Science of the Total Environment 466–467 (2014) 841–848

Contents lists available at ScienceDirect

Science of the Total Environment

j ourna l homepage: www.e lsev ie r .com/ locate /sc i totenv

Banded applications are highly effective in minimising herbicidemigration from furrow-irrigated sugar cane

Danielle P. Oliver a,⁎, Jenny S. Anderson a, Aaron Davis b, Stephen Lewis b, Jon Brodie b, Rai Kookana a

a CSIRO Land and Water, Water for a Healthy Country National Research Flagship, PMB 2 Glen Osmond, South Australia 5064, Australiab Centre for Tropical Water & Aquatic Ecosystem Research (TropWATER), Australian Tropical Science and Innovation Precinct, Building 145, James Cook University, Townsville,Queensland 4811, Australia

H I G H L I G H T S

• Shielded sprayers able to target application of herbicides in furrow-irrigated sugarcane• Restricted application of PSII herbicides to only raised beds decreased the load moving off-site by 90%.• Low concentrations of glyphosate, an alternative herbicide for use in furrows, detected in drainage water

⁎ Corresponding author. Tel.: +61 8 83038434; fax: +E-mail addresses: danni.oliver@csiro.au (D.P. Oliver), j

(J.S. Anderson), aaron.davis@jcu.edu.au (A. Davis), stephejon.brodie@jcu.edu.au (J. Brodie), rai.kookana@csiro.au (R

0048-9697/$ – see front matter. Crown Copyright © 2013http://dx.doi.org/10.1016/j.scitotenv.2013.07.117

a b s t r a c t

a r t i c l e i n f o

Article history:Received 3 June 2013Received in revised form 22 July 2013Accepted 29 July 2013Available online xxxx

Editor: D. Barcelo

Keywords:AtrazineDiuron2,4-DGlyphosateGreat Barrier ReefBanded spraying

Runoff from farm fields is a common source of herbicide residues in surface waters in many agriculturalindustries around theworld. In Queensland, Australia, the runoff of PSII inhibitor herbicides (in particular diuronand atrazine) is amajor concern due to their potential impact on the Great Barrier Reef. This study compared theconventional practice of broadcast application of herbicides in sugarcane production across the whole field withthe banded application of particular herbicides onto raised beds only using a shielded sprayer. This study foundthat the application of twomoderately soluble herbicides, diuron and atrazine, to only the raised beds decreasedthe average total load of both herbicides moving off-site by N90% compared with the conventional treatment.This was despite the area being covered with the herbicides by the banded application being only 60% lessthan with the conventional treatment. The average total amount of atrazine in drainage water was 7.5% of theactive ingredient applied in the conventional treatment compared with 1.8% of the active ingredient applied inthe banded application treatment. Similarly, the average total amount of diuron in drainage water was 4.6% ofthat applied in the conventional treatment compared with 0.9% of that applied in the banded applicationtreatment. This study demonstrates that the application of diuron and atrazine to raised beds only is a highlyeffective way of minimising migration of these herbicides in drainage water from furrow irrigated sugarcane.

Crown Copyright © 2013 Published by Elsevier B.V. All rights reserved.

1. Introduction

Overmany years there have been increasing concerns about herbicidetransport from agricultural lands in Queensland to neighbouring waterbodies and particularly the potential impact on the Great Barrier Reef(GBR). Herbicide residues (particularly those in the photosystem II(PSII) inhibiting group) have been detected in waterways of the GBRcatchment area (Bainbridge et al., 2009; Davis et al., 2008, 2012; Ham,2007; Lewis et al., 2009; Mitchell et al., 2005; Packett et al., 2009; Smithet al., 2012; Stork et al., 2008) and in waters of the GBR lagoon(Kennedy et al., 2012a,b; Lewis et al., 2009, 2012; Shaw et al., 2010,

61 8 83038565.enny.anderson@csiro.aun.lewis@jcu.edu.au (S. Lewis),. Kookana).

Published by Elsevier B.V. All rights

2012). These concerns about pollutants from land-based activities inthe catchments (including suspended sediment and nutrients as well aspesticides) have led to the development of government policies requiringindustries in catchments draining into the GBR lagoon to meet waterquality targets through Reef Plan (Brodie et al., 2008, 2012). In addition,the use of diuron has been heavily restricted since 2012 in theGBR Catch-ment Area by the Australian Government pesticide regulatory body, theAustralian Pesticides and Veterinary Medicines Authority (APVMA)(APVMA, 2012).

The Burdekin region lies within Queensland's dry tropics and is thelocation of the largest sugarcane (Saccharum officinarum) growing regionin Australia(Davis et al., 2008) accounting for approximately 100,000 ha(approximately 30% of the total sugarcane area in the GBR catchments,Furnas, 2003; Thorburn et al., 2011). The dominant industries in theGBR catchment area since the late 19th century have been sugarcaneand beef cattle grazing. The introduction of minimum tillage coupled

reserved.

842 D.P. Oliver et al. / Science of the Total Environment 466–467 (2014) 841–848

with the expansion of the sugarcane industry have resulted in a 3 to 7fold increase in herbicide use over the last 30 to 40 years (Johnsonand Ebert, 2000). Sugarcane production in this region is also particularlyproblematic compared with other crops/regions because it is grownwith the highest application of irrigation water, via furrow irrigation,of any sugarcane producing region in Australia (Thorburn et al., 2011).Irrigation tailwater runoff from furrow irrigation has been found inother studies to contain substantial concentrations of nutrients, sedi-ment and pesticides (Davis et al., 2008, in press; Oliver and Kookana,2006a; Silburn et al., in press; Spencer et al., 1985).

