Chemosphere 61 (2005) 717–725 www.elsevier.com/locate/chemosphere
Herbicide losses in runoff events from a field with a low slope: Role of a vegetative filter strip Monica Vianello a,*, Costantino Vischetti b, Luciano Scarponi c, Giuseppe Zanin d a
Dipartimento di Agronomia Ambientale e Produzioni Vegetali, Universita` di Padova, Agripolis, Viale dell’Universita´ 16, 35020 Legnaro (PD), Italy b Dipartimento di Scienze Ambientali e delle Produzioni Vegetali, Universita` Politecnica delle Marche, via Brecce Bianche, 60131 Ancona, Italy c Dipartimento di Scienze Agroambientali e Produzione Vegetale, Universita` degli Studi, Borgo XX Giugno 72, 06121 Perugia, Italy d CNR—Istituto di Biologia Agroambientale e Forestale, Sezione di Legnaro—Malerbologia, Agripolis, Viale dell’Universita` 16, 35020 Legnaro (PD), Italy Received 26 July 2004; received in revised form 12 March 2005; accepted 19 March 2005 Available online 28 April 2005
Abstract Herbicide runoff and the effects of a narrow vegetative filter strip (VFS) were studied on an arable field in the lowlying plains of the Veneto Region (north-east Italy). Cultivated plots were compared with and without a 6 m wide VFS composed of trees, shrubs and grass. Natural and simulated runoff were monitored during 2000 and 2001. Herbicides applied on the field were: metolachlor (2184–2254 g ha1), terbuthylazine (1000–1127 g ha1) and isoproturon (1000 g ha1). The VFS reduced both runoff depth (10.2–91.2%) and herbicide losses (85.7–97.9%) in the monitored rainfall events. Total herbicide loss with runoff was low (0.69–3.98 g ha1 without VFS, less than 0.27 g ha1 with VFS), but concentrations were sometimes very high, especially of terbuthylazine and isoproturon during the first events after treatment. In these events there was a high probability of exceeding the ecotoxicological endpoint for algae, but the VFS helped to reduce the potential risk. Two VFS effectiveness mechanisms were identified: (i) dilution, and (ii) a ‘‘sponge-like’’ effect, which temporarily trapped chemicals inside the VFS before releasing them. 2005 Elsevier Ltd. All rights reserved. Keywords: Agricultural non-point pollution; Ecotoxicology; Metolachlor; Terbuthylazine; Isoproturon
1. Introduction Relatively small herbicide loads are carried by surface runoff water in relation to the amount applied to a
* Corresponding author. Tel.: +39 0498272822; fax: +39 0498272839. E-mail address:
[email protected] (M. Vianello).
cropped field (from less than 0.5% up to 5%) (Wauchope, 1978), yet they can be a potential environmental risk. The main pathway for herbicide losses is surface runoff, and a rainstorm shortly after application can determine high chemical concentrations in the runoff (Brown et al., 1995; Ng and Clegg, 1997), which can lead to serious consequences for water quality and wildlife habitats. A vegetative filter strip (VFS) is proposed as a means to reduce surface water contamination caused by
0045-6535/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2005.03.043
718
M. Vianello et al. / Chemosphere 61 (2005) 717–725
agricultural non-point pollution (Gilliam, 1994; Daniels and Gilliam, 1996; Lee et al., 2000; Rankins and Shaw, 2001). VFS acts as a natural dam or terrace and, by reducing runoff, the water has more time to penetrate and incorporate the pollutants in the soil and thus prevent off-site movement (Webster and Shaw, 1996). VFS alters flow hydraulics, reducing runoff speed and increasing water infiltration (Misra et al., 1996). The filter thus enhances sediment deposition and filtration by vegetation, pollutant adsorption onto the soil and dead and living plant materials, and uptake of soluble pollutants by plants (Misra et al., 1996; Blanche et al., 2003). Infiltration was found to be the most important herbicide removal mechanism associated with VFS, especially for soluble or weakly adsorbed pesticides (Klo¨ppel et al., 1997). Watanabe and Grismer (2001), investigating diazinon transport within a VFS, found that the pesticide was trapped on its surface and in the root-zone, where further adsorption, attenuation and presumably degradation may occur. Nevertheless, enhanced infiltration could cause more leaching and the herbicide reach the water table, changing the ecotoxicological impact from surface to subsurface water. Delphin and Chapot (2001) evaluated the leaching and investigated the fate of atrazine and de-ethylatrazine transported in runoff effluents and trapped by a grass filter strip. The plants in the VFS confer a higher organic matter content to the filter zone than in the adjacent cultivated field. This organic matter accumulation should increase adsorption capacity and microbial activity for herbicide degradation, so reducing the amount of herbicide in surface runoff and leaching (Staddon et al., 2001). Higher herbicide dissipation in the VFS soil is due both to enhanced degradation and the formation of non-extractable (bound) residues, which can become a long-term store inside the filter (Benoit et al., 2001). The north-east of Italy is mainly formed by the alluvial Po Valley, a flat, intensively-farmed plain. The most common arable crops are maize, winter wheat, soybean and sugarbeet. In these conditions, narrow VFS have demonstrated their effectiveness in reducing sediment, N and P runoff (Borin et al., 2005), and in a preliminary study (Borin et al., 2004), also in reducing pesticides in subsurface water coming from cropland. The objectives of this study were to assess: (i) the importance of herbicide runoff in low slope conditions; (ii) the effects of a narrow VFS on herbicide loss in surface runoff from an arable field.
