obtained from each plot, but not enough for chemical analysis. Therefore, atrazine loss with sediment could not be included in the calculation.
2.6. Data analysis Atrazine loss rate was expressed as the percentage of applied atrazine lost. Data were analyzed using the Mixed Procedure (SAS Ins., Chicago, USA) according to the method of Piepho et al. (Piepho et al., 2003). Probability of 0.05 or less was considered significant, and the results were expressed as mean ± standard error. SPSS optimal scaling (SPSS Inc., Chicago, USA) was performed to determine the comparative importance of the grass hedge (defined as unordered categorical variable), rain intensity and slope gradient to the surface runoff and atrazine loss. Optimal scaling is a useful method in the analysis of nominal and ordinal variables, and detailed information about this method can be found in Young et al. (1976).
3.3. Atrazine in soil After application atrazine residue levels in the top 15 cm soil were significantly influenced by the rain intensity, grass hedge, slope 5
Agriculture, Ecosystems and Environment 253 (2018) 1–10
Q. Wang et al.
Fig. 5. Atrazine residue in soil of under rainfall intensity of 30 mm h−1 (A) in the control plots under the slope gradient of 15%; (B) in the plots with Melilotus hedge under the slope gradient of 15%; (C) in the plots with Pennisetum hedge under the slope gradient of 15%; (D) in the control plots under the slope gradient of 20%; (E) in the plots with Melilotus hedge under the slope gradient of 20%; and (F) in the plots with Pennisetum hedge under the slope gradient of 20%. *and ** indicate significance of the t-statistic between the slope gradient of 15% and 20% for the control, or Melilotus or Pennisetum hedges at the 0.05, and 0.01 levels, respectively. Different letters indicate significant differences between the control, Melilotus or Pennisetum hedges at P < 0.05 under the same slope gradient according to Duncan’s test. Error bars indicate SE of the mean (n = 3).
degraded under the above rainfall conditions.
gradient, slope position and soil depth (Table 1, Figs. 5 and 6). Soil atrazine residue at 30 mm h−1 rain intensity was significantly higher than that at 45 mm h−1 (P < 0.001). Soil in the plots with grass hedges contained notably higher atrazine residues compared to the plots without grass hedges (P < 0.001), and residue levels in soils under Pennisetum hedge were significantly higher than those under Melilotus hedge (P = 0.021). The soils with 15% gradient or at bottom slope position maintained significantly higher residue level in comparison with the plots with the slope gradient of 20% or at middle slope position (P < 0.001). Atrazine residue levels showed a decreasing trend with soil depth, and differences were significant between all soil layers (P < 0.001). The rain intensity, grass hedge, slope gradient and slope position had significant interaction effects with soil depth on atrazine residue content (P = 0.003, < 0.001, < 0.001, and 0.032, respectively). Under conditions of heavier rain intensity or steeper slope, atrazine residues were significantly lower at 0–5 cm depths (P = 0.043, 0.017, respectively), but there was no significant difference at 5-10 (P = 0.623, 0.106, respectively) and 10–15 cm depths (P = 0.281, 0.247, respectively). Conversely, soil atrazine residues at 5-10 and 10–15 cm depths within grass hedges were higher than those in plots without grass hedges (P = 0.016, 0.019, respectively), while the significant difference was not found in soil layer of 0–5 cm (P = 0.272). As can be seen in Table 2, atrazine in the soil at 0–15 cm depth was higher for the plots with grass hedges than the control plots. Particularly for the plots with Pennisetum hedge, 68%-89% of applied atrazine still remained in the top 15 cm soil after rainfall events at an intensity of 30 mm h−1 for 1 h. The unaccounted-for atrazine in the mass balance analysis mainly included the quantity leaching to deeper soil layers (below 15 cm depth) and the quantity degraded. It was lower in grass hedge treatments than the control. Only 10%-31% of applied atrazine in Pennisetum hedge treatments was leached to lower depths and