The lower Burdekin sugarcane growing region is upstream ofthe Bowling Green Bay Ramsar wetland complex. The Barratta Creek,which drains much of the Burdekin River Irrigation Area (BRIA), drainsto Bowling Green Bay through the Ramsar site (Davis et al., 2012). Inthedry season almost the entireflowof Barratta Creek ismade up of tail-water flows from sugarcane andother cropping. Under these conditions,for about seven months of the year (July to January), herbicides, includ-ing atrazine and diuron, exceed Australian water quality guidelines on adaily basis (Davis et al., 2012, in press). Following its passage throughthe Ramsar site, waters from Barratta Creek and the adjacent HaughtonRiver discharge into the Great Barrier Reef World Heritage Area(GBRWHA) and subsequently into the Great Barrier Reef Marine Park(GBRMP) (Fig. 1). Thus two sets of environmentally important areasfor Australia and a national Marine Park are being impacted by pesti-cides, specifically PSII herbicides such as diuron, atrazine, ametryn andhexazinone.

Various management strategies have been suggested for minimizingoff-site pesticide migration including the use of polyacrylamide (Oliverand Kookana, 2006b), the addition of farm waste or sewage sludge tosoil (Antonious, 2011), managing the placement and applicationmethodof herbicide (Masters et al., in press; Shipitalo et al., 2008; Silburn et al., in

Fig. 1. Map of the lower Burdekin floodplain indicating the nationally significant and RATropWATER, JCU.

press), divertingwater through filter socks filledwith compostedmateri-al (Shipitalo et al., 2010), the addition of enzymes to water recycle pits toenhance pesticide degradation (Scott et al., 2010), retention ofsugarcanemulch after harvest or incorporation ofmulch into the furrows(Selim et al., 2003) or installing mechanisms for trapping sediment andattached contaminants as they move off-site (e.g. buffers or wetlands)(Gregorie et al., 2009; Rose et al., 2006; Schulz and Peall, 2001;Shipitalo et al., 2010). The Australian government has provided financialincentives to assist sugarcane farmers with the adoption of best manage-ment practices to minimise migration of herbicides off-site (Brodie et al.,2012; Masters et al., in press). One management strategy that has beensuggested is to restrict the application of the more mobile herbicides,such as atrazine and diuron, to raised beds only (banded application)and replace their application in furrows with alternative herbicides thathave relatively shorter half-lives and/or are less soluble in water. Thereare limited data available to validate and quantify the effectiveness ofthis proposed strategy for minimising migration of herbicides fromtheir source in sugarcane production, particularly under standard growerpractices. There has been one other study of the effect of bandedapplication of herbicides in sugarcane (Masters et al., in press). Theyused a rainfall simulator on small scale plots (810 m2 and 1080 m2)and studied the effect of a heavy rainfall event 1 day and 21 days afterherbicide application on herbicide transport. We are not aware of anyother published study comparing herbicide transport from broadcastand banded application practices from the first irrigation event usinggrower practices that are commonly used in the Burdekin region. Nor isthere a published study investigating the off-site movement of anychemicals that may be used in the furrows as replacements for diuronand atrazine.

The objective of this field experiment was to quantify the effective-ness of banded application of herbicides by shielded sprayers tominimise

MSAR wetlands and the experimental location. Courtesy of Maughan and Harkness,

843D.P. Oliver et al. / Science of the Total Environment 466–467 (2014) 841–848

the off-site migration of atrazine and diuron that moves with irrigationdrainagewater from furrow-irrigated sugarcane comparedwith the con-ventional practice of broadcast application of herbicides. Also the studyaimed to assess the off-site transport of herbicides applied in the furrowsas alternatives to atrazine and diuron.

Fig. 2. Shielded sprayer used to apply the herbicides in the Banded treatment. Inset showsa close-up of the shielded sprayer.

2. Materials and methods

2.1. Study area and experimental plan

The lower Burdekin region's wet-dry tropical climate exhibits strongseasonality in the form of pronounced wet and dry seasons. Highintensity summer wet season rainfall (occurring between Novemberand April) contrasts with minimal rainfall over the remainder of theyear (Dight, 2009). The region's dry-tropical monsoonal climate,coupled with low and variable rainfall (average annual rainfall approx-imately 1000 mm, Petheram et al., 2008; BOM, 2013) has seen that thesugarcane industry relies on irrigation since floodplain lagoons werefirst used to complement local rainfall in the 1880s (Kerr, 1994). Annualirrigation inputs in the region are typically around 2000 mm per crop(Charlesworth et al., 2002; Thorburn et al., 2011), and are almost entirelyapplied by furrow systems involving lay-flat surface fluming and flowcontrolling cups or syphons located at the headof paddocks. Paddock tail-water is in some cases recaptured and recycled, or otherwise leaves thefarm via drainage channels. Generally in the Burdekin, sugarcane isplanted between April and May and harvested 15 to 18 months later.The crop is then allowed to re-grow and is mechanically harvestedannually (Thorburn et al., 2011). A map of the lower Burdekin region isshown in Fig. 1.

The experiment was conducted on a commercial sugarcane farm inthe Mulgrave district of the Burdekin-Haughton Water Supply Schemeregion near Clare (Fig. 1). Water for irrigation is supplied from theBurdekin Falls Dam and this grower recycled his irrigation water bycapturing drainage water from irrigation in on-farm storage pits. Atthis site the soil texture was a medium clay throughout the profile andthe site had been laser levelled prior to raised bed formation.

The treatments were a conventional broadcast application of herbi-cides across the beds and furrows from boom sprayers (Conventional)and a banded application using a shielded sprayer (Banded) wherecertain herbicides were applied to furrows only and other herbicides toraised beds only. Details of the herbicides applied are given in Table 1.The herbicides studied were chosen because they are commonly usedin sugarcane production in the Burdekin region. Glyphosatewas includedsince it is being considered as an alternative herbicide to diuron andatrazine for use in the furrows. Two herbicide treatments were appliedin duplicate on November 24, 2012 to designated blocks. In this studyeach block consisted of 6 furrows and 6 raised beds (910 mlong × 9.12 mwide), giving a total treatment area of 0.83 ha. The shields

Table 1Details of herbicide application rates and amount of active ingredient (a.i.) applied for each tre

Treatment Product Applicat-ionrate(kg or L ha−1)

Acting(a.i

Conventional Applied Atradex 1 AtrAcross Diurex 1 DiuWhole Amicide 700 1 2,4Area Gramoxone 1.3 Par

Banded Furrow Round-up attack 1.5 GlyFurrow Amicide 700 1 2,4Bed Atradex 1 AtrBed Grammoxone 1.3 ParBed Amicide 700 1 2,4Bed Diurex 1 Diu

were 900 mmwide and the over top sprays were 620 mmwide. Photosof the field application are shown in Fig. 2.