ley, north-east Italy (4512 0 N, 1158 0 E, altitude 6 m a.s.l.). The soil is classified as Fulvi–Calcaric Cambisol (FAO-UNESCO, 1990). It is silty–loam textured (11.8% clay, 44.9% silt, 43.3% sand), rich in limestone, with sub-basic pH (pH = 8.11), good organic carbon content (0.92%) and medium–low hydraulic conductivity (4.7 · 104 cm s1). According to the De Martonne classification, the climate in the area is sub-humid (De Martonne, 1926): annual rainfall is about 805 mm, mainly during spring and autumn. Weather data were collected at the weather station on the farm. The experimental site consisted of 20 m · 35 m plots with a 1.8% slope downwards to a ditch. Two treatments were compared: plots cultivated to the edge of the ditch and plots with a 6 m wide VFS between the cropland and ditch. Each treatment (with or without VFS) had two replicates. The VFS was composed of grass and two shrub-tree rows. The grass was Festuca arundinacea Schreber, with regularly alternating shrubs Viburnum opulus L. and trees Platanus hybrida Brot in the rows. The two rows were 1.5 m and 4.5 m from the ditch, planted 2 m apart in the centre of 1.2 m wide ethylene–vinyl acetate (EVA) film. The grass was sown (at a density of 30 kg ha1), and shrubs and trees planted in 1997. The grass was cut at least twice a year during the growing season. In 2000, plots were cropped with maize and two herbicide treatments applied: metolachlor and terbuthylazine at pre-emergence and terbuthylazine at postemergence. After maize harvesting, the plots were cropped with winter wheat and sprayed with isoproturon, and then (summer 2001) cropped with soybean and sprayed with metolachlor. The soybean was sodseeded on the winter wheat stubble. Herbicides were applied using a tractor-mounted Hardy LY–HY sprayer equipped with a 12 m boom and 4110-16 fan nozzles. Standard agronomic practices were followed (Table 1). Weather conditions were optimal during herbicide treatments, assuring good distribution uniformity. The physico-chemical properties and ecotoxicological endpoints of the herbicides are reported in Table 2. Simulated rainfalls using a sprinkler irrigation system were applied to all plots 1 day after the second terbuthylazine treatment on 25 May 2000 (52 mm rainfall at 43 mm h1 intensity) and 28 days after the metolachlor application on 31 July 2001 (87 mm rainfall at 82 mm h1 intensity).
2. Materials and methods
2.2. Runoff monitoring and water samples collection
2.1. Experimental site
A collector system with multi-pipe divisor (Morari et al., 2001) was designed and built to measure runoff volumes and collect water samples. There was a sampler for each plot, located in the ditch. The system allowed
The study was done during 2000 and 2001 at the Padova University Experimental Farm in the Po Val-
M. Vianello et al. / Chemosphere 61 (2005) 717–725
719
Table 1 Crops, cultural practices and weather conditions during herbicide applications (treatment length: 30 min) Crop
Maize
Winter wheat
Soybean
Ploughing Sowing Harvest
5/10/99 27/4/00 25/9/00
4/10/00 25/10/00 25/6/01
Sod seeding 3/7/01 8/11/01
Fertiliser treatment Date 27/04/00 Nutrient N Dose (kg/ha) 32
27/04/00 P 20
24/5/00 N 150
Herbicide treatment Date 28/4/00 a.i. M Dose (g/ha) 2254
28/4/00 T 1127
24/5/00 T 1000
25/10/00 N 32
25/10/00 P 20
19/2/01 N 100
3/7/01 N 16
27/10/00 I 1000
3/7/01 M 2184
Weather conditions during herbicide treatment Wind speed (m/s) Mean 0.7 1.5 Min 0.2 0.0 Max 1.2 4.0
2.1 0.8 3.5
1.5 0.0 4.0
Air humidity (%) Mean 66.3 Min 61.0 Max 75.0
71.9 46.0 100.0
85.2 75.0 98.0
64.0 42.0 94.0
Temperature (C) Mean 21.5 Min 20.6 Max 22.5
18.5 10.0 24.7
15.0 13.0 16.8
23.7 18.1 29.2
Sun radiation (W h/m2) Mean 492.1 Min 209.3 Max 676.8
311.1 0.0 892.6
236.1 62.8 425.6
301.7 0.0 1018.1
3/7/01 P 10
M = metolachlor, T = terbuthylazine and I = isoproturon.