3.4. Regression between atrazine loss and related factors Using the optimal scaling in SPSS, we evaluated the comparative importance of the grass hedge, rain intensity and slope gradient to the runoff and atrazine loss (Table 3). Their contributions followed decreasing effect on runoff were rain intensity, grass hedge and slope gradient; on atrazine loss were grass hedge, rain intensity and slope gradient. The analysis result confirmed the importance of grass hedges in the management of atrazine loss from sloping fields, also revealed a significant influence of the grass hedge and rain intensity on runoff and atrazine loss (P < 0.001, P = 0.009, respectively). The slope gradient was significantly related to runoff (P = 0.005) but not to atrazine loss (P = 0.057). Additionally, a strong positive linear relationship existed between the atrazine loss amount and runoff volume (Fig. 7A), but no statistically significant relationship existed between atrazine loss and atrazine concentration in runoff (Fig. 7B). 4. Discussion Rainfall and grass hedges were dominant factors influencing atrazine loss. The importance of the slope gradient was less. So the slope gradient was not discussed in a separate section. 4.1. Effects of the rainfall on atrazine loss The good linear correlation between atrazine loss and surface runoff indicated that greater loss of atrazine caused by higher rainfall intensity and/or steep slope mainly resulting from runoff increase. The linear relationship was consistent with the published studies in both 6
Agriculture, Ecosystems and Environment 253 (2018) 1–10
Q. Wang et al.
Fig. 6. Atrazine residue in soil of under rainfall intensity of 45 mm h−1 (A) in the control plots under the slope gradient of 15%; (B) in the plots with Melilotus hedge under the slope gradient of 15%; (C) in the plots with Pennisetum hedge under the slope gradient of 15%; (D) in the control plots under the slope gradient of 20%; (E) in the plots with Melilotus hedge under the slope gradient of 20%; and (F) in the plots with Pennisetum hedge under the slope gradient of 20%. *and ** indicate significance of the t-statistic between the slope gradient of 15% and 20% for the control, or Melilotus or Pennisetum hedges at the 0.05, and 0.01 levels, respectively. Different letters indicate significant differences between the control, Melilotus or Pennisetum hedges at P < 0.05 under the same slope gradient according to Duncan’s test. Error bars indicate SE of the mean (n = 3).
Table 2 Atrazine mass balance in the grass hedge system. Atrazine amount (g ha−1) −1
Rainfall intensity (mm h
)
30
Slope gradient Control
Surface runoff Soil layer
Surface runoff Soil layer
Surface runoff Soil layer
Unaccounted
30
45
20%
15%
20%
15%
20%
15%
20%
0–5 cm 5–10 cm 10–15 cm Sum
29.7 391.8 168.3 95.4 655.5 597.3
36.6 253.3 113.9 82.3 449.6 796.3
48.8 363.9 160.2 79.5 603.6 630.1
62.6 242.5 100.9 67.0 410.5 809.4
2.3 30.5 13.1 7.4 51.1 46.6
2.9 19.8 8.9 6.4 35.1 62.1
3.8 28.4 12.5 6.2 47.1 49.1
4.9 18.9 7.9 5.2 32.0 63.1
0–5 cm 5–10 cm 10–15 cm Sum
13.0 355.6 224.4 129.6 709.6 559.9
20.0 270.5 212.2 140.6 623.2 639.2
23.7 312.9 218.4 114.4 645.8 613.0
39.7 237.6 174.9 135.6 548.0 694.7
1.0 27.7 17.5 10.1 55.3 43.7
1.6 21.1 16.5 11.0 48.6* 49.8*
1.9 24.4 17.0 8.9 50.4 47.8
3.1 18.5 13.6 10.6 42.7* 54.2*
0–5 cm 5–10 cm 10–15 cm Sum
7.2 600.9 261.6 283.5 1146.0 129.3
12.0 488.4 200.4 179.1 867.9 402.6
12.6 344.8 250.5 212.2 807.5 462.4
26.1 299.4 201.3 199.9 700.5 555.9
0.6 46.9 20.4 22.1 89.4* 10.1*
0.9 38.1 15.6 14.0 67.7* 31.4*
1.0 26.9 19.5 16.5 63.0* 36.1*
2.0 23.3 15.7 15.6 54.6* 43.3*
Unaccounted Pennisetum
45
15%
Unaccounted Melilotus
Percentage to total applied atrazine (%)
Note: In the calculation of atrazine mass balance, soil residue in each layer at middle slope position represents the average residue level (the atrazine residue in the soil at different slope position is different). * means significant differences between grass hedge treatments and the control (P < 0.05).