Usually water is applied to the rows for approximately 12 h withexcess water either running off the end of the field into drains and fromthere into nearby creeks and rivers. On-farm storage pits to collect runoffare common in the region, but the proportion of runoff that can be storedvarieswith pit size (Thorburn et al., 2011). The bayswere irrigated simul-taneously commencing on the evening of November 26, 2012 and waterflowed out of the furrows on November 27, 2012. Each end of the drain-age channel for each baywas blocked off so thewater draining from eachbay was isolated and constrained to flow out through a pipe. Watersamples were collected as soon as flow started (Time 0), then at thefollowing time intervals: 0.5, 1, 2, 4, 6, 8, 10 and 12 h thereafter.

2.2. Monitoring flow discharge

Water flow was monitored from bays 1 and 3 using a pre-calibratedV-notch weir crate designed to cope with low flow discharges (USDepartment of Interior/US Department of Agriculture, 2001). The depthof water in the crate was regularly recorded by reading the water levelheight and thiswas related toflow rate fromapreviously determined dis-charge rating curve. In the other two bays water flow was monitoredmanually by timing how long it took to fill a 20 L bucket at varioustimes throughout the flow event. This technique was also used in bays 1and 3 and allowed comparison between the data from the V-notch weircrates with that from the bucket technique. Generally good agreementwas obtained between the twomethods of measuring flow but occasion-ally it was necessary to predict the flow data by extrapolation processes.

atment.

iveredient.)

Fraction ofa.i.

Amount of a.i.applied(kg or L ha−1)

Areasprayed(ha)

a.i.applied(kg or L)

azine 900 g kg−1 0.9 0.83 0.747ron 900 g kg−1 0.9 0.83 0.747-D 700 g L−1 0.7 0.83 0.58aquat 250 g L−1 0.325 0.83 0.270phosate 570 g L−1 0.855 0.49 0.419-D 700 g L−1 0.7 0.49 0.343azine 900 g kg−1 0.9 0.34 0.306aquat 250 g L−1 0.325 0.34 0.111-D 700 g L−1 0.7 0.34 0.238ron 900 g kg−1 0.9 0.34 0.306

Table 2Total volume of water that moved off-site from each treatment bay.

Treatment Bay Total volume of water(kL)

Area of bay(ha)

Volume per area(kL ha−1)

Conventional 1 150 0.83 180.7Conventional 2 123 0.83 148.2Banded 3 155 0.83 186.7Banded 4 127 0.83 153.0

844 D.P. Oliver et al. / Science of the Total Environment 466–467 (2014) 841–848

2.3. Processing water samples

Water samples were collected in pre-cleaned Nalgene bottles andstored on ice until transported to the laboratory the following day. Sam-ples were then stored at 4 °C until they were partially processed in anearby laboratory by loading onto solid phase extraction (SPE) cartridges.The cartridges were then transported to the analytical laboratory forextraction and herbicide analysis. In addition, three blank samples(50 mL high purity MilliQ water) were spiked individually with aknown concentration of each herbicide (50 μL × 200 mg L−1 stocksolution) and taken to the field to check on any herbicide losses duringstorage and transport to the laboratory. Irrigation water was collectedjust before it was applied to the field and analysed for herbicide residuesand to ensure there was no phosphorus present that could interfere withglyphosate analyses on LC-ICP-MS. The spiked sampleswere treated in anidentical manner to the field samples. The SPE cartridges were thentransported to the analytical laboratory on ice bricks and eluted usingsolvent followed by herbicide analysis.

2.4. Herbicide analysis

2.4.1. Glyphosate and AMPAThe sample (100 μL) was injected onto an Agilent 7700 ICP-MS

connected to an Agilent 1260 Infinity HPLC tuned to detect mass 31(P) using Helium as the collision gas. The column was a HamiltonPRP-X100 (250 mm × 4.1 10 μm). The flow rate was 0.8 mL min−1.The outlet of the separation column was connected to the concentricnebulizer of an Agilent 7700 ICP-MS by use of a 50 cm length of PEEKtubing. The conditions used for LC-ICP-MS were RF power 1550 W,plasma gas (Ar) flow 1.09 L min−1, sampling depth 8 mm andintegration time 0.1 s. The mobile phase was 20 mM citric acid.Retention times were glyphosate 9.6 min and AMPA 2.7 min. Thechemicals were separated from P on the HPLC column prior to injectiononto the ICP-MS. The analytical detection limit on the LC-ICP-MS was0.3 mg L−1 which equated to 0.015 mg L−1 in the field-collectedwater samples prior to concentration on the SPE cartridges.

2.4.2. Atrazine, DIA, DEA, 2,4-D and diuronThe sample (20 μL)was injected onto theAgilent 1100 LC. The column

was an Alltech Altima C 18 (250 mm × 4.6 mm). The flow rate was1.0 mL min−1. An isocratic mobile phase was used consisting of 65%acetonitrile:35% 0.2% phosphoric acid. Atrazine, DIA, DEA and 2,4-Dwere measured at 230 nm and diuron was measured at 254 nm. Reten-tion times were: DIA 5.65 min, DEA 4.2 min, atrazine 7.6 min, diuron8.35 min and 2,4-D 9.0 min. The analytical detection limits on the HPLCwere 0.2 mg L−1 for atrazine, DEA, 2,4-D and diuron and 0.8 mg L−1

for DIA which equated to 0.01 mg L−1 for atrazine, DEA, and diuron,0.04 mg L−1 for DIA and 0.005 mg L−1 for 2,4-D in the field-collectedwater samples prior to concentration on the SPE cartridges. Althoughparaquat was applied, it was not analysed in the water samples.