samples to be collected and measured even if runoff volume was very low, as volumes ranging from less than 1.0 l to 2352 l could be collected. The latter corresponds to a maximum runoff depth of 50 mm, which is an extreme event with a return time of 20 years. Water volume was measured after each runoff event and samples collected for determining herbicide concentration. Runoff volume and herbicide concentration data were used to compute the mass transport at the end of the VFS and cropland. Samples were stored until analysis in the dark at 4 C, for a maximum of 8 months. 2.3. Soil sampling After spraying, soil samples were taken to assess initial herbicide concentration in the cultivated soil (soil concentration at t0 time), then at increasing times from treatment to evaluate dissipation kinetics. The field half life (DT50) was calculated for all herbicides using first order kinetics.
A 2.5 cm diameter soil core was used to collect soil samples, which were stored in the dark at 4 C until chemical analysis. At each sampling time, 20 cores were collected and the analyses performed within 8 months. 2.4. Herbicide extraction and analysis To determine herbicide concentrations in the soil, samples of 50 g were mechanically shaken for 1 h with 100 ml of a 1:1 methanol:water mixture. After shaking, the suspensions were centrifuged at 6000g for 15 min and the supernatants filtered through extra-rapid filter paper. The procedure was then repeated. The filtrates were collected and extracted twice with chloroform (2 · 40 ml), evaporated in a rotary evaporator at 40 C and re-dissolved in hexane (1 ml). To determine herbicide concentration in water, water samples (1.5 l in volume) were enriched and clean-up by SPE (solid phase extraction). Water was filtered through filter paper to remove suspended solids. For SPE, a C-18 (EC) 500 mg/6 ml (Isolute) cartridge from IST, and a
720
M. Vianello et al. / Chemosphere 61 (2005) 717–725
Table 2 Main physico-chemical parameter (Nicholls, 1994), experimental DT50 in study site soil, R2 of first-order kinetics, and ecotoxicological endpoints of the herbicides
Physico-chemical parameters Molecular weight Water solubility (mg/l) Vapor pressure (Pa) Koc range (l/kg) DT50 (days) Experimental DT50 (days) R2 Ecotoxicological endpoints Fish Bluegill sunfish LC50 (96 h) (mg/l) Rainbow trout LC50 (96 h) (mg/l) Carp LC50 (96 h) (mg/l) Algae Scenedesmus subspicatus EC50 (72 h)b (mg/l) Scenedesmus vacuolatus EC50 (72 h)b (mg/l) NOEC (mg/l)
Metolachlor
Terbuthylazine
Isoproturon
284 530 1.70E03 90–130 12–130 16.1(2000) 12.1(2001) 0.92**(2000) 0.90**(2001)
230 8.5 1.50E04 400–800 30–100
206 55 3.30E06 130–180 6–62
26.9a
30.7
0.78**
0.65**
10.0+ 3.9+ 4.9+
7.50+ 3.8–4.6+ 7.0+
>100+ 37.0+ 193.0+
0.016++ 0.0022++
0.047+++ 0.0048+++
0.1+
a: Terbuthylazine DT50 was calculated after the second treatment. b: EC50 (growth inhibition); NOEC = no observed effect concentration; ** = statistical significance at p < 0.01; + = after Tomlin, 2000; ++ = after Faust et al., 2001; +++ = after Backhaus et al. (2004).