conventional-tillage and strip-tillage fields (Potter et al., 2014). The atrazine loss rate caused by the single rainfall event with an intensity of 45 mm h−1 in our study approached the highest annual loss (4.7%) in the 9-year period investigation of Shipitalo and Owens (2006). They
observed that most of this loss (78%) was also the result of a single runoff event that began 2 d after herbicide application. Their another long-term study further confirmed that yearly loss was only 0.014% of the applied atrazine when the first runoff occurred 51 days after 7
Agriculture, Ecosystems and Environment 253 (2018) 1–10
Q. Wang et al.
Table 3 The contribution of grass hedge, rain intensity and slope gradient to surface runoff, atrazine loss based on optimal scaling regression analysis. Factor
Surface runoff Atrazine loss
Model ANOVA
Grass hedge
Rain intensity
Slope gradient
R2
p
Standardized coefficient
p
Importance
Standardized coefficient
p
Importance
Standardized coefficient
p
Importance
0.992
0.008
0.446
0.001
0.201
0.869
< 0.001
0.761
0.193
0.005
0.038
0.957
< 0.001
0.800
< 0.001
0.669
0.483
0.009
0.244
0.290
0.057
0.088
Fig. 7. Relationships of atrazine loss to overland flow (A) and atrazine concentration in runoff (B).
effectively reduced exported herbicide load, including dissolved-phase and sediment bound transport (Lin et al., 2011; Lafrance et al., 2013). Furthermore, the contour buffers planted in cultivated fields decreased the loss amount of pollutants mainly by inhibiting their mobilization (Dosskey, 2001). Pennisetum exhibited significantly higher effectiveness than Melilotus, particularly at the higher rain intensity. Previous report confirmed that the role of plants in controlling water and soil loss was possibly related to their root characteristics (Baets et al., 2007), and plants with denser stems and richer roots proved to be more effective (Xiao et al., 2010). The difference of efficacy between Melilotus and Pennisetum hedges may be explained by differences in their root systems: Pennisetum has a fibrous root system with more fine roots (diameter < 1 mm), a larger root surface area, and faster root growth (Cheng et al., 2008); while Melilotus has a taproot system with a smaller root surface area and less fine roots in the topsoil layer. The main mechanisms of grass hedges for reducing water and soil loss has been reported: 1) grass roots may increase soil porosity, thus forming water channels for infiltration; 2) dense and stiff stems of grass hedges may catch surface soil particles, thus enabling increased sediment deposition (Gilley et al., 2000; Blanco-Canqui et al., 2004). Consequently, grass hedges control transport of herbicides lost from fields through reducing water and sediment carriers of atrazine. The result of mass balance analysis indicated that grass hedge, especially Pennisetum, intercepted more amount of atrazine to retain in the topsoil and hence held promise for protecting ground water, which may be related to the increase of soil organic matter content facilitating strong sorption of the herbicides in grass hedge system. From these results, we concluded that grass hedges demonstrated their abilities to decrease atrazine export from agricultural lands mainly through reducing the runoff amount, as well as showed their potential for reducing the hazard of atrazine leaching. But the efficiency for decreasing the herbicide loss was limited by the slope gradient. Therefore, although grass hedges action alone maintained a certain protective function, other control measures (e.g. water harvesting, buffer strips on the edges of fields) may be