2.5. Quality assurance/quality control

An initial check on selected herbicide (atrazine, DEA, DIA, diuron and2,4-D) concentrations in the water samples was made by analysing allwater samples directly on the HPLC (with no processing through SPEcartridges) using the method for analysing 2,4-D as outlined above.Regression of the data obtained by direct analysis of water samples onthe HPLC and by the SPE technique gave excellent agreement for atrazine(R2 = 0.98) and diuron (R2 = 0.90) and reasonable agreement for 2,4-D(R2 = 0.68).

Three samples (MQ water) were spiked in the field with knownconcentrations of atrazine, DEA, DIA, diuron, 2,4-D and glyphosateand treated exactly as the experimental water samples. The expectedconcentration after processing the field-spiked water samplethrough the SPE cartridge was 2 mg L−1 for atrazine, DEA, DIA,

diuron and 2,4-D and 5 mg L−1 for glyphosate. The average recoveriesof these field spikes were 98% for atrazine, 75% for DEA, 93% for DIA,99% for diuron, 89% for 2,4-D and 112% for glyphosate. To check on therecoveries of the herbicides after processing through the SPE cartridgessix water samples (three MQ water and three irrigation water collectedfrom the head channel) were also spiked individually with atrazine,DEA, DIA, diuron, 2,4-D (all 2 mg L−1) and glyphosate (5 mg L−1) andstored overnight at 4 °C before processing through the SPE cartridge.The average recoveries for the laboratory spiked samples (as assessmentof recovery through the SPE processing)were 107% for atrazine, 78.6% forDEA, 97.9% for DIA, 111.1% for diuron, 83.7% for 2,4-D and 120.2% forglyphosate. All concentration data were corrected for the respectedrecoveries from processing through the SPE cartridges.

2.6. Calculation of pesticide load

Amounts (mg) of pesticides leaving each irrigation bay for the desig-nated time interval were determined by multiplying concentrations(mg L−1) by the total volume (L) of water leaving the irrigation bayduring that time interval. The total volume of runoff water leaving thefield for a designated time increment was determined by calculatingthe area under the hydrograph. Since there were only two replicatesfor each treatment the data were firstly assessed for equality ofvariances between the two treatments and then statistical differencesbetween the two groups were determined using the appropriate t-testin Excel i.e. assuming equal or non-equal variances.

3. Results and discussion

3.1. Flow measurements

The total volume of water leaving each irrigation bay is given inTable 2 and ranged from 123 to 155 kL which equated to 148 to187 kL ha−1. The measurement of flow using the calibrated V-notchweir crates and the manual bucket measurements agreed quite well forthe two bays (Bays 1 and 3) when both techniques were used. An exam-ple of theflowdata is shown for Bay 1 in Fig. 3. This good agreement gaveconfidence in the flowdata generated for Bays 2 and 4where themanualbucket measurements were used.

3.2. Concentrations of herbicides

Atrazine, DEA, DIA, diuron, glyphosate or AMPA was not detected inthe irrigation water applied to the field. Low concentrations of 2,4-D(0.016 mg L−1) however were detected and this was deducted fromthe concentrations measured in drainage water leaving the field. Theconcentrations of atrazine, diuron and 2,4-D in water were highest inthe first 2–3 samples but generally decreased exponentially over time.A similar result was found by Masters et al. (in press) with rainfall sim-ulation studies from sugarcane fields. The range of concentrations inwater from the Conventional bays in this study was 0.02–1.31 mg L−1

for atrazine, 0.07–0.60 mg L−1 for diuron and 0.10–1.59 mg L−1 for2,4-D with average concentrations for the irrigation event being0.52 mg L−1, 0.31 mg L−1 and 0.50 mg L−1 for atrazine, diuron and2,4-D, respectively (Fig. 4A, C and D). The ANZECC/ARMCANZ (2000)trigger value for protection of 99% of freshwater species is 0.7 and

845D.P. Oliver et al. / Science of the Total Environment 466–467 (2014) 841–848

140 μg L−1 for atrazine and 2,4-D, respectively. There is no trigger valuelisted for diuron. The average concentrations measured leaving theirrigation field were 700 and 3.5 times the ANZECC/ARMCANZ (2000)trigger value for atrazine and 2,4-D, respectively. This indicates thateven after taking into consideration that these concentrations arehigher than expected to be found in rivers since they are edge-of-fieldmeasurements these herbicides are highly likely to pose a significantadverse effect on freshwater aquatic ecosystems.

The range of concentrations in water from the Shielded sprayer bayswas 0.01–0.12 mg L−1 for atrazine, b0.01–0.06 mg L−1 for diuron and0.10–0.6 mg L−1 for 2,4-D with average concentrations for the irrigationevent being 0.04 mg L−1, 0.02 mg L−1 and 0.35 mg L−1 for atrazine,diuron and 2,4-D, respectively (Fig. 4A, C and D). The concentrations ofatrazine and diuron (Fig. 4A and C) were statistically (P b 0.001) greaterin water samples from Conventional bays than from Banded bays.However, there was no significant difference between treatments in theconcentrations of 2,4-D in water (Fig. 4D), which reflects that there waslittle difference in the amount of 2,4-D applied in both treatments(Table 1). Davis et al. (in press) found that the maximum concentrationsin irrigation water lost from furrow-irrigated sugarcane fields were0.78 mg L−1 for atrazine, 0.12 mg L−1 for diuron and 0.28 mg L−1 for2,4-D.