Baker spe-12 G vacuum column processor were used. The cartridge was preconditioned with 5 ml of methanol and washed twice with 5 ml of distilled water. The water sample was then flowed through the cartridge at a rate of approximately 10–15 ml h1 under vacuum. After solid-phase extraction, the cartridge was aspirated with dry air to remove residual water. The cartridge was eluted with 5 ml of methanol to extract the herbicides adsorbed at the stationary phase. The extract was then diluted with 50 ml of distilled water, extracted twice with chloroform (2 · 40 ml), evaporated in a rotary evaporator at 40 C and re-dissolved in hexane (1 ml). Gas chromatographic and HPLC techniques were used to detect herbicide residues in the samples. Analytical-grade metolachlor (96.1% purity), terbuthylazine (99.9% purity) and isoproturon (99.5% purity) (Dr. Ehrenstorfer GmbH, D-86199 Augsburg) were used as analytical standards. Metolachlor and terbuthylazine analyses were performed by gas chromatography, using an HRGC MEGA 2 series gas chromatograph equipped with an NPD detector and SE-54 fused silica, 30.0 m · 0.32 mm i.d., 0.25 lm df. The injector was an on-column type. A 1.0 ll sample volume was manually injected each time. The initial column oven temperature of 50 C was held for 1 min and then ramped at 25 C min1 to the final temperature of 260 C, which was maintained for 10 min. Helium was used as carrier (1.5 ml min1) and make-up gas (35 ml min1). Isoproturon analyses were performed by HPLC, using a Perkin–Elmer Series 410 chromatograph,
equipped with a C18 inertsil column, polymer-based, 25 cm · 4.6 mm i.d., 5 lm particle size, a UV detector at 230 nm, a mobile phase CH3CN/H2O 70/30 and flow rate of 0.7 ml min1. Retention times were 14.1 min, 12.2 min and 10.5 min, and method sensitivity was 2 lg kg1, 1 lg kg1 and 8 lg kg1, for metolachlor, terbuthylazine and isoproturon, respectively. Pesticide concentrations in soil and water samples were quantified by comparison with a pesticide external standard. Recoveries of the three pesticides from extracted soil and water samples were performed in triplicate at three initial concentrations (10, 100 and 1000 lg kg1 for soil samples and 10, 100 and 1000 lg l1 for water samples). They were always satisfactory, varying from 80.5% (terbuthylazine at 10 lg kg1) to 94.4% (isoproturon at 1000 lg kg1) for soil samples, and from 95.7% (terbuthylazine at 10 lg l1) to 99.8% (metolachlor at 1000 lg l1) for water samples. The detection limit of the method, determined at a signal to noise ratio of two, was 5 ng for terbuthylazine and 8 ng for metolachlor (gas chromatography), and 10 ng for isoproturon (HPLC). 2.5. Statistical analysis Data were not normally distributed, so non-parametric procedures were used. Mann–Whitney U-test at 0.05 a level was performed to assess VFS effectiveness on
M. Vianello et al. / Chemosphere 61 (2005) 717–725
without VFS, 0.58 mm with VFS), but in four minor events it appeared to have no effect. In 2001 the first metolachlor runoff event was 17 DAT (Table 4), and only four events occurred after herbicide application in soybean. VFS was not as effective as for the other herbicides: runoff volume was reduced by 10.2% (11.3 mm without VFS, 10.2 mm with VFS), and did not reduce the volume at the first event (5.88 mm without VFS, and 9.17 mm with VFS). Soybean was sod-seeded so it can be assumed that the cracks created during sowing, together with the wheat residues, were more effective than the VFS in limiting surface runoff and increasing infiltration when there was a dramatic rainfall event (60 mm). The U-test highlighted significant reductions in runoff depth for terbuthylazine in 2000, while for metolachlor the p-level was very close to being statistically significant (p = 0.075). No significant reduction was observed in 2001. This is probably related to the dramatic loss of grass cover beneath the VFS trees in 2001: only about 45% of the ground was grass covered. This was caused by both increased tree shading and an accumulation of plant litter.
water quality and amount of total losses. For running the U-test, n.d. herbicide concentration values were recorded as 0. The values presented are the means of two data. STATISTICA, version 6.1, (StatSoft, Inc., 2004) was used for the statistical analysis.