atrazine application (Shipitalo and Owens, 2003). This data implies that the rainfall intensity had greater effects on the total atrazine loss, and a few extreme rainfall events close to application time were responsible for most of the herbicide loss. These findings further support the conclusion of Potter et al. (2015) that the runoff close to pesticide application time was the determinant source in total pesticide loss. High rain intensity increased total atrazine runoff loss, but the mean atrazine concentrations in runoff were higher under lower rain intensity. This was mainly attributed to enough runoff volume provided by the higher rainfall intensity to reduce the concentration (Pote et al., 1994). As found in the present study, under 45 mm h−1 the runoff was nearly 5 times higher than under 30 mm h−1, while the former caused less than 2 times as much atrazine loss as the latter. The significant variation of atrazine residue at 0–5 cm soil depth under different rain intensities and slope gradients suggested that this soil layer was the key soil segment for herbicide loss after rainfall events. The deeper soil layer of 10–15 cm contained significantly lower atrazine with intensive rainfall. A study by Ouyang et al. (2016) reported an opposite observation that the herbicide leaching load increased under the highest rain intensity. The difference may have resulted from no surface runoff being generation in their soil column study. These results imply that the rain intensity strongly changed physical redistribution of residues by affecting the occurrence and volume of runoff. Therefore, in farming practice, soil application of atrazine should be likely to avoid favorable conditions for surface runoff formation, such as saturated soil and heavy rainfall.
4.2. Effects of grass hedges on atrazine loss Both Melilotus and Pennisetum grass hedges significantly reduce surface runoff and consequent atrazine loss, and their impacts on the control of atrazine loss is more important than rain intensity and slope gradient. As reported in other pesticide runoff management studies under simulated or natural rainfall conditions, grass filter strips 8
Agriculture, Ecosystems and Environment 253 (2018) 1–10
Q. Wang et al.
Baets, S., Poesen, J., Knapen, A., Barberá, G.G., Navarro, J.A., 2007. Root characteristics of representative Mediterranean plant species and their erosion-reducing potential during concentrated runoff. Plant Soil 294, 169–183. Baker, J.L., Mickelson, S.K., 1994. Application technology and best management practices for minimizing herbicide runoff. Weed Technol. 8, 862–869. Blanco-Canqui, H., Gantzer, C.J., Anderson, S.H., Alberts, E.E., 2004. Grass barriers for reduced concentrated flow induced soil and nutrient loss. Soil Sci. Soc. Am. J. 68, 1963–1972. Bouldin, J.L., Farris, J.L., Moore, M.T., Smith, S., 2006. Hydroponic uptake of atrazine and lambda-cyhalothrin in Juncus effusus and Ludwigia peploides. Chemosphere 65, 1049–1057. Bradley, P.M., Journey, C.A., Romanok, K.M., Barber, L.B., Buxton, H.T., Foreman, W.T., Furlong, E.T., Glassmeyer, S.T., Hladik, M.L., Iwanowicz, L.R., Jones, D.K., Kolpin, D.W., Kuivila, K.M., Loftin, K.A., Mills, M.A., Meyer, M.T., Orlando, J.L., Reilly, T.J., Smalling, K.L., Villeneuve, D.L., 2017. Expanded target-chemical analysis reveals extensive mixed-organic-contaminant exposure in U.S. streams. Environ. Sci. Technol. 51, 4792–4802. Cheng, X., Chen, W., Gu, B., Liu, X., Chen, F., Chen, Z., Zhou, X., Li, Y., Huang, H., Chen, Y., 2008. Morphology, ecology, and contaminant removal efficiency of eight wetland plants with differing root systems. Hydrobiologia 623, 77–85. Devlin, D.L., Regehr, D.L., Barnes, P.L., 2000. Managing to minimize atrazine runoff. Kansas State University Agricultural Experiment Station and Cooperative Extension Service MF2208. Kansas State University, Manhattan, KS. Dosskey, G.M., 2001. Toward quantifying water pollution abatement in response to installing buffers on crop land. Environ. Manage. 28, 577–598. Gilley, J.E., Eghball, B., Kramer, L.A., Mooreman, T.B., 2000. Narrow grass hedge effects on runoff and soil loss. J. Soil Water Conserv. 