Studies with rainfall simulators in the adjacent MackayWhitsundayregion also found a 54% and 56% decrease in the eventmean concentra-tion of atrazine and diuron, respectively frombanded applications of theherbicides compared with broadcast applications (Masters et al., inpress), whichwas attributed to the decrease in area sprayed by approx-imately 50%. Other rainfall simulation studies of the banded applicationof pesticides in cotton found that the total concentrations of pyrithiobacsodium, metolachlor and diuron in runoff were decreased by 38, 22 and50%, respectively (Silburn et al., in press).

Although diuron and atrazine have similar solubilities in water(35 and 35.6 mg L−1, respectively) (Footprint Pesticides PropertiesDatabase, 2013) and were applied at the same rate of 1 kg ha−1 inboth treatments, the atrazine concentrations in drainage water werealmost double those of thediuron concentrations. Thismost likely reflectsthe differences in sorption behaviour of the two herbicides since thesorption coefficient (Koc) of atrazine is only 100 L kg−1 (range of 89 to513 L kg−1) while diuron is 813 L kg−1 (range of 498 to 1358 L kg−1)(Footprint Pesticides Properties Database, 2013). The less strongly sorbedof the two herbicides, atrazine, was detected in much higher concentra-tions than diuron. Herbicide runoff partitioning studies on the lowerBurdekin floodplain have similarly highlighted the relative propensityfor diuron to sorb to particulate matter (Davis et al., 2012).

0

1

2

3

4

5

6

7

4:48 7:12 9:36 12:00 14:24 16:48 19:12

Flo

w L

/sec

Time

manual bucketmeasurements

V-notch weirmeasurements

Fig. 3. Example of the flow data from Bay 1 derived from the V-notch weir crate that hadbeen calibrated prior to the field experiment and from the manual measurements madeusing volumes of water collected in the bucket at designated times throughout theirrigation event.

The atrazinemetabolite, DEA, was only detected at very low concen-trations in water from Conventional bays (Fig. 4D). Neither atrazinemetabolite, DEA or DIA, was detected in water from Banded bayswhere atrazine had been applied to the raised bed only. Studies ofcontaminants in drainage water from furrow irrigation in sugarcane inthe Ord Irrigation Area (north-west Australia) also found no DIA andvery low concentrations of DEA in drainage water (Oliver, unpublisheddata). Although glyphosate was only applied in the furrows in theBanded treatments, the concentrations of glyphosate and the mainmetabolite of glyphosate, AMPA, were below detection limits(0.015 mg L−1) in all water samples (data not shown). The triggervalue for glyphosate for protection for 99% of species in freshwateris 370 μg L−1 (ANZECC/ARMCANZ, 2000), which is approximately24 times the detection limit in the water samples in this study(15 μg L−1). This indicates that the strongly sorbed glyphosate(Koc 1435 L kg−1, Footprint Pesticides Properties Database., 2013)is a good alternative herbicide to diuron and atrazine for use in thefurrows.

Since samples were collected directly at edge of field it would beexpected that the concentrations would be higher than those foundduring monitoring of river waters or of inshore waters of the GBR.Davis et al. (2012, in press) noted an order of magnitude or moreincrease in concentrations of pesticides in water collected at paddockscale compared with those from the creek systems. Recent sub-catchment scale water quality monitoring programmes on the lowerBurdekin floodplain have documented concentrations of atrazine upto 27 μg L−1, and diuron up to 8.5 μg L−1 in Barratta Creek sites duringtailwater dominated low flow (dry season) conditions (Davis et al., inpress). Moreover, dry season concentration values for herbicides suchas atrazine and diuron are consistently above relevant ecosystemhealthguidelines. Atrazine concentrations in low flow samples were above thetrigger value for protection of 99% of freshwater species (0.7 μg L−1,ANZECC/ARMCANZ, 2000) in 91%, 73% and 89% of samples at Upper,East and West Barratta Creek sites, respectively. Diuron concentrationsin low flow samples were above the ‘low reliability’ trigger value forprotection of 99% of freshwater species (0.2 μg L−1, ANZECC/ARMCANZ,2000) in 91%, 73% and 89% of samples at Upper, East and West BarrattaCreek sites, respectively (Davis et al., in press).

3.3. Loads of herbicides transported off-site

There was no significant difference in the average total load of 2,4-Dmoving off-site in drainage water from the Banded bays (46.98 g or56.60 g ha−1) compared with the Conventional bays (42.10 g or50.72 g ha−1) (Table 3). The amount of 2,4-D lost in drainage wateras a fraction of the active ingredient applied in this experiment was7.3% and 8.0% from the Conventional and Banded bays, respectively.This is due to the same amount of 2,4-D being applied to both treat-ments even though in the Conventional bays itwas applied by broadcastapplication and in the Banded bays it was applied in two passes by ashielded sprayer. This indicates that placement of 2,4-D in furrows,and not the method of application, affected the amount transportedoff-site.

The load of atrazine, diuron and 2,4-D followed the trend of the flowdata, increasing to a maximum of approximately 4–5 h after flowstarted (Fig. 5A, C and D). The load of DEA however, generally increasedin a linear trend for the duration of the irrigation event, albeit fromConventional bays only (Fig. 5B), which most likely reflected thegradual transformation of atrazine over time.

There was a 90% and 92% decrease in the average load of atrazineand diuron, respectively, moving off-site in water from Banded bayscompared with Conventional bays (Table 3). However, atrazine anddiuron were applied only to raised beds in the Banded bays withglyphosate and 2,4-D applied to the furrows, while in the Conventionalbays the herbicide treatmentswere applied across thewhole area. Thus,less atrazine and diuron were applied to the total area of the Banded

0

0.2

0.4

0.6

0.8

1

1.2

1.4

Con

cent

ratio

n m

g/L

Time (h) after flow started

Conventional Rep 1

Conventional Rep 2

Banded Rep 1

Banded Rep 2

a

0

0.005

0.01

0.015

0.02

0.025

Con

cent

ratio

n m

g/L

Time (h) after flow started

Conventional Rep 1

Conventional Rep 2b

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Con

cent

ratio

n m

g/L

Time (h) after flow started

Conventional Rep 1

Conventional Rep 2

Banded Rep 1

Banded Rep 2

c

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

0 2 4 6 8 10 12 0 2 4 6 8 10 12

0 2 4 6 8 10 12 0 2 4 6 8 10 12

Con

cent

ratio

n m

g/L

Time (h) after flow started

Conventional Rep 1

Conventional Rep 2

Banded Rep 1

Banded Rep 2

d

Fig. 4. Concentration of a) atrazine, b) DEA, c) diuron and d) 2,4-D in discharge water for the four irrigation bays over the irrigation event.