3. Results and discussion 3.1. Runoff volume In both 2000 and 2001, annual rainfall was much lower than the long-term average, being 686.2 mm and 664.3 mm, respectively, compared to the annual mean of 805 mm. The wettest periods were autumn 2000 and spring 2001 (Fig. 1), with maximum rainfall intensity during the natural runoff events ranging from 2.6 to 30.2 mm h1. Runoff events and volume reductions during 2000 are reported in Table 3: the first monitored event was a simulated rainfall (52 mm) at 26 days after treatment (DAT) for metolachlor and 1 DAT for the second treatment with terbuthylazine. VFS reduced runoff by 65.6% and this behaviour was common in both major and minor events for metolachlor and terbuthylazine. The highest runoff depth was caused by simulated rainfall: 12.2 mm without VFS and 4.2 mm with VFS, which represented 90% and 95%, respectively, of the total amount measured during the crop cycle. The first event after isoproturon treatment was 6 DAT (Table 4). Unlike the other herbicides, isoproturon was distributed in autumn, i.e. the rainiest period of the year and thus was monitored in more events, even if runoff volume was not as high as for the other herbicides. VFS reduced total runoff depth by 91.2% (6.40 mm
M+T T
rainfall (mm)
3.2. Herbicide concentrations and loads In 2000, metolachlor and terbuthylazine losses were reduced by 85.7% and 91.9%, respectively; concentration reductions for the two herbicides ranged from 50% to 56%, and 84.9% to 100%, respectively (Table 3). In 2001, isoproturon and metolachlor losses were reduced by 97.9% and 93.0% respectively, with concentration reductions for the two herbicides of 18.7–80.2%, and 319–100%, respectively (Table 4). Statistical analysis highlighted significant differences only in isoproturon
M
I Ma
90 80
721
S
W
70 60 50 40 30 20 10 01/12/2001
01/11/2001
01/10/2001
01/09/2001
01/08/2001
01/07/2001
01/06/2001
01/05/2001
01/04/2001
01/03/2001
01/02/2001
01/01/2001
01/12/2000
01/11/2000
01/10/2000
01/09/2000
01/08/2000
01/07/2000
01/06/2000
01/05/2000
01/04/2000
01/03/2000
01/02/2000
01/01/2000
0
Fig. 1. Rainfall pattern during 2000 and 2001, crop presence (delimited area; Ma = maize, W = wheat, S = soybean) and herbicide treatment (arrow; M = metolachlor, T = terbuthylazine; I = isoproturon) in the field. Discontinuous lines are simulated rainfall.
722
M. Vianello et al. / Chemosphere 61 (2005) 717–725
Table 3 Rainfall, runoff depth, herbicide concentrations and loss in plots with and without VSF and reduction during 2000 Descriptive statistics Event (no.)
1
2
3
4
5
26
76
101
130
133
52.0S 43.0
33.6 18.6
22.8 4.2
18.2 11.2
10.2 7.6
136.8
0.862 0.083 90.4
0.022 0.003 86.4
0.340 0.039 88.5
0.209 0.032 84.7
13.633 4.357
Herbicide concentration (lg/l) No VFS 6.20 1.20 VFS 2.70 0.60 Reduction (%) 56.4 50.0
n.d. n.d. n.d.
n.d. n.d. n.d.
n.d. n.d. n.d.
Herbicide loss (g/ha) No VFS 0.67 VFS 0.12 Reduction (%) 82.1
0.02 0.0006 97.0
0.00 0.00 0.0
0.00 0.00 0.0
0.00 0.00 0.0
1
50
75
104
107
131
52.0S 43.0
33.6 18.6
22.8 4.2
18.2 11.2
10.2 7.6
33.2 5.8
170.0
0.862 0.083 90.3
0.022 0.003 86.4
0.340 0.039 88.5
0.209 0.032 84.7
0.138 0.102 26.1
13.771 4.459
Herbicide concentration (lg/l) No VFS 25.10 8.90 VFS 3.80 0.40 Reduction (%) 84.9 95.5
1.00 n.d. 100.0
0.60 n.d. 100.0
n.d. n.d. 0.0
0.40 n.d. 100.0
Herbicide loss (g/ha) No VFS 2.17 VFS 0.18 Reduction (%) 91.7
0.0002 0.00 100.0
0.002 0.00 100.0
0.00 0.00 0.0
0.0007 0.00 100.0
Metolachlor DAT Rainfall mm mm/ha
Runoff depth (mm) No VFS 12.200 VFS 4.200 Reduction (%) 65.6
Terbuthylazineb DAT Rainfall mm mm/ha
Runoff depth (mm) No VFS 12.200 VFS 4.200 Reduction (%) 65.6
0.05 0.0004 99.2
6
Sum
0.69 0.12
2.22 0.18
U-test results
Mean
S.D.
Median
Valid number
U
p-level
2.727 0.871 68.1
5.305 1.861
0.340 0.039 88.5
10 10
26.5
0.08
1.48 0.66
2.69 1.17
0.00 0.00
10 10
48.5
0.89
0.14 0.02 85.7
0.30 0.05
0.00 0.00 0.0
10 10
47.5
0.82
2.295 0.743 67.6
4.861 1.694
0.275 0.061 77.8
12 12
35.5
0.04
6.00 0.70
9.96 1.66
0.80 0.00
11 9
33.0
0.19
0.37 0.03 91.9
0.88 0.07
0.00 0.00 100.0
12 9
42.5
0.39
Statistical results and parameters at U-test analysis are also indicated. DAT is days after treatment; S is simulate rainfall; a maximum rainfall intensity during the event; b is DAT counted after second treatment.