55, 190–196. Gong, D., Shi, P., Wang, J., 2004. Daily precipitation changes in the semi-arid region over northern China. J. Arid Environ. 59, 771–784. Humphry, J.B., Daniel, T.C., Edwards, D.R., Sharpley, A.N., 2002. A portable rainfall simulator for plot-scale runoff studies. Appl. Eng. Agric. 18, 199–204. Hussain, S., Siddique, T., Arshad, M., Saleem, M., 2009. Bioremediation and phytoremediation of pesticides: recent advances. Crit. Rev. Env. Sci. Tec. 39, 843–907. Jablonowski, N.D., Ffer, A.S., Burauel, P., 2011. Still present after all these years: persistence plus potential toxicity raise questions about the use of atrazine. Environ. Sci. Pollut. Res. 18, 328–331. Juergens, M.D., Chaemfa, C., Hughes, D., Johnson, A.C., Jones, K.C., 2015. PCB and organochlorine pesticide burden in eels in the lower Thames River (UK). Chemosphere 118, 103–111. Lafrance, P., Caron, E., Bernard, C., 2013. Impact of grass filter strips length on exported dissolved masses of metolachlor, atrazine and deethylatrazine: a four-season study under natural rain conditions. Soil Use Manage. 29, 87–97. Lin, C., Lerch, R.N., Goyne, K.W., Garrett, H.E., 2011. Reducing herbicides and veterinary antibiotics losses from agroecosystems using vegetative buffers. J. Environ. Qual. 40, 791–799. Mao, M., Ren, L., 2004. Simulating nonequilibrium transport of atrazine through saturated soil. Ground Water 42, 500–508. Mei, M., Du, Z., Cen, Y., 2011. QuEChERS-ultra-performance Liquid Chromatography Tandem Mass Spectrometry for Determination of Five Currently Used Herbicides. Chinese J. Anal. Chem. 39, 1659–1664. Milan, M., Vidotto, F., Piano, S., Negre, M., Ferrero, A., 2013. Buffer strip effect on terbuthylazine, desethyl-terbuthylazine and S-metolachlor runoff from maize fields in Northern Italy. Environ. Technol. 34, 71–80. Ouyang, W., Huang, W., Wei, P., Hao, F., Yu, Y., 2016. Optimization of typical diffuse herbicide pollution control by soil amendment configurations under four levels of rainfall intensities. J. Environ. Manage. 175, 1–8. Piepho, H.P., Büchse, A., Emrich, K., 2003. A Hitchhiker's Guide to Mixed Models for Randomized Experiments. J. Agron. Crop Sci. 189, 310–322. Pote, D.H., Daniel, T.C., Edwards, D.R., Mattice, J.D., Wickliff, D.B., 1994. Effect of drying and rainfall intensity on cyromazine loss from surface-applied caged-layer manure. J. Environ. Qual. 23, 101–104. Potter, T.L., Bosch, D.D., Strickland, T.C., 2014. Comparative assessment of herbicide and fungicide runoff risk: A case study for peanut production in the Southern Atlantic Coastal Plain (USA). Sci. Total Environ. 490, 1–10. Potter, T.L., Bosch, D.D., Strickland, T.C., 2015. Tillage impact on herbicide loss by surface runoff and lateral subsurface flow. Sci. Total Environ. 530–531, 357–366. Reichenberger, S., Bach, M., Skitschak, A., Frede, H., 2007. Mitigation strategies to reduce pesticide inputs into ground- and surface water and their effectiveness: a review. Sci. Total Environ. 384, 1–35. Richter, C.A., Papoulias, D.M., Whyte, J.J., Tillitt, D.E., 2016. Evaluation of potential mechanisms of atrazine-induced reproductive impairment in fathead minnow (Pimephales promelas) and Japanese medaka (Oryzias latipes). Environ. Toxicol. Chem. 35, 2230–2238. Ries, J.B., Seeger, M., Iserloh, T., Wistorf, S., Fister, W., 2009. Calibration of simulated rainfall characteristics for the study of soil erosion on agricultural land. Soil Till. Res. 106, 109–116. Rippy, M.A., Deletic, A., Black, J., Aryal, R., Lampard, J., Tang, J.Y., McCarthy, D., Kolotelo, P., Sidhu, J., Gernjak, W., 2017. Pesticide occurrence and spatio-temporal variability in urban run-off across Australia. Water Res. 115, 245–255. Schwarzenbach, R.P., Egli, T., Hofstetter, T.B., von Gunten, U., Wehrli, B., 2010. Global water pollution and human health. Annu. Rev. Environ. Resour. 35, 109–136. Sharpley, A., Kleinman, P., 2003. Effect of Rainfall simulator and plot scale on overland flow and phosphorus transport. J. Environ. Qual. 32, 2172–2179. Shipitalo, M.J., Owens, L.B., 2003. Atrazine, deethylatrazine, and deisopropylatrazine in surface runoff from conservation tilled watersheds. Environ. Sci. Technol. 37, 944–950.