846 D.P. Oliver et al. / Science of the Total Environment 466–467 (2014) 841–848

bays compared with the Conventional bays. For both diuron and atrazine747 g (900 g ha−1) was applied to the Conventional bays comparedwith 306 g (370 g ha−1) to the Banded bays (Table 3). When the differ-ences in amount applied to the two treatments were taken into account,therewas still a significant (P b 0.001) decrease in the amount of atrazineand diuron moving off-site from Banded bays. The average amount ofatrazinemoving off-site as a fraction of the amount applied in this exper-iment was 7.5% from the Conventional bays compared with 1.8% fromBanded bays. Including the load of the atrazine metabolites moving indrainage water had a marginal effect on the total load deriving from theoriginal parent compound in drainage water from the Conventionalbays where 58.87 g (7.9%) of total atrazine and metabolites was lost insurface drainage water under this treatment.

Similarly, the average amount of diuron moving off-site as a fractionof the amount applied in this experiment was 4.6% from theConventional bays compared with 0.9% from the Banded bays. Theseresults indicate that applying atrazine and diuron to the raised bedsonly (Banded bays), rather than applying across the whole field(Conventional), significantly (P b 0.001) decreased the load of atrazine

Table 3Total average (±standard deviation) load (g) and (g/ha) of herbicides in drainage waterfor irrigation event for the two treatments and the fraction migrating off-site as % ofherbicide applied.

Treatment Atrazine Diuron 2,4-D

Load of herbicide applied (g) to each treatmentConventional 747 747 580Banded 306 306 581

Average load (g) in drainage waterConventional 56.09 (±6.53) 34.64 (±4.64) 42.10 (±14.43)Banded 5.46 (±1.70) 2.85 (±0.87) 46.98 (±21.22)

Average load (g ha−1) in drainage waterConventional 67.58 (±7.87) 41.73 (±5.58) 50.72 (±17.38)Banded 6.57 (±2.05) 3.44 (±1.04) 56.60 (±25.57)% difference between treatments 90% 92% 10%

Average load in irrigation water as % herbicide applied to treatment bayConventional 7.5% 4.6% 7.25%Banded 1.8% 0.9% 8.0%

and diuron moving off-site. This result is due to the delay in water“wetting up” the raised beds resulting in a lack of or a decrease incontact between the herbicide on the raised bed and the irrigationwater draining from the irrigated field through the furrows. Ultimatelythis resulted in a decrease in the herbicidesmoving off-site in dischargingirrigationwater from the Banded bays. By contrast, when herbicideswithhigher water solubilities and lower sorption coefficients are placed in thefurrows they are more likely to move off-site due to the direct contactwith the water travelling through the furrows. Davis et al. (in press)found the amount of herbicide lost during an irrigation event soon(2–6 days) after application was 0.27–4.8% for atrazine, 0.03–1.9% fordiuron and 0.48–3.4% for 2,4-D. A rainfall simulator study found thatthe loads of herbicide lost (as percent of applied active ingredient)1 day after application were 9.8% for hexazinone, 7.9% for atrazine, 6.1%for diuron and 5.0% for ametryn (Masters et al., in press).

Furthermore, although glyphosate was applied in the furrows in theBanded bays the glyphosate concentrations in the drainagewater in thisstudywere just below detection limit, suggesting that this is potentiallya good alternative herbicide to replace atrazine and diuron in thefurrows. Both the Banded and Conventional treatments provided goodweed control suggesting that both positive agronomic as well asenvironmental outcomes should be possible from adoption of precisionpesticide application technologies by the industry.

3.4. Implications of the use of banded spraying for herbicide concentrationsin Barratta Creek

Given that the high and unacceptable concentrations of diuron andatrazine found in Barratta Creek derive from herbicide use in sugarcanecultivation in the catchment of the creek (Davis et al., 2012, in press) wecan model likely improvements of the condition of Barratta Creek andthe Ramsar wetland if the shielded sprayer methodology describedabove was applied widely in the catchment. The herbicides atrazineand diuron are quite water soluble and have long half-lives, hencethey move easily by surface flow from the field to the nearest streamthrough the drainage network. Previous research on off-site herbicidelosses under conventional practices on the lower Burdekin floodplainfound that approximately 80–90% of the total annual loss occurs in the

0

5000

10000

15000

20000

25000

Load

mg

Time (h) after flow started

Conventional Rep 1

Conventional Rep 2

Banded Rep 2

Banded Rep 2

a

0

100

200

300

400

500

600

700

800

Lo

ad m

g

Time (h) after flow started

Conventional Rep 1

Conventional Rep 2b

0

2000

4000

6000

8000

10000

12000

14000

Load

mg

Time (h) after flow started

Conventional Rep 1

Conventional Rep 2

Banded Rep 1

Banded Rep 2

c

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

0 5 10 15 0 5 10 15

0 5 10 15 0 5 10 15

Lo

ad m

g

Time (h) after flow started

Conventional Rep 1

Conventional Rep 2

Banded Rep 1

Banded Rep 2

d

Fig. 5. Load of a) atrazine, b) DEA, c) diuron and d) 2,4-D in discharge water for the four irrigation bays over the irrigation event.