concentration and loads: this was probably due to the small dataset and high frequency of the zero value, but is nonetheless an important result. These results are in agreement with those reported by other authors (Webster and Shaw, 1996; Klo¨ppel et al., 1997; Rankins and Shaw, 2001). For a correct evaluation of the influence of different mechanisms and phenomena in the VFS on the reduction of pesticide concentrations in runoff water, the physico-chemical characteristics of the pesticides should also be taken into account (Table 2). Terbuthylazine is the least soluble and most adsorbed compound, isoproturon is the opposite. DT50 is also important because fast degradation implies less herbi-
cide in runoff water. The data in Table 2 suggest that terbuthylazine could also migrate in surface water with sediments, while for metolachlor and isoproturon solubilisation in runoff water could be the main transport mechanism to surface water. Calculated DT50 (Table 2) agree with values reported in the literature, and confirm the short persistence of metolachlor. Terbuthylazine DT50 was calculated from the second treatment, considering the residual amount of the first, so the high summer temperature could explain the shorter persistence of this compound. Suppositions on herbicides behaviour are confirmed by the data in Tables 3 and 4: the reductions in concentrations of terbuthylazine
Table 4 Rainfall, runoff depth, herbicide concentrations and loss in plots with and without VSF and removal, during 2001 Descriptive statistics Event (no.)
2
3
4
5
6
7
8
6
24
89
137
160
162
174
204
58.0 9.6
23.0 3.2
19.2 3.8
17.6 3.4
12.4 2.6
14.4 5.6
22.6 5.6
27.0 4.2
194.2
Runoff depth (mm) No VFS 0.20 VFS 0.08 Reduction (%) 60.0
0.10 0.10 0.0
0.03 0.06 100.0
3.32 0.09 97.3
n.d. 0.03 100.0
274 0.04 98.5
n.d. 0.08 100.0
0.01 0.10 900.0
6.40 0.58
Herbicide concentration (lg/l) No VFS 48.4 66.7 VFS 17.6 13.2 Reduction (%) 63.6 80.2
35.2 27.0 23.3
35.2 28.6 18.7
n.d. n.d. 0.0
22.5 5.4 76.0
n.d. n.d. 0.0
n.d. n.d. 0.0
Herbicide loss (g/ha) No VFS 0.09 VFS 0.03 Reduction (%) 66.7
0.05 0.01 80.0
0.02 0.02 0.0
2.34 0.03 98.7
0.00 0.00 0.0
1.23 0.004 99.7
0.00 0.00 0.0
0.00 0.00 0.0
17
21
28
110
60.2 30.2
18.4 13.8
87.0S 82
29.4 7.2
195.0
0.03 0.00 100.0
5.40 1.00 81.5
0.03 0.02 33.3
11.34 10.19
Herbicide concentration (lg/l) No VFS 42.5 7.3 VFS 1.2 n.d. Reduction (%) 97.2 100.0
3.1 13.0 319.3
0.7 n.d. 100.0
Herbicide loss (g/ha) No VFS 3.96 VFS 0.09 Reduction (%) 97.7
0.02 0.18 800.0
0.0004 0.00 100.0
Isoproturon DAT Rainfall mm mm/ha
Metolachlor DAT Rainfall mm mm/ha
Runoff depth (mm) No VFS 5.88 VFS 9.17 Reduction (%) 56.0
0.004 0.00 100.0
Sum
3.73 0.09
3.98 0.27
Mean
S.D.
Median
Valid number
U
p-level
1.07 0.07 91.2
1.39 0.03
0.07 0.08 14.3
13 8
38.0
0.31
416 18.4
16.8 9.7
35.2 17.6
13 8
14.5
0.005
0.47 0.01 97.9
0.87 0.01
0.04 0.01 75.0
13 8
15.5
0.007
2.84 2.55 10.2
3.24 4.44
2.72 0.51 81.2
7 7
18.5
0.438
13.4 7.1
19.6 8.3
5.2 7.1
6 4
9.5
0.593
1.00 0.07 93.0
1.98 0.09
0.01 0.05 400.0
7 7
20.0
0.549
723
Statistical results and parameters at U-test analysis are also indicated. DAT is days after treatment, S is simulate rainfall, a is maximum rainfall intensity during the event.