needed to achieve sufficient protection under the steep slope conditions (particularly in frequent intense rain regions). The present finding may be helpful in controlling pesticide losses from sloping cropland. 4.3. Limitation and prospective In present study, the loss of atrazine adsorbed on soil particulates in runoff was not taken into consideration, which somewhat influenced accurate evaluation of atrazine loss. But the underestimate is unlikely to be marked, because only 1.0-2.1 g sediment was obtained after being filtered and dried from 1 L total runoff. Research typically found that most of the atrazine transported was associated with the water fraction because of the greater amounts of water lost as compared with soil, the total amounts lost in water were about 7 and 10 times greater than in the soil (White et al., 1967). Further, Devlin et al. (2000) confirmed that more than 90 percent of atrazine loss occurred in runoff water. Our study focused on atrazine due to its worldwide use with a long application history, relative long-term persistence, and potential threat to human and ecosystem health. So the findings are limited to the loss of pesticides with similar properties as atrazine. Pesticide loss problems are usually associated with natural conditions of weather. Though simulated rainfall has been widely used to conduct research on runoff generation and soil erosion on agricultural land, there are performance limitations due to its inability to simulate all aspects of natural storms, particularly in terms of drop fall velocity and rainfall distribution on the plot (Humphry et al., 2002; Ries et al., 2009). Therefore, under natural rainfall conditions, more comprehensive studies are needed to better estimate the loss of pesticides with widely different properties within the grass hedge system. 5. Conclusions The current plot study evaluated the effectiveness of grass hedges in reducing surface runoff and accompanying atrazine loss, and compared the importance of the grass hedge, rainfall intensity and slope gradient to atrazine loss. The main conclusions are as follows: (1) Atrazine loss in a single rainfall induced surface runoff was less than 5% of applied amount under the experimental conditions. The intense rainfall and steep slope increased atrazine loss. (2) Both Pennisetum and Melilotus hedges were efficient in controlling atrazine loss in Northern China. Pennisetum provided more effective protection, and can be preferentially selected when designing grass hedges in similar climatic regions. Under heavy rainfall and steep slope conditions, other engineering measures incorporating with grass hedges should be devised for sufficient reducing pesticide runoff. (3) The grass hedge and rain intensity were significant factor for atrazine loss, and had more important effects than slope gradient. Further researches are necessary to fully evaluate the efficiency of grass hedges in reducing pesticide runoff under natural rainfall and broader scales conditions, and a comprehensive analysis of protective mechanisms of grass hedges against runoff losses of atrazine are needed. Acknowledgments The research was supported by the National Natural Science Foundation (31370540) and National Key Technologies R & D Program (2012BAD14B02) of China. References Atwood, D., Paisley-Jones, C., 2017. Pesticides industry sales and usage: 2008-2012 market estimates. US Environmental Protection Agency, Washington, DC.