847D.P. Oliver et al. / Science of the Total Environment 466–467 (2014) 841–848

first 1–2 irrigations following herbicide application (Davis et al., in press).While there will certainly be additional losses of herbicides following thefirst irrigation frombanded applications in follow-up irrigations or subse-quent rainfall events, the marked decreases in losses documented in thisstudy could translate to marked improvements in off-site water quality,particularly during the dry-season. Given current concentrations of diu-ron and atrazine in Barratta Creek average 1.9 μg L−1 and 5.3 μg L−1,respectively, throughout the year (Davis et al., in press) and are abovewater quality guidelines, a reduction of 90% would theoretically bringthese down to 0.19 μg L−1 and 0.5 μg L−1, which, on average would bebelow available guidelines. This study suggests that the adoption ofbanded herbicide application technologies could remedy, in large part,one of the major water quality issues for the lower Burdekin floodplainand the associated Bowling Green Bay RAMSAR site.

4. Conclusions

This field experiment demonstrated clearly that banded applicationof highly soluble herbicides (atrazine and diuron) to raised beds is anefficient management practice for minimising off-site movement ofherbicides from furrow-irrigated sugarcane compared with the currentpractice of broadcast application. This study found that the bandedapplication of atrazine and diuron decreased the average total loadmoving off-site in drainagewater by N90% comparedwith conventionalapplication treatments. Even after allowing for the difference betweenthe two treatments in area sprayed with the atrazine and diuron therewas still a highly significant (P b 0.001) decrease in the average totalload of the two herbicides in drainage water between the Banded andthe Conventional treatments. Additional monitoring of these technolo-gies across different soil types, weed species and possibly with differentcombinations of chemicals would provide valuable additional testing ofthe approach from an industry perspective. In addition, this study foundthat glyphosate concentrations in drainage water from the Bandedtreatment bays were below the detection limit suggesting that itwould be a good alternative herbicide to atrazine and diuron for usein furrows. However, other studies have found detectable concentra-tions of glyphosate in tailwater draining from furrow-irrigated sugar-cane (Davis, unpublished data), which would suggest that furtherwork on different soil types is required before glyphosate can be fullyendorsed as an alternative to diuron and atrazine in the furrows.

Acknowledgments

This work was financially supported by Australian Federal Govern-ment through Reef Rescue Research and Development funding fromCaring for Our Country Initiative. The field experiment was madepossible by the generosity of growers. The assistance of DominiqueO'Brien during the field experiment was gratefully appreciated.

References

Antonious GF. Mitigation of dimethazone residues in soil and runoff water fromagricultural field. J Environ Sci Health B 2011;46:76–83.

ANZECC, ARMCANZ. Australian and New Zealand Guidelines for Fresh and MarineWater Quality. Australian and New Zealand Environmental and ConservationCouncil, Agriculture and Resource Management Council of Australia andNew Zealand: Canberra. http://www.environment.gov.au/water/publications/quality/nwqms-guidelines-4-vol2.html, 2000. [accessed Feb 2013].

APVMA. Review of diuron use in Australia. http://www.apvma.gov.au/products/review/completed/diuron_labels.php, 2012. [accessed Feb 2013].

Bainbridge ZT, Brodie JE, Faithful JW, Sydes DA, Lewis SE. Identifying the land-basedsources of suspended sediments, nutrients and pesticides discharged to the GreatBarrier Reef from the Tully Basin, Queensland, Australia. Mar Freshw Res 2009;60:1081–90.

Brodie J, Binney J, Fabricus K, Gordon I, Hoegh-Guldberg O, Hunter H, et al. Synthesisof evidence to support the Scientific Consensus Statement on Water Qualityin the Great Barrier Reef. Brisbane: The State of Queensland (Departmentof the Premier and Cabinet), Reef Water Quality Protection Plan Secretariat;2008 [64 pp. http://www.reefplan.qld.gov.au/resources/assets/reef-consensus.pdf (accessed Feb 2013)].

Brodie J, Kroon F, Schaffelke B, Wolanski E, Lewis S, Devlin M, et al. Terrestrial pollutantrunoff to the Great Barrier Reef: an update on issues, priorities and managementresponses. Mar Pollut Bull 2012;65:81–100.

Bureau of Meterology (BOM). Monthly climate statistics 2013. http://www.bom.gov.au/climate/data/index.shtml?bookmark=200, 2013. [accessed Feb 2013].

Charlesworth PB, Bristow KL, Chinn C, Ham G. Healthy crop and healthy groundwater —sugarcane in the Burdekin Delta. Contributed paper to the IAA Conference, 20 to 24May 2002, Sydney, Australia; 2002.

Davis A, Lewis S, Bainbridge Z, Brodie J, Shannon E. Pesticide residues in waterways of thelower Burdekin region: challenge in ecotoxicological interpretation of monitoringdata. Aust J Ecotoxicol 2008;14:89–108.

Davis AM, Lewis SE, Bainbridge ZT, Glendenning L, Turner R, Brodie JE. Dynamics ofherbicide transport and partitioning under event flow conditions in the lowerBurdekin region, Australia. Mar Pollut Bull 2012;65:182–93.

Davis AM, Thorburn PJ, Lewis SE, Bainbridge ZT, Attard SJ, Milla R, et al. Environmentalimpacts of irrigated sugarcane production: herbicide run-off dynamics from farmsand associated drainage systems. Agric Ecosyst Environ 2013. [in press].

Dight I. Burdekin Water Quality Improvement Plan. Townsville: NQ Dry Tropics; 2009.Footprint Pesticide Properties Database. http://sitem.herts.ac.uk/aeru/iupac/index.htm,

2013. [accessed Feb, 2013].

848 D.P. Oliver et al. / Science of the Total Environment 466–467 (2014) 841–848

Furnas MJ. Catchments and corals: terrestrial runoff to the Great Barrier Reef. Townsville:Australian Institute of Marine Science and CRC Reef; 2003.