M. Vianello et al. / Chemosphere 61 (2005) 717–725
1
U-test results
724
M. Vianello et al. / Chemosphere 61 (2005) 717–725
were the highest in all events, due to its physico-chemical characteristics, while the reductions of metolachlor and isoproturon were lower. Herbicides weakly adsorbed on soil are transported in runoff water, but VFS changes the soil hydraulic characteristics, with an increase of infiltration in large pores and transport in the deep soil layer. VFS was particularly effective during the first runoff events after treatment, when potential transfer is maximum. A dilution mechanism is assumed comparing runoff volume and herbicide removal in isoproturon event 3 and metolachlor event 1, during 2001: in both cases runoff depth was higher with VFS than without VFS, but herbicide concentrations and consequently loads were lower with VFS (Table 4). Trends of measured concentrations also suggest that VFS had a ‘‘sponge-like’’ effect, absorbing chemicals during the first runoff events, and releasing them in the subsequent events. During 2001, this phenomenon was clear for isoproturon and metolachlor, at 89 and 28 DAT, respectively (Table 4). However, herbicide losses were at all times very low, less than 0.5% of the applied dose, confirming a low risk of runoff events in gently sloping fields. Risk assessment associated with herbicide runoff has been expressed by a risk quotient calculated as the ratio of the measured concentration to the effect concentration. The ecotoxicological endpoints for fish and algae are given in Table 2. When the quotient value is higher than 1, the herbicide concentration exceeds the toxicological endpoint and indicates a risk. None of the runoff events were dangerous for fish, but algae have been shown to be more sensitive. In particular, isoproturon and terbuthylazine, both inhibitors of photosynthetic electron transport at PSII were likely to have the highest
1.8 1.6 1.4 1.2 risk
1.0 0.8 0.6 0.4 0.2 0.0 0
1
2
3
4 5 event number
6
7
8
Fig. 2. Risk assessment of herbicides transported by runoff water, referred to algae susceptibility. Risk is a quotient calculated as measured concentration/effect concentration. Black points = ‘‘no VFS’’ treatment, white points = ‘‘VFS’’ treatment. Circle = isoproturon, rectangle = terbuthylazine, rhombus = metolachlor-2000 and triangle = metolachlor-2001.
probability of exceeding the ecotoxicological threshold. The VFS reduced the potential risk for damage to algae (Fig. 2). Moreover, a potential ecotoxicological impact was revealed only the first events after treatment, with the higher herbicide concentrations.
4. Conclusions The rainfall pattern during the monitored period did not promote runoff, so the dataset was not large enough to statistically highlight VFS effectiveness. Nevertheless, VFS reduced both runoff depth (10.2–91.2%) and herbicide losses (85.7–97.9%) in the monitored events, with percentages similar to those observed in other studies conducted under different climatic, pedological and agricultural conditions. Total herbicide losses by runoff were low (0.69– 3.98 g ha1 without VFS, less than 0.27 g ha1 with VFS). Very high concentrations were sometimes obtained, especially of isoproturon and terbuthylazine during the first rainfall events after treatment, which could have an ecotoxicological effect on algae. The VFS reduced the runoff events with herbicide concentrations higher than ecological endpoints. Low herbicide concentrations, below NOEC values, may also play an important rule in the aquatic environment: recent studies have demonstrated that low, non-significant effect concentrations of single herbicides contribute to the overall toxicity (Faust et al., 2001; Backhaus et al., 2004). In north-east Italy, farmers use a wide range of herbicides, so herbicides mixtures are usually found in aquatic environments, increasing the risk to aquatic primary producers. Different mechanisms were identified that determine the effect of VFS on surface runoff: (i) dilution, which reduces herbicide concentration in runoff water passing through the filter, and (ii) the ‘‘sponge-like’’ effect, with chemicals temporarily trapped inside the filter, then released after having been degraded. A relevant factor influencing VFS effectiveness in reducing runoff depth was grass cover: the denser the grass is, the greater is the hydraulic resistance and the lower the runoff volume compared to cropped soil. Finally, physical and chemical properties determine herbicide behaviour in the environment and consequently their interaction with the VFS. For these reasons, different climatic conditions and cropping systems could require different types of VFS to assure the best performance in reducing herbicide runoff.
Acknowledgements The research was supported by the Italian National Research Council (CNR) as part of the activities of
M. Vianello et al. / Chemosphere 61 (2005) 717–725
the Institute of Agro-Environmental and Forestry Biology, Weed Science Division of Legnaro (PD). The authors wish to thank Prof. M. Borin and Dr. F. Morari (DAAPV, University of Padova, Italy) for their suggestions regarding the organization of this research; Mr. A. Coletti (ex Centro Chimica Biochimica Fitofarmaci CNR, Perugia, Italy) for his help in laboratory analyses; and Azienda Regionale per la Prevenzione e Protezione Ambientale del Veneto for weather data.