9
Agriculture, Ecosystems and Environment 253 (2018) 1–10
Q. Wang et al.
Wille, K., Kiebooms, J.A.L., Claessens, M., Rappé, K., Vanden Bussche, J., Noppe, H., Van Praet, N., De Wulf, E., Van Caeter, P., Janssen, C.R., De Brabander, H.F., Vanhaecke, L., 2011. Development of analytical strategies using U-HPLC-MS/MS and LC-ToF-MS for the quantification of micropollutants in marine organisms. Anal. Bioanal. Chem. 400, 1459–1472. Wu, J.Y., Huang, D., Teng, W.J., Sardo, V.I., 2010. Grass hedges to reduce overland flow and soil erosion. Agron. Sustain. Dev. 30, 481–485. Xiao, B., Wang, Q., Wang, H., Dai, Q., Wu, J., 2011. The effects of narrow grass hedges on soil and water loss on sloping lands with alfalfa (Medicago sativa L.) in Northern China. Geoderma 167–168, 91–102. Xiao, B., Wang, Q., Wu, J., Huang, C., Yu, D., 2010. Protective function of narrow grass hedges on soil and water loss on sloping croplands in Northern China. Agr. Ecosyst. Environ. 139, 653–664. Young, F.W., de Leeuw, J., Takane, Y., 1976. Regression with qualitative and quantitative variables: An alternating least squares method with optimal scaling features. Psychometrika 41, 505–529. Zhong, Y., Jia, Y., Li, Z., 2013. Spatial and temporal changes of maximum 1 h precipitation intensity in Beijing region in last 53 years. J. China Hydrol. 33, 32–37 (in Chinese with English abstract).
Shipitalo, M.J., Owens, L.B., 2006. Tillage system, application rate, and extreme event effects on herbicide losses in surface runoff. J. Environ. Qual. 35, 2186–2194. Stehle, S., Schulz, R., 2015. Pesticide authorization in the EU-environment unprotected? Environ. Sci. Pollut. Res. 22, 19632–19647. Stone, W.W., Gilliom, R.J., Ryberg, K.R., 2014. Pesticides in US streams and rivers: occurrence and trends during 1992-2011. Environ. Sci. Technol. 48, 11025–11030. Vianello, M., Vischetti, C., Scarponi, L., Zanin, G., 2005. Herbicide losses in runoff events from a field with a low slope: role of a vegetative filter strip. Chemosphere 61, 717–725. Vieira, D.A.N., Dabney, S.M., 2012. Two-dimensional flow patterns near contour grass hedges. Hydrol. Process. 26, 2225–2234. Vymazal, J., Březinová, T., 2015. The use of constructed wetlands for removal of pesticides from agricultural runoff and drainage: a review. Environ. Int. 75, 11–20. Walter, E.S., Lyle, A.D., 1962. Effects of environment upon herbicides applied preemergence. Weeds 10, 304–307. Wei, W., Chen, L., Fu, B., Huang, Z., Wu, D., Gui, L., 2007. The effect of land uses and rainfall regimes on runoff and soil erosion in the semi-arid loess hilly area, China. J. Hydrol. 335, 247–258. White, A.W., Barnett, A.P., Wright, B.G., Holladay, J.H., 1967. Atrazine losses from fallow land caused by runoff and erosion. Environ. Sci. Technol. 1, 740–744.
10