Gregorie C, Elsaesser D, Huguenot D, Lange J, Lebeau T, Merli A, et al. Mitigation ofagricultural non-point source pesticide pollution in artificial wetland ecosystems.Environ Chem Lett 2009;7:205–31.

Ham G. Water quality of the inflows/outflows of the Barratta Creek system. Proc Aust SocSugar Cane Technol 2007;29:149–66.

Johnson AKL, Ebert SP. Quantifying inputs of pesticides to the Great Barrier Marine Park—

a case study in the Herbert River catchment of north-east Queensland. Mar Pollut Bull2000;41:302–9.

Kennedy K, Schroeder T, ShawM, Haynes D, Lewis S, Bentley C, et al. Long termmonitoringof photosystem II herbicides — correlation with remotely sensed freshwater extent tomonitor changes in the quality of water entering the Great Barrier Reef, Australia.Mar Pollut Bull 2012a;65:292–305.

Kennedy K, Devlin M, Bentley C, Lee-Chue K, Paxman C, Carter S, et al. The influence of aseason of extreme wet weather events on exposure of theWorld Heritage Area GreatBarrier Reef to pesticides. Mar Pollut Bull 2012b;64:1495–507.

Kerr JD. Black snow and liquid gold: a history of the Burdekin Shire. Ayr, Queensland:Burdekin Shire Council; 1994.

Lewis SE, Brodie JE, Bainbridge ZT, Rohde KW, Davis AM, Masters BL, et al. Herbicides: anew threat to the Great Barrier Reef. Environ Pollut 2009;157:2470–84.

Lewis SE, Schaffelke B, Shaw M, Bainbridge ZT, Rohde KW, Kennedy KE, et al. Assessingthe additive risks of PSII herbicide exposure to the Great Barrier Reef. Mar PollutBull 2012;65:280–91.

Masters B, Rohde K, Gurner N, Reid D. Reducing the risk of herbicide runoff in sugarcanefarming through controlled traffic and early-banded application. Agric Ecosyst Environ2013. [in press].

Mitchell C, Brodie J, White I. Sediments, nutrients and pesticide residues in event flowconditions in streams of the Mackay Whitsunday Region, Australia. Mar Pollut Bull2005;51:23–36.

Oliver DP, Kookana RS. On-farmmanagement practices to minimise off-site movement ofpesticides from furrow irrigation. Pest Manag Sci 2006a;62:899–911.

Oliver DP, Kookana RS. Minimising off-site movement of contaminants in furrow irrigationusing polyacrylamide (PAM) I. Pesticides. Aust J Soil Res 2006b;44:551–60.

Packett R, Dougall C, Rohde K, Noble R. Agricultural lands are hot-spots for annual runoffpolluting the southern Great Barrier Reef lagoon. Mar Pollut Bull 2009;58:976–85.

Petheram C, Bristow KL, Nelson PN. Understanding andmanaging groundwater and salinityin a tropical conjunctive water use irrigation district. Agric Water Manage 2008;95:1167–79.

Rose MT, Sanchez-Bayo F, Crossan AN, Kennedy IR. Pesticide removal from cotton farmtailwater by a pilot-scale ponded wetland. Chemosphere 2006;63:1849–58.

Schulz R, Peall SKC. Effectiveness of a constructed wetland for retention of non-pointsource pesticide pollution in the Lourens River catchment, South Africa. Environ SciTechnol 2001;35:422–6.

Scott C, Lewis SE, Milla R, Taylor MC, Rogers A, Dumsday G, et al. A free-enzyme catalystfor the bioremediation of environmental triazine contamination. J Environ Manage2010;91:2075–8.

SelimHM, Zhou L, Zhu H. Herbicide retention in soil as affected by sugarcanemulch residue.J Environ Qual 2003;32:1445–54.

Shaw M, Furnas MJ, Fabricius K, Haynes D, Carter S, Eaglesham G, et al. Monitoringpesticides in the Great Barrier Reef. Mar Pollut Bull 2010;60:113–22.

Shaw CM, Brodie J, Mueller JF. Phytotoxicity induced in isolated zooxanthellae byherbicides extracted from Great Barrier Reef flood waters. Mar Pollut Bull 2012;65:355–62.

Shipitalo MJ, Malone RW, Owens LB. Impact of glyphosate-tolerant soybean andglufinate-tolerant corn production on herbicide losses in surface runoff. J EnvironQual 2008;32:1445–54.

Shipitalo MJ, Bonta JV, Dayton EA, Owens LB. Impact of grassed waterways and compostfilter socks on the quality of surface runoff from corn fields. J Environ Qual 2010;39:1009–18.

Silburn DM, Foley JL, deVoil RC. Managing runoff of herbicides under rainfall and furrowirrigation with wheel traffic and banded spraying. Agric Ecosyst Environ 2013. [inpress].

Smith R, Middlebrokk R, Turner R, Huggins R, Vardy S, Warne M. Large-scale pesticidemonitoring across Great Barrier Reef catchments — Paddock to Reef IntegratedModelling and Reporting Program. Mar Pollut Bull 2012;65:117–27.

Spencer WF, Dliath MM, Blair J, LeMast RA. Transport of pesticides from irrigated fields insurface runoff and tile drain water. USDA-ARS conservation research report 31.Washington DC: US Government Printing Office; 1985.

Stork PR, Bennett FR, Bell MJ. The environmental fate of diuron under a conventionalproduction regime in a sugarcane farm during the plant cane phase. Pest Manag Sci2008;64:954–63.

Thorburn PJ, Biggs JS, Attard SJ, Kemei J. Environmental impacts of irrigated sugarcaneproduction: nitrogen lost through runoff and leaching. Agric Ecosyst Environ2011;144:1–12.

US Department of Interior/US Department of Agriculture. The water measurementmanual. Washington, DC: U.S. Government Printing Office 20402; 2001 [www.usbr.gov/pmts/hydraulics_lab/pubs/wmm/ (accessed October 2012)].