References Backhaus, T., Faust, M., Scholze, M., Gramatica, P., Vighi, M., Grimme, L.H., 2004. Joint algal toxicity of phenylurea herbicides is equally predictable by concentration addition and independent action. Environ. Toxicol. Chem. 23, 258– 264. Benoit, P., Barriuso, E., Vidon, P., Re´al, B., 2001. Isoproturon movement and dissipation in undisturbed soil cores from a grassed buffer strip. Agronomie 20, 297–307. Blanche, S.B., Shaw, D.R., Massey, J.H., Boyette, M., Cade Smith, M., 2003. Fluometuron adsorption to vegetative filter strip components. Weed Sci. 51, 125–129. Borin, M., Bigon, E., Zanin, G., Fava, L., 2004. Performance of a narrow buffer strip in abating agricultural pollutants in the shallow subsurface water flux. Environ. Pollut. 131, 313– 321. Borin, M., Vianello, M., Morari, F., Zanin, G., 2005. Effectiveness of a buffer strip in removing runoff pollutants from a cultivated field in north-east Italy. Agr. Ecosyst. Environ. 105, 101–114. Brown, C.D., Hodgkinson, R.A., Rose, D.A., Syers, J.K., Wilcockson, S.J., 1995. Movement of pesticides to surface waters from a heavy clay soil. Pestic. Sci. 43, 131–140. Daniels, R.B., Gilliam, J.W., 1996. Sediment and chemical load reduction by grass and riparian filters. Soil Sci. Soc. Am. J. 60, 246–251. Delphin, J.-E., Chapot, J.-Y., 2001. Leaching of atrazine and deethylatrazine under a vegetated filter strip. Agronomie 21, 461–470. De Martonne, E., 1926. Une nouvelle function climatologique: lÕindice de aridite´. Me´te´or. Nouv. 2, 449–458. FAO-UNESCO, 1990. Soil map of the world. Revised Legend. FAO, Rome. Faust, M., Altenburger, R., Backhaus, T., Blanck, H., Boedeker, W., Gramatica, P., Hamer, V., Scholze, M., Vighi, M.,
725
Grimme, 2001. Predicting the joint algal toxicity of multicomponent s-triazine mixtures at low-effect concentrations of individual toxicants. Aquat. Toxicol. 56, 13–32. Gilliam, J.W., 1994. Riparian wetlands and water quality. J. Environ. Qual. 23, 896–900. Klo¨ppel, H., Ko¨rdel, W., Stein, B., 1997. Herbicide transport by surface runoff and herbicide retention in a filter strip— rainfall and runoff simulation studies. Chemosphere 35, 129–141. Lee, K.-H., Isenhart, T.M., Schultz, R.C., Mickelson, S.K., 2000. Multispecies riparian buffers trap sediment and nutrients during rainfall simulations. J. Environ. Qual. 29, 1200–1205. Misra, A.K., Baker, J.L., Mickelson, S.K., Shang, H., 1996. Contributing area and concentration effects on herbicide removal by vegetative buffer strips. T. ASAE 39, 2105–2111. Morari, F., Vianello, M., Borin, M., Zanin, G., 2001. Un misuratore–campionatore del deflusso superficiale in presenza ed assenza di fasce tampone. Riv. Irrig. Dren. 48, 24– 30 (in Italian). Ng, H.Y.F., Clegg, S.B., 1997. Atrazine and metolachlor losses in runoff events from an agricultural watershed: the importance of runoff components. Sci. Total Environ. 193, 215– 228. Nicholls, P.H., 1994. PETE—Physico-chemical evaluation: the environment—expert system for pesticide preregistration assessment. In: Proceedings of the BCPC, Pests and Diseases, Brighton, UK, pp. 1337–1342. Rankins, J.A., Shaw, D.R., 2001. Perennial grass filter strips for reducing herbicides losses. Weed Sci. 49, 647–651. Staddon, W.J., Locke, M.A., Zablotowicz, R.M., 2001. Microbiological characteristics of a vegetative buffer strip soil and degradation and sorption of metolachlor. Soil Sci. Soc. Am. J. 65, 1136–1142. StatSoft, Inc., 2004. STATISTICA (data analysis software system), version 6.1, statsoft.com. Tomlin, C.D.S., 2000. The Pesticide Manual. British Crop Protection Council. Farnham Surrey GU9 7PH, UK. Watanabe, H., Grismer, M.E., 2001. Diazinon transport through inter-row vegetative filter strips: micro-ecosystem modelling. J. Hydrol. 247, 183–199. Wauchope, R.D., 1978. The pesticide content of surface water draining from agricultural fields—a review. J. Environ. Qual. 7, 459–472. Webster, E.P., Shaw, D.R., 1996. Impact of vegetative filter strips on herbicide loss in runoff from soybean (Glycine max). Weed Sci. 44, 662–